Work up to rule out perioperative myocardial infarction: is it...

Document technical information

Format pdf
Size 2.1 MB
First found Jun 9, 2017

Document content analysis

not defined
no text concepts found


James Joyce
James Joyce

wikipedia, lookup

Jack Waldman
Jack Waldman

wikipedia, lookup



Work up to rule out perioperative myocardial infarction: is it overused?
SK Appavu, TR Haley, A Khorasani and SR Patel
Departments of Surgery and Anesthesiology, Cook County Hospital and the Department of Surgery, University of Illinois College
of Medicine, Chicago, Illinois, USA
The awareness of the diagnostic difficulty and the documented high mortality risk of perioperative myocardial
infarction (PMI) has led to the wide use of work up to
rule out PMI after major noncardiac operations. This has
caused stable postoperative patients to be kept in monitored hospital beds for extended periods of time and to
be subjected to additional tests. We hypothesized that
the mortality of PMI is high and, therefore, the wide use
of postoperative work up to identify these patients is justifiable. We performed the following study to prove our
hypothesis. All patients in the recovery room after major
noncardiac operations who underwent work up to rule
out PMI were identified and followed. The PMI work up
included care in an electronically monitored unit, physical
assessment, continuous ECG monitoring, and three 12lead electrocardiograms and cardiac enzymes obtained
at six to eight hour intervals. Data collection included
patient demographics; preoperative cardiac risk factors;
incidence of intraoperative hypotension, hemorrhage and
ECG changes; type of anesthesia and operative procedures and their durations; postoperative ECG and
cardiac enzyme results; the incidence of PMI and patient
Two hundred patients were studied; 85 males and 115
females. Their mean age was 62.9 years. Preexisting conditions included hypertension in 162 patients, peripheral
arterial disease in 102, diabetes mellitus in 97, angina in
30, previous myocardial infarction in 41, and smoking in
107. Of 200 patients, 164 had an abnormal preoperative
ECG. Vascular operations were performed in 104
patients, nonvascular abdominal operations in 48, and
other operations in the remaining 48. Intraoperatively,
hypotension occurred in 29 patients, blood loss of
>500 ml in 25 and ECG changes in 10. There were no
deaths. PMI occurred in 5/200 (2.5%) patients. Four had
undergone vascular operations and one had had an
abdominal operation. The mean age of the patients with
PMI was 64.2 years. The duration of operation and blood
loss were similar to those of patients without PMI. None of
these patients developed cardiac failure or cardiogenic
shock and none of them died.
Conclusion: The incidence of PMI among patients undergoing noncardiac surgery is low and its mortality is negligible.
Physicians should become more selective in the use of monitored beds and in the ordering of a work up to rule out PMI.
Relationship between cardiac troponin I (cTnI) release during cardiac operations and myocardial
cell death
JF Vazquez-Jimenez*, Ma Qing†, B Klosterhalfen‡, O Liakopoulos*, G von Bernuth†, BJ Messmer* and M-C Seghaye†
Department. of Cardiac Surgery*, Pediatric Cardiology†, Pathology‡, Aachen University of Technology, Pauwelsstrasse. 30,
D-52057, Aachen, Germany
Aims: To study the relationship between myocardial release
of cTnI and myocardial cell death as assessed by the
amount of apoptosis and necrosis after cardiac surgery.
Methods: Eighteen young pigs were operated on with
standardized cardiopulmonary bypass (CPB). Release of
cTnI in the cardiac lymph (CL), coronary sinus (CS), and
arterial blood (A) was related to postoperative myocardial
cell death by both necrosis and apoptosis. Apoptotic cells
were detected by a TUNEL detection kit. Necrotic cells
were counted by light microscopy.
Results: In all animals, cTnI was significantly released and
reached peak values observed simultaneously in A (cTnI,
20.1±2.6 ng/ml) (mean ±SEM), CS (19.5±3.2 ng/ml) and
CL (5202±2500 ng/ml). Percentage of total myocardial
cell death was 3.1±0.5%, including 1.2±0.35% necrosis
and 1.9±0.5% apoptosis. cTnI release during and after
CPB did not correlate with the degree of myocardial apoptosis or necrosis.
Conclusion: Cardiac operations with CPB are related to
myocardial cell damage including myocardial cell death
due to both necrosis and apoptosis. As the loss of cTnI is
not related to the amount of cell death, our results suggest
that increased cardiac myocyte membrane permeability
more than cell death is responsible for intraoperative and
postoperative cTnI release.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Evaluation of thoracic fluid contents in patients with acute myocardial infarction
H Okawa, A Suzuki, I Sakai, H Tsubo, H Ishihara and A Matsuki
University of Hirosaki School of Medicine, Department of Anesthesiology, 5 Zaifucho, Hirosaki, 036-8562 Japan
It is common to evaluate patients with acute myocardial
infarction (AMI) according to the Forrester classification. A
high PCWP value is a good predictor of pulmonary congestion although there is no documented correlation
between PCWP and the degree of pulmonary congestion
in patients with normal PCWP. In this study, we sought to
investigate the correlation between (1) PCWP and thoracic fluid contents (TFC) as an indicator of pulmonary
congestion, (2) PCWP and cardiac index (CI) and (3) CI
and TFC in patients with AMI with PCWP values
<18mmHg. The thoracic electrical bioimpedance (TEB)
method was used to measure TFC and CI. CI was also
measured by the thermodilution (TD) method.
Thirty-six patients with a diagnosis of AMI were included.
Patients were monitored with the TEB device (BioZ
System, CardioDynamics International Co, USA) and the
pulmonary artery (PA) catheter, in addition to conventional
monitoring. We measured CI by means of TEB and thermodilution (TD) methods (CI-TEB and CI-TD, respectively)
and compared values obtained from each patient with
those two methods. We also measured PCWP with the
PA catheter, and TFC values as an indicator of pulmonary
congestion with the TEB device. The correlation between
PCWP and CI, PCWP and TFC, CI and TFC were also
Patients were 29 males and 7 females and age was
65±11 years (mean ±SD). All the patients had AMI in the
left ventricle. CI-TEB and CI-TD were positively correlated
(r2=0.75). CI-TD was inversely correlated with PCWP
(r2=0.22). TFC was weakly correlated with PCWP
(r2=0.26). Neither CI-TEB nor CI-TD had significant correlation with TFC (r2=0.01 and 0.02, respectively).
In this study, we confirmed that TEB method might be clinically useful for CI measurement in patients with AMI.
PCWP was negatively correlated with CI-TD indicating
that decreased CI in our patient group was not due to systemic dehydration but to decreased cardiac contractility.
The weak correlation between PCWP and TFC means
that the tendency of pulmonary water retention exists with
an increasing PCWP even in patients with normal PCWP.
The clinical importance of this increase in TFC is not clear.
TFC values had no significant correlation with CI values.
This means that the degree of pulmonary congestion is
unpredictable from CI provided that PCWP remains
normal. This is consistent with the Forrester classification.
Is the volume of a pleural effusion predictable using the thickness of the pleural lamella
measured by sonography as a reference?
M Cardon, N Müller, M Van de Velde, J Ghekiere, K Vandevelde, L Steyaert, E Vandermeersch and M Bourgeois
Introduction: The aim of this study was to quantify the
volume of pleural effusions (PEs) in the critically ill using
ultrasound. PE was suggested on the daily postero-anterior chest radiography [1] in the semirecumbent position.
All patients with suspected PE were investigated with
ultrasound. We hypothesize that there is a strong correlation between the maximal width of the fluid lamella along
the lateral chest wall (seen on sonography) and the
volume of pleural fluid punctured.
Materials and methods: The study was approved by the
hospital Ethics Committee. Eighty-seven consecutive critically ill patients underwent a pleural puncture when ultrasound analysis revealed a lamella of more than 2 cm [2]. A
total of 138 pleural punctures was performed in 87 individuals. Ultrasound was performed using the ACUSON,
sequoia 512, with the patient in the semirecumbent position. The deepest possible puncture side in this position
was marked for pleural puncture. The PE was gradually
drained, 200 ml at a time, until the fluid was completely
evacuated. The width of the pre-puncture effusion lamella
as measured with ultrasound was compared case by case
with the actual punctured volume. Statistical analysis was
performed using linear regression analysis and Spearman
(rank) correlation coefficient.
Results: Due to technical difficulties or missing data, 7
punctures had to be excluded. 131 punctures (67 on the
left hemithorax and 64 on the right) remained for analysis.
No complications were encountered as a consequence of
the pleural puncture.
The sonographic measurements correlated very well with
actual effusion volume on the left (rs=0.83) and the right
left (cm)
left (ml)
right (cm)
right (ml)
Std. Deviation
Percentile 25
Percentile 50
Percentile 75
Volume Right
Volume Left
Pleura Left (cm)
Pleura Right (cm)
Left: rs=0.83, P<0.001, y=208.77×–317.12. Right: rs=0.77, P<0.001, y=178.38×–159.7.
side (rs=0.77). For both sides the level of statistical significance was taken as P<0.001.
The thickness of the fluid lamella was taken as the independent variable, while the actual effusion volume was
taken as the dependent variable. The linear sonographic
y=208.77×–317.12 (left), and y=178.38×–159.7 (right).
Y is the predicted effusion volume in milliliters and×is the
sonographically measured thickness of the effusion
lamella in centimeters. The mean predicted error was
199.7 ml for the left side and 285.3 ml for the right side.
Discussion: There is a strong correlation between the
sonographic measurements and the actual effusion
volume which was punctured. Unlike the findings of other
authors, we were not able to predict the punctured
volume, based on the width of the lamella measured by
sonography probably due to the wide spread of obtained
data (scatter plot).
Müller NL: Imaging of the pleura. Radiology 1993, 186:297–309.
Eibenberger KL, Dock WI, Ammann ME, Dorffner R, Hörmann MF,
Grabenwöger F: Quantification of pleural effusions: sonography
versus radiography. Radiology 1994, 191:681–684.
Evaluation of a new wireless ECG electrode built in a telemeter in critically ill patients
Y Sato, K Ikegami, F Morimoto, Y Maekawa, T Sekiguchi*, F Matsumura* and H Sakata*
Trauma and Critical Care Center, Koshigaya Hospital, Dokkyo University School of Medicine, 2-1-50 Minami-Koshigaya,
Koshigaya, Saitama, Japan; *R and D Center, Nihon Kohden Corporation, Tokyo, Japan
Introduction: Aiming for the improvement of ‘spaghetti syndrome’ caused by many leadwires for the measurements of
biological signals such as electrocardiogram (ECG), BP,
SpO2, EtCO2 and lines for intravenous infusion, we developed the new small wireless ECG electrode (size:
3cm×10cm) consisting of a battery-operated telemeter
equipped with two electrodes. The purpose of this study
was to evaluate the ECG with the wireless ECG electrode
placed on critically ill patients in the emergency room.
Methods: ECGs were recorded with both a wireless ECG
electrode placed on the right subclavian area and a conventional three-lead ECG (II) in 30 critically-ill patients
whose ECGs were not only of normal rhythm but also
arrhythmic. The 12-lead ECG (II) was also recorded. The
wireless ECG electrode was also evaluated during cardiopulmonary resuscitation.
Results: The wireless ECG electrode was quickly and
easily placed on the patients compared with the conventional ECG electrode. ECGs were recorded very well in all
patients utilizing the wireless ECG electrode. It was also
possible to detect various arrhythmias such as VPC,
SVPC, Af, VF, VT with the wireless ECG electrode as
clearly as with the conventional ECG electrode. The influence of body movement was less than for the conventional ECG electrode. The wireless ECG electrode was
operative even during DC shock.
Conclusion: It was found that the wireless ECG electrode
was superior in its handling, ECG performance, and durability, suggesting its possible contribution toward the
improvement of ‘spaghetti syndrome’ in the emergency
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Immediate complications of central venous cannulation in ICU
AM Varvinski and GP Findlay
Critical Care Directorate, University Hospital of Wales, Heath Park, Cardiff, Wales, UK, CF4 4XW
Introduction: Catheterization of the central veins is one of
the most widespread procedures in any Intensive Care
Unit (ICU). Despite the availability of different literature on
anatomy, techniques and the introduction of special ultrasound devices, complications are frequent. Complication
rates range from 7–20% according to the literature.
95.8%). There were 21 complications (complication rate,
17.5%). They were intra-arterial insertion/puncture 9
(7.5%), perivascular insertion 3 (2.5%), pneumothorax 1
(0.83%), bleeding requiring blood transfusion 1 (0.83%),
arrhythmias requiring intervention 1 (0.83%), surgical
emphysema 1 (0.83%), unsuccessful 5 (4.16%).
Methods: This is a prospective study to assess the complication rate from central venous cannulation in the ICU.
A separate registration form for every attempted catheterization was used. Data collected were: patient’s demographics, grade of operator, approach used, number of
attempts, indication for insertion, use of ultrasound probe,
type of catheter and complications.
Conclusion: The complication rate of central venous
catheterization remains high even in experienced hands.
Fatal or serious complications are rare. The use of the
ultrasound probe may decrease the complication rate
even if used for checking the anatomy only. Intra-arterial
puncture/insertion, considered by many as a relatively
innocent complication, can lead to unnecessary morbidity
and even fatal complications, especially in the ITU population with abnormal clotting. Formal teaching should be
provided for all new staff in training starting their appointments on the Intensive Care Units.
Results: In the first three months, 82 patients were
involved in the study. There were 120 attempted catheterizations and 115 lines were inserted (success rate,
A comparison of pulmonary artery occlusion pressure (PaOP) measurements using pressure
controlled ventilation (PCV) versus airway pressure release ventilation (APRV)
LJ Kaplan and H Bailey
Medical College of PA Hospital, Departments of Surgery and Emergency Medicine, Division of Trauma and Critical Care, 3300
Henry Avenue, Philadelphia, PA 19129, USA
Purpose: To determine the optimal time within the APRV
phase cycle to accurately measure PaOP.
Methods: Ten consecutive patients with acute lung injury
(ALI) managed with PCV and a pulmonary artery catheter
(PAC) were studied. Demographic data was recorded.
Patients served as their own controls and were ventilated
by a Drager Evita 4 Pulmonary Workstation. No patients
received paralytics. PCV settings (AC mode) achieved a
pCO2 of 35–45(torr) and a pO2>60 (torr) on 60% O2;
PEEP was not controlled. Hemodynamic profiles were
recorded 30 min after achieving the above pCO2 and pO2
values. Patients were then changed to APRV to achieve
the same pCO2 and pO2 values and hemodynamic measurements were repeated at 30 min. All medications were
held constant. PaOP tracings (mmHg) were recorded and
compared to the downloaded flow-time trace from the
ventilator (Evitaview software). The PCV PaOP was
recorded at the end of exhalation and served as the standard for comparisons. PaOP was recorded during the
APRV phase cycle (positive pressure and release) and
compared to the PCV value. Data are shown as means
±standard deviation and were compared using a twotailed paired t-test; significance assumed for P<0.05.
Results: Principal diagnoses were trauma (66%), abdominal sepsis (32%), and other (2%). Mean age was 54±6.2
years. PCV blood gas values were pH 7.34±0.04, pCO2
39.3±3.8, pO2 77.4±9.5. APRV blood gas values were
pH 7.37±0.03, pCO2 35.5±2.8, pO2 98±11, (P<0.05 vs
PCV). The PCV PaOP was 16.3±3 on a PEEP of
13.6±2.2 cmH2O with a CI of 3.2±0.5 L/min/m2 and an
SvO2 of 76.8±4.5% at a hemoglobin of 9.6±1.04 gm%.
The APRV PaOP during the positive pressure phase was
21.2±3.3 (initial), 19±2.5 (mid), and 20.5±2.8 (end);
P<0.01 for all versus PCV. The APRV PaOP during the
release phase was 19±2.7 (initial, P<0.05), 17.7±2.3
(mid, P=0.09), and 16.4±2.6 (end, P=0.9). CI was significantly increased at 3.6±0.4 (P<0.01 vs PCV) while SvO2
was unchanged at 79.1±4.1 (P>0.05 vs PCV).
Conclusions: APRV increases the measured PaOP
during the positive pressure phase. PaOP may be reliably
measured at the midpoint or end of the release phase of
APRV. APRV increases oxygenation and cardiac index
compared to PCV in patients with acute lung injury.
Effective pulmonary capillary pressure during hyperdynamic porcine endotoxemia
T Handschuh, D Vasilev, M Georgieff, P Radermacher and B Šantak
Department of Anesthesia, University Hospital, D-89073, Ulm, Germany
Introduction: The pulmonary microvascular hydrostatic or
effective capillary pressure (PCPeff) is the principal determinant of the rate of edema formation [1,2]. Since scarce
data are available only on the natural development of
PCPeff, independently from vasoactive treatment, we
investigated the time course of both PCPeff and the longitudinal distribution of pulmonary vascular resistance
during long-term hyperdynamic porcine endotoxemia.
Methods: Over 9 h anesthetized and ventilated pigs
received continuous intravenous endotoxin (ETX, n=7) or
placebo (Sham, n=5). Hydroxyethylstarch was infused to
maintain MAP >65 mmHg together with a sustained
30–50% increase in cardiac output. Mean (MPAP) and
pulmonary artery occluded (PAOP) pressures were
recorded before and 3, 6, and 9 h after the start of ETX.
After signal filtering via a low-pass filter, PCPeff was
derived from a bi-exponential fit to the post-occlusion
pressure decay curve, back-extrapolating the slower part
to the instant of occlusion determined by simultaneously
recording the balloon pressure [1]. The contribution of the
venous resistance to the pressure drop over the pulmonary vascular bed was calculated as PVRven=(PCPeffPAOP)/(MPAP–PAOP) [1].
Results: See Table.
Conclusion: Early endotoxemia results in a pronounced
increase of PCPeff both due to increased downstream (i.e.
left atrial) pressure as well as increased contribution of the
postcapillary pressure drop. Hypoxia and/or early mediator
release (i.e. thromboxane) probably assume importance in
this context [1]. The later normalization of both PCPeff and
PVRven may be due to increased NO formation [3].
Acknowledgement: D Vasilev is supported by the Deutscher Akademischer Austauschdienst.
Cope et al.: Critical Care Med 1992, 20:1043.
Grimbert: Eur Respir J 1988, 1:297.
Hinder et al.: J Appl Physiol 1997, 83:1941.
Before ETX
3 h ETX
6 h ETX
9 h ETX
22 (18/23)
19 (17/22)
47 (42/52)#*
19 (18/21)
44 (41/52)#*
22 (20/25)
34 (32/35)#*
22 (21/23)
7 (5/10)
6 (4/7)
11 (10/13)#*
6 (5/7)
9 (7/12)
5 (4/5)
8 (7/11)
6 (5/7)
10 (8/11)
8 (6/9)
24 (20/26)#*
7 (6/7)
17 (11/24)*
8 (6/8)
14 (12/16)
9 (7/10)
23 (21/27)
19 (18/22)
37 (34/39)*
15 (11/18)
24 (23/27)
16 (15/19)
22 (19/26)
16 (14/20)
Median (25/75%), # P<0.05 vs 0h (Friedman-Anova); *P<0.05 ETX vs Sham (Mann-Whitney)
A comparison between oesophageal Doppler and continuous thermodilution for the
measurement of cardiac output in critically ill patients
JR Wilkinson, C Ruff, SK Patel, RM Venn, A Rhodes and ED Bennett
Department of Intensive Care, St George’s Hospital, Tooting, London, SW17 OQT, UK
Introduction: There have been concerns raised over the
safety of pulmonary artery (PA) catheters. Previous studies
have shown oesophageal Doppler (OD) monitoring provides a reliable, easy alternative method of assessing
cardiac output (CO) in an ICU setting.
Methods: After ethical approval we studied 12 ICU patients
who required PA catheters for monitoring of therapy. A continuous cardiac output (CCO) PA catheter (Vigilance
monitor, Baxter) and OD probe (CardioQ monitor, Deltex),
were inserted in each patient. Insertion time and complications were noted for both techniques. Measurements of CO
by thermodilution and OD were recorded. Data was
analysed using regression analysis and Bland-Altman for the
difference between the two techniques.
Results: Twelve patients were analysed with a total of
128 pairs of cardiac output measurements. There were
four complications due to PA catheters and none due to
OD. Mean insertion time was 52.9 min for PA catheter
and 5.8 min for OD (P<0.0001). Thermodilution cardiac
outputs were obtained between 2.9 and 15.5 L/min. Pearson’s correlation coefficient (r) for PA catheter vs OD was
0.689 (P<0.0001). The mean difference in CO between
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
the two methods was–0.17 L/min with limits of agreement
of –4.016 to +3.676.
Discussion: To our knowledge, this is the first study to
compare continuous cardiac output by thermodilution
with cardiac output by OD (CardioQ monitor). OD
probes are faster to insert, easier to use and safer.
However, our study shows that there are major differences in the cardiac output values obtained by the two
Comparison of pulmonary arterial and arterial trans-cardiopulmonary thermodilution cardiac
output in porcine septic shock
B Vangerow, M Cobas Meyer, J Ahrens, T Schuerholz, G Marx, M Moeller, M Leuwer and H Rueckoldt
Department of Anaesthesiology, Hannover Medical School, D-30625, Hannover, Germany
Introduction: Despite its invasiveness and inherent risks,
the pulmonary artery catheter (PAC) is still regarded as
the clinical standard for cardiac output (CO) determination. Arterial trans-cardiopulmonary thermodilution is a less
invasive method (PAC not necessary) for CO monitoring.
The aim of this study was to compare arterial trans-cardiopulmonary thermodilution to conventional pulmonary
arterial thermodilution for CO determination during substantial hemodynamic variations in a sepsis model.
Methods: In a prospective study 24 anaesthetized,
mechanically ventilated pigs (19.7±1.6 kg) with peritonitis-induced septic shock were investigated. Cardiac
output was determined using a 7.5F thermodilution
catheter placed in the pulmonary artery and a 4F thermistor tipped catheter (Pulsion Medical Systems, Germany)
inserted into the right carotid. Nine sets of corresponding
CO determinations were obtained during a period of 8 h in
each animal, all measurements were performed in triplicate. Data were analyzed using Bland-Altman analyses,
linear regression and correlation.
Results: During the period from induction of peritonitis to
profound septic shock, major variations in heart rate
(range: 48–310 beats/min) and systemic vascular resistance (range: 7400–1340 dyne×sec×cm-5) were
observed. 196 sets of CO determinations were yielded
with a mean CO measured by pulmonary arterial thermodilution (PATD) of 2.0±0.7 L/min (range: 0.7–5.2 L/min).
Mean CO measured by arterial trans-cardiopulmonary
thermodilution (ATPTD) was 2.17±0.6 L/min (range:
1.0–4.4 L/min). Linear regression equation was:
ATPTD=0.711×PATD +0.82; r2=0.68. The mean bias
was 0.17 L/min (95% confidence interval: 0.125–0.215
L/min), with limits of agreement of–0.61 to 0.95 L/min and
a precision of 0.34 L/min.
Conclusion: Arterial trans-cardiopulmonary thermodilution
CO correlates acceptably with pulmonary arterial thermodilution CO even during pronounced hemodynamic
Comparison of pressure-related performance data with data measured by thermodilution in heart
failure and sepsis
L Engelmann, J Otto and HE Kunig*
University of Leipzig, Center of Internal Medicine, Department of Intensive Care, Philipp-Rosenthal- Straße 27a, 04103 Leipzig,
Germany; * P.O. Box 192, Saltsburg, PA 15681-0192, USA
Introduction: The performance diagrams evaluate the cardiocirculatory function by means of blood pressure, pulse
rate and body surface area (BSA). The basics are represented by Kunig et al. in this symposium. The performance
data predict nonsurvival with a sensitivity of 94% and a
specificity of 84%.
Aim of study: Comparison of hemodynamic data measured by thermodilution with pressure-related performance diagrams and their use in differentiation of critical
cardiocirculatory states.
Method: On-line measured performance data are compared with discontinuously determined thermodilution
data in patients with heart failure (CI<2.5 l/min/m2;
PCWP>20 mmHg) and sepsis (CI>4.5 l/min/m2; PCWP
>16 mmHg).
Results: In 9 patients with cardiac failure (CI 1.94±0.44
l/min/m2; PCWP 23.3±5.5 mmHg) the efficiency (EF[P])
declines to 0.27±0.03, the diastolic blood pressure (DBP)
and systolic blood pressure (SBP) range slightly above
the basal values (DBP* 2.64±0.58; SBP* 3.61±0.73) and
the pulse pressure rate (PP*) below the critical border of 1
(0.93±0.30). In 20 patients with sepsis (CI 5.22±1.32
l/min/m2; PCWP 19.4±4.4 mmHg) the efficiency is normal
(0.48±0.07), while SBP* and DBP* are remarkably
increased (5.16±1.23 and 2.69±0.85). The PP* ranges is
2.52±0.61. The differences in CI, EF(P), PP* and SBP*
are strongly significant (P<0.01) in sepsis compared with
heart failure.
Conclusions: (1)The pressure-related performance data
show changes in case of abnormal deviations of volumetrically determined hemodynamic parameters.(2) A differenti-
parameters from basal values signals a cardiocirculatory
deterioration, which should be analyzed by conventionally
measured hemodynamic parameters at present.
ation of hemodynamic parameters between sepsis and
severe heart failure succeeds by means of CI, EF(P), PP*
and SBP*. (3) A remarkable deviation of performance
Compliance or failure and improvement or deterioration diagnosis of patients from performance
HE Kunig, SV Kunig, L Engelmann*, U Pilz*, J Otto*, L Mende*, E Huettemann†, TJ Gallagher‡, J Elledge-Nauman‡,
JK Stene§, P Tassani#, U Jaenicke**, H Patrick†† and MR Pinsky‡‡
Dept. Bioengineering, University of Washington, Seattle, USA; *Dept. Int. Care Med. University of Leipzig; †Dept. Crit. Care
Med., University of Jena; ‡Dept. Crit. Care Med., University of Florida, USA; §Dept. Crit. Care Med., Pennsylvania State
University, USA; #German Heart Center, Munich; **Dept. Anesth., University of Munich, Germany; ††Dept. Crit. Care Med.,
Jefferson Med. College, Philadelphia, USA; ‡‡Dep. Crit. Care Med., University of Pittsburgh USA.
deterioration or improvement when the trend of two successive measurements departs from, or points towards,
the normal values. Data from 213 patients were retrospectively analyzed using standard statistical methods including sensitivity and specificity determinations, binomial
scoring, and t-testing.
Introduction: Blood pressure and heart rate data displayed in a performance diagram (PD) may diagnose
accurately compliance or failure and improvement or deterioration while traditional hemodynamics would indicate a
patient to be stable.
Materials and methods: A PD plots the parameters pressure efficiency (EF[P]) versus time in an upper graph and
arterial pulse pressure (PP*), systolic pressure (SBP*),
and diastolic pressure (DBP*) versus time in a lower
graph. EF(P)=PP*/SBP* in analogy to the volume ejection
fraction EF(V)=SV/EDV, where SV=stroke volume and
EDV=end-diastolic volume. The asterisk (*) indicates conversion of events per beat into events per time and standardization to body surface area (BSA), f. e.,
PP*=(PPxHR)/BSA, analogous to the conversion of SV to
cardiac index (CI), where CI=(SV*HR)/BSA. PDs suggest:
(1) compliance when all parameters equal or exceed
normal values (N), (2) failure without immediate danger of
death when normal values of EF(P),or SBP*, or DBP* are
not maintained, (3) failure with immediate danger of death
(critical illness), when normal PP* is not maintained, (4)
Results: Based on statistical analysis, PDs predicted
compliance and failure with a sensitivity of 94% and a
specificity of 84%. PDs also predicted improvement and
deterioration at P<0.01. As illustrated in the figures, traditional hemodynamics diagnoses a patient as stable as late
as 10 min before the occurrence of flash edema. The PD
diagnoses a failing and deteriorating patient as early as
2.5 h prior to occurrence of flash edema and diagnoses
the patient as critically ill 25 min prior to the occurrence of
the flash edema.
Discussion: This study suggests utility of PDs in accurately
diagnosing compliance or failure which would allow early
intervention and monitoring of the effects of intervention in
real time as compared to traditional hemodynamic evaluation.
Performance Diagram
Traditional Hemodynamics
HR 1/min
EF(P) %
1 HR
2 HR
2 HR
DBP* mm Hg/(m2 sec)
PP* mm Hg/(m2 sec)
1 HR
SBP* mm Hg/(m2 sec)
flash edema
flash edema
SBP mm Hg
DBP mm Hg
PP mm Hg
1 HR
2 HR
1 HR
2 HR
Critical Care
20th International Symposium on Intensive Care and Emergency Medicine
CI/SVRI relationship during different phases of inflammatory response
R Kula, P Sklienka, I Petrašovicová, L Kolár and J Jahoda
Department of Anaesthesia and Intensive Care, Faculty Hospital, Ostrava, Czech Republic
Figure 1
r2 = 0.6932
Introduction: High cardiac index (CI) and low systemic
vascular resistance (SVRI) are frequently observed in
septic patients and are essential in consideration of further
therapeutic interventions. As these findings could be
observed even in healthy individuals (i.e. during hard exercise) we decided to analyse the relationship between CI
and SVRI in patients without apparent clinical signs of the
inflammatory response and in patients with the different
phase of the inflammatory response.
Kula et al.: Int Care Med 1997, 23(Suppl.1):S74.
Figure 2
r2 = 0.7876
Cardiac index
CI/SVRI relationship during period of the localized inflammatory
Figure 3
r = 0.7621
CI/SVRI relationship during period of the generalized inflammatory
Results: See Figures
Conclusion: The relationship between CI and SVRI
shows a different course in the phase of the generalized
inflammatory response compared to the phase of the
localized inflammatory response. The course of the relationship between CI and SVRI in patients without inflammatory response is nearly identical to that in patients with
symptoms of the localized inflammatory response. On the
basis of these findings, we presume that assessment of
the relationship between CI and SVRI (but not isolated CI
and/or SVRI) is needed for accurate therapeutic decision
Cardiac index
Methods: Thirty-one critically ill patients with pulmonary
artery catheter inserted were included in this prospective
study (average age 48±16 years, average sum of SOFA
score 2.4±1.36, etiology: 60% traumatic patients), all of
them with apparent clinical signs of the inflammatory
response (heart rate >90 bpm, WBC >12000 or <4000,
BT >38ºC or <36ºC, CRP >50mg/l) with microbiologically
confirmed infectious etiology. In accordance with our findings (published in Int Care Med 1997, 23(Suppl.1.):S72)
we identified 16 periods of generalization of the inflammatory response (fall in platelet count, antithrombine-III activity
and serum albumin, positive fluid balance, hemodynamic
instability). Within these periods, 101 hemodynamic measurements were performed, measured parameters are
shown in Fig. 1. In periods without symptoms of generalized
inflammatory response we performed 115 hemodynamic
measurements, results of which are shown in Fig. 2. 130
hemodynamic measurements in patients without apparent
clinical signs of the inflammatory response (25 patients, age
51±17 years, pulmonary artery catheter inserted in setting
of preoperative hemodynamics optimization) we used as the
control group. Results are shown in Fig. 3.
Vol 4 Suppl 1
Cardiac index
CI/SVRI relationship in patients without the inflammatory response
Would the blood volume analyzer-100 (BVA-100) be an alternative to pulmonary artery catheter
(PAC) in critically ill patients
S Alrawi, W Cholewczynski, R Raju, A Acinapura and J Cunningham
Department of Surgery, Trauma office, Lutheran Medical Center, 150 55th Street, Brooklyn, New York, NY 11220, USA
Background: Measurements of urine output, mean arterial
pressure (MAP), heart rate (HR), pulmonary wedge pressure (PWP) and cardiac index (CI) reflect the status of
central blood volume including heart, brain, and lungs but
do not reflect the status of peripheral blood. The accurate
measurement of total blood volume provided by a certain
noninvasive technique compared to a ‘guesstimate’ given
by hematocrit reading or by PAC with all its possible complications could mean the difference between life and
death. The standard tests—hemoglobin and hematocrit—do
not provide an accurate measure of blood volume in
sudden acute blood loss situations.
Material and methods: Twenty-five critically ill patients in
the intensive care unit (ICU; surgical and medical) with
PAC and fourteen normal persons (control) were evaluated
after IRB approval for total blood volume measurement
(BVM) by BVA-100. The instrument is used in conjunction
with a special patented syringe, which delivers a precise,
quantitative injection of Albumin I-131 isotope into the
patient. Six consecutive blood samples were then taken. All
ICU patients were classified according to the APACHE II
severity of disease classification system, PAC parameters
were considered and fluid balance, ventilatory settings,
vasopressors and treatment were evaluated. Serum lactate,
arterial blood gas and mixed venous blood gas were determined. All these parameters were correlated with total
blood, red cell and plasma volume.
Results: We find no significant correlation between blood
volume (BV), pulmonary capillary wedge pressure
(PCWP) and central venous pressure (CVP) readings.
There is also no correlation between (BV) and the other
variables in the study including base excess from arterial
blood gas, mixed venous blood gas saturation, cardiac
output, cardiac index, systemic vascular resistance, serum
lactate and oxygen delivery (A-VO2).
Conclusion: Clinical decision-making in the management
of critically ill patients regarding blood volume status
should be measured by direct blood volume analysis
rather than by the indirect parameters of PCWP and CVP.
Sepsis increases accumulation of cell-free hemoglobin in intima of submucosal arterioles in rats
AW Sielenkämper, K Eicker, H Kloppenborg, R Dreyer*, WJ Sibbald† and H Van Aken
Department. of Anesthesiology and Intensive Care Medicine, *Department of Pathology, University of Münster, 48129 Münster,
Germany; †AC Burton Vascular Biology Laboratory, University of Western Ontario, London, Canada
Introduction: It was reported from in vitro studies that
interruption of the endothelial barrier increases endothelial
permeability for cell-free hemoglobin solutions[1]. We
studied the effects of sepsis on the abluminal distribution
of intravenously infused cell-free hemoglobin (DCLHb,
Baxter Corp.) in submucosa of rat ileum.
Methods: Rats were randomized to sham laparotomy
(n=6) or cecal ligation and perforation (CLP, n=7) to
induce sepsis. Twenty-four h later, 300 mg of DCLHb
labeled with Evans-Blue dye was infused. The laparotomy
was reopened, and material was obtained from the ileum
90 min after DCLHb infusion for histological examination.
Using fluorescence microscopy, the concentration of cellfree hemoglobin in intima and smooth muscle of submucosal arterioles and in submucosal tissue was graded by
two blinded investigators using a score (0, no; 1, minor; 2,
moderate; 3, strong; 4, very strong fluorescence). Additional studies (n=3) were performed to exclude that the
tracer alone accounted for the observed effects.
Results: DCLHb fluorescence scores were low in submucosal tissue and in smooth muscle of arterioles for both
groups (all <0.75 points.). In the intima of arterioles, considerable fluorescence was observed, with scores
increased in the septic group as compared to the sham
group (3.2±0.5 vs 1.4±0.4 points, mean ±SEM, P<0.05).
Infusion of Evans-Blue dye alone resulted in scores <0.75
points in all tissues.
Conclusion: Sepsis increased the accumulation of cellfree hemoglobin in the intima of submucosal arterioles.
This finding may be relevant for the previously documented effects of cell-free hemoglobin to modify perfusion
pressure and microvascular blood flow in sepsis [2].
Acknowledgement: This work was supported by Baxter Healthcare Corp.,
Nakai et al.: J Lab Clin Med 1998, 132:313–319.
Sielenkämper et al.: Crit Care Med 2000, in press.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Diaspirin cross-linked hemoglobin (DCLHb) ensures tissue oxygenation during hemodilution
below the critical hematocrit
G Kemming*, F Meisner, O Habler*, M Kleen*, J Tillmanns, J Hutter, A Pape, J Meier, C Wojtczyk, D Bottino and K Messmer
Institute for Surgical Research and *Clinic for Anesthesiology, Klinikum Großhadern, Marchioninistrasse 15, Ludwig-MaximiliansUniversity, D-81377 Munich, Germany
Background: Normovolemic hemodilution (HD) is an
effective blood sparing strategy in the perioperative
setting. Despite physiologic compensation for reduced
hematocrit (Hct) and O2 content (CaO2), tolerance to HD
is limited. At a Hct called ‘critical’ (Hctcrit), tissue hypoxia
Objective: To assess, whether HD can be extended
below Hctcrit when using an hemoglobin-based O2 carrier
as diluent.
Methods: Twelve anesthetized, mechanically-ventilated
pigs were hemodiluted by 1:1 exchange of blood with either
10% DCLHb (DCLHb, n=6) or 8% albumin (HSA, n=6)
oncotically matched to DCLHb (diluents were provided by
Baxter Healthcare Inc., Boulder, CO). In both groups, measurements were performed at baseline and after HD to (1)
preset Hct-values of 15%, 8%, 4% 2%, 1% or (2) until the
individual Hctcrit, defined by the onset of myocardial
ischemia (ST-segment depression), was reached. Total
body O2-delivery and O2-uptake (DO2I, VO2I), as well as
local tissue O2 partial pressure (tPO2, MDO-Electrode,
Eschweiler, Kiel, Germany) on the surface of liver and skeletal muscle were assessed during every measurement
(Median ± [Q1–Q3]/2; rANOVA, MWU-test P<0.05).
Results: In the HSA-group, Hctcrit was 6.1±0.9%, at
which all animals died. In the DCLHb-group, Hctcrit was
not reached despite hemodilution to Hct 1.2%. At Hct 4%
in the DCLHb-group, all parameters (CaO2, DO2I, tPO2:
P<0.05 vs HSA; VO2I n.s.) were less altered than at Hctcrit
in the HSA-group.
Conclusion: During HD with DCLHb, O2-transport and
tissue oxygenation were fully preserved at Hct 4%, while
all HSA-animals had died at Hct 6.1%. The difference is
due to the O2 transporting capacity of DCLHb.
Anesth Analg 1997, 85:953–958.
Adv Microcirc 1972, 4:1–77.
Anesth Analg 1992, 75:818–821.
Decrease of red blood cell sialic acid membrane content in septic patients
K Zouaoui Boudjeltia*, M Piagnerelli†, P Piro*, D Bastin*, E Carlier†, P Lejeune† and M Vanhaeverbeek*
*Experimental Medicine Laboratory and †Department of Intensive Care, A. Vésale Hospital, ULB, 6110, Montigny-le-Tilleul, Belgium
Background: Septic shock is a complex physiopathologic
model with alterations in the microcirculation, occurring
together with changes in biochemical and physiologic
characteristics of the blood. Red blood cells (RBCs) of
septic patients are characterized by a decreased deformability and an increased aggregatability. These events are
also noted in RBCs of diabetic patients secondary to a
reduction in sialic acid (SA) RBC membrane content.
Objective: To study the SA RBC membrane content in a
septic population.
Methods: We studied blood samples in ICU patients
without documented sepsis (n=20, age 69±12 years) and
with documented sepsis (n=20, age 65±16 years).
SA was measured on isolated RBC membrane protein by
a fluorimetric technique in high precision liquid chromatography (HPLC; Anumula).
Results: We observed a significant decrease in SA in
septic patients compared with nonseptic patients: SA
1.94±0.79 µg/100 µg protein and SA: 2.2±0.39 µg/100
µg protein (P=0.013), respectively for septic and nonseptic patients.
Conclusions: RBC of septic patients are characterized by
a decrease in SA membrane content as described in diabetic patients. These data highlight another possible biochemical modification of RBC in sepsis.
Preload assessment in septic shock
A Donati, G Conti, S Loggi, S Falcetta, E Adrario, P Pelaia and P Pietropaoli
Istituto delle Emergenze Medico-Chirurgiche, Università di Ancona, Ospedale Regionale Torrette, via Conca, 60020 Torrette
(Ancona), Italy
Background: The accuracy of intra thoracic blood volume
(ITBV) as a preload index, instead of central venous pressure
and wedge pressure, has been demonstrated by Lichtwark-
Ashoff in mechanically-ventilated patients with acute respiratory failure [1,2]. The aim of our work was: (1) to verify ITBV
as a preload index in patients suffering from septic shock
and (2) to relate measured data (CVP, WP, ITBV, extra vascular lung water [EVLW]), with PaO2/FiO2 ratio.
Methods: 15 patients suffering from septic shock were
studied. Admission criterion was septic shock refractory to
standard therapy with cathecolamines. All patients were
monitored with a Swan-Ganz catheter and Cold system
(Pulsion, Münich) and received a bolus of methylene blue
(MB) at the rate of 3 mg/Kg. Hemodynamic and
oxyphoretic data were measured before MB infusion, and
20 min, 1 h and 2 h after the end of MB infusion, so that 4
sets of data for each patient were obtained to give a total
of 60 data points.
Results: There is a significant correlation between ITBV
and stroke volume (SV) and between ITBV and cardiac
index (CI). There is also a correlation between ITBV and
EVLW. No correlation was found between WP and SV
and CI. Then we plotted WP vs the PaO2/FiO2 ratio and
found that when the PaO2/FiO2 ratio was <200 there was
no difference in EVLW between patients with WP ≤16
mmHg and patients with WP >16 mmHg. On the contrary, plotting ITBV vs PaO2/FiO2, with PaO2/FiO2 ratio
>200, EVLW was very significantly higher (P<0.001) if
ITBV was >1100 ml/mq, than if it was ≤1100 ml/mq.
Conclusions: In septic patients, when a respiratory failure
happens, ITBV can be normal or high (>1100 ml/mq). A
high ITBV is related to a high EVLW, while a high EVLW
with a normal ITBV means a normal preload. In this case, it
is not useful to decrease preload to increase the
PaO2/FiO2 ratio. ITBV together with PaO2/FiO2 ratio is
useful to optimize hemodynamic therapy during respiratory
failure in septic patients and allows us to identify the
patients that need diuretic therapy.
Lichtwark-Ashoff M et al.: Intensive Care Med 1992, 18:142–147.
Lichtwark-Ashoff M et al.: J Crit Care 1996, 11:180–188.
The clinical relevance of the fluid balance in critically ill patients
I Petrašovicová, P Sklienka, L Kolár, J Jahoda and R Kula
Department of Anaesthesia and Intensive Care, Faculty Hospital, Ostrava, Czech Republic
Introduction: Fluid management based on the assessment of heart filling pressures (CVP, PCWP) to reach
hemodynamic stability is an important point in management of the critically ill patient. In the case of impaired permeability induced by inflammation, this strategy usually
results in excessive interstitial fluid sequestration. In this
situation, consequent hemodynamic stability is often at the
expense of positive fluid balance and oedema formation,
which can cause tissue oxygenation to deteriorate resulting in organ failure progression. The aim of our study was
to analyse the relationship between fluid intake and fluid
balance together with development of organ dysfunction
in surviving and nonsurviving critically ill patients over the
course of ten consecutive days.
Materials and methods: 117 critically ill patients admitted to our department from 1996 to 1998 were included
in this retrospective study. Criterion for inclusion was hospitalisation longer than 3 days. Patients were divided in
two groups: survivors (n=70, average age 49±16 years,
54% multiple trauma, 24% elective surgery, 26% medical,
average duration of ICU stay 15±8 days) and nonsurvivors
(n=47, average age 51±17 years., 48% multiple trauma.,
20% elective surgery, 32% medical, average duration of
ICU stay 17±11 days). In both groups, fluid intake and
fluid balance were evaluated during a ten day interval
(from 3rd until 13th day of ICU stay, Fig. 1, Fig. 2)
together with intensity of inflammatory response (average
CRP value during whole period of evaluation) and organ
dysfunction development (sum of SOFA score on 3rd and
13th day of ICU stay, Fig. 3).
Results: See Figures.
Conclusion: In the nonsurvivors group a much more positive fluid balance was found within the interval from 3rd
until 13th day of ICU stay, together with increasing extent
of the organ dysfunction. The stronger correlation
between fluid intake and fluid balance in nonsurviving
patients compared to survivors (Fig. 1) points to the need
for careful fluid management in critically ill patients, especially in those with an intensive inflammatory response.
Figure 1
Relationship between fluid intake and fluid balance in nonsurvivors
(average fluid intake, 5810±1800 ml; average fluid balance,
1580±1900 ml; average value of CRP, 125±62 mg/l).
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Figure 2
Figure 3
Relationship between fluid intake and fluid balance in survivors
(average fluid intake, 5539±1160 ml, not significant compared to
nonsurvivors; average fluid balance, 523±1480 ml, P<0.05 compared
to nonsurvivors; average value of CRP, 92±60 mg/l, P<0.05
compared to nonsurvivors).
Organ dysfunction development.
Effects of volume replacement on plasma volume and albumin escape rate in a porcine model of
fecal peritonitis
G Marx*, B Vangerow*, M Cobas Meyer*, T Schuerholz*, R Sümpelmann*, L Wilkens†, KF Gratz‡, M Leuwer* and H Rueckoldt*
of Anaesthesiology, †Pathology and ‡Nuclear Medicine, Hannover Medical School, D-30625 Hannover, Germany
Objectives: We investigated effects of 4% and 8% gelatinbased solutions, hydroxyethyl starch (HES) and ringer’s
solution on albumin escape rate (AER) and maintenance of
plasma volume (PV) in a porcine fecal peritonitis model.
Design: This was a prospective randomized, controlled
animal laboratory study.
Measurement and main results: Fasted, anaesthetized,
mechanically ventilated and multi catheterized pigs
(20.8±1.8 kg) received 1g of feces per kg of body weight
into the abdominal cavity to induce sepsis and were
observed over eight h. At induction, animals were randomized to resuscitation with gelatin 8% (n=5), gelatin 4%
(n=5), 6% HES 200/0.5 (n=5), ringer’s solution (n=5) and
compared to a nonseptic control group (n=5) receiving
6% HES 200/0.5. The infusion rate was set to maintain a
central venous pressure of 12 mmHg. Red blood cell
volume was analyzed using chromium-51-tagged erythrocytes (RBC:51Cr). PV was determined from standard formulae. AER was calculated using 99m-labeled technetium
serum albumin. Systemic hemodynamics and oxygenation
Gelatin 8%
Gelatin 4%
Control (HES)
RBC: 51Cr [ml/kg]
20.1 ± 1.9
20.2 ± 1.7
21.1 ± 2.6
19.8 ± 2.8
22.1 ± 2.5
21.1 ± 2.0
23.2 ± 2.6
18.9 ± 1.4
24.7 ± 1.6
24.6 ± 1.5*
PV: 51Cr [ml/kg]
60.0 ± 5.1
73.9 ± 7.9*
57.3 ± 5.9
54.8 ± 9.5
59.0 ± 8.1
59.1 ± 11.1
61.0 ± 2.2
32.9 ± 2.9*
65.5 ± 7.7
79.8 ± 6.8*
160 ± 25
146 ± 43
401 ± 84*
108 ± 31
Infusion volume [ml/kg] 8h 105 ± 15
SvO2 [%]
71 ± 7
37 ± 20
73± 4
37 ± 13
73± 5
55 ± 19*
75± 5
42 ± 17
78± 5
74 ± 5*
AER [%]
24 ± 9
11 ± 8
24 ± 3
10 ± 3
22 ± 4
21 ± 2
3± 3
-2 ± 4*
*P < 0.05, intergroup differences using univariate analysis of variance (UNIANOVA).
were obtained before induction of sepsis, and 4 and 8 h
Conclusion: PV was increased by gelatin 8% only. This suggests a better remaining volume in circulation during sepsis
as there were no differences between infusion volume in all
colloid groups. On the other hand, systemic oxygenation was
significantly better in animals treated with HES than with
gelatin 8% or 4%. In this respect, it is interesting to note that
histological investigation of lung and kidney revealed a bluish
blubbery fluid in small blood vessels, glomerula and tubules
of the animals treated with gelatin 8% or gelatin 4%.
The formation of platelet microvesicles in septic pigs treated with different kinds of volume
T Schuerholz*, G Marx*, B Vangerow*, M Cobas Meyer*, M Ballmaier†, J Heine*, HJ Schuberth‡, H Rueckoldt* and R Sümpelmann*
*Department of Anesthesiology, †Pediatric Hematology and Oncology, Hannover Medical School, 30625 Hannover, Germany;
‡Immunology Unit3, Hannover School of Veterinary Medicine, Hannover, Germany.
Introduction: Bacterial lipopolysaccharide (LPS) is believed
to be one of the major pathogenic factors of Gram-negative
sepsis. Sepsis may be accompanied with thrombocytopenia, platelet activation and elevated platelet-derived
microvesicles. Higher formation of microvesicles has been
associated with higher mortality in endotoxemic pigs.
Design: Prospective, randomized animal laboratory study.
Methods: Twenty anesthetized pigs divided in four groups
received feces (1 g/kg body-weight) installed in the
abdomen to induce sepsis. Group 1 was treated with
ringer’s solution (RS), group 2 with 6% hydroxyethyl
starch 200/0.5 (HES), group 3 with modified fluid gelatin
4% (MFG4) and group 4 with modified fluid gelatin 8%
(MFG8), each to maintain a central venous pressure of
12 mmHg. Before induction of sepsis (baseline) and after
2 (+2 h), 4 (+4 h), and 8 (+8 h) h, a citrated blood sample
was drawn. Flow cytometry was used for determination of
microvesicles. Platelets and microvesicles were identified
with an anti-platelet-mAb (Donor: DH Sachs, Charlestown,
USA) and a secondary antibody labeled with phycoeryP22
thrin (goat-anti-mouse, Dako SA, Denmark). Microvesicles
were determined as smallest 1–3% positive cells in
forward scatter.
Results: Baseline values were considered as 100%. In
the RS-group sepsis was associated with an increase of
microvesicles to 137±58% (Mean ±SD) after 2 h, leading
to 179±45% (+4 h) up to 210±121% (+8 h). In group 2
(HES) there was an increase at +2 h (109±25%) and a
decrease at +4 h (89±26%) and +8 h (73±19%). In
Group 3 and 4 a decrease of microvesicles after induction
was revealed at +2 h (51±25%, MFG4 and 45±16%,
MFG8) +4 h (62±10%, MFG4, 79±58%, MFG8) and
+8 h (63±41%, MFG4, 53±17% MFG8). Formation of
microvesicles was significantly higher in the RS treated
group compared to all colloid groups (P<0.05).
Conclusion: In this porcine sepsis model, volume replacement with colloid solutions seemed to reduce the
detectable number of platelet derived microvesicles,
which could be seen after the use of ringers solution. This
may be a sign of less activation with colloids.
The changes in platelet function in SIRS, sepsis and MODS — a tight connection to the changes
in the immune and hemostatic system
U Leonhardt, M Koksch, G Röthig and L Engelmann
University Leipzig, Department of Internal Medicine I, Philipp-Rosenthal-Strasse 27, 04103 Leipzig, Germany
Background: In previous studies platelet dysfunction was
described as a part of disseminated intravascular coagulation (DIC). These findings are also important in septic
patients. In this present study, the association of the
platelet function with the systemic inflammation and the
development in different parts of the septic process in
patients of an internal intensive care unit were investigated.
Methods: Twenty-five patients fulfilling clinical, hemodynamic and blood chemistry results of a systemic inflammation were included. The investigations were done in a
two-day period. The characterization of the immune-state
with IL-6, TNF-α and Procalcitonin used standard
methods. The platelet activation marker P-selectin (GMP140) was analyzed by flow cytometric detection ex vivo
and after stimulation using 5 µM ADP and 10 µM TRAP-6.
DIC was also characterized by standard laboratory results
(platelet count, aPTT, AT III, fibrinogen, TAT, D-Dimer). The
APACHE II-score evaluated the clinical situation.
Results: The activation status of platelets was significantly
associated with the process of inflammation. Pre-activated
platelets (ex vivo) were seen in all patients with systemic
inflammation (PCT P<0.03). During the measurement over
a two-day period (five results for each patient) the pre-activation and the reagibility of TRAP-6-stimulation were significantly correlated to plasma levels of IL-6 (P<0.01) and
TNF-α (P<0.04). Furthermore the detection of changes in
platelet function started earlier then the measured results
of common laboratory tests in DIC.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Conclusion: Platelet function was tightly associated
with the process of systemic inflammation. The dysfunction of the cellular part of coagulation could be an
important marker of changes in the hemostatic system
and the development of the disseminated intravascular
Newly developed thrombocytopenia in medical intensive care patients
R Strauss, M Wehler, K Mehler, D Kreutzer, A Mueller, P Martus* and EG Hahn
Departments of Medicine I and *Medical Informatics, Biometry and Epidemiology, University Erlangen–Nuremberg,
Krankenhausstrasse 12, D-91054 Erlangen, Germany
Introduction: Thrombocytopenia is a major concern in
intensive care medicine. The incidence is greatly varying
depending on the clinical setting.
Objective: To determine the incidence, severity, prognosis
and therapeutic implications of thrombocytopenia in our
12-bed medical (noncoronary) ICU.
Methods: We evaluated, in a prospective observational
study over 13 months (1.11.1997–30.11.1998), all patients
who stayed longer than 48 h in the ICU. Thrombocytopenia
was defined as a platelet count below 150 000/µl.
Results: We studied 243 patients (63.3% male) with a
mean age of 57.9 (±16.34, 18–96) years, a mean
APACHE II score of 21.2 (±10.06, 0–49) and a mean
length of stay of 13.5 days. Thrombocytopenia was
observed frequently: only 82 patients (33.7%) were never
thrombocytopenic. Ninety-nine (40.7%) patients were
thrombocytopenic on admission. From the 144 patients
who had normal platelets on admission 62 (45.2%)
developed thrombocytopenia. In this group of patients,
ICU mortality was significantly correlated with the degree
of thrombocytopenia: 66.7% in patients with platelets
<20 000/µl, 55.6% in patients with platelets
20 000–50 000, 47.4% in patients with platelets
50 000–100 000, 9.7% in patients with platelets
100 000–150 000 (P=0.003 χ2).
Conclusion: 45.2% of patients became thrombocytopenic during their ICU stay. The degree of newly developed thrombocytopenia was highly correlated with
mortality, length of stay, initial APACHE II score and the
consumption of blood products.
Platelets during the ICU stay
Packed red blood cells
100 001–150 000
50 001–100 000
20 000–50 000
<20 000
SDAP (platelets)
on admission
Length of stay
All data mean ± standard deviation. ANOVA was used for testing.
Effects of Antithrombin III on body cavity effusions, fluid balance, colloid osmotic pressure and
hemodynamics in porcine septic shock
M Cobas Meyer, B Vangerow, J Ahrens, T Schuerholz, G Marx and H Rueckoldt
Department of Anaesthesiology, Hannover Medical School, D-30625 Hannover, Germany
Introduction: It has been suggested that Antithrombin III
(AT III) might have an anti-inflammatory effect in addition
to its well known anti-thrombotic properties. The aim of
this study was to investigate the effects of AT III on
abdominal and pleural effusions, fluid balance, colloid
osmotic pressure (COP) and hemodynamics, in a porcine
fecal peritonitis model.
Methods: This was a prospective randomized, controlled
animal laboratory study. Fasted, anaesthetized, mechanicallyventilated and multi-catheterized pigs (20.6±1.9 kg)
received 1g/ kg body weight feces into the abdominal
cavity to induce sepsis and were observed over 8 h. Before
induction, animals were randomized to treatment with 200
IE/ kg AT III (Kybernin® P, Centeon Pharma GmbH;
Marburg, Germany) (n=4) or to controls (NaCl 0.9%) (n=4).
Hemodynamic treatment scheme with ringer´s solution
was based on ensuring baseline intrathoracic blood
volume (ITBV). Hemodynamics and COP were measured
before sepsis induction and 4 h and 8 h afterwards.
Cumulative fluid balance was calculated after 4 and 8 h.
Abdominal and pleural effusion volumes were obtained
after sacrificing the animals.
Results: See Table.
+8 h
Abdominal effusion [ml]
Pleural effusion [ml]
COP [mmHg]
+4 h
CVP [mmHg]
PAWP [mmHg]
MAD [mmHg]
Cumulative fluid balance [l]
ITBV [ml]
(*) P<0.05; difference between the groups (Student’s t-test with Bonferroni Holm correction).
Conclusion: In the AT III treated group both abdominal
and pleural effusion volumes were lower than in the controls. COP values 4 h after sepsis induction were higher
and the cumulative fluid balance lower. This may indicate
that AT III has a positive effect on capillary leakage in
septic shock.
Antithrombin (AT) improves inflammation induced microcirculatory disturbance in rat mesentery
B Leithäuser, S Lendemans, J Schumacher, H Tillmanns and FR Matthias
Medical Clinic I, Department of Cardiology–Angiology, Justus-Liebig-University, Klinikstrasse 36, 35385 Giessen, Germany
Background: Impairment of the microcirculation is one of
the main causes for multiple organ failure (MOF) in critical
care patients. Trials with AT substitution in patients with
sepsis showed a positive effect on the development and
course of MOF. Recent experimental studies indicate that
there might be specific AT-effects independent of the
coagulation cascade, which can modulate endothelial–
leukocyte interaction and vascular permeability.
Methods: The effect of AT substitution on LPS-induced
microvascular leakage (ML) and leukocyte adhesion (LA) in
the rat mesentery was investigated by means of intravital
microscopy. Male CD rats were infused with 0.5 mg/kg LPS
(E. coli) over 80 min. Vascular leakage was detected with
FITC-marked rat serum albumin by fluorescence microscopy
and evaluated by grey-value analysis with image processing
software. Light microscopy was used to evaluate leukocytes
adherence to the vessel wall. Two treated groups received
500 U/kg AT either 20 min prior to or 20 min after the beginP26
ning of LPS-infusion. Animals not infused with LPS, either
untreated or treated with placebo (albumin), served as controls. One pre-treated group additionally received heparin at
a clinically used dosage. Furthermore, interleukin-2 (IL-2)
was used to induce ML and treated with AT prior to infusion.
Results: LPS-infusion led to a significant increase of ML
and LA compared to controls. Both effects were reduced
to the level of controls by the substitution of AT. No significant differences were found between the pre-treated or
the post-treated group. IL-2 induced ML was also reduced
by treatment with AT.
Conclusion: Substitution of AT, even when given after
LPS, ameliorates vascular leakage and leukocyte adhesion
to the vessel wall. Together, these effects improve flow
conditions in the microcirculation. Hence IL-2-induced
leakage could also be modulated by AT and a direct effect
by AT on leukocyte adhesion seems to be likely.
G-protein- and phosphodiesterase-dependent regulation of neutrophil migration by antithrombin
III involving CXC-receptor 1
S Dunzendorfer, A Rabensteiner, N Kaneider, J Römisch* and CJ Wiedermann
Department of Internal Medicine, University of Innsbruck, Innsbruck, Austria; *Centeon Pharma GmbH, Research, Marburg, Germany
Background: Antithrombin III (ATIII) is a serpin with a
newly uncovered regulatory role in neutrophil (PMN)
chemotaxis. ATIII and platelet factor-4 (PF4) differentially
inhibited migration of PMN toward the CXC-chemokines,
interleukin-8 (IL-8) and GRO-α, and the classical attractants, fMet–Leu–Phe (fMLP) and C5a.
Aim: To determine signalling events and involvement of
CXC-receptors (CXCR) 1 and 2 in ATIII-induced inhibition
of directed PMN migration, we studied the in vitro effects
of ATIII in the presence of specific CXCR1 and CXCR2
antibodies (mAb), pertussis toxin or various blockers of
signalling enzymes of PMN.
Methods: As in the absence of other attractants, highly
purified ATIII has been shown to itself induce PMN migration, effects were determined in modified Boyden chemotaxis chambers by the micropore-filter leading front assay
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
in a 48-well system, after pretreatment of PMN with antibodies, pertussis toxin or enzymes blockers.
Results: Preincubation of PMN with effective concentrations of pertussis toxin, staurosporine or 3-isobutyl-1methylxantine completely blocked ATIII-induced migration
whereas treatment with bisindolylmaleimide (GFX), a
selective protein kinase C blocker, wortmannin or tyrphostin-23, had no effect. In assays of IL-8-induced migration, PMN responses to the enzyme blockers were
comparable, whereas PF4-induced responses differed as
they were also sensitive to GFX. Migration of PMN toward
ATIII was significantly reduced by pre-treatment of cells
with CXCR1 but not CXCR2 mAb, whereas migration
toward IL-8 was antagonised by both CXCR1 and
CXCR2 mAb. The mAbs had no effect on fMPL-induced
PMN migration. A mAb to IL-1R, which was used as a
control, was inactive as well.
Conclusion: Effects of ATIII on PMN migration appear to
involve specific signalling pathways including CXCR1, Gproteins and phosphodiesterase and staurosporine-sensitive enzymes other than protein kinase C (i.e. protein
kinase A or phospholipase D).
Inhibition of neutrophil migration by the serpin antithrombin III
S Dunzendorfer, A Rabensteiner, C Reinisch, J Römisch* and CJ Wiedermann
Department of Internal Medicine, University of Innsbruck, Innsbruck, Austria; *Centeon Pharma GmbH, Research, Marburg, Germany
Background: Beyond the well-known regulatory role of
antithrombin III (ATIII) in haemostasis, anti-inflammatory
effects have been described such as the induction of
prostacyclin release by endothelial cells or the reduction
of tissue accumulation of neutrophils (PMN) in animal
models of ischemia-reperfusion and sepsis.
Aim: To determine a role of ATIII beyond its ability to
reduce leukocyte-mediated events induced by thrombin,
we studied direct effects on the in vitro migration and respiratory burst activity of PMN by the ATIII concentrate
Kybernin®P, the monoclonal antibody-purified ATIII thereof
and its α and β isoforms.
Methods: PMN of healthy donors were isolated by lymphoprep® followed by hypotonic lysis of contaminating
erythrocytes. Chemotactic activity was determined in modified Boyden chambers by the micropore-filter leading
front assay in a 48-well system. Respiratory burst activity
of PMN was measured fluorimetrically.
Results: Pre-incubation in vitro of human PMN with
Kybernin®P, immune-adsorbed ATIII or both isoforms significantly inhibited the migration toward fMet–Leu–Phe (fMLP),
interleukin-8 (IL-8), and ATIII itself, in a concentration-dependent manner following a bell-shaped curve. Maximum ATIII
effects were seen at a concentration of 1 U/ml, which is the
level of ATIII found in normal human plasma. Purified ATIII
did not deactivate PMN chemotaxis toward GRO-α,
whereas fMLP, IL-8, platelet factor-4 (PF4) and GRO-α
itself did. Checkerboard analyses indicate that the inhibition
of migration was not due to a general impairment of the
chemotactic activity of the cells but most probably involved
receptor regulatory processes. Deactivation of chemotaxis
toward IL-8 but not GRO-α suggests differential interaction
at the CXC receptors 1 and 2. At comparable concentrations purified ATIII and Kybernin®P did not affect resting or
stimulated respiratory burst activity of PMN.
Conclusion: Our results suggest that ATIII may be a physiologic regulator of the acute inflammatory response by
protecting cells from premature activation.
Lipopolysaccharide induced procoagulant activity and cytokine production: influence of
E Gray, P Souter, J Römisch* and S Poole
NIBSC, Blanche Lane, South Mimms, Potters Bar, Herts EN6 3QG, UK; *Centeon Pharma GmbH, 35002 Marburg, Germany
Sepsis-induced disseminated intravascular coagulation is
usually caused by endotoxin evoked production of cellular
procoagulant tissue factor (TF) and pro-inflammatory
cytokines, which further perpetuate the generation of TF.
Large doses of antithrombin (AT) reduce mortality and morbidity in septic patients and there is increasing evidence to
suggest that AT has anti-inflammatory properties in addition to its anticoagulant properties. In the present study, we
investigated the effect of AT in three in vitro cellular
systems: citrated whole blood, human umbilical vein
endothelial cells (HUVECs) and mononuclear cells
(MNCs). The cells were stimulated with lipopolysaccharide
(LPS) for 4–6 h in the presence and absence of AT. Procoagulant tissue factor (TF) activity was estimated by a TFdependent clotting or chromogenic assay and interleukin-6
(IL-6) was measured by ELISA. In all three systems, 5–40
IU/ml AT was found to inhibit TF and IL-6 production in a
dose-dependent manner. This inhibitory effect was not
attributable to excipients or co-purified components of AT.
Experiments with chemically modified AT and a low heparin
binding fraction of AT indicated that binding to heparin
and/or cell surface glycosaminoglycans is important for the
inhibitory activity. Up to 40 µM of a specific thrombin
inhibitor, r-hirudin, did not inhibit the production of TF or IL-
6 in either of three cellular systems, suggesting that inhibition of thrombin might not be the main mechanism by
which AT prevents the production of TF and IL-6. The
results of this study have shown that, apart from the inhibiP29
tion of thrombin and other activated coagulation factors, AT
may also downregulate the cellular expression of proinflammatory responses and therefore may have an added
value in the treatment of sepsis-induced DIC.
Effects of antithrombin III (ATIII) treatment (high dose) in severe pre-eclampsia and HELLP
syndrome with alterations of coagulation inhibitors and inflammatory markers: a preliminary report
A Giarratano, G Cuccio* and S Mangione†
Istituto Materno Infantile, †Cattedra di Anestesia e Rianimazione, *Cattedra di Terapia Intensiva, Università degli Studi di
Palermo, via C Rampolla 1, 90142-Palermo, Italy
Objective: Several investigations and our recent experience indicate that the intravascular inflammatory response
and the clotting alterations registered during severe preeclampsia and HELLP syndrome are not an epiphenomenon but the cause of the clinical syndrome. The aims of this
study were to investigate the effects of the ATIII substitutive treatment on cytokine plasma concentrations, assumed
as a marker of endothelial damage and of the associated
systemic inflammatory response. A secondary objective
was to correlate the ATIII treatment with the evolution of
both single or multi-organ dysfunction syndrome (MODS).
Materials and methods: The study involved four patients
with severe pre-eclampsia and three HELLP syndrome
patients. Diagnostic criteria for severe pre-eclampsia
were those published by the American College of Obstetricians and Gynecologists (diastolic blood pressure
>110 mmHg and proteinuria ≥0.5g/l). Diagnostic criteria
for HELLP syndrome required, in association with hypertension and proteinuria, thrombocytopenia (<150 000
cells/µl), evidence of hepatic dysfunction (aspartate
aminotransferase [AST] and alanine aminotranserase
[ALT] levels >40 IU/L with lactate dehydrogenase [LDH]
level of >600 IU/l) and evidence of hemolysis (increased
LDH and anemia). Plasma levels of tumor necrosis factor
(TNFα and interleukins [IL1β and IL6]) were measured by
enzyme-linked immunoadsorbent assay (ELISA). Plasma
concentrations of Antithrombin III (ATIII) and Protein C (PC) were measured by a chromogenic assay. ATIII was
administered after the first sampling (admission) by a
loading dose of 3.000 IU in bolus infusion and a maintenance dose of 1500 U/12 h over 4 days. Results were
compared using the Mann-Whitney test. Maternal parameters and clinical data were compared using unpaired
Student t-test.
Results and discussion: Clinical data are listed in Table 1
and results are shown in Table 2. These results demonstrate significant changes in fibronectin and ATIII concentrations between levels at admission compared with levels
after ATIII treatment. It is very interesting to register a
decrease in blood levels of cytokines and a reduced proinflammatory activity. The pro-inflammatory activity was
increased in women with severe pre-eclampsia and HELLP
syndrome patients and this result could also be explained
by specific effects which are independent from the coagulation cascade (anti-inflammatory actions).
Conclusions: The preliminary results of this study confirm
our hypothesis as well as the utility of marker monitoring
and substitutive treatment. We believe that it is necessary
to develop a large study (as a phase III trial) to confirm our
hypothesis and to achieve other significant results.
Redman CWG. et al.: Am J Obstet Gynecol 1999, 180:499–506.
Okajima K et al.: Thromb Res 1998, 24:27–32.
Souter et al.: Int Care Med 1999,19
Table 1
Clinical data
Diastolic pressure
AST/ALT (IU) (Admission)
AST/ALT (IU) (Peak)
Proteinuria (dipstick)
Gestational age at admission (w)
Gestational age at delivery (w)
Data were presented as mean(±sd).
Table 2
Markers and cytokines levels
Levels at Admission
Levels after
ATIII treatment (n=7)
IL1β (pg/mL)
IL6 (pg/mL)
TNFα (pg/mL)
91** (78–96)
Fibronectin (µg/ml)
AT III (%)
IL=Interleukin; TNFα=tumor necrosis factor α; AT III=antithrombin III;
PC=Protein C. Data are presented as median concentration
(range).*P<0.01 compared with levels at admission (Mann Whitney U
test).**P<0.001 compared with levels at admission (Mann Whitney U
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Decreased Protein C, Protein S and Antithrombin III Levels are predictive of poor outcome in
Gram-negative sepsis caused by Burkholderia pseudomallei
SP Larosa*, SM Opal†, B Utterback*, B Yan*, J Helterbrand*, AJ Simpson‡, N White‡ and CJ Fisher*
*Lilly Research Laboratories, Indianapolis, IN, USA; †Brown University, Providence, RI, USA; ‡Mahidol University, Bangkok, Thailand
The acute septicemic form of Burkholderia pseudomallei
infection or Melioidosis is associated with substantial
release of endotoxin, TNF-α and IL-1. This inflammatory
response leads to endothelial injury, activation of the
extrinsic coagulation cascade, depletion of naturally occurring anticoagulants, microvascular thrombosis, organ
failure and death. Plasma samples drawn at baseline and
at time points during the illness from 30 patients with
Melioidosis were assayed for D-dimer levels, Protein C
and Protein S antigen levels and Antithrombin III functional
activities. Results of samples drawn during the illness
were averaged for each patient. Baseline and continued
deficiencies of Protein C, Protein S and Antithrombin III
were predictive of poor outcome in a statistically significant fashion by logistic regression.
Endothelial injury as a result of inflammatory response to
Burkholderia pseudomallei infection leads to coagulopathy and depletion of the natural anticoagulants Protein C,
Protein S and Antithrombin III. Early and continued deficiency of these factors is predictive of poor outcome.
Replacement therapy of depleted factors to achieve
normal levels may be a worthwhile strategy for patients
with Gram-negative sepsis.
*±std Assay
Protein C (%) ag
nl. (74–140)
Protein S (%) ag
nl. (55–116)
Anti-thrombin III (%)
nl. (84–124)
Hypercoagulability indicated by elevated blood TFPI (tissue factor pathway inhibitor) levels is
closely related to severity of septic patients with glucose intolerance
M Hoshino, Y Haraguchi*, M Sakai, K Hayashi, N Horita, N Miyayama, H Saegusa and H Ohsawa
Department of Intensive and Critical Care Medicine, Tokyo Police Hospital, Fujimi 2-10-41, Chiyoda-ku 102-8161, Tokyo,
Japan; *Tokyo Disaster Medical Center, Tokyo, Japan
Background and purpose: Hypercoagulability and
endothelial cell activation and/or injury are mutually related
and often found in acutely ill septic patients, and recently
they have been reported to be related to multiple organ
dysfunction syndromes (MODS).It is not clear, however,
which parameters indicating coagulopathy are most
closely related to MODS. In this report, we analyze correlations among the severity of the illness including MODS
and parameters related to coagulopathy including TFPI in
acutely ill septic patients, in order not only to ascertain
their close relationships but also to find sensitive and predictive markers of the severity of septic patients.
Materials and methods: Five acutely ill septic patients
with glucose intolerance were analyzed. Their blood
glucose levels were strictly controlled by means of a
bedside-type artificial pancreas (AP), STG-22, manufactured by NIKKISOH corporation in Japan. We selected
septic patients in whom blood glucose levels were strictly
controlled with STG-22, because some parameters
related to coagulopathy, including Plasminogen Activator
Inhibitor-1 (PAI-1), are influenced by blood glucose levels.
Analyzed items were: (1) regarding the severity of the
illness, MOF score as the indicator of the degree of
MODS (calculated from the MOF criteria of Japanese
Association for Critical Care Medicine), blood thrombomodulin (TM) levels as the indicator of endothelial cell
injury; (2) regarding coagulopathy, disseminated intravascular coagulation (DIC) score (calculated from the DIC criteria of the Ministry and Welfare of Japan), platelet count
(PLT), fibrinogen, FDP, prothrombin time, TAT, D-dimer,
PIC, plasminogen (PLG), Antithrombin (AT), Protein C
activity, protein S activity, total PAI-1, tissue plasminogen
activator (tPA)-PAI-1 complex, and total TFPI; (3) regarding glucose tolerance, daily mean blood glucose level
(BSm), M value (measured by the euglycemic hyperinsulinemic glucose clamp method with AP. The clamped
blood glucose level was 80 mg/dl, with the insulin infusion
rate of 1.12 mU/kg.min normal range 5–10 mg/kg.min).
Results: (1) Mean of M value and BSm measured within a
few days of admission (and during ICU stay) were 4.7±2.7
mg/kg.min (n=4), 176±10.7 mg/dl (n=5) [5.7±3.1
mg/kg.min (n=12), 183.4±19.4 mg/dl (n=65)], respectively. (2) There was strong positive relationship between
MOF score and TM (correlation coefficient r=0.95, n=32,
P<0.005). (3) MOF score was correlated with TFPI
(r=0.91, n=30), PLT (0.85, n=78), DIC score (0.84,
n=78), tPA-PAI (0.76, n=33), AT (0.66, n=32), PLG
(0.65, n=33), and PIC (0.53, n=33), (P<0.005).(4) TM
was correlated with TFPI (0.89, n=29), tPA-PAI (0.79,
n=32), PLT (0.77, n=32), DIC score (0.75, n=32), PLG
(0.63, n=32), PIC (0.58, n=32), (P<0.005). (5) DIC score
was correlated with PLT (0.73, n=78), TFPI (0.72, n=30),
tPA-PAI (0.69, n=33), AT (0.61, n=32), (P<0.005). (6)
TFPI was correlated with tPA-PAI (0.76, n=30), PLT
(0.70, n=30), PIC (0.50, n=30), (P<0.005).
Interpretation and conclusions: The severity of disease
in acutely ill septic patients, indicated by the progression
of MODS and endothelial cell injury, was closely related to
coagulopathy characterized by hypercoagulability with
decreased fibrinolysis. Treatment for hypercoagulability
seemed to be justified in severely septic patients.
Although the mechanism was unclear, TFPI, which
seemed to be a parameter of hypercoagulability and
endothelial cell activation, was thought to be a sensitive
and possibly predictive marker of the disease severity in
septic patients.
Endotoxin-induced capillary perfusion failure and leukocyte adhesion is not prevented by the
thrombin inhibitor hirudin: an intravital microscopic study
JN Hoffmann*, B Vollmar†, D Inthorn*, FW Schildberg* and MD Menger†
*Department of Surgery Klinikum Großhadern, Ludwig-Maximilians-University Munich, 81377 München, Germany; †Department
of Clinical and Experimental Surgery, University of Homburg/Saar, 66421 Homburg, Germany
Besides its central role in coagulatory pathways, thrombin
is thought to be a key mediator of macrophage and granulocyte activation. During recent years the concept of coagulatory inhibition by the specific thrombin inhibitor hirudin
has been established to treat septic disorders. Since
basic mechanisms of sepsis include leukocyte–endothelial
cell interaction as well as deterioration of capillary perfusion, we hypothesised that hirudin would modulate leukocyte activation and microvascular injury. In a skinfold
preparation of Syrian hamsters, severe endotoxemia was
induced by intravenous administration of endotoxin (LPS,
E. coli, 2 mg/kg). Hirudin (0.25 mg/kg) was substituted
intravenously during 4h after the induction of endotoxemia
[n=7, hirudin group (hirudin)]. In control animals [n=6,
Control group (control)], LPS was given without hirudin
substitution. By intravital fluorescence microscopy leukocyte–endothelial cell interaction and functional capillary
density (FCD, a measure of capillary perfusion) were
analysed during a 24 h period after the LPS injection.
Hirudin effectively normalised thromboplastin time and
antithrombin activity when compared to controls (P<0.05,
ANOVA). However, hirudin did not attenuate LPS-induced
arteriolar and venular leukocyte adherence during 24 h
(P>0.05, MANOVA). In parallel, hirudin caused significant
deterioration in FCD over time when compared to controls
(hirudin: baseline=171±19 vs 16±9 at 24 h; control:
baseline =150±20 vs 62±18 at 24 h; P<0.05 MANOVA).
The decrease of FCD in hirudin animals was associated
with a significant increase of wet-to-dry weight ratios in
different organs (lung, kidney, muscle, and small intestine;
P<0.05 vs control, ANOVA). Thus, our experiment does
not indicate a protective effect of hirudin on microcirculation during endotoxemia, despite an improvement of coagulatory parameters. This result may at least in part explain
the lack of efficacy of hirudin efficacy on lethality during
endotoxemia and sepsis.
Heparin vs recombinant hirudin for anticoagulation in continuous renal replacement therapy
O Vargas Hein†, C v Heymann†, J Nissen†, M Lips†, M Welte*, WJ Kox† and C Spies†
Departments of Anesthesiology and Intensive Care, *University Hospital Charite, Campus Mitte, Humboldt University Berlin,
Schumannstr. 20/21, 10117 Berlin, Germany; †Benjamin Franklin Hospital, Free University Berlin, Hindenburgdamm 30, 12200
Berlin, Germany
Introduction: Heparin as standard for anticoagulation in
continuous renal replacement therapy (CRRT) has a
bleeding incidence of 20% and is contraindicated in
patients with heparin induced thrombocytopenia II [1].
Recently recombinant hirudin (rhirudin) has been reported
to be superior as an anticoagulant to heparin for intermittent hemodialysis [2]. The aim of the study was to
compare heparin and rhirudin as anticoagulants in CRRT
regarding hemofiltration efficacy and possible bleeding
Methods: After ethical committee approval and written
informed consent from the relatives 16 critically ill patients
with an indication for CRRT were randomly allocated to 2
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
groups: heparin (9 patients) initially 250 IU/h; target activated clotting time (ACT) 180–210 s, 125 IU/h stepwise
heparin dose change; hirudin (8 patients) initially 10
µg/kg/h; target ecarin clotting time (ECT) 80–100 s, 2
µg/kg stepwise dose change. Every four h, thrombin time
(PT), partial thromboplastin time (PTT), hemoglobin (Hb)
and thrombocytes were determined. A bleeding complication was defined as an Hb decrease >2 g/dl. The observation time was 96 h. The ocurrence of filter clottings were
Statistical analysis: Mann-Whitney-U-Test, P ≤0.05.
Results: Basic patient characteristics did not differ
between groups.
Three bleeding complications were observed in three
patients in the rhirudin group, but none were seen in the
heparin group. The bleedings were observed 60 h after
study initiation. Platelets were 76, 14 and 62/nl at bleeding time. The ECT was 83, 65 and 85 s respectively. PTT
was higher than 60 s and PT was lower than 60 s at this
point in all three patients. rhirudin application was interP34
Filter clotting and run time; median (range)
Patients with filter clotting (n)
Filter clottings/filter changes (n)
Filter run time (h)
12 (4–36)
20 (4–56)
rupted 4 and 8 h before bleeding time point in two
patients, the third patient received 5 µg/kg/h rhirudin.
Conclusion: rhirudin prevented filter clotting more effectively than heparin. Low platelet count together with pathological PT and PTT, despite aimed or low ECT in patients
with bleeding complications, indicated that improved clotting monitoring and dose adjustment must be implemented before rhirudin can be used safely.
Vanholder R et al.: Kidney International 1994, 45:1754–1759.
vWyk V et al: Kidney International 1995, 48:1338–1343.
Using fraxiparin in patients with gestosis
DV Elioutine and DV Marshalov
Saratov State Medical University, B. Kazachia 112, Saratov, 410026, Russia
Introduction: The problem of treating the coagulate disorders in cases of gestosis remains of topical interest in
obstetrics; heparin is dangerous because of the possible
development of thrombocytopenia, heparin-induced thrombosis, osteoporosis, and increased consumption of antithrombin III.
Methods: The investigation was carried out in 2 groups of
pregnant women with gestosis. To the first group (consisting of 34 patients aged 18–32), the injections of fraxiparin
(2850 ME) were given once a day. To the second group
(consisting of 32 patients aged 19–30), the injections of
heparin (5000) were given 4 times a day. The effectiveness of the drug was checked clinically and according to
the investigations of coagulate and thrombocyte gemostaP35
sis the day before the operation and on the 1st, 4th and
7th days after the operation.
Results: In the group of patients with gestosis who
received heparin, reduction of the number of thrombocytes to 21.4% (P<0.01) and growth of their aggregation
activity to 18.3% (P<0.05) were observed on the 7th day
of the postoperative period in comparison with the period
prior to operation. In the group of pregnant women who
received fraxiparin, a change in the number of thrombocytes and an increase in their aggregation activity were
not observed.
Conclusion: The use of fraxiparin is more effective than
treating with heparin and, as a result, it can be used for prophylaxis and treatment of thrombosis in pregnant women.
The effect of using a heparin-free flush system for central venous and pulmonary artery catheters
on a general medical and surgical intensive care unit
C Bradley and P Munro
ITU, St George’s Hospital, Blackshaw Road, London, SW17 OQT, UK
Introduction: Heparin flush systems prevent thrombus formation on invasive monitoring devices. Unfractionated
heparin increases the risk of thrombocytopenia, and drug
interactions in lines. Low dose heparin, given to most
patients on the unit to prevent deep vein thrombosis (DVT),
decreases thrombus formation on central venous catheters
(CVC). Pulmonary artery catheters (PAC) can be maintained with non-heparinised solutions but evidence for arterial lines is conflicting. Due to the high incidence of
thrombocytopenia in our patients and the routine use of
low molecular weight heparin to prevent DVT, we decided
to investigate the use of a heparin-free flush system.
Method: The study took place over two 4 week periods:
one of heparin-free flush and one using a heparin flush. All
patients on the unit were recruited into the study. Line
care was according to unit practice. Duration of catheter
patency and complications were recorded. Results were
analysed using Chi squared and Fisher exact 2 tailed
The numbers of thrombus associated problems were:
Arterial Catheter
With Heparin
Heparin Free
There was no statistical difference in the thrombus-associated complication rates between the two groups.
Results: The numbers of catheters in each group were:
With Heparin
Heparin Free
Discussion: This study demonstrated that the removal of
heparin from the flush systems of invasive monitoring lines
does not lead to an increase in thrombus-associated complications, prevents the exposure to the risks of unfractionated heparin, which include thrombocytopenia and drug
interactions, thereby improving patient care.
Is the ecarin clotting time (ECT) a valid monitoring parameter for r-hirudin-based anticoagulation
in continuous renal replacement therapy?
C v Heymann*, O Vargas Hein*, S Ziemer†, P Sinha†, T Schröder*, J Nissen*, M Lips* and C Spies*
*Department of Anaesthesiology and Intensive Care, †Institute of Laboratory Medicine and Pathological Biochemistry, University
Hospital Charité, Charité Campus Mitte, Schumannstrasse 20-21, D-10117 Berlin, Germany
Introduction: For continuous renal replacement therapy,
systemic anticoagulation is required. Besides heparin, the
current anticoagulant of choice, the direct acting thrombin-inhibitor r-hirudin was reported to be of advantage in
chronic intermittent haemodialysis due to reduced filter
clotting. The ecarin clotting time (ECT) was suggested to
be a valid monitoring parameter with a linear correlation to
r-hirudin concentrations in plasma [1]. To our knowledge,
no results of hirudin-based anticoagulation during continuous renal replacement therapy in intensive care have been
published yet. The objective of our study was to evaluate
whether the ECT correlated to r-hirudin concentrations in
plasma in this setting.
ECT [sec]
Arterial Catheter
Methods: Seven patients on continuous veno-venous
haemofiltration, due to acute renal failure, were included in
this prospective study. Anticoagulation was titrated to
achieve ECT-values between 80–100 s and parameters
(ECT, PT, aPTT, TT, Fibrinogen, AT III, platelet count, ACT)
were monitored by four-hourly determinations. R-hirudin
concentrations in plasma were measured 24 and 48 h
after starting therapy. Statistical analysis was performed
by Spearman correlation.
Results: 5 of 14 measurements of hirudin-concentrations
in plasma were found to be below the cutoff point of the rhirudin test (<0.1 µg/ml). Six measurements were
between 0.1 and 0.2 µg/ml and three results >0.3 µg/ml.
For all measurements of our study we found a weak but
statistically significant correlation between ECT and rhirudin levels (r=0.52; P=0.048). Within the narrow range
Hirudin [µg/ml]
of r-hirudin levels between 0.1 and 0.2 µg/ml, applying to
11 of 14 test results, a broad range of ECT values
between 60 and 105 s. were found.
Conclusion: During CRRT very low plasma levels of rhirudin ranging from <0.1–0.35 µg/ml were observed. For
low plasma r-hirudin concentrations up to 0.2 µg/ml,
including the majority of the investigated patients, the ECT
may not be reliable for guiding anticoagulation during
CRRT. It is not clear whether this is related to methodological problems due to laboratory determinations or to
interference with pathological conditions of other coagula-
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
tion disorders. However, a more valid monitoring for rhirudin based anticoagulation in continuous renal replacement therapy is required.
Nowak G et al.: Seminars in Thrombosis and Hemostasis 1997,
Impact of CVVH on monocyte activation in the critically ill
R Rokyta, M Holub*, M Matejovic, J Hanzlikova, I Novak, P Hora, K Balihar and V Sramek
ICU, Medical Department I, Charles University Hospital, Pilsen; *Department of Infectious Diseases, University Hospital Bulovka,
Introduction: Monocyte activation (HLA-DR+ CD14+) was
suggested to be a useful marker of immune system function. As extracorporeal circuit might influence the immune
response (e.g. bioincompatibility and/or elimination of
pro/anti-inflammatory cytokines) we investigated the
impact of CVVH initiation on monocyte activation in critically ill patients.
Figure 1
P = 0.07
non-septic (n = 7)
septic (n = 13)
Methods: Fifteen septic (4 ICU survivors) and 7 nonseptic
(1 ICU survivor) critically ill patients in whom CVVH (ultrafiltration 1–2 l h-1) was initiated because of acute renal
failure as a part of MODS. Expression of HLA-DR+ CD14+
was assessed by flow-cytometry before CVVH initiation
and at 4 h and 24 h afterwards. The Friedmann two-way
ANOVA and Mann-Whitney U tests were used when
appropriate. Data are presented as means ±SD; P<0.05
was considered significant.
Results: Baseline HLA-DR+CD14+ values in septic and nonseptic patients were not different (57±21% and 71±21%
respectively, P=0.23). Time course of HLA-DR+CD14+
during the first 24h of CVVH are shown in Fig. 1.
P = 0.15
4 hrs
24 hrs
Conclusion: CVVH initiation has no impact on monocyte
HLA-DR expression in both septic and nonseptic patients.
Acknowledgement: Supported by a Grant of IGA MZ (4972-3).
Effect of continuous hemofiltration (CVVHF) on inflammatory parameters
KH Smolle, W Wonisch, F Tatzber, P Kaufmann, G Koschsorur and R Aigner
Introduction: Hemofiltration and hemodialysis are very
important life saving procedures in the management of
critically ill patients. The ability of these techniques to
remove low molecular weight toxic substances from the
circulation is very well documented, but little is known
about their influence on inflammatory or lipid peroxidation
parameters. Procalcitonin, PMN-elastase and Neopterin
can be used as diagnostic and prognostic parameters to
monitor critically ill patients and to control therapeutic
interventions. For these reasons the aim of this preliminary
investigation was to evaluate the effect of continuous
veno-venous hemofiltration (CVVHF) versus intermittent
hemodialysis (IHD) on plasma concentrations of inflammatory (Neopterin, PMN-elastase, Procalcitonin) and lipid
peroxidation parameters (Malonic dialdehyde, Peroxides,
Antibodies to oxidised LDL).
Patients and methods: We investigated 20 consecutive
mechanically-ventilated critically ill patients (n=12, 25–86
years, 7 male, 5 female) under CVVHF (blood flow
100–150 ml/min, ultrafiltrate 1–2 l/h). Cellulose triacetate
membranes (FB-21U Biorena) (n=3) and polyacrilonitrile
membranes (AN69 Hospal) (n=9) were used in this group.
In the IHD group (n=8, 31–86 years, 5 male, 3 female)
(blood flow 200–250 ml/min, dialysate flow 500 ml/min)
polysulfone membranes (F8HPS Fresenius) were applied.
Blood samples were drawn at 0, 10, 60 and 240 min after
starting the procedure from before and after the membrane. Serum and plasma samples were drawn at 0, 10, 60
and 240 min after the onset of CVVHF or IHD. After 30 min
at 4°C, the blood samples were centrifuged at 3000 × g.
Malonic dialdehyde (MDA) determination was done by
HPLC according to the method of Wong et al Neopterin
(NPT), Elastase, Procalcitonin (PCT) and antibodies to oxidised LDL (oLAb) were determined by commercially available ELISA methods. Determination of peroxides was
performed by an enzymatic method based on the peroxidase reaction with tetramethylbenzidine as a chromogenic
Table 1
Concentrations of inflammatory and lipid peroxidation parameters in CVVHF and IHD patients
0 pre
0 post
10 pre
10 post
60 pre
60 post
240 pre
240 post
CVVHF (n=10)
IHD (n=5)
1, Median (No Gauss distribution of data); 2, Values >50 eliminated; 3, >10 and <500 included. Pre, Concentration before the membrane; Post,
Concentration after the membrane.
Statistical analysis: In the case of a Gaussian distribution
the statistical analysis was done by the t-test. Other distributions were assayed by the χ2 test.
Results: PCT and oLAb remained more or less constant in
all observations. No statistically significant differences were
observed before and after membrane passage or between
the CVVHF and the IHD group. PMN-elastase concentrations increased constantly from start to the end of CVVHF
and IHD. MDA levels reached the maximum 10 min after
the onset of CVVHF and IHD. Again lower concentrations
were observed in the IHD group compared with CVVHF.
Peroxides were not detectable in most of the samples of
the CVVHF group. In the IHD group, there were statistically
significant differences in peroxide concentrations before
and after the membrane passage. Neopterin concentrations decreased significantly after membrane passage in
both CVVHF and IHD group. The decrease was more proP39
nounced in the IHD group. Details in data for all parameters
in both groups are presented in Table 1.
Discussion: For inflammation parameters (oLAb, NPT,
PCT) no significant increases in concentrations could be
observed. This indicates that no pronounced inflammatory
reactions arise due to the dialysis procedure. The
decrease in Neopterin concentrations after membrane
passage shows efficient clearance of low molecular
weight substances during CVVHF and IHD. In contrast,
PMN-elastase concentrations increased from start to the
end of CVVHF and IHD thus indicating PMN activation
rather than activation of macrophages. In case of lipid peroxidation parameters, we found a peak in MDA concentrations 10 min after onset of CVVHF and IHD, which might
be a consequence of PMN activation after contact with
the membranes. In conclusion, our data suggest that even
beneficial procedures like CVVHF or IHD may activate
inflammatory and lipid peroxidation processes.
Prospective randomized study of hemodialysis membrane biocompatibility in acute renal failure
A Jörres*, GM Gahl*, C Dobis*, A Kahl*, R Schindler*, U Frei*, C Guenther†, C Grossmann‡, W Gaus‡ and J Hoegel‡
*Nephrology and Medical Intensive Care, UK Charité, Campus Virchow-Klinikum, Augustenburger Platz 1, D-13353 Berlin,
Germany; †Membrana GmbH, Wuppertal; ‡Biometry and Medical Statistics, University of Ulm, Germany
Introduction: There is considerable controversy in the
current literature as to whether HD membrane biocompatibility may influence the mortality of patients with acute
renal failure (ARF). We performed a prospective randomized (central telephone randomization) study in patients
with dialysis-dependent ARF treated either with Cuprophan® (CUPRO) or polymethylmethacrylate (PMMA) lowflux membranes. The data presented are the unicentric
evaluation of a larger international multicenter trial that was
published recently [1].
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Methods: In the study center in Berlin, a total of 104
patients were randomized. Of these, 94 patients
(35f/59m) with ARF were evaluable (CUPRO 45, PMMA
49 patients). Forty-six (49%) of these were
surgical/trauma patients. Mean age was 63.6 (19–87)
years, mean APACHE II score was 26.3 (11–42). Sixtyseven (71%) patients required mechanical ventilation, 81
(86%) received parenteral nutrition.
Results: Overall, 56 patients (60%) survived, 25 (56%) in
the CUPRO group and 31 (63%) in the PMMA group
(P=n.s., Fisher’s two-sided Exact Test). The odds ratio
(OR) for not surviving for CUPRO vs PMMA was 1.38 with
a 95% CI ranging from 0.56 to 3.42. Moreover, no differP40
ence between CUPRO and PMMA was detected when
age and APACHE II score were entered as possible confounders in a logistic regression model. There was also no
difference between the two study groups regarding time on
dialysis, number of dialysis sessions required, need for
mechanical ventilation, or total parenteral nutrition.
Conclusion: In summary, this controlled, prospective randomized trial did not reveal any differences in the outcome
of dialysis-dependent ARF patients treated with CUPRO
vs PMMA dialyser membranes.
Jörres et al.: Lancet 1999, 354:1337–1341.
Carbon dioxide elimination during continuous venovenous hemodiafiltration (CVVHD)–laboratory
J Ruzicka, I Novak, R Rokyta, M Matejovic, P Hora, M Hadravsky and V Sramek
ICU and Department of Biophysics, Medical Faculty of Charles University, Pilsen, Czech Republic; Department of Biophysics,
Karlovarska 36, Pilsen 301 66, Czech Republic
Objective: In a previous experiment with crystalloid solution we have shown that during CVVH(D) CO2 losses are
proportional to the fractional ultrafiltration (UF) rate of circulating solution, with only a marginal impact of dialysis
and temperature [1]. We report here results of a similar
experiment with packed RBC.
Methods: Packed RBC were heparinized, diluted with
crystalloid solution with bicarbonate to hematocrit of 0.35
and continuously saturated in a special mixing chamber to
targeted pCO2 value of 5 kPa. After equilibration, the
mixing chamber was connected to a PRISMA monitor
(Prisma, Hospal, France) at a blood rate (Qb) of
150 ml hr-1 and different CVVH(D) settings were tested
for CO2 elimination at two temperature levels (37 and
40°C). Blood samples were drawn in triplicate at each
setting before and after the filter for blood gases analysis
(ABL 520, Radiometer, Denmark). CO2 content was calculated using standard formula [2] and CO2 removal was
expressed as pre–postfilter CO2 content difference.
Table 1
Blood temperature 37°C
Blood temperature 40°C
UF 1000 UF 3000
UF 1000 UF 3000
Dial 0
Dial 0
Dial 1000
Dial 1000
Dial 2500
Dial 2500
Dial–dialysis [ml/hr], UF–ultrafiltration [ml/hr]
5 ml/min, i.e. a negligible part of the whole body CO2 production (VCO2). A significant CO2 loss (about 15% of
VCO2) can be expected using high volume ultrafiltration
Acknowledgement: This paper was supported by GAUK, Czech Republic.
Results: CO2 removal values (in % of prefilter CO2
content) are summarised in Table 1 (means).
Conclusion: The often used CVVH setting (UF 1000 ml/hr,
Qb 150 ml/min) leads to CO2 elimination at about
Ruzicka J, Šrámek V, Novák I, Hadravský M, Rokyta R, Hora P, Matejovic M, Nalos M: Gas exchange during continuous extracorporeal
elimination techniques (CRRT). In Anaesthesia, Pain, Intensive Care
and Emergency Medicine Selected Papers. Edited by Gullo A. 1998,
Giovannini I, Chiarla C, Boldrini G, Castagneto M: Calculation of
venoarterial CO2 difference. J Appl Physiol 1993, 74:959–964.
The combination of lactate and bicarbonate buffers in continuous venovenous hemodiafiltration
and its impact on serum lactate levels and homeostasis
M Balik*, A Kazda†, M Kolar* and A Hendl†
*Department of Intensive Care and Anaesthesiology, University Hospital Kralovske Vinohrady, Prague 10, Srobarova 50, Czech
Republic; †Department of Clinical Biochemistry, Postgraduate Medical School, Faculty Policlinic, Prague, Czech Republic
Introduction: The buffer substitution of dialysis and
hemofiltration is usually performed using either lactate or
bicarbonate solutions. Bicarbonate works as a diuretic
and it has certain renal tubular protective properties. The
use of lactate containing solutions can interfere with the
monitoring of lactate as a tissue oxygenation parameter.
Methods: Twelve septic ventilated patients were treated
with continuous venovenous hemodiafiltration (CVVHDF).
They were monitored before the treatment was started and
after 24 and 48 h. The dialysis fluid contained 35 mmol/l of
Na-lactate as a buffer. The hemofiltration substitution fluid
was Ringer’s solution and 8.4% bicarbonate was infused
at the average rate of 16.0 ml/h. The dose was calculated
initially and it was adjusted later to reach positive BE and
mild hypernatremia up to 150 mmol/l. In all patients the
dose of furosemide was reduced to 1 mg/ after commencing of the treatment. Hemodynamic monitoring was
performed with transesophageal echocardiography in all
patients. Systemic oxygen delivery (DO2) was calculated,
oxygen extraction was approximated according to central
venous pO2 (cvpO2).
Results: Lactate levels were found to be significantly elevated 24 h and 48 h after commencing the treatment:
1.77±0.68 mmol/l vs 2.70±0.88 mmol/l (P<0.001) vs
2.52±1.06 mmol/l (P<0.03). No significant relationship
was found between arterial lactate and cvpO2 before the
treatment, but there was an indirect relationship between
arterial lactate and cvpO2 during CVVHDF (correl. coeff.
r=–0.5 P=0.01 n=23). The regression equation of the
estimate of arterial lactate using cvpO2 was derived:
lactate =6.9–0.81*cvpO2.Average DO2 changed little
(777.2±187.0 ml/min vs 741.5±171.8 ml/min vs
760.1±149.6 ml/min). Average sodium did not change
but the standard deviation of sodium levels decreased
(147.6±11.4 vs 148.7±6.9 vs 146.2±1.9) as well as the
standard deviation of base excess (0.83±5.29 vs
2.24±4.00 vs 1.74±3.48). Eight patients (67%) had preserved sufficient residual renal functions. The average time
of CVVHDF was 3.83 (3–5) days in six patients who survived (50%).
Conclusion: Lactate remains a valuable marker of tissue
hypoxia in patients treated with CVVHDF with constant
input of lactate by dialysis. The stabilization of serum tonicity and acid-base balance was recorded. Continuous
bicarbonate infusion during early CVVHDF can probably
reduce the dose of other diuretics and promote the residual renal functions.
Efficacy of the endotoxin absorption method (PMX) in patients with septic shock associated with
intraperitoneal infections
T Ikeda, K Ikeda, T Onizuka, H Suzuki, K Suzuki and H Maruya
Division of Critical Care and Emergency Medicine, Tokyo Medical Center of Tokyo Medical University, 1163 Tatemchi, Hachiojicity, Tokyo, 193 Japan
Many points remain to be clarified concerning the mechanism of the endotoxin absorption method (PMX). In the
present study, we investigated the efficacy of PMX in 13
patients who met all of the following four conditions: (1)
patients in septic shock associated with an intraperitoneal
infection; (2) patients who underwent SIRS; (3) patients
with endotoxin (ET) levels of at least 10 pg/ml at the start of
PMX; and (4) patients with circulatory dynamics requiring
administration of catecholamine. The 13 patients consisted
of nine men and four women with an average age of 62±10
years. The outcome after four weeks was survival of nine
patients and death of four patients. PMX was performed for
2 h and nafamosat mesylate was used as anticoagulant.
The ET value, blood pressure, WBC, pH, PaO2/FiO2,
amount of DOA used, IL-6, IL-1ra, NOx, PAI-1, thromboP43
modulin (TM), ICAM-1 and ELAM-1 were measured immediately before, immediately after and 24 h after PMX.
Results: The ET value showed a definite decrease from
29.7±23.4 to 14.9±11.2 pg/ml. The mean blood pressure
increased from 79.7±13.8 to 97.4±16.8 torr. Among the
cytokines, no significant differences were observed in IL-6
before and after PMX but IL-1ra tended to decrease. PAI1 also decreased after PMX in the same way as IL-1ra.
NOx, ICAM-1, ELAM-1 and TM showed no changes.
Conclusion: The inhibition of excessive increases in antiinflammatory cytokines by PMX was considered beneficial
for the body’s defense from the standpoint of preventing
progression to CARS.
Effects of epinephrine and norepinephrine on endotoxin-induced tissue factor expression in
blood monocytes
L de Rossi, T Scholz*, T Eckermann†, U Temmler*, S Rußwurm†, K Reinhart† and W Lösche*
Clinic of Anaesthesiology, Rheinisch-Westfälische Technische Hochschule Aachen, Pauwelsstr. 30, D-52074 Aachen,
Germany; †Clinic of Anaesthesiology and Intensive Care Therapy; and *Centre for Vascular Biology and Medicine/Erfurt,
Friedrich-Schiller-University, Bachstr. 18, D-07740 Jena, Germany
Background: Tissue factor (TF) is the most important initiator of intravascular coagulation. In monocytes the
expression of pro-inflammatory cytokines and TF is con-
trolled by the same transcription factors. Catecholamines,
frequently used in sepsis therapy of critically ill patients,
have been shown to inhibit endotoxin-induced expression
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
of monocyte cytokines, such as TNF-α or IL-6 [1]. The aim
of our study was to prove whether epinephrine (EPI) and
norepinephrine (NOREPI) may also affect TF expression.
Methods: To induce TF and TNF-α expression in monocytes we incubated citrated human whole blood (WB) for
four hours with LPS (E. coli 055:B5; 50 ng/ml). TF expression on monocytes was determined by flow cytometry,
TNF-α secretion was measured by ELISA.
Results: Incubating WB samples with 50 ng/ml LPS we
observed an increase in the number of TF-positive monocytes from 3.9±0.7% to 42.9±2.8%. 5.5 nM EPI inhibited
LPS-induced TF expression by about 16% (P=0.03), but
the inhibitory effect was attenuated at higher EPI concentrations (>550 nM). LPS-induced TNF-α secretion
(4.2±0.5 ng/ml) was inhibited by about 60% in presence
of 5.5 nM EPI, and maximum inhibition of 75% was
reached at 55 nM. Compared to EPI, higher concentrations of NOREPI were needed to get a significant inhibition of monocyte activity. Inhibition of TF-expression by
13% was observed at 550 nM and the same inhibition
was observed at 5,500 nM. TNF-α secretion was inhibited
in a clear dose-dependent manner and amounted at 5.5
nM NOREPI to about 35% and at 5500 nM to about 75%.
Conclusion: EPI is more effective than NOREPI to inhibit
LPS-induced monocyte TF expression and TNF-α secretion. At higher concentrations EPI also seemed to have a
stimulatory effect on TF expression, which could be due to
an interaction of EPI-activated platelets [2] with monocytes.
van der Poll et al.: Infect Immunol 1994, 62:2046–2050.
Amirkhosravi et al.: Thromb Haemost 1996, 75:87–95.
Influence of enoximone and dobutamine upon liver perfusion and function in fluid optimized
septic patients
H Kern, T Schröder, M Kaulfuss, WJ Kox and C Spies
Department of Anesthesiology and Intensive Care Medicine, University Hospital Charité (Campus Mitte), Humboldt-University
Berlin, Schumannstr. 20/21, 10117 Berlin, Germany
Introduction: Splanchnic mucosal perfusion abnormalities
have been implicated in the development of sepsis and
multiple organ failure. In clinical and experimental settings,
administration of dobutamine can increase hepatosplanchnic perfusion [1,2]. The administration of enoximone
during and after cardiopulmonary bypass diminished
endotoxin levels in liver venous blood indicating a beneficial effect on tissue perfusion and barrier function of the
gut [3]. The aim of this study was to investigate the impact
of dobutamine and enoximone upon liver perfusion and
function in fluid optimized septic patients.
Methods: After approval by the local ethical committee
and obtained written informed consent by next of kin, 48
septic patients were included in this study. After fluid
resuscitation according to an optimal left atrial filling pressure established by plotting left ventricular stroke work
index against pulmonary artery occlusion pressure,
patients were randomly treated with dobutamine (initially
5 µg/kg/min, increasing dose up to a maximum of
20 µg/kg/min) or enoximone (initially 2.5 µg/kg/min,
increasing dose up to a maximum of 10 µg/kg/min) for
10 h up to a dosage where no further increase in left ventricular stroke work index was achieved. Hemodynamics,
liver blood flow (LBF), hepatic tumor necrosis factor (TNF)
and monoethylglycinexylidide (MEGX) kinetics to assess
hepatic function were performed within the first 12 h of
sepsis, before administration of inotropic support as well
as 12 h and 48 h after treatment. Statistical analysis was
performed using Wilcoxon signed rank sum and Friedman
Results: Basic patient characteristics (age, sex, APACHE
III) did not differ between groups. Oxygen delivery and
oxygen consumption increased in both groups without significant difference between groups.
Conclusion: Administration of enoximone in fluid optimized septic patients may be favorable in comparison to
dobutamine. Enhanced DO2 and VO2 in both regimens
resulted in an increased regional perfusion and improved
splanchnic function only in the enoximone group indicated
12 h treatment 48 h treatment
12 h treatment 48 h treatment
LBF [L/min/m²]
3.3±0.8 b
3.2±0.7 b
3.0±0.8 b
2.9±0.8 b
LTNF [pg/mL]
43±31 ab
62±36 ab
MEGX [µg/L]
33±26 ab
29±20 b
21±22 a
32±17 b
Results are presented as means ±standard deviation; a) P<0.05 between groups; b) P<0.05 within groups compared to baseline.
by a diminished release of the pro-inflammatory hepatic
TNF and increased MEGX.
J Surg Res 65:93–100.
Am J Resp Crit Care Med 160:839–845.
Intensive Care Med 23:267–275.
Renal blood flow in critically ill cardiac patients: effect of dobutamine on total flow and selective
redistribution of the cardiac index
H Sherif, AW Andraos, A Abdel Fattah, A Amer, A Eltaweel, O Elhennawy and MS Mokhtar
Baseline regional distribution of cardiac output (CO)
depends on the physiological determinants of the endorgans. In congestive heart failure (CHF), this distribution
is modified in favour of the coronary, brain and kidney circulations, depending on the main regulatory mechanisms
of the vasomotor system, including the sympathetic
nervous system, the vascular endothelium and the local
non-endothelial mechanisms.
In this study, we measured the renal blood flow (RBF) and
CO in 20 critically ill cardiac patients (17 males, mean age
49±10 years), utilizing both techniques of color coded
duplex and echocardiography. All patients had CHF due
either to dilated cardiomyopathy (12 patients) or ischemic
heart disease (8 patients). A control group of 14 males
(mean age 41±8.6 years) was also included. Following
clinical evaluation, including the history and 12-lead ECG,
all patients and control group were subjected to duplex
assessment of RBF (ml/min/m2), in post-absorptive state
by measuring the internal diameters of both renal arteries
at their aortic origins, and then estimation of the renal
artery pulsed flow wave to measure the time average
velocity, (the average of the velocity spectrum in one
second) and then multiplied by 60 to express the flow in
one min. The CO was measured echocardiographically
through the pulsed wave Doppler trans-aortic flow. The
whole procedure was repeated, first under low dose
(5 µg/kg/min), and then under high dose (20 µg/kg/min)
dobutamine, with 24 h apart. The regional renal percentage (RRP) was calculated and expressed as percentage
of the cardiac index.
Compared to the control group, patients, at baseline,
showed statistically significant lower CO, RBF and RRP
(mean =16.7±4.3% vs 19.3±4.1%, P<0.003). On low
dose dobutamine, they showed a significantly higher CO
and RBF with an insignificant increase in RRP (mean
=17.3±2.3%) due to renal vasodilatation. On the other
hand, reaching the higher dose of dobutamine, CO
showed a further statistically significant increase, whereas
RBF and RRP decreased significantly (mean =
14.2±3.4%, P<0.02), due to renal vasoconstriction.
Conclusion: The effects of different doses of dobutamine
on the CO were not paralleled by similar effects on the
RBF and the RRP. In patients with CHF, the adjustment of
the proper dose of dobutamine should follow the regional
renal blood flow rather than the absolute values of the CO.
Influence of iloprost on hepato splanchnic metabolic activity and energy balance in patients with
septic shock
P Kiefer, I Tugtekin, H Bracht, C Altin, J Vogt, H Wiedeck, M Georgieff and P Radermacher
Universitätsklinik f. Anästhesiologie, Universität, D-89070 Ulm, Germany
Introduction: Septic shock is characterized by increased
splanchnic blood flow due to enhanced metabolic activity.
Endogenous prostacyclin may be crucial to maintain liver
function and gastric mucosal integrity [1]. Therefore we
studied the effect of intravenous. iloprost on hepatosplanchnic metabolic activity and energy balance in
patients with septic shock.
lactate uptake (Fick principle), endogenous glucose production rate (stable isotope approach), as well as hepatic
venous lactate/pyruvate and acetoacetate/β-OH-butyrate
ratios. Measurements obtained after 90’ of hemodynamic
steady-state were recorded before, after and during iloprost infusion titrated to obtain a 20% increase in CI
(0.9–3 ng/kg×min-1).
Patients/methods: Twelve patients with septic shock
(cardiac index, CI ≥3 l/min×m2, all requiring noradrenaline
≥0.03 µg/kg×min-1 to maintain mean arterial pressure
(MAP ≥70 mmHg) were studied. In addition to routine systemic hemodynamics and gas exchange we inserted a
Swan Ganz catheter into a hepatic vein (hv) to measure
splanchnic blood flow using primed continuous infusion of
indocyanine-green. Moreover, we assessed splanchnic
Results/statistics: See Table. Friedmann test/StudentNewman-Keuls: *P<0.05 vs baseline.
Conclusion: While maintaining hepato-splanchnic lactate
clearance, iloprost reduced the endogenous glucose production rate, hence the hepatic O2 requirements. The
unchanged regional O2 uptake therefore suggests shifting
of O2 consumption to other energy demanding processes.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Sys. DO2 ml/min/m2
Spl. DO2 ml/min/m2
Spl. VO2 ml/min/m2
Spl. Lactate balance µmol/kg/min
Glucose production µmol/kg/min
Sys. VO2
Hv. Lactate/pyruvate
Hv. Acetoacetate/β-OH-butyrate
Acknowledgement: Supported by Deutsche Forschungsgemeinschaft (Ra
396/4-1); Schering - Germany
Brinkmann A et al.: Crit Care Med 1996; 24:1293–1301.
Epinephrine is more effective than other sympathomimetics in correcting cerebral hypoperfusion
associated with mesenteric ischemic reperfusion insult
MM Zayek, CR Hamm, KT O’Donnell and FG Eyal
Pediatrics, University of south Alabama, Mobile, AL, USA
Objective: To assess the efficacy of sympathomimetics in
maintaining cerebral blood flow when cardiac function is
Background: Sympathomimetics are frequently used to
treat hypotensive newborns with one of the goals being
the preservation of cerebral blood flow. No animal model
has substantiated the efficacy of this practice. Although
inotropes have been extensively studied in healthy animals,
little is available concerning their efficacy in “sick” hemodynamically impaired newborn animals.
Design/methods: A laparotomy was performed in anesthetized piglets (10±1 days old, n=40) to clamp a major
branch of the superior mesenteric artery for 30 min. One
hour after, a persistent state of impaired cardiac function
was produced. Cardiac output remained 24±2% below
initial baseline. There was a parallel decrease in carotid
blood flow (21±5%)(CBF) while the decrement in systemic mean blood pressure (BP) was small (73±2 to
67±2 mmHg; P<0.01). Those parameters remained the
same for the next two hours when treatment consisted
% change
only of intravenous normal saline at a rate of 35 ml/kg/h.
One hour after the start of mesenteric reperfusion, animals
were randomized to additionally receive: Dobutamine
(DOB), Dopamine (DP), Epinephrine (EPI) or no inotropes
(CONT). The sympathomimetics (SYMP) were given at
20 min intervals, using a randomized latin squares design,
at a rate of 5, 10, 20, 40 (DOB and DP) or 0.5, 1, 2 and
4 µg/kg/min (EPI). The animals were instrumented to
measure: aortic, pulmonary artery and superior sagittal
sinus blood gases and lactate, cerebral oxygen extraction
(CBF ext); mean systemic and pulmonary artery pressures(PAP); cardiac output (thermodilution),carotid blood
flow (ultrasonic flow transducers).
Results: See Table. Mean ±s.e.m of the % change from
levels immediately prior to treatment.
Conclusions: Depressed cardiac output is associated
with a significant decrease in cerebral blood flow even
though BP is minimally reduced. EPI may be more efficient
than the other inotropes in supporting neonatal CBF when
CO is impaired.
CBF ext
*:10Þ20,40 or 1Þ2,4,**:SYMPÞCONT, §:EPIÞCONT,DOB,DP; P<0.05; RM Anova. All SYMP similarly improved cardiac output. EPI was the
only one to increase BP and improve cerebral blood flow. The interruption of all SYMP resulted in a large drop in CBF (27±5%) and BP (28±3%)
while CO remained unchanged.
Norepinephrine-induced blood pressure increase to pre-shock levels worsens regional flow
distribution in porcine septic shock
MM Treggiari-Venzi, J-A Romand, D Burgener*, M Svensson†‡, PM Suter and A Aneman‡
Surgical Intensive Care Unit Division, *Division of Gastroenterology, University Hospital, 1211 GENEVA 14–Switzerland;
†Departments of Physiology, Göteborg University, and ‡Anesthesiology and Intensive Care, Sahlgren’s University Hospital,
Göteborg, Sweden
Septic shock is characterized by hypotension, persisting
after fluid replacement and requiring vasopressors. Mean
arterial pressure (MAP) as endpoint of treatment providing adequate regional perfusion is unclear. In an acute
endotoxic shock model, norepinephrine was used to
reverse hypotension in seven fluid-resuscitated pigs,
anesthetized with α-chloralose and equipped with flow
probes around the portal vein and renal artery, renal and
jejunal mucosal laser Doppler flowmetry and jejunal
tonometry. MAP was increased by 10 and 20 mmHg
above the shock level with norepinephrine. Seven
shocked, fluid-resuscitated only animals served as
control. Measurements were performed before 2 h-endoP49
toxin infusion and at the end of each increased MAP level.
Raising MAP with norepinephrine by 10 mmHg increased
significantly cardiac output, systemic oxygen extraction,
portal vein blood flow, improved metabolic balance and
tended to restore renal and jejunal mucosal flows to preshock levels. Increasing MAP by 20 mmHg further
increased cardiac output and oxygen delivery but
reduced portal vein blood flow and tended to decrease
renal and jejunal mucosal flows. In conclusion, using norepinephrine to increase MAP by 10 mmHg in volumeresuscitated, acutely septic animals improved systemic
and regional perfusion. Higher MAP increase did not add
any benefit despite an enhanced cardiac output.
Inhibition of inducible nitric oxide synthase (iNOS) reduces multi-organ failure (MOF) in the
thioacetamide (TAA) rat model
TM Rahman and HJF Hodgson
Centre for Hepatology, Department of Medicine, Royal Free School of Medicine, Rowland Hill Street, London NW1, UK
Introduction: The role of iNOS in MOF has been controversial, studies having presented contradicting results,
however, the use of more specific inhibitors of iNOS may
provide benefit.
pre-treated with the NO precursor L-Arginine (300 mg/kg IP)
once daily and Group 3 was pre-treated with NO synthase
inhibitor aminoguanidine (100 mg/kg SC) for three days.
Results: See Table. n=10 Mean ±SD.
Aim: To examine the effects of iNOS inhibition in our
model of TAA induced MOF.
Method: MOF was induced by two intraperitoneal (IP) injections of TAA (500 mg/kg) eight hours apart. Three groups
were studied, Group1 receiving TAA only. Group 2 and 3
followed thr protocol for Group 1, however, Group 2 was
Conclusion: Inhibition of iNOS using aminoguanidine significantly improves the incidence of MOF and mortality in
the TAA model of MOF.
Creatinine (umol/l)
Lactate mmol/l
Survival %
Time (h)
AST (iu/l)
Histology: The histological sections show markedly less
organ damage in the aminoguanidine group (Group 3).
Effect of iNOS inhibition on exhaled NO and serum nitrate in hyperdynamic porcine endotoxemia
MM Theisen, D Vasilev, J Altherr, M Matejovic*, I Tugtekin, M Georgieff and P Radermacher
Department of Anesthesia, University Hospital, D-89073 Ulm, Germany; *Department of Intensive Care, Charles-University, CZ30460 Plzen
Introduction: Controversial data have been reported on
the pivotal role of excessive nitric oxide (NO) production
for sepsis-induced vasodilatation in higher species [1,2,3].
Since both the measurement of plasma nitrate (NO3) and
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
12 h
18 h
24 h
A nitrate mol/l
1400 W
24 (13;31)
28 (16;29)
21 (16;33)
21 (17;26)
22 (13;31)
17 (14;23)
18 (11;31)
15 (13;23)
PV nitrate mol/l
1400 W
20 (11;25)
18 (14;28)
25 (16;28)
20 (15;22)
23 (13;30)
19 (13;22)
20 (16;28)
16 (14;18)
HV nitrate mol/l
1400 W
25 (16;34)
25 (16;28)
24 (15;34)
19 (16;27)
23 (14;32)
20 (15;23)
21 (10;29)
18 (14;20)
NO expired pmol/kg/min
1400 W
8 (4; 12)
9 (5; 16)
25 (19;34)*
36 (27;55)*
18 (13;25)*
13 (9;18)
22 (12;27)*
17 (12;21)
exhaled NO levels are used as a surrogate for excess NO
formation, we investigated the effect of selective iNOS
inhibition on these parameters during longterm porcine
endotoxemia [4].
Methods: 12 h after shock induction with continuous i.v.
endotoxin (LPS) anesthetized pigs were randomly
assigned to receive either LPS alone (CNT, n=13) or LPS
plus the selective iNOS inhibitor 1400W (1400W, n=9)
tritrated to keep MAP at pre-shock levels (range 0.1–0.5
mg/kg/h). Hydroethyl starch was infused to maintain sustained increase in cardiac output. Before, as well as 12,
18 and 24 h after, the start of LPS we measured expired
NO levels and arterial (A), portal (PV) and hepatic venous
(HV) NO3 concentrations (after chemical reduction to NO)
using chemiluminescence. Exhaled NO was calculated as
described in [5].
Results: Data are median (25/75% quartiles), P<0.05 *vs
pre-shock (RM ANOVA on ranks, SNK test for multiple
comparison), †1400W vs CNT (Mann-Whitney Rank Sum
Test). There was no significant relationship between
exhaled NO and blood NO3 concentration.
Conclusions: Remaining plasma NO3 levels do not reflect
excess NO release during sepsis, probably due to variable
volume of distribution resulting from interstitial edema
and/or flux into third spaces such as ascites. Exhaled NO
may be more appropriate for on-line monitoring of NO production in sepsis.
Acknowledgement: Supported by the Deutsche Forschungsgemeinschaft,
Deutscher Akademischer Austauschdienst and ESICM.
Šantak B et al.: Br J Pharmacol 1997, 122:1605–1610.
Pastor CM et al.: AJP 1999, in press.
Mehta S et al.: Crit Care Med 1999, 27:385–393.
Šantak B et al.: Br J Pharmacol 1998, 124:1689–1697.
Mehta S et al.: Chest 1997, 111:1045–1049.
Effects of mercaptoethylguanidine during long-term hyperdynamic porcine endotoxemia
F Ploner*, I Tugtekin, M Matejovic†, M Theisen, A Stehr‡, C Szabo§, UB Brückner# and P Radermacher
*Department of Anesthesia, District Hospital, I-39042 Brixen; †Department of Intensive Care, Charles-University, CZ-30460
Plzen; ‡Department of Surgery, University Hospital, D-93042 Regensburg; §Inotec Inc., Beverly, MA 01915, USA; Division of
Pathophysiology in Anesthesia and #Surgical Research, University Hospital, D-89073 Ulm, Germany
Introduction: Excess NO production due to iNOS activation and cellular toxicity resulting from peroxynitrite
(ONOO–) may contribute to organ dysfunction in septic
shock. Therefore, we studied the effect of the combined
ONOO–scavenger and selective iNOS-inhibitor mercaptoethylguanidine (MEG) [1] on hepato-splanchnic hemodynamics and energy metabolism during long-term
hyperdynamic porcine endotoxemia [2].
Methods: 12 h after starting continuous intravenous
endotoxin, pigs received either no drug (CNT; n=9) or 3
mg/kg/h MEG (n=7). Hydroxyethyl starch was infused to
maintain a sustained increase in cardiac output [2].
Before, as well as 12, 18, and 24 h after, the start of LPS
we assessed expired NO formation (chemiluminescence),
systemic (CO) and liver (Doppler ultrasound flow probes)
blood flow, arterial-ileal mucosal PCO2-gap (fiberoptic
sensor), portal (pv) and hepatic venous (hv) lactate/pyruvate (L/P) ratios and hepatic lactate clearance.
Results: See Table.
Conclusion: MEG allowed for hemodynamic stabilization
due to blunting of the progressive endotoxin-induced fall in
MAP while maintaining CO but did not influence the parameters of hepato-splanchnic energy metabolism. Ongoing
oxidative stress resulting from inadequate dosage of the
compound may account for this result [3].
Acknowledgement: Supported by Provinz Bozen-Südtirol (Italy), ESICM
and Deutsche Forschungsgemeinschaft.
Szabó et al.: Crit Care Med 1999, 27:1353.
Šantak et al: Br J Pharmacol 1998,124:1689–1697.
Whiteman et al: Br J Pharmacol 1999, 126:1646.
Before LPS
12 h LPS
18 h LPS
24 h LPS
Expir. NO
6 [2;8]
7 [6;14]
16 [13;21]#
20 [16;34]#
13 [10;18]#
8 [6;16]
15 [14;19]#
10 [3;19]
Data are median
[25;75% quartile]
101 [91;106]
99 [90;105]
85 [75;95]#
88 [82;97]
79 [68;90]#
94 [90;111] §
82 [61;85]#
102 [93;106] §
108 [98;122]
115 [104;128]
160 [147;168]#
154 [135;165]#
150 [128;174]#
159 [136;175]#
159 [134;198]#
160 [149;186]#
L/P ratio
15 [13;17]
13 [11;14]
19 [18;21]#
18 [17;21]#
20 [17;28]#
21 [18;26]#
25 [21;30]#
24 [20;27]#
15 [8;17]
14 [10;22]
17 [15;26]
22 [17;25]
20 [16;27]
15 [13;20]
20 [10;31]
14 [10;26]
Liver blood
flow ml/kg
28 [23;31]
24 [23;27]
31 [27;34]
26 [24;33]
31 [23;36]
31 [27;36]
33 [24;35]
26 [24;37]
15 [13;23]
12 [10;16]
3 [–4;6]#
5 [2;9]#
–4 [–8;0,6]#
4 [-6;10]#
–8 [–12;-6]#
0,4 [–15;7]#
L/P ratio
15 [13;19]
14 [12;16]
22 [16;30]#
20 [17;22]#
29 [20;75]#
22 [19;31]#
64 [24;100]#
50 [17;94]#
# P<0.05 vs Before
LPS (Kruskal-Wallis
ANOVA on Ranks)
P<0.05 MEG vs
CNT (MannWhitney Test).
Endotoxin-induced pulmonary hyporesponsiveness to inhaled nitric oxide is improved by nitric
oxide synthase 2 inhibition
C Bopp, A Holzmann, H Schmidt, MM Gebhard* and E Martin
Department of Anaesthesiology and *Department of Experimental Surgery, Ruprecht-Karls-University, INF 110, 69120
Heidelberg, Germany
Introduction: In isolated-perfused lungs of endotoxin-challenged rats, inhibition of nitric oxide synthase 2 (NOS2) by
aminoguanidine (AG) improved responsiveness to inhaled
nitric oxide (NO) [1]. In dogs treated with lipopolysaccharide (LPS), S-methylisothiourea (SMT) and AG prevented
worsening of hemodynamics and gas exchange [2]. In this
study, inhibition of NOS2 by SMT and AG modulates
responsiveness to inhaled NO in lungs of endotoxin (LPS)challenged rats in a dose-dependent manner.
Methods: Sprague-Dawley rats (400–450 gm BW) were
injected i.p. with 0.5 mg/kg E. coli 0111:B4 LPS (n=42)
or were controls (n=9). Four hours later, LPS-treated rats
were injected i.p. with 3, 10, and 30 mg/kg AG or 0.01,
0.1, 1 and 10 mg/kg SMT. 18 h later, lungs were isolatedperfused in situ using Hank’s solution. Pulmonary artery
pressure (PAP) was elevated by 6–8 mmHg using the
thromboxane analogue U46619. Decrease of PAP in
response to inhaled NO (0.4, 4, and 40 ppm NO in
random order) was measured. Serum and lung perfusate
nitrate/nitrite levels were measured with the Griess-reaction.
0.4ppm NO
∆PAP (mmHg)
4ppm NO
LPS+3mg/kg AG
–4.98±1.69* –6.16±2.55*
LPS+10mg/kg AG
–4.02±2.16* –6.22±1.65*
LPS+30mg/kg AG
–3.37±1.80* –5.53±1.27*#
LPS+0.01mg/kg SMT –0.44±0.33#
–1.60±1.23# –2.64±1.71
LPS+0.1mg/kg SMT
–1.28±0.97# –1.96±1.51#
LPS+1mg/kg SMT
–3.70±0.79* –5.75±2.47*
LPS+10mg/kg SMT
–0.50±0.51# –1.77±0.40#
–0.64±0.68* –0.98±0.65#
*P<0.05 vs treatment with LPS alone; # P<0.05 vs control; data are
mean ±SEM (Scheffe-test).
ment of NO responsiveness correlated with nitrate/nitrite
levels in serum and lung perfusate.
Results: See Table.
Conclusions: AG is more effective in improving NO
responsiveness than SMT. Responsiveness to inhaled NO
appears to be dependent on the degree of NOS2 inhibition in lungs of LPS-challenged rats, because improve-
40ppm NO
Holzmann A: Anaesthesiology 1999, 91(I):215–221.
Numata M.: J Immunol 1998, 160(6):3031–3037.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
In contrast to CD64 expression, CD14 expression is persistently downregulated on
polymorphonuclear cells (PMN) but not on monocytes of patients with septic shock
E Barth, J Wollmeyer, G Fischer, T Weckmann, M Appel, EM Schneider, LL Moldawer*, M Georgieff and M Weiss
Department of Anesthesiology, University of Ulm, Steinhoevelstr. 9, 89070 Ulm, Germany; *Department of Surgery, University of
Florida, Gainesville, USA
Objectives: To investigate the time course of surface
CD14 and CD64 on PMN and on monocytes, in postoperative/post-traumatic patients with septic shock, from the
first day of septic shock onwards.
Methods: Over a six month period, in 22 patients admitted to the intensive care unit, a daily analysis of CD14 and
CD64 surface expression on PMN and monocytes was
performed by flow cytometry (Becton Dickinson, USA). All
patients suffered from septic shock for three days at least.
Box plots of the mean intensity (MnI) values of surface
CD14 and CD64 are given for days 1–14. Statistical
analysis was performed by Wilcoxon rank sum test
(P<0.05; significant intragroup differences between the
day indicated by a star [*] and the following days denoted
by a line over the respective days [I–I]). In addition, normal
range of controls [C] is presented by a box plot.
Results: CD64 is significantly upregulated on PMN and
monocytes compared to healthy volunteers, whereas
CD14 expression is significantly decreased on all days.
Regarding the time course, a significant decrease in
CD14 and CD64 expression on PMN occured within the
first three days. In monocytes, compared with the first two
days, CD64 expression was diminished from day 3
onwards, however, CD14 expression on day 4, only.
Conclusions: Upregulation of CD64 expression on PMN
and monocytes indicates leukocyte activation in patients
with septic shock. In contrast to CD64, surface CD14 on
PMN decreased and is continuously downregulated over
time, whereas no decrease in surface CD14 on monocytes is observed. This may be due to different membrane
anchors and/or different LPS–CD14 signaling in PMN and
monocytes in patients during septic shock.
Respiratory burst activity of PMN is increased in most patients during septic shock
E Barth, G Fischer, F Börner, LL Moldawer*, EM Schneider, M Georgieff and M Weiss
Department of Anesthesiology, University of Ulm, Steinhoevelstr. 9, 89070 Ulm, Germany; *Department of Surgery, University of
Florida, Gainesville, USA
Objectives: To investigate whether burst activity of polymorphonuclear cells (PMN) of postoperative/post-traumatic patients during septic shock is predominantly
increased, unchanged or decreased compared to that of
healthy volunteers.
Methods: Over a period of six months, 12 patients (four
survivors and eight nonsurvivors) with septic shock for at
least three days duration were studied during their period
of septic shock on the intensive care unit. E. coli induced
burst activity of granulocytes was determined by flow
cytometry (Bursttest™, Orpegen, Heidelberg, Germany).
The PMN burst activity of 12 healthy volunteers served as
normal range.
Number of patients with more than 50% of days during septic
shock within a distinct range of PMN burst activity compared
to the normal range
Results: PMN burst activity in healthy volunteers ranged
from 5–18 MnI (median 9.2). Compared to this normal
range, burst activity of patients with septic shock was predominantly (i.e. on more than 50% of days in septic shock)
greater than that of healthy volunteers, in all patients, survivors and non-survivors.
Conclusions: PMN burst activity is predominantly
increased in most patients with septic shock compared to
healthy volunteers, indicating an activation of PMN during
septic shock. Moreover, burst activity of granulocytes
All patients
during septic shock does not discriminate survivors from
Differences in phagocytosis by polymorphonuclear leucocytes (PMNs) and monocytes (Mos) in
endotoxemia and endotoxin (ET) tolerance
K-H Staubach, J Nolde, L Song, H Brade* and H-P Bruch
Department of Surgery, Medical University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, *Research Center Borstel, Medical
University of Lübeck
an ET shock with ET from Salmonella friedenau at a dose
of 250 ng/kg BW/h, like the controls who were not pretreated.
Flow cytometry offers a rapid method to quantitate phagocytosis by PMNS and Mos. Since the PMN is a major
defense beside the Mos against bacterial infection we
compared both systems in a porcine endotoxin shock
model in control animals (C) and endotoxin-tolerant
animals (T). A commercial test kit (phagotest, Orpegen,
Heidelberg, Germany) was used for demonstrating ingestion of opsonized fluorescent E. coli (FI) and phagocytosing cells (PC). Seven pigs were pretreated over five days
with increasing doses of atoxic endotoxin from
Rhodopseudomonas sphaeroides (T) before induction of
Results of both groups were expressed in fluorescence
intensity (FI) and summarized in the Table.
The functional assessment of Mos and PMNS in our shock
model revealed an enhancement of phagocytic activity in
endotoxin-tolerant animals both in Mos and PMNS with
the highest increase in the latter.
Days (immunization period)
Hours (ET shock)
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Role of platelet activating factor (PAF) on leukocyte-independent plasma extravasation and mast
cell activation during endotoxemia
A Walther, N Yilmaz, W Schmidt, A Secchi, MM Gebhard*, E Martin and H Schmidt
Department of Anesthesiology, and *Department of Experimental Surgery, Ruprecht-Karls-University, Im Neuenheimer Feld 110,
69120 Heidelberg, Germany
Introduction: Independently from leukocyte adherence,
endothelial factors and mast cell activation promote
microvascular permeability [1]. The platelet activating
factor (PAF) has been shown to play a significant role in
endotoxin-induced leukocyte adherence [2]. The aim of
our study was to investigate if there is also a role of PAF in
mediating leukocyte-independent microvascular permeability changes and activation of mast cells during endotoxemia. Therefore, during endotoxemia, microvascular
permeability and mast cell activation were determined
after inhibition of the L-selectin mediated leukocyte-adherence by fucoidin and after inhibition of PAF effects by the
the PAF receptor antagonist BN52021.
dure in group A, group B (n=10) received BN52021 (5
mg/kg b.w.) after baseline measurements. Group C
(n=10; control group) only received equivalent volumes of
NaCl 0.9%. Statistical analysis was performed using student’s t-test. A P-value <0.05 was considered significant.
Methods: In male Wistar rats, microvascular permeability
(MP), leukocyte adherence (LA) and mast cell activation
(MCA) were determined in mesenteric postcapillary
venules using intravital microscopy at baseline and at 60
and 120 min after the start of a continuous infusion of
endotoxin (ETX; 2 mg/kg/hr, E. coli O26:B6) (group A,
n=10). Leukocyte–endothelial interaction was blocked
using fucoidin (25 mg/kg b.w.). In addition to the proce-
Conclusions: The results of this study demonstrate that
PAF plays a significant role in the initiation of endotoxininduced leukocyte-independent plasma extravasation and
mast cell activation.
Results: In groups A and B, fucoidin prevented LA, and
so there were no significant differences in LA between
these groups. In group A, MP and MCA significantly
increased starting at 60 min (P<0.05 vs baseline). There
was no significant increase in MP and MCA in groups B
and C. Differences between groups A and B in MP and
MCA were significant at 120 min.
Kubes P: Am J Physiol 1996, 271:H2438.
Schmidt H: J Surg Res 1996, 60:29.
α: a possible mediator of remote tissue injury after viper envenomation
O Szold, AA Weinbroum, R Ben-Abraham, TE Englender, D Ovadia, S Marmor, M Sorkine, R Flaison and P Sorkine
General Intensive Care and the Departments of Oncology and Pathology, Tel Aviv Sourasky Medical Center and the Sackler
Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; and Unite des Venins, Institute Pasteur, Paris,Cedex 15, France
Introduction: TNF is a potent inflammatory promoting agent
that can potentiate organ and tissue injury. A possible role
of TNF in causing local tissue damage following snakebite
was recently demonstrated in an intact rat model. We compared the systemic effects of TNF on the hearts and lungs
of rats following an intramuscular injection of a sub-lethal
dose of Vipera asis venom (500 µg/kg, experimental
groups) to equivalent injections of saline (control group).
Results: Systemic TNF activity, heart rate and blood pressure as well as lung permeability and neutrophil sequestration were then evaluated in both groups. The venom
caused a significant reduction in heart rate and arterial
blood pressure, and the serum TNF levels peaked after
two hours. These values remained unchanged in the
control group. In contrast, lung microvasculature permeability and neutrophil sequestration were not significantly
different between the experimental and control groups.
Conclusions: This study of the systemic and inflammatory
effects of Vipera aspis venom showed that intramuscular
injection of the poison results in systemic effects that are
possibly mediated in part by TNF. These findings may
have therapeutic implications in the treatment of patients
with severe systemic manifestations after snakebites in
terms of the possible benefits of blocking TNF activity.
Phagocytosis of granulocytes is decreased in most patients with severe sepsis or septic shock
G Fischer, E Barth, R Remmele, LL Moldawer*, EM Schneider, M Georgieff and M Weiss
Department of Anesthesiology, University of Ulm, Germany, and *Department of Surgery, University of Florida, Gainesville, USA
Objectives: To investigate the phagocytotic activity of
granulocytes during septic shock in postoperative/posttraumatic patients compared to that of healthy volunteers.
Methods: Over a six month period 19 patients with septic
shock were monitored on a daily basis during their stay in
the intensive care unit. Phagocytotic activity of granulocytes was determined by flow cytometry (Phagotest™
Orpegen, Heidelberg, Germany). The normal range of
phagocytosis (expressed as mean intensity, MnI) was
defined by a control group of 11 healthy volunteers.
Results: Phagocytosis of healthy volunteers was in the
range of 34 MnI and 149 MnI. All survivors and 7 of 10
nonsurvivors (i.e. 84% of all patients) had a lower phagocytotic activity on more than 50% of days in septic shock
compared to that of healthy volunteers. No patient had a
phagocytosis greater than 149 MnI, the upper value of the
normal range.
Conclusions: Phagocytosis is decreased in most patients
during septic shock compared to healthy volunteers. Thus
diminished phagocytotic activity may contribute to an
impaired mechanism of bacterial elimination or a reduced
Number of patients with more than 50% of days during septic
shock within a distinct range of phagocytosis during septic
shock compared to the normal range.
Phagocytosis (MnI)
compared to normal
resolution of infection in these patients. Moreover, phagocytotic activity of granulocytes does not discriminate survivors and nonsurvivors of septic shock.
Triiodothyronine (T3) ameliorates the cytokine storm in rats with sepsis
T Yokoe*, M Negishi*, Y Iino* and Y Morishita†
*Department of Emergency and Critical Care Medicine, †Second Department of Surgery, Gunma University School of Medicine,
3-39-15 Showa-machi Maebashi Maebashi, Gunma 371-8511, Japan
We reported that T3 eliminated acidosis and improved survival rate in a rat model of sepsis. The aim of this study is
to clarify the effect of T3 on cytokine levels.
Male Sprague-Dawley rats, weighing 350–420 g, were
ligated in the cecum with puncture (CLP method) under
anesthesia with pentobarbital (40 mg/kg). After CLP, 50
ml/kg of saline was injected subcutaneously for fluid
resuscitation. Rats were assigned two groups; no treatment group (control group, n=15) and a T3 treated group
(n=10). The T3 treated group was given 3 ng/hr of T3
using a osmotic pump embedded subcutaneously. Survival rate, levels of IL-1β, TNF-α, IL-6, IL-8, and IL-10 were
studied 24 h after surgery.
Three animals died in the control group, whereas no
animals died in T3 treated group. There was no significant
difference on mean value of IL1-β and TNF-α between two
groups. Mean levels of IL-6 and IL-8 in the control group
were 237.4 pg/ml and 5342.7 pg/ml, respectively,
however, those in the T3 group were suppressed to the
level of 183.5 pg/ml and 55.2 pg/ml, respectively
(P<0.05). Mean IL-10 level (6.0 pg/ml) in the T3 treated
group was lower than that (72.4 pg/ml) in the control
group (P<0.05).
It is concluded that T3 prevents the cytokine storm and
improves the survival rate in this rat model of sepsis.
Interleukin-17 stimulates intraperitoneal neutrophil infiltration through the release of the
α from peritoneal mesothelial cells
chemokine GROα
J Witowski*†, K Pawlaczyk†, A Breborowicz†, A Scheuren*, M Kuzlan-Pawlaczyk†, J Wisniewska†, A Polubinska†, H Friess‡,
GM Gahl*, U Frei* and A Jörres*
*Nephrology and Medical Intensive Care, UK Charité, Campus Virchow-Klinikum, Augustenburger Platz 1, D-13353 Berlin,
Germany; †Pathophysiology, University Medical School, Poznan, Poland; and ‡Visceral and Transplant Surgery, University of
Bern, Inselspital, Bern, Switzerland
Introduction. IL-17 is a newly discovered cytokine implicated in the regulation of hematopoiesis and inflammation.
Since IL-17 production is restricted to activated T lymphocytes the effects exerted by IL-17 may help to understand
the contribution of T cells to the inflammatory response.
We investigated the role of IL-17 in leukocyte recruitment
into the peritoneal cavity.
Methods: Leukocyte infiltration in vivo was assessed in BalB/
CJ mice. Effects of IL-17 on chemokine generation in vitro
were examined in human peritoneal mesothelial cells (HPMC).
Results: Intraperitoneal administration of IL-17 resulted in
a selective recruitment of neutrophils into the peritoneum
and increased levels of KC chemokine (murine homologue
of human GROα). Pre-treatment with anti-KC antibody
significantly reduced the IL-17-driven neutrophil accumulation. Primary cultures of HPMC expressed IL-17 receptor
mRNA. Exposure of HPMC to IL-17 led to a dose- and
time-dependent induction of GROα mRNA and protein.
Combination of IL-17 together with TNFα resulted in an
increased stability of GROα mRNA and synergistic
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
release of GROα protein. Anti-IL-17 antibody blocked the
effects of IL-17 in vitro and in vivo.
Conclusions: IL-17 is capable of selectively recruiting neutrophils into the peritoneal cavity via the release of neutrophilspecific chemokines from the peritoneal mesothelium.
α (TNF-α
α) in men: relationship to cytokine
The ex vivo production of tumor necrosis factor-α
balance during cardiopulmonary bypass
Y Tabardel*, L Dumont†, D Schmartz*, JL Vincent‡ and J Duchateau
Departments of *Anesthesiology and ‡Intensive Care, Erasme University Hospital, Departments of †Anesthesiology and
‡Immunology, Brugmann University Hospital, Free University of Brussels, Brussels, Belgium
Objectives: To discriminate patients with strong or weak
ex vivo TNF-α synthesis in response to whole blood stimulation with endotoxin (LPS) and to relate this ex vivo
response to cytokine balance during and after cardiac
surgery with cardiopulmonary bypass (CPB).
Patients and methods: 19 patients scheduled for coronary artery bypass graft (CABG) with CPB and 29 adult
volunteers were entered into the study. Whole blood was
stimulated with LPS, incubated for 16 h and TNF-α concentrations determined in the culture supernatant. Results
obtained in both patients and volunteers were pooled. Individuals showing TNF-α concentrations lower or higher than
the median value of 1052 pg/ml were defined as low and
high responders, respectively. In the 19 patients, cytokine
balance during and after CPB was evaluated by the plasma
concentrations of TNF-α, interleukin (IL)-8 and IL-10.
Results: In patients undergoing cardiac surgery, perioperative cytokine plasma levels were not significantly different
in both low and high responders. In these two groups ex
vivo TNF-α production did not correlate with either TNF-α
or IL-8 plasma levels measured during and after CPB. In
contrast, ex vivo TNF-α production inversely correlated
with IL-10 plasma concentrations observed four hours
after the end of CPB in the low responder group (Spearman, –0.76, P=0.01) and tended to do so in the high
responder group (Spearman, –0.62, P=0.1).
Conclusion: In our series, the ex vivo production of TNFα does not allow us to predict the magnitude of the systemic inflammatory response related to cardiac surgery,
as evaluated by the plasma levels of TNF-α and IL8
during and after CPB. The negative correlation observed
between the ex vivo TNF-α production and the production of IL-10 after CPB suggests that patients with a low
ex vivo capability to synthesize the pro-inflammatory
TNF-α have a high potential to synthesize the anti-inflammatory cytokine IL-10 in vivo. Whether this profile can be
considered a risk factor for morbidity after cardiac
surgery, as it has been shown for infectious diseases
remains to be assessed.
α and IL10 release during cardiac operations and the development of
Relationship between TNFα
M Qing*, J Vazquez-Jimenez†, B Klosterhalfen‡, K Schumacher*, A Pütz-Ebert*, J Duchateau§, G von Bernuth*
and M-C Seghaye*
*Department of Pediatric Cardiology, †Cardiac Surgery, ‡Pathology, Aachen University of Technology, Pauwelsstrasse 30,
D-52057, Aachen, Germany; §Department of Immunology, Hopital Brugman, Brussels, Belgium
Aim: To analyze the role of tumor necrosis factor (TNF)α
and interleukin (IL)10 produced during cardiac operations
on the development of apoptosis in vital organs and the
effect of core temperature during CPB on these finding.
Methods: 16 young pigs were assigned to a temperature
(T°) regimen during standardized CPB: normothermia (T°
37°C; n=8) and moderate hypothermia (T° 28°C; n=8).
TNFα and IL10 were determined by a pig specific ELISA.
Six hours post-operative, tissue probes of the heart, liver,
lung, and kidney were taken for standard- and immuno-histological examinations. Apoptotic cells were detected by
an in situ apoptosis detection kit (TUNEL). Necrotic cells
were counted by light microscopy.
Results: Pigs operated on in normothermia showed significantly higher TNFα production and lower IL10 production
during and after CPB, and higher degree of cellular necrosis but lower degree of cellular apoptosis than pigs operated on in moderate hypothermia. In all animals,
postoperative IL10 levels were negatively correlated with
the production of myocardial apoptosis (P=0.005) and
TNFα levels correlated significantly with necrosis in the
heart and liver (P<0.05).
Conclusion: Cardiac operations are related to cell death
via both the necrotic and the apoptotic pathways. In this
setting, IL10 could be an inductor of apoptosis. By
increasing the production of IL10 and the ratio apoptosis/necrosis, moderate hypothermia during CPB might
therefore provide organ protection.
Elevated levels of adhesion molecules in septic patients with pre-existing coronary artery disease
R Wittich*, H Kern, WJ Kox, G Haring* and C Spies
Department of Anesthesiology and Intensive Care, University Hospital Charité, Humboldt-University, D-10098 Berlin, Germany;
*Department of Anesthesiology and Intensive Care, Carl-Thiem-Hospital, Thiemstrasse 111, D-03048 Cottbus, Germany
Introduction: Sepsis, as well as coronary artery disease
(CAD), is associated with high levels of circulating adhesion molecules [1,2]. Elevated plasma levels of these molecules may have clinical significance as markers of
endothelial injury [2]. The purpose of the study was to
investigate whether CAD influences the liberation of
plasma adhesion molecules, cytokines and markers of the
NO-system in septic patients and if these parameters may
serve for identification of patients at greater risk for cardiovascular complications in sepsis.
Table 1
CAD n=24
non-CAD n=20
sE-selectin (ng/ml)
135 (33–352)
98 (34–472)
141 (56–320)
49 (25–175)*
sICAM-1 (ng/ml)
614 (207–1616)
536 (203–1479)
446 (207–867)*
351 (202–764)*
cGMP (mmol/l)
17 (6–46)
35 (8–51)
13 (6–44)
17 (3–52)*
median and range, * P<0.05 between groups.
Methods: After approval by the local ethical committee
and obtaining informed consent from relatives, 44 septic
patients were included in this study, of whom 24 patients
met the criteria of CAD [3]. Measurements were performed within 12 h (early sepsis) of onset of sepsis and
after 72 h (late sepsis). Soluble adhesion molecules sEselectin, and soluble intercellular adhesion molecule
(sICAM)-1 were analyzed by commercially available
ELISA. NO-markers were measured by the Griess reaction
and cyclic guanosine monophosphate (cGMP) by competitive radioimmunoassay.
Results: Basic patient characteristics did not differ
between groups. CAD patients showed significantly
increased sICAM-1 levels in early and late sepsis (Table 1)
but this failed to be predictive for cardiovascular complications. ICAM-1 and cGMP correlated positively in both
groups with the intravenously administered norepinephrine
dosage in early and late sepsis. Cardiac index, oxygen
delivery and mixed venous oxygen saturation were significantly decreased in CAD patients in early and late sepsis
(P<0.05), whereas VO2 did not differ significantly between
both groups. Mortality was increased in tendency in CAD
patients (75% vs 50% in non-CAD).
Conclusion: Plasma levels of sICAM-1, sE-selectin and
cGMP were elevated in CAD but may not serve as
markers for cardiovascular complications in sepsis. The
tendency in increased mortality rate could be due to an
impaired endothelial reserve in CAD patients.
Spies C et al.: Chest 1998; 113:1055–1063.
Hwang SJ et al.: Circulation 1997; 96: 4219–4225.
Mangano D: Anesthesiology 1990; 72:153–184.
Are IL-6, IL-10 and PCT plasma concentrations more reliable than APACHE-III or SAPS-II for the
individual mortality risk prediction in severe sepsis?
C Müller, G Drüge, O Eichelbrönner and N Roewer
Bayerische Julius-Maximilians-Universität Würzburg, Klinik für Anaesthesiologie, Josef-Schneider-Strasse 2, D-97070 Würzburg,
Introduction: Scoring systems such as APACHE-III and
SAPS-II often fail to reliably predict the individual mortality
risk particularly in patients with severe sepsis. Cytokines
and procalcitonin (PCT) have been shown to play a crucial
role both in the pathogenesis of sepsis and as diagnostic
tools for the inflammatory process. The objective of this
study was to investigate the relationship between plasma
cytokine and PCT concentrations and the actual and predicted mortality using APACHE-III and SAPS-II in patients
with severe sepsis.
Material and methods: The SCCM/ACCP selection criteria were used to identify the patients with severe sepsis.
After severe sepsis was diagnosed SAPS-II and
APACHE-III scores were calculated. In addition, blood
samples were taken daily for the analysis of IL-6, IL-10 and
PCT plasma concentrations for three consecutive days. IL6 and IL-10 were analysed using a sandwich type of a
immunoenzymatic assay (Immunotech, Marseille, France)
PCT plasma concentrations were measured using a specific, ultra sensitive immunoluminometric assay (LUMItest
PCT assay, BRAHMS Diagnostica, Berlin, Germany).
Results: Out of the 29 patients of this study, 13 patients
died leading to a mortality rate of 44.8%. The APACHE-III
and SAPS-II scoring systems produced almost identical
results for the predicted mortality risks, but the areas
under the ROC-curves (AUC) formed by the plot of possible pairs of false-positive and true-positive rates exhibited
a poor prognostic validity for both scores (day 1: 0.5612;
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
day 2: 0.5593; day 3: 0.5732). The plasma PCT concentrations were significantly higher in non-survivors compared to survivors (day 1: P<0.01; day 2: P<0.07; day 3:
P<0.015). However, a discrimination value, to simply distinguish between survivors and non-survivors, was not
found. The determination of plasma IL-6 concentrations on
the first three days after inclusion into the study showed
no significant difference between survivors and non-survivors. The IL-10 plasma concentrations of the patients
who died were significantly higher compared with those of
the survivors measured on the first two days (day 1:
P<0.001; day 2: P<0.005). The degree of correlation
between PCT-values and the APACHE-III scores on all
days of study was high (day 1: P<0.0004; day 2:
P<0.001; day 3: P<0.03). For the PCT plasma concentrations and the SAPS-II score, a good correlation was only
found on the first day (day 1: P<0.005). The IL-6 concentrations showed a good correlation with the APACHE-III
scores only on the first day (day 1: P<0.04) and the IL-10
concentrations correlated well with the APACHE-III score
at day 1 and day 3 (day 1: P<0.006; day 3: P<0.026). No
correlation was found between the interleukin concentrations and the SAPS-II scoring system.
Conclusion: Our results demonstrate that the APACHEIII and the SAPS-II scoring systems evaluate a similar
mortality rate in severe sepsis. However, in terms of individual mortality risk prediction, the reliability and precision of both scores is limited and does not allow a
precise statement about the mortality risk. In contrast,
the plasma concentrations of PCT and IL-10 were significantly higher in the non-survivors and showed a good
correlation with the actual mortality. IL-6 was not useful
in this context. Thus, daily analyses of plasma PCT and
IL-10 concentrations might be a helpful tool to improve
the individual mortality risk prediction in patients with
severe sepsis, possibly in combination with scores
assessing the severity of illness.
α levels in patients with burns
Tumor necrosis factor receptor I, II and tumor necrosis factor-α
S Endo*, Y Yamada*, K Inada† and S Sato†
Critical Care and Emergency Center, Department of Bacteriology, School of Medicine, Iwate Medical University, 19-1 Uchimaru,
Morioka 020-8505, Japan
Objects: The present study was undertaken to examine
the relationship between the severity of burns and TNFRI
and TNFRII levels in the acute post-burn period and the
entire follow-up period.
Materials and methods: Tumor necrosis factor α (TNF-α)
and TNF receptor I and II (TNFRI and TNFRII) were
studied in 24 burn patients who had a total burn surface
area (TBSA) of 50.2±20.4%. TNFRI and TNFRII were
quantified by enzyme-linked immunosorbent assay (ELISA)
using a kit (Amersham, Buckinghamshire, UK). TNF-α was
quantified by ELISA (Medogenix, Fleurus, Belgium).
Results: Immediately after the injury, both the TNFRI and
TNFRII levels correlated significantly with TBSA
(r=0.7344, P<0.0001; r=0.6074, P=0.0012). The TNFRI
and TNFRII levels immediately after the injury were significantly higher in the 11 patients who later died of their
burns than in the 13 patients who survived (0.8±0.4 ng/ml
vs 1.8±0.7 ng/ml, P=0.0002; 2.3±1.1 ng/ml vs 4.5±1.6
ng/ml, P=0.0009). The TNF-α levels immediately after the
injury did not differ significantly between the group that
survived and the group that died. The TNFRI and TNFRII
values for the entire follow-up period also correlated significantly with TBSA. Peak TNFRI and TNFRII levels were
significantly higher in the group that died than in the group
that survived (6.0±4.7 ng/ml vs 14.1±7.8 ng/ml,
P=0.0009; 7.0±5.1 ng/ml vs 16.7±5.2 ng/ml, P=0.0003).
Conclusion: The TNF-α levels correlated significantly with
both the TNFRI and the TNFRII levels. The TNFRI and
TNFRII levels thus closely reflected the severity of the
burns in both the acute post-burn period and the subsequent follow-up period. In other words, these parameters
well reflected the severity and outcome of the burns, irrespective of the presence or absence of accompanying
Cytokines and sepsis – just black smoke?
S Marum, JP Ribeiro, E Arranhado*, H Lage*, L Mota, P Marcelino, AP Fernandes, J Oliveira and MR Silva
Hospital de Curry Cabral, Unidade de Cuidados Intensivos, *Laboratório de Imunologia/Nefrologia, Rua da Beneficiência, nº8,
1069-166, Lisbon, Portugal
Introduction: Sepsis is a clinical syndrome resulting from
an inflammatory response of the host to infection, mediated by the production of cytokines in circulation. The
need to understand better the mechanisms that lead to
multiple organ failure, the authors (AA) began a study of
cytokines in these patients.
Material and methods: In 1998 the AA began a prospective study of soluble cytokines, consisting at the moment of
54 patients, 36 with criteria of Sepsis/MODS (Multiple
organ dysfunction syndrome) and 18 with SIRS (Systemic
inflammatory response syndrome). The soluble cytokines
(TNFα, IL-6, IL-8 and IL-10) were measured in serum
samples by the quantitative sandwich enzyme immunoassay
(solid phase) — RD systems, Inc. Minneapolis (CITOMED,
PORTUGAL) — with a previously established protocol.
Results: In the sepsis/MODS group, 7 patients with IL-6
<300 pg/ml survived; among the ones with IL-6 >300
pg/ml, only the five patients submitted to pulses of methylprednisolone and/or emergent surgery survived. The
patients with SIRS always presented low concentrations
of cytokines, and 14 of them survived.
Conclusion: In the patients with Sepsis/MODS, the clinical severity was always related to high serum levels of IL6, and/or its increase. Serum values of IL-6 >300 pg/ml
were predictive of death. Although TNFα and IL-8 values
were high in some patients, these showed a non-uniform
behaviour, and the correlation of these values with clinical
severity or outcome was not statistically significant as was
IL-6. In those patients who survived, the clinical improvement was always followed by a remarkable decline in the
serum values of IL-6. IL-6 is an important severity serum
marker in equal or superior value to others described and
accepted world-wide for patients with sepsis.
The GH/IGF-I axis in children with a meningococcal septic shock: striking differences between
survivors and non-survivors
F de Groof, ED de Kleijn, JA Hazelzet, P Uitterlinden, J v Doorn, ACS Hokken-Koelega and KFM Joosten
Department of Paediatrics, Sophia Children’s Hospital, Dr Molewaterplein 60, 3015 GJ Rotterdam, The Netherlands
Introduction: Meningococcal septic shock is a life-threatening clinical syndrome that may lead to pronounced
neuro-endocrine and metabolic alterations. In critically ill
adults a GH resistant state has been well documented.
Less information is available in critically ill paediatric
patients. In the present study we evaluated the time
course and pulsatile pattern of the GH secretion and the
levels of serum IGF-I, IGFBP-1 and IGFBP-3.
Methods: Children with a septic shock and
petechiae/purpura requiring intensive care treatment were
enrolled in this study. The paediatric risk of mortality
(PRISM) score was used to monitor the severity of
disease. A GH profile was done in the first 6 h after admission and at day 3. The serum IGF-I, IGFBP-1 and IGFBP-3
levels were measured on admission (T=0), and 24 and 48
h after admission. The serum levels of IGF-I and IGFBP-3
were corrected for sex and age, expressed as SDS-levels.
Results: Twelve patients fulfilled the inclusion criteria, 7
boys and 5 girls, with a median age of 22 months. The
median PRISM score was 23. Three of the 12 patients
died. The serum levels of IGF-I and IGFBP-3 were
decreased in all patients, the median SDS being respec-
tively –2.6 and –5.7. There was a significant difference
between survivors and non-survivors regarding median GH
levels during the first 6 h (7 mU/l vs 131 mU/l, P=0.01),
serum IGF-I levels at T=0 (43 ng/ml vs 7 ng/ml, P=0.03),
IGFBP-3 SDS levels at T=0 (–5.6 vs –7.7, P=0.05), and
PRISM score (18 vs 34, P=0.01). The levels of serum
IGFBP-1 at T=0 were increased in the non-survivors in
comparison with the survivors (median being respectively
1315 and 262 ng/ml), but this difference didn’t reach significance (P=0.15). Non-survivors showed very high GH
levels without GH peaks during the first 6 h after admission. In survivors, the GH secretion pattern at day 1
showed a larger number of GH peaks and larger GH peak
amplitudes than at day 1. A significant increase in serum
IGF-I levels was found in all patients between T=0 and
T=24 (P=0.04), followed by a significant decrease in
serum IGF-I levels between T=24 and T=48 (P=0.01).
Conclusion: Mortality in children with a meningococcal
septic shock is associated with extremely increased serum
GH levels, no oscillatory GH secretion, high levels of
IGFBP-1 and a high PRISM score. In those who survived
there are significant changes during the first 48 h in the
GH/IGF-I axis, suggesting a decrease in the GH resistance.
Secretory phospholipase A2 (sPLA2), procalcitonin (PCT) and C-reactive protein (CRP) for the
diagnosis and differentiation of septic shock and non-septic shock
O Anglès, MH Chabannier, E Bauvin, J Fauvel, Y Coulais, B Cathala and M Génestal
Service de Réanimation Adultes, Hôpital Purpan, F31059 Toulouse, France
Introduction: Serum sPLA2 and CRP levels increase in
patients with systemic inflammatory response syndrome
(SIRS). High serum levels of PCT have been detected in
patients with inflammatory conditions from invasive bacterial and fungal infections. The aim of this study was to
determine the diagnostic value of sPLA2, CRP and PCT
in septic shock.
Patients and methods: Consecutive patients admitted to
the ICU with shock were included. Patients were
screened in two groups: septic shock (American college
of chest physicians criteria) and non-septic shock. Upon
admission, serum sPLA2, CRP and PCT were simultaneously analysed. Data, expressed as means ±SD, were
analysed by an independent investigator not involved in
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Critical Care
Mann-Whitney test
ICU. The catalytic activity of sPLA2 was detected by fluorimetric assays (normal 10 mU/ml). PCT was analysed by
commercially available Lumitest® kit (BRAHMS, Berlin).
Results: Fifty-nine patients were included (20 women, 39
men), thirty-nine with septic shock (13 women, 26 men,
mean age=56±18) and twenty with non-septic shock (7
women, 13 men, mean age=51±22). Total mortality=40%. Septic shock mortality=51% and non-septic
shock mortality=20%.
sPLA2, CRP and PCT values were significantly higher in
patients with septic shock. The areas under the curve
(ROC) of sPLA2, CRP and PCT were respectively 0.896,
0.792 and 0.765. The area under the curve of sPLA2 was
significantly higher than PCT area (P<0.05).
Conclusion: PCT does not appear to be a better marker
to discriminate septic shock and non-septic shock than
sPLA2 and CRP.
How does major abdominal surgery induce procalcitonin and IL-6 in the postoperative period?
EK Karpel
Department of Anaesthesiology and Intensive Therapy, Silesian University of Medicine, ul. Medyków 14, 40-752 Katowice, Poland
Introduction: Major abdominal surgery is often complicated by systemic inflammatory response (SIRS), local or
general infection, sepsis and even septic shock. It is
important to evaluate the risk of these serious complications to prevent them. It has been demonstrated that circulating markers of inflammation can be helpful in early
diagnosis of surgical infection in the postoperative period.
The aim of my study was to evaluate how major abdominal
surgery induces interleukin 6 and procalcitonin by itself
and to determine the usefulness of these markers in diagnosis of infectious complications.
Methods: Blood samples were taken during induction of
anaesthesia and on days 1, 3, 5 and 7 after surgery. IL-6
measurements were performed by immunoradiometric
assay (IL-6-IRMA, BIOSOURCE) and procalcitonin was
measured by illuminometric method (PCT LUMItest,
Brahms). Number of SIRS criteria and SOFA score were
assessed at the same time points.
Study design: Prospective clinical study approved by
local Ethics Committee.
In two patients with general infection, the highest levels of
PCT were 8.41 and 7.95 with IL-6 levels of 3500 and
3700 respectively.
Patients: Thirty patients, ASA II–IV, undergoing elective
surgery of the gastrointestinal system were studied. Surgical procedures included: gastrectomy or oesophagus
resection (N=10), Whipple’s operation (N=10), bowel
resection (N=5), partial liver resection (N=2), others
(N=3). In all patients, antibiotic prophylaxis (24–48h) was
applied. Eleven patients developed local infection (wound
N=6, lung infection N=3) or general infection (N=2). None
of the analysed patients died; the total period of hospital
treatment varied from 15 to 110 days.
Results: We observed low plasma levels of both IL-6 and
PCT prior to surgery and a slight increase in PCT at the first
postoperative day in patients with signs of infection accompanied by increases in SIRS and SOFA assessment.
Infection (N=11)
Time of
Before surgery
no infection (N=19)
1 day
3 day
5 day
7 day
Conclusions: Major abdominal surgery slightly
increases procalcitonin serum levels in patients with
infectious complications only, with the peak level
24–48 h after operation. Increase of IL-6 serum levels
occurs 24–48 h after surgery in all patients and correlates with the assessment of inflammation measured by
the number of SIRS criteria. PCT is a marker of infection
rather than inflammation.
Role of decisive markers in diagnosis and outcome of patients with septic shock
A Rizk, F Khalil, F Ragab, L Hamed and S Mokhtar
Critical Care Medicine Department, Cairo University Hospital, Cairo, Egypt
Introduction: Of the few biochemical parameters available to monitor critically ill patients and to control the
course of therapy in severe inflammation, procalcitonin
(PCT) and polymorphonuclear (PMN) elastase enzyme
stand as innovative diagnostic parameters with features
different from other presently routine indicators of the
inflammatory response.
Methods: Our work aims at assessing the potential role of
PCT and PMN elastase enzyme in early diagnosis and
early prediction of prognosis in patients (pts) with sepsis
and septic shock. To achieve this goal we studied 20 pts
with septic shock (16 male and 4 female, mean age 50.15
years), together with a second group comprising 10 pts (9
male, mean age 49.2 years) with systemic sepsis without
shock. A third group including 20 healthy volunteers
matching with age and sex, served as controls. Serum
PCT and PMN elastase enzyme levels were estimated on
admission for both pts and control groups with other laboratory investigations and clinical parameters. A multivariate, discriminate analysis was performed using the
following variables: PCT, PMN elastase enzyme, albumin,
α-1-antitrypsin, α-2-macroglobulin, and C-reactive protein
(CRP), as independent parameters.
Results: Comparing the three groups (septic shock,
sepsis, and healthy control subjects) all together our
results exhibited significantly higher PCT levels (59.7±14,
97.2±14 vs 0.4±0.1ng/ml, P<0.0000), higher PMN elastase enzyme (183.4±101, 132.1±117 vs 6.5±3 µg/l,
P<0.0000), lower serum albumin (2.1±0.6, 2.6±0.9 vs
4.7±0.5gm%, P<0.05), higher α-1-antitrypsin (3.9±1,
6.2±3 vs 3.6±0.6mg/L, P<0.0001), higher α-2-macro-
globulin (2.27±0.8, 2.4±1 vs 0.7±0.2mg/l, P<0.0000),
higher CRP (52.2±36, 46.8±31 vs 5.4±0.4mg/l,
With further patient subdivision to 6 survivors and 24 nonsurvivors only PCT could be identified as independent predictor for short term prognosis in patients with sepsis with
an overall predictive accuracy 80% and with cut off value
of ≥78ng/ml. While PMN elastase enzyme could be used
as a parameter inside a model including the whole previous parameters with overall predictive accuracy 76.7%,
and with cut off value of ≥142 µg/l.
Conclusion: Serum PCT and PMN elastase enzyme are
independent useful diagnostic markers for early detection
of systemic inflammatory response syndrome with or
without shock. However PCT has the advantage over the
above mentioned parameters in being significantly predictive of short-term prognosis, with overall predictive accuracy of 80%. Procalcitonin, PMN elastase enzyme,
α-1-antitrypsin, α-2-macroglobulin, CRP, and albumin, is a
model which could be used for early prediction of complications of sepsis patients with overall predictive accuracy
of 76.7%.
Gramm HJ et al: Procalcitonin — ein neuer marker der inflammatorischen wirtsantwort, longitudinalstudien bei patienten mit
sepsis und peritonitis. Chir Gastroenterol 1995, 11(suppl.):51–54.
Bohuon C et al: Blood procalcitonin is a new biological marker of
the human septic response. New data on the specificity. Suppl
Clin Intens Care 1994, 5:2.
Enscnauer M, Quintel R, Kattermann J Aufenanger. Clin Investig
1994, 72:843–849.
Gardinali M et al.: Arch Surg 1992, 127(10):1219–1224.
A new diagnostic and severity marker in patients with SIRS, sepsis, severe sepsis and septic
shock; procalcitonin
S Çelebi, P Ergin Özcan, S Tugrul, N Çakar and F Esen
Department of Anesthesiology and Intensive Care, University of Istanbul, Medical Faculty, Istanbul, Turkey
Although the exact origin of procalcitonin (ProCT) in infection remains unclear, its association with the presence and
severity of infection has been demonstrated. This prospective study aims to evaluate ProCT as a diagnostic marker
of infection in critically ill patients with SIRS, and to investigate the relationship of ProCT levels with sepsis severity
and outcome.
The study included 85 consecutive adult patients, fulfilling
the criteria of inflammatory response, admitted to the ICU
at the Istanbul University Hospital. The clinical state of the
patients was defined on the basis of clinical, laboratory
and bacteriological findings as SIRS, sepsis, severe
sepsis and septic shock according to the consensus criteria. Data were collected including temperature, heart rate,
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Table 1
Severe Sepsis
Septic Shock
Sepsis vs
Septic Shock
Severe Sepsis vs
Septic Shock
11 287±1667
12 538±7275
16 740±8949
respiratory rate, arterial pressure, WBC count, ProCT, and
CRP on each ICU day. The severity of sepsis was
recorded by SOFA score and the organ system failure by
MODS score. Correlation of ProCT values with severity
indices were done by Spearman correlation analysis and
to compare independent samples we used Mann Whitney
U-test. ROC curves and area under the curves of ProCT,
CRP, and WBC in diagnosing infection and prognosticating sepsis outcome, were calculated.
Compared with CRP, ProCT had a higher sensitivity (0.35
vs 0.73) and specificity (0.42 vs 0.81) in identifying infecP72
Sepsis vs
Severe Sepsis
tion in SIRS with a best cutoff level of 1.31ng/ml for
ProCT and 13.9 mg/dl for CRP. Parallel to SOFA and
MODS scores, there was a consistent increase in ProCT
levels with sepsis severity (Table 1).
Plasma ProCT values were higher in nonsurvivors in septic
patients, and area under the ROC curve of ProCT in prognosticating outcome was higher than for both CRP and
WBC. The present data demonstrated ProCT as a better
diagnostic marker of infection in patients with inflammatory
response. This laboratory procedure also seemed to be
closely correlated with sepsis severity and outcome.
Procalcitonin serum concentrations do not differentiate severity of septic shock in postoperative
M Weiss*, E Barth*, G Fischer*, H Wiedeck*, EM Schneider*, M Georgieff* and G Steinbach†
Departments of *Anesthesiology and †Clinical Chemistry, University Hospital, D-89070 Ulm, Germany
Objectives: To investigate whether procalcitonin (PCT)
serum concentrations differentiate severity of disease in
postoperative/post-traumatic patients with septic shock, in
which severity of shock is defined by different ranges of
dosages of norepinephrine, necessary for cardiovascular
Methods: Over a six month period, 192 patients admitted
to the intensive care unit (ICU) were studied. Out of them,
18 patients were in a first phase of septic shock lasting at
least four days. PCT serum concentrations in these
patients were measured on days 1, 2, 4, 6, 8, 10, 12 and
14 after onset of septic shock until death or discharge from
ICU. PCT was measured using a commercial immunoluminometric assay (BRAHMS Diagnostica, Berlin).
Results: Nine patients survived the episode of septic
shock lasting 4–10 days. Nine patients died after 5–55
days. In six patients, maximal dosage of norepinephrine
was ≤0.1 µg/kg×min (group 1), in five patients between
>0.1 and ≤0.2 µg/kg×min (group 2), and in seven patients
>0.2 µg/kg×min (group 3). Ranges of PCT serum concentrations overlapped between group 1, 2, and 3, with
1.2–67.6, 1.8–194.3, and 0.4–46.1 ng/ml of PCT,
respectively, resulting in no significant differences of PCT
values during septic shock between the three groups.
Conclusions: PCT serum concentrations are not a marker
of severity of disease in patients with septic shock, if
severity of shock is defined by certain ranges of dosages
of norepinephrine.
Expression and regulation of procalcitonin in different human cells
S Rußwurm, I Stonans, M Wiederhold, M Oberhoffer, M Meisner, PF Zipfel and K Reinhart
Clinic of Anesthesiology and Intensive Care Medicine, University of Jena, Bachstr. 18, D-07740 Jena, Germany
Introduction: Procalcitonin (PCT) was recently forwarded
as a diagnostic marker of systemic bacterial infection and
sepsis. The biological function(s) and biochemical properties of this protein are poorly defined and the cellular
sources of plasma PCT remains yet to be established.
Methods: Primary human cells (peripheral blood monocytes, umbilical vein endothelial cells) and cell lines (liver,
renal parenchymal and lung fibroblastic lines) were cultivated under standard conditions. Basal and stimulated
mRNA expression of PCT was investigated using a semi-
quantitative reverse transcriptase polymerase chain reaction (RT-PCR). Intracellular PCT protein expression was
verified by Western blotting and surface-enhanced laser
desorption/ionization (SELDI). Experiments elucidating the
intracellular location of PCT were performed after protein
fragmentation in different fractions by secondary immunofluorescence and laser scan confocal microscopy.
Results: (1) A basal and inducible mRNA expression of
PCT was found only in human peripheral blood monocytes.
(2) In these cells, a distinct influence of various proinflammatory mediators was observed. (3) Western blotting of
monocyte lysates using various primary antibodies directed
against PCT showed a strong intracellular protein expression. (4) Experiments with SELDI revealed a molecular
weight for PCT in monocytes of 12.1 kDa. (5) Human
monocytes express PCT protein in association with
cytoskeleton. No PCT was found in cytoplasmic fractions.
Conclusions: Since human peripheral blood monocytes
produce PCT and its expression depends strongly from
sepsis-related mediators, we conclude, that this cell population is one important source of elevated PCT serum
levels during sepsis. Further experiments analyzing the
role of Kupffer cells and liver parenchymal cells are in
Procalcitonin helps to discriminate between septic and non-septic underlying disease at
admission in ICU
C Lepousé, B Hamou Ouali, J Cousson, P Raclot, JL Suinat and A Leon
Hôpital R. Debré, 51092 Reims cedex, France
Introduction: Differentiation between acute bacterial
infection from other types of inflammation is often difficult
in ICU. Procalcitonin (PCT) was reported to be a new
potential specific marker for infection. The aim of this
study was to assess, at the admission, a PCT cutoff of
infection in ICU patients.
Methods: In a prospective study, we studied all patients
admitted to our ICU between January 1999 and July
1999. PCT was measured at admission (J0) and on
second day (J2). Patients were grouped according to
Bones classification (SIRS, sepsis, severe sepsis, septic
shock, SDMV). Statistical analysis was performed using
Results: 377 consecutive patients (234 males and 143
females, mean age 60.3±16.9 years, mean IGS II
34.12±17.3) were admitted to ICU. Mortality rate was 26%,
mean length of ICU stay was 7.14±8.1 days. At admission
and J2, mean PCT for all of them was 16.4±57 ng/ml and
27.1±100 ng/ml. According to Bones classification, mean
positive value
J0: PCT ≥1.5
J0: PCT ≥3
J2: PCT ≥1.5
J2: PCT ≥3
PCT was: without SIRS=1.74 ng/ml (n=213), SIRS=
1.51 ng/ml (n=33), sepsis=14.48 ng/ml (n=43), severe
sepsis=35.56 ng/ml (n=41), septic shock= 70.85 ng/ml
(n=30), SDMV=94.95 ng/ml (n=13). The PCT cut off of
infection (Bones ≥ sepsis) is shown in the Table.
Discussion: At admission to ICU, PCT seems to be an
interesting marker of early diagnosis of infection when the
level is higher than 1.5 ng/ml. On the second day, a PCT
level of 3 ng/ml seems to be more effective for diagnosis
of infection than 1.5 ng/ml.
Procalcitonin concentrations in blood after long-distance running
A Leon*, C Lepousé*, T Floch*, X Schwartz* and N Schneider**
*Laboratoire d’Anesthésie-Réanimation, Faculté de Médecine and **Laboratoire de Radioanalyse et Immunologie, Institut Jean
Godinot, Reims, France
Incidence of injuries and other health problems is known
during prolonged running and recent studies suggest that
exercise-induced damage may trigger production of proinflammatory molecules.
The aim of the study: The present study investigates to
what extent prolonged strenuous exercise influences the
plasma concentrations of procalcitonin as well as
cytokines inhibitors and anti-inflammatory cytokines.
Methodology: After informed consent, 4 marathon
runners and 10 half-marathon runners were recruited and
investigated 12 h before running, immediately after the
end of the run and 24 h later. Blood samples were collected at each time and PCT concentrations in plasma
were measured with a luminometric method (Brahms). The
plasma concentrations of TNF α, IL-1 β, IL-6, IL-1 ra,
sTNF-r1, sTNF-r2 were measured by ELISA. Statistical
analysis used Systat (Jandel).
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Results: The mean duration for marathon running was
235±18 min and 105±10 min for half-marathon running.
The highest concentration of IL-6 and TNF α was found
immediately after the run. Time course (1 =12 h before; 2
=immediately after and 3 =24 h later) of blood PCT, ng/ml
is presented in the figure (P<0.001).
Discussion: This study suggests that long-distance
running may trigger production of TNF α and pro-inflammatory cytokines, which stimulate the production of PCT.
As it has been demonstrated in sepsis, PCT may be an
early marker of elevated cytokines and muscle and/or
splanchnic damage.
The role of gallium-67 scintigraphy in diagnosing sources of fever in critically ill patients in ICU
I Rozet, S Ben-Haim* and S Burzstein-De Myttenaere
Intensive Care Unit, Department of Anesthesia, and *Nuclear Medicine, Carmel Medical Center, Haifa, Israel
Introduction: The role of radionuclide imaging to diagnose sources of infection is rather confusing [1].The aim
of this study was to evaluate the diagnostic value of
gallium-67-scintigraphy in critically ill febrile patients and
its impact on treatment and outcome.
Methods: Forty whole-body Ga-67 scans performed in 37
patients in a five year period (1994–1999) were retrospectively analyzed.
Result: 34 patients were surgical and 3 medical with a
median age 69 years and median APACHE II 23. All
patients were febrile with systemic inflammatory response
syndrome (SIRS) and all, except two, were ventilated for
more than three days. Ga-67 scan was performed either
within the first two weeks (median 7 day) because of no
improvement of sepsis (group 1) or later than two weeks
(median 22 day) because of the resurgence of SIRS
(group 2). Group 1 consists of 24 patients with APACHE
II 21, multisystemic organ failure (MSOF) in 6 patients and
5 exitus. Four from the 24 Ga-67 scans showed positive
findings (two soft tissue abscesses, one intra-abdominal
collection and one osteomyelitis) leading to changes in
treatment and improvement of outcome. Group 2 consists
of 13 patients with an APACHE II 24, MSOF and exitus in
7 patients. On the 16 Ga-67 scans performed in this
group, only one showed retrosternal Ga fixation and had
no impact on treatment and outcome.
False negative results were observed in 9 of 15 patients
with known lung and thoracic pathology proved by other
investigations. In spite of rather high sensitivity in patients
with abdominal pathology, the results were non-specific.
Conclusion: Ga-67 scan in the critically ill may be helpful
in localizing a pyrogenic focus in severe unimproving
septic patients in the early phase after their admission.
However, in patients with resurgence or sepsis in the late
phase of hospitalization, Ga-67 scan appeared to be of
very little diagnostic help and its value remains questionable.
Peters A: Br J Radiol 1998, 71:252–261.
Bacteremia and systemic inflammatory response syndrome (SIRS) in confirmed post-partum
P Myrianthefs, K Varvaresou, C Ladakis, S Pactitis, V Lappas, J Stamatiou, A Carousou, J Pavlides and J Baltopoulos
Department of Intensive Care Unit, “IASO” Maternity Hospital, Kifissias 37–39, Marousi, Athens, Greece
Background: Postpartum endometritis is the most
common infectious complication of pregnancy, which has
a serious impact on women’s health and can lead to
severe sepsis.
Material and methods: We encountered 150 women
with clinical evidence of early (1–48 h) post-partum
endometritis and SIRS (fever, tachycardia, tachypnea and
increased white blood cell count and C-reactive protein)
after either vaginal delivery or caesarian section. We
obtained blood cultures (aerobic and anaerobic) at
38.5°C and before administration of antibiotics.
Results: Fifteen samples (10%) were positive proving
bacteremia of which 12 (80%) were indicative of
aerobic and 3 (20%) of anaerobic bacteria. There were
also 7 cases of contamination due to Staphylococcus
epidermidis. Isolated aerobic Gram negative bacteria
Pseudomonas cepacia (1), Enterobacter aerogenes (1),
Salmonella group C (1), Klebsiella pneumoniae (1) and
Gram positive cocci (2/15, 13.3%) were Enterococcus
faecalis (1) and Streptococcus group A (1). Anaerobic
bacteria (3/15, 20%) were Bacteroides fragilis (2) and
Clostridium sp (1). All patients were treated empirically
until antibiotic sensitivity tests were available with ceftriaxone 2 gr×1 iv, netilmicin 300 mg×1 iv and metronidazole 500 mg×3 iv. Outcome was successful in all
patients. None of the isolated bacteria was multi-drug
Conclusion: Bacteremia in early post-partum endometritis
was confirmed in 10% of cases. Isolated bacteria were
similar to those referred to in the literature (80% Gram
negative bacteria, 20% Gram positive and anaerobic bacteria). Early recognition and appropriate management
ensures successful outcome and prevention of complications such as severe sepsis.
Evolution of micro-organisms isolated in nosocomial bloodstream infections in Belgian Intensive
Care Units, 1992–1999
B Jans, C Suetens, H Carsauw, I Morales, E Leens and P Selway
Scientific Institute of Public Health (IPH), J. Wytsmanstreet 14, 1050 - Brussels, Belgium
Objectives: To compare the distribution of micro-organisms from nosocomial bloodstream infections (BSIs) in
Belgian intensive care units (ICUs) during two surveillance
periods: from October 1992 until June 1996 and from
October 1998 until September 1999.
Table 1
Methods: Since 1992, Belgian hospitals have been
invited to participate on a voluntary basis in the national
surveillance of nosocomial BSIs (NSIH). Surveillance data
are transmitted by the hospitals to the IPH for analysis,
national benchmarking and feedback. The surveillance
was discontinued in July 1996 and started up again on 1
October 1998. We compared the distribution of
pathogens isolated in BSIs occurring >48 h of hospital
stay before and after the interruption of the surveillance.
Distribution of micro-organisms isolated in nosocomial BSI’s
in Belgian ICUs, 92/96 and 98/99:
Gram-positive cocci
Other gram-negative bacilli
Yeast and fungi
Gram-positive bacilli
Other micro-organisms
Results: During the first period a total of 3563 isolates
were reported from 104 hospitals, for the second period
925 isolates from 62 hospitals. Table 1 shows the distribution of the micro-organisms by type. The most frequently
isolated micro-organisms were coagulase-negative
staphylococci (1992–1996: 22.6% and 1998–1999:
22.8%), Staphylococcus aureus (13.1%;11.3%), enterococci (7.8%;9.8%), Escherichia spp (7.7%;8.6%), and
Enterobacter spp (7.1%;9.4%). The proportion of Enterobacter spp. increased significantly from 7.1% in
1992–1996 to 9.4% in 1998–1999 (P=0.017). This
increase was due to an increase of Enterobacter aerogenes isolates (from 3.2% to 4.7%, P=0.025). The proP79
portion of Enterobacter cloacae isolates remained constant (3.1% to 3.6%). A significant decrease was
observed for Citrobacter spp (0.9% to 0.2%, P<0.03) and
Acinetobacter spp (2.2% to 1.2%, P=0.043).
Conclusions: Overall, the distribution of micro-organisms
in BSIs in Belgian ICUs has remained quite stable since
1992. The earlier reported increasing trend in Enterobacter aerogenes BSIs was again confirmed.
Ten years of blood cultures surveillance in an intensive care unit of a tertiary hospital in Israel
R Kitzes*, S De Myttenaere-Bursztein, A Sheskin and Z Stoler
*Clinical Pharmacology and Infectious Diseases Unit, Intensive Care Unit#, Department of Community Medicine and
Epidemiology Carmel Medical Center, The Rappaport Faculty of Medicine, Technion, Haifa, Israel
Carmel Medical Center is a tertiary hospital with 430
beds. The intensive care unit (ICU) has 7 beds and
admits patients from all hospital departments. The objectives of this study were to analyze the etiologies of blood
cultures (BC) in ICU and to compare them to those
obtained from the departments. The study period
included retrospectively ten years from 1988 to 1998,
excluding 1990. During this period 4497 positive BC in
different patients were reported from the hospital. 3976
BC including 835 from ICU were analyzed. These BC
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
represent 90% of most commonly found Gram positive
and negative bacteria. The distribution of Gram negative/Gram positive BC was respectively 61%/39% for
the ICU and 53%/47% for the departments. This difference is statistically significant. In decreasing order of
incidence the most important bacteria in the ICU were:
Klebsiella 23.8%, Pseudomonas 15.6%, MRSA 11.9%,
MRSE 10.5%, Acinetobacter 8.7%, Stapylococcus
aureus 6.7%, E. coli 5.4%, Enterococcus 4.7% and
Enterobacter 4.7%. The most important bacteria from
departments were: E. coli 19.9%, Klebsiella 13.2%,
Staphylococcus aureus 11.4%, Streptococcus 9.2%,
MRSE 8.2%, Pseudomonas 7.1%, Acinetobacter 5.3%,
Enterococcus 5.2%, Staphylococcus epidermidis 4.9%,
MRSA 4.1%.
Conclusion: In spite of continuous transfer of patients
from departments to ICU and from ICU to departments,
the etiology of BC is different. E. coli and Staphylococcus
aureus are the most predominant Gram negative/positive
bacteria from the hospital departments, reflecting community acquisition. Klebsiella and MRSA are the most predominant Gram negative/positive bacteria in the ICU. The
etiologies of BC remain stable throughout the years for
the majority of bacteria excluding outbreaks of MRSA and
Acinetobacter in the ICU.
Intraluminal antibiotic treatment of central venous catheter infection in patients receiving total
parenteral nutrition
V Alivizatos, S Rathosis, D Skarpetas and P Athanasopoulos
Department of Surgery, Nutrition Unit, “St. Andrew” General Hospital, Patras, Greece
Catheter guidewire exchange has been suggested as an
effective method of treatment of catheter-related infection
(CRI) in nontunneled central venous catheters. However,
the old line is usually intraluminally colonised, so contamination of the new line through the guidewire occurs early.
To minimise this event, we evaluated a method of treatment consisting of catheter guidewire exchange followed
by intraluminal antibiotic administration.
Methods: Eight consecutive patients with CRI were
studied. They received TPN through a subclavian vein
catheter for a period of 8–19 days prior to CRI. In the
suspicion of CRI, the catheter was removed and
exchanged over a guidewire with a new catheter, followed by catheter tip culture and peripheral blood
culture. TPN administration was stopped for 2–3 days
until isolation and identification of the responsible bacteria. After identification of the bacteria, TPN administration
was started again and the appropriate antibiotic, as indicated by microbiological sensitivity, was administered
through the new central venous catheter for a period of
10 days. All new catheters were removed after 10 days
and cultured.
Results: In all the patients was noted defervescence of
the clinical signs of infection following catheter removal.
The responsible pathogens were Staphylococcus (S. epidermidis in 4, S. aureus in 3, all sensitive to vancomycin),
and Enterococcus in one (sensitive to ampicillin). There
were no further problems with the new catheters in 7
patients, and the new cultures were sterile after 10 days
of intraluminal antibiotic treatment; however, one patient
developed a new episode of CRI in the 9th day of treatment, and the catheter tip culture showed colonisation by
Proteus. In this patient, CRI was finally treated by removal
of the catheter and replacement at a different site.
Conclusion: Our results suggest that, in the majority of
cases, CRI can be successfully treated by exchange of the
catheter over a guidewire followed by administration of the
appropriate antibiotic through the new catheter for at least
10 days.
Urgent orotracheal intubation induces transient bacteremia in critically ill patients
BJA Rijnders, A Wilmer, J Van Eldere and E Van Wijngaerden
UZ gasthuisberg, Inwendige ziekten, Herestraat 49, B-3000 Leuven, Belgium
Objectives: To test the hypothesis that urgent or semiurgent orotracheal intubation can induce short-lived bacteremia with oral flora. To find predictive factors for
post-intubation bacteremia.
Methods: Prospective study in a 17 bed medical ICU.
Patients in need of orotracheal intubation (OI) could be
included if no cardiopulmonary resuscitation was performed. One aerobe and anaerobe blood culture (BC)
was taken immediately before OI, as soon as possible
(preferable less than 10 min) after and 60 min after OI. If it
was impossible to take a BC before intubation, patients
could be included if both BCs after OI were taken. The
indication for OI, ease of OI, experience of the doctor
(<10,<100 or >100 OI previously) and the antibiotics
used before OI were registered.
Results: During the study period (16 weeks) 70 patients
underwent OI. Twenty-eight patients were included. Lack
of inclusion was always due to the urgency of the situation, which caused delay in taking the first post-intubation
BC. Four patients (4/28 or 14%) had streptococcal bac-
teremia immediately after OI (mean of 11 min, SD 2.5). No
patient (0/28) had streptococcal bacteremia 60 min after
OI (P=0.03). Three of the four patients showed streptococcal bacteremia after OI had to be performed by a
second doctor because of difficulties experienced during
OI whereas this was the case in only 4/24 in the control
group (P<0.001). Two of the four doctors intubating the
four patients with subsequent streptococcal bacteremia
had previous experience of fewer than 10 OIs. Four of 24
doctors in the control group had previous experience of
less than 10 OIs (P=<0.001). In other terms, three of the
seven patients (43%) that needed to be intubated by a
second doctor due to difficulties during OIs showed transient streptococcal bacteremia.
Conclusion: Non-elective (urgent or semi-urgent) OI can
cause transient bacteremia with Streptococcus species in
a significant proportion of medical intensive care patients.
The observed frequency of bacteremia is higher than previously reported after elective OI. The difficulty of OI might
be a major predisposing factor.
Tetanus – following up 285 patients in ICU
JS Brauner, SRR Vieira and CA Franke
ICU at Nossa Senhora da Conceição Hospital and ICU at Clinicas Hospital of Porto Alegre, Porto Alegre, RS, Brasil
Introduction: Tetanus is a toxic disease generally prevented by immunization, but it is still present in the Third
World and it shows high mortality rate.
Objective: To evaluate morbidity and mortality of tetanic
patients in ICU during 18 years of follow up.
Methods: This trial is an historical and a prospective
cohort that studied 285 patients from October 1981 to
October 1999. They were classified in two groups
according to modified Ablett’s scale: not severe (mild and
moderate) tetanus and severe tetanus. The following variables were compared: incubation period (IP), onset period
(OP), symptomatic period (SP), age, period of autonomic
instability (AI), period of administration of benzodiazepine
(Pbenzo), curare (Pcur), mechanical ventilation (PMV),
clinical and infectious events and mortality during ICU
follow up. In statistical analyses, the continuous variables
were presented as the mean and standard deviation.
Student t test was used to compare the two groups. The
level of significance was P<0.05.
Results: This trial evaluated 71 patients with not severe
tetanus and 214 patients with severe tetanus. The group
with severe tetanus had a smaller IP and OP (P=0.002
and P=0.003, respectively), but a longer Pbenzo, Pcur
and PMV (P=0.001) than the not severe group. There was
no difference concerning the mean age between the two
groups (P=0.26). The most common clinical events in
tetanic patients were cardiorespiratory arrest (25.6%),
pneumothorax (10.17%) and acute renal failure (10.17%).
The incidence of respiratory, urinary and catheter infections were 76.14%, 38.24% and 7.71%, respectively. The
mortality rates were 4.28% in not severe tetanus and
28.9% in severe tetanus (P<0.001).
Conclusions: The severe tetanus group presents a high
mortality rate, probably as a result of autonomic instability,
despite intensive care. Infections related to prolonged
mechanical ventilation and invasive procedures were the
most frequent events noticed in this study.
Leptospirosis in intensive care units: report of 33 cases
SRR Vieira, JS Brauner, F Fernandes and C Moraes
ICU at Clinicas Hospital of Porto Alegre and ICU at Nossa Senhora da Conceição Hospital, Porto Alegre, RS, Brasil
Introduction: Leptospirosis is, in general, a self-limited
disease but it can be associated with important complications such as multiple organic dysfunction and high mortality [1].
multiple organ dysfunction and their mortality rate. We
have also compared survivors with non-survivors. The
quantitative variables have been compared by unpaired
t-tests and the qualitative variables by a Chi squared test.
Objective: The goal of this paper is to evaluate the clinical
characteristics and the morbimortality of severe leptospirosis in general Intensive Care Units from two general
Results: We described 33 adult patients, aged 40±16
years, of which 27 were men and 6 women. The most frequent clinical manifestations were fever (n=30), myalgias
(n=29), jaundice (n=27) and dyspnea (n=27). All patients
showed some level of organic dysfunction such as respiratory (n=29), renal (n=24), hepatic (n=24) and cardiovascular (n=20). The mortality rate was 52% (n=17). The
comparison of non-survivors with survivors showed they
have higher incidences of respiratory, renal, cardiovascular
Methods: All cases with the diagnosis of leptospirosis
confirmed by blood macroagglutination test and admitted
from 1990 to 1999 were studied. We have analyzed their
clinical and laboratory characteristics, the occurrence of
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
and neurological failures as well as higher levels of acidosis (P<0.05).
Conclusions: In endemic regions leptospirosis has to be
considered as a cause of multiple organic dysfunction
with a high mortality rate mainly when respiratory, renal,
cardiovascular or neurological failures are present.
Ko AI et al.: Urban epidemic of severe leptospirosis in Brazil.
Lancet 1999:820–825.
Falciparum malaria in ICU
A Aaron and B Avinash
Parsee General Hospital, Cumballa Hill, Mumbai 400036, India
Objectives: To study and compare the various treatment
modalities and the attendant complications of Falciparum
malaria with or without concomitant Vivax in the Intensive
Care Unit between 1993 and 1999.
Methods: Retrospective study of patients hospitalized
between 1993 and 1995 and prospectively from 1996 to
Results: See Table.
Conclusions: 1) Over the period of years, early use of
Quinine has given good and consistent results (10 mg/kg,
up to 600 mg thrice daily for 5–10 days) and can cure all
stages of Falciparum and other Plasmodia, save the hyp-
nozoites of Vivax. 2) Parasites have developed resistance
to Chloroquine, Sulfadoxine- Pyremethamine. Quinine is
poorly tolerated when given after Mefloquine due to the
addition of adverse effects especially those of GIT. 3) If
patients who have been treated with chloroquine or other
drugs, including if radical cure has been attempted, continue to spike, it is best to restart with quinine even if a
species other than Falciparum is the culprit. 4) Hypoglycemia, otological complications were not really severe
enough or irreversible to warrant stopping Quinine.
Although ECG(QTc) or other cardiac disturbances were
not seen in those group,a daily ECG is a must. 5) In
severely toxic patients, a simultaneous intramuscular dose
of 60 mg of Artesunate on each buttock followed by a
daily dose of 60 mg for the next 5 days reduces the para-
Total admissions in the ICU:
Quinine group: 40 patients
Non Quinine group:
All Started with Quinine 10 mg/kg
(up to max 600 mg thrice a day for 5–10 days)
IV Fluids esp. Dextrose
Antibiotics esp. Doxycycline
56 patients
Started either with Chloroquine, SulfadoxinePyremethamine, Mefloquine with or without
Halofantrine, Doxycycline, Norfloxacin
Artemesnin derivatives and others
Other parameters like Mechanical ventilatory support, Blood transfusions, Positive Ionotropes were used indiscriminately for all patients, as the
situation demanded
CNS - Drowsy, altered sensorium responded
well by the 3rd dose of Quinine
CNS – Disorientation, Unconsciousness,
coma, death
Respiratory- Reversible with prompt treatment
and others recovered well after mechanical ventilaton
Respiratory – Respiratory Failure, ARDS
Renal Failure: High Creatinine due to microvascular
ischemia returned to baseline over a period of time
Renal Failure – High creatinine, Anuria, ATN
DIC, Shock, Multi organ failure were not encountered
Hepatic failure, Multi Organ Failure, Hemolysis, Circulatory failure, DIC
Ominous signs of Falciparum
1) Peripheral smear: due to invasion of young RBCs and sequestration in tissues, a large number of parasites may be invisible on the smear,
thereby giving false assurance of a “low parasite index”.
2) Premunition: in holo and hyperendemic areas, due to some degree of acquired immunity, patients are largely asymptomatic early on. It is almost
always late when symptoms manifest overtly.
3) Cerebral Malaria is a “diffuse symmetrical encephalopathy” and therefore sparing mild neck rigidity, signs of meningeal irritation are largely
absent. Mild delerium, obtundation, convulsions etc. are to be taken seriously.
4) Pregnant women are spared gross parasitemia for a long time due to heavy “parasitisation of the placenta”.
minimum 72 h before discontinuing quinine or switching to
an alternate regime. 7) Intravenous Quinine may be used
only in the most toxic and orally intolerant population. 8)
Primaquine must be given in Chloroquine resistant Vivax
even after usage of Quinine for radical cure.
site load as it is relatively new and thereby less resistant
and is an effective schizonticide. It is, however, devoid of
any action on the gametocytes. This was the only other
group of drugs which promised future potential for concomitant use with Quinine. 6) It is best to wait for a
Fungaemia in the intensive care unit: Baragwanath experience
B Zeippen and S Madhi
Intensive Care Unit, Chris–Hani Baragwanath Hospital, University of the Witwatersrand, Johannesburg, PO Bertsham 2013, RSA
Introduction: Despite advances in the care of the critically
ill, fungaemia often remains undiagnosed and caries a high
were reviewed (eight medical, twelve surgical, six paediatrics).
Results: See Table.
Objectives: To assess the demographics, the morbidity
and mortality of patients infected with Candida species.
Conclusion: Despite treatment, fungaemia carries a high
mortality.The actual mortality of 69% is higher then predicted by the APACHE II/PRISM scores.
Materials and methods: The intensive care unit records
of 26 patients diagnosed with fungaemia during 1998
Total (n=26)
Alive (n=8)
Dead (n=18)
Age (years)
Median (range)
47 (8d–76y)
26 (10m–53y)
48.5 (8d–76y)
APACHE II (n=20)
Mean (SD)
16.9 (7.4)
11.6 (6.1)
18.6 (7.0)
PRISM (n=6)
Mean (SD)
12.7 (5.8)
8.6 (5.0)
16.6 (3.1)
ICU stay (days)
Median (range)
13.5 (1–115)
15.5 (3–71)
11 (1–115)
Delay to Rx (days)
Mean (SD)
2.6 (3.6)
8.0 (1.5)
1.2d (2.6)
Mean time to initial fungaemia following ICU admission was 6.5 days (25th–75th: 2–14). The mean time of isolation of fungi from tracheal aspirate
(n=9), catheter tip (n=7), urine (n=6) was 4.4 (SD 4.4), 12.3 (SD 13) and 21.8 (SD 18.2) days respectively.
Response to therapy at 2 weeks:
Died during therapy
≤5 risk factors
High incidence of positive fungal DNA in the blood of critically ill patients
K Ikegami, T Sakai*, F Morimoto, T Seki, K Hirose and Y Sato
Trauma and Critical Care Center, Koshigaya Hospital, Dokkyo University School of Medicine, 2-1-50 Minamikoshigaya
Koshigaya-shi, Saitama, Japan 343-8555; *Department of Molecular Biology, Kyorin University School of Medicin, Japan
Introduction: Systemic fungal infection is not uncommon
in critically ill patients. Because of difficulty in the diagnosis, early antifungal therapy is not always possible.
Although positive blood culture for fungus is well recognized as a hallmark of systemic therapy, sensitivity of blood
culture is low. Detection of fungal DNA in the blood is
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
expected to serve as a new clue to initiate antifungal
Purpose: To assess the incidence of positive fungal DNA
in the blood of critically ill patients.
Methods: 34 blood samples were obtained from 24
patients. Blood samples were processed for polymerase
chain reaction (PCR) to detect fungal DNA. DNA extraction was performed with the modified methods of
Buchman. PCR primers to amplify conserved DNA
sequences of 18S rRNA genes shared by most clinically
important fungi were used. The amplification was performed in a thermal cycler through the cycles as follows.
Results: Fungal DNA was detected from the 16 samples
taken from 22 patients with SIRS and from seven samples
taken from 12 patients without SIRS. When patients
developed SIRS after seventh hospital day, fungal DNA
was detected in 80% of samples taken from these
Conclusions: Incidence of positive fungal DNA was more
than 50% and the incidence reached 80% when patients
developed SIRS after seventh hospital day. High incidence of positive fungal DNA suggests that there are
more candidates for systemic antifungal therapy than
Frequency, mortality and risk factors of candidemia at a tertiary care hospital
J Garbino, P Rohner, Th Kinge, P Pichna, D Pittet and D Lew
Infectious Diseases Division University Hospital of Geneva, Geneva, Switzerland
Introduction: The incidence of fungal infections has been
increasing for the last three decadess, especially in neutropenic, cancer and critically ill patients. It is associated
with high mortality rates.
Methods: We retrospectively reviewed medical charts of
adult patients with fungemia from 1989 to 1998 at the
University Hospital of Geneva. For patients with candidemia, demographic information, risk factors (antibiotic
treatment, preceding surgery, cancer, intravascular
devices, ICU stay), outcome, and the incidence of fungal
species was analyzed as well as the overall consumption
of fluconazole.
Results: Of 291 patients with fungemia, we reviewed 278
(96%) medical records, 184 (66%) were from male
patients. The mean age was 60 years and the overall mortality 45.5%. Candida albicans (61%) was the most commonly identified species followed by C. glabrata (15%),
and C. parapsilosis (6%). The incidence of candidemia
ranged from 0.2 to 0.56 per 10 000 patient-days with the
highest incidence in 1993 and the lowest in 1997. Of the
risk factors analyzed only ICU stay was statistically significant (P<0.0001) (OR 3.42; CI 1.98–5.92). When candidemia was revealed, 38% of the patients were
hospitalized in an ICU. Most candidemias occurred
several days after admission, that is 52% after two weeks
and 70% after three weeks.
Conclusion: The regular use of fluconazole since 1990
may explain a decreasing trend of candidemias due to C.
albicans, whereas those due to other Candida spp
remained stable. The predisposing risk factors analyzed
associated with a high mortality were similar to those
described in other series. For the last two years, fungemia
remained low and stable in our institution. No significant
changes in the incidence of fungal species during the 10
year study period were observed.
Hospital urinary infection epidemic by Candida famata: reduction of mortality after therapy with
fluconazole and identification of the risk factors related to the infection
MS Lapchik*, E Andrioli*, M Melo†, P Junior†, LC Leiro* and EF Silva*
*Hospital Evaldo Foz, Av. Vereador J. Diniz, 3505, São Paulo, Brazil, CEP–04603-901; †Laboratório SAE, São Paulo, Brazil
Background: The more recent studies have revealed that
infections by Candida spp. have increased their participation
as a hospital infection. Some clinical studies have revealed
reduction of the mortality rates in critical patients treated for
candiduria with fluconazole. In 1997 and 1998, we detected
an epidemic outbreak of hospital urinary infection by C.
famata, which determined an epidemiological investigation.
Objectives: Identify the risk factors for hospital candiduria
by C. famata, the lethality attributed to this infection, as
well as evaluate the impact of early treatment with fluconazole on the lethality of candiduria.
Results: This matched paired case-control study detected
19 cases of hospital infection by C. famata, fulfilling 100%
of the pairing criteria with 38 patients of the control group.
The univariate analysis revealed the following variables as
risk factors for hospital candiduria by C. famata (P<0.05):
Post-operative of abdominal surgery, previous use of antimicrobials with coverage for aerobic Gram-negative bacilli and
anaerobic bacteria, the prolonged use of vesicular catheter
(>4 days) and of antimicrobials. The greater time of use of
antimicrobials was considered the only independent risk
factor, obtained by multivariate analysis (OR=1.1). The
lethality attributed to hospital candiduria by C. famata was
42% (P<0.05), with identification of secondary infection in
the blood stream in two patients. All the patients who did
not make use of fluconazole or other anti-fungal therapy for
treatment of the candiduria (7 patients) evolved to obit.
(P<0.05) No potential sources of infection were identified,
with an endogenous source being attributed to as most
probable for the occurrence of the outbreak.
Conclusion: The diagnosis of candiduria in critical patients
remains as a characteristic related to the high morbidity and
lethality. The institution of preventive measures such as the
reduction of time of use of vesicular catheter, rational use of
antimicrobials, and the early treatment of candiduria
episodes in critical patients, are measures which can reduce
the lethality related to this infection, considering candiduria
as a predisposing factor to systemic candidiasis. The
research of C. famata as etiological agent in hospital urinary
infection must be taken into account in future studies.
Usefulness of vancomycin serum concentration monitoring in the critically ill patient
M Bertolissi, A Di Silvestre, F Pea*, M Furlanut* and F Giordano
Department of Anesthesia and ICU 2° General Hospital, * Department of Clinical Pharmacology University, 33100 Udine, Italy
Introduction: The aim of this study was to compare two
methods, the serum concentration monitoring (SCM) and
the nomogram of Moellering (NM), employed to calculate
the dose of vancomycin (V) in the critically ill patient.
Table 1
Methods: We studied 32 patients admitted to our ICU and
treated with V for Staphylococcus aureus infections. The
patients were randomly divided into two groups: A (16) and
B (16). In group A the dose of V was calculated by the
SCM (fluorescence polarization immunoassay technique)
on two blood samples drawn daily just before the first
administration of the drug [to determine the minimal serum
concentration (C min)] and 30 min after the infusion of V
has been stopped [to determine the maximal serum concentration (C max)]. In group B the daily dose of V was calculated on the basis of the NM even though the SCM was
also performed but its result was not known. We fixed the
therapeutic range of V between 5 and 10 mg/l for C min [1].
C min <5 mg/l
5 mg/l < C min <10 mg/l
C min >10 mg/l
Results: In group A, C min of vancomycin was inside the
therapeutic range in all patients: In group B it was above
10 mg/l in 4 patients (14.3±2.7) and below 5 mg/l in 3
patients (3.9±0.3) (Table 1). C max was 20.4±3 mg/l in
group A and 20.1±6 mg/l in group B.
Discussion: The results obtained show that the SCM is a
useful tool to maintain the serum concentration (SC) of
vancomycin inside the therapeutic range, while the NM is
not so effective. The possibility of maintaining the right SC
of V is essential to prevent both the side effects from overdosing (oto, nephro-toxicity) and underdosing the drug
(uneffective antibiotic activity, bacterial resistance), but it
Number of patients with the minimal serum concentration (C
min) of V below, between and above the limits of the
therapeutic range.
can be problematic in the critically ill patient whose physiopathological characteristics may interfere with the pharmacokinetics of the drug. In the 4 patients who showed a
C min of vancomycin above 10 mg/l we did not find any
correlation between the elevated level of V and the clinical
conditions; in the 3 cases with C min below 5 mg/l we
observed an increase in the volumes of distribution
because of peritonitis and fluid overload.
We conclude sustaining the usefulness of the serum concentration monitoring to establish the right dose of vancomycin in the critically ill patient, and underlining that the
clinical conditions characterized by the increase in the
volumes of distribution represent a risk factor of underdosing the drug.
Leader WG et al.: Pharmacokinetic optimisation of vancomycin
therapy. Clin pharmacokinet 1995, 28(4):327–342.
Vancomycin as a single daily dose in critically ill patients on an intensive care unit
S Grundler, N Stallard, A Paull and MN Smithies
University Hospital of Wales, Heath Park, Cardiff, Wales, UK, CF4 4XW
Purpose: The purpose of the study was to assess the
peak and trough levels of 2 g of vancomycin (VAN), given
once daily (od), in critically ill patients, irrespective of renal
function, instead of lg VAN twice daily.
Methods: Prospective, observational, open study. 2 g of
VAN, diluted in 500 mls of 0.9% normal saline, infused over
four hours. VAN blood-level assays were done at 1, 2 and
12 h after completion of the infusion, and 12 hourly there-
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
after or until the next dose of VAN. (Random VAN levels are
done routinely on the ICU at 08h00 daily.) Another dose of
VAN was given if the random level was below 10 mg/l.
assess the effect on renal function, since most of our
patients requiring VAN were already on continuous renal
replacement therapy.
Results: Five subjects (2 male, 3 female) aged 16–64
(mean 56±23.5) years, weighing 65–105 (83±21.4) kg
were enrolled. The assay used to measure VAN levels was
the FPIA using the FLX system. The peak levels 1 and 2 h
after the infusion were between 21.5–50.6 (35.6±12.9
mg/l) and 20.4–48 mg/l respectively.
Concerns have been voiced regarding the short postantibiotic effect (PAE) of VAN (<2 h). In aminoglycides the
larger, od dose resulted in a prolonged PAE. The PAE of
once daily VAN is unknown and at present we are maintaining trough levels above 10mg/l.
Conclusion: Our peak levels did not exceed 80 mg/l, a
level which has been associated with ototoxicity. The incidence of nephrotoxicity with VAN alone is about 5% and
is unrelated to serum VAN concentrations. We did not
Our results show that od VAN 2 g should not cause an
increase in ototoxicity as peak VAN levels are well below
80 mg/l and if trough levels are maintained above the
minimum inhibitory concentration of the organism, then
concerns regarding the short PAE are eliminated.
Risk factors associated with vancomycin-resistant enterococcal colonization in a general ICU
and the effect of surveillance and prevention over respiratory tract and nosocomial infections
PS Martins, FR Perrechi, MH Kai, RC Zanella, F Valdetaro and S Blecher
Hospital Santa Marcelina, São Paulo, Brasil
Objectives: 1) To determine the risk factors associated
with vancomycin-resistant enterococcal (VRE) colonization
in a general ICU. 2) To observe the effect of infection
control measures over respiratory tract and nosocomial
infections in that unit.
Design: Point-prevalence surveys for VRE colonization of
patients were carried out over the period from March
1998 to June 1998.
Setting: General medical and surgical ICU in a tertiary
teaching hospital in the city of São Paulo, Brasil.
Measurements: The first case of VRE colonization in the
unit was related in March 1998. Point-prevalence surveys
were then carried out to evaluate the main risk factors
related to VRE colonization (enteral feeding, severity of
illness, length of ICU stay, colonization pressure, antimicrobial use, and others). A univariate analysis was made and
the variables that demonstrated P≤0.10 were then entered
into a discriminant analysis to identify the β coefficients.
Measures against VRE colonization were established
based on CDC criteria (MMWR 44(RR12);1–13) and the
effects of these measures over the respiratory tract infec-
tion rate and the nosocomial infection rate according to
the NNIS (National Nosocomial Infections Surveillance)
system were measured. Strain typing was characterized
using pulsed-field gel electrophoresis (PFGE).
Main results: Thirty-eight patients were evaluated. 18
VRE and 11 vancomycin-susceptible enterococcal (VSE)
colonizations were identified in anal swab, wound and
venous catheter. Out of 18 enterococcal strains isolated,
13 strains were E. faecium, distributed in 5 different
groups according to PFGE. The other 5 strains were E.
fecalis, distributed in 4 groups according to PFGE as well.
The results concerning the univariate analysis and the discriminant analysis (Table 1), the respiratory tract infection
rates and the nosocomial infection rates (Table 2) are given.
Conclusions: The colonization pressure, vancomycin use
and length of ICU stay ≥29 days were the main risk factors
associated with VRE colonization.
Measures against VRE colonization resulted in a significant decrease in the respiratory and nosocomial infection
rates. This decrease was probably due to reduction of
other pathogens colonization pressure.
Table 1
Univariate analysis
Discriminant analysis
OR (CI 95%)
7.07 (0.92–6.77)
Cutoff ≥5
12.5 (1.7–326)
Sensitivity 94%
Colonization pressure
36 (3.19–691)
Specificity 81.8%
Risk factors
Length of ICU stay ≥29 days
Table 2
Respiratory infection
Nosocomial infection
Risk factors of infection and molecular typing in ICU colonized patients with Enterobacter
E Carlier*, M Piagnerelli*, A Deplano‡, P Lejeune* and D Govaerts†
*Department of Intensive Care and †Department of Microbiology, A. Vésale Hospital, 6110 Montigny-le-Tilleul; ‡Department of
Microbiology, Erasme University Hospital, 1070 Brussels, ULB, Belgium
Objectives: To identify risk factors for Enterobacter aerogenes (Ea) infection in colonized intensive care patients
and compare by Pulsed-Field Gel Electrophoresis (PFGE)
infection strains to colonization strains.
Methods: During a prospective 8 month period, prescription order of perineal swabs has been undertaken to all
ICU patients hospitalized more than 48 h on the admission
day, discharge and on a weekly basis. The following risk
factors were studied by univariate and multivariate analyses: age, sex, underlying pathologies, antibiotic and corticoid therapies, SAPS II score, intravascular and urinary
catheters, surgical drain, nasogastric tube, use of mechanical ventilation, bronchoscopy, length of ICU stay, length
between hospital and ICU admission (H/ICU) and mortality rate. Antimicrobial susceptibility and clonal analysis by
PFGE were performed in colonized and infected groups.
Results: 204 patients were included: 170 without Ea colonization/infection (83.3%), 24 colonized (11.8%) and 10
infected (4.9%). The Ea infection in colonized patients
was associated with younger age: respectively 76.5 years
in colonized and 63.5 years in infected patients
(P=0.039); higher ICU stay (9 days versus 31 days;
P=0.007) mechanical ventilation (P=0.039) and bronchoscopy (P=0.026). PFGE profiles demonstrated two
major clones (profiles 1 and 2) in ICU colonized and
infected groups (45/52 strains tested). We calculated an
attack rate infection/colonization of 28% for clone 1 and
20% for clone 2.
Seven of eight infected patients (87.5%) had the same
genomic profile for colonization and infection using
Conclusions: Our results indicate that younger age, long
length of ICU stay, use of mechanical ventilation and bronchoscopy were risk factors for Ea infection in colonized
patients. Ea colonization is associated with a very high risk
of Ea infection by the same genomic strain. This may
guide antibiotic treatment in this particular population of
ICU patients.
The new way of postoperative pulmonary complications prophylaxis in oesophageal cancer
S Sviridova, I Nekhaev, M Kiselevsky, A Pleskov and E Gorobets
Russian Cancer Research Centre, 115478, 24 Kashirskoye shosse, Moscow, Russia
Objective: The efficacy of Neupogen (granulocyte colonystimulating factor) in prevention of pulmonary complications after transthoracic subtotal oesophagectomy (TSE).
Design: Prospective randomized double-blind clinical trial.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Materials: Twenty-eight oesophageal cancer patients
were treated with Neupogen 5 µg/kg s.c. one day prior
surgery, and on the 1st and 2nd days after TSE. Thirty-five
patients received placebo on the same days. No one
patient was basically neutropenic. The therapeutic
regimen, including choice of antibacterial agents, was
similar in both groups. The incidence of postoperative
pneumonia (according to CPIS criteria), sepsis and multiple organ failure syndrome (MOFS) was registered.
Results: Four (14%) patients in the Neupogen group
developed pneumonia postoperatively. There were no
cases of sepsis and MOFS in this group. White blood
cells (WBC) in peripheral blood samples on the third postoperative day reached 24.5±2.9×103 per µl.
On the contrary, severe pneumonia occurred in 16 (46%)
patients of control group including 5 (14%) cases of sepsis
and MOFS. WBC level was 8.1±3.4 ×103 per µl. The difference between the two groups was significant (P<0.05).
Conclusion: The stimulation of granulocytes with Neupogen decreases greatly the incidence of pneumonia and
sepsis after transthoracic oesophageal resections in
cancer patients, and probably after extended cancer surgical procedures in general.
The impact of late-onset ventilator-associated pneumonia on mortality in a Saudi-Arabian
W Djazmati, GA Oni, ZA Memish, G Cunningham, M Itani, Y Beyene and R McIntyre
King Fahad National Guard Hospital, ICU 1425, PO Box 22490, Riyadh 11426, Saudi Arabia
Introduction: Ventilator-associated pneumonia (VAP) has
been recognized as a major factor affecting mortality in
intensive care units (ICU). This study compared the relative impact on mortality of early- versus late-onset VAP in
an adult medical surgical ICU of a 550-bed tertiary care
hospital in Riyadh, Saudi Arabia.
Methods: All patients mechanically ventilated for more
than 48 h between November 1996 and December 1997
were included prospectively. VAP was diagnosed according to the definition of the Center for Disease Prevention
and Control (CDC), Atlanta. Early- or late-onset VAP was
defined respectively as the occurrence of VAP within or
after five days of intubation. Pathogens isolated were compared between the two groups. The mortality in the two
groups were also compared using univariate and multivariate (logistic regression) analysis.
Results: A total of 202 patients were included in the
study. Our incidence density rate of VAP was 16.8 per
1000 person days of ventilation. Fifty-one patients
(25.2%) developed VAP, of whom 22 (43.1%) were lateonset. The mortality rates for the early- and late-onset VAP
groups were 10.3% (95% CI=2.7 to 28.5%) and 36.4%
(95% CI=18.0 to 59.2%), respectively. In a univariate
analysis, the relative risk of mortality in the latter group was
3.5 (95% CI=1.1 to 11.7; P=0.029). S. aureus was significantly more common in early- compared to late-onset
VAP (44.8% vs 13.9%; P=0.02), while Acinetobacter was
more common in late- compared to early-onset VAP
(40.9% vs 17.1%; P=0.06). In the logistic regression in
which adjustment was made for the effects of pathogens,
age and the administration of H2 receptor antagonists,
late-onset VAP remained significantly associated with mortality (risk ratio=6.0, 95% CI =1.1 to 33.9; P=0.042).
Conclusion: Our study shows significant association of
late-onset VAP with mortality. This supports what has
been observed in previous studies. It is possible to use the
time of occurrence of VAP as an important indicator of
mortality, and may be used to identify ‘at risk’ patients and
initiate more aggressive therapy at an earlier stage.
Effect of mechanical ventilation and respiratory tract colonization on bronchial immunoglobulin
A and G levels in ICU patients
A Kalopissi, X Katsenos, E Kremasmenou, A Anthi, K Mandragos, K Tergoglou and P Behrakis
Intensive Care Unit, Red Cross Hospital, Athens, Greece
Objective: The aim of this study was to evaluate the
potential effect of mechanical ventilation and colonization
of respiratory tract on bronchial IgA and IgG levels.
Patients and methods: Fifteen patients under mechanical ventilation, 10 males and 5 females, sustaining cerebrovascular injury, were included in the study. All
patients underwent blood and bronchial immunoglobulin
determinations on day 1 (intubation day) and day 4. At
the same time bronchial cultures were obtained. None of
the patients developed lung infection during the study
Results: Mean IgG and IgA plasma levels were 757.72
mg/dl (day 1) 738.81 mg/dl (day 4) and 163.45 mg/dl
(day 1), 185.18 mg/dl (day 4), respectively.
Mean IgG and IgA bronchial levels were 35.95 mg/dl (day
1), 29.69 mg/dl (day 4) and 17.84 mg/dl (day 1),
14.76mg/dl (day 4), respectively. No significant differ-
ences were observed between the two measurements of
immunoglobulins (day 1,4). In contrast, bronchial cultures
were positive (>104 cfu/ml) in 13 out of 15 patients on
day 4. Bronchial cultures obtained on day 1 were negative
in all the patients.
Conclusion: The colonization of the respiratory tract
occurring under mechanical ventilation doesn’t alter significantly the bronchial immunoglobulin levels. Although the
slight decrease observed may account for the high incidence of pneumonia in ICU patients.
Effects of bronchoalveolar lavage (BAL) fluids of patients with ventilator-associated
bronchopneumonia (VAP) on alveolar cells in culture
M Nys*, P Damas*, G Deby*†, M Mathy†, S Legrand‡ and M Lamy*†
*Service of Anaesthesiology and Intensive Care, University Hospital of Liege; †Center of Oxygen Research and Development,
University of Liege; ‡Laboratory of Fundamental Virology, University of Liege, B4000, Sart-Tilman, Liege, Belgium
Lungs of patients with VAP or ARDS are invaded by neutrophils which are activated and degranulate in situ,
releasing granulocytic enzymes and reactive oxygen
species (ROS) that could be responsible for alterations of
neighbouring alveolar cells. When clinical and radiological
signs suggesting VAP were encountered, we performed
71 bronchoalveolar lavages (BAL) in 58 ventilated
patients (42 men, 16 women with a mean age of
55.59±16.87 years; an APACHE II score at entry of
15.85±5.49; a mean intubation duration of 19.93±15.49
days; a mean ICU stay of 29.45±23.99 days; and 17
deaths). VAP was confirmed by quantitative bacteriological culture >104 CFU/ml of BAL and non-infectious
ARDS was confirmed by classical criteria including
PaO2/FiO2 ratio <200. In these BAL fluids, we measured
the concentrations of nitrated proteins (NTP) as an indicator of oxidative activity (ELISA technique), and the concentrations of active myeloperoxidase and elastase, two
markers of phagocyte degranulation (enzymatic measure).
The effects of BAL fluid on human alveolar cells (A549) in
culture were analyzed by measuring the release of 51Cr
pre-incorporated by the cells (cytotoxicity test) and correlated with NTP, myeloperoxidase and elastase values
(Pearson’s correlation and Mann-Whitney test with
P<0.05 being significant). Preliminary studies were also
performed to determine the mechanism of cytotoxicity of
BAL by analysing the capacity of these fluids to activate
the nuclear transcription factor NFκB.
VAP and/or ARDS was diagnosed in 38 patients
(VAP/ARDS group: 29 VAP and 9 ARDS), and the 20
others formed the control group. No significant difference
was observed between the two groups for clinical parameters. There was no significant difference in the mean
neutrophils × 106/ml
group (range)
control group
P value
NTP µg/ml
MPO U/ml
elastase U/ml
IC %
protein value of BAL between the two groups
(1.49±0.17 mg/ml in VAP/ARDS versus 1.38±0.34 in
controls). On the contrary, a significant difference was
found for the neutrophil count, NTP, myeloperoxidase and
elastase values, and for cytotoxicity (IC: cytotoxicity index
expressed in %).
The cytotoxicity of BAL was correlated with the values of
NTP (r=0.92; P<0.001) and MPO (r=0.88; P<0.001). In
19 BAL, NFκB was found to be activated, and this activation was correlated with the BAL value of IL8 (P<0.005).
From these data, we concluded that MPO and elastase
were released in the alveoli by activated neutrophils and
that NTP were formed in situ by the oxidant activity of
stimulated neutrophils (in situ production of peroxynitrite
and/or activity of MPO on nitrite or peroxynitrite). This
intra-alveolar oxidant activity led to the production of BAL
fluids which were cytotoxic on alveolar cells and which
enable the activation of the signal transduction pathway.
However, the exact consequences of this NFκB activation
and the particular compounds of BAL responsible for
cytotoxicity remain to determine.
Cardiopulmonary effects of bronchoalveolar lavage in critically ill patients with ventilatorassociated pneumonia
A Koroneos, I Kalomenidis, F Moraitou, P Polakis, S Papanikolaou and G Katsaris
ICU, Nikaia-Piraeus General State Hospital, Nikaia, Fanarioton Street, Greece
Purpose: The purpose of this study was to evaluate the
cardiopulmonary effects of bronchoalveolar lavage (BAL)
in critically ill patients with ventilator-associated pneumonia (vap).
Method: Ten sedated and paralyzed critically ill patients
suspected of having vap and requiring invasive hemodynamic monitoring were studied. A pulmonary artery
catheter with continuous SvO2 and CO display was
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
inserted to titrate administration of fluids and vasopressors. The FiO2 was adjusted to 1.0, 30 min before the performance of BAL for bacteriologic diagnosis and remained
at this level for 6 h after the procedure. The following parameters were determined (1) before, (2) immediately after,
(3) 2 h after and (4) 6 h after the BAL: compliance, Ppeak,
Pplat, Pi, Rsmax, Rsmin, ∆Rs, PO2, PCO2, pH, SaO2, SvO2,
AP(S/D/M), PAP(S/D/M), CVP, PCWP, CI, heart rate,
CvO2, C(a-v)O2, ∆PCO2, O2ER, Qs/Qt.
The paired t-test method was used to compare the results
at the four different time frames. Patients with necessary
manipulation in the infusion rate of fluids-vasopressors
and/or the parameters of mechanical ventilation during the
6 h period were excluded from data analysis.
Results: The results of eight patients were analyzed.
Compliance showed after a nonsignificant fall from (1) to
(2), an increase from (2) to (3) (P=0.038) and from (2) to
(4) (P=0.017). Pplat increased slightly from (1) to (2) and
decreased afterwards significantly from (2) to (4)
(P=0.017). PCO2 increased from (1) to (2) (P=0.015)
and remained high afterwards with parallel drop in pH
from (1) to (2) (P=0.006). The CI showed, after a fall from
(1) to (3) (P=0.031), a gradual improvement. PVRI
demonstrated, after a nonsignificant fall from (1) to (2), a
gradual increase after (2) which was significant at (4)
(P=0.026). No further significant differences were recognized in every other parameter studied.
Conclusions: BAL in our patient population was a safe
procedure. It did not affect oxygenation and pulmonary
shunt. Compliance and Pplat showed after a nonsignificant
initial deterioration gradual improvement. The delayed fall
of CI did not persist at the end of the study. The increase
in PVRI needs further evaluation.
The influence of empiric antimicrobial therapy on acquired pulmonary infection in patients with a
chest injury on ICU
I Chytra, E Kasal, R Pradl, J Voborníková and R Sviták
Department of Anaesthesiology and Intensive Care, University Hospital, Alej svobody 80, 30460 Plzen, Czech Republic
Introduction: Infections acquired in the intensive care unit
(ICU) commonly complicate the course of critical illness.
The administration of empiric antibiotics was identified as
an independent risk factor for the development of subsequent nosocomial infection [1]. The aim of our study was
to evaluate the influence of empiric antimicrobial therapy
on the acquired pulmonary infection in patients with a
chest injury on interdisciplinary ICU.
Table 1
Admission APACHE II
ISS (Injury Severity Score)
Length of stay (days)
4 (17.4%)
Mechanical ventilation (days)
Methods: We analyzed a prospectively-collected database of 64 patients with a chest injury admitted for at least
four days to the interdisciplinary ICU of university hospital
in 1999. Patients were retrospectively divided into a group
receiving, from admission to the ICU, antimicrobial therapy
(ATB group, n=41) and a group without antimicrobial
therapy during first 48 h after admission (No ATB group,
n=23). The age, APACHE II, ISS, TRISS, length of ICU
stay, duration of mechanical ventilation, rate of isolated
chest injury and associated diagnoses (multiple injury,
abdominal injury, head injury), ICU mortality, and rate of
acquired pulmonary infection were evaluated. χ2 test,
Mann-Whitney and unpaired t-test were used accordingly;
P<0.05 was considered statistically significant.
Number of acquired
pulmonary infection
Results: No differences in most of the parameters were
found between groups, except for the length of ICU stay,
17 (41.5%)
Data expressed as mean ±SD, *=P<0.05, NS=non-significant.
duration of mechanical ventilation and the incidence of
acquired pulmonary infection (Table 1).
Conclusions: Empiric antimicrobial therapy in patients
with the chest injury on ICU may increase the rate of pulmonary infections, prolong duration of mechanical ventilation and extend the time of ICU stay. There is a need for
controlled studies to define the role of antibiotics in
empiric therapy in trauma patients.
Kollef MH et al.: Chest, 1997, 112:666–675.
Non invasive mechanical ventilation (NINMV) in cardiac surgery
M Ferrante, L Bianciardi, E Conti, A Quacquarelli, V Praštalo and B Amari
Department of Cardiac Anaesthesia and Intensive Care, Poliambulanza Hospital, Via Bissolati 57, 25124, Brescia, Italy
Objective: To evaluate the efficacy of NINMV in cardiac
surgical patients. In order to verify if, beside already established advantages in critical care patients [1] (less intubation related complications, minor discomfort for patients
and less need for sedation), NINMV can be useful in managing the weaning process in the postoperative cardiac
surgical course.
Design: Retrospective survey. We used NINMV with two
indications: (1) patients with unexpected postoperative
respiratory failure requiring reintubation; (2) patients extubated early and with incomplete postoperative recovery of
cardiorespiratory stability.
Setting: Five bed intensive care unit at a cardiothoracic
surgery centre.
Patients: Twenty NINMV patients representing 4.1% of
the 484 patients operated on from October 1998 to September 1999 in our centre. Urgent surgery occurred in
12% of cases, emergency surgery in 4.75% of cases.
Overall in-hospital mortality was 14/484 (2.2%).
Interventions: We use a nasal or facial mask (Respironics Inc., Murrysville, Pennsylvania, USA) connected to a
Servo Ventilator 300 (Siemens-Elema, Sweden).
Patients were ventilated by Pressure Support (10–15
cmH20) with PEEP (5–8 cmH2O). Intermittent periods
of NINMV were alternated with periods of spontaneous
breathing accordingly to patients needs. Sedation was
obtained by Remifentanyl infusion in the range
0.03–0.07 γ kg-1min-1.
Measurements and main results: Twenty patients. were
treated by NINMV (two with indication 1 and 18 patients.
with indication 2). The two patients with indication 1 were
successfully weaned and discharged. None of the other
484 patients in this series needed reintubation.
Table 1 characterises patients with indication 2 and summarises results.
Only 3 patients had to be reintubated (n°3 for unexpected
haematological complication, n°8 for psychotic disturbances, n°18 failed full haemodynamic recovery).
Table 1
Haemodynamic support
NINMV days
Ascend. Aorta Replacement, AR,MR
MR, Pulm. Hypert., postop renal failure
CABG, postop acute leukaemia
CABG, postop redo CABG
CABG, postop PTCA
AR, MR, Pulm Hypert.
MR,CABG,cardiogenic shock
AR,CABG, postop psycomot. agitation
MR, Pulm Hypert
CABG, Pulm. Fibrosis
CABG, obesity
CABG,postop renal failure
AR,pulm Hypert
CABG, Hypertrophic. Myocard.
AR=aortic replacement, MR=mitral replacement, CABG =coronary artery bypass grafting, COPD=chronic obstructive pulmonary disease,
PTCA=percutaneous coronary angioplasty, W=weaned, R=reintubated.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Conclusions: NINMV patients with indication 2 were
weaned from respiratory support earlier than usual during
a complicated postoperative course. The good results of
our series suggest that NINMV indication 2 can be viewed
as a weaning model. The most severe forms of circulatory
impairment and the non-collaborative patient seem to be,
accordingly to literature, exclusion criteria. A wider study
would better precise inclusion and exclusion criteria in
cardiac surgical patients.
Slutsky AS: ACCP consensus conference. Mechanical ventilation.
Chest 1993, 104: 1833–1859.
An audit of unplanned extubations in a medical intensive care unit
A Johan
Tan Tock Seng Hospital, 11 Jalan Tan Tock Seng, Singapore 308433, Republic of Singapore
Aim: To study the factors associated with and the
outcome of unplanned extubations (UE) in an adult
medical intensive care unit (MICU).
Methods: Prospective observational study of all patients
admitted to an 8 bed university-affiliated MICU from
January to December 1998. All patients requiring orotracheal intubation and mechanical ventilation were enrolled
at time of the first UE. Information on demographics,
disease severity, sedation, ventilator parameters, nursing
factors and outcome was captured. Patients were followed up until in-hospital death or hospital discharge.
Results: 312 (57.5%) of the 543 patients admitted over the
12 months required intubation and mechanical ventilation.
There were 27 (8.7%) events of UE. The mean age of all UE
patients was 57.8±21.0 years with a mean APACHE II
score of 18.8±9.8. The mean number of days prior to UE,
mean MICU length of stay (LOS) and mean hospital LOS
were 3.2±3.3, 9.5±10.4 and 26.0±34.7 respectively There
were observed differences in MICU LOS (12.7±11.4
versus 3.2±2.6) and hospital LOS (31.4±41.2 versus
15.1±10.8) between the groups who did and did not
require reintubation. All 6 deaths came from the group that
was reintubated. These 6 patients were not weaning when
UE occurred. 17 (63%) of the UE events were deliberate
despite 16 (59%) and 20 (74%) of this cohort being on
chemical and physical restraints respectively. Nursing ratios
of 1 nurse to 2 or 3 patients, depending on shift, were
acceptable by our hospital standards and UE events
occurred during period of decreased nursing acuity during
night shift, change of shift and break time.
Conclusion: 8.7% UE, the majority being deliberate, was
observed in our MICU. Patients who extubated during ventilator weaning had a better outcome. Windows of
decreased nursing acuity may have contributed to the UE.
Manpower redistribution during these windows, the adoption of improved sedation titration protocols and protocols
to screen patients for readiness for extubation in our MICU
may further reduce the risk for deliberate unplanned extubations.
Timing of extubation after esophagectomy
FHY Yap, ACW Chan and GM Joynt
Department of Anaesthesia and Intensive Care, Prince of Wales Hospital, Shatin, HKSAR
Introduction: Prior to 1995, it had been our standard
practice to ventilate patients overnight after esophagectomy. Since then, however, this practice has changed. If
patients are warm, hemodynamically stable, and comfortable, we have elected to extubate them early.
Aim: To study the association between time to extubation
and outcome after esophagectomy.
Methods: We retrospectively reviewed the records of all
esophagectomies between 1990 and 1998. Early extubation was defined as extubation within 12 h. The primary
outcomes evaluated were respiratory complications, hospital mortality and length of ICU stay. Outcomes between
the two time periods 1990–94 (first period) and 1995–98
(second period) were compared. In the second time
period, the relationship between time to extubation and
outcomes was also evaluated.
Results: There were 78 esophagectomies in the first
period (1990–94) and 84 in the second period
(1995–98). In the second period, more patients received
epidural analgesia (64% vs 34%, P=0.001) and more
patients were extubated early (41% vs 6%, P<0.001)
compared to the first period. There were no significant differences in the rates of respiratory complications (27% vs
34%) and hospital mortality (6% vs 10%) when comparing the second with the first time period.
In the second time period (1995–98), 31 of the 84 patients
were extubated early (median time to early extubation was
6.8 h postop). More patients in the early extubation group
received epidural analgesia (90% vs 50%, P=0.001). There
were again no significant differences in the rates of respiratory complications (29% vs 30%) and hospital mortality
(6.7% vs 7%), but the late extubation group stayed longer in
the ICU (median stay 2 days vs 1 day, P=0.006).
Conclusions: Our experience suggests that early extubation after esophagectomy in stable patients is safe, and
may decrease the length of ICU stay with cost-saving
implications. We believe that thoracic epidural analgesia
may facilitate early extubation by enabling patients to
cough and cooperate with physiotherapy in the early postoperative period.
Evaluation of different CPAP systems in ICU population
D Chiumello, P Pelosi, N Bottino, E Calvi, P Taccone and L Gattinoni
Institute of Anaesthesia and ICU, Policlinico Hospital, IRCCS, University of Milan, Italy
The aim of this study was to evaluate the effects of different CPAP delivery systems in ICU patients. We tested
four systems: 1) a home-made CPAP low flow circuit with
an inspiratory reservoir and expiratory water valve (LW); 2)
the same system with an expiratory mechanical valve (Vital
Sign’s) (LM); 3) a high flow CPAP system (Jet Venturi)
with an expiratory mechanical valve (Vital Sign’s) (HM); 4)
the CPAP of the mechanical ventilator Servo 300
(Siemens); 5) the CPAP of Servo 300 plus three additional breaths per min delivered in pressure control ventilation (i.e. plateau airway pressure of 30 cmH2O)
(ServoPC). In 13 intubated patients (9 males, PEEP level
of 6±2 cmH2O and PaO2/FiO2 250±10) during the
weaning phase we measured the breathing pattern, work
of breathing, gas exchange and functional residual capacity. Data are expressed as mean ±sd.
We found no differences between the low flow, either
water or mechanical valve, and the high flow CPAP
systems. The Servo 300 reduced the work of breathing,
while the same CPAP system plus three additional breaths
caused a further reduction in work of breathing and
improvement in gas exchange and FRC.
TV ml
PTP cmH2O*sec/min
P.01 cmH2O
PaO2 mmHg
RR bpm
* P<0.05 vs others.
A comparison of the effects of conventional and balloon laryngoscopy on the sagittal
dimensions of the space available for the cord at the level of the occipitoatlantoaxial complex
SD Mentzelopoulos, MJ Tzoufi and EP Papageorgiou
Department of Anesthesiology of Egion General Hospital, Egion, Greece
Background: The purpose of this study was to compare
the sagittal surface areas (SSAs) of the space available
for the cord (SAC) at the occiput (OCC)-first cervical vertebra (C1) and C1-C2 levels determined at neutral head
position, and during conventional and balloon laryngoscopy [1,2].
Methods: Anesthesia was induced in eight ASA I, Mallampati I, elective surgery patients. Cross-table lateralview radiographs were taken at neutral head position
(NHP) and upon maximal laryngeal exposure (MLE) with
a #4 standard and a #4 modified blade carrying two 10
Foley catheters. MLE during balloon laryngoscopy was
achieved by right catheter-balloon inflation with 2 ml air
and blade-elevation [1,2]. The radiographs were
scanned in Photoshop 5.0, the X-ray magnification was
determined, and the OCC-C1 and C1-C2 SAC-SSAs
were measured with Autocad 2000. The OCC-C1-SAC
and C1-C2-SAC anatomic landmarks are shown in
Fig. 1. Data were analyzed with Kolmogorov-Smirnov test
and ANOVA.
Results: At the OCC-C1 level, the SAC-SSA was significantly reduced during conventional laryngoscopy (MLEvalue =0.386±0.104 in2, NHP-value =0.501±0.11 in2,
P=0.0095). During balloon laryngoscopy, there was no significant SAC-SSA reduction (MLE-value =0.423±0.128 in2,
P=0.07 vs NHP-value), while the SAC-SSA was significantly greater than the SAC-SSA determined during conventional laryngoscopy (P=0.044). At the C1-C2 level, the
SAC-SSA-changes were nonsignificant.
Conclusions: The OCC-C1-SAC is significantly less
affected by balloon laryngoscopy.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Figure 1
Mentzelopoulos SD et al: Anesth Analg 1999, 88:1425–1426.
Mentzelopoulos SD et al: Crit Care 1999, 3(suppl 1):3–4.
ABC=inferior margin of occiput. DE=C1-spinous process base. E=most caudal
point of C1-spinous process base. F=radiographic intersection of C1 by
odontoid-posterior cortex. G=most cephalad point of odontoid-posterior cortex.
HI=C2-spinous process base. I=most caudal point of C1-spinous process base.
FJ=posterior cortex of C2-vertebral body
Use of Combitube airway to protect the airway from methylene blue dye
MH Mercer
Department of Anaesthesia, Frenchay Hospital, Bristol BS16 1LE, United Kingdom
Introduction: The Combitube airway (Kendall UK Ltd.,
Basingstoke, England) is included in the European Resuscitation guidelines for the management of the emergency
airway [1]. In the trauma patient tracheal soiling is usually
from the upper airway [2]. Although the proximal, large cuff
of the Combitube may protect the airway, this has not
been formally assessed.
manually cleaned with absorbent gauze. The Combitube
was removed and the airway examined laryngoscopically.
Method: After local ethics committee approval, 10 ASA 1
and 2, starved patients undergoing routine elective
general anaesthesia involving non-depolarising neuromuscular blockade and mechanical ventilation of the lungs
gave informed consent. Only patients with grade 1 direct
laryngoscopic views were included to ensure adequate
views of the vocal cords. The small adult sized Combitube
was placed into the oesophagus, all patients being of the
appropriate height, and the cuffs inflated. 10 ml of 0.1%
methylene blue dye was instilled into the mouth. Full monitoring was used, and adequate ventilation ensured
throughout. At the end of surgery all dye was suctioned
away and the oral cavity and proximal Combitube cuff
Conclusion: The Combitube therefore protects the majority of patients airways from aspiration of dye from the oral
cavity. In the trauma setting blood and oral debris may
therefore be prevented from entering the trachea. The one
failure was probably a result of patient movement by surgeons as this same patient developed a ventilatory leak
mid-operation. However, patient movement would be
expected in the trauma population.
Results: One patient had an initial airway leak that settled
with minor Combitube realignment prior to dye instillation.
No laryngeal dye was detected in 9 patients. One patient
had significant blue staining of the vocal cords.
European Resuscitation Council: Resuscitation 1996, 31:201–230.
Lockey DJ, Coats T, Parr MJA: Aspiration in severe trauma: a
prospective study. Anaesthesia 1999, 54:1097–1098.
Translaryngeal tracheostomy: prospective experience in two Canadian tertiary intensive care units
MD Sharpe*, L Parnes†, C Harris* and J Drover‡
*Departments of Anaesthesia and †Otolaryngology, London Health Sciences Centre, 339 Windermere Road, London, Ontario
Canada N6A 5A5; ‡Program in Critical Care Medicine, University of Western Ontario, and Department of Surgery, Kingston
General Hospital, Queen’s University, Kingston, Ontario, Canada
Introduction: Bedside percutaneous tracheostomy has
supplanted open surgical tracheostomy as the procedure
of choice in many Intensive Care Units for patients who
require prolonged mechanical ventilatory support. The
new translaryngeal tracheostomy (TLT) method described
by Fantoni et al. [1] has inherent advantages over other
percutaneous techniques. We present our prospective
review of 111 cases using this new technique.
Methods: During a 26 month period, a TLT was performed on 111 patients who required an elective tracheostomy in our two ICUs. Under general anesthesia, the
TLT was performed at the bedside using the Translaryngeal Tracheostomy kit (Mallinckrodt, Italy). This technique
is performed under direct visualization with either a flexible
bronchoscope or a rigid 30° endoscope. The tracheostomy tube with its dilator is pulled from within the
trachea using a guide wire that is initially placed during
direct visualization. After rotating the cut tube into place,
bronchoscopic visualization confirms its correct position
within the trachea.
Results: All 111 tracheostomies were performed successfully. Mean duration of the procedure was 28±11 min
with a mean oxygen saturation of 98±3%; the lowest saturation occurring was 81%. Desaturations occurred in 20
patients; however, they were short lived and caused no
harm. Mean blood pressure was 77±26 mmHg. Blood
loss was minimal (<5 mls) in all cases. Mean platelet
count was 277×106±135; range 39–818). Mean INR
was 1.2±0.3; range 0.9–2.8. Mean PTT was 48±58 s;
range 22–554. All patients tolerated the procedure well.
Postoperative wound infections did not occur. One patient
was noted to have a pneumothorax the day following the
procedure while on positive pressure ventilation.
Conclusion: This technique induces minimal trauma due
to its ‘dilation’ of tissue as it passes through the anterior
wall of the trachea. This new dilatational technique is easy,
safe and reliable. Bleeding is minimal, even in patients with
coagulopathies due to the tight approximation of the tube
to the tissue. Ventilation and protection of the airway is
maintained throughout the procedure. It has become our
procedure of choice for patients requiring an elective tracheostomy.
Fantoni et al.: Int Care Med 1997, 23:386–392.
The use of percutaneous tracheostomy in patients with severe acute or chronic liver disease
N Parnell, W Bernal and J Wendon
The Institute of Liver studies, King’s College Hospital, Denmark Hill, London, UK
Introduction: Percutaneous tracheostomy (PT) is recognised as a safe and effective method for prolonged ventilatory support, avoiding the complications of lengthy
endotracheal intubation. Bleeding has been reported as
the most common perioperative complication in most
series [1]. Patients with severe liver disease frequently
have a coagulopathy and thrombocytopaenia. We wished
to assess the safety of PT in this group of patients, who
are at high risk of haemorrhagic complications.
Methods: A retrospective study of PT’s performed in
patients admitted to a dedicated liver intensive care unit in
a supraregional liver centre over two years. All procedures
followed the Ciaglia technique [2] and were performed by,
or under the supervision of, experienced operators. All
patients had liver disease, which included decompensated
chronic disease, hyperacute and acute liver failure and
liver transplantation with severe postoperative complications.
Patients considered to be at particularly high risk for
bleeding, based on platelet count (Plts) and international
normalised ratio (INR) received transfusions of platelets
and/or fresh frozen plasma prior to the procedure.
Results: The median age was 46.5 years with 20 males
and 14 females. Five of the thirty-four patients suffered
complications, none of which were fatal, and included
minor bleeding only (requiring no treatment). The MannWhitney U test was used to compare differences in INR
and Plts between those with and without complications.
The median pre-correction INR for the group without complications was 1.1 (0.82–1.79) and for those with complications was 1.34 (0.92–1.76) and this did not reach
statistical significance. There was a trend towards lower
Plts in those with a complication (median 44; range
21–84) compared to those without (median 76; range
3–510) but again this was not statistically significant
Conclusions: Although this is a small study, based on the
data presented, PT appears to be a safe procedure in this
group of patients, at high risk of bleeding, when standard
precautions are followed.
Petros S: Percutaneous tracheostomy. Crit care 1999, 3:R5–R10.
Ciaglia P, Firsching R, Syniec C: Elective percutaneous dilatational
tracheostomy. Chest 1985, 87:715–719.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
In vitro study of a new vaporization humidifying device: DAR HC 2000®
N Bottino, L Chiaravalli, M Panigada, F Carrieri, E Allegritti, P Pelosi* and L Gattinoni
1st. Anestesia e Rianimazione, Osp. Maggiore Policlinico-IRCCS, via F. Sforza, 35, 20122, Milan, Italy; *Dip. di Scienze Cliniche
e Biologiche, Univ. dell’Insubria, Varese, Italy
In long-term mechanical ventilation, inspired gas should
reach the endotracheal tube with a temperature of
30–32°C, and an absolute humidity (HA) of approximately
28–30 mg H2O/l.
We tested a humidifier (DAR HC 2000®) composed of: 1)
a vaporising cartridge, containing a hydrophobic membrane that separates the gas flow from the heated water;
2) a heating wire that runs all along the inspiratory line,
maintaining the heat and moisture conditions of the
inspired gas from the cartridge to the patient; 3) a double
temperature control that permits the setting of the desired
heating level both at the cartridge (T°c) and at the patient’s mouth (T°p).
At first, we studied the system’s ability to condition
inspired gas at different min ventilation (VE=6, 10, 14
l/min), tidal volume (VT=250, 500, 750 ml), and inspiratory
flow (Vi=0.5, 1 l/sec) with different settings of T°c and T°p
(32–32, 34–34, 36–36, 34–32, 36–32, 36–34, 38–34,
39–34°C, respectively). We found that increasing VE from
6 to 10 to 14 l/min led to a progressive decrease of HA
from 33.0±4.4 to 27.5±4.4 to 23.4±3.6 mg/l (P<0.01),
while VT and Vi had no effects on humidification.
Secondly, we tested the humidifier at different VE (from 4
to 16 l/min, 2 l/min per step) with different T°p (32, 34,
36°C; T°c initially set at the same level as T°p, and then
increased at 1°C at a time until condensate appeared in
the thermoregulated tube). Adequate humidification levels
(HA=30 mg/l) can be obtained for each VE and T°p by
increasing the temperature set at the cartridge enough.
The third part of the study was to verify the decrease of
microbiological contamination by simulating a contaminated patient (using a reservoir containing Micrococcus
luteus) and analysing the inspiratory line with the heating
wire switched first OFF, then ON: keeping the thermoregulated tube OFF, as supposed, a bacterial growth was
seen in the inspiratory line, while switching it ON, no M.
luteus colonies were observable.
In conclusion, DAR HC 2000® humidifier allows us to
obtain adequate levels of humidity of inspired gas for each
ventilatory set, just by increasing the cartridge temperature. Besides, the hydrophobic membrane and the heating
wire avoid, respectively, the gas–water contact and the
condensate formation in the inspiratory line resulting in a
reduction of microbiological contamination of the circuit.
Non-invasive two-point estimation of arterial PCO2 and alveolar deadspace
E Michel* and JA Peper*†
*Emma Children’s Hospital, PICU, and †Department of Anaesthesiology, Academic Medical Centre, University of Amsterdam,
Meibergdreef 9, NL-1105 AZ Amsterdam, The Netherlands
Introduction: Often end-tidal PCO2 (etPCO2) gives a good
estimate of arterial PCO2 (PaCO2). However, there are conditions, namely diseases with pulmonary ventilation/perfusion mismatch, where PaCO2 is severely underestimated.
Objective: We aimed to develop a method for the accurate non-invasive estimation of PaCO2 from etPCO2.
Method: The method is based on the model assumption
that etPCO2 is a mixture of PCO2 from both the ventilated
and perfused alveoli, and the ventilated but non-perfused
ones (‘alveolar deadspace’). Knowing the etPCO2 under
conditions of inspiratory gases with different CO2 content
(PCO2_a; PCO2_b), it is possible to calculate the PCO2
of the ventilated and perfused alveoli (PCO2alv). Knowledge of PCO2alv allows to calculate the fraction of alveolar deadspace ventilation. As long as para-alveolar (i.e.
cardiac) shunting is less than 20%, PCO2alv accurately
reflects PaCO2.
PCO2alv =
(etPCO2_a • PCO2_b)–(etPCO2_b • PCO2_a)
PCO2_b–etPCO2_b–PCO2_a + etPCO2_a
The model was applied on ventilated adult patients while
inpiring different levels of PCO2. Estimates of PaCO2 and
alveolar deadspace were compared to the results of invasive determination.
Results: Preliminary results indicate a good correlation
between invasively and non-invasively determined PaCO2
and alveolar deadspace (Table 1).
Conclusions: Using gas void of CO2 as one of the two
inspiratory gases further simplifies the formula. Our preliminary findings show that a difference in the levels of
inspired PCO2 of at least 3 kPa gives an estimate of
PaCO2 with an accuracy of 10% or better. We provide a
non-invasive method for the accurate estimation of PaCO2
Table 1
PaCO2 (kPa)
difference (%)
and alveolar deadspace ventilation. We suggest its implementation in ventilators for the close monitoring of pulmonary treatment response.
Hemodynamic effects of the inspiratory flow rate in patients with septic shock
A Koroneos, J Kalomenidis, F Moraitou, P Polakis, G Leptidis, V Kekeris and G Katsaris
ICU, Nikaia-Piraeus General State Hospital, Fanarioton Street, Nikaia, Greece
Purpose: The purpose of the study was to assess the
effects of changes in inspiratory flow (IF) on hemodynamic
parameters of mechanically-ventilated patients with septic
Method: Eight patients with septic shock (pneumonia
n=2 and abdominal sepsis n=6) were included in the
study. A pulmonary artery catheter with continuous SvO 2
and CO display had been inserted to titrate fluid and
vasopressor therapy. The patients were on volume
control ventilation with a square flow waveform. Three
different levels of IF were applied, each for a period of
thirty min (IF1=40l/min, IF2=50l/min, IF3=60l/min). At the
end of each period the following parameters were
LVSWI, RVSWI, SaO2, SvO2, DO2. The paired t-test
method was used to compare the results at the three different levels of IF.
Results: Two patients were withdrawn from the study,
due to necessary manipulation in the infusion rate of fluids
and/or vasopressors. In six patients, a significant increase
in LVSWI from IF1 to IF3 (5.22±1.38 g.m/m2 [mean ±SE],
P=0.013) was observed. No further significant differences
were recognized.
Conclusion: The change in the IF from 40 to 60l/min did
not have any negative effect in the hemodynamic profile of
patients with septic shock. On the contrary, it led to an
increase of the LVSWI.
Automatic tube compensation combined with pressure support ventilation – improved work of
breathing pattern and less work
P Singer, J Cohen, A Abraham and E Grozovsky
General Intensive Care Unit, Rabin Medical Center, Beilinson Campus, Petah Tikwa 49100, Israel
Introduction: Liberation from mechanical ventilation may
be difficult due to resistive properties of the endotracheal
tube (ETT), which increases work of breathing. PSV alone
may not overcome this problem as the increased resistance largely limits expiratory gas flow–inspiration and
expiration. The benefits of combining ATC with PSV have
not been assessed.
Method: We included 34 patients ventilated with three
levels of PSV (<15, 15–20, 20–30 CMH2O). ATC with
100% conpensation was then added for 30 min, stopped
for 30 min and re-instituted for 30 min. At the end of each
period we measured respiratory rate (RR), tidal volume
(TV), minute ventilation (MV), PEEPi, HR, VO2, VCO2 and
energy expenditure (EE).
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Results are given as median and ±SD
With ATC
Without ATC
With ATC
RR/ mnt
TV (ml)
MV (L/mnt)
EE (Kcal/D)
No significant differences were found for the parameters
studied at other levels of PSV.
Conclusions: The addition of ATC to low levels of PSV
(<15CM of water) results in improved breathing patterns
with metabolic benefits to the patient.
Proportional pressure support in acute lung injury: an observational study
C Capra and S Manera
Intensive Care Unit, Saronno, P.zza Borella 1, Saronno (VA), Italy
Proportional Pressure Support (PPS™) is a partial ventilatory support that overcomes patient respiratory system
elastance (Ers) and resistance (Rrs) by applying pressure
‘proportional’ to volume (volume assist, VA) and flow (flow
assist, FA) respectively. The aim of this study was to evaluate its clinical use at the bedside in critically ill intubated
patients affected by acute lung injury (ALI).
Methods: Thirty previously healthy patients affected by
post-traumatic (emopneumothorax and lung contusion,
N=21), infective (pneumonia, N=8) and inflammatory (pancreatitis, N=1) ALI (American-European Consensus Conference on ARDS) were studied. After several days (range
2–7, mean 4) of assist/controlled (A/C) ventilation
(BIPAP™), patients were switched to PPS once they were
stable. Ers and Rrs were determined during a short period
of volume control ventilation (square flow wave, tidal
volume 10 ml/kg) using the inspiratory hold technique. On
the ventilator (EVITA 4, Draeger) FA and VA were set to
80% value of patient Ers and Rrs respectively. The Automatic Tube Compensation (ATC™) was also used at 100%
compensation. O2 inspiratory fraction (FiO2) and positive
end-respiratory pressure (PEEP) were the same as in A/C
mode. Progressive reduction of FA, VA, PEEP and FiO2
was used as the weaning technique, during the healing
phase of the disease. The following parameters were
ascertained daily: arterial blood gas analysis, respiratory
rate (RR), tidal volume (Vt), minute ventilation (Ve), patients’
comfort, number of respiratory distress that needed
changes in ventilatory mode and failure to wean. The FA
and VA values at which the patients were able to sustain
spontaneous breathing and extubation were registered.
Results: PaO2/FiO2 ratio remained unchanged or slightly
better compared to A/C mode, and progressively returned
to normal during recovery. PaCO2 ranged between normal
values and pH ranged between 7.45 and 7.40 for each
patient during the study period. The respiratory pattern
changed greatly in terms of RR and Vt from one patient to
another and in the same patient during the days, while Ve
remained more stable. All patients, except two, were
always comfortable and they were successfully weaned
and extubated at mean values of FA=7±2 cmH2O/l/sec
and VA=8±3 cmH2O/l. The two patients that experienced
respiratory distress were affected by pneumonia and
needed a change of ventilatory mode.
Conclusion: PPS™ with ATC™ may be used to assist the
ventilatory needs of patients with post-traumatic, infective
or inflammatory ALI. The clinician must not be frightened
by an unusual respiratory pattern in the presence of a
comfortable patient, because with PPS™ patients are able
to breathe as they like to maintain a normal pH, most likely
because the neuroventilatory coupling is improved by this
mode, as described by Younes [1]. Extensive studies are
needed to evaluate the role of PPS™ with ATC™ among the
ventilatory modes now available to support the patient
with ALI.
Younes M: Proportional Assist Ventilation. In Principles and Practice
of Mechanical Ventilation. Tobin MJ: Mc Graw-Hill, Inc.;
Selected biochemical values and organ dysfunction assessment in prediction of difficult to wean
V Cerny, P Zivny*, P Dostal and R Parizkova
Department of Anesthesiology and Intensive Care, *Department of Biochemistry, Charles University, Faculty of Medicine, 50005
Hradec Králové, Czech Republic
Introduction: Metabolic status and severity of illness play
an important role in developing ventilatory dependency
during long term ventilatory support in critically ill patients.
The aim of the study was to evaluate selected clinical and
biochemical values in order to identify ‘difficult to wean’
and ‘easy to wean’ patients during ventilatory support.
Methods: After institutional approval, 40 critically ill
patients were prospectively studied during ventilatory
support and weaning. All patients were weaned according
to standard weaning protocol. Blood samples were drawn
daily and collected until analysis. APACHE II score, organ
failure score (Goris), sepis organ failure assessment score
(SOFA), ventilatory days, and ‘weaning’ days were
recorded. After successful weaning, patients were divided
into two groups according to the length of weaning (W):
group S (W ≤3 days, n=15), group L (W >3 days, n=22).
Serum levels of albumin, prealbumin, cholinesterase, CRP,
magnesium, phosphorus and urine levels of myoglobin
were measured during ventilatory support (including
weaning trials). SigmaStat Statistical Software was used
for analysis, r=corr. coeff., *P<0.05.
Results: There were no significant differences in any biochemical values between groups during ventilatory
support. In group L there were positive significant correlations between SOFA score and length of ventilatory
support and weaning (r=0.473, P=0.035) and between
Goris score and length of ventilatory support (r=0.626,
Conclusion: Assesment of severity of organ dysfunction,
but no special biochemical values, seems to be from
‘cost/effectiveness point-of-view’ the best way to identify
patients with need for long term ventilatory support and
who are difficult to wean.
Respiratory mechanics studied by multiple regression and the end-inspiratory pause technique
during mechanical ventilation
MP Vassiliou, Ch Psarakis, A Amygdalou, K Kalopissi, M Moukas and PK Behrakis
Pulmonary Function Laboratory, ‘St. Savas’ Hospital and Laboratory of Experimental Physiology, Medical School, University of
Athens, Athens, Greece
Aim of the study: In the present study, a comparative
evaluation of the multiple linear regression analysis
(MLRA) and the end-inspiratory pause technique (EIP) is
attempted for the investigation of respiratory system (RS)
mechanics during mechanical ventilation.
Methods: Airway pressure (Pao) and flow (V′) data were
digitally obtained from 25 ICU patients, who were
mechanically ventilated (CMV mode) and under sedation
and muscle relaxation. Volume (V) was calculated by
numerical integration of V′. Data were analysed on a cycle
per cycle basis with the aid of MLRA according to:
Pao=PE + Ers.V + Rrs.V′, where Ers represents the RS
Elastance, Rrs the RS resistance and PE the end-expiratory pressure. Ers and Rrs were also calculated with the
aid of the EIP, while end-expiratory pressure was measured on the actual Pao signal. Ers(EIP), Rrs(EIP) and
PE(EIP) were used to reconstruct the pressure, according
to EIP. Predicted Pao according to MLRA and EIP were
correlated to the measured Pao. The error of correlations
measured as the root mean square difference (RMSD)
was used for the comparative evaluation of the two techniques. Student paired t-test was used for the comparison
of the calculated mechanical coefficients (P=0.05).
Results: Results are presented in the table (*denotes statistically significant difference at P=0.05).
Ers (hPa/L)
Rrs (hPa/L/s)
PE (hPa)
RMSD (hPa)*
Conclusions: Both MLRA and EIP offer an almost equal
and linear approach to respiratory mechanics during
mechanical ventilation. MLRA gives more accurate results
according to the better correlation (lower RMSD) of MLRA
pressure than EIP pressure to the actually measured Pao.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Open lung strategies in patients with intracranial lesions and acute lung injury
S Wolf, L Schürer, HA Trost and Ch B Lumenta
Department of Neurosurgery, Academic Hospital Munich-Bogenhausen, Englschalkinger Strasse 77, 81925 Munich, Germany
Objective: For patients with acute lung injury, a recent
randomized controlled trial showed the superiority of a
ventilator strategy with low tidal volumes and elevated
PEEP levels after performing recruiting maneuvers
(together termed the ‘open lung approach’) compared to a
conventional approach with tidal volumes of 12 ml/kg
body weight [1]. However, patients with intracranial
lesions were excluded from this and other ARDS studies
due to the potential risk of exacerbation of intracranial
cantly from 130±104 before to 344±63 measured 24 h
after first recruiting (P<0.001). Mean PEEP after the first
recruitment maneuver was 14.9±2.7 mmHg. ICP values
during the first 24 h rose slightly (not significant) in seven
of the eight patients, requiring standard treatment with
immediate response in two of them. In one patient, the ICP
decreased from 33 mmHg to 25 mmHg after the first
recruitment maneuver. Relevant change of ptiO2 values
was not noticed, as the FiO2 was consequently lowered,
according to provide an arterial paO2 around 100 mmHg,
and with respect of a ptiO2 above 20 mmHg.
Methods: We present the clinical course of eight patients
with known intracranial pathology and concomitant acute
lung injury, which was treated according to the open lung
concept. Primary diagnosis was either traumatic brain
injury (n=3), subarachnoid hemorrhage (n=2), cerebellar
hemorrhage (n=1), cerebellar infarction (n=1) or frontal
meningeoma (n=1). ICP monitoring was present in all
patients, brain tissue oxygen (ptiO2) monitoring in two of
them. Data was collected with multimodal monitoring
devices with a sampling rate of 1 per min.
Conclusion: Although this study was done as a feasibility
analysis and lacks a control group, our results encourage
us to propose open lung strategies, which had been
restricted up to now, for neurosurgical patients with acute
lung injury. The potential risk of intracranial deterioration
due to critical ICP values seems far lower than the achievable benefit and decrease of mortality due to improved
Results: Ventilation according to the open lung approach
improved the mean oxygenation index (paO2/FiO2) signifi-
Amato MB et al.: N Engl J Med 1998, 338:347–354.
Relationship between superimposed pressure and pleural pressure gradient in an experimental
model of ARDS
N Bottino, G Eccher, P Pelosi†, A McKibben*, A Adams*, M Goldner*, JJ Marini* and L Gattinoni
1st. Anestesia e Rianimazione, Osp. Maggiore Policlinico-IRCCS, via F. Sforza, 35, 20122, Milano, Italy; *University of
Minnesota, St. Paul, MN, USA; †Dip. di Scienze Cliniche e Biologiche, Univ. dell’Insubria, Varese, Italy
Introduction: During Acute Respiratory Distress Syndrome (ARDS), the lung density increases along a gravity
gradient, causing a hydrostatic pressure gradient, even if
the edema is homogeneously distributed. In an oleic acid
induced ARDS experimental model, we studied by CT
scan the influence of lung hydrostatic pressure on regional
pleural pressure.
Methods: ARDS was induced in eight sedated, paralyzed
and mechanically-ventilated dogs, by an oleic acid dose
(0.075 mg/kg) injected in the pulmonary circulation.
Before and after ARDS induction, a CT scan basal section
was taken, and pleural pressure was directly measured
throughout two flat pressure sensors (wafers) positioned
into the pleural space, in the most upper part and in the
most dependent part of the lung. We calculated the
superimposed pressure (SP) by mean density (ρ) and
height (h) of the CT scan section, according to the following formula: SP=ρ×h [1]. SP represents an estimation of
hydrostatic pressure that weighs at the lowest level of the
CT section.
Results: After oleic acid, SP is significantly increased
(from 5.0±0.3 to 10.5±1.9 cmH2O, P<0.01), and pleural
pressure gradient (calculated as the difference between
the pressure measured by the lower and the upper wafer)
has a similar trend (from 6.5±2.0 to 10.0±4.2 cmH2O,
P<0.01). The SP sterno-vertebral gradient (∆SP) is significantly correlated with the pleural pressure gradient (∆Ppl)
(P=0.01; r=0.76), according to the following equation:
∆Ppl [cmH2O]=(1.33±1.21) + (0.74±0.23)* ∆SP
Conclusion: Pleural pressure seems to change as a function of superimposed pressure, and both these pressures
change as a function of the sterno-vertebral level.
However, the slope between ∆Ppl and ∆SP is lower than
1 because also other variables (thorax shape, regional
compliance, etc.) probably influence the SP effect on the
pleural pressure changes.
Pelosi P et al.: Am J Respir Crit Care Med 1994, 149:8–13.
Effects of a lung recruitment maneuver keeping PEEP before and after L-Pflex on gas exchange
in child ARDS patients
AMAS Gaudêncio, EJ Troster, LF Faria, JBS Borges, FAC Vaz, CRR Carvalho, MBP Amato and CSV Barbas
Respiratory and Pediatric ICU, Hospital das Clínicas, University of São Paulo, S.P. Brazil, Av. Dr. Eneas de Carvalho Aguiar
155- 8 andar-bloco 3, CEP-05403-900, São Paulo, SP, Brazil
level associated with a recruitment maneuver (PCV=15
cmH20, PEEP between 30 and 40 cmH20, according to an
immediately previous thoracic CT-scan for each child that
showed less than 5% of the lung area between -100 and
+100 HU). Then the children were ventilated for 30 more
min with PEEP 2 cmH20 above the L-Pflex without a recruitment maneuver, and finally 30 min with the same PEEP level
associated with the recruitment maneuver.
PEEP titrated by the L-Pflex of static PxV curve of respiratory system associated with recruitment maneuvers
improved oxygenation and mortality in ARDS patients. In
order to study the effects of PEEP 2 cmH20 below and
above the L-Pflex without and with a recruitment maneuver
on gas exchange and hemodynamic parameters, we analyzed foyr children (7 to 14 months) with ARDS criteria:
acute bilateral RX infiltrates, PaO2/FIO2 <200, echocardiography without left cardiac dysfunction, (<5 days of installation). All the children were sedated and paralyzed. After a
ventilatory control period (TV=10 ml/kg, PEEP=5 cmH20), a
constant flow (0.2 l/min) PxV curve was done for each child
and the L-Pflex was calculated. Then the children were ventilated in VCV (6 ml/kg, RR=20/min- Servo-Siemens-300)
for 30 min with PEEP 2 cmH20 below the L-Pflex without
recruitment maneuver, then 30 min with the same PEEP
Conclusion: A recruitment maneuver according to the
findings of the thoracic CT-scan improved oxygenation
mainly after PEEP set 2 cmH20 above the L-Pflex of the
PxV curve, improved CO2 exchange without hemodynamic
impairment, in child ARDS patients.
30 min
PEEP 2 cmH20
<L-Pflex without R.M.
30 min
PEEP 2 cmH20
<L-Pflex with R.M.
30 min
PEEP 2 cmH20
>L-Pflex without R.M.
30 min
PEEP 2cmH20
>L-Pflex with R.M.
PaCO2 (mmHg)
Patients n=4
MAP (mmHg)
Results: See Table.
Effects of artificial changes in chest wall compliance on respiratory mechanics and gas exchange
in patients with acute lung injury (ALI)
N Bottino, M Panigada, D Chiumello, P Pelosi* and L Gattinoni
1st. Anestesia e Rianimazione, Osp. Maggiore Policlinico–IRCCS, via F. Sforza, 35, 20122, Milan, Italy; *Dip. di Scienze
Cliniche e Biologiche, Univ. dell’Insubria, Varese, Italy
In ALI patients, the prone position induces a decrease in
chest wall compliance significantly correlated with oxygenation improvement. We studied the effects on respiratory mechanics and oxygenation of artificial change of
chest wall compliance (obtained positioning a 10 kg ‘sand
bag’ (SB) upon the upper thorax) in 11 supine mechanically ventilated ALI patients (volume control mode,
PEEP=14±3 cmH2O, i.e., 2–5 cmH2O above the PV
curve inflection point (‘Pflex PEEP’), VT=556±137 ml,
RR=17±4 br/min, VE=9.1±1.3 L/min, FiO2=80±22%).
We measured gas exchange, hemodynamics, total respiratory system compliance (CRS) partitioned into its lung
(CL) and chest wall (CCW) components (end-inspiratory
occlusion and esophageal balloon technique), end-expi-
ratory lung volume (EELV; helium dilution), abdominal
ventilation distribution (Vabd/Vtot; Respitrace®, SensorMedics) and intraabdominal pressure (IAP; intrabladder technique).
SB positioning led to an EELV decrease (from 1.22±0.75
to 0.99±0.66 l, P<0.05) and, interestingly, to an increase
of CCW and CRS (from 43±23 to 58±35 and from 34±15
to 40±18 ml/cmH2O, respectively, P<0.05). Mean PaO2,
CL and Vabd/Vtot did not change, even if patients who
improved oxygenation were the same who reduced their
CCW (r=0.78, P<0.01) and increased CL (r=0.64,
P<0.05) and Vabd/Vtot (r=0.68, P<0.05). Moreover,
patients with lower baseline IAP had a greater CCW reduction (r=0.60, P<0.05).
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
We tested SB positioning also at an higher PEEP level (5
cmH2O above ‘Pflex PEEP’): increasing PEEP induced an
increase of EELV (from 1.22±0.75 to 1.37±0.82 l,
P<0.05), Vabd/Vtot (from 41±19 to 50±20%, P<0.05) and
PaO2 (from 100±24 to 121±38 mmHg, P<0.05) but, at
this higher PEEP level, after SB application, oxygenation
decreased in all patients (from 121±38 to 105±30
mmHg, P<0.05).
Conclusion: Effects of artificial change of chest wall compliance depend on PEEP level and the patient’s pathophysiological characteristics: in patients with low IAP
values, SB positioning could decrease the regional compliance of upper thorax and favor a caudal displacement of
the diaphragm, increasing the ventilation distribution to the
abdominal compartment. This can result in an oxygenation
improvement, probably due to a decrease of ‘true’ shunt
and/or a more homogeneous ventilation to perfusion ratio.
The lower inflection point of the inspiratory pressure-volume curve overestimates optimal PEEP
in surfactant-treated immature lambs
J Ingimarsson*, LJ Björklund†, A Larsson‡ and Olof Werner*
Departments of *Anaesthesiology and †Pediatrics, University Hospital, Lund, Sweden; and ‡Department of Anaesthesiology,
Gentofte University Hospita, Hellerup, Denmark
Introduction: Static inspiratory pressure-volume (P-V)
curves are typically S-shaped with a lower (LIP) and an
upper inflection point (UIP), probably because of successive recruitment of lung units. It is often said that PEEP
should be set above LIP to avoid derecruitment during
expiration. However, as shown in a rabbit model [1], the
effect of PEEP on lung volume is determined by the expiratory part of the P-V loop, which has a markedly different
Methods: Eleven lambs with gestational age 119–141
days (term =145days) delivered by cesarean section were
given surfactant and subsequently mechanically ventilated.
Using an interrupter technique, static P-V loops were
recorded between an end-expiratory pressure of a) 0
cmH2O (ZEEP) or b) 5cmH2O and an inspiratory pressure
of 35 cmH2O.
Results: Loops obtained during insufflation from ZEEP
had a large hysteresis; LIP was 16–21 cmH2O, while UIP
was above 30 cmH2O, if seen at all (Fig. 1 loop A) and the
expiratory limbs showed a high deflation stability down to
a pressure of 6–11 cmH2O with a steep segment at lower
pressures. Loops from 5 cmH2O had a nearly linear inspiratory limb, and much less hysteresis (Fig. 1 loop B).
Conclusion: In preterm lambs with fully-recruited lung
volume, derecruitment can be avoided by a much lower
Figure 1
Loops at 15 min of age in a 132 day GA lamb. A: loop from ZEEP; B:
loop from positive end-expiratory pressure (5 cmH2O).
PEEP than indicated by LIP. During mechanical ventilation
of immature subjects, the upper inflection point of the expiratory limb (UIPexp) is probably a better indicator of optimal
PEEP than LIP of the inspiratory curve.
Rimensberger et al.: Crit Care Med 1999, 27:1946–1952.
Effect of different tidal volumes and PEEP levels on gas exchange and FRC
D Chiumello*, P Pelosi*, E Calvi*, P Taccone*, A Aliverti, R Dellaca’, S Lanzi and L Gattinoni*
Dipartimento di Bioingegneria, Politecnico di Milano, Italy; *Istituto di Anestesia e Rianimazione, Ospedale Policlinico, IRCCS,
Universita’ di Milano, Italy
The effect of PEEP on improving oxygenation and FRC is
well known [1], not so the effects of PEEP and tidal
volumes on gas exchange and FRC.
The aim of this study was to evaluate the effect of three
different levels of tidal volume (300, 600, 900 ml) and
PEEP (5, 10, 15 cmH2O) on gas exchange and FRC. We
studied nine intubated sedated and paralyzed patients (5
males) in controlled volume ventilation with respiratory
failure in the early phase of the disease. Baseline clinical
characteristics: age 60±19 years, weight 72±10 kg,
PaO2/FiO2 240±101, PEEP level of 10 cmH2O. FRC was
300 ml
600 ml
900 ml
5 cmH2O
102±41 mmHg
870±392 ml
10 cmH2O
130±63 mmHg
1189±613 ml
15 cmH2O
147±82 mmHg
1435±670 ml
ANOVA two way RM: P<0.05 either PEEP or TV, no significant
interaction between PEEP and TV.
measured with the Helium dilution technique during an
end-expiratory pause. Data are expressed as mean ±sd.
Moreover, we found a significant correlation between
PaO2 and FRC in every considered conditions (r =0.61,
Our results showed that both the level of PEEP and the
level of tidal volume, but not their interaction, have a significant effect on improving FRC and gas exchange.
Ranieri et al.: Am Rev Respir Dis 1991, 144:544–551.
Beneficial effects of lung recruitment with PEEP after cardiac surgery
PT Morley, K Rees, JJ Presneill, M Rowlands, JF Cade and J Tatoulis
Cardiothoracic Intensive Care, The Royal Melbourne Hospital, Victoria, 3050, Australia
Introduction: Delayed extubation and pulmonary atelectasis increase resource utilization after cardiac surgery. Lung
recruitment techniques may help prevent these problems.
Objectives: To study the cardiovascular and respiratory
effects of a recruitment maneuver using PEEP in postoperative cardiac surgical patients.
Methods: A prospective, randomized, double-blind study
was performed. All patients were stabilized on a standard
ventilatory protocol (10 ml/kg at 10/min; with FIO2 of 0.8
and PEEP=5 cmH2O). PEEP was either left at 5 cmH2O
(control) or increased (1 cmH20/min) until 15 cmH20
(recruitment), when cardiovascular endpoints were
reassessed. PEEP was then returned to baseline at the
same rate. Primary endpoints were the proportion of
patients achieving predetermined oxygenation criteria for
extubation (PaO2 ≥70 mmHg on ≤50% FiO2) at 1 h, and
time to extubation.
Results: 100 patients (mean age 63 years; 79% male)
were randomized to either recruitment (n=55) or control
(n=45). In the recruitment group, at peak PEEP there was
a small but statistically significant increase in PAWP and
CVP, with a small decrease in mean BP and cardiac index.
Time to extubation was similar in each group. Oxygenation
criteria for extubation at 1 h were achieved by 93% of the
total recruitment group versus 82% of controls (P=0.13).
In the subgroup with an initial PaO2/FIO2 ratio <300, 94%
achieved oxygenation criteria for extubation at 1 h versus
61% of controls (P<0.01). Chest radiographs taken after
completion of the protocol showed significantly less basal
pulmonary collapse in the recruitment group (20% versus
53% in controls; P<0.01).
Conclusion: The use of this incremental PEEP maneuver
after cardiac surgery improves pulmonary gas exchange,
probably via recruitment of otherwise collapsed pulmonary
segments. These benefits are more pronounced in those
patients with initially poor gas exchange.
Effects of PEEP and tidal volume on elastances and distribution of volume changes of the
different chest wall compartments
A Aliverti, R Dellacà, A Lo Mauro, E Carlesso, W Del Frate, P Pelosi*, D Chiumello* and A Pedotti
Dipartimento di Bioingegneria, Politecnico di Milano, Italy; *Istituto di Anestesia e Rianimazione, Università di Milano, Ospedale
Maggiore IRCCS, Milano, Italy
We used optoelectronic plethysmography [1] to analyze
six sedated and paralyzed ALI/ARDS patients receiving
continuous positive pressure ventilation (CPPV) with different settings of PEEP (5, 10, 15 cmH2O) and tidal
volume (TV, 300, 600, 900 ml). The aim of the present
work is to study the effects of PEEP and TV on volume
distribution in three different chest wall compartments
(upper thorax, VUT, lower thorax, VLT and abdomen, VAB)
and on their elastances (EUT, ELT and EAB, respectively)
assessed by the end-inspiratory occlusion technique.
Results are expressed in the table as mean ±SD.
We found that both PEEP and TV, but not their interaction, have an effect on volume distribution in the UT and
AB (P<0.05). At increasing PEEP and TV, the increased
abdominal and decreased upper thoracic contribution to
inspired volume were associated to a decreased compliance of the UT. Finally, we found significant (P<0.001)
Critical Care
Vol 4 Suppl 1
PEEP (cmH2O)
TV (ml)
20th International Symposium on Intensive Care and Emergency Medicine
Vut (% TV)
Vlt (% TV)
Vab (% TV)
Eut (cmH2O/L)
Elt (cmH2O/L)
Eab (cmH2O/L)
exponential relationships between UT and AB contributions to TV and EUT.
We conclude that at high PEEP and TV, the elastance of
the upper thorax decreases and it causes a more ‘abdominal’ distribution of TV.
Aliverti A, Dellacà R, Pelosi P, Chiumello D, Pedotti A, Gattinoni L:
Opto-electronic plethysmography in intensive care patients. Am J
Resp Crit Care Med (in press).
Effects of PEEP above the L-Pflex on gas exchange, hemodynamic and gastric tonometry in
ARDS patients
CSV Barbas, E Silva, C Hoelz, AG Garrido, EC Meyer and E Knobel
ICU, Hospital Israelita Albert Einstein, São Paulo,SP-Brazil, Av. Albert Einstein 627/701-5 andar-CTI-adultos-CEP-105651-901,
São Paulo, SP-Brazil
PEEP titrated by the L-Pflex of static PxV curve of respiratory system improved oxygenation and mortality in ARDS
patients. In order to study the acute effects of PEEP on
oxygenation, CO2 exchange, hemodynamic parameters
and gastric tonometry we analyzed six ARDS patients (<5
days of installation) after 30 min on PEEP of 5 cmH20,
then 30 min on PEEP 2 cmH20 above the L-Pflex and then
30 min after PEEP of 5 cmH20 again. They were all
sedated and paralyzed. A Swan-Ganz catheter with a continuos cardiac output (Baxter®) and a continuos gastric
tonometer (Tonocap®) were inserted in each patient. The
tidal volume and respiratory rate were kept constant (8
ml/kg and 20/min- VCV-Ventilation-Servo-Siemens-300).
L-Pflex was titrated by the static PxV curve of the respiratory system (random volumes).
Results: See Table.
Conclusion: PEEP titrated by the L-Pflex of the static PxV
curve of the respiratory system improved oxygenation
without impairing global as well as regional hemodynamic
parameters in ARDS patients.
30 min after
PEEP=5 cmH20
30 min after
PEEP 2 cmH20 >L-Pflex
30 min after
PEEP =5 cmH20
Minute ventilation (L/min)
PaCO2 (mmHg)
PgCO2 (mmHg)
Arterial Lactate (mg/dL)
Cardiac Index
Mean arterial pressure (mmHg)
Influence of different PEEP levels and tidal volumes on the regional nonaerated tissue:
experimental study
N Bottino, G Eccher, P Pelosi†, A McKibben*, A Adams*, M Goldner*, JJ Marini* and L Gattinoni
1st. Anestesia e Rianimazione, Osp. Maggiore Policlinico-IRCCS, via F. Sforza, 35, 20122, Milano, Italy; *University of
Minnesota, St Paul, MN, USA; †Dip. Scienze Cliniche e Biologiche, Univ. dell’Insubria, Varese, Italy
Introduction: In the Acute Respiratory Distress Syndrome
(ARDS), although the edema is homogeneously distributed, a hydrostatic pressure gradient squeezes the gas
out from the alveoli, causing a lung density increase along
a gravity gradient. We studied by CT scan, in an oleic acid
induced ARDS experimental model, the influence of different VT and PEEP levels on the regional nonaerated tissue
during volume-controlled ventilation.
Methods: ARDS was induced in six sedated, paralyzed
and mechanically-ventilated dogs by an oleic acid dose
(0.075 mg/kg) injected into the pulmonary circulation.
After ARDS induction, the volume-controlled ventilation
was maintained, changing three tidal volumes (VT) (12, 24
and 36 ml/kg) and two PEEP levels (5 and 15 cmH2O).
During each of the six randomized steps a CT scan
section has been taken, at end-expiration and end-inspiration, 5 cm above the diaphragm. For data analysis, we
divided the CT scan section into an upper and a lower
part. The nonaerated tissue mass of the upper or lower
region was calculated from the densities (ρ) (–200 H <CT
<+100 H) and the volume (V), according to the formula
m=ρ×V [1]; then it was standardized for the area of each
lung region. To evaluate the effect of VT, PEEP level and
lung region (upper/lower) on the amount of nonaerated
tissue, we performed analysis of variance, in condition of
end-expiration and end-inspiration, respectively.
Results: The behavior of nonaerated tissue (g/cm2) is
shown in the table (data expressed as mean ±SD).
The amount of nonaerated tissue is greater in the lower
region than in the upper one, in both end-expiration and
end-inspiration (P<0.01). In both regions PEEP and VT
independently affect the amount of nonaerated tissue
(P<0.01). There is interaction between PEEP and VT only
at end-inspiration (end-expiration: P=n.s.; end-inspiration:
Conclusion: The effects of PEEP level and VT on lung
recruitment are greater in dependent lung regions and at
Pelosi P et al.: Am J Respir Crit Care Med 1994, 149:8–13.
5, low VT
5, normal VT
5, high VT
15, low VT
15, normal VT
15, high VT
Computer tomography of chest organs in the diagnosis of acute respiratory distress-syndrome of
patients with severe brain injury
NV Agafonova, SA Kravtsov and EP Rodionov
State Scientific and Clinical Center of Coal Miners’ Health Protection, Leninsk-Kuznetsky, Russia, 652509
Purpose: Early diagnosis of acute injuries of lungs in
patients with severe brain injury (SBI) during acute period
providing, with the comparison of results in dynamic
control, reduction of radial load and of examination time.
Materials and methods: We examined 25 patients with
SBI (in admission the severity rate was less than 8 points
according to the Glasgow scale) at 1, 2, 3, 5, 7 and 14
days. We developed the algorhythm of patients examination: high resolution computer tomography (HRCT) of four
levels with assessment of densitometric index in four
points of each level.
Results: Roentgenologic signs of 96% patients during 1
day; from these patients 62.5% had first rate with the
increase of density index by 13.7%, 37.5% patients had
second rate with the increase of density index by 25.2% in
comparison with density index. Main localization of pathologic changes registered in posterior basal parts of the
lungs. Pleural exudate was diagnosed in 50%, hydrothorax
in 4%, aspiration pneumonia in 8% patients.
Conclusion: The proposed method, in contrast to standard CT, allows the appreciation of the revealed changes,
to diagnose the early symptoms of ARDS on the first day,
to analyze the development of pathologic in dynamics and
to conduct the treatment correction in time.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Partial liquid ventilation combined with two different gas ventilation strategies in acute lung
injury in piglets
G Zobel, S Rödl, B Urlesberger, I Knez and D Dacar
Departments of Pediatrics, Neonatology, and Cardiac Surgery, University of Graz, Austria
Objectives: To investigate the effects of two different gas
ventilation strategies during partial liquid ventilation (PLV)
on gas exchange and lung mechanics in acute lung injury.
and expiratory airway resistances and arterial blood
gases were measured. Data were obtained at baseline,
after lung injury and at 60 and 120 min of PLV. See
Table 1.
Design: Prospective, randomized, experimental study.
Setting: University research laboratory.
Subjects: Fourteen piglets weighing 8.9±0.2 kg.
Interventions: Acute lung injury was induced by the infusion of oleic acid and repeated lung lavages with 0.9%
NaCl (20 ml/kg). After a stabilization period of 30 min the
animals were randomized to two groups with different
PEEP levels and tidal volumes (Group A: Vt >12.5 ml/kg,
PEEP 6 cmH2O; Group B: Vt 9 ml/kg, PEEP 12 cmH2O).
The pre-oxygenated and warmed perfluorocarbon liquid
(30 ml/kg) was instilled into the trachea over 15 min
without changing the ventilator settings.
Measurements and main results: Airway pressures,
tidal volumes, static respiratory compliance, inspiratory
Data analysis: Values are given as mean ±SEM. Comparisons were made by ANOVA for repeated measures. A Pvalue of <0.05 was considered significant. The infusion of
oleic acid (0.07 ml/kg) combined with five lung lavages
induced a significant reduction of PaO2/FiO2 from
486±24 torr to 60±3.8 torr (P<0.01).
Conclusions: During PLV both gas ventilation strategies
significantly improved oxygenation in an animal model of
acute lung injury. Static compliance of the respiratory
system was not different between both gas ventilation
strategies. Inspiratory airway resistance was significantly
higher during the low PEEP-high Vt gas ventilation strategy, whereas the high PEEP-low VT strategy resulted in
significantly higher PaCO2 values.
Table 1
Blood gases and respiratory variables during PLV with two different strategies of gas ventilation.
PaO2/FiO2 (torr)
PaCO2 (torr)
64±2.4** ##
66±2.1** ##
Group A: Vt >12.5 ml/kg, PEEP 6 cmH2O, Group B: Vt 9 ml/kg, PEEP 12 cmH2O; *P<0.05 vs ALI, **P<0.01 vs ALI, BL: baseline; ALI: acute lung
injury; #P<0.05 vs group A, ##P<0.01 vs group A.
Oxygenation response to prone positioning in pulmonary and extrapulmonary ARDS
O Akinci, M Cetin, P Ergin, F Esen, L Telci and N Cakar
Department of Anesthesiology and Intensive Care, University of Istanbul, Medical Faculty, Istanbul, Turkey
Introduction: Various etiologic factors change clinical response, morbidity, and mortality rates in ARDS. In this study,
we investigated the oxygenation response of pulmonary
and extrapulmonary ARDS patients to the prone position.
Methods: Thirty-two ARDS patients (Group1: n=16 with
pulmonary causes and Group2: n=16 with extrapulmonary
causes) were evaluated retrospectively. ARDS was
defined according to the criteria established by the Ameri-
Table 1
Table 2
Group 1 (PaO2/FiO2)
Group 2 (PaO2/FiO2)
Supine P
Prone P 1st hour
Prone P 4th hour
Prone P 24th hour
can-European Consensus Conference on ARDS [1].
PaO2/FiO2 ratios, PaCO2, PCWP values in the supine
position (baseline) and then the same values in the prone
position at the 1st, 4th and 24th h were obtained from the
medical records. ARDS, APACHE II, and MOF scores on
admission, and the first day of prone positioning were calculated and compared. Statistical evaluations were performed by student t-test.
Results: Murray, APACHE II and MOF scores were similar
in both groups during admission and there was no difference in oxygenation response to prone positioning
between two groups (Table 1). However, the MOF score
was higher in extrapulmonary ARDS on the day of prone
positioning. Mortality rate was higher in the extrapulmonary group and on discharge this group had higher
Murray, APACHE II and MOF scores (Table 2).
Group 1
Group 2
on admission
on discharge
on admission
on discharge
on admission
on discharge
Conclusion: Our retrospective data suggest that oxygenation response to prone positioning was similar in both
groups, but mortality rates were higher in extrapulmonary
ARDS patients, which correlates with MOF scores.
Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy
M, Le Gall JR, Morris A, Sprag R, and the consensus committee:
American- European Consensus Conference on ARDS. Definition,
mechanism, relevant outcomes and clinical trial coordination. Am
J Respir Crit Care Med 1994, 149:818–824.
Prospective study to evaluate the kind of prone position concerning nursing, clinical outcome
and material and personnel resources
TR Neubert, R Stiletto and L Gotzen
Centre of Operative Medicine, Department of General Surgery, University of Marburg
Introduction: Acute Respiratory Distress Syndrome
(ARDS) is one of the most common, potentially lethal
disease processes encountered in critical care with
extremely high mortality of about 60%. Researchers have
found that a significant improvement in gas exchange
often occurs when ARDS patients are turned from the
supine to the prone position. Different reasons are discussed for this effect: reduction of oxygen toxicity, recruitment of alveolar space and optimisation of postural
drainage. But there are a lot of difficulties in nursing these
patients and these depend on the kind of prone position.
The dangers of pressure damage and oedema formation
increase in the prone position. Also the possibility of
suction and observation is decreased. So the major goal is
to find out the best kind of prone position.
The available study is designed by physicians and nurses.
Besides the question of patients’ benefit concerning gas
exchange, handling, the acceptance of the nurses, and
economic consequences are proved.
Method/material: Patients with ARDS, or those patients
identified as requiring to be nursed in the prone position
with a Horowitz-Quotient (PaO2/FiO2) <250, were turned
over into the prone position. In a randomised procedure
the patients were placed into a 180° prone position (face
down), a 135° prone position (near side position) or they
were treated in a Rotation bed (RotorestR). Gas analyses
defined the clinical effect of the position on gas exchange.
Also changes in skin integrity, skin status and the clinical
outcome of proning were documented.
Additionally, the number of nurses/physicians being
involved in positioning the patient and the time taken was
The prone position interval is fixed at 4–6 h depending on
clinical data and the personnel situation.
End of the positioning-treatment is defined by clinical data
and a Horowitz quotient >300.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Results: Twenty-two patients were positioned in the near
side prone position. Eight patients were positioned in the
180°-position and five patients were treated in the rotation
The distribution of the patient identified as requiring to be
nursed in prone position shows the preferences of the
prone position. In handling, nursing and observing, the
near side position is the preferred prone position. The clinical outcome is comparable to the other forms of pronepositioning. In comparison with the other kinds of position
the risk of complication (skin damages, oedema formation,
lost of catheter or tube) is very small. On average you
need one physician and two very well-introduced nurses
for this positioning. So at every point of time, position
changing can occur. For the face down position (180°) on
average you need four to five well-introduced persons at
minimum for a position changing. Additionally you need
sufficient place to range two beds side by side. The possibility of observing, nursing and suctioning is less in the
face down position than in the near side position. Our findings show a similar clinical effect concerning the gas
exchange in the 180° position as in the 135° position. But
for a clear position the number of researched patient is too
Also the patients treated in rotation bed shows a similar
clinical outcome concerning the gas exchange.
But here also the handling and the economic resources
necessary are incomparably high and at every time you
need a special introduced nurse.
Conclusion: This study is being continued to get evident
data for a clear point of view. The target is to develop
evident criteria for the kind of prone position, not only in
dependence of the clinical outcome but also concerning
the economic and personnel possibilities in an ICU. A clinical treatment is not only orientated in doing the best for
the patient but it is also limited by personnel acceptance
and economic resources. The kind of treatment is established when the benefit for the patient is recognisable,
when the handling is simple, does not need complicated
equipment and a lot of personal resources. It must be safe
and show a small quantity of complications.
These criteria, as shown by our study, are applicable for
the near side position. Our goal is to develop criteria for
the different kind of positioning possibilities concerning
the clinical problems in gas exchange during the disease
Inhaled nitric oxide in infants and children with ARDS
G Zobel, S Rödl, M Trop and HM Grubbauer
Department of Pediatrics, University of Graz, Austria
Objective: To evaluate the effects of inhaled nitric oxide
on gas exchange and outcome in pediatric patients with
acute respiratory distress syndrome (ARDS).
Interventions: Initially inhaled (NO) was applied at 10
ppm using a microprocessor based system. A positive
response after 15 min of NO inhalation was defined as an
increase in arterial oxygen saturation >5%.
Design: Case series report.
Setting: Pediatric intensive care unit of a tertiary care children’s hospital.
Patients: Seventeen pediatric patients with ARDS requiring mechanical ventilation with an FiO2 >0.5 at a positive
end-expiratory pressure >6 cmH2O, and whose PaO2/FiO2
ratio was <100 torr were enrolled in this study.
Measurements and main results: At the start of NO
inhalation the oxygenation (OI=Paw*FiO 2*100/PaO2) and
ventilation (VI=PaCO2*PIP*RR/1000) indices were
29±2.8 and 43±4.1, respectively, the PaO2/FiO2 ratio
was 69±7 torr, and the static compliance of the respiratory system 0.51±0.04 ml/cmH2O/kg. An initial positive
response to inhaled NO was observed in 59% of
Table 1
Change in oxygen saturation 15 min after starting NO inhalation, duration of MV, number of ECMO support, and outcome of
pediatric patients with ARDS (n=17).
Number of patients
SpO2 (%)
Duration of MV (d)
ECMO support (n)
S/NS (Mortality
*P<0.01; S=survivors; NS=nonsurvivors; MV=mechanical ventilation; ECMO=extracorporeal membrane oxygenation.
Conclusion: Inhaled NO significantly improves oxygenation in 59% of infants and children with ARDS. However,
an initial positive response to inhaled NO does not reduce
the need for ECMO support and does not influence duration of mechanical ventilation and outcome of pediatric
patients with severe ARDS.
Role of hyperbaric oxygen therapy (HBOT) in recovery of cardio-pulmonary function: survival of
patients developing ARDS following closed chest trauma (CCT)
GG Rogatsky and EG Shifrin
Faculty of Life Sciences, Bar-Ilan University, 52900 Ramat-Gan and Department of Vascular Surgery, Sourasky Medical Center,
Sackler School of Medicine, Tel-Aviv University, 64239 Tel-Aviv, Israel
Many years mortality in ARDS, often complicating CCT, is
still very high and reaches 70–80%. In ARDS patients,
hypoxia is a result of a ‘vicious circle’ when acute respiratory insufficiency intensifies acute cardiac insufficiency
which, in turn, deepens respiratory failure.
tory functions were checked using a gas-analyzer. All
ARDS patients were divided into three groups: group A–4
surviving patients after conventional therapy; group B–14
deceased patients after conventional therapy, and group
C–3 patients surviving after addition of HBOT.
Patients and methods: The study is based on the data
obtained from 45 patients with moderate–severe CCT.
During conventional treatment in ICU, after 48–72 h, in 21
patients (48%) an ARDS developed, and three of these
patients were treated with HBOT in a mono-placed
barochamber. All treatments were conducted under
1.6–2.0 ata for 40–60 min each every day during 4–15
days, depending on progress of the recovery. Monitoring
of cardiac function was carried out using non-invasive
impedance cardiography (IC). Stroke Volume Index (SVI)
and Cardiac Output Index (COI) were calculated. Respira-
Results: Obtained data was exposed to statistical analysis using Student’s unpaired t-test and the results are presented in Table 1.
Conclusions: This study clearly shows that: 1) in ARDS
patients after CCT the state of cardiac function is the
factor determining development of respiratory hypoxia; 2)
HBOT is a decisive treatment improving cardio-respiratory
function which leads to the favorable outcome; 3) patients
with CCT must be treated with HBOT just after the
trauma, before ARDS has developed.
Table 1
PaO2 (mmHg) -mean
SVI (ml m-2) - mean
COI (I min-1m-2) - mean
78. 0**##
The differences between the means was considered significant if P<0.05 (* or
vs group B.
HR(b min-1) - mean
P<0.01 (** or
* - vs phase I; ** - vs phases I and II; # or ## -
The incidence of ARDS, interim results of the East Anglian ARDS Registry
JM Dixon and KEJ Gunning
John Farman Intensive Care Unit, Box 93, Addenbrooke’s NHS Trust, Hill’s Road, Cambridge, CB2 2QQ, UK
Background: There is a wide variation in the reported incidence of acute respiratory distress syndrome (ARDS) due
to the use of different diagnostic criteria. The publication
by the American European Consensus Conference in
1994 of diagnostic criteria for ARDS has made comparisons of the incidence of ARDS more reliable. Only one
study has looked at the incidence of ARDS in the UK. This
was a retrospective survey in 1988 [1] that reported an
incidence of 4.5/100 000 population/year.
Method: We report a prospective observational study of
the incidence of ARDS in 10 ICUs in 9 hospitals covering
a population of 1.89 million people over the age of 15
years. The consensus conference criteria were used to
diagnose ARDS. Acute onset was defined as occurring
within five days of onset of illness. A study co-ordinator in
each participating ICU identified patients and there was
regular feedback from the study organisers to ensure all
cases were identified.
Results and discussion: Over the first 10 months of data
collection, 68 patients met the criteria for the diagnosis of
ARDS, giving an incidence of ARDS of 4.3/100 000 population/year. The mean age was 50.9 years. The mean
length of stay on ICU of survivors was 16.9 days and 7.8
days for non-survivors, the intensive care mortality was
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
38%. These interim results give an incidence of ARDS
that is similar to that reported by the previous UK study
and recent studies from other countries [2]. This figure
would now appear to be a more accurate estimate of the
incidence of ARDS in an unselected population. The study
is ongoing and it is planned to follow up the survivors on a
long-term basis.
Webster NR, Cohen AT, Nunn JF: ARDS, how many cases in the
UK? Anaesthesia 1988, 43:923–926.
Valta P, Uusaro A, Nunes S et al.: Acute respiratory distress syndrome: frequency, clinical course and costs of care. Crit Care Med
1999, 27:2367–2374.
ARDS before and after the start of mechanical ventilation
A Anzueto, A Esteban, I Alía, L Brochard, T Stewart, F Frutos and MJ Tobin
International Mechanical Ventilation Study Group
We determined if outcome differed in patients who developed ARDS before and after the institution of mechanical
ventilation (MV). A prospective study of 5156 patients
who underwent MV >12 h was done in 361 ICUs from 20
countries. Among those patients, 261 (5%) had ARDS at
the onset of MV (prior-group), and 313 (6%) developed
ARDS 48 h or more after the onset of MV (after-group).
The mean time from onset of MV to development of ARDS
was 4±2 days. In the after-group, MV was initiated for the
following conditions: sepsis (19%), aspiration (16%),
pneumonia (15%), and trauma (15%). MV parameters
were similar in both groups. The patient characteristics
and outcomes are shown in Table 1 (mean±SD).
Patients who developed ARDS after onset of MV were
more likely to have ventilator associated pneumonia (38%
vs 26%, P<0.05) and coagulopathy (31% vs 23%,
Table 1
P value
Age (years)
SAPS II score
Duration MV (days)
In-ICU mortality
In-hospital mortality
P=0.03). In summary, patients who developed ARDS after
the onset of mechanical ventilation were more likely to
develop complications, required longer duration of ventilator support, and had higher ICU and hospital mortalities.
Prognostic factors in severe ARDS-patients
U Friess, E Münch, H Roth, M Quintel and K van Ackern
Department of Anaesthesiology and Intensive Care, Universitätsklinikum Mannheim, University of Heidelberg, Germany
Aim of the study: To evaluate the prognostic value of different parameters on the outcome of conventionally treated
ARDS-patients, admitted to the ICU as potential extra-corporal-membrane-oxygenation (ECMO) candidates.
Material and methods: Between 1995 and 1999, 115
patients with severe ARDS were transferred to the ICU of
a University hospital as potential ECMO candidates. 22
patients underwent ECMO-therapy, 6 died soon after
admission. 66 patients were treated with conventional
therapy. In these 66 patients, we evaluated epidemiological characteristics, days in hospital before admission, days
on the ventilator before admission, length of ICU-stay,
days on the ventilator at the ICU, ventilator parameters
(PEEP, peak airway pressure, mean airway pressure,
PaO2/FiO2-ratio, compliance), biochemical parameters
(haemoglobin, hematocrit, lactate, CRP, leukocytes,
thrombocytes, creatinine), APACHE II and fluid balance on
admission (1), 24 (2), 48 (3) and 72 (4) h after admission.
Data are given as mean ±SD. The Mann-Whitney U-test
was used to identify significant differences between the
Results: Forty-four of 66 patients survived, with an
average Glasgow-outcome-scale of 3.88±0.44. The mortality rate was 33% and mainly caused by MODS due to
sepsis. None of the patients died of hypoxemia. The nonsurvivors were significantly older (44.64±21.2 years) than
the survivors (32.39±12.2 years, P<0.001), and had more
concomitant diseases (28 vs 12), and a longer course of
disease before hospitalisation (7.62±9.95 vs 2.90±10.33
days, P<0.005). For hospital stay (12.54±9.99 vs
10.0±7.88 days) or days on the ventilator before transfer
(8.74±9.41 vs 6.98±6.48 days) no difference could be
shown. After admission the nonsurvivor group received
more days of controlled ventilation (9.71±5.81 days) than
the survivor group (5.21±5.69 days, P<0.001), with
regard to ICU stay no difference could be shown
(19.09±14.99 vs 18.34±11.31 days, ns). Significant differences between survivors and nonsurvivors are shown in
Table 1.
No significant differences could be demonstrated for biochemical parameters, fluid balance, and compliance.
Table 1
Significant differences between survivors and non-survivors, level of significance P<0.001
PEEP [cmH2O]
PEAK [cmH2O]
MEAN [cmH2O]
PaO2/ FiO2 [mmHg]
Conclusion: High APACHE II values, high PEEP, peak,
and mean-airway-pressure levels and a low PaO2/FiO2ratio are demonstrated to be of high prognostic value in a
P133 13C-stable
non-ECMO treated severe ARDS population. None of the
patients included in this survey died of hypoxemia, poor
outcome was mainly determined by MODS due to sepsis.
isotope analyses technique for ICU patients
U Hölscher, T Gruchmann, Ch Goeters*, C Schulzki* and H Van Aken*
Faculty of Applied Physics and Engineering, University of Applied Sciences, Hüfferstr. 27, D-48153 Münster; * Department of
Anaesthesiology and Intensive Care Medicine, University of Münster, Münster, Germany
Introduction: By 13C isotope spectroscopy, a noninvasive
diagnostic tool for specific functions of different organs is
available [1]. A 13C labelled tracer is administered and is
metabolised by a specific pathway to 13CO2. The increase
of exhaled 13CO2 is analysed and gives an indication of a
specific function of the organ under test. The availability of
sensitive (resolution 0.3‰) and stable isotope-selective
infra-red (IR) spectrometers has replaced expensive massspectrometers. The present IR analysers rely on two conditions: 1) on co-operating patients who collect breath
samples into containers and; 2) on equilibrated test gas
Methods: Online monitoring of ventilated ICU patients over
a long period requires an automated collection of the
exhaled gases and an equilibration of the CO2 content
without compromising the lung ventilator set-up. The
sample should only contain exhaled gases, a dilution would
increase the necessary sensitivity of the analyser. A collection and mixing device has been developed that collects
the exhaled gases and does not interfere with the safety
and monitoring concept of the lung ventilator. The unit collects the sample gas from an Evita ventilator (Dräger) and
supplies it to an Infra-Red Isotope-selective Spectrometer
(IRIS, Wagner) which has been modified to analyse a continuous gas stream. The expiratory concentrations of
13CO and 12CO , the ratio of 13CO /12CO as well as the
cumulated hourly percentage of the recovery of the administered 13C label are determined by evaluation IRIS software (V2.3 1998, Wagner) and stored on a computer.
Results: The risk analysis and technical set-up have been
reviewed and certified by the TÜV-Rheinland. The collection unit connecting the lung ventilator to the analyser has
proven not to interfere with the functionality of the therapeutic device. The set-up was validated in volunteers and
in ICU patients. Results of a first study determining the
gastric emptying by isotope techniques will be discussed
in parallel on this congress.
Conclusion: The stable IR isotope spectroscopy can be
applied to ventilated patients. This allows a non-invasive
focus e.g. on liver enzyme activities and on other organ
specific problems.
Ghoos Y: 13CO2 breath tests at the laboratory ‘digestion-absorption’. University Hospital Gasthuisberg, Leuven 1996.
Metabolism in abdominal organs, as evaluated by microdialysis, in experimental severe acute
P Ederoth, ZW Sun and R Andersson
Departments of Anesthesiology and Surgery, Lund University Hospital, S-221 85 Lund, Sweden
Introduction: Septic states might induce regional variations in metabolism. The aim of the present study was to
evaluate if regional metabolic differences could be identi-
fied in a model of severe acute pancreatitis in the rat, with
the use of a microdialysis technique.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Methods: Under full anesthesia, microdialysis probes
were inserted in the parenchyma of the pancreas and liver
and under the serosa of the small intestine. Microdialysate
was collected every 10 min. After a baseline period of 60
min, acute pancreatitis was induced by intraductal injection of 0.20–0.29 ml 5% sodium taurodeoxycholate. The
animals were studied for three h after induction of pancreatitis after which they were killed. Arterial blood samples
were taken every 60 min. The microdialysis fluid and blood
were analyzed for glucose and lactate. Six groups were
studied with six animals in each group; Sham, Pancreatitis
and four groups with treatment given 15 min after induction of acute pancreatitis (early treatment). Treatment
administered included N-acetylcystein (NAC; 200 mg/kg
i.v), a platelet-activating factor (PAF) antagonist (lexipafant,
5 mg/kg) and monoclonal antibodies against the adhesion
molecules ICAM-1(0.2 mg) and PECAM-1 (0.2 mg).
Results: The levels of lactate and glucose in pancreas
and lactate in blood were higher (P<0.05) in the pancreatitis group compared with the sham group. The
increases of glucose and lactate in the pancreas were
higher than in the blood, demonstrating an earlier and
stronger change in metabolism in the pancreas as compared with the rest of the body. This finding was higher
(P<0.05) in the pancreatitis group than in the sham
group. There were no clear differences between the
various pancreatitis groups, with or without treatment.
and no differences in between the treatment groups.
Within the sham group, glucose showed no regional difference, but there was a significant increase in the pancreas in pancreatitis animals as compared to the liver and
intestine. Regarding lactate, there were differences
between the pancreas, liver and intestine in both groups
with higher levels seen in the pancreas.
Conclusion: An increase in pancreatic lactate concentrations in pancreatitis animals was seen and this increase
seemed to have its origin in the pancreas and not in the
whole body. Similar changes were also found for glucose.
Early treatment, as described earlier, had no obvious
effect on these parameters. The results imply that microdialysis could be of potential future value in monitoring
metabolic locoregional differences in critical illness.
Acute necrotizing pancreatitis — histological study focusing on necrosis and apoptosis
Y Haraguchi, Y Tomoyasu, N Shindo, J Inoue, Y Itoh, M Hoshino, S Hasegawa, T Ishihara and T Nagata
National Hospital Tokyo Disaster Medical Center, Tokyo, Japan
Acute necrotizing pancreatitis is histologically studied focusing on the degree and extent of necrosis and apoptosis.
Materials and methods: Evaluation was done by lightmicroscopic examination (Hematoxylin-Eosin stain) and
electron-microscopic examination. In order to judge the
existence of apoptosis, the TUNEL method was applied.
Only exocrine tissue was surveyed. The number of specimens is three, which were obtained by distal pancreatectomy in the early phase of acute necrotizing pancreatitis
because they resisted intensive care. The normal control
was one specimen in which 1) negative control and; 2) positive control (TACS-Nuclease-treated control) were used.
Results: Pancreatic tissue was divided into three areas by
light- and electron- microscopy before TUNEL: 1) microP136
scopically normal tissue; 2) microscopically apparently
necrotized area; 3) minimal or almost normal tissue by the
light microscope, although electron-microscopically mild
or moderate destructive change of the cytoplasm and
nuclear structure was apparent. The TUNEL method
revealed the existence of apoptosis in both the the necrotized area (2) and in the minimally destroyed area lightmicroscopically (3).
Discussion and conclusions: From our study, the existence of apoptosis in acute necrotizing pancreatitis is
thought to exist. Several important themes seem to be
pointed out: 1) new classification of acute pancreatitis,
including apoptosis may be necessary; 2) apoptosis will
influence the prognosis or mortality of necrotizing pancreatitis, hitherto believed to depend on the ratio of necrosis;
3) the criteria of severity should be revised.
Inflammatory response, prognostic scores and mortality in acute pancreatitis
MV De la Torre Prados, MA Herce Muñoz*, A Soler García, C Reina Artacho, A García Alcántara, A Poullet Brea
and M Morell Ocaña*
Servicio Medicina Intensiva, *Servicio de Bioquímica, H.U. Virgen Victoria, 29010-Málaga, Spain
Objective: The object of the study was to relate, in the
first 72 h following hospital admission, biochemical
markers of inflammatory response and prognostic scores
to mortality in acute pancreatitis (AP).
Materials and methods: For a year (November 1997 to
October 1998) a prospective study was made consecutively of 30 patients admitted to the Emergency Department where AP was diagnosed; 83.3% were in the
Intensive Care Unit (ICU) with three or more Ranson criteria and 24 h APACHE score >8. At 12, 36, 48 and 72 h
12–30 h
30–48 h
48–72 h
221 (17)
213 (29)
210 (10)
190 (34)
162 (43)
163 (45)
CPR mg/dl
11 (7)
10 (8)
*33 (7)
*19 (12)
29 (15)
23 (12)
IL-6 pg/ml
123 (67)
138 (80)
348 (97)
301 (85)
184 (86)
182 (82)
IL-8 pg/ml
*107 (92)
*55 (43)
*87 (62)
*40 (31)
28 (19)
52 (32)
TNF pg/ml
20 (12)
13 (5)
*45 (32)
*13 (6)
*29 (10)
*15 (10)
Elastase µg/l
24 h
48 h
72 h
Values are shown as means(SD). NS=Nonsurvival, S=Survival. (*) T-test, P<0.05.
biochemical parameters were measured: elastase, C-reactive protein (CRP), interleukin (IL)-6, IL-8 and Tumor
Necrosis Factor (TNF). At the same time 24, 48 and 72 h,
APACHE scores and 24 and 48 h Ranson scores were
56.7% had some complications and the hospital mortality
was 16.7%. In relation to mortality there was a significant
difference (T-test) in the number of dysfunctional organs
3.8(1.3) vs 1.7(1.2) and in the Multiple Organ Dysfunction
(MOD) score 8.8 (6) vs 3 (2.2).
Results: Thirty patients with a mean age of 62.7 (13.5)
years were studied. 60% were women and 80% had
biliary etiology. The median interval from pain until hospital
admission was 16.6 h. 67% showed C or D grade Balthazar Scale. 60% had surgery at 38 days median period.
Conclusions: In the nonsurvival group, IL-8 increased significantly in the first 30 h and CPR, IL-8 and TNF between
30 and 48 h. Between 48–72 h, TNF continued to
increase. There were differences in APACHE and Ranson
Scores, but these were only significant at 48 h.
Indocyanine green clearance monitoring during liver transplant for fulminant hepatic failure:
preliminary results
G Della Rocca, MG Costa, S Monaco, D Romboli, S Costini and P Pietropaoloi
Istituto di Anestesiologia, University of Rome “La Sapienza”, Viale del Policlinico n°155-00161 Rome, Italy
Introduction: The COLD System (Pulsion Medical
System, Munich, Germany) allows measurements of
cardiac index (CI), intrathoracic blood volume (ITBV),
extravascular lung water (EVLW) and liver function
through the plasma disappearance rate of indocyanine
green dye (ICG) (PDRdy n.v.=20–30% min).
Methods: Six patients (6F), with a mean age of
42.2±11.7 years and with a mean BSA of 1.7±0.1 m2,
affected by fulminant hepatic failure (FHF), undergoing
orthotopic liver transplantation (OLT) were studied with
the COLD System to evaluate PDRdy as a marker of
native and transplanted liver function and volumetric data.
After anesthesia induction, a pulmonary artery catheter
(Intellicath CCO; Baxter, Irvine, CA, USA) was placed via
the right subclavian vein and a 3Fr thermistor-tipped
fiberoptic catheter for thermal dye dilution was inserted
into the descending aorta via a 4Fr introducer sheath in
the femoral artery and connected to the COLD-Z021. All
data were collected during the following steps: after anesthesia induction (A), during anhepatic phase (B), after
graft reperfusion (C), at the end of surgery (D) and 6 h
after the end of surgery (E).
Results: Results are reported in the table. Four patients
had a rapid weaning, were extubated, discharged from
ICU and are still enjoying good health. Two patients,
despite a good liver function recovery, died with MOF.
Conclusion: The evaluation of cardiac preload is a useful
tool to guide fluid replacement and vasoactive drug admin-
856±101 692±124 838±204
818±116 764±123
3.2±2.1 21.6±6.6
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
istration during anesthesia for a major surgical procedure
such as OLT. In selected cases with FHF, PDRdy monitorP138
ing allows one to obtain an accurate bedside evaluation of
graft function recovery.
Clinical and economic study of the use of somatostatin 14 AA (amino-acids) in the treatment of
enterocutaneous and pancreatic fistulas
FA Allaert*, F Yahchouchy†, O Sol‡, M Billardon‡ and PL Fagniez§
*Centre d’Evaluation et de Normalisation, CHU du Bocage, 21000 Dijon; †Chirurgie générale, Hôpital de Poissy, 78303 Poissy;
‡UCB Pharma, 21 Rue de Neuilly, 92003 Nanterre; §Hôpital Henri Mondor, 94010 Créteil, France
Purposes of the study: To describe the use of somatostatin 14 AA (amino-acids) by patients with post-surgical
enterocutaneous and/or pancreatic fistulas, its clinical
results, and the analysis of the economic benefits resulting
from such treatment.
Material and methods: Nature of the study: open
medico-economic observational study. Follow-up period of
the patients: from somatostatin 14 AA treatment initiation
to the closure of the fistula (fistula output ≤10 ml/24 h) or
to a maximum of 28 days. Criteria of selection: presence
of one or more entero-cutaneous and/or pancreatic postsurgical fistulas. Evaluation criteria: 1) closure rate and
time to closure of the fistulas; 2) hospital expenses.
Results: Eighty patients were included in the study. The
average age of the patients was 58.4±16.6 years (mean
±SD). 87.7% (n=64) of the patients had single fistula and
12.3% (n=9) had multiple fistulas. 37.0% (n=27) were
ileal, 17.8% (n=13) jejunal and 45.2% (n=33) pancreatic.
60.8% fistulas were closed after 15.2±7.6 days and
11.5±5.1 days of treatment. The percentage of closure is
comparable whatever the clinical characteristics (age, sex,
associated pathology) of the patients, the conditions of
the surgical operation (first intervention or re-intervention,
regular or emergency surgery) and the type of fistula (ASA
classification, morphology, topography). The percentage
of closure varies significantly according to the response at
48 h. Respectively, among the patients who have a
decrease of 50% and 70% of their fistula output at 48 h,
80% (n=32) and 91.7% (n=22) have an output less than
10 ml/24 h at the end of the study. The rate of closure of
the fistula does not differ from those described in the literature for treatments without somatostatin 14 AA but the
time of closure is halved from 31 days to 15.2 days. At
comparable therapeutic success rate, the use of somatostatin 14 AA provides greater comfort for the patient and
an important decrease in costs (30%) arising from the
reduction in the duration of the hospital stay.
Conclusion: The use of somatostatin 14 AA shortens the
time of closure of the enterocutaneous and pancreatic fistulas and allows an important reduction of the hospital
costs. The evaluation of the therapeutic response 48 h
after the beginning of the treatment is a fundamental criterion for optimising the use of this drug.
Decrease in urine output during mesenteric traction syndrome is an early predictor of multiorgan
P Garcia-Saura, A Marquez, C Serrano, J Prieto, A Maldonado, MS Serrano-Atero and F Peramo
Unidad de Reanimacion, Departamento de Anestesiologia, Reanimación y Terapia del Dolor, Hospital Clinico Universitario de
Granada, Spain
Introduction: Six retrospective case-control studies showed
that the only difference between those patients who underwent elective infrarenal aortic aneurysm repair (IAAR) developing mesenteric traction syndrome (MTS) in the operation
room (OR) and multiorgan dysfunction syndrome (MODS)
during their ICU stay, was the decrease in urine output
before aortic cross-clamping. Two of them presented the socalled ‘hepatohemorrhagic syndrome’ (HHS) [1] characterised by the association of disseminated intravascular
coagulation (DIC) and acute ischemic hepatitis (AIH).
Methods: Data from six patients with MTS was obtained
from the anaesthesia records (1995–1998) and clinical
evolution during their ICU stay.
Results: After the placement of the intestinal traction
device, all patients developed MTS (facial flushing,
reduced mean arterial pressure (MAP) and systemic
vascular resistance (SVR) with increased heart rate
(HR) and Cardiac Index (CI). They were treated with
volume replacement until they reached the previous
MAP before cross-clamping. Diuresis showed an
increase that remained elevated during the intervention
in all the patients but two. These two developed DIC
intra-operatively, with bleeding before unclamping and
posterior MODS. We believe that the initial release of
PGI2, as the main product triggering the MTS, is also
responsible for the increase in diuresis, but dysregulation of the renal function by an unknown mechanism [2]
in the context of the prostanoid disbalance, would
explain the two cases of DIC and posterior MODS in
the two patients which, after a little increase in urine
output, suddenly without any explanation turned to fall
cause-effect relationship between the presence, or not, of
MTS during the anaesthesia period, and the posterior
course in which urine output fell. Second, we also think
that this syndrome (at least the two HHS we have seen)
would be included in the context of MODS. Third, in this
setting, the kidneys are the first organs in dysfunction by
unknown mechanism, and this is due to the impaired renal
response to the disbalance of prostanoids.
DIURESIS (ml /h)
Conclusions: First, although it has been postulated that
HHS can not be explained by the operative course before
the release of the aortic cross-clamp, we found a close
Levy PJ, Tabares AH, Olin JW, et al: Disseminated intravascular
coagulation associated with acute ischemic hepatitis after elective aortic aneurysm repair: comparative analysis of 10 cases. J
Cardiothoracic Vascular Anaesthesia 1997, 2:141–148.
Gelman S: The pathophysiology of aortic cross-clamping and
unclamping. Anesthesiology 1995, 82(4).
Maldonado AJ, Costela JL, Serrano-Atero MS, et al: Novedades
anestésicas en angiología y cirugía vascular. In Novedades Farmacológicas en las Vasculopatias. Edited by Ros Die E. Barcelona: J
Uriach and Cia. SA, Barcelona.
Increased intestinal permeability accompanies the development of sepsis in ICU patients
M Hersch, I Skorochod, L Kanter, M Gurevitz and R Goldstein
Department of Anesthesiology Intensive care and Medicine, Shaare Zedek Medical Center, Jerusalem, Israel
Introduction: The concept of bacterial/toxin translocation
from the gut as the source of sepsis is well recognized.
Continuous monitoring of gut permeability in correlation
with clinical signs of sepsis is poorly documented. We
measured daily intestinal permeability of patients admitted
to the ICU and correlated it with signs of sepsis.
Methods: Design: prospective, observational study. Population: twelve consecutive patients admitted to the ICU
with an expected admission of three days or more. Three
patients were septic on admission and nine were not. Six
out of the nonseptic group developed signs of sepsis
during their ICU stay. Permeability measurement: each of
the 12 patients received 5 g of Mannitol and 10 g of Lactulose daily (from ICU admission up to seven days),
through their NG tube. Urine was collected every 24 h,
and an aliquot of 40 ml was kept at–20°C for subsequent
high pressure liquid chromatography, to measure the
secretion of Mannitol and Lactulose. The ratio of Lactulose/Mannitol (L/M) was calculated. Signs of sepsis
(according to Bone’s criteria: Ann Int Med, 1991) were
recorded daily.
Results: A total of nine septic patients showed a significant increase of L/M compared to nine non-septic patients
(21.2±0.3 vs 4.3±0.4, P<0.05). The six patients who
developed sepsis during their ICU stay showed a significant rise of L/M while becoming septic (28.7±0.2 during
sepsis, vs 4.8±0.3 when not septic, P<0.05).
Conclusions: Increased intestinal permeability is associated with clinical signs of sepsis, thus substantiating the
role of gut permeability as a major defect in the evolvement of sepsis and its sequelae.
Hypocapnic and normocapnic ventilation in hepatic encephalopathy: a comparision
AK Baronia, CK Pandey and N Mathur
SGPGIMS, Lucknow, 226 014, India
Purpose: Encephalopathy with raised intracranial pressure is the major cause of death in hepatic failure [1]. We
investigated role of arterial hypocapnia, achieved by hyperventilation, in the management of hepatic encephalopathy.
had PaCO2 in the range of 25–30 mmHg (group I) and
35–40 mmHg (group II). Tidal volume was kept in the
range of 8–10 ml per kg and respiratory rate in the
range of 12–20 per min.
Methods: In a prospective study of twenty-two patients,
with fulminant hepatic failure with grade IV
encephalopathy receiving mechanical ventilation, two
groups of twelve (group I) and ten (group II) patients
Results: The demographic data was comparable in both
the groups. The shortest duration of recovery was 8 h in
group I and 30 h in group II.. Survival was 20% group I
and 41.66% group II. No adverse effects, secondary to
hypocapnia, were noticed in group I.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Conclusion: Improved survival in the normocapnic group
may indicate that normocapnic ventilation may be preferred over hypocapnic ventilation, particularly if the
patients of hepatic encephalopathy are being managed
without cerebral perfusion pressure and ICP monitoring.
Ede RJ, Williams R: Semin Liver Dis 1986, 6:107–118.
Emergency department evaluation of orthotopic liver transplant recipients
G Couto, S Marum, J Oliveira, M Rui Silva and J Palmeiro Ribeiro
Unidade de Cuidados Intensivos do Hospital de Curry Cabral, Lisboa, Portugal
Introduction: The intensive care unit (ICU) has an important role in the early management of orthotopic liver transplantation (OLT) recipients.
Patients and methods: We reviewed the records of the
83 consecutive patients (91 procedures) who underwent
OLT from September 1992 to June 1999 in our centre,
collecting sample characteristics, reviewing the early complications after OLT and the changes in medical care of
these patients over the past seven years.
Results: Most frequent indications for primary OLT and
for retransplantation were familial amyloidotic polyneuropathy (37.3%) and hepatic artery thrombosis (75%),
respectively. The most challenging complications seen
were acute renal failure (31.3%), postoperative bleeding
(15.6%), infections (9.6%), acute cellular rejection
(9.6%) and warm ischemia (8.4%), with lethal primary
nonfunction ensuing it in three cases, the most common
cause of death in ICU. There was considerable reduction
in operative time (7.1±1.36 vs 5.8±0.83 h; P<0.001),
blood transfusions requirement (8.1±4.6 vs 4.9±2.7
units; P<0.01), ventilation time in the routine patient
(51±30 vs 30±30 h; P<0.05) and in ICU length of stay
(6.2±2.7 vs 3.9±2.5 days; P<0.002) between the two
time periods in which we equally divided our sample
(92–96 and 97–99) with a sensible reduction in ICU
morbidity (48% vs 24% of moderate-to-severe complications; P<0.02).
Discussion and conclusions: The reduction in procedure
and ICU invasiveness throughout the years allowed for a
reduction in morbidity and may confirm that appropriate
medical care can overcome the adverse influences of
many of the negative predictor factors suggested to influence OLT recipients outcome.
Timing and incidence density of upper gastrointestinal bleeding acquired by critically ill children
M Chaïbou, M Tucci, M A Dugas, CA Farrell and J Lacroix
Pediatric Intensive Care Unit, Department of Pediatrics, Sainte-Justine Hospital, Université de Montréal, Montréal, Québec, Canada
Objectives: To determine in critically ill children: 1) when
upper gastrointestinal bleeding (UGIB) and clinically significant UGIB (CS-UGIB) occur; 2) the mean incidence
density and the cumulative incidence of UGIB and CSUGIB.
Design: Prospective epidemiological study.
Site: Multidisciplinary pediatric intensive care unit (PICU)
in a tertiary care teaching hospital.
Methods: UGIB was considered to be present if
hematemesis occurred or if blood was present in the
gastric tube. UGIB was qualified as clinically significant if
2 out of 3 experts independently concluded that at least 1
of 6 complications (transfusion, decreased hemoglobin
concentration, hypotension, surgery, multiple organ
system failure or death) was attributable to it. The mean
time from PICU admission to onset of UGIB or CS-UGIB
was obtained by the summation of the number of events
multiplied by duration of PICU stay for each event, then
divided by the total number of events. The mean incidence
density was calculated by dividing the total number of
patients with UGIB or CS-UGIB by the cumulative number
of patient-days.
Results: The cohort included 1006 consecutive patients.
103 UGIB (10.2%) were diagnosed, including 16 CSUGIB (1.6%). The mean time from PICU admission to
UGIB and CS-UGIB onset was 4.5 days (95% CI:
1.4–7.6) and 3.9 days (95% CI: 1.4–6.4) respectively.
The mean incidence density was 24.6/10 000 patientdays for UGIB and 3.8/10 000 patient-days for CS-UGIB.
The estimation of the risk of bleeding by actuarial method
indicates that the probability of occurrence of UGIB and
CS-UGIB was highest on the second day of PICU stay:
4.7% and 1.2% respectively.
Conclusion: UGIB and CS-UGIB occur soon after admission to PICU. Prophylaxis to prevent CS-UGIB should be
given early to patients at risk of developing this complication.
Does metoclopramide prevent bacterial translocation (BT)?
F Cansizoglu, M Tulunay, M Oral, Z Alanoglu, A Karaaslan, D Gerceker and I Palabiyikoglu
Ankara University Medical Faculty, Anesthesiology and Reanimation Department, Sihhiye, Ankara, Turkey
One of the mechanisms causing BT is impaired host
defense. Metoclopramide (M) restores the depressed
immune function after hemorrhage [1]. To our knowledge,
there is no study investigating the effect of M on BT.
Procedures: We investigated the administration of M on
BT in a dog model of ischemia-reperfusion (IR) injury
induced by thoracic aortic cross-clamping and declamping. Twenty-two mongrel dogs were randomized into three
groups: sham-operated group (n=7) (without cross-clamping), M group (n=8) and placebo group (n=7). Placebo
and M group received placebo and M (0.15 mg/kg iv),
respectively, before cross-clamping. During 45 min of
ischemia and 30 min reperfusion arterial blood gases and
hemodynamic data were continuously recorded. 72 h later
dogs were re-operated on, and peritoneal swabs, blood
samples and specimens from duodenum, jejunum, ileum,
colon, mesenteric lymph node (MNL), heart, spleen,
kidney, liver and lung were obtained for bacteriological
analysis. Blood samples were also obtained before the
first operation.
Results: No dogs in the sham group showed BT. Two
ischemia-reperfusion groups were found to have BT in
various organs; however, M prevented BT to the MLN (χ2)
without producing any hemodynamic changes (MannWhitney U-test, Kruskal Wallis one way ANOVA, Friedman
two way ANOVA) organs (χ2). Both gram (–) and gram (+)
microorganisms translocated. One dog in the placebo
group and two dogs in M group developed bacteremia
after the first operation.
Conclusion: M prevents BT to the MLN. Whether higher
doses of M prevent BT and the clinical importance of
these results requires further studies.
Zellweger R, Wichmann MW, Ayala A, Chaudry IH: Metoclopramide:
a novel and safe immunomodulating agent for restoring the
depressed macrophage immune function after hemorrhage. J
Trauma 1998, 44:70–77.
Is redistribution of microcirculatory blood flow within the small intestinal wall the cause of
prolonged paralytic ileus in the critically ill?
L Hiltebrand, V Krejci, D Erni*, A Banic* and GH Sigurdsson
Autoregulation of microcirculatory blood flow (MBF) and
redistribution of flow within the intestinal wall (from the
muscularis to the mucosa) are known mechanisms to
maintain O2 delivery in the gut during hypovolemia. During
sepsis, however, autoregulation of blood flow is impaired
[1]. The aim of this study was to monitor the distribution of
MBF within the intestinal wall in different parts of the gastrointestinal tract during the development of septic shock.
Materials and methods: We measured systemic (CI),
regional (mesenteric artery; SMA) and microcirculatory
blood flow (MBF) in 11 sedated and ventilated pigs. MBF
was measured with multichannel laser Doppler flowmetry
in the gastric, jejunal, and colonic mucosa and the corresponding muscularis layers. Septic shock was induced by
faecal peritonitis. After 240 min, i.v. fluids were administered to alter hypodynamic shock to hyperdynamic septic
Results and discussion: During the first 240 min (hypodynamic shock) CI, SMA and MBF in the stomach mucosa
decreased by 50% (Figs 1 and 2), while MBF in the
jejunal and colonic mucosa remained virtually unchanged
(Figs 3 and 4). In the muscularis of the jejunum and colon,
on the other hand, MBF decreased significantly more than
CI and SMA (Figs 2 and 4). Administration of i.v. fluids at
240 min resulted in a significant increase (above baseline)
in CI and SMA (Fig. 1) as well as in MBF in the mucosa of
the stomach, jejunum and colon. The ‘overshooting’
increase in MBF in the jejunal and colonic mucosa, indicates that there was some hypoperfusion present in the
mucosa during the hypodynamic phase despite maintained blood flow. This is supported by a significant
decrease in jejunal pHi (Fig. 5) during the same time
period. There was virtually no increase in MBF in the
jejunal and colonic muscularis after fluid administration
indicating a prolonged, perhaps endothelin induced,
Conclusion: It was concluded that MBF in the mucosa of
the jejunum and colon remained unchanged, despite a 50%
decrease in systemic and regional flows, suggesting a
largely intact autoregulation during septic shock. MBF in the
jejunal and colonic muscularis decreased significantly more
than systemic and regional flows, suggesting that there is
an active mechanism to maintain adequate blood flow to the
mucosa during septic shock. Redistribution of blood flow
from the muscularis to the mucosa causes severe hypoperfusion of the muscularis which contributes to intestinal atony
frequently observed in critically ill patients.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Figures 1–5
% of baseline
% of baseline
MBF Colon
*** ***
300 min
% of baseline
% of baseline
MBF stomach
300 min
300 min
MBF Jejunum
pHi Jejunum
300 min
300 min
Onset of peritonitis at 0 min and i.v. fluid administration at 240 min. Data is presented as mean ± SEM. Statistics: ANOVA for repeated
mesurements. P < 0.05 was considered statistically significant.
Ileal microcirculation and mucosal acidosis during hyperdynamic porcine endotoxemia
I Tugtekin, M Matejovic*, A Stehr†, M Theisen, F Ploner‡, K Träger, M Georgieff and P Radermacher
Department of Anesthesia, Univ. Hospital, D-89073 Ulm, *Department of Intensive Care, Charles-University, CZ-30460 Plzen;
†Department of Surgery, Univ. Hospital, D-93053 Regensburg; ‡Department of Anesthesia, District Hospital, I-39042 Brixen
Introduction: The tonometric determination of the arterialmucosal PCO2-gap (∆a-rPCO2) is used to monitor adequacy of gastrointestinal perfusion. As PrCO2 also integrates
other phenomena [1], we analyzed the influence of the
intestinal villus microcirculation on increased ∆a-rPCO2
during long-term hyperdynamic porcine endotoxemia.
Material and methods: Anesthetized and ventilated pigs
received continuous i.v. endotoxin (ETX, n=12) for 24 h or
placebo (Sham, n=6). Hydroxyethylstarch was infused to
maintain MAP >65 mmHg together with a sustained
increase in cardiac output [2]. Before the start of ETX (0
h), as well as 12 and 24 h afterwards, portal venous blood
flow (Qpv, ultrasound flow probes) and lactate/pyruvateratios (L/P pv), ileal mucosal ∆a-rPCO2 (fiberoptic sensor)
and bowel wall capillary Hb-O2-saturation (Hb-O2cap,
remission spectrophotometry) were assessed together
with intravital video records of the ileal mucosal microcir-
0 h ETX
12 h ETXk
24 h ETX
Qpv ml/kgxmin
Hb-O2cap % (±S.D.)
∆a-rPCO2 mmHg
L/P pv
n (Villi) perfused/not perfused
32(23;38)/22(10;35) #§
culation (number of perfused/not-perfused villi) by orthogonal polarization-spectrometry (Cytoscan™) [3] obtained
via an ileostomy.
sion. We can only speculate about the putative additional
influence of a disturbed intracellular O2-utilisation.
Results: See Table. Median(25/75%); #P<0.05 vs 0 h
(Friedman-Anova); §P<0.05 ETX vs Sham (Mann-Whitney).
Acknowledgement: Supported by *ESICM, Deutsche Forschungsgemeinschaft and ‡Provinz Bozen-Südtirol (Italy). Cytoscan™ is a trademark of
Cytometrics Inc., Philadelphia, PA.
Conclusion: Taking into account the unchanged portal
venous blood flow, the progressive increase in ∆a-rPCO2
is mainly due to the heterogeneity of capillary villus perfu-
Brinkmann et al.: Intensive Care Med 1998, 24:542.
Träger et al.: AJRCCM 1999, 159:1758.
Groner et al.: Nat Med 1999, 10:1209.
Relationship between systemic arterial pressure (SAP) and EtCO2 in patients with hemorrhagic
Š Grmec*, Z Zabavnik†, M Batišta†, P Klemen* and Š Mally*
*EMS-PHU Maribor; †Education Hospital Maribor; Ljubljanska 5, 2000 MB, Slovenia
Introduction: Capnometry may be useful in monitoring
critical ED patients for early signs of shock or pulmonary
embolus. However, EtCO2 levels must be interpreted in
the context of other parameters giving information about
the patient’s clinical status. A fall in EtCO2 mav indicate
decreased lung perfusion. If ventilation has not changed,
the clinician should be alerted to reevaluate the patient for
early signs of shock. The aim of this study was to examine
the relationship between the changes in systemic arterial
pressure (SAP) and in EtCO2 in hemodynamic status
usually encountered during hemorrhagic shock.
Methods: Data were collected prospectively in EMS-PHU
from February 1998 until August 1999. The study population was composed of adult (at least 18 years of age; mean
age 47.8±13.4; rang 18–76 years) 68 patients (25 female,
43 male) with 3rd or 4th stage of hemorrhagic shock
(trauma, GI or respiratory hemorrhage). We measured SAP
and EtCO2 for 1 min first 5 min, and then each 5 min). For
each patient we determined MEES 1, MEES 2, APACHE II,
∆MEES. SAP and EtCO2 were measured with PROPAQ
encore, Vital Sings Monitor, Protocol Systems, INC, Beaverton, OR, USA. We measured for each patient ∆EtCO2 and
∆SAP in interval from prehospital setting to hospitalization
(674 simultaneous measurements were recorded). All
EtCO2 and SAP values are reported in mmHg.
Results: The mean and standard deviation of the ∆EtCO2
was 12.52±6.36 and of the ∆ was 39.95±18.23. Results
are presented in the Figure. Statistics on the 68 individual
∆SAP 0 0.18 + 3.63×∆EtCO2; r=0.89; n=674.
correlations indicate ∆SAP=0.06 (±0.18) + 3.97 (±0.63)×
∆EtCO2 with r=0.91±0.08.
Conclusion: A linear relationship does exist between the
changes in SAP and in EtCO2 in patients with hemorrhagic
shock. A 1mmHg change in EtCO2 is associated with ~4
mmHg change in SAP. It suggests that small changes in
EtCO2 may be useful sign of significant changes in SAP
(cardiac output). We conclude that the measurement of
EtCO2 during hemorrhagic shock may be useful as a indicator of SAP and successful volume resuscitation.
Continuous monitoring of gastric carbon dioxide with optical fibers: European project
COMOCADOF, in vitro evaluation
A Lueger, D Scherr, G Brunner, W Trettnak*, A Mencaglia† and F Baldini‡
Department of Int. Med., Karl-Franzens-University, Auenbruggerplatz 15, 8036 Graz, Austria; *Joanneum Research, 8010 Graz,
Austria; †Prodotec SRL, 50127 Firenze, Italy; ‡IROE-CNR, 50127 Firenze, Italy
Introduction: Gastric mucosal PCO2 is an important parameter to evaluate the splanchnic perfusion and to estimate
the prognosis of critically ill patients. Saline and gas
tonometry are the existing measurement methods, with air
tonometry being the golden standard. The main disadvantage of this method is the relatively long equilibrium time,
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
which only allows intermittent registrations and therefore
cannot reflect rapid changes of PrCO2.The target of the
COMOCADOF project was the development of an optical
fiber sensor capable of a continuous monitoring of PCO2
inside the stomach, with high accuracy and long-term stability, able to track rapid PrCO2 changes.
Objectives: Evaluation of the feasibility of a measurement
device using an optical fiber sensor for continuous measurement of gastric PCO2.
Methods: The COMOCADOF system consists of a
fiberoptical catheter and the measurement unit. A CO2sensitive polymer layer is inserted into a probe head,
which is fixed at the end of the optical fiber. The in vitro
tests were performed in a glass test-tube containing 20 ml
of sample solution, which was maintained at a temperature
of 37°C. The sample solution (either physiological saline
or a physiologically adjusted hydrochloric acid of pH 1.0)
was flushed with 6 gas mixtures containing 0.0 to 15.0%
CO2 for defined time intervals and at a gas flow rate of 50
ml/min. The sampling rate was typically 0.8 s. The tests
were performed over a period of up to 48 h. In addition,
the influence of gastric juice, colored or turbid samples
(such as e.g. enteral nutrition) on the stability of the optochemical sensor was tested.
Results: The table shows the original minimum requirements (target at the beginning of the project) and the
Measurement range (mmHg)
Resolution at 0 mmHg (mmHg)
Resolution at 120 mmHg (mmHg)
Response time (T90) (min)
Accuracy in range 40–105 mmHg (mmHg)
Accuracy in range 0–40 mmHg (mmHg)
Measurement period within accuracy (h)
*=PCO2 >105 mmHg not investigated, but possible.
achievements obtained up to now with the optical sensor
for the detection of gastric carbon dioxide.
No significant influence of gastric juice, physiologically
adjusted hydrochloric acid of pH 1.0, colored or turbid
samples (such as e.g. enteral nutrition) was seen.
Conclusion The COMOCADOF System allows direct
continuous measurement of CO2 in gastric juice. The main
targets of the project (response time, long-term stability
and accuracy) were achieved. The system seems to
perform better than the existing air tonometry method.
Comparative studies with an air tonometry system both in
vitro and in vivo are in progress.
Assessment of a handy-type blood gas analyzer for measuring intramucosal pH (pHi)
H Dohgomori*, K Arikawa*, H Takamatsu* and Y Kanmura†
*Division of Emergency Medicine, †Department of Anesthesiology and Critical Care Medicine, Kagoshima University Hospital,
Sakuragaoka, Kagoshima, 890-8520, Japan
Introduction: pHi is important in the evaluation of critically ill
patients. When pHi data is needed quickly, a handy-type
blood gas analyzer may be useful but it must be accurate.
We therefore assessed the accuracy of a new handy-type
blood gas analyzer (Opti, AVL, Switzerland; hereafter called
V-O) in measuring pHi by comparing it with three other analyzers: (1) ABL520 (B-5), (2) ABL300 (B-3), (both from
Radiometer, Holland), and (3) Compact 2 (AVL; V-C).
Methods: We infused normal saline into the balloon of a
gastrointestinal tonometer (Instrumentarium Corp, Finland)
located in the patient’s stomach. After a 60 min equilibration time, the saline was withdrawn and the CO2 (PrCO2)
measured by the four gas analyzers (the analysis of each
sample of normal saline by the four analyzers were completed within 5 min). At the same time, we withdrew and
analyzed arterial blood. Then, pHi was calculated from the
Henderson-Hasselbalch equation, and the CO2 gap
(PrCO2–PaCO2) was also calculated.
Results: Twenty-five pHi and CO2 gap values were obtained
for each machine. The CO2 gaps for the various machines
were as follows: B-5: +4.5±2.6 (mean ±SD: mmHg); B-3:
+1.5±3.1; V-C: +2.9±2.3; and V-O: –13.2±4.1. The calculated pHi values were 7.336±0.056 (mean ±SD),
7.332±0.051; 7.330±0.051, and 7.542±0.089, respectively. With respect to the (B-5) results for CO2 gap, the
values for correlation coefficient (r), precision (P) and bias
(B) were as follows: (B-3) r, 0.900; P, 1.3; B, -3.1; (V-C) r,
0.911; P, 1.4; B, -1.5; and (V-O) r, 0.790; P, 2.9; B, -18.3. A
statistically significant relationship with B-5 was observed for
each of these three groups, but the discrepancy between VO and B-5 was large.
Discussion: We can now use blood gas analyzers at the
bedside and get results quickly. However, the value of pHi
obtained differs according to the kind of analyzer used, a
point that needs to be kept in mind. Our important finding
is that the data obtained with the handy analyzer tested
here differed quite considerably from those obtained using
the other three. We conclude that although the new handy
analyzer is convenient, it does not allow accurate calculations of pHi and CO2 gap.
Gastric mucosal tonometry: should we adjust PiCO2 and CO2-gap to temperature?
H Vogelsang, T Uhlig, C Seifert and P Schmucker
Department of Anaesthesiology, Medical University of Luebeck, Ratzeburger Allee 160, D-23538 Luebeck, Germany
Introduction: It has been under discussion for many
years, whether acid-base-equilibrium during hypothermia
should be managed using the alpha-stat or the pH-stat
concept. Several studies in cardiac patients undergoing
hypothermic cardiopulmonary bypass (CPB) led to controversial results. Until now there are only few data available,
on whether to adjust PiCO2 and CO2-gap to temperature
or not. Thus we calculated PiCO2 and CO2-gap in cardiac
surgery patients with and without temperature correction
looking for differences in the prediction of postoperative
Methods: After IRB approval, we studied 69 patients
undergoing elective aortocoronary bypass surgery, ASA
class II–III. In addition to standard monitoring each patient
received a pulmonary artery catheter and an air-filled nasogastric tonometry catheter connected to a Tonocap. Documentation of gastric-intramucosal PCO2 (PiCO2) followed
at 15-min-intervals, documentation of hemodynamics, arterial and mixed-venous blood gas analyses and lactate followed seven times until extubation (T1–T7). We used mild
hypothermic CPB and the alpha-stat concept. PaCO2 and
PiCO2 were adjusted to temperature by the following algoP151
rithm: PCO2t=PCO2×10 0.0185×(t-37) [mmHg] (t=body temperature; PCO2t=temperature adjusted PCO2).
Statistical analyses were done using the paired t-test,
P<0.05 was regarded as significant.
Results: There was a significant difference of about
3–4mmHg in temperature corrected PiCO2 and PaCO2
compared to uncorrected values. The CO2-gap ([PiPa]CO2) differed significantly at T3, T6 and T7 in all
patients and at T3 and T4 in patients with postoperative
cardiopulmonary resuscitation (n=5). However, the time
course of measurements was equal.
Conclusion: For clinical applications, it is not necessary to
adjust PiCO2 and CO2-gap to body temperature during
mild hypothermic CPB with alpha-stat concept since no
further information could be gained from the correction. In
laboratory settings or deep hypothermic CPB with pH-stat
concept temperature adjustment of tonometric parameters
may be helpful. In addition, the correction may also be
useful in hyperthermia (septic patients). However, this has
to be investigated in further studies.
Evaluation of PCO2 measurements using a gastric air tonometer located in non-mixing solution
Y Sato, H Inoue, K Ikegami, T Seki, F Morimoto, Y Nagai*, T Sekiguchi* and M Nakamura*
Trauma and Critical Care Center, Koshigaya Hospital, Dokkyo University School of Medicine, 2-1-50 Minami-Koshigaya,
Koshigaya, Saitama, Japan; *R and D Center, Nihon Kohden Corp., Tokyo, Japan
Introduction: Many studies reported that measurements
of gastric PCO2 with the air tonometer method were more
precise than those obtained with saline tonometer. Gastric
tonometers, however, were usually validated in vitro in a
solution bubbled with CO2. In this study, we evaluated the
accuracy and equilibration characteristics of PCO2 in the
balloon of a gastric air tonometer located in saline solution
either with or without bubbling of CO2.
Methods: A gastric tonometer was immersed in a 0.9%
saline solution maintained at 37ºC, in which certified calibration gases at three different CO2 concentrations (5%,
9%, and 12%) were bubbled. When solution was saturated with CO2 gas, PCO2 in the balloon was measured
every 10 min for 30 min by an infra-red analyzer with or
without CO2 bubbling. Bias and precision were calculated
from the measured and expected PCO2 values.
Results: Under CO2 bubbling, relatively small bias
(–4%~9%) was seen at all three CO2 concentrations. On
the contrary, large negative bias (26%–39%) was
observed without bubbling.
Conclusion: It has been demonstrated that quick equilibration of PCO2 between inside and outside of the
tonometer balloon needs the mixing of sample solution.
This result strongly suggests that measurements of
PCO2 with the air tonometer would underestimate
gastric PCO2 when it is located in the gastric juice with
little mixing.
Influence of prone position on gastric mucosal–arterial PCO2 gradient
P Kiefer, A Morin, H Wiedeck, M Georgieff and P Radermacher
Universitätsklinik f. Anästhesiologie, Universität, D-89070 Ulm, Germany
Introduction: In patients with acute lung injury (ALI), the
prone position is a well-established method to improve
gas exchange by reopening atelectasis and secret
drainage. This approach might lead, however, to increased
intra-abdominal pressure and thereby impaired gastric
mucosal perfusion. Therefore, we studied the effect of the
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
prone position on both intragastric pressure and gastric
mucosal–arterial PCO2 gradients in patients with ALI.
Patients/methods: So far, 10 patients with ALI
(PaO2/FiO2 <250 mmHg) and clinical indication for
mechanical ventilation in the prone position were studied.
In addition to ventilator settings, systemic hemodynamics
and gas exchange, we semi-continuously measured the
gastric mucosal PCO2 (PCO2gm) (Tonocap) via a nasogastric tube as well as the intragastric pressure continuously via one lumen of the nasogastric tube using a
common pressure transducer. After 60 min of stable conditions baseline measurements were made. Then the
patient was turned to prone position and further measurements were obtained 60 and 120 min later.
Results/statistics: See Table. Friedmann test/StudentNewman-Keuls: #P<0.05 vs baseline.
Conclusion: Despite major individual variability, we could
not find a significant overall change in gastric
mucosal–arterial PCO2 gradient after turning to prone
60 min
120 min
Ci l/min/m2
MAP mmHg
CVP mmHg
PCO2gm mmHg
PCO2art mmHg
Gastric pressure mmHg
PCO2gm–PCO2art mmHg
position. Increased intragastric pressure (>5 mmHg),
however, was associated with an increased gastric
mucosal–arterial PCO2 gradient despite stable systemic
hemodynamics. Positioning induced variations of intraabdominal pressure, hence, may impair gastric mucosal
energy balance possibly due to compromised regional
blood flow.
Systemic and splanchnic hemodynamics, metabolism and PCO2-gap during septic shock induced
by live E. coli infusion in dogs
C Lagoa, LFP Figueiredo, R Cruz Jr, E Silva and M Rocha e Silva
Heart Institute (InCor), University of São Paulo Medical School, Brazil. Av. Dr. Enéas de Carvalho Aguiar, 44 CEP 05403-000
São Paulo–SP, Brazil
Background: The effects of intravenous live bacteria infusion on the time course of splanchnic oxygenation variables
have not been adequately defined, as well as the effects of
large volume crystalloid infusion on these variables.
Methods: Twenty-seven anesthetized mongrel dogs
(17.1±1.77 kg) were challenged by a 15 min intravenous
infusion of live E. coli (6×109 CFU/kg) and followed for 90
min. The animals were then randomized in two groups over
60 min: CT (controls, n=13), no fluid infusion or LR (lactated Ringer’s 32 ml/kg/h, n=14). Cardiac index (CI in
l/min/m2), mean arterial pressure (MAP in mmHg), mesenteric blood flow (MBF in ml/min, ultrasonic flowprobe),
oxygen-derived variables, lactate levels (in mmol/l) and
gastric PCO2 (gastric tonometry in mmHg) were assessed
throughout the 165 min experimental protocol. PCO2-gap
was defined as the difference between gastric and arterial
PCO2 and lactate flux was defined by the standard formula:
([portal lactate levels–arterial lactate levels]×MBF).
Results: Data are presented as mean ±SE. E.coli infusion
significantly reduced MAP, CI, MBF, and increased PCO2gap in both groups. Fluid infusion increased CI, systemic
DO2 and stabilized PCO2-gap. Although MBF was similar
in both groups, only controls presented gut lactate pro-
Table 1
Time course of MAP, CI, MBF, Lactate Flux and PCO2-gap at baseline, 15, 105 and 165 min.
Lact. flux
15 min
105 min
165 min
- 30.6±39
duction. Table 1 shows the time course of MAP, CI, MBF,
lactate flux and PCO2-gap of the two groups.
it seems to prevent gut lactate production in a canine
model of experimental septic shock.
Conclusion: Although large-volume crystalloid infusion
fails to restore mesenteric and gastric mucosal perfusion,
Acknowledgement: This study was suported by FAPESP, São Paulo,
Brazil. Grants # 98/06458-7 and # 98/06459-3.
Routine blood gas analysis does not provide information on regional metabolism of the stomach
obtained by gastric tonometry in patients with left ventricular failure
J Graf, E Karassimos and U Janssens
Medical Clinic I, RWTH Aachen, Pauwelsstr. 30, 52057 Aachen, Germany
Purpose of the study: To define the association between
variables reflecting regional perfusion and metabolism of
the stomach obtained by gastric air tonometry with routine
blood gas analysis and global hemodynamics in patients
with cardiac failure.
Methods: In 50 mechanically-ventilated patients with
severe left ventricular failure gastric intramucosal PCO2
(PgCO2) was assessed utilising the Tonocap. Simultaneously hemodynamics were measured and routine blood
gas analysis was carried out. Independent variables were
correlated and the discriminative power to predict a low
pHi <7.32 was calculated using Receiver Operating Characteristics (ROC).
Results: 564 paired measurements of cardiorespiratory
and gastric tonometry variables were collected. Neither
global hemodynamics nor arterial lactate values correlated
with one of the tonometric parameters or patients’
outcome at any timepoint. Correlation of arterial PCO2,
arterial pH, arterial bicarbonate, and base excess with
PgCO2 was only weak. According to the areas under
ROC prediction of pHi <7.32 was excellent for PgCO2
and CO2 Gap (difference of PgCO2 and arterial PCO2)
(areas under ROC curve 0.90±0.01 and 0.92±0.01, cutoff values 51 mmHg and 13 mmHg, respectively).
Conclusion: Both PgCO2 and CO2 Gap are reliable predictors of low pHi, thought to reflect gastric mucosal hypoperfusion. None of the cardiorespiratory parameters were
capable of predicting a pHi <7.32. Thus, in patients with
severe left ventricular failure assessment of regional perfusion indices using gastric tonometry cannot be replaced by
one of the tested routinely applied monitoring modalities.
Gastric-arterial pCO2 gradient, but not lactate levels, is related to multiple organ dysfunction
assessed by SOFA score in septic patients
E Silva, M Kai, M Assumpção, PC Martins, AG Garrido, S Blecher and E Knobel
Albert Einstein Hospital, Av. Albert Einstein, 627 5th Floor, São Paulo, 05651-90, Brazil
Introduction: The gut hypoperfusion could contribute to
development of multiple organ dysfunction in septic
patients. However, there is no definitive study correlating
the time course of gastric mucosal PCO2 and organ dysfunction.
Methods: We have studied prospectively 27 adult severe
septic patients consecutively admitted in two large mixed
ICUs. Each patient had a nasogastric tonometer and a pulmonary artery catheter. Every 8 h, systemic hemodynamic
and oxygen variables, plasma lactate levels and gastric
mucosal pCO2 (PgCO2) were measured. All these variables were measured on inclusion day (Day 0) and on the
1st, 2nd and 3rd days. pCO2-gap was calculated as the
difference between PgCO2 and PaCO2. Daily, we measured the SOFA score to characterize organ dysfunction
evolution. We used the median pCO2-gap and lactate
values of each day and time course of these variables to
correlate to organ dysfunction and outcome (mortality on
day 28). Changes over time were analyzed using a
Kruskal-Wallis test and the relative risk (RR) was calculated.
Results: The median age was 55 years and median
APACHE II score was 18.5. The overall mortality rate was
52%. Table 1 shows the RR and CI 95% of each variables
on day 0, 1, 2 and 3. The best cutoff values of pCO2-gap,
lactate values and SOFA score were 15 mmHg, 2 mEq/l
and 11, respectively. On the 0, 1st, 2nd and 3rd days, the
patients that had pCO2-gap values of more than
14 mmHg, either the SOFA score remained high or the
Table 1
RR and CI 95% of each variable to estimate outcome on 28th
Day 0
Day 1
Day 2
Day 3
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
patients died on the 10th day (P<0.05). In contrast, on the
same days, lactate levels did not discriminate either organ
dysfunction development or death on the 10th day.
Conclusions: There was a good correlation between pCO2gap and multiple organ dysfunction assessed by SOFA
score. Lactate levels were not able to predict outcome.
The relationship between data of gastric tonometry by determination of polymorphonuclear
leukocytes (PMNs, i.e., myeloperoxidase activity) and gut mucosal hypoperfusion in elective
cardiac surgical patients
A Pietsch, T Uhlig, H Vogelsang and P Schmucker
Department of Anesthesiology, Medical University of Luebeck, Ratzeburger Allee 160, 23538 Luebeck, Germany
Introduction: The expression of polymorphonuclear leukocytes (PMNs) as a result of local and systemic inflammation in patients undergoing cardiac surgery by using the
CPB is considered to be an important cofactor in tissue
hypoxia and the pathogenesis of multiple organ failure [1].
Therefore, an early detection of tissue hypoxia is very
useful for the prevention of MOFS. It has been shown that
gastric tonometry improves clinical applicability of gut
mucosal oxygenation monitoring [2]. In a previous paper
the correlation between gastric tonometry and PMNs was
warranted [3]. This was done in the following study.
Methods: Following institutional approval, 66 adult male
patients undergoing elective cardiac surgery using cardiopulmonary bypass were studied. They were monitored
with the Tonocap™ in addition to standard monitoring
(ECG, CVP, arterial pressure) and the pulmonary artery
catheter. The PiCO2 was measured every 15 min after
admission to the ICU. In addition, documentation of CO2gap, arterial and pulmonary arterial hemodynamics, arterial
and mixed-venous blood gas analyses and lactate followed four times until extubation. The patients were
shifted into three groups by using the terciles of preopera-
tive expression of the PMNs (group I <1451 µg/l; group II
≤2487 µg/l; group III >2487 µg/l).
Results: The course of CO2-gap was parallel to the
PMNs. Both parameters increased after admission to the
ICU. Using the Wilcoxon-Wilcox test the increase in CO2gap and PMNs was significant in all groups. These differences did not occur in the parameters obtained from
blood gas analyses (BE, lactate, mixed venous saturation)
and in hemodynamic parameters.
Conclusions: Gastric tonometry is superior to other techniques in detecting dysfunctions of gut oxygenation if
PMNs are used as a biochemical indicator. Questions of
therapeutic interventions in increasing CO2-gaps and
PMNs need further investigations.
J Thorac Cardiovasc Surg 1983, 86:845–857.
Intensive Care Medicine 1997, 23:276–281.
Critical Care Medicine 1999, 27:2038–2039.
Detection of perfusion failure during open heart surgery with sublingual PCO2
H Inoue, I Sato, N Hitosugi, Y Takano, M Miyake, Y Sato, K Ikegami, Y Nagai*, T Sekiguchi* and M Nakamura*
Koshigaya Hospital Dokkyo University School of Medicine, 2-1-50, Minami-Koshigaya, Koshigaya, Saitama, Japan; *Nihon
Kohden Corp, Tokyo, Japan
Background: Gastric tonometry has been used to detect
intramucosal acidosis. However, there remains controversies about the methodology and interpretation of the
results. In recent years, sublingual PCO2 (PSLCO2) has
been introduced as a noninvasive method to detect perfusion failure. Yet, there has not been enough evidence to
prove that the PSLCO2 is clinically of use.
Purposes: To define the efficacy of PSLCO2.
Methods: We investigated 6 patients who underwent
open-heart surgery. PSLCO2 were recorded continuously
with a PCO2 sensor based on an ion-sensitive field effect
transistor (ISFET, NIHON KOHDEN). Arterial lactate was
also measured.
Result: A substantial increase of PSLCO2 came into our
notice from 377 mmHg, at the time of induction to 5712
mmHg, during the cardiopulmonary bypass (CPB).
Gradual decrease of PSLCO2 after the aorta declamping
was observed down to 485 mmHg, 6 h after the operation. Significant increase in PSLCO2 indicates that oxidization is not fairly sustained to the tissue during CPB
despite enough pump flow. The increase in blood lactate
concentration was followed parallel to PSLCO2 while
CPB was performed.
Conclusion: PSLCO2 is an attractive parameter to be
served for the swift detection of perfusion failure during
open-heart surgery.
Is muscular mass affected by L-Carnitine levels in critically ill patients?
M Moukas, P Dimou, A Chalazonitis, A Patsalides, G Tsimitselis, S Georgakopoulou, N Koudounis, M Vasiliou and P Behrakis
Intensive Care Unit and Department of Radiology, Hellenic Red Cross Hospital, 1 Erythrou Stavrou str. Athens 11526, Greece
Introduction: The aim of our study is to investigate the
change of blood and muscle Carnitine levels in patients of
the Intensive Care Unit and investigate how the change in
Carnitine levels affects muscular mass.
Method: In our double blind clinical study, forty-six ICU
patients were randomly divided into two groups. Patients
with either renal or liver insufficiency were excluded. Both
groups of patients were on enteral or parenteral nutrition
with 30 kcal/kg/day intake with 1 g/kg/day protein intake.
The mean duration of hospitalization was 23.6±8 days for
group A and 26.9±9 days for group B. All patients were
under physical therapy for 60 min/day. Only group B
patients received intravenous supplements containing
100 mg/kg/day L-Carnitine. Total and free Carnitine were
measured from the blood of each patient at the time of
admission and at the time of exit from the ICU. Muscle
Carnitine levels were also studied with a triceps muscle
biopsy at the same time. Muscle mass thickness was estimated by the same utrasonographer in all cases. A 10
MHz frequency, linear high-resolution transducer was held
longitudinal to the biceps muscle in the plane of maximum
Results: All values are expressed as mean ±SD on admission and exit from the ICU. Blood results for Group A
patients are: total Carnitine (nmol/ml) 78.14±23.17 and
50.56±21.23 (P<0.001) and free Carnitine (nmol/ml) was
69.41±23.77 and 42.34±18.00 (P<0.001) respectively.
Group B has the following results: total Carnitine
(nmol/ml) 74.08±24.51 and 529.92±228.45 (P<0.001)
and free Carnitine (nmol/ml) 65.59±25.07 and
459.43±162.12 (P<0.001).
Muscle biopsy results are as follows: Group A patients
had free Carnitine 1.50±0.33 on admission and
1.06±0.19 on exit (P<0.0001), and total Carnitine
1.70±0.29 and 1.10±0.21 respectively (P<0.0001). In
Group B patients, free Carnitine was 1.53±0.31 on
admission and 1.99±0.15 on exit (P<0.0001), and total
Carnitine 1.67±0.29 and 2.23±0.16 respectively.
Muscle mass (cm) measurement results for group A
patients are: 2.6±0.5 and 2.3±0.6 (P<0.01) whereas for
group B patients they are 2.09±0.5 and 2.31±0.1
(P<0.001) respectively.
Conclusions: In group A patients, blood and muscle Carnitine are decreased. Muscle mass is concomitantly
decreased. The opposite results are obtained from group
B (L-Carnitine supplementation). We conclude that L-Carnitine supplementation prevents muscle atrophy in critically ill ICU patients.
Pseudocholinesterase and albumin in infants with cirrhosis
LJ Popoviê, A Antabak, B Æupanëiê and V Vegar
Surgical Clinic and Department of Anaesthesiology, Clinical Hospital Centre Rebro, Kispaticeva 12, 10000 Zagreb, Croatia
Objectives: Pseudocholinesterase (PChE) is an enzyme
(EC and like albumin is synthesised in the liver.
Although the two proteins are synthesised in the liver they
are not interdependent and both have been used independently as an index of liver function. Because of the contrary literature data, the aim of the present study was to
measure PChE activity and albumin concentration in
healthy infants and compare them with adult values [1,2].
We also compared PChE and albumin values between
healthy infants and infants with biliary atresia and cirrhosis.
PChE activity and albumin concentration were measured
in 25 infants with biliary atresia and accompanying liver
cirrhosis before Kasai operation, in 10 healthy infants, as
well as in 10 healthy adults undergoing elective surgery.
Methods: PChE activity was determined by the spectrophotometric method of Ellman, using butyrylthiocholine
as substrate. Plasma albumin concentration was measured by electrophoresis.
Results: In infants with liver cirrhosis, PChE values were
about 50% lower then in healthy infants of the same age
i.e. 2–6 months (P<0.001). In healthy infants PChE activity was lower by 20% than in adults (P<0.001). Significantly lower albumin concentration in all infants than in
adults has also been established (P<0.05).
Conclusion: Our results showed diminished synthesis of
PChE and albumin in liver cells of infants (2–6 months).
These observations would suggest that infants might be
expected to differ significantly from adults in the extent to
which drugs might be protein bound and in the rates of
hydrolysis of ester-type drugs. It is well known that PChE
is involved in the hydrolysis of muscle relaxants succinyl-
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
choline and mivacurium. According to our results, 50%
lower PChE activity in infants with biliary cirrhosis may
cause prolonged duration of mentioned drugs.
Ecobichon DJ, Stephens DS: Perinatal development of human
blood esterase. Clin Pharmacol Ther 1973, 14:41–47.
Strauss AA, Mondanlou HD: Transient plasma cholinesterase deficiency in preterm infants. Dev Pharmcol Ther 1986, 9:82–87.
Serum albumin and mortality risk in critically ill patients
FHY Yap, GM Joynt and TA Buckley
Department of Anaesthesia and Intensive Care, Prince of Wales Hospital, Shatin, HKSAR, China
Introduction: It had been shown that serum albumin measured within 48 h of ICU admission is as accurate as the
APACHE II score in predicting hospital mortality. If this
result could be confirmed, then albumin would be a
cheaper and more convenient predictor of mortality than
Aim: To evaluate serum albumin as a predictor of hospital
mortality in critically ill patients.
Methods: All consecutive adult ICU admissions over an
18 month period (April 1997 to September 1998) were
reviewed. The data were retrieved from a computerised
database. Serum albumin concentrations between survivors and nonsurvivors were compared. Accuracy of
outcome prediction using albumin was assessed by area
under the curve (AUC) of the receiver operator characteristics (ROC) curve.
Results: 1003 records were reviewed. There were 647
(64.5%) survivors and 356 (35.5%) nonsurvivors. The sur-
vivors had a significantly higher serum albumin both on
ICU admission and in the first 72 h after admission
(P<0.001) (see Fig. 1). In both groups, the serum albumin
fell markedly after ICU admission. The ROC curves for
hospital mortality are shown in Fig. 2. The AUC of
APACHE II (0.79) is significantly higher than that of admission albumin (0.63) and albumin at 24 h (0.64) (P<0.001).
We used regression analysis to combine APACHE II with
admission albumin (AUC 0.80) and APACHE II with
albumin at 24 h (AUC 0.79) to predict hospital mortality.
The addition of albumin did not significantly improve the
accuracy of outcome prediction over that of APACHE II
Conclusion: There is an association between serum
albumin concentration and hospital mortality. However,
as a single variable, serum albumin is a poor predictor of
hospital outcome. Combining APACHE II with admission albumin or albumin at 24 h did not improve the
accuracy of outcome prediction over that of APACHE II
score alone.
Figures 1 (left) and 2 (right)
ROC Curves
Serum Albumin (Mean ± 2SE)
Day 1
Day 2
Day 3
Albumin D1
APACHE + Albumin
APACHE + Albumin D1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Parenteral nutrition (PN) depresses hepatic albumin synthesis in septic rats
MJ O’Leary*, M Koll†, CN Ferguson*, CJ Hinds*, VR Preedy†, JH Coakley* and PJ Garlick‡
*Department of Intensive Care, St. Bartholomew’s Hospital, London EC1A 7BE, UK; †Department of Nutrition and Dietetics,
King’s College London, London SE1 8WA, UK; ‡Department of Surgery, SUNY, Stony Brook, New York, USA
Serum albumin concentration is frequently used as an indicator of nutritional response in patients receiving PN. The
acute-phase response following injury and sepsis is characterised by increased hepatic synthesis of specific
secreted proteins while production of albumin is
decreased. We have investigated the effect of acute
administration of PN on absolute and relative rates of
hepatic albumin synthesis during sepsis in vivo. Two
groups of male Wistar rats (mean weight 239 g) underwent caecal ligation and puncture (CLP), with a third
group as unoperated controls allowed free access to
chow and water (ad lib). Between 18 and 24 h later CLP
survivors were infused by tail vein with 0.9% NaCl or a PN
solution delivering 33% of calculated daily energy and
protein requirements. Total hepatic protein synthesis rate
(TPS), albumin synthesis rate (TAS) and relative albumin
synthesis rate (RAS) were determined with a flooding
dose of 3H-phenylalanine and the use of anti-rat albumin
antibody to isolate the protein for radioactive counting.
Results: (mean ±SEM) are displayed in the table, with
statistical analysis by ANOVA.
Group (n)
ad lib (5)
0.9%NaCl (8)
PN (9)
TPS (mg/day)
TAS (mg/day)
RAS (%)
*P=0.0011 compared with ad lib; #P=0.1, •P=0.0411
compared with 0.9% NaCl.
We have demonstrated that whereas TPS is increased in
sepsis, TAS falls. Provision of PN has no effect on TPS,
but significantly further depresses RAS. Reduced hepatic
albumin synthesis in sepsis is not reversed by substrate
provision, hence changes in serum albumin concentration
are unlikely to be a useful monitor of efficacy of PN.
Acknowledgement: MJ O’Leary was supported by The British Journal of
Anaesthesia, BMI/Columbia Healthcare Ltd., and The Joint Research Board
of St. Bartholomew’s Hospital.
Measurement of gastric emptying in ICU patients with 13C-acetate
Ch Goeters, C Schulzki, A Wenn, U Hölscher* and H Van Aken
Department of Anaesthesiology and Intensive Care Medicine, University of Münster, Albert-Schweitzer-Str. 33, D-48149
Münster; *Faculty of Applied Physics and Engineering, University of Applied Sciences, Münster, Germany
Introduction: In clinical practice, gastric emptying and the
success of gastric feeding is assessed by measurement of
gastric residuals. A new setting easily allows bedside
measurements in ventilated ICU-patients by 13C-technology to evaluate gastric transport.
% recovery
Methods: 13C-acetate added to tube feeding will be
absorbed into the blood stream from the duodenum and
oxidised to 13CO2 in the liver. The ratio of 13CO2/12CO2 in
the expired air can be determined by infrared isotopeselective spectrometry (IRIS). Increasing expiratory concentrations of 13CO2 indicate transport of the applied
tracer from the stomach to the duodenum. In 11 healthy
volunteers breathing spontaneously via a tight CPAP mask
on a Dräger Evita respirator system, 13CO2 enrichment in
the expired air was measured for two hours after ingestion
of 150 mg 13C-acetate in 50 ml tube feeding. In nine ventilated ICU patients measurements were performed to
determine expiratory 13CO2/12CO2 ratios after administration of the same amount of tracer and tube feeding via
gastric tube. Gastric reflux and the total amount of tube
feeding were recorded before measuring.
time [min]
Cumulative recovery (mean ± SD) in volunteers (t) and ICU-patients
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Results: The new setting was validated in volunteers.
About 20 to 40 min after ingestion of the tracer, peak
values (% of 13C dose/h) were recorded. Within two
hours the cumulative recovery was 20%. In the ICU
patients only a small quantity of the applied tracer (<5%)
was recovered independent of reflux amounts.
Conclusion: We could not quantify gastric emptying by
this new setting in ICU patients. Failure might be due to
changes in transport mechanisms, resorption in the duodenum, metabolism and pool exchange compared to
healthy volunteers.
Propofol-induced inhibition of intestinal peristalsis involves enteric opioidergic pathways
MK Herbert, W Berg, P Holzer* and N Roewer
Department of Anaesthesiology, University of Wuerzburg, Josef-Schneider-Strasse 2, D-97080 Wuerzburg, Germany;
*Department of Exp. and Clin. Pharmacology, University of Graz, Austria
Introduction: In recent years, the short acting propofol is
increasingly used for sedation in intensive care units (ICU)
patients. Inhibition of intestinal peristalsis is a major
problem in those critically ill patients and an inherent
feature of sedatives such as barbiturates or midazolam.
Since little is known about the action of propofol on
intestinal peristalsis, this study investigates the effect of
propofol on the ileal peristaltic reflex in vitro.
Methods: Ileal segments of adult guinea-pigs were
mounted in silanized organ baths that contained oxygenated Tyrode solution (30 ml, 37°C). Prewarmed
Tyrode solution was infused into the intestinal lumen, the
infusion rate being 0.5 ml min-1. The fluid passing the gut
lumen was directed into a vertical outlet tubing which
ended 4 cm (reflecting a pressure of 400 Pa) above the
fluid level of the organ bath. This arrangement caused
gradual filling of the intestine. When the intraluminal pressure reached a threshold an aborally moving wave of circular muscle contraction, measured as a spike-like
increase in intraluminal pressure, propelled the intraluminal
fluid to leave the system and thus caused emptying of the
segment. The pressure threshold for eliciting peristaltic
waves (peristaltic pressure threshold - PPT) was used to
quantify the effects of drugs on peristaltic activity. Inhibition of peristalsis was reflected by an increase of PPT.
After registration of normal peristalsis the segments were
exposed to propofol (0.1–100 µM), which was administered cumulatively into the bath, i.e., to the serosal surface
of at least 6 intestinal segments from 6 different guineapigs either alone, after application of vehicle (DMSO
diluted with Tyrode solution), the GABAA receptor antagonist bicuculline (BIC, 10 µM), or the opioid receptor antagonist naloxone (NAL, 0.5 µM).
Results: Propofol (0.1–100 µM) concentration-dependently increased the PPT, however, never caused complete inhibition of intestinal propulsion in the doses tested.
While BIC (10 µM) failed to prevent the inhibitory effect of
propofol, the increase in PPT was absent after pretreatment with NAL (0.5 µM).
Conclusions: Propofol inhibits intestinal peristalsis by
increasing the PPT, but is less potent and effective compared to barbiturates or clonidine since it did not abolish
peristaltic activity. The inhibitory effect of propofol seems
not to be mediated through binding to GABAA receptors,
but seems to involve enteric opioidergic pathways.
Clonidine concentration-dependently inhibits intestinal peristalsis in vitro
R Roth-Goldbrunner, MK Herbert, P Holzer* and N Roewer
Department. of Anaesthesiology, University of Wuerzburg, Josef-Schneider-Strasse 2, D-97080 Wuerzburg, Germany, and *
Department of Exp. and Clin. Pharmacology, University of Graz, Austria
Introduction: The α-adrenoceptor agonist clonidine provides a good alternative for sedation in patients in intensive care units (ICUs). Inhibition of intestinal transport, the
development of ileus, and subsequently other complications are major side effects of sedatives used in ICU
patients. This study examines whether clonidine exerts an
inhibitory effect on intestinal peristalsis.
Methods: Ileal segments of adult guinea-pigs were
mounted in silanized organ baths that contained oxygenated Tyrode solution (30 ml, 37°C). Prewarmed
Tyrode solution was infused into the intestinal lumen, the
infusion rate being 0.5 ml min-1. The fluid passing the gut
lumen was directed into a vertical outlet tubing which
ended 4 cm (reflecting a pressure of 400 Pa) above the
fluid level of the organ bath. This arrangement caused
gradual filling of the intestine. When the intraluminal pressure reached a threshold an aborally moving wave of circular muscle contraction, measured as a spike-like
increase in intralumal pressure, propelled the intraluminal
fluid to leave the system and thus caused emptying of the
segment. The pressure threshold for eliciting peristaltic
waves (peristaltic pressure threshold - PPT) was used to
quantify the effects of drugs on peristaltic activity. Inhibition of peristalsis was reflected by an increase of PPT.
After registration of normal peristalsis the segments were
exposed to clonidine (1 nM–100 µM), which was administered cumulatively into the bath, i.e., to the serosal surface
of 6 intestinal segments from 6 different guinea-pigs either
alone, after application of Tyrode solution (vehicle), or of
the adrenoceptor antagonist yohimbine (1 µM).
yohimbine (1 µM) prevented clonidine (0.1–100 nM) from
having any inhibitory effect on PPT.
Results: Clonidine concentration-dependently increased
the PPT. While Tyrode solution (vehicle) and 0.1–1 nM
clonidine were without any effect on the PPT,
10 nM–10 µM clonidine caused an increase of PPT and
complete abolition of peristalsis occurred after 100 µM
clonidine in 4 of 4 segments tested. Pretreatment with
Conclusions: Clonidine concentration-dependently inhibits
the ileal peristaltic reflex in vitro through adrenoceptors
located in the intestine. It is assumed that clonidine also
affects propulsive peristalsis in ICU patients and thus
might contribute to other complications such as ileus and
multiple organ failure in those patients.
Early enteral nutrition after Pancreas Kidney Transplantation (PKTx) with enteral drainage (a pilot
study in five patients)
D Agthé*, AR Mueller†, H Full*, I Sauer†, WO Bechstein†, A Kahl†, KJ Falke* and P Neuhaus†
*Department of Anesthesiology and Operative Intensive Care and †Department of General and Transplant Surgery, Charité
Campus Virchow, Augustenburger Platz 1, D-13353 Berlin, Germany
Introduction: The use of early enteral nutrition, in particular formulas containing glutamine, arginine and omega-3fatty acids (immunonutrition), may have an impact on the
post-operative course of critically ill patients. Previous
research has demonstrated a decrease in complication
rates, morbidity, and mortality. Little data exist, so far,
about the use in pancreas-kidney-transplant recipients
with enteral drainage. Because of the enteral anastomosis,
most surgeons restrain from feeding the patients enterally,
and postpone the start of enteral nutrition until at least five
days after the procedure. No data is published on early
enteral immunonutrition in this type of patient. The aim of
this preliminary study, is to analyze the feasibility and complication rates of early enteral immunonutrition in pancreas-kidney-transplant recipients with enteral drainage.
graphic and clinical characteristics including age, sex,
SAPS II score, ICU and hospital mortality were recorded.
The patients were screened for the occurrence of rejection,
opportunistic infections and surgical complications.
Patients and Methods: Prospective analyzes of five
patients after PKTx was performed. Nasogastric feeding
with Impact® (Novartis, Switzerland), 25 cc/h was started
immediately after surgery. This was continued for the first
five days after PKTx. Additionally, patients recieve a parenteral standard nutrition program and substitution of fluids
as needed to maintain diuresis. Further medication included
propulsive agents (metoclopramide 20 mgs i.v. every 6 h,
cisapride 10 cc p.o. every 6 h) and immunosuppression
(tacrolimus 0.05 mgs/kg or cyclosporine 4 mgs/kg, ATG 5
mgs/kg, prednisolone 20 mgs/day, mycophenolatmofetile
2 g/day). During the study for each patient basic demo-
Discussion and conclusion: The number of five patients
is to small to make a valid conclusion and the results are
rather descriptive. They show that early enteral feeding in
this group of patients is possible without an increase in
postoperative complications. Furthermore, early regeneration of mucosal gut barrier function may decrease infectious complications and improve patient outcome. In order
to confirm this hypothesis further research should be performed to investigate the impact of this strategy. Variables
like infection rates, surgical complications, and rejection,
as well as length of ICU stay, morbidity and mortality will
be analyzed.
Results: Early enteral immunonutrition was tolerated by all
the patients. No surgical complications, especially no
leakage at the enteric anastomosis, were observed. There
was also no occurrence of transplant pancreatitis and a
reduced incidence of bowel distention and vomiting.
Except for one CMV-infection, no opportunistic infections
or other infectious complications were observed. All
patients developed rejection; rejection was mild and
steroid-sensitive in four patients, one patient required OKT
3 treatment and tacrolimus rescue therapy.
Impact of carrier solution on biological insulin availability
Z Rusavý, V Srámek, R Suchá, E Langhamerová, R Rokyta and O Topolcan*
Medical Dpt 1, *Medical Dpt 2, Charles Uni Hospital, 304 60 Pilsen, Czech Republic
Objectives: When added into all-in-one (AIO) bags, total
needs of exogenous insulin decrease compared to
amounts given by perfusor. We studied the influence of
carrier solution on insulin availability.
Materials and methods: A 20 ml polyvinylchloride (PVC)
syringe was filled either with saline or AIO solution aspirated from original 3 l bags (Nutrimix, Brown, Melsungen,
Germany). Then 8 IU of insulin (Actrapid HM, Novo-
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Nordisk, DK) was added into syringes (calculated concentration 400 IU×l-1) and a PVC perfusion line (5 ml volume)
was filled. Baseline samples (0.5ml) for immunoreactive
insulin assay (IRI) were taken. Perfusor rate was set at 2
ml×hr-1 and samples taken directly from the hose at 5, 10,
30, 60, 90 and 210 min, placed on ice and stored at 4°C
before analysis. Fifteen sets of measurements were done
both for saline and AIO solution. Values are presented as
means ±SD. MANOVA, ANOVA for repeated measures
and paired T-test with Bonferroni correction were used
when appropriate; P<0.05 was considered significant.
Figure 1
insulin concentration [IU/l]
Critical Care
all in one
Results: Figure 1 shows IRI concentrations in both saline
and AIO solutions during the experiment.
IRI differed significantly giving higher IRI yield in AIO solution compared to saline (MANOVA group by time effect
P<0.001). Changes in IRI concentration depending on
time was seen in saline only (ANOVA time effect P<0.001
for saline; P=0.26 for AIO). When separate time points for
saline were analyzed major changes were seen at the very
beginning of the experiment (a decrease from 206±34 at
baseline to 152±33 IU*l-1 at 5 min; P<0.001). The mean
yield of IRI expressed as ratio of measured to calculated
Time [min]
concentration at baseline was 0.52 for saline and 0.70 for
Conclusion: Insulin bioavailability in AIO solution is better
than in saline with respect to the yield and stability.
Delivery of parenteral nutrition by ‘all in one bag’ (AIO) versus single bottle system (SBS) in
critically ill patients: a prospective randomized trial
C Ebener*, JM Harms†, B Thümmer†, R Buhl‡, A Röhrborn†, PE Goretzki† and HD Röher†
*Department of General Surgery, University of Regensburg, Franz Josef Strauß Allee 10, 94053 Regensburg; †Department of
General and Trauma Surgery, Heinrich Heine University Moorenstraße 5, 40225 Düsseldorf; ‡Department of Anaesthesiology,
Heinrich Heine University, Moorenstraße 5, 40225 Düsseldorf, Germany
Rationale: AIO has become an easy and practical way to
deliver total parenteral nutrition (TPN) in patients on a
regular ward. However no data are available yet concerning the use of AIO in critical ill patients. Here the SBS is
still the gold standard.
Method: A prospective randomized trial was conducted in
surgical patients who were expected to need TPN for more
than five days after major surgery or trauma. Patients were
randomized to receive for at least five days an isocaloric
and isonitrogenic TPN delivered as AIO or as SBS. The
TPN met the caloric requirements, which were determined
by indirect calorimetry. Substrates delivered, numerous laboratory parameters, metabolic disturbances as a result of
the nutrition, as well as expenses for equipment and manpower were documented daily. Forty patients are required
by the protocol to complete the study.
Results: 37 patients have entered the study since 2/98.
3/37 refused to continue after the first day (drop out). 16
patients received AIO (86 days recorded) and 18 pat
received SBS (101 days recorded). There was no difference in age, sex, diagnosis, risk factors between the treatment groups. Three patients died during the hospital
course (18, 14 and 123 days after completing the study),
all in the SBS group. Cause of death was in no case
related to the TPN. The nutrition could be delivered
without problems in both groups. Reduction of the TPN
for hyperglycaemia was necessary only in one patient in
each treatment group. Reduction of the TPN for hypertriglyceridaemia was necessary in one patient in each
group. The SBS required more equipment e.g. infusion
pumps and disposable infusion systems as well as more
manpower e.g. to fix technical alarms.
Conclusion: AIB seems to be an useful alternative to SBS
in critically ill and metabolically stable surgical patients.
Glutamine-enriched parenteral nutrition during postoperative catabolic state
O Ahlers, K Harndt, R Hodek, C Kietzmann, M Pettersson, K Ruland, Kv Fournier, P Neuhaus, K Falke and H Gerlach
Department of Anesthesiology and Intensive Care Medicine, Charité - Campus Virchow Clinic, Humboldt-University, 13344
Berlin, Germany
Background: Previous clinical studies led to the hypothesis
that glutamine-enriched total parenteral nutrition (TPN) has
beneficial effects on clinical course and on changes in the
immune system of patients in postoperative catabolic state.
Patients and methods: 20 patients with esophageal carcinoma were enrolled in a controlled, prospective, randomized and double-blinded study after esophageal resection:
10 patients received glycyl glutamine-enriched (0.15 g glutamine/kg/d) TPN (Glamin®, Baxter Inc.) from day 2 postoperatively, 10 patients served as control (TPN without
glycyl-glutamine: Vamin®, Fresenius Inc). Clinical scores
and immunological parameters were observed before and
during the first 8 days after surgery. Lymphocyte subpopulations were measured using flow cytometry analyses
(FACS-Flow®, Becton Dickinson), plasma cytokine levels
were determined using standard ELISA technique.
Results: No significant differences were found, neither in
clinical nor in immunological parameters: SAPS II scores
as well as GORIS scores showed similar courses in both
groups and ICU-days did not differ. Postoperative alterations of the immune system, measured by plasma IL-6
and procalcitonin levels as well as changes in lymphocyte
counts and T4-T8-ratio, were nearly similar.
Conclusion: A beneficial effect of glutamine-enriched
TPN could not be found in our study population. Further
investigations are necessary to find out if higher doses
and/or other patient groups are able to demonstrate
similar effects as shown in former studies.
Acknowledgement: Supported by Baxter Inc., Munich, Germany.
Muscular mass and subcutaneous fat changes according to duration of ICU hospitalization with
corticosteroid and muscle relaxant intake
M Moukas, AN Chalazonitis, P Dimou, A Patsalides, Cr Kouraziotis, E Xinis, C Mandragos, M Vasiliou and PK Behrakis
Intensive Care Unit and Department of Radiology, Hellenic Red Cross Hospital, 1 Erythrou Stavrou str. Athens 11526, Greece
Introduction. In the present study we profit from the high
U/S sensitivity and specificity in order to study the muscular mass (MM) and subcutaneous fat (SBF) changes
during Intensive Care Unit hospitalization. In addition, we
investigate the influence of the administration of muscle
relaxants and corticosteroids.
Materials and methods: Thirty patients (19 male aged
51.1±14.15 years hospitalized for 18.84±3.5 days and
11 female aged 64.36±12.4 years hospitalized for
19.18±4.6 days) entered the study. Clinical examination
and computerized tomographic scan proved intracerebral
hemorrhage in 14 of them and brain infarction in 16 of
them. The patient’s nutrition schedule included 30
Kcal/kg/day with protein uptake of 1 g/kg/day. Physiotherapy (passive movements of the limbs) was applied for 1
h/day in all patients. Muscle relaxation was applied to 9
patients with atracurium (0.6–0.7 mg/kg). The mean duration was 2.77±0.44. Dexamethasone (24 mg/d iv) was
applied to 14 patients for a mean duration of 4 days. Two
ultrasonographic measurements were done for each
patient; one on the first and another on the last hospitalization day in the ICU. Upper arm biceps MM and SBF thickness were assessed by M-mode ultrasonography using a
10 MHz linear-array high resolution transducer. All scans
were performed with the same ultrasonographer. MM and
SBF thickness are expressed in cm. MM, and SBF
changes were calculated as differences of the 2nd mea-
surement minus the 1st and they were expressed as
∆MM[cm] and ∆SBF[cm], respectively.
Results: A significant loss of MM and SBF was observed
in all patients included in the study. SBF loss (mean
∆SBF%=41.1±16.70) was significantly higher than corresponding MM reduction (∆MM%=21.4±4.90). A significant correlation was found between ∆MM% and the
duration of hospitalization (19±4 days), (r=0.62, P<0.05).
On the contrary ∆SBF% was poorly correlated with the
duration of the hospitalization (r=0.10, P>0.05). Neither
∆MM% nor ∆SBF were correlated with patients’ age.
Patients who took corticosteroids (n=5) showed a nonsignificantly higher percentage of MM loss in comparison
to those who did not take corticosteroid (n=15; corresponding ∆MM%: 22.1±5.22 versus 20.0±4.83%). Nevertheless, ∆MM% for the patients, who did not take
muscle relaxation was slightly different (20.4±4.53) from
∆MM% of patients who did not take corticosteroids or
corticosteroids and muscle relaxants. A significantly higher
reduction of MM was observed in patients who took corticosteroids and muscle relaxants in comparison to those
who did not take either corticosteroids or muscle relaxants
(corresponding ∆MM%: 23.8±4.90 versus 19.8±4.47).
SBF did not differ significantly between patients who took
and did not take corticosteroids (corresponding ∆SBF%:
47.5±26.47 versus 40.6±18.91). Not significant was also
the SBF difference between those who took corticos-
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
teroids and muscle relaxants (∆SBF%: 35.9±17.39) in
comparison to those who did not take either corticosteroids or muscle relaxants (∆SBF%: 42.0±12.69).
∆SBF% for patients who did not take muscle relaxation
(43.2±15.95) did not differ significantly from corresponding ∆SBF values of the other formatted groups.
Conclusions: Our study shows that MM loss is correlated
with the duration of hospitalization as well as with the synchronous corticosteroids and muscle relaxants administration. On the contrary SBF, although decreased during the
period of our study, is not statistically correlated with any
of the studied factors.
Incidence of cortisol deficiency in patients with traumatic brain injury
MA Staber, DJ Cooper, O Montero and L Murray
Intensive Care Department, The Alfred Hospital, Prahran, Melbourne 3181, Australia
Cortisol deficiency due to hypothalamic or pituitary injury
has been uncommonly reported after traumatic brain injury
(TBI), and the incidence is unknown. Diabetes insipidus
after TBI, which usually reflects posterior pituitary injury, is
however frequently recognized.
Blood cortisol concentrations are usually high in ICU
patients — often 2–3 times greater than the normal
morning peak values of 500 nmol/l. Cortisol has a relatively short half-life and a permissive action on catecholamine efficacy, so cortisol deficiency could manifest
rapidly, and cause hypotension.
In 80 patients with severe TBI (GCS <9), and in 12 controls (severe multi-trauma without TBI), we measured
blood cortisol and ACTH concentrations on days 1,3 and
7 of ICU admission. Fifteen of 80 patients (19%) had
absolute cortisol deficiency with morning blood cortisol
concentrations <140 nmol/l (normal 140–500) and ACTH
<10 ng/l (normal <65) on days 3 or 7 of ICU admission.
These patients had a mortality of 21% compared to 4% in
the others. Twenty-four of the 80 patients (30%) had
blood cortisol concentrations <250 nmol/l. None of the
controls had a blood cortisol concentration <140 nmol/l.
TBI patents with blood cortisol concentrations
<250 nmol/l had greater elevation of intracranial pressure
32±45 vs 9±23 h >20 mmHg (P<0.05), and greater peak
vasopressor doses 18±14 vs 9±16 µg/min (P<0.05) than
the others.
Absolute cortisol deficiency occurs during the first 7 days
of ICU admission in 19% of patients with severe TBI, and
is likely to be of hypothalamic or pituitary origin. These
patients had greater vasopressor doses suggesting that
cortisol deficiency may contribute to vasopressor-resistant
hypotension in patients with severe TBI.
Treatment of hypophosphataemia in critically ill patients with a two day dosing regimen
LJ Giles, A Jennings, B Ng, G Creed, L Gallard, D Pierre, R Beale and A McLuckie
Department of Intensive Care, Guy’s and St. Thomas’ Hospital Trust, St. Thomas’ Hospital, London, SE1 7EH, UK
Introduction: Standard phosphate replacement regimens
[1,2] are only 21–30% effective at correcting serum concentrations in severely hypophosphataemic critically ill
patients, with 60–85% of patients redeveloping hypophosphataemia. We evaluated a two day regimen employing a
rapid infusion rate.
Method: Fifty-two intensive care patients developing
hypophosphataemia (<0.8 mmol/l) were evaluated. Intravenous phosphate doses on days 1 and 2 were based on
the morning serum phosphate concentration: 0.8–0.6
mmol/l: 20 mmol and 20 mmol, 0.6–0.4 mmol/l: 40 mmol
and 20 mmol, <0.4 mmol/l: 50 mmol and 20 mmol.
Table 1
Mean (SD) serum
PO4 (mmol/l)
Patients % successfully
corrected (≥0.8 mmol/l)
Mean (SD) change in
ionised Ca (mmol/l) from
pre-treatment value
Serum PO4 mmol/l
dose 1
dose 2
dose 1
dose 2
dose 2
dose 1
dose 2
Sodium phosphate (0.1 mmol/ml) was infused at 10
mmol/h. Serum phosphate and ionised calcium concentrations and calcium phosphate product were determined
pre-infusion, immediately post-infusion and 6 h post-infusion for each dose and after 24 h. Twenty-two patients
were receiving CRRT.
Results: See Table 1.
Only seven calcium phosphate product values were
slightly raised (range: 4.9–5.7 mmol2/l2). Twenty-two
patients had ionised calcium concentrations of ≤1.05
mmol/l on occasions during the regimen but the extent of
the hypocalcaemia was mild (mean [SD] ionised Ca: 0.99
(0.07) range: 0.68–1.05 mmol/l [n=50]) and usually transient. Furthermore the degree of change in ionised
calcium concentrations from pre-treatment values for all
patients was not significant (see Table 1). No clinical evidence of hypocalcaemia was noted.
Conclusion: The two day intravenous phosphate regimen
described is safe and highly effective at correcting
hypophosphataemia in critically ill patients.
Perreault MM, Ostrop NJ, Tierney MG: Ann Pharmacother 1997,
Clark CL, Sacks GS, Dickerson RN et al.:Crit Care Med 1995,
Hypoglycemia masquerading as acute psychosis in young age
P Klemen, Š Grmec and D Cander
EMS–PHU ZD Maribor, Ul. Talcev 9, 2000 Maribor, Slovenia
Introduction: The classic presentation of hypoglycemia
includes tachycardia, anxiety, extreme hunger, tremor, palpitations, pallor, weakness and an altered level of consciousness. The hypoglycemic patient’s presentation may
also lead to suspect another condition such as cerebrovascular accident, seizure, head injury or acute psychosis.
Manifestations of Hypoglycemia
Age (Years)
Focal Neurol. Findings or Seizure
A clinical study was done to determine whether there is a
correlation between the manifestations of hypoglycemia
and the patient’s age.
Methods: We studied 191 prehospital patients in the
period from February 1998 to September 1999. For each
patient we studied: the level of blood glucose, age and
clinical manifestations of hypoglycemia.
Results: 99 (51.4%) patients presented with coma, 24
patients (12.6%) presented with focal neurological deficit
or seizure activity, 56 (29.2%) were confused and 12
(6.8%) presented as acute psychosis.
We found out that there existed a statistically important
difference in patient’s age between clinical manifestations
of hypoglycemia (Table).
Conclusion: In our study we found out that psychosis as a
manifestation of hypoglycemia presents in significantly
younger patients than other manifestations (Student t-test
<0.05). In elderly patients hypoglycemia presents more
often with focal neurological deficit, seizure, confusion and
coma. Our cases illustrate the importance of considering
hypoglycemia in all patients who present with alterations
in mental status even when the clinical findings seem to
be explained initially by other etiologies. Emergency physicians must be aware of such a presentation of hypoglycemia and the need for rapid testing, since testing is
easily performed and therapy is most often curative with a
good outcome.
Plasma Cl:Na ratio: a simple alternative to Stewart’s Strong Ion equation for detection of
elevated lactate or unmeasured anions in metabolic acidosis
S Skellett, A Mayer, A Durward, S Tibby and IA Murdoch
Guy’s Hospital, London SE1 9RT, UK
Introduction: There has been a resurgence of interest in
the interpretation of acid base abnormalities using Stewart’s Strong Ion Theory [1], central to which is the solving
of a complex quadratic equation that accurately quantifies
the contribution of both organic and inorganic anions to
pH. In accordance with this theory, we hypothesised that
an increase in organic anions (lactate and unmeasured
anions) during metabolic acidosis would cause a compensatory fall in inorganic chloride thus preserving electroneutrality. The aim of this study was to determine if this
compensatory hypochloraemia, expressed as a fall in the
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Cl:Na ratio, could be a useful method to detect these elevated organic anions in patients with metabolic acidosis.
Patients and methods: Paired blood samples were taken
on admission and at 24 h from 341 patients (median
weight 8kg, IQ range 3–15) admitted to a regional PICU,
from which serum electrolytes, lactate and blood gases
were measured and the strong ion gap calculated to quantify unmeasured anions (UMA). A metabolic acidosis was
defined as being present if the base deficit (BD) was
≥5mmol/l and was chosen rather than pH since it better
defines the metabolic component of an acidosis [2].Using
the above criteria for acidosis, 136 samples were selected
and further divided into two subgroups using a Cl:Na ratio
≤0.74 as the cutoff value and compared for acid base
parameters, lactate and UMA concentrations. The cut off
value of 0.74 represents the 1st quartile for Cl:Na ratio
derived from analysis of the 341 paired samples (median
0.76, IQ range 0.74–0.78). This approach was necessary
since there are no normal values for Cl:Na ratio.
Table 1
Median values; (1st and 3rd quartiles)
Cl:Na ≤0.74
Cl:Na >0.74
P- value
7.22 (7.13–7.31)
7.28 (7.21–7.32)
pCO2 (kPa)
4.4 (3.7–6.7)
4.6 (3.9–5.3)
BD (mmol/l)
10 (6–13)
7 (6–11)
LAC (mmol/l)
7 (2.2–9)
1.3 (0.8–3.6)
UMA (mmol/l)
12 (8–15)
3 (–1–7)
Table 2
ROC curve values for Cl:Na ratio ≤0.74. Values in mmol/L
(95% CI)
Sensitivity Specificity
Results: In the episodes characterised by a BD ≥5mmol/l
(n=136), the blood lactate and UMA concentrations were
significantly greater in the subgroup with a Cl:Na ratio
≤0.74 despite having similar conventional acid base parameters (pH, pCO2 and BD).
0.75 (0.64–0.85)
0.82 (0.75–0.9)
Lactate + UMA
0.86 (0.8–0.93)
Conclusion: A low plasma Cl:Na ratio (≤0.74) is a simple
alternative test for detecting the presence of abnormal elevations in lactate (>2.4 mmol/l) or UMA (>4.3 mmol/l)
concentrations in patients with metabolic acidaemia. This
obviates the need to solve the complex quadratic equa-
tions required for calculation of UMA using the Strong Ion
Theory of Stewart.
Stewart: Can J Physiol Pharmacol 1983, 61:1444.
Balasubramanyan: N: Crit Care Med 1999, 8:1577–1581.
Effects of angiotensin II on the renal excretion of urodilatin in the isolated rat kidney
M Heringlake*, K Wagner*, L Bahlmann*, S Klaus*, J Schumacher*, H Pagel† and P Schmucker*
*Klinik für Anaesthesiologie, and †Institut für Physiologie, Medizinische Universität Lübeck, D - 23538 Lübeck, Germany
Background: The kidney-derived peptide urodilatin has
been suggested to be important in the regulation of natriuresis and diuresis. This hormone has been used clinically
as a drug for the treatment of incipient renal failure in
cardiac surgery patients and is known to be increased in
these patients if renal function is preserved [1]. However,
the mechanisms involved in the regulation of the production of urodilatin are far from being clear. This study was
designed to determine, if Angiotensin II may influence the
renal excretion of urodilatin (VUROU).
Materials and methods: Isolated kidneys from SpragueDawley rats were perfused for 180 min with a pressure of
100 mmHg in a recirculating perfusion system. After an
equilibration time of 90 min (baseline-period), the following drugs were added to the perfusion medium:
Angiotensin II (AII: n=5); Losartan (LOSA: n=5), and both
drugs combined (LOSA-AII: n=5) and the kidneys were
perfused for another 90 min (intervention-period). Five
kidneys served as controls (CON). Measurements were
performed every 30 min. Data of both observation periods
were averaged (baseline: 0–90 min; intervention:
91–180 min); baseline data were normalized to 100%.
Statistical analyses were performed with ANOVA and
Fisher’s PLSD.
Results: Data for the intervention period are given in
percent of average baseline values (Table 1). RVR: renal
vascular resistance; VU: urine flow; VNaU: urinary excretion
of sodium; GFR: glomerular filtration rate; VUROU urinary
excretion of urodilatin.
Conclusions: These observations suggest that VUROU is
affected —directly or indirectly — by vasoconstrictory concentrations of Angiotensin II and may help to explain
recent findings on the regulation of the renal excretion of
urodilatin; especially the positive relationships of VUROU
with plasma sodium [2] and renal perfusion pressure [3].
Table 1
*Denotes a significant difference (P<0.05) vs control.
Sehested et al.: J Thoracic Cardiovasc Surg. 114:666–671.
Drummer et al.: Am J Physiol 262:F744–F754.
Heringlake et al.: Am J Physiol 277:F347–F351.
Plasma endothelin-1 and natriuretic peptide levels and antioxidant capacity of heart transplanted
I Olejarova, K Danova, H Minarova, V Fischer, V Rendekova, J Fabian and I Pechan
Slovak Institute of Cardiovascular Diseases, Pod Krasnou Horkou 1, Bratislava, Slovak Republic
There is evidence that a transplanted heart does not
provide the recipient with functionally normal cardiac
function and many differences in resting haemodynamics
and alterations in electrophysiology and neurohumoral
disbalance can be seen. Therefore, the plasma levels of
functionally important metabolites — endothelin-1 (ET-1)
as well as A- and B-type natriuretic peptides (ANP and
BNP) — are released into the blood bed in response to
fluid retention and vasoconstriction in heart disorders.
Besides these haemodynamically-interesting parameters,
in various intervals after the heart transplantation (HTx),
the actual endogenous antioxidant capacity of patients
was studied. Before HTx and successively three times
during the first week and then weekly during the first
month after HTx, seven transplanted patients were carefully clinically and biochemically investigated. In their
blood plasma (or erythrocytes) the total antioxidant status
(TAS), activities of superoxide dismutase (SOD) and glutathione peroxidase (GPx) as well as plasma levels of
thiobarbituric acid reactive substances (TBARS), ET-1,
ANP and BNP were estimated.
During the whole investigation period the gradual
decrease of TAS was noted. Compared with the pretransplant levels, a gradual decrease of plasma levels of
ET-1 and successive rise in ANP and BNP levels were
observed, especially during the first week after the HTx. In
conclusion, we have shown that HTx-patients demonstrated decreased total antioxidant capacity and characteristic movements of plasma levels of analysed peptides
reflects the response of the cardiovascular system to the
haemodynamically different situation.
Management and outcome of symptomatic hyponatremia in emergency department
J Kato and Y Tokuda
Okinawa Chubu Hospital, 208-3 Miyazato, Gushikawa city 902, Okinawa, Japan
Introduction: Severe symptomatic hyponatremia was
thought to be extremely dangerous. It is recommended
that acute hyponatremia should be treated without delay
and rapidly at a rate of at least 1 mmol/l/hour. Symptomatic hyponatremia was said to lead to death or permanent brain damage. However, some authors did not
support quick correction and suggested that most deaths
were caused by underlying diseases. We evaluated the
clinical outcome of cases with symptomatic hyponatremia
in terms of different management strategies in the emergency department.
Method: In the emergency outpatient department of a
community hospital, Okinawa, Japan, we retrospectively
collected adult cases of symptomatic hyponatremia
(serum sodium <130mmol/l) from April 1995 to October
1999. Etiology, treatment and clinical outcomes were
evaluated. Neurological sequelae were assessed in all
Result: Eighty-two cases of hyponatremia were determined (mean age, 54 years old, 48 males and 34
females). The mean sodium level was 118 mmol/l (the
lowest 104 mmol/l). There were 45 cases with consciousness disturbance and 33 cases with seizure. Fifty-three
cases were acute and 29 chronic. Underlying etiology
included water intoxication, SIADH and hypoadrenalism.
Death occurred in four cases and all were caused by
underlying diseases. Thirty-five cases were treated with
isotonic saline, 15 cases with hypertonic saline, and 32
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
cases with no active fluid treatment. Although correction
rate for hyponatremia was not rapid (<0.6 mmol/l/h) for all
cases, there were no cases of mortality and neurological
sequelae from hyponatremia.
sary for treatment of symptomatic hyponatremia just
because the serum concentration is extremely low.
Conclusions: Slow correction of serum sodium was not
complicated with permanent brain damage from acute
symptomatic hyponatremia. Rapid correction is not necesP177
Sterns RH: Severe symptomatic hyponatremia: treatment and
outcome. Ann Intern Med 1987, 107:656–664.
Ayus JC, Krothapalli RF, Arieff AI: Treatment of symptomatic
hyponatremia and its relation to brain damage. N Engl J Med
1987, 317:1190–1195
Variation of early auditory evoked potentials (EAEP) in severe hyponatremia
C Dusson, T Van der Linden, P Cabaret, X Fournet, P Gallois, G Forzy, YM Robin and B Lepoutre
Intensive care unit, CH Saint-Philibert, LOMME 59462, France
Severe hyponatremia induces neurological symptoms in
connection with its rate of appearance. The aim of this
study was to measure the variations of EAEP during
hyponatremia and to determine its prognostic interest.
Patients and method: 65 patients (44 women and 21
men, mean age 67.3±15.7 years) with hyponatremia
<125 mmol/l at their admission to the emergency room or
intensive care unit, were included. The EAEP (monaural
stimulation by alternative clicks of 0.1 ms at 90 Db) were
recorded daily until hyponatremia correction was achieved
(250 records). EAEP parameters (I, III and V peaks
latency, I–III, III–V and I–V intervals) were studied according to the level and the mechanism of hyponatremia, and
according to the neurological symptomatology and survival of the patients.
Results: The mean value of the initial natremia was 118±7
mmol/l, by dilution in 78% of cases, with neurological
signs in 69%. Correction speed was 4.8±2.2 days.
Fifteen patients died but no cases of centropontine myelinolysis occurred. There was a linear correlation between
natremia decrease and brainstem conduction lengthening
(III and V peaks, intervals, P<10-4). These variations were
corrected by the normalization of natremia (P<10-5).
Hyponatremia was associated with a lengthening of the
I–V interval in 74.5% of patients. The mechanism of the
hyponatremia, the presence of neurological signs or the
patient outcome had no effect on EAEP parameters.
Conclusion: Severe hyponatremia must be considered a
cause of EAEP lengthening. EAEP evaluation does not
constitute a prognostic factor of hyponatremia.
Combined central diabetes insipidus and cerebral salt wasting after craniotomy
TF Loh, WRW Lee, YCK Goh and WT Seow
Department of Paediatric Medicine and Department of Neurosurgery, KK Women’s and Children’s Hospital, 100 Bukit Timah Rd,
Singapore 229899
Introduction: Sodium disturbances and polyuria in children after craniotomy for intracranial lesions are not
uncommon. Diabetes insipidus (DI) of central origin is
often cited as a cause of high serum sodium [1] while
cerebral salt wasting (CSW) is a recognised cause of
hyponatraemia in these patients [2]. DI and CSW can
both result in polyuria; patients with central DI suffer from
free water loss while patients with CSW have severe salt
loss. Both syndromes may occur at various times in the
postoperative period. A combination of the above mentioned diseases could occur in the clinical setting and this
can lead to difficulty in diagnosis and cause problems in
maintaining normal fluid and sodium status. Proper differentiation is essential, as the treatment for each entity is different. CSW and DI in the same clinical setting have not
been previously described in the paediatric age group.
We report two cases of combined DI and CSW in the
immediate postoperative period.
MRI of the brain showed a large suprasellar tumour with
hydrocephalus. A ventriculo-peritoneal shunt was inserted
immediately. She needed two operations to completely
excise the tumour. Histology showed pilocytic astrocytoma.
Postoperatively, she developed evidence of central diabetes
insipidus and required intravenous pitressin before urine
output was successfully controlled. She remained stable till
the fifth postoperative day (POD) when she developed
hyponatraemia that was very resistant to treatment despite
hypertonic saline replacements. She was also clinically dehydrated. Urine sodium was 296 mEql/l/l and FeNa was 10.8.
With aggressive saline replacements, hyponatraemia was
corrected and kept within normal limits on a regimen of
enteral feeds and intravenous saline replacements according
to urine output and sodium measurements. Fludrocortisone
was started. After tenth POD, urinary sodium began to
decline consistently below 150 mEql/l by the fifteenth POD
and less than 70 mEql/l by a month POD.
Case reports: Case 1. CXH is a 10 year old girl who presented with 1 month’s history of visual and gait disturbances.
Case 2. NWS presented at 1.5 months of age with rhinorrhea and nasal mass. MRI showed a nasal mass with
intracranial extension. Consent for surgery was not given
until a year and a half later. By that time, a repeat MRI
showed the tumour had extended into the left orbit and
ethmoid sinus. A lateral rhinotomy, craniotomy with tumour
excision and craniofacial reconstruction was performed.
Histology revealed a low-grade nasal glioma. Postoperatively, he developed central diabetes insipidus requiring
intravenous pitressin before urine output was controlled.
However, serum sodium continued to drop even though
urine output did not change significantly. On the second
POD, the child developed a generalised tonic clonic
seizure that was aborted with intravenous Valium and
dilantin. Computer tomography scan of the head showed
residual tumour in the left orbit and suprasellar region, evidence of CSF leak but no haemorrhage or cerebral
oedema. Central venous pressure had dropped to
+3cmH2O. Serum and urine sodium was 117 mEql/l and
176 mEql/l. FeNa was 9.6. Hyponatraemia was corrected
gradually with normal saline and hypertonic saline replacements. However, urinary sodium levels continued to rise
over the ensuing days, reaching a peak of 295 mEql/l on
the sixth POD and there was massive solute diuresis. He
required aggressive replacements with hypertonic saline
to keep serum sodium within limits. However, urinary
sodium persisted between 120–135 mEql/l thereafter and
on the tenth POD, fludrocortisone was started at
10 µg/kg/day. Subsequently, urinary sodium dropped to
56 mEql/l two weeks after surgery and remained below
30 mEql/l after three weeks post surgery. Intravenous
pitressin was successfully weaned off and the patient
started on intranasal DDAVP.
Discussion: CSW syndrome is characterised by hyponatraemia (<130 mEq/l), dehydration, and inappropriate
urinary sodium loss that responds to fluid and saline
replacements. Osmotic diuresis often accompanies this
syndrome. The main feature of central DI is massive diuresis of dilute urine with low sodium content thus resulting in
hypernatraemia and dehydration. Both clinical states are
characterised by diuresis but in CSW, hyponatraemia is
present while in DI, the patient experience hypernatraemia.
In the neurosurgical intensive care setting, ensuring
sodium homeostasis is important to maintain an environment without major flux in osmolality. CSW may develop in
patients with an established diagnosis of DI. Early recognition can be achieved by measuring osmolality and sodium
in the plasma and urine as well as plasma ANP levels. The
cornerstone of CSW management remains replacing
sodium and water loss with normal or hypertonic saline. In
our experience, mineralcorticoid therapy appears to be an
effective adjunct [3].
Lugo N et al.: J Ped Endo Met 10:633–639.
Harrigan MR et al.: Neurosurgery 1996, 38:152–160.
Ishikawa S et al.: Ann Int Med 1987, 106:187–191.
Changes in the C-type Natriuretic peptide (CNP) and NO derivatives in the plasma and CSF of the
patients with subarachnoid hemorrhage (SAH)
K Ikeda, T Ikeda, T Onizuka, H Terashi, T Nakanishi*, T Fukuda* and S Endo†
Division of Critical Care and Emergency Medicine and *Department of Neurosurgery, Hachiouji Medical Center of Tokyo Medical
University; †Department of Emergency Medicine, Iwate Medical University, Japan
The aim of this study was to determine a role of CNP and
NO for cerebral vasospasm following subarachnoid hemorrhage (SAH). Both of them are endothelium-derived
relaxing factor (EDRF), and CNP exists in the brain as a
cerebral vasodilator. We conducted a one week monitoring of CNP and NOx levels in the plasma and cerebrospinal fluid (CSF) in 26 patients who received clipping
operation within 24 h after the occurrences of SAH, and
classified to the Group A (angiographical spasm +), and
Group B (angiographical spasm –). And we examined
CNP and NO‚ levels in the CSF of the patients who
received spinal anesthesia for a small operation as a reference patients.
Results: CNP levels in the CSF on day 1 of hospitalization
was higher than that of the reference patients and the
level decreased significantly on day 7, but we did not
observe any significant difference between the groups.
CNP levels in the plasma did not change.
NOx levels in the CSF in the week following SAH were
higher than the levels of the reference patients, and the
level in Group A was higher than in Group B on day 1.
Only in Group A did NOx levels decrease significantly
from day 3. NOx levels in the plasma did not change significantly.
Conclusions: CNP levels in the CSF were high in the
acute phase following SAH, while plasma CNP levels
were maintained at a constant. But any specific role of
CNP for cerebral vasospasm was not indicated from our
We considered that NOx in the CSF has a role as an
inhibitor for cerebral vasospasm following SAH, because
NOx levels decreased significantly only in the vasospasm
group, while plasma NOx levels did not change significantly.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Acute weakness in the ICU–electrophysiological and pathological findings
J Manák*, Z Lukᚆ, M Schreiber‡, P Kunc‡, H Matulová‡, F Pára‡, L Sobotka* and Z Zadák*
*Dept. Metabolic Care, University Hospital Hradec Králové, Czech Republic; †Dept. Pathology II, Childrens Hospital, University
Brno, Czech Republic; ‡Dept. Neurology, University Hospital Hradec Králové, Czech Republic
Objective: To describe electrophysiological and histopathological findings in patients with acute weakness,
which developed in the ICU during the course of critical
Materials and methods: During a 15 month period, all
patients with clinical signs of acute onset quadruparesis
were studied prospectively. In nine critically ill patients of
mean age 55 (21–78) years, clinical examination, electrophysiological studies (conduction studies of peripheral
motor and sensory nerves, needle electromyography,
repetitive stimulation) and skeletal muscle biopsies were
performed. None of the patients had any neuromuscular
abnormality in anamnesis.
Results: Clinical examination: all of our patients had signs
of quadruparesis and generalized hypo- or areflexia of
tendon reflexes and prolonged dependence on the ventilator. Electrophysiology: compound muscle action potentials (CMAP) were markedly reduced (mean 18±15% of
normal values) or totally absent in all patients. Sensory
nerve action potentials (SNAP) were reduced or absent in
six patients, only two patients had normal SNAP. Motor
nerve conduction velocities were normal or slightly
reduced in six patients, two patients had significantly
lower motor nerve conduction velocities. Sensory nerve
conduction velocities were normal, there was only slight
reduction in one patient. Needle electromyography
showed fibrillations and positive sharp waves in eight
patients. No pathology in repetitive stimulation was found.
Histopathology: all skeletal muscle biopsies showed
pathologic alterations. Fiber diameters varied from atrophy
to normal. In muscle fibers, internal nuclei, fiber splitting
and focal regressive changes with increased activity of
lysosomal acid phosphatase were found. In 50% of cases,
regenerating fibers were present. Endomysial fibrosis was
common. All the changes were classified as myogenic, no
convincing signs of neurogenic atrophy were noticed. One
patient revealed disuse atrophy.
Conclusions: We found electrophysiologic signs of
axonal sensory-motor polyneuropathy in all our patients. In
spite of that, no neurogenic changes in the muscle were
found in the biopsies, the changes in skeletal muscles had
a myopathic pattern. These data suggest that in acute
quadruplegia of the critically ill, both the nerve and the
muscle are affected simultaneously.
Acknowledgement: This study was supported by grant IGA MZ CR No.
NB 5197-3.
A prospective study of the incidence of critical illness polyneuropathy
V Segura, ME Arteaga and F Delgado
Unidad de Medicina Critica, Departamento de medicina, Hospital Medico Quirurgico (ISSS), Alameda Juan Pablo II, San
Salvador, El Salvador
Introduction: Critical illness polyneuropathy (CIP) has
been associated with sepsis and multiorgan dysfunction
syndrome (MODS). Studies have reported an incidence of
CIP that oscillated among 21–81.8% [1,2,3].We conducted this study to assess the incidence of CIP in our
intensive care unit.
Setting: Polyvalent ICU.
Method: During the period of 05/01/1999 to 10/30/1999
359 patients entered our unit, from this group 154 needed
mechanical ventilation for more then five days. The
APACHE II score and GORIS score of multiorgan failure
was evaluated. When they were transferred from the unit
EMG study was realized according to the protocol published before [1].
Patients with a history of polyneuropathy, chronic renal
failure, diabetes mellitus, chronic alcoholism and thus
patients with polytraumatism of lower extremity, in which it
was not possible to realize the study, were excluded.
Results: See Table.
Our results show a smaller incidence (6.6%) of CIP than
other studies [1,2].These studies had a major proportion
of patients with neurotrauma and polytrauma and our
sample was only 13%. Nevertheless, we can observe a
score of FMO larger than or equal to 5, and a more
extended mechanical ventilation tim,e which can also
explain the difference that was found.
Leizten FSS, De Weere AW: Critical illness polyneuropahy in multiorgan dysfunction syndrome and weaning from the ventilator.
Intensive Care Med 1996, 22:856–861.
Mohr M: Effects of early treatment with inmunoglobulin on critical
illness polyneuropathy following multiple organ failure and gran
negative sepsis. Intensive Care Med 1997, 23:1144–1149.
Berek K, Maugreiterj: Polyneuropathy in critically ill patients a
prospective evolution. Intensive Care Med 1996, 33:849–855.
Mechanical ventilation
Segura V
4 months
14.1 days
Berek K
1 year
Leizten F
18 months
33.6 days
2 years
23 days
Neurosurgical patients admitted to intensive care units (ICU) exhibit a systemic inflammatory
DF McAuley†, K McCallion, DW Harkin, GE Brown*, PJ Erwin, G Lavery†, MI Halliday and KR Gardiner
Departments of Surgery, Queen’s University of Belfast, Belfast, UK; *Beth Israel Deaconess Medical Center, Boston, USA;
†Intensive Care Unit, Royal Victoria Hospital, Belfast, UK
Methods: Over a period of 3 weeks, all neurosurgical
patients (elective and emergency) admitted to a regional
ICU had blood taken on the 1st and 3rd morning after
admission. Blood was also taken from controls. Neutrophil
(PMN) respiratory burst activity was assayed over 17.3
min using a BioOrbit 1251 Luminometer to detect diluted
(1:20) whole blood chemiluminescence (CL) in the presence of luminol. Circulating and maximal CL were measured by stimulation with phorbol 1,2-myristate
1,3-acetate in the absence or presence of tumour necrosis factor alpha (TNF) respectively (Fig. 1). The peak
signal for each was obtained and the maximal value dived
by the circulating value. This ratio represents the capacity
to increase circulating respiratory burst activity. Soluble
p55 TNF receptor (anti-inflammatory marker) and interleukin 6 (IL6, pro-inflammatory marker) were measured.
Clinical parameters were recorded.
Results: Median [interquartile range, IQR] APACHE II
scores were 22 (17–26) and 24 (11–27) on days 1 and 3
CL (mV)
Aims: Neurosurgical patients are at risk of multi-organ
dysfunction. It was hypothesised that those patients
requiring ICU admission would exhibit evidence of a systemic inflammatory response.
Time from stimulation (minutes)
Example of chemiluminescent assay from one patient on day 3.
CL Ratio
p55 TNF (ng/ml)
IL6 (pg/ml)
Day 1 (n=15)
2.2 [1.4–2.4] 7.7 [6.6–10.7] 270.1 [81.5–388.1]*
Day 3 (n=7)
4.0 [1.5–4.0]† 9.0 [6.3–12.8] 216.9 [36.0–294.0]†
Control (n=24)
2.2 [1.5–2.2] 7.8 [5.0–12.6]
11.0 [11.0–11.0]
Analysis by Mann-Whitney U. *P=0.002, day 1 vs control; †P<0.05,
day 3 vs control.
Conclusion: Patients admitted to ICU with neurosurgical
insults have raised levels of pro-inflammatory cytokines
with an increased capacity for PMN activation. Assays of
PMN activation can be used as a measure of the balance
between pro- and anti-inflammatory mediators.
Bacterial versus viral meningitis: comparison of the old and the new clinical prediction models
Y Tokuda
Department of Medicine, Okinawa Chubu Hospital, Gushikawa city, Okinawa, Japan
Introduction: Accurate initial diagnosis is the cornerstone
for therapeutic decision making of acute bacterial meningitis (ABM). A previously reported statistical model based
on a combination of four parameters (total polymorphonuclear cell count in cerebrospinal fluid (CSF), CSF/blood
glucose ratio, age and month of onset) appeared effective
in differentiating acute viral meningitis (AVM) from acute
bacterial meningitis in western countries. The objectives of
this study were to validate this model on a independent
sample of patients with acute meningitis seen in Okinawa,
a tropical region of Japan, and to build a new model based
on this sample.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Methods: Retrospective review was performed for
medical records of all persons aged more than 15 years
for the management of community-acquired acute meningitis treated at a our hospital between 1985 and 1998.
The criterion standard for bacterial meningitis was a positive CSF or blood culture. For viral meningitis, it was a discharge diagnosis of viral meningitis with no other etiology
evident. A new prediction model was developed by using
stepwise logistic regression analysis.
Results: Forty-five cases of bacterial meningitis and 101
cases of viral meningitis were confirmed. The discriminatory power of the old model as measured by the area
under the receiver operating characteristic curve (AUC)
was 0.695 (95% Confidence Interval or CI 0.545–0.845).
The best representative model selected four slightly different independent variables: age, mental confusion, neck
stiffness, and total CSF polymorphonuclear cell count. The
area under the ROC curve was 0.986 (95% CI
0.949–1.0), and significantly better than the old (P<0.05).
Conclusion: In differential diagnosis of acute bacterial
and viral meningitis, the old prediction model may not be
clinically useful when applied to a geographically distinct
population in a different ethnicity. Older age, mental confusion, neck stiffness, and higher CSF polymorphonuclear
cell count were identified as the most accurate logistic
model predicting the likelihood of bacterial meningitis on
our patients.
Spanos A, Harrel FE, Durack DT: Differential diagnosis of acute meningitis. JAMA 1989, 262:2700–2707.
McKinney WP, Heudebert GR, Harper SA, Young MJ, McIntire DD: Validation of a clinical prediction rule for the differential diagnosis of
acute meningitis. J Gen Intern Med 1994, 9(1):8–12.
Protection of spinal cord ischemic injury with the β-agonist Clenbuterol
MC DaSilva and BY Lee
New York Medical College, WCMC, Valhalla, NY, USA
We studied Clenbuterol, a selective β-2 agonist, in preventing ischemic injury to the spinal cord.
Methods: We used 30 New Zealand rabbits for their
unique segmental arterial supply to the spinal cord. Fifteen
animals in the study arm were given Clenbuterol 9 mg PO
24 h before clamping. Fifteen (control arm) received water
only. Heparin (70 IU/kg) was given 5 min before the incision. Ischemic injury was reproduced by cross-clamping
the infrarenal aorta for either 22 or 30 min. Direct aortic
blood flow measurements were taken with the transonic
flowmeter before and after cross-clamping. Paralysis was
assessed by the Tarlov’s neurological scale.
Results: All animals (2 control, 2 study arm; n=4) subjected to a clamping time of 30 min developed complete
paraplegia. Of the control group with 22-min clamping
time 77% developed paraplegia (Tarlov 0, n=9; Tarlov 1,
n=1) and 23% did not (Tarlov 3, n=1; Tarlov 4, n=2). In
the experimental (Clenbuterol) group with 22-min clamping time 38% developed paraplegia (Tarlov 0, n=3; Tarlov
1, n=2) and 62% did not (Tarlov 3, n=2; Tarlov 4, n=6). A
consistent observation was that animals which developed
paraplegia had a marked increase in post-declamping
ß-agonist Clenbuterol
Control with Water
22-minute Clamping Time
22-minute Clamping Time
blood flow, i.e., hyperperfusion, suggesting that variations
in aortic blood flow post-declamping may play a role in the
development of paraplegia.
Conclusion: We concluded that Clenbuterol had a protective effect on the spinal cord in 62% (experimental group)
vs 23% (control) in animals subjected to a clamping time of
22 min. Furthermore, controlling variations in aortic blood
flow post-declamping may help to prevent paraplegia.
Therapy of malignant intracranial hypertension by lumbar cerebrospinal fluid drainage
E Münch*, P Vajkoczy†, C Bauhuf†, M Quintel* and P Schmiedek†
*Departments of Anaesthesiology and Intensive Care Medicine and †Neurosurgery, Faculty of Clinical Medicine Mannheim,
University of Heidelberg, Mannheim, Germany
Objectives: The use of lumbar drainage in uncontrollable
intracranial hypertension is considered to be contraindicated due to the possibility of inducing transtentorial or
tonsillar herniation. The objective of this study was to
investigate the effect of controlled lumbar CSF drainage
on ICP and to evaluate the potential side effects of this
new therapeutic modality in adult patients with therapyrefractory intracranial hypertension.
Methods: Twenty-three patients (age 36.09±17.4 years)
with severe traumatic brain injury (TBI) (n=12) or
vasospasm after subarachnoid haemorrhage (SAH)
(n=11) were included prospectively. Patients were considered for lumbar CSF drainage if they suffered persistent
intracranial hypertension (>25 mmHg) for a period of more
than 15 min and failed to respond to high intensity treatment. Lumbar CSF drainage was not performed in
patients with tight basal cisterns. After institution of the
lumbar drain, 5–20 ml CSF were initially aspirated and
then continuous CSF drainage was maintained. ICP and
CPP before and after bolus-aspiration were documented.
The neurological outcome of the patients was scored
according to the Glasgow Outcome Scale (GOS) 6
months after injury.
Results: Lumbar drainage was installed 4.8±3 days after
TBI/SAH and the average period of drainage was 6.6±5.1
days. All patients showed an immediate and lasting
decrease of ICP (–17.4±11.4 mmHg; P<0.0001) and an
increase of CPP (+14.3±10.9 mmHg). Ten patients
(44%) showed a favourable outcome (good recovery or
moderate disability), 4 patients (17%) survived with a
severe permanent neurological deficit, 1 patient (4%)
remained in a persistent vegetative state and 8 patients
(35%) died.
Conclusions: Controlled lumbar CSF drainage reduces
therapy-resistant intracranial hypertension significantly.
The hazard of transtentorial or tonsillar herniation might be
limited by considering lumbar drainage only in the presence of discernible basilar cisterns.
Central neuroaxial blockade improves case-mix adjusted mortality of the critically ill surgical
AJ Prabhu, KR Burchett and MC Blunt
Queen Elizabeth Hospital, Gayton Road, King’s Lynn, Norfolk, PE30 4ET, UK
Introduction: Central neuroaxial blockade is known to
provide good postoperative analgesia, hastens the return
of gut function and may attenuate the ‘stress response’ to
surgery [1]. However, there may be haemodynamic instability in critically ill patients. Despite improvements in
many physiological parameters, there is little published
evidence of improvement in outcome following surgery
with such a block. Comparison of mortality outcome in
intensive care units must take into account differences in
the case-mix of patients in order to be meaningful. We
investigated the effect of neuroaxial blockade on mortality
standardised for case-mix using the APACHE 2 prognostic indicator (SMR) [2].
Methods: 205 adult patients admitted to the ICU for >8 h
following major abdominal surgery were categorised
according to whether they had received neuroaxial blockade (block: 91; no-block 113) in a retrospective contemporaneous cohort-controlled design. APACHE 2 scores,
calculated risk of death, types of surgery, ICU lengths of
stay and hospital length of stay were collected from the
ICU database, and SMRs were calculated for each group.
Demographic data was assessed using ANOVA and SMR
by Poisson distribution.
Results: There was a significant improvement in SMR in
the block group (P<0.01). The APACHE 2 scores and calculated risks of death were significantly lower in this
group. Both hospital and ICU length of stay were significantly shorter in the block group. There were significant
differences in the type of surgery performed between the
two groups (Table).
No Block
15.0 (13.9–16.0)**
16.7 (15.5–17.9)
Risk of Death (%)
19.1 (16.1–22.0)***
27.6 (23.8–31.6)
1.3 (0.4–13.9)**
1.9 (0.4–28.7)
12.5 (1–107)*
17 (1–95)
0.46 (0.23–0.92)**
1.41 (1.05–1.90)
ICU length of stay (days)
Time to hospital
discharge (days)
Standardised Mortality
Ratio (95% CI)
Data presented as mean (95% Confidence intervals) for age, APACHE
and risk of death; median (range) for time data. *P<0.05; **P<0.01,
Conclusions: Within the methodological restrictions of
the non-randomised design this study demonstrated an
improvement in outcome in critically ill patients who
receive neuroaxial blockade. The use of SMR as the
primary end-point ensures a meaningful comparison of the
groups despite the obvious difference in severity of illness
between the groups as it allows outcome comparison for
critically ill patients against a previously defined standard.
Buggy DJ, Smith G: BMJ 1999, 319:530–531.
Rowan KM, Kerr JH, Major E, McPherson K, Short A, Vessey MP: BMJ
1993, 307:972–977.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Alteration of current perception threshold in severely injured patients
M Miyake, N Hitosugi, K Ikegami, H Inoue, Y Sato and I Sato*
Trauma and Critical Care Center, Koshigaya Hospital, Dokkyo University School of Medicine; *Dept of Anesthesiology, 2-1-50
Minamikoshigaya Koshigaya-shi, Saitama, Japan 343-8555
Background: Measurements and assessment of current
perception threshold with Neurometer in patients with
chronic pain is well established. However, pathogenesis of
acute pain in severely injured patients is not fully understood.
Purposes: To measure current perception threshold in
severely injured patients with and without intravenous fentanyl and to determine if pain threshold is increased with
sufficient analgesia.
Patients and methods: Twenty severely injured patients
with normal mental status were studied. Patients were
given bolus injection of fentanyl (1 g/kg) upon arrival at the
emergency room. After resuscitation and surgery, patients
were transferred to ICU and given continuous intravenous
fentanyl started at 50 g/h and increased up to 150 g/h on
patient’s demand. Pain threshold (PT) was measured with
a Neurometer. VAS (visual analogue scale) and VRS
(verbal rating scale) were also measured before and after
bolus injection of fentanyl and during ICU care (every hospital day until seventh).
Findings: Before bolus injection of fentanyl, PT of the
patients was significantly lower than that of control (106
mA vs 221 mA; P<0.01). After continuous intravenous
fentanyl in ICU, VAS decreased significantly. However, no
change was observed in PT of the patients during continuous analgesia with fentanyl.
Conclusions: Severely injured patients were more sensitive to nociceptive stimuli than normal healthy controls.
Fentanyl is an effective analgesia in trauma patients and
decreased VAS significantly in the patients. But pain
threshold itself did not show any change even with sufficient analgesia.
The early management of pain in casualty wards
F Van Trimpont and M Genard
Ambroise Pare University Hospital, Mons, Belgium
Objectives: Pain is often the main complaint of patients
coming into casualty wards but its management is rarely
initiated early. Pain management must be considered as
an aspect of the general management of the patient
coming into the emergency ward [1,3]. It is consequently
essential to relieve the pain quickly while continuing the
diagnosis [2]. The treatment of pain is a progressive
process and there should be no hesitation in combining
different drugs with a maximum analgesic benefit while
controlling the analgesia in order to avoid secondary
effects. The following protocol will allow rapid and reassuring analgesia.
Method: Upon arrival in the ward, the adult patient is dealt
with by a nurse who evaluates the analgesic need using
the Visual Analogue Scale (VAS). After the consent of the
doctor by signing a standard protocol, proparacetamol (2
g IV) and diclofenac (75 mg IV) are administered every 6
and 12 h respectively, irrespective of the VAS level. If the
VAS is greater than 3 in the half hour following the initial
administration of proparacetamol and diclofenac, the analgesia is completed by subcutaneous morphine (the dose
is defined on the written protocol in relation to the weight
and age of the patient). This administration must be
repeated with an interval of 4 h if the VAS is greater than
3. The contraindications for the different drugs must be
respected: chronic renal insufficiency, arterial hypertension with heart failure and gastric ulcers for diclofenac;
hepatic insufficiency, atopy, nasal polyps, asthma and
eczema for proparacetamol; chronic respiratory insufficiency and drug addiction for morphine.
Results: The cases collected (n=200) show a demographic equivalence (45% women, 55% men) and an
average age of 49 years. The distribution of the cases was
mainly orthopedic (72%), with renal colic representing
9.6%. Other less common indications made up the
remaining 18.4% (neuralgia, arthritis, knife wounds, gout,
colitis). The Table shows the VAS over time, from admission up to 28 h (for 63 patients).
Standard deviation Quantity
% VAS >3
Hour 0 (H0)
The average VAS on admission was 6.02±2.06.
Conclusions: Using a simple standardized method this
protocol allows effective and early management of pain
with combined analgesia in a casualty ward.
Monitoring of sedation with BIS Index, comparison with Ramsay and Cook sedation scores
R Šplechtna, L Pokorný, E Hušková and D Nalos
ICU, Masaryk Hospital, Pasteurova 9, 401 13 Ústí nad Labem, Czech Republik
Introduction: Sedation and analgosedation are an integral
part of intensive care medicine. The level of sedation is
often very important for the final result of the treatment.
Different sedation scores are used to asses this level
(Ramsey, Cook, etc.). But these scores can be very subjective, dependent on a observer and can show quite
great interindividual differences. The Bispectral index
(BIS) is a processed EEG parameter and shows continuously the level of inhibition of brain function during application of drugs for anesthesia and sedation.
Monitoring of all patients by the monitor ‘Aspect A–1000
TM’. The values of the BIS index were recorded to protocol
at any time of change in the level of sedation. At the same
time this level of sedation was evaluated by two physicians
(to prevent interindividual differences) by the use of
Ramsay and Cook scores. All 75 examinations were
recorded to study protocol. Results were assessed by help
of means (median) + SD (range) and graphs.
We have tried to find if there is any relation between the
value of the BIS index and the Ramsay or Cook score. In
the case of a positive result, there can be a chance of
exact titration of drugs and possible economical profit.
Conclusion: Figures 1 and 2 demonstrate that a narrow
span of BIS index values includes a very wide scale of a
depth of sedation and analgosedation assessed by
Ramsay and Cook scores. These results and especially
our experiences with continual monitoring show that to
rely only on BIS index or to titrate drugs according to
these values is impossible. We admit that bigger number
of patients is needed for better evaluation.
Methods: The group of 20 patients hospitalized in our ICU
(October 1998–September 1999). Including criteria–GCS
15 and the need of continual sedation or analgosedation.
Results: See Figs 1 and 2.
Figure 1
Figure 2
BIS index
Bis index
Berthier F, Potel G, Le Conte P, Tonze MD, Baron D: Comparative
study of methods of measuring acute pain intensity in an emergency department. AM J Emerg Med 1998, 16:132–136.
Berthier F, Le Conte P, Garrei P, Potel G, Baron D: Analyse de la
prise en charge de la douleur aiguë dans un service d’accueil et
d’urgence. Réan Urg 1998, 7:281–285.
Blettery B, Bhrahim L, Honart D, Aube H: Les échelles de mesure de
la douleur dans un service d’accueil des urgences. Réan Urg
1996, 16:691–697.
Ramsay score
4 5 6 7 8 9 10 11 12 13 14 15 16 17
Cook score
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Bispectral index (BIS) monitoring of ICU patients on continuous infusion of sedatives and
paralytics reduces sedative drug utilization and cost
LJ Kaplan and H Bailey
Medical College of PA Hospital, Departments of Surgery and Emergency Medicine, Division of Trauma and Critical Care,
3300 Henry Avenue, Philadelphia, PA 19129, USA
Purpose: To determine whether BIS-monitored titration of
sedatives in ICU patients on continuous infusions of sedatives and paralytics was cost effective and reduced the
incidence of the recall phenomenon.
Methods: Four consecutive months of patients in a SICU
on continuous infusions of paralytics and sedatives were
studied. Patients with severe brain injury (GCS ≤8) were
excluded. In months one and two, sedatives were titrated
to patient comfort guided by vital sign changes after stimulation. Months three and four titrated sedative infusions
to a BIS value between 70 and 80. Paralytic agents were
titrated to a train-of-four response (2/4 at 50 mA) on a
protocol. Demographic data, drug usage, as well as drug
cost were abstracted. On ICU discharge, patients were
queried about recalling pain or fear while they were on
continuous infusions of paralytics and sedatives. Data are
means ±standard deviations and were compared by
unpaired t-test or χ2 as appropriate; significance assumed
for P<0.05.
Results: During months one and two, the cost per patient
(n=31) for sedatives was $819±2.045. Patient diagnosis
Conclusion: Bispectral index monitoring of ICU patients
on continuous infusions of paralytics and sedatives
reduces sedative drug cost as well as the recall phenomenon. Significant under-sedation may occur using subjective analysis of sedation in the ICU.
The change in Bispectral Index with stimulation indicates depth of sedation in intensive care
AT Dewhurst, S Chieveley-Williams and J Goldstone
Centre for Anaesthesia, UCL Hospital, London W1N 8AA, UK
Introduction: Bispectral Index (BIS) is an EEG parameter
which gives a continuous reading between 0–100
depending on depth of anaesthesia. It has been used in
ICU as a monitor of sedation [1]. Clinical assessment of
sedation requires the patient to be stimulated, which may
cause elevation of the BIS. We proposed that this change
in BIS in response to stimulation is important when
assessing depth of sedation. The aim of this study was to
assess the feasibility of BIS as a monitor of sedation in
ICU and to quantify within different sedation groups the
changes in BIS score before and after stimulation.
Method: ICU patients being ventilated and sedated were
studied. BIS was recorded on an Aspect A2000 monitor.
A nurse blinded to the BIS reading assessed sedation
using the observer’s assessment of alertness/sedation
score (OAAS 1–5: 1=no response, 5=fully awake). Sedation was assessed hourly, average unstimulated BIS
(BISu) measured before clinical assessment of sedation
was compared to average stimulated BIS (BISs). Spearman’s correlation coefficient was used to assess association between BIS and OAAS, Mann-Whitney U was used
to compare differences between BIS in OAAS groups
4–5 (lightly sedated), OAAS groups 1–3 (deeply sedated)
Change in BIS with stimulation
10th, 25th, 75th, 90th percentile
was 66% trauma, 22% abdominal sepsis, 9% vascular,
and 5% other. The most frequently used medications and
their percent contribution to total costs were morphine
58% (5.8% cost), lorazepam 54.8% (63.2% cost), midazolam 12.9% (4.2% cost), and propofol 12.9% (24.6%
cost). 18% of the patients recalled a frightening or painful
experience during their period of continuous sedation.
During months three and four, the cost per patient (n=26)
for sedatives was $669±1.362 - an 18% reduction in cost
versus months one and two (P>0.05). Only 4% of patients
recalled frightening or painful events (P<0.05). Demographics were similar compared to months one and two
(P>0.05). Significant reductions occurred in lorazepam
(18%), midazolam (18%) and propofol (47%) usage
(P<0.05). 15.4% of patients who appeared sedated were
under-sedated by BIS analysis, while 54% of patients
required less sedative than initially predicted by the staff.
Sedation Group
and in the change of BIS (∆BIS) with stimulation in all
OAAS groups.
Results: A total of 139 observations in 30 patients were
recorded. There was a significant difference (P<0.05)
between BIS score in the lightly sedated group and
deeply sedated group. Correlation coefficient BISu/OASS
r=0.53, BISs/OASS r=0.73. ∆BIS was larger in OAAS
3–4 than in OAAS 1,2 and 5.
Conclusion: BIS can distinguish between lightly and
deeply sedated patients. BISs correlated more closely
with OAAS than BISu. On stimulation ∆BIS was greatest
in OAAS 3–4. BIS can be used in ICU patients but readings should be interpreted over time and in response to
Simmons L et al.: Critical Care Medicine 1999, 27:1499–1504.
Dexmedetomidine for sedation in the medical ICU
RM Venn, J Ball, A Steele, PJ Newman and RM Grounds
St George’s Hospital, Blackshaw Road, London SW17 0QT, England
The highly selective alpha-2 agonist,
dexmedetomidine, has been shown to be an effective and
safe agent for the sedation of postoperative patients
requiring mechanical ventilation on the ICU [1]. It possesses analgesic and sympatholytic properties and its
ability to sedate without significant respiratory depression
may benefit other patient groups requiring ICU sedation.
This study reports our experience of dexmedetomidine for
the sedation of critically ill medical patients.
Methods: Twelve critically ill medical patients (mean
APACHE II score 21) requiring artificial ventilation,
received dexmedetomidine for up to seven days with
rescue sedation and analgesia provided by propofol and
morphine, respectively, if clinically indicated. Dexmedetomidine infusion was commenced with a loading infusion of
1 µ over 10 min followed by a maintenance infusion
of 0.2–0.7 µg/kg/h to maintain a Ramsay sedation score
≥3. Following experience with the first four patients, the
maintenance rate of infusion could be increased to a
maximum of 2.5 µ Heart rate, systolic and diastolic arterial pressures, central venous pressure and,
where possible, cardiac output by thermodilution were
monitored continuously and recorded at 10 min intervals
for the first 30 min and then hourly.
Results: The mean duration of dexmedetomidine infusion
for all patients was 33 (SD=19.7) h, and mean
dexmedetomidine infusion rate in the latter eight patients
was 1.0 (SD=0.7) µ Only one out of the eight
patients proved difficult to sedate with dexmedetomidine
and required an additional propofol infusion of up to
100mg.h-1. Arterial pressure, heart rate and cardiac output
were well-maintained during dexmedetomidine infusion.
Patients requiring vasoactive drugs prior to commencement of dexmedetomidine showed diminishing requirements during the course of the study. Following
discontinuation of dexmedetomidine infusion there was no
clinically important rebound phenomenon in haemodynamic measurements. Adverse cardiovascular events
(hypotension±bradycardia) were nearly all confined to the
initial loading infusion of dexmedetomidine.
Summary: Dexmedetomidine provides safe, titratable
sedation with cardiovascular stability for critically ill
medical patients requiring artificial ventilation.
Bradshaw CJ, Venn RM, Spencer R: Cardiovascular effects of
dexmedetomidine for ITU sedation: UK results of a multi-centre
study. Critical Care 1999, 3:P233.
Respiratory effects of dexmedetomidine in the ICU
J Hell, RM Venn, R Cusack, A Rhodes and RM Grounds
St George’s Hospital, Blackshaw Road, London SW17 0QT, England
Introduction: The incidence of haemodynamic disturbance and myocardial ischaemia in the ICU can be
reduced if sedation is continued over the peri-extubation
period [1]. Dexmedetomidine has been shown to cause
only minimal ventilatory depression in human volunteers
[2] and also possesses analgesic and sympatholytic properties and so may be a useful agent for this peri-extubation
Methods: Twenty-five patients involved in a randomised,
placebo-controlled, double-blind study were ventilated for
at least 6 h postoperatively and were sedated with
dexmedetomidine or placebo infusions. They were then
extubated whilst receiving these respective infusions. Following extubation, infusions (Dexmedetomidine 0.2–0.7
µg/kg/h) were continued for six hours to maintain a
Ramsay sedation score ≥2. Rescue sedation and analgesia was provided by midazolam and morphine respectively,
if clinically required. Arterial blood gas estimation, oxygen
saturation, respiratory rate and analgesic requirements
were recorded for the peri-extubation period. Results are
reported as mean (SD) and analysed using MannWhitney-U and ANOVA for repeated measures where
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
2 h post-extubation
4 h post-extubation
P value
Resp. rate
Results: Patient demographics and degree of sedation
and analgesia was equivalent in the two groups but the
placebo group required three times more morphine
(P=0.04). The average dexmedetomidine infusion rate was
0.15 µ following extubation. No adverse respiratory events were seen in either group.
6 h post-extubation
Summary: Dexmedetomidine is safe to use in extubated
spontaneous breathing intensive care patients and does
not cause clinically significant respiratory depression.
Br J Anaesth 1998, 80(6):834–836.
Anesthesiology 1992, 77(6):1125–1133.
Lactic acid clearance in the emergency department prognosticates multisystem organ failure and
BP Knoblich, EP Rivers, HB Nguyen, MT Mullen, B Rittinger, G Hays, A. Muzzin, B Sheridan, M Jankowski
and MC Tomlonovich
Departments of Emergency Medicine, Internal Medicine and Surgery, Henry Ford Hospital/Case Western Reserve University,
2799 West Grand Blvd. Detroit, Michigan 48202 USA
Background: Lactic acid (LA) indicates anaerobic metabolism (global tissue hypoxia). While single LA levels are
helpful diagnostically and therapeutically; the change in
LA levels over time (LA-clearance) has better prognostic
value in predicting morbidity or multisystem organ failure
(MSOF) and death in the intensive care unit (ICU).
Objective: To examine the relationship of LA-clearance in
the Emergency Department (ED) and the development of
multisystem organ failure and death.
Methods: This was a prospective case series of critically
ill patients presenting to a large urban ED in shock (systolic blood pressure <90 mmHg for thirty min after a
40 cc/kg volume challenge) or a LA >2 mM/l on arrival.
Patients presenting in hemorrhagic shock, trauma, requiring immediate surgery, do not resuscitate orders, seizures
or end-stage disease not benefiting from ICU care were
excluded. Patients were managed by ED physicians by
standard ED management. LA and Multisystem Organ
Dysfunction Scores (MODS) were obtained on ED arrival,
discharge and every 12 h for 72 h. Patients were grouped
by the lactic acid clearance during the ED stay: Group 1
(no clearance, <0 mM/l/h, actual increase in lactic acid),
Group 2 (intermediate clearance, >0 and <1 mM/l/h), and
Group 3 (high clearance, >1 mM/l/h). Student’s T-test
and ANOVA were used to compare the MODS score and
LA over time, with Bonferroni correction for multiple comparisons.
Results: A total of 142 patients were studied. The mean
age was 65.9±17.2 years. The lactic acid clearances for
groups 1, 2 and 3 were –0.45±0.55, 0.48±0.48,
1.44±0.38 mM/l/h respectively P<0.0001. The mean
MODS scores over 72 h were 8.63±3.35, 6.16±4.23,
5.12±3.76 for groups 1, 2 and 3 respectively P<0.02.
The in-hospital mortality was 50%, 23%, 12% in groups 1,
2 and 3 respectively P<0.045.
Conclusions: The duration of global tissue hypoxia or
decreased LA clearance in the ED is associated with
MSOF and death. This pathogenic link suggests that diagnostic and therapeutic intervention should begin immediately in the ED.
Earthquake, rescued victims, crush injury and their follow up
N Ünal, M Oral, AA Yilmaz and M Tulunay
Department of Anaesthesiology and Reanimation, University of Ankara, Faculty of Medicine, Ankara, Turkey
Thirty complicated crush injury patients were accepted to
our ICU after the Big Marmara (Turkey) earthquake in
August 1999. Results of 18 patients were reviewed and
are presented below.
The mean ages of 10 male and 8 female patients were
31.1±12.6 (range: 15–55) years. Time from earthquake to
salvage and first medical aid were 12.5±9.4 (range:
4–40) and 17.5±10.9 (range: 6–41) h, respectively.
Victims were accepted into our unit 5±8.4 (range: 1–35)
days after the disaster. Their initial problems were acute
renal failure (13), respiratory failure (5), disseminated
intravascular coagulopathy (DIC) (4), bleeding (3), long
bone fractures (2), pelvic fracture (2), lung contusion (1),
haemodynamic instability (1), sepsis (1), in addition to
crush injuries in the lower extremities (14) and upper
extremities (6). Decompressive fasciatomies had been
performed for all patients with crush injuries before acceptance to our ICU. Patients were treated for acute renal
failure (14); DIC (13); sepsis (13) [originating from
wounds (11), urinary system (9) and respiratory system
(7)], respiratory failure (10), hepatic dysfunction (7) and
cardiac dysfunction (3) during their stay in the ICU. Mean
dialysis requirements for patients with renal failure were
5.1±3.6 (range: 3–13) for a period of 10.1±6.8 (range:
6–23) days. Venovenous hemofiltration was performed
twice for two patients and once for four patients because
of volume overload. Fresh frozen plasma, 9.5±10.4
(range: 0–37) units; platelet suspension, 5.2±7.1 (range:
0–21) units; cryoprecipitate, 3±6.6 (range: 0–26) units;
and erythrocyte suspension, 18.6±10.6 (range: 6–39)
units were transfused because of preDIC/DIC and
massive bleeding especially after excision of debris.
Antithrombin III replacement was performed in 8 patients.
Ten patients with respiratory failure were mechanically
ventilated for 9±14.8 (range: 3–55) days. Amputations
were required in six patients.
Eventually two patients with MOF died and 16 patients
were discharged to other clinics with some sequels like
neuropathies (10) and cardiomyopathy (1). One patient
was discharged for hyperbaric oxygen therapy because of
deep soft tissue infection unresponsive to medical therapy.
Experience of medical examination and health consultation for public exposed to neutrons
caused by criticality incident in Japan
Y Haraguchi, Y Tomoyasu, T Satoh, K Neriishi, N Oohashi, M Hoshino and T Arai
National Hospital Tokyo Disaster Medical Center, Tokyo, Japan
A criticality incident occurred September 30th, in Tokaimura
in Japan. Neutron radiation was emitted in a widespread
area. In addition to three patients, who were exposed to
high dose neutron radiation, people who were living around
the incident plant (within 350 m) were urged to take shelter
according to an evacuation order, and people who were
living within 10 km‚ were also recommended to remain
within their houses for three or four days.
We had an opportunity to examine and give advice to the
people who feared the radiation exposure. The experience
is reported.
Results: The criticality incident continued for nearly one
day. From the next day after the end of criticality was
declared, we started health examination. During the consecutive three days, a total of 1852 people were examined. The complete blood count, including the hemogram
or lymphocyte counts, and blood and urine chemistry were
evaluated in all the patients. The survey of radiation contamination was done when patients requested. The history
of the health condition and estimation of the radiated
amount was evaluated by our members (physicians) in all
the patients. Although many people complained of the fear
of the after effects of radiation, the degree of fear was
extremely strong in children and pregnant women. Regarding the results of blood and urine examination, there were
no patients who had abnormalities that were thought to be
caused by high amounts of radiation exposure, although
there were several patients who had decreased lymphocyte counts.
Discussion and conclusions: Consultation only seems
to be justified because, so far as we are informed officially (from the Japanese Agency of Science and Technology), the amount of radiation exposure was within the
safe range. However, several patients who had
decreased lymphocyte counts are recommended to
receive repeated examination of the blood counts.
Mental care and long-term follow-up may be necessary
for people who are fearing the exposure and for children/pregnant woman.
Establishment of emergency medical services in Addis Ababa, Ethiopia
O Benin-Goren*, P Halpern*, S Amir† and MB Tesfay‡
*ED Tel Aviv Sourasky Medical Center Israel; †Sapir Medical Center Israel; ‡Armed Forces General Hospital, Addis Ababa
In February 1998, an Israeli Medical Delegation that
included two ED Directors and one Deputy Head Nurse of
the ED, was founded in order to introduce the emergency
medicine (EM) profession and to help in establishing an
EM system in Addis Ababa. The program was produced in
coordination with an Ethiopian physician who visited the
ED in his fellowship and was very impressed with the
Israeli system.
With the approval of the International Relation of the
Foreign Bureau of the State of Israel, the team traveled to
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Ethiopia to consider the possibility that the program could
be changed according to the needs in the field.
The program consisted of several issues in EM: administrative concepts of ED building; organization of the ED;
the role of Emergency Nurses in the team; trauma and the
treatment of critical care; study and practices of Basic Life
Support; advanced Cardiac Life Support; and Trauma Life
The participants on the program included twenty physicians and ten nurses who work in emergency services.
The first meeting was dedicated to study the existing
system in Addis Ababa. With poor conditions, lack of necessary equipment, medication and resources, and with
lack of basic knowledge, the health system in Addis
Ababa cannot start re-organization without help.
Yet there is a new movement of physicians who are ready
to take responsibility and be part of the change. The Israeli
team gave the recommendation to the authorities in Israel
to continue education among the Ethiopian staff-nurses,
physicians, technicians, and Paramedics.
This paper examines EM in Addis Ababa, the course, and
recommendation for the future. Furthermore, in addition to
the program that has been developed until now, the Israeli
team with the cooperation of some American physicians
plans to continue the program in December 1999 that will
include studies and meetings with policy makers in Addis
A prospective trial of surgical resident performed focused abdominal sonography for trauma
(FAST): an easy training method
DM Bridge, M Bonta, J Blackford, W Wiand, D Giammar and LM Foley
Riverside Methodist Hospitals, 3535 Olentangy River Road, Columbus, Ohio 43214, USA
Introduction: Ultrasound is a quick, easy tool for evaluating trauma patients. A focused abdominal sonogram for
trauma (FAST) is a limited sonographic evaluation easily
learned by those with minimal experience. Our study
attempts to prove that surgical residents of all postgraduate levels can be trained with a short didactic session followed by hands-on experience, studying patients with
known intra-abdominal fluid. With this method, surgical
residents can be quickly and cost-effectively taught to
perform FAST examinations with proficiency rates matching those of the literature .
Materials and methods: In this prospective randomized
trial, thirteen surgical residents were trained to perform
FAST scans. Residents attended a 3-hour didactic course
and then performed three normal FAST scans followed by
four scans on patients with known intra-abdominal fluid.
Lastly, trauma patients over a six-month period were
studied. FAST results were compared to CT, DPL, or
serial examination results. Statistical analysis was performed.
Results: In 93.8% of patients no fluid was detected.
There was one false positive and one false negative exam.
The sensitivity and specificity were 80% and 99%,
respectively. Positive and negative predictive values were
80% and 99%, respectively. A kappa statistic of 0.79 was
calculated indicating a substantial amount of agreement
between the FAST scan group and the CT/DPL/serial
exam group.
Discussion: Sonography guidelines for the training of surgical residents are needed. By maximizing positive examinations, the number of required proctored examinations
can be lowered and the learning curve shortened. In a
short time, surgical residents can learn to perform the
FAST scan with proficiency rates matching that of the literature.
Conclusion: By training on patients with known intraabdominal fluid, in a short time, surgical residents can
learn the FAST scan. With our method, resident proficiency can match that of more experienced personnel.
Data from the U.S. Army field surgical hospital intensive care unit in post-war Kosovo
SM Grosso
FACS, Major, Medical Corps, U.S. Army
Objective:To review the admissions to the intensive care
unit in a military field surgical hospital to determine the
relationships, if any, between demographics, severity of
illness, length of stay, ventilatory duration, and other parameters for operative and non-operative trauma, nontrauma surgical, and medical admissions.
Design: Prospective observational study.
Setting: U.S. Army 67th Combat Support Hospital intensive care unit deployed to a forward field location in the
U.S. Sector in post-war Kosovo in support of the NATO
peacekeeping mission in that region.
Patients: All fifty-five patients admitted to the intensive
care unit from July through November, 1999 were enrolled
in the study.
Interventions: None.
Measurements and main results: Therapeutic Intervention Scoring System (TISS) Scores for each patient were
determined at admission and at discharge. Other data collected included demographic data, including nationality,
age, and gender; diagnosis; length of stay; duration of
ventilatory support, if applicable; and disposition at discharge.
Conclusions: The majority (68%) of patient admissions
were for trauma. Patient age ranged from 1 day to 76
years. Sixty-two percent of patients were 42 years of age
or younger. Mean length of stay was 1.70 days. Seventyone percent of patients were local national (Kosovo)
patients, and 29% were NATO personnel. Mean TISS
scores for trauma, non-trauma surgical, and medical
admissions were 21.67, 17.00, and 9.53, respectively.
There was a positive correlation between TISS score at
admission and length of stay for all patient categories.
Mean delta TISS scores between admission and discharge were 12.14, 8.00, and 5.00 for trauma, non-trauma
surgical, and medical admissions respectively. Patients
who remained on the ventilator for prolonged periods had
higher TISS scores at admission (mean 36.71) than the
population overall (mean 18.23). Length of stay was determined by many variables, including severity of illness, diagnosis, availability of host nation facilities for local patient
transfers, and urgency of evacuation for military and other
NATO personnel.
Management of children with severe traumatic brain injury in an adult ICU: evaluation of
G Tsaousi, A Lavrentieva, D Setzis, Ch Skourtis, E Anastasiou and M Yannakou
Department of Anaesthesiology and ICU, AHEPA University Hospital, St. Kiriakidi 1, 54006,Thessaloniki, Greece
Introduction: Injuries are the leading cause of death in
children and brain injury is the most common cause of
paediatric traumatic death. Shortage of paediatric intensive care units imposes children’s occasional admission in
an adult ICU. We attempted to evaluate the effectiveness
of their management in a nonspecialised environment
(adult ICU).
Methods: Data of 31 children with severe traumatic brain
injury (STBI), admitted in our ICU, were reviewed retrospectively. Cases were analysed according to the following criteria: 1) age in years, 2) GCS (upon admission), 3)
CT- scan grade, 4) injury severity score (ISS), 5) paediatric trauma score (PTS), 6) paediatric risk of mortality
(PRISM III-24h), 7) length of ICU stay in days and 8)
outcome at ICU discharge (Glasgow outcome score GOS). The demographic data of the studied group are
shown in the Table.
Results: 1) Mean age was 6.4 years (SD±5). 2) Trafficrelated accidents were responsible for 67.7% (n=21) of
cases, 29.1% (n=9) were attributed to fall from height and
3.2% (n=1) to gunshot. 3) The average length of ICU stay
was 5.1 days. 4) Eighteen patients (58.1%) fully recovered (GOS 5), 7 patients (22.5%) presented moderate
neurological dysfunction (GOS 3–4) and 6 patients
(19.4%) had a poor outcome (GOS 1) while predicted
mortality according to PRISM III-24h was estimated to be
21.8%. Among those with poor outcome, four cases of
brain death due to secondary brain oedema were
Demographic data *mean ±SD
Group patients (all)
Group patients (deaths)
included. All fatalities were attributed to the severity of
injury (Table). 5) In addition to GCS (r=0.78, P<0.01)
early hints for poor prognosis were acquired from CT findings (r=–0.84, P<0.01) and PTS (r=0.57, P<0.01).
Conclusion: Despite the adverse environment related to
the inherent impediments of an adult ICU, the results of
this study are indicative of a rather efficient treatment of
children with STBI compared to relative outcome results
of specialised paediatric units.
Feickert HJ, Drommer S, Heyer R: Severe head injury in children:
impact of risk factors on outcome. J Trauma 1999, 47(1):33–38.
Ward JD: Pediatric issues in head trauma. New Horiz 1995,
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Outcome after major trauma: 12-month follow-up
A Anthi, S Georgopoulou, A Kalopissi, E Boukouvalas, K Mandragos and P Behrakis
Intensive Care Unit, Red Cross Hospital, Athens, Greece
Objective: The objective of this study was to describe the
functional status of surviving major trauma victims after
management on an ICU.
Patients and methods: During 1998, 86 trauma victim
patients with an average injury severity score (ISS) of 32,
were admitted to the ICU of Red Cross Hospital in
Athens. After ICU treatment, 50 patients survived and
were discharged from the hospital. Follow-up 12 months
later was completed for 41 patients (82%). Functional
capacity was assessed using Rosser's Disability Categories (RDC) .
Results: The mean age of surviving trauma victims was
36±16 yrs. The mean ISS was 25±8 and the mean
length of ICU stay was 11±19 days. Nine patients fell
into Rosser's Disability Category 1, 10 patients into
RDC 2, seven patients into RDC 3, six patients into
RDC 4, five patients into RDC 6, and four patients into
RDC 7. No significant correlation was found between
severity of trauma as expressed by ISS and Disability
Conclusions: 32% of survivors experienced severe social
disability and a modest to severe work limitation (RDC
3,4). 22% of survivors were confined to a wheelchair or
bed (RDC 6,7). 12-month follow-up after major trauma
clearly demonstrates a profound level of functional limitation in more than 50% of patients. The severity of trauma
alone does not seem to determine the functional capacity
of surviving major trauma victims.
The Bair Hugger patient warming system in prolonged surgery
JKC Huang, E Shah, N Vinodkumar, MA Hegarty* and RA Greatorex
Department of Surgery and *Department of Microbiology, Queen Elizabeth Hospital, Gayton Road, King’s Lynn, Norfolk, PE30
4ET, England, UK
Introduction: The Bair Hugger patient warming system is
widely used and the advantages of avoiding hypothermia
for patients undergoing surgical procedures are well
established; however, prolonged exposure to the exhaust
of the warming blanket may force the patients’ skin organisms into the theatre atmosphere, and then into the surgical field and more importantly, onto prosthetic material
such as vascular grafts, increasing the risk of devastating
infection. This danger had not been previously studied.
Methods: Sixteen patients undergoing aortic surgery with
graft insertion were studied. Air and wound specimens
were taken from seven separate sites using standard
methods with the centrifugal air sampler and agar touch
plates. Two readings were taken from each site — one
when the warming blanket was first applied at the start of
the operation, and again at the end of the operation.
Results: The mean duration of surgery and use of the
warming blanket was 234 min (180–270 min). Operating
theatre air showed a significant decrease in colony counts
at the end of surgery. The exhaust air colony counts
remained similar. All wound specimens were sterile. None
of the patients had postoperative wound-related infections.
Conclusions: The use of the Bair Hugger patient warming
system during prolonged surgery does not lead to
increased contamination of the surgical field nor increase
the risk of wound and graft infection.
Induced moderate hypothermia markedly exacerbates pulmonary hypertension and dysoxia in a
neonatal piglet model of elevated pulmonary vascular resistance
MM Zayek, CR Hamm, KT O’Donnell and FG Eyal
Pediatrics, University of South Alabama, Mobile, AL, USA
Objective: To assess the cardiopulmonary effects of
induced mild to moderate hypothermia in a neonatal
animal model of pulmonary hypertension.
Background: Induced hypothermia has been proposed
for neuroprotection following perinatal asphyxia. This latter
condition is often associated with other pathophysiologi-
cal derangements such as pulmonary hypertension,
depressed cardiac function and decreased mesenteric
blood flow. In those instances, hypothermia may become
detrimental because of its potential to further increase pulmonary vascular resistance.
Design/methods: Anesthetized piglets (7±1 days old,
n=27) were randomized to receive a thromboxane A2
mimetic (U46,619, 0.09±.07 µg/kg/min)(TX) or nothing
(CONT). One hour later, both groups were subjected to
mild (34–36°C) or moderate (32–34°C) whole body
hypothermia. Rewarming was started one hour later. At all
times mechanical ventilation was modified to preserve normocapnea, pH >7.30 and Pa02 >100 torr. The following
parameters were measured every 30 min: arterial, pulmonary and superior sagittal sinus blood gases; mean systemic (BP) and pulmonary artery pressures (PAP); cardiac
output (thermodilution)(CO); internal carotid (CBF) and
superior mesenteric (MBF) blood flow (ultrasonic flow transducers); lung mechanics and blood lactate concentration.
Pulmonary (PVR) and systemic vascular resistance (SVR),
alveolar–arterial oxygen gradient (A-aDO2), shunt fraction,
oxygen index (mPaw xFiO2/PaO2)(OI) and oxygen extraction and O2 consumption (VO2) were derived from those
Results: Administration of TX increased (P<0.01) baseline PVR by 137±7%, while CO, MBF and CBF
decreased by 23±4%, 33±1% and 29±7% respectively.
The table shows the mean ±s.e.m % changes from their
respective prehypothermia levels for each group at the
end of the one hour hypothermic period.
Hypothermia caused a decrease in CBF (CONT:
25±8%,TX:44±9%), cerebral (VO2) (23±17% vs
34±11%) and MBF (23±7 vs 29±16%). CONT and TX
were not statistically different. The physiological
responses to mild or moderate hypothermia were similar.
No excessive lactate production was observed.
Conclusions: Induced therapeutic hypothermia may exacerbate the adverse pulmonary hemodynamic effect associated with newborn pulmonary hypertension and lead to
consequent hypoxemia.
% Changes:
↑Shunt F.
21 ± 18
40 ± 18
11 ± 6
42 ±12
26 ± 7
–27 ± 3
98 ± 28
220 ± 70
55 ± 20
27 ± 4
49 ± 7
53 ± 7
–16 ± 3
Accidental hypothermia and active rewarming: the metabolic changes observed above and below
JJ McInerney, B Breakell, W Madira, T Davies and PA Evans
Accident and Emergency department, The Leicester Royal Infirmary, Leicester, LE1 5WW, UK
Background: In accidental hypothermia the underlying
physiological mechanisms responsible for poor outcome
during rewarming through 32°C remain obscure. Deleterious changes in plasma calcium (Ca2+), magnesium
(Mg2+), and acid-base concentrations have been reported
in other forms of tissue injury, whilst parathyroid hormone
(PTH) and interleukin 6 (IL-6) levels are elevated. Accordingly this study investigated the changes which occur in
Ca2+, Mg2+, pH, PTH and IL-6 during rewarming.
Methods: Eight patients, 4 males and 4 females, age 45
to 85 years, admitted with core temperatures <35°C were
included in the study. Patients were rewarmed with dry
warm blankets and fluid replaced by crystalloid at 40°C.
Blood for Ca2+ and Mg2+, pH, PTH and IL-6 were collected at presentation, during rewarming, and at 24 h.
Results: Four patients were admitted with mild
hypothermia (32–35°C) and four with moderate
(28–32°C) hypothermia. All patient clinical parameters
responded favourably to rewarming. Rewarming to 32°C
had no significant effect on the presenting acidosis
(P=0.1740), although above 32°C pH increased with
temperature (P<0.0001). There was a negative correlation between pH and both Ca2+ (P=0.0005) and Mg2+
(P=0.0488) below 32°C; above this temperature the
relationship was significant only for Ca 2+ (P=0.0494).
PTH and Ca2+ correlated positively (P=0.0041) and
negatively (P=0.0039) in moderate and mild hypothermia respectively. IL-6 and Ca2+ also correlated positively (P=0.0039) and negatively (P=0.0018) when
presentation temperature was below and above 32°C
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Conclusions: During rewarming pH remains unchanged
until patient temperature reaches 32°C. Ca 2+ and Mg2+
decline is associated with this pH change. Poor
outcome is associated with presentation temperature
(<32°C), nonphysiological correlation between PTH and
Ca2+, and age (≥84 years). IL-6 levels mirror this PTH
change but further study is required to explain these
Treatment of severe heat stroke in an intensive care unit: clinical course, acute outcome and
implications for management
P Sorkine, O Szold, P Bidermann, S Bulocinic and P Halpern
General Intensive Care, Emergency Department Sourasky Medical Center. 6 Weizmann St. Tel Aviv 64239, Israel
Introduction: During the August 1998 heat wave in Tel
Aviv, many patients were referred to the Tel Aviv Medical
Center for acute heat-related illness. Of these, six patients
were admitted to the Intensive Care Unit for severe heat
stroke. We describe the clinical course during their first
five days of hospitalization, response to treatment and
implications for future management of this disorder.
Results: The six patients were admitted to the Intensive
Care Unit in critical condition (mean APACHE II score
30±3.5), coma (mean Glasgow Coma 3.5±0.5), hypovolemic shock and respiratory failure necessitating mechanical ventilation. Despite early effective therapy (in all
patients core temperature was reduced to less than 39°C
in less than 1 h from admission to the hospital), one
patient died (mortality 15%) and four patients required
further intensive care for life-threatening multiple organ
Conclusion: During a heat wave such as that experienced
in Tel Aviv during August 1998 a significant number of
referrals for acute heat-related illness, possibly overwhelming the admission capacity of regional intensive-care units,
must be anticipated. Severe heat stroke complicated by
multiple organ failure may occur and is not necessarily
related to physical activity prior to symptom onset.
Although important in determining prognosis, early treatment does not prevent severe complications in this disorder. Mechanisms regulating body heat may remain
disturbed for days following early treatment and apparent
stabilization, mandating careful observation of ample duration in hospital.
Severe mushroom poisoning due to amatoxin in children
M Uher, M Pisarcíková, J Filka, L Podracká and M Kurák
University Hospital, Department of Pediatrics, PICU, Košice, Slovak Republic
Objectives: Mushroom poisoning is characterized by a
very variable clinical picture. The amatoxins, the main toxic
component of these fungi, are responsible for gastrointestinal symptoms as well as hepatic and renal failure.
The authors present the clinical picture of the phalloid syndrome and its treatment.
Methods: Within a period of four years, 12 children with
mean age of 10.75 years (4.5–16 years) including 6
males and 6 females with severe Amanita mushroom poisoning were treated at the PICU. The severity of poisoning
was graded according to serum transaminase elevations
and prolongation of prothrombin time. The diagnosis was
based on anamnestic data, the clinical picture and, in the
some patients, on mycological examinations of mushrooms residues, and their spores.
Results: The mean duration of hospitalization in the PICU
was 8 days (3–19). 11/12 patients improved and were
discharged from the hospital asymptomatic, one patient
died. Our protocol for conservative treatment consisted
of: fluid and electrolyte replacement, oral activated char-
coal and lactulose, i.v. penicillin G, all patients received
i.v. thioctic acid and i.v. silibinin, all received a special
diet, and some patients received frozen fresh plasma and
IV Fibrinogen. 3/12 children were given conservative
therapy with hemodialysis and hemoperfusion This combination of treatment modalities was used to accelerate the
elimination of amatoxin from the patients’ bodies. The
mean time between the consumption of the mushrooms
and the first gastrointestinal symptoms was 9.84 h
(4–18) and the mean time between the consumption and
the admission to the hospital (the start of therapy) was
28 h (4–72). The duration of therapy according to the
protocol of conservative therapy was 5.45 days
(3.5–7.5). The increased liver enzymes SGPT and SGOT
were found in 10/12 patients. In 7/12 patients the prothrombin time was prolongated. In 3/12 patients we used
hemoperfusion combined with hemodialysis and in 2/12
Conclusions: We conclude that intensive combined treatment applied in our patients is highly effective in improving
patients with both moderate and severe amanitin poisoning.
Results of transplantation with kidneys from non-heart-beating donors
W Hordijk*, RGWL Tiggeler*, JA van der Vliet† and LB Hilbrands*
*Department of Nephrology, †Department of Surgery, University Hospital Nijmegen “St. Radboud”, P.O. Box 9101, 6500 HB
Nijmegen, The Netherlands
Introduction: During the last twenty years, thirty kidneys
from non-heart-beating donors (NHBD) were used for
transplantation in our hospital. We examined the results of
these transplantations.
Methods: For every recipient from a NHBD kidney, we
selected two recipients who received a kidney from a
heart-beating cadaveric donor, as a control group (60).
The control group was selected by matching the following
items: first or retransplantation, HLA-AB or HLA-DR compatibility, cold ischaemia time, immunosuppressive medication and date of transplantation.
Results: According to “the Maastricht categories” classification from the NHBD resulted in: 1) death on arrival
(n=3); 2) unsuccessful resuscitation (n=2); 3) awaiting
cardiac arrest (n=14); 4) cardiac arrest while brain dead
(n=11). The mean duration of follow-up was 4.5 years.
Graft survival rates at 1 and 5 years after transplantation
were 87% and 57% respectively in the NHBD group and
85% and 72% in the control group (NS). The incidence of
acute tubular necrosis(ATN) was 53% in the NHBD group
and 30% in the control group (P<0.05). With regard to
graft function there was no difference between both
groups. In transplantations with NHBD-kidneys, warm
ischaemia time (WIT) shorter than 30 min was accompanied by a better graft survival than WIT longer than 30 min
(5 year graft survival 81% versus 33%, P<0.05). The incidence of ATN in kidneys with WIT shorter than 30 min
was 40% versus 67% in kidneys with WIT of more than
30 min (NS).
Conclusion: Our results confirm that the use of kidneys
from NHBD is a valuable contribution to a kidney transplantation program. A short WIT clearly leads to better
The role of selected biochemical values in graft quality prediction in brain death organ donors
V Cerny, P Zivny, R Parizkova and P Dostal
Department of Anesthesiology and Intensive Care, Charles University, Faculty of Medicine, 50005 Hradec Kralové, Czech
Introduction: The aim of the study was to evaluate
selected biochemical values in brain death kidney donors
and to compare these results with number of rejection
episodes in recipients.
Methods: After institutional approval 35 patients were
prospectively studied. According to insult leading to brain
injury, patients after brain death diagnosis (angiography)
were divided into the two groups: donors with traumatic
intracranial hemorrhage (Group D1) and donors with nontraumatic intracranial hemorrhage (Group D2). Blood
samples were drawn immediately after brain death diagnosis. Lymphocyte count, CD3, CD4, CD8, CD19,
CD56/16, CD25, ALT, AST, immunoglobulins (IgG, IgM,
IgA), malondialdehyde, erythrocyte glutathione levels
(GSH) and erythrocyte glutathione-peroxidase activity
(GPX), selenium, sIL-2R, IL-2, IL-6, IL-8 and TNF-α were
measured. Results as mean (SD), *P<0,05.
Group D1 (n=17)
Group D2 (n=18)
Recipients no rejection
Rejection treated by steroids
Rejection resistant to steroids
AST (µkat/L)
ALT (µkat/L)
GPX (mg/g Hb)
GSH (mg/g Hb)
MDA (µmol/L)
Serum IL-6 (pg/mL)
Selenium (µmol/L)
Results: Selected results are presented in the Table.
Conclusions: Selenium levels correlated with graft prognosis, significant differences in selected values were
found between trauma and non-trauma brain-stem death
organ donors. High selenium levels in donors seem to be
of value in prediction of later graft rejection.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Experience with brain death and organ donation in a tertiary care hospital in Riyadh, Saudi
W Djazmati*, B Attar†, Y Arabi*, R McIntyre* and FA Shaheen†
*King Fahad National Guard Hospital, ICU 1425, PO Box 22490, Riyadh, 11426, Saudi Arabia and †Saudi Center for Organ
Transplantation (SCOT)
Introduction: The Kingdom of Saudi Arabia has a national
organ procurement organization (SCOT) with rigid criteria
for the definition of brain death and a standard method of
approaching the families of potential organ donors. As one
of the major ICUs in Riyadh, we have reviewed our experience with organ donation.
approached the relatives of all patients that were confirmed as brain dead and was successful in obtaining
consent for organ donation from 14 (35%) of these cases.
Eight of these were Saudi and 6 were non-Saudi. Twentysix kidneys, 10 livers and 22 corneas were successfully
transplanted and 13 hearts were used for valves.
Methods: Retrospective review of all confirmed brain
deaths in the 14 bed Adult ICU during the two-year period
from January 1997 to December 1998.
Conclusion: 1) The organ retrieval rate of 35% compares
favorably with other national and international data. This
rate is probably related to the active involvement and coordination by SCOT. 2) Our incidence of brain death (19%)
is above average and possibly related to the high incidence of road traffic accidents due to the unique traffic
conditions with extreme heat and lack of seat belt use. 3)
Family refusal for organ donation remains the main reason
for failure to donate and greater effort is still required to
increase public awareness in the Kingdom about the
concept of brain death and the importance of organ donation.
Results: Out of 210 deaths in the ICU, there were 40
(19%) confirmed brain deaths using the SCOT criteria.
Mean age was 21.7±12.1 years with male:female ratio of
9:1 reflecting the fact that 24 (60%) of the cases resulted
from road traffic accidents. Other trauma accounted for 4
(10%), cerebro-vascular accidents for 7 (18%) and brain
anoxia for 5 (20%). Thirty-one (78%) patients were Saudi
nationals and 9 (22%) were non-Saudi. SCOT
Epidemiology of cardiac arrhythmias in a medical-cardiologic intensive care unit: single center
G Heinz, P Reinelt, G Delle Karth, A Geppert and P Siostrzonek
Department of Cardiology, University of Vienna, Austria
Aim of study: To obtain frequency and distribution of
various types of significant, sustained arrhythmia
(ARRHY), its time of occurrence and factors influencing
the occurrence of ARRHY in a prospective observational
study in a medical-cardiologic ICU.
Results: There were 310 ARRHY episodes in 133 patients
(94 m, 39 f; age 65±12 years) during 11/1996 and 7/1999
(2.91 ARRHY episodes/patient, range 1–14). 278
episodes were tachyarrhythmias (TACHY; narrow-QRS
complex [N-QRS] n=108, wide-QRS complex [W-QRS]
TACHY n=169; regular n=179, irregular n=98) and 32
bradycardias. One case of primary ventricular fibrillation was
classified as TACHY but not with respect to QRS width and
regularity. The number of patients showing significant
ARRHY was relatively constant over the years (1996: 4/28
[14.3%], 1997: 50/302 [16.5%], 1998: 51/286 [17.8%],
22/140 (15.7%) until 7/1999). 135/278 (48.6%) W-QRS
TACHY were assigned ventricular TACHY (VT). 12 W-QRS
TACHY episodes could not be classified with certainty.
There were 13 episodes of Torsade de pointes (4.8%). NQRS TACHY episodes were atrial fibrillation (AFIB) 83,
atrial flutter 10, supraventricular TACHY 21, ectopic junctional TACHY 1. The occurrence of ARRHY followed a circadian pattern showing a peak during daytime and a decline
Figure 1
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Daytime (hours)
in the evening hours and early morning (Fig. 1). TACHY
were treated electrically (n=19), pharmacologically (n=139)
or combined (n=77). The antiarrhythmic drugs most frequently used were amiodarone (n=129), diltiazem (n=47),
ibutilide (n=21), lidocaine (n=24) and digitalis (n=18).
Proarrhythmia occurred due to haloperidol (n=4), cisapride
(n=2), ibutilide (n=1), and amiodarone (n=4). Sedoanalge-
sia (P=0.64), mechanical ventilation (P=0.76) or catecholamine treatment (P=0.63) had no influence on the
diurnal distribution of ARRHY. During 270/310 (90%)
episodes there was an elevated C-reactive protein (CRP), in
176/310 (56.8%) an elevated leukocyte count [L] and in
230/310 (74%) episodes were there elevated fibrinogen
levels (FGEN). These inflammation parameters on the day of
ARRHY did not differ significantly when compared to the
respective values 24 and 48 h before ARRHY onset (CRP
16.6±11, 16.3±11, 16±1mg/dl, P=0.9; L 11.4±5, 11.5±5,
11.9±5.7 G/l, P=0.7; FGEN 545±221, 565±208,
582±221mg/dl, P=0.26).
Conclusions: 1) Clinically significant, sustained ARRHY
occurred in ~1/5 of patients in this medical-cardiologic
ICU. 2) VT and AFIB were the single most frequent
ARRHY. 3) ARRHY followed a circadian pattern irrespective of the presence of sedoanalgesia, mechanical ventilation or catecholamine support. 4) The vast majority of
ARRHY occurred while there were signs of inflammation
without preponderance to the ascending or descending
limb of inflammation.
Transthoracic cardioversion with damped biphasic waveform shocks
VA Vostrikov, KV Razumov, PV Kholin and AL Cyrkin
Department of Cardiology, Moscow Medical Academy, Hospital N1 and Hospital N 81, Moscow, Russia
Introduction: The biphasic waveform has been shown to
have high efficacy for transthoracic ventricular defibrillation
[1,2]. The objective of this prospective study was to evaluate the clinical efficacy of biphasic waveforms for cardioversion of atrial fibrillation.
Methods: The pulse is an asymmetric quasi-sinusoidal
biphasic waveform. The peak current of the second phase
was approximately half that of the first phase. Transthoracic cardioversion (29 emergent, 71 urgent and 41 elective) were performed in 141 patients who were receiving
antiarrhythmic drugs (e.g., amiodarone). Ischemic heart
disease was the most common (about 90%) etiology.
Shocks were delivered through 11.5 cm paddles in the
antero-apical position. The maximum delivered energy was
195 J.
Results: See Table.
Delivered Energy
Cumulative Success
95% Confidence Interval
≤90 J
113/141 (80%)
≤140 J
121/141 (86%)
≤195 J
132/141 (94%)
Conclusions: Our clinical results demonstrate that the
biphasic waveform with a delivered energy of ≤195 J was
highly effective in cardioverting atrial fibrillation.
Poole J et al.: Resuscitation, 1998, 37(2):S13,(P–11).
Vostrikov V et al.: Resuscitation, 1998, 3(2):S42,(O–16).
Evaluation of transesophageal atrial pacing in the prone and lateral position
NM Schwann, DP Maguire, SE McNulty and JV Roth
Departments of Anesthesiology, Thomas Jefferson University Hospital, Jefferson Medical College, and the Albert Einstein
Medical Center, USA
Introduction: Transesophageal atrial pacing (TEAP) is
used for temporary treatment of hypotension and/or low
cardiac output caused by sinus bradycardia or atrioventricular junctional rhythm. It can also be used for temporary
overdrive pacing of reentrant tachycardias. A pacing
esophageal stethoscope (PES) is easy to place in supine
intubated patients. However, no guidelines exist for PES
placement in the prone or laterally positioned intubated
Methods: After IRB approval and written informed
consent, an 18 Fr PES with 1 cm depth markings (Model
550 CardioCommand, Inc) was inserted to a depth of
insertion (DOI) of 44 cm from the edge of the maxillary
alveolar ridge into the esophagus of 30 adult intubated
patients in the supine position. With the pacing rate set
10–20 beats per min faster than the patient’s intrinsic
heart rate and the current output on the pulse generator
(Model 2A) set at its lowest current setting (5.5 mA), the
current was gradually increased until atrial capture was
achieved. The minimum current producing continuous
atrial capture throughout the respiratory cycle was
recorded as the pacing threshold. The PES was then withdrawn 1 cm at a time and pacing thresholds were redetermined for each DOI from 44 to 25 cm inclusive. After
data collection in the supine position, each patient was
then positioned into either the prone (P), right (RLD) or left
lateral decubitus (LLD) position. Data were collected in
the same manner after the patients’ position was changed.
The width of the “Region of Capture” (ROC) represents
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
the number of DOI locations where TEAP could be
achieved with thresholds 40 mA. The “Optimal” DOI represents where the minimum TEAP current threshold
(THmin) occurred. Data is presented as mean ±standard
Results: TEAP was accomplished in all 30 patients both
in the supine and prone or lateral positions. The Optimal
DOI increased on average only 1.2 cm (from 34.0 to 35.2
cm) after positioning. Similarly, the average width of ROC
increased only by 0.5 cm (from 12.4 to 12.9 cm). The
THmin (mean 14.4±2.7 mA) never decreased after placement of a supine patient into the prone position. It either
remained unchanged or increased (minimum to maximum
individual difference 0–8 mA). However, there was considerable variability in how the THmin, Optimal DOI, and
ROC changed with the lateral position. Neither phrenic
nerve pacing, brachial plexus stimulation, nor induction of
arrhythmias was observed in any patient.
Discussion: This study demonstrates that TEAP with a
PES can be successful in patients who are subsequently
turned into the prone or lateral positions with minimal
adjustments. TEAP may be initiated in the prone or lateral
positions with the same insertion recommendations as
those for patients in the supine position.
Cardiopulmonary resuscitation with cardiopulmonary bypass after cardiac surgery
I Rodrigus, BJ Amsel, B Stockman, R De Paep and AC Moulijn
University Hospital Antwerp, Wilrijkstraat 10, 2650 Edegem, Belgium
Background: Emergency cardiopulmonary bypass (CPB)
can be life-saving in cardiac arrest (CA) after cardiac
surgery, when cardiopulmonary resuscitation (CPR) and
open chest massage do not suffice. Outcome of these
patients is, however, questionable. A retrospective study
on these patients was done to investigate predictors of
Material and methods: From January 1995 to October
1999, 18 patients experienced CA or ventricular fibrillation (VF) after a cardiac surgical procedure, refractory to
CPR and open chest massage (i.e. 0.7% of our total adult
cardiac surgery population). Mean age was 64.8 years
(range 36–75) with 14/18 male patients. Preceding operative procedure was coronary artery bypass graft (CABG)
(n=1), urgent redo CABG (n=2), ascending aorta replacement for dissection (n=1). CA most often occurred some
hours after surgery in intensive care (n=14), or with the
patient still in the operation theatre (n=3). Duration of
CPR is therefore always less than 1 hour. All patients
were immediately put on conventional CPB. Events
leading to CA were ischemia (n=11), bleeding with or
without tamponade (n=3), aortic dissection, electromechanical dissociation, right ventricular failure or shock
(each in 1 patient). In 8 of the 11 patients with ischemia,
redo CABG was performed. In the other patients (n=7),
bleeding sites were handled or reperfusion was given.
Results: Of 8 patients with redo CABG as a resuscitative
procedure, 5 could be weaned from CPB with or without
intra-aortic balloon pumping (IABP) and survived. The
remaining 3 patients were put on IABP (n=1) or biventricular assist device (BVAD) (n=2), but died. Three patients
with reperfusion for ischemia did not survive, despite IABP
(n=2) or BVAD (n=1). Of 7 patients without ischemia, only
2 survived.
Postoperative complications in the 7 surviving patients
(38%) were pneumonia (n=5), renal failure with hemodialysis (n=3), slow neurological recovery (n=1). Hospital
length of stay was 20.5 days (range 5–43). All but one
patient are still in good health (mean follow up 26 months).
Conclusion: The incidence of CPR necessitating emergency CPB was low. Only aggressive treatment with
emergent bedside resternotomy and CPB can save one
third of these patients. Best results were achieved when a
correctable cause for the CA was identified.
Functional status and quality of life (QoL) in long-term survivors of cardiac arrest after cardiac
A Anthi*, I Dimopoulou*, A Michalis† and GE Tzelepis*
*Intensive Care Unit and †2nd Cardiac Surgery Department, Onassis Cardiac Surgery Center, Athens, Greece
We assessed long-term survival, functional status and
QoL in cardiac surgery patients who survived after cardiac
arrest during the immediate postoperative period. Of
3982 patients undergoing cardiac surgery over a 30month period, 29 patients (0.7%) had a sudden cardiac
arrest. Of these, 27 patients (93%) were successfully
resuscitated with closed or open cardiopulmonary resuscitation (CPR). Overall, four patients (14%) survived the
CPR but died in the ICU, and 23 patients (79%) survived
to hospital discharge. At reevaluation four years later, 16
patients were still alive, yielding an overall four-year survival rate of 55%, while seven patients had died; of these
last seven patients, three (43%) died during the first year
and the remaining at various time intervals. The New York
Heart (NYHA) classification and a questionnaire based on
the Nottingham Health Profile (NHP) were used to investi-
gate functional status and QoL in these 16 long-term survivors. Of these, 12 patients (75%) were in NYHA class I,
three patients (19%) in NYHA class II and one patient
(6%) in NYHA class III. None of the patient reported any
emotional problems, signs of social isolation or difficulties
in physical mobility. Six patients (38%) had exertional
angina and six patients (38%) reported mild sleep disturbances, such as early awaking. All but two patients had
retired (taking advantage of the available laws), 15
patients (94%) were able to look after their home, 10
patients (63%) had a sexual activity, 13 patients (81%)
were active with their special interests or hobbies, and 12
patients (75%) had gone on holidays. In conclusion, more
than 50% of post-cardiac surgery patients who survive
after cardiac arrest occurring in the immediate postoperative period have a good functional status and a preserved
quality of life up to 4 years following cardiac arrest and
successful CPR.
Health-related quality of life after heart surgery
JK Ohukainen, PK Kaukoranta and TI Ala-Kokko
Department of Anesthesiology, Oulu University Hospital, PL 22, 90220 Oulu, Finland
SF-36 health survey has been validated to measure QoL
after intensive care [1]. A Finnish version of SF-36 is available with values for normal population [2]. This was used
to evaluate QoL after heart surgery and postoperative
intensive care.
Methods: Questionnaire was sent six months after ICU
discharge to 454 patients who had undergone heart
surgery during 01.01.1998–31.08.1998 in Oulu University Hospital.
Results: 375 patients, of whom 73% were male, completed and returned the questionnaire. CABG was performed in 81% of the cases. 71% of the operations were
elective. QoL of the patients was worse than that of the
normal population, especially for physical functioning and
general health. Instead general mental health and social
functioning were at the same level. General well-being
was worse than 12 months earlier in only 12% of the
patients. Physical functioning (P<0.001) and role limitations due to personal or emotional problems (P=0.04)
were worse in patients over 65 years old compared to
those under 65. Pain worsened QoL more on younger
patients (P=0.04). Female gender (P=0.002), emergency
operation (P<0.001), type of surgery (P=0.02), neurological complications (P=0.02) and preoperative HigginsCleveland classification (P<0.001) were associated with
low QoL. Instead postoperative infections, perioperative
myocardial infarction and prolonged intensive care had no
effect on QoL.
Conclusions: Most patients experience a better or equal
QoL six months after heart surgery than before the operation. Better QoL can be achieved by optimising preoperative status and by elective surgery. Measures to avoid
perioperative neurological complications must be applied
to patient care.
Chrispin PS et al.: Anaesthesia 1997, 52:15–23.
Hagman E: SLL 1996, 51:3534–3539.
Is postoperative ICU care needed for high-risk patients after major elective surgery?
J Cohen, M Shapiro, J Sulkes*, E Grozovski and P Singer
*Department of General Intensive Care and Epidemiology, Rabin Medical Center, Beilinson Campus, Petah Tikva 49100;
Sackler School of Medicine, Tel Aviv University, Israel
Introduction: Patients are frequently admitted to ICU following elective surgery either because of surgery-specific
or patient-specific risk. However, ICU beds are expensive
and not always available. We assessed whether there
were significant differences in outcome between patients
admitted to an ICU and a similar group of patients
managed in general surgical wards.
1.9+3.7 h in the recovery room (group 2, n=47). We
recorded preoperative factors (demographics, reason for
Surgery–specific risk
Serum creatinine (mg%)
Methods: All patients undergoing elective surgery over a
17-month period and for whom a place in ICU was
requested were included in this prospective study. The
anesthetic and surgical departments were not made aware
of the study. Following surgery, patients were either admitted to ICU if a bed was available (group 1, n=97) or transferred to the general surgical wards after a mean stay of
Group 1 (n=97)
Group 2 (n=47)
Ward complications
Length of stay (days)
15.4% (15/97)
3.84% (4/47)
28-day mortality
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
admission, type of procedure, medical history, hemoglobin,
serum creatinine, ASA, NYHA), intraoperative factors (type
of anesthesia, intraoperative complications, units of blood
transfused, length of surgery) and postoperative factors
(duration of ventilation, APACHE II, TISS scores). Outcome
variables were the occurrence of postoperative complications, duration of hospitalization and 28-day mortality.
Results: See Table.
Conclusion: We found no difference in three outcome
variables in high-risk patients who were either admitted to
the ICU or treated in a general surgical ward. Criteria for
ICU admission need to be reassessed in these patients.
Abdominal sepsis in patients 65-onwards after non-scheduled laparotomy
P García-Saura, F Peramo, C Serrano, J Maldonado, A Márquez, MS Serrano and J Prieto
Department of Anaesthesiology, Resuscitation and Pain Management, Hospital Clínico Universitario “San Cecilio”, Granada,
Introduction: Abdominal sepsis after emergency laparotomy has a high mortality rate because is associated with
a great number of serious complications. The aim of this
study was to assess abdominal sepsis associated mortality and to determine which variables could predict the possibility that abdominal sepsis occur.
Materials and methods: We prospectively studied 128
patients, 65 onwards, who underwent emergency laparotomy (abdominal resection, perforation or ischaemia). In
every patient we recorded several variables of the pre,
intra and postoperative period. By patient’s history, radiographic and scan studies, laboratory tests specific for the
type of surgery performed, wound inspection and surgical
reexploration, we determine the presence of abdominal
infection. We also studied mortality associated with this
type of infection and variables that could predict its
appearing (multiple logistic regression analysis, using a
model in which probability an event P will occur is
1/(1 + ez), where z is the linear combination, z=B0+
B1X1+ B2X2+.....BPXP; B0, B1, B2,...Bp are coefficients
estimated from the data,×is the independent variable and
e is the base of the natural logarithms, approximately
2.718). Statistical analysis was made with SPSS for
Windows 5.01® (multiple logistic regression analysis and
Fisher exact test as required) and P<0.05 was regarded
as significant.
Results: We evaluated 52 women and 76 men and
patients’ characteristics were (mean ±standard deviation):
age 75.27±7.52 and weight 70.34±16.35. Abdominal
sepsis was recorded in 15 patients (11.7%) and 10 died
(66.67%). Mortality associated with abdominal infection
was greater than mortality in patients without abdominal
sepsis. Indeed, variables that can predict abdominal
sepsis were diffuse vascular peripheral disease, acute
renal failure, ARDS, surgical reexploration and a hospital
stay greater than 7 days.
Conclusions: Emergency laparotomy has a greater mortality in patients with abdominal sepsis (66.67 vs 27%).
Variables that can predict its appearing are: decreasing
O2 tissue delivery factors, events that increase the possibility of nosocomial infection or those circumstances
favouring multisystemic organ failure (lung or kidney).
Cook TM, Day CJE: Hospital mortality after urgent and emergency
laparotomy in patients aged 65 years and over: Risk and prediction
of risk using multiple logistic regression analysis. Br J Anaesth 1998,
An experience with surgical admissions to a Paediatric ICU (PICU) in Harare Zimbabwe
I-E Pazvakavambwa
Department of Paediatrics and Child Health, University of Zimbabwe, P.O. Box A178 Avondale, Harare, Zimbabwe
The PICU in Parirenyatwa Hospital in Harare is a 5bedded unit catering for both surgical and medical
patients. A retrospective review was undertaken to evaluate the surgical ICU experience in the Unit. The study
period was from January 1997 to December 1998.
Results: During this period there were 147 surgical
admissions to the unit constituting just under half of the
total admissions. The ages ranged from 1 day to 14 years
with a median age of 36 months. Seventy-six (52%) of the
patients were male. The mean stay in the PICU was 3.9
days with a range of 28 days. Cardiac patients made up
45% of the patients followed by respiratory tract surgery
(15%). The overall ICU mortality was 17%. Twenty-five
percent (25%) of the patients were hypothermic on arrival
to PICU. Factors associated with a high mortality were
abnormal temperature on arrival in the PICU, young age
and being ventilated. Three patients had accidental extubation on arrival to PICU.
Conclusions: There is a rather high incidence of
hypothermia in surgical patients admitted to the PICU in
Harare and this has adverse effect on mortality. Closer
attention to temperature control during surgery and transport to PICU could reduce mortality. Further studies are
needed in this regard.
Organization and staffing of intensive care units in Brazil
J Livianu, JMC Orlando, A Giannini, RGG Terzi, M Moock, C Marcos and N David
AMIB, r.Domingos de Morais 814 bloco II cj.23, CEP 04010-100, São Paulo, SP, Brazil
Introduction: The ‘1° Brazilian Census of ICUs’ was
developed by the Brazilian Society of Intensive Care
(Associação Medicina Intensiva Brasileira - AMIB) to delineate the ICU profile in this country.
Methods: Data collection was done through a questionnaire sent to all hospitals. Through a software package,
these data were gathered, creating a comprehensive database with ICU organizational and resource information.
Results: To be accredited as a training center by AMIB,
the ICU must run a special program under specific conditions. At this census, 31.3% of the ICUs had a critical care
training program but only 10.4% were accredited by
AMIB. Despite the continental dimension of the country
and the large number of units, they are concentrated
(79%) in south-east region. In Brazil, critical care medicine
has largely been considered a second specialty by the
physician: 61.1% were clinicians, 22.6% pediatricians,
9.8% surgeons and only 3.1% anesthesiologists. Just
23% of the intensivists are certified as specialists by
AMIB. 72.9% of the physicians work on duty (12 or 24 h
shifts) and 18.1% work on a daily basis. 84.7% of the
respondent ICUs had a chief-nurse exclusive to the ICU
and 67.4% had therapists 12 h per day. 71.5% performed
scientific meetings with ICU staff regularly. 51.5% had a
computerized registry of admitted patients but only 33.4%
classified admitted patients according to a scoring
system. 74.8% had written admission and procedures
rules and 64.4% had written therapeutic orientation rules.
83.3% performed evaluation of adverse patient occurrences and 53.1% of the ICUs elaborated an annual
report about their activities.
Conclusion: This study was the first step to recognize the
structure and distribution of ICUs and exposed facts that
must be improved, for instance, the need to increase the
number of specialists through the creation of new training
centers all over the country.
Acknowledgement: This study was supported by Bristol-Myers Squibb
Baseline audit of manipulation and management of intravenous therapy delivery systems
C Martinsen, A Hughes and M Smithies
Critical Care Services, University Hospital of Wales, Cardiff CF14 4XW, UK
Background: We are developing local evidence-based
guidelines on the management of intravenous delivery
systems in a 14-bedded Teaching Hospital General ICU1.
A baseline audit was carried out to assess current practice prior to the publication of our proposed guidelines,
and re-audit.
Methods: We performed an observational audit of the
setup or change of an intravenous infusion and the management of intravenous delivery systems. ICU staff were
aware that an observational audit was in progress but
blind to what was being observed. The observations were
carried out over a three-week period.
Set up or change of an intravenous infusion (n=13)
Are hands washed before the procedure?
Are hands washed with soap and water?
Are hands washed for at least 15 s?
Is a clean plastic apron worn?
Are clinically clean gloves worn?
Was alcohol swab used prior to disconnection?
Was connection allowed to dry before disconnection?
Was set disposed of as unit policy?
Was alcohol swab used on valve prior to connection?
Was valve allowed to dry prior to connection?
Set up or change of an intravenous infusion (n=10)
Results: See Table.
Conclusion: Practice is short of the stringent care necessary to avoid infection related to the delivery of intravenous
therapy. The lack of consistency is also an issue. This
audit highlighted the need for clinical guidelines and also
the value of clinical audit of practice. Guidelines will now
be introduced and developed in the light of appraisal of
evidence, further audit and user feedback.
Do all connections have a luer lock?
Are connected lines labelled?
Are all labelled lines within 72 h of date/time?
Do venous catheter hubs have Smartsite valve connected?
Do all catheters have a transparent dressings?
Are the catheter entry sites visible?
Does the dressing need to be changed?
Are all iv lines secured to the patient with tape?
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Collaborative practice and clinical outcomes in the ICU
RE Nap, M Silva Alvaro, V Fidler and D Reis Miranda
Health Services Research Unit, University Hospital Groningen, P.O. Box 30.001, 9700 RB Groningen, The Netherlands
Objective: To evaluate the effect of organizational and
managerial intervention, aimed at increasing the professional collaboration between medical and nursing staff,
upon the clinical outcomes in the ICU.
Design: Prospective, randomized, multi-center and multinational study.
Setting: 47 ICU’s (22 control vs 25 intervention group)
from 9 European countries.
Methods: All consecutive admissions were enrolled,
during two periods of two months each.
Patient data: Admission data; first day SAPS II score;
hourly registration of Critical Events (CrEv) defined as the
duration (in hours) of out-of-range measurements of four
parameters (heart rate, blood pressure, urine output and
oxygen saturation); Sequential Organ Failure Score
(SOFA) at admission and then every 24 h; ICU outcome.
Measurements: Final outcome: ICU mortality. Intermediate
outcomes: SOFA, CrEv (expressed as the % of time spent
in CrEv; single or combined), and length of stay (LOS).
Intervention: 1) training of nursing and medical staff of 25
ICU’s in inter-professional collaboration; 2) the use of a
specific manual of instructions, supported by the daily use
of two protocols covering awareness of processes of
care, and professional dialogue (6 months).
Analysis: The effect of the intervention upon final and intermediate outcomes was evaluated by comparing the values
of these variables in the experimental and the control
groups of ICU’s, before and after the intervention (Table
below). SOFA computations included total daily score and
Delta(∆)-SOFA [1]. A ∆-CrEv variable was also constructed. The analyses were made on Day 1, 3 and 5, in the
group of patients staying ≥3 and ≤10 days in the ICU. χ²test was used for comparing means; significance if P<0.05
Results: Data on 1,633 patients were obtained; there
were no significant differences in the 4 ‘study cells’ for
Age (median=69), SAPS II (median=35) and LOS
(median=4 days). A significant decrease in ICU mortality
was observed for the intervention group (16% control
group; 9% intervention group).
Control group
n=343 vs 331
SOFA day1
SOFA day3
SOFA day5
CrEv day1
CrEv day3
CrEv day5
Intervention group SOFA day1
n=603 vs 356
SOFA day3
SOFA day5
CrEv day1
CrEv day3
CrEv day5
1.12±1.74 0.00
1.66±1.98 0.00
0.23±0.45 0.000
0.21±0.48 0.000
*Indicates significant differences between intervention and control
groups. BI, before intervention; I/C, intervention/control.
Conclusion: Collaborative practice has a significant and
beneficial effect upon clinical outcomes in the ICU.
Moreno R et al.: Intensive Care Med 1999, 25:686–696.
Prolonging ICU stay reduces the number of ward deaths: evaluation of a model for triaging ICU
KJR Daly, RJ Beale and RWS Chang*
Department of Intensive Care, St Thomas’ Hospital, London SE1 7EH; *Department of Renal Transplantation, St George’s
Hospital, London SW17 0QT, UK
Introduction: A significant number of patients discharged
alive from the intensive care unit (ICU) die subsequently
on the general wards. A predictive model using data from
the patients’ last day in the ICU prior to discharge, and a
0.6 cutoff, correctly identified 66% of ward deaths. We
tested the model’s ability to identify those patients who
may benefit from a further 24 to 48 h stay in ICU.
Patients and methods: All ICU survivors discharged
between 1st June 1990 to 31st December 1996 who
stayed for more than three days, in whom the predictive
model applied within 48 h of ICU discharge, were studied.
562 patients were classified into three groups: Group (0)
311 patients last predicted to be at risk of ward death on
the day of ICU discharge; Group (1) 188 patients last predicted at risk 24 h prior to ICU discharge; Group (2) 63
Hospital outcome
Group 0
(333 patients)
Group 1*
(188 patients)
Group 2**
(63 patients)
256 (82%)
55 (18%)
159 (85%)
29 (15%)
59 (94%)
4 (6%)
*χ2 versus Group 0 = not significant; **χ2 versus Group 0 P<0.024.
patients last predicted at risk 48 h prior to ICU discharge.
The model was further evaluated using another two independent data sets.
Results: See Table. Similar findings were found for the
two other data sets.
Conclusion: There was a significant improvement in hospital survival for those patients who stayed in the ICU an
additional 48 h following the prediction of ward death. If
this can be confirmed in a prospective study, it will have a
major impact on the provision of ICU beds in the United
Are we allocating limited resources to patients in most need?
T Nolin
Intensive Care Unit, Hospital, 291 85 Kristianstad, Sweden
Introduction: We aimed to examine this question by studying the correlation between severity of illness, outcome and
the nurse workload (major determinant of cost).
Methods. We did a retrospective analysis of all intensive
care patients admitted during 1992–1998 to the 10-bed
general ICU. APACHE II [1] was used to determine the
hospital mortality risk (MR). Patients were grouped into
risk bands, in steps of 20%. Standardised Mortality Ratio
(SMR=observed hospital mortality/calculated hospital
mortality) was used in each stratum to define clinical efficacy. As a proxy for resource consumption, a modified
form [2] of the nursing care recording (NCR) system [3]
was used. Workload per patient, per survivor, per non-survivor and ‘effective’ workload (workload all patients/
number of survivors) was calculated within each stratum.
vivors/non-survivors was 12.7/23.4, with estimated MR of
18.0/47.9%. NCR per patient was 2.3 times higher for
deceased (385) compared to survivor’s (165). In survivors, NCR increased in a linear fashion with increasing
MR up to 40<60%. NCR in deceased patients was
highest in those patients with lowest MR and lowest in
those with highest MR. Effective work load increased
gradually in a linear fashion with increasing MR.
Discussion. Resources were allocated to patients in most
need. It was difficult to quantify the degree of efficacy.
Effective NCR is an indicator of clinical efficacy in relation
to severity of illness.
Results: 4395 patients were admitted. 306 were children
and 342 had missing values in scoring. APACHE II for surP224
Crit Care Med 1985, 10:818–829.
Acta Anaesthesiol Scand 1992, 36:610–614.
Svensk Anestesi och Intensivvård 1997, 1:16–38.
Mortality and outcome in different subgroups of patients admitted in an ICU
A Marinho, J Luzio, F Prospero, L Lopes, M Fernandes, A Azevedo and M Brandão
Intensive Care Service, General Hospital Stº Antonio, Oporto, Portugal
Introduction: SOFA and SAPS II scoring is used for monitoring of organ function and outcome prediction in intensive care. We evaluated the use of these scores to analysis
the different subgroups of patients admitted to our ICU.
Methods: Over a period of nine months (between
October 1998 and June 1999), the first five days scores
of SOFA and SAPS II, were prospectively calculated for
all consecutive patients (length of ICU stay >12 h) admitted to our 10-bed medical-surgical ICU. We stratified
patients using three different subgroups: 1) medical (M);
2) surgical, elective (E); 3) surgical, unscheduled (U).
Organ failure if SOFA score ≥3. Mortality was assessed at
ICU discharge. Data analysis and statistics were performed using the Statistical Package for Social Sciences
(SPSS) version 6.0.1 for Windows.
Results: 278 patients, aged 58.5±18.4 years, were
included in the study. In this cohort of patients, the median
length of ICU stay was 7.8±14.6 days, SAPSII 29.1±13.5
(median 27), SOFA 5±3.75 (4), and the ICU mortality rate
was 19.8%. Relation between SOFA/SAPSII at admission
and ICU discharged: non-survivors–SOFA 8.3±4.3 (8),
SAPSII 40.6±13.9 (39); survivors–SOFA 4.25±3.13 (3.5),
SAPSII 26.3±11.8 (24.5). In the three different subgroups
we have: (M)–73 patients (26.2%), length of stay
13.78±20.98 (median 6), SAPSII 33.9±12.6 (32), SOFA
5.37±3.7(4), mortality rate 30.1%; (E)–150 patients (54%)
length of stay 3.89±8.05 (1), SAPSII 23.3±10.2 (22),
SOFA 4.43±3.5 (4), mortality rate 10%; (U)–55 patients
(19.8%) length of stay 10.47±15.47 (5), SAPSII
38.7±4.07 (38), SOFA 6.29±4.07 (6), mortality rate
32.7%. The mortality rate increased with the number of
failing organs present at the time of admission (7.6% with
no organ failure, and 100% with 5 organs failure). The evaluation of a subgroup of 100 patients who stayed in the ICU
for at least 5 days showed: length of stay 18.53±20.4,
SAPSII 35.49±10.9, SOFA 6.76±3.75 (7), mortality rate
34%. Subgroups: (M)–57%; (E)–25%; (U)–28%.
Conclusion: We have in our ICU two subgroups of
patients: 1) medical (M) and unscheduled surgery (E) with
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
higher length of stay, higher scores, and higher mortality
rate, compared with the second group–elective surgery
(E). This is especially important when we try to compare
the performance of different ICUs.
Pre-admission functional status and outcome in medical intensive care
M Wehler, R Strauss, A Bost, A Geise, A Mueller, P Martus* and EG Hahn
Departments of Medicine I and *Medical Informatics, Biometry and Epidemiology, University Erlangen-Nuremberg, PF 3560,
91023 Erlangen, Germany
Introduction: Guidelines for intensive care unit (ICU)
admission recommend taking the premorbid functional
status and comorbidity of patients into consideration when
selecting those who are likely to benefit from ICU care [1].
We analyzed the prognostic value of pre-admission functional status, comorbidity and daily APACHE II scores in
medical ICU patients.
Patients and methods: All patients ≥18 years of age who
stayed >24 h in our non-coronary medical ICU were eligible for inclusion. Functional status referred to the two
months prior to admission and was assessed using the
domains of basic physiologic activities and activities of
daily life of the PAEEC quality of life questionnaire [2].
Comorbidity was categorized using the Charlson Index
(CI) [3]. ICU, hospital and six-month follow-up mortality
rates were documented. Statistical analysis used the cox
model for censored data.
Results: During the 12-month study period, 325 patients
were included. 55% were male, mean age was 58±17
(±SD) years, range 19–95 years. Mean ICU length of stay
was 10.4±15 days, range 2–127 days. Mean APACHE II
score (first 24 h) was 23±10, range 0–51. 75% of our
patients had one or more chronic diseases, mean CI was
2.4±1.9, range 0–11. Functional status was reduced in
80% of our patients, with 40% of them demonstrating
severe impairments. Cumulative mortality rates were: ICU
24%, hospital 34.5%, six-month follow-up 40.3%. One
patient was lost to follow-up. Multivariate analysis showed
that pre-admission functional status (ICU/hospital mortality: P=0.008/P=0.0009), CI (P=0.003/P=0.0005) and
APACHE II score (P=0.012/P=0.012) had an highly significant association with ICU and hospital mortality.
Conclusion: In our population of critically ill medical
patients with a high rate of preexisting chronic diseases,
pre-admission functional status and comorbidity are independent variables with a highly significant association with
on-unit and hospital mortality.
Crit Care Med 1999, 27:633.
Rivera Fernandez R et al.: Intensive Care Med 1996, 22:1034.
Charlson ME et al.: J Chronic Dis 1987, 40:373.
Therapeutic activity in elderly medical intensive care patients
M Wehler, R Strauss, A Mueller, U Reulbach*, P Martus* and EG Hahn
Departments of Medicine I and *Medical Informatics, Biometry and Epidemiology, University Erlangen-Nuremberg, PF 3560,
91023 Erlangen, Germany
Introduction: Previous studies showed a limitation of therapeutic activity in elderly critically ill patients. We analysed
whether the age of patients admitted to our medical intensive
care unit (ICU) influenced the quantity of therapy provided.
Patients and methods: All patients admitted to our 12bed non-coronary medical ICU between 8/95 and 7/99
were enrolled in the study. Prospectively the reason for ICU
admission, acute diagnoses, co-morbid illness, diagnostic
procedures, daily APACHE II and TISS scores, ICU and
hospital mortality rates were documented. Contingency
tables were analysed using chi square test, continuous
variables were compared using Mann-Whitney U test.
Results: 2127 patients were studied; mean age was
57±18.6 (±SD) years, median 59, range 18–101 years,
45% were female. Mean ICU length of stay was 7±12.8
days, median 3. Mean APACHE II score (first 24 h) was
18.6±12.3. Mean TISS score (mean value of all days spent
on the ICU) was 25.8±13.3. ICU mortality was 23.6%,
hospital mortality was 31.7%. 432 patients (20.3%) were
≥75 years of age. Compared with younger patients these
patients had higher APACHE II scores (23.2±11.8 vs
17.5±12.1, P=0.0005), higher modified APACHE II
scores (age points subtracted) (17.2±11.8 vs 15.2±11.6,
P=0.001), higher mean TISS scores (28.3±12.2 vs
25.2±13.5, P=0.0005), a higher ICU mortality rate (21.5%
vs 31.7%, P=0.0005) and hospital mortality rate (28.4% vs
44.4%, P=0.0005). No differences between the two age
groups were noted in the number of days spent on a respirator or renal replacement therapy and in the frequency of
diagnostic procedures, like ultrasound and endoscopy.
Multivariate analysis showed that after adjustment for
disease severity, mechanical ventilation and mortality rates
patients ≥75 years of age received the same amount of
therapy as assessed by daily TISS points.
Conclusion: In contrast to other investigators we found
no limitation of therapeutic activity in our sample of elderly
medical ICU patients.
Ward death following ICU discharge — can it be predicted? Development and validation of a
predictive model
KJR Daly, RJ Beale and RWS Chang*
Department of Intensive Care, St Thomas’ Hospital, London SE1 7EH; *Department of Renal Transplantation, St George’s
Hospital, London SW17 0QT, UK
Introduction: A significant number of patients discharged
alive from the intensive care unit (ICU) die subsequently
on the general wards; a problem compounded by early
discharge due to pressure on ICU beds. We analysed
data from the patients’ last day in the ICU prior to discharge, during their first ICU admission, to derive a predictive model that would identify those patients who were at
risk of dying on the ward.
Patients and methods: All ICU survivors discharged
between 1st June 1990 to 31st December 1996 were
included in the study. Of the 5475 survivors there were 200
(3.7%) ward deaths. Physiological and treatment data, collected prospectively from the Intensive Care Database
(Riyadh Intensive Care Program, Medical Associated Software House Ltd, London, UK), for the patients’ last day in
the ICU were analysed to determine which variables had a
significant influence on death following ICU discharge. Multiple logistic modelling was used to select which variables
produced the ‘best’ predictive model of ward death. The
model was then validated using two independent data sets.
Results: A higher mean acute physiology score (APS)
(10.1 vs 8.0), increasing age (65 vs 59 years), presence
of chronic ill health (39% vs 22%), longer ICU stay (8 vs 3
days), and whether the patient had cardiac surgery, were
selected for inclusion in the model. The APS, age, chronic
ill health and length of ICU stay were positively associated
with death whilst having had cardiac surgery was positively associated with survival. Using a 0.6 cutoff the predictive model demonstrated good discrimination
(sensitivity 66%, specificity 88%, ROC curve 85.6%) and
calibration (χ2 7.5982; P=0.4737). External validation of
the model produced a ROC curve of 86.8% (sensitivity
90%, specificity 68%) (data set 1) and 77.5% (sensitivity
75% specificity 69%) (data set 2).
Conclusion: Using patient data obtained at ICU discharge it is possible to develop a discharge triage model
which identifies those patients who are at risk of dying on
the ward following ICU discharge.
Withdrawing life-prolonging medical treatment in 246 patients
T Nolin
Intensive Care Unit, Hospital, 291 85 Kristianstad, Sweden
Introduction: In this study, made in a 10-bed general ICU
in a 500-bed district hospital, we examined 1) how often
withdrawal (WD) decisions are made;, 2) the reasons and;
3) the results of the decisions.
Methods: We followed up medical charts of 246 patients
in whom we had documented decisions to withdraw lifesupport, according to our written policy. The study period
was between 1994–1998.
Results: The incidence of WD decisions was 7.7%. 41%
of patients dying in the ICU, and 40% of those later dying
in the ward, had a decision made of WD. Of the 246
patients with a WD decision, 57% died in ICU, 35% later
in the general ward and 8% were discharged alive from
hospital. Median age was 76 years in those with a WD
decision, as compared with 64 years in those without a
decision. Reasons for WD of therapy were autonomy in
6%, prognosis of acute disease in 34%, prognosis of
coexisting, chronic disease in 19% and failure to respond
to therapy in 41%. Median time from admission to ICU to
WD was 2.8 days, and from decision to death in ICU 1.9
days. In four cases renewed consideration was done. We
found one case of discordance with relatives concerning
the WD decision.
Discussion: Our results were similar to Sjøqvist et al. in
Sweden [2] and show that 1) these decisions are
common (8%); 2) the reasons for the decision were mainly
made from failure to respond to therapy and prognosis of
the acute and underlying disease and; 3) few patients
survive the hospital stay after such decisions.
BMA: Withholding and Withdrawing Life-prolonging Medical Treatment. BMJ Books 1999.
Sjökvist P et al.: Limiting life support. Acta Anaesthesiol Scand
1998, 42:232–237.
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Quality of life 6 and 12 months after discharge from the intensive care unit
HF O’Neill, GG Lavery, P Donnelly, M O’Mahony and B Alexander
Regional Intensive Care Unit, Royal Victoria Hospital, Belfast, BT12 6BA and Department of Psychology, The Queen’s University
of Belfast, Northern Ireland
Introduction: Mortality alone is a commonly used but nondiscriminating outcome measure after critical illness.
Although we now view physical, psychological and social
outcomes as important [1,2], which assessment tools to
apply and when to apply them remains unresolved.
Method: Over a period of one year, we prospectively
studied all patients who spent more than 72 h in a tertiary
referral intensive care unit (ICU). All eligible patients were
then visited by an ICU nurse at their normal residence 6
and 12 months after ICU discharge. They were scored
using several scales: functional limitations profile (FLP),
acceptance of illness (AOI), generalised self efficacy
(GSES), health value (HVS), satisfaction with life (SWLS)
and hospital anxiety (HAS) and depression (HDS).
Results: During the study period, there were 557 ICU
admissions. ICU and hospital mortality rates were 21%
(n=119) and 32% (n=185). Study entry criteria were satisfied by 227 patients of whom 58 were excluded (<18
years old, no longer resident in region, refused to participate, died before interview). FLP scores (mean ±SEM) for
the 169 patients (M:F 99:70) at 6 and 12 months were
23.5±1.20 and 18.2±1.03 respectively, showing a signifiP230
cant increase in QOL (P<0.001). AOI, HAS and GSES
scores also improved between 6 and 12 months while
HVS, SWLS and HDS were unchanged. The most significant factors determining QOL at both 6 and 12 months
after discharge from ICU were levels of perceived physical
and psychosocial limitation as measured by FLP. Age,
admission APACHE II score and length of ICU stay did
not influence QOL outcomes.
Conclusion. Measures of QOL outcomes in post-ICU
patients vary with the assessment tool applied. Several
show a significant increase in QOL between 6 and 12
months after ICU discharge and suggest that longer
periods of follow-up than previously suggested [3] should
be used.
Rogers J. Ridley S. Chrispin P. Scotton H. Lloyd D: Reliability of next
of kins’ estimates of critically ill patients quality of life. Anaesthesia
1997, 52:1137–1143.
O’Boyle CA. Quality of life assessment: A paradigm shift in healthcare? The Irish Journal of Psychology 1997, 18:51–66.
Bell D and Turpin K. Quality of life at three months following
admission to intensive and coronary care units. Clinical Intensive
Care 1994, 5:276–281.
Change in the provision of out-of-hours consultant cover improves case-mix adjusted mortality in
a district general (university-affiliated) hospital intensive care unit
MC Blunt and KR Burchett
Queen Elizabeth Hospital, Gayton Road, King’s Lynn, Norfolk, PE30 4ET, UK
Introduction: In the United Kingdom the national recommendations [1] for intensive care units (ICUs) include the
24-h availability of consultants with a sessional commitment to intensive care (intensivists). In practice this has
not been possible outside the larger teaching hospitals,
and normal practice in other hospitals has been for the
consultant anaesthetist to provide cover for the ICU out of
routine hours. Following introduction of 24-h intensivist
cover in our hospital we wished to assess whether there
was an improvement in mortality standardised for case-mix
using the APACHE 2 prognostic calculation [2] (SMR).
Methods: The 465 patients admitted to ICU in the 18
months following introduction of 24-h intensivist cover
(Intensivist) were compared with the 387 patients admitted to ICU in the 18 months immediately preceding the
change (Non-Specialist) in a historically controlled study.
APACHE 2 scores, calculated risk of death, age, ICU
lengths of stay and hospital length of stay (survivors) were
collected from the ICU database, and SMRs were calculated for each group. Patients who were under 16 years of
Age (years)
57.6 (55.6–59.6)
59.7 (57.7–61.7)
18.1 (17.4–18.8)*
19.3 (18.5–20.1)
Risk of Death (%)
29.6 (27.4–31.8)
31.0 (28.5–33.5)
1.5 (0.1–6.6)
1.8 (0.2–5.3)
13 (4–23)
14 (4–19)
ICU length of stay (days)
Time to hospital
discharge (days)
Standardised Mortality
Ratio (95% CI)
0.81 (0.66–0.98)** 1.11 (0.93–1.34)
Data presented as mean (95% Confidence intervals) for age, APACHE
and risk of death; median (interquartile range) for time data.
*P<0.05; **P<0.01.
age or who stayed in ICU<8 h (Intensivist: 49; Non-Specialist: 59) were excluded. Demographic data was
assessed using ANOVA and SMR by Poisson distribution.
Results: There was a significant improvement in SMR in
the intensivist group (P<0.01). The APACHE 2 scores in
the patients in the intensivist group were significantly
lower (P<0.05) (Table).
ingful comparison of the groups despite the lower
APACHE 2 scores in the intensivist group.
Conclusions: Within the methodological restrictions of
the historical control design this study supports the introduction of 24-h intensivist cover in all intensive care units.
The use of SMR as the primary end-point ensures a mean-
Standards for Intensive Care Units. Intensive Care Society; London.
Rowan KM, Kerr JH, Major E, McPherson K, Short A, Vessey MP. BMJ
1993, 307:972–977.
A comparison between some standardized scales of evaluation of anxiety and depression in
nursing staff
D Chiumello, P Caironi, P Pelosi*, S Losappio, M Mondino, R Malacrida†, M Tomamichel§ and L Gattinoni
1st. di Anestesia e Rianimazione, Università degli Studi, Osp. Maggiore di Milano–IRCCS, Italy; *Dip. di Scienze Cliniche e
Biologiche, Università dell’Insubria, Varese, Italy; †Ospedale Civico di Lugano, Switzerland
Anxiety and depression can be evaluated through different
methods. We studied the current associations between
the commonest evaluation methods that are present in scientific literature.
Materials and methods: We analyzed a population of
1392 nurses working in intensive care units (ICUs) and in
general medicine units (GMUs), distributed in 101 Italian
hospitals (with a mean age of 32±6.9 years, 73.8 %
female). We considered the following evaluation scales: the
Hospital Anxiety and Depression scale, divided in anxiety
(HAD A) and depression (HAD D) status; the STAI scale,
divided in acute anxiety (Y-1) and chronic anxiety (Y-2)
status; the Maslach Burnout Inventory–Human Services
Survey (MBI.), divided in Emotional Exhaustion (EE), Depersonalization (DP) and Personal Accomplishment (PA).
Assuming the HAD as a reference scale, we evaluated the
influence of the others scales to determine HAD. The population was divided, distinctly for anxiety and depression, into
three groups, according to standardized parameters of HAD
- ‘non-cases’ (HAD 0–7), ‘doubtful cases’ (HAD 8–10), and
‘cases’ (HAD 11–21). We used multiple linear regression
models; statistical significance was accepted as P<0.05.
Results: The regression coefficients of the multiple linear
regression models are expressed in the table, with the
variables that result in statistical significance. For depression, we considered doubtful cases and cases together
(last being only 39).
Conclusions: 1) Besides HAD A, STAI Y-2 also seems to
be useful for evaluating anxiety, whatever the degree of it.
2) Besides HAD D, STAI Y-1 and MBI EE also seem to be
useful for evaluating depression, whatever the degree of it.
r = 0.75
P < 0.001
HAD A 0-7
r = 0.46
P < 0.001
HAD A 8-10
r = 0.27
P < 0.001
HAD A 11-21
r = 0.55
P < 0.001
S.T.A.I. Y-1
S.T.A.I. Y-2
P < 0.001
P < 0.001
P < 0.001
P < 0.001
S.T.A.I. Y-1
S.T.A.I. Y-2
P < 0.001
P < 0.001
P < 0.001
S.T.A.I. Y-2
P < 0.05
P < 0.05
S.T.A.I. Y-1
S.T.A.I. Y-2
P < 0.001
P < 0.05
P < 0.05
S.T.A.I. Y-1
S.T.A.I. Y-2
P < 0.001
P < 0.001
P < 0.001
P < 0.001
P < 0.001
S.T.A.I. Y-1
S.T.A.I. Y-2
P < 0.001
P < 0.001
P < 0.001
P < 0.001
S.T.A.I. Y-1
P < 0.05
P < 0.05
r = 0.69
P < 0.001
HAD D 0-7
r = 0.52
P < 0.001
HAD D 8-21
r = 0.49
P < 0.001
Stress in nursing staff: a comparative analysis between intensive care units and general
medicine units
D Chiumello, P Caironi, P Pelosi*, S Losappio, R Malacrida†, M Tomamichel§ and L Gattinoni
1st.. di Anestesia e Rianimazione, Università degli Studi, Osp. Maggiore di Milano–IRCCS, Italy;*Dip. di Scienze Cliniche e
Biologiche, Università dell’Insubria, Varese, Italy; †Ospedale Civico di Lugano, Switzerland
It’s a current belief that stress is an outstanding feature of
intensive care units, in particular within nursing staff. The
aim of this study was to compare some variables belong-
ing to stress (i.e. anxiety, depression and ‘Burnout’ syndrome) between nurses working in intensive care units
(ICUs) and general medicine units (GMUs).
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
Table 1
A 11-21
S.T.A.I. Y-1
S.T.A.I. Y-2
EE ≥27
Table 2
7 (5,9)
3 (2,6)
33 (30,39)
36 (31,42)
19 (13,29)
6 (2,11)
36 (29,41)
5 (3,7)
6 (3,8)
4 (2,6)
4 (2,6)
5 (2,7)
3 (1,5)
5 (2,7)
7 (5,10)
4 (2,6)
35 (31,41)
37 (32,43)
25 (16,34)
6 (1,11)
37 (32,41)
6 (4,8)
5 (2,7)
4 (2,6)
7 (5,9)
7 (4,9)
5 (3,7)
5 (2,8)
Materials and methods: We studied a population of 883
nurses working in ICUs, distributed in 79 Italian hospitals
(70.1 % female) and 509 nurses working in GMUs, distributed in 35 Italian hospitals (80.2 % female). We asked
them to fill in a form including: 1) general data and his/her
work environment; 2) different evaluation standardized
scales — the Hospital Anxiety and Depression Scale,
divided into anxiety (HAD A) and depression (HAD D)
status 0–7 ‘non cases’, 8–10 ‘doubtful cases’, 11–21
‘cases’; the S.T.A.I. scale, divided into acute anxiety (Y-1)
and chronic anxiety (Y-2) status; the Maslach Burnout
Inventory–Human Services Survey (MBI.) divided into
Emotional Exhaustion (EE), ≤18 ‘low’, 19–26 ‘average’,
≥27 ‘high’, Depersonalization (DP) and Personal AccomP233
plishment (PA). We also evaluated the different reasons of
anxiety through individual questions (higher value, more
anxiety): A1, a critically ill patient; A2, a young patient; A3,
an old patient; A4, a suicidal patient; A5, a terminal
patient; A6, presence of mechanical supports; A7, relationship with patients’ relatives. The comparison between
the two groups was performed by the Mann-Whitney Rank
Sum test and z-test; statistical significance was accepted
as P<0.05.
Results: The results, expressed as median value, with
25th and 75th percentile in brackets, are shown in Tables
1 and 2.
Table 1 also shows the proportions of nurses that had a
highest value of HAD A and M.B.I. EE.
Conclusions: Pathologic anxiety and emotional exhaustion
are more prevalent in nurses working in GMUs. Thus, contrary to a common belief, ‘stress’ is a more distinctive peculiarity of general medicine units than intensive care units.
Anxiety and depression in family members of ICU patients: ethical considerations regarding
decision-making capacity
F Pochard, E Azoulay, S Chevret, F Lemaire, P Hubert, P Canouï, JR Le Gall, JF Dhainaut and B Schlemmer
For the FAMIREA group, Saint-Louis hospital and University Paris 7, 1 av. C Vellefaux, 75010 PARIS, France
Background: Anxiety and depression have a major impact
on the ability to make decisions. Characterization of symptoms reflecting anxiety and depression in family members
visiting ICU patients may be of major relevance to the
ethics of involving family members in decision-making, particularly about end-of-life issues.
Methods: Prospective multi-center study in 43 French ICUs
(37 adult and 6 pediatric). Each unit included 15 patients
admitted for longer than two days. ICU characteristics and
data on the patient and family members were collected.
Family members completed the Hospital Anxiety and
Depression Scale (HADS) to allow evaluation of the prevalence and potential predictors of anxiety and depression.
Findings: 637 patients were included in the study and
920 family members completed the HADS. All items were
completed in 836 HADS questionnaires, which formed
the basis for this study. The prevalences of anxiety and
depression in family members were 69.1% and 35.4%,
respectively. Anxiety or depression were present in 72.7%
of family members and 84% of spouses. Factors predictive of anxiety in a multivariate model included patientrelated factors (absence of chronic disease), family-related
factors (spouse, female gender, desire for professional
psychological help, help being received by usual doctor)
and caregiver-related factors (absence of physician-nurse
meetings on a regular basis, absence of a room used only
for meetings with family members). The multivariate model
also identified three groups of factors predicting depression: patient-related (age), family-related (spouse, female
gender, not of French descent), and caregiver-related (no
waiting room, perceived contradictions in the information
provided by caregivers).
Interpretation and conclusion: More than two-thirds of
family members visiting ICU patients suffer anxiety or
depression. Their involvement in end-of-life decisions
should be carefully discussed.
Acknowledgement: This abstract is funded by the French Society of Critical Care Medicine.
Meeting the needs of ICU patient families: a multicentre study
E Azoulay, F Pochard, S Chevret, F Lemaire, M Mokhtari, J R LeGall, JF Dhainaut and B Schlemmer
For the FAMIREA group, St-Louis hospital and University Paris 7, 1 av. Vellefaux, 75010 PARIS, France
Background: ICU caregivers should specifically seek to
develop a collaborative relationship with ICU patient family
members, based on a rich and free exchange of information and aimed both at helping family members to cope
with their distress and at providing them with the opportunity to speak for the patient if necessary.
Methods: We conducted a prospective multicentre study
of family member satisfaction as evaluated using the Critical Care Family Needs Inventory (CCFNI) developed by
the Society for Critical Care Medicine. Forty-three French
ICUs (37 adult and 6 paediatric) participated in the study;
each was required to include 15 consecutive patients hospitalised for longer than two days. ICU characteristics,
demographic data on patients and family members and
questions about satisfaction were collected. The additive
score of all CCFNI items was the dependent variable.
Factors associated with dissatisfaction were identified
using multivariate analysis.
Results: 637 patients were included in the study, and
920 family members completed the questionnaire. The
multiple regression model identified 10 factors associated
with family dissatisfaction: (i) two patient-related factors,
namely age (OR 1.23) and admission for acute respiratory
failure (OR 1.62); (ii) three family-related factors, namely
younger age (OR 0.98), not the spouse (OR 1.81), and
not of French descent (OR 4.34); and (iii) five caregiverrelated factors, namely perceived contradictions in information given by caregivers (OR 3.20), information not
always provided by the same physician (OR 2.71), no
knowledge of the occupational identity of each caregiver
(OR 2.32), no help from the family’s own doctor (OR
1.45), and insufficient caregiver time spent giving information (OR 1.35).
Conclusion: Among family members visiting the patient,
spouses were more satisfied. The caregiver-related factors
identified in this study are amenable to intervention and
deserve to be the focus of future studies aimed at improving the satisfaction of ICU patient family members.
Acknowledgement: This abstract is funded by: the French Society of Critical Care Medicine.
Acute psychological reactions in relatives of patients treated in the intensive care unit (ICU)
A Melissaki*, T Paparrigopoulos†, A Efthymiou*, Ch Vadala*, G Kribeni* and E Pavlou‡
*ICU, Agios Savvas General Hospital, 171 Alexandras Ave., GR - 11522, Athens, Greece; †Athens University Medical School,
Department of Psychiatry, Eginition Hospital, ‡ICU, KAT General Hospital, Greece
Objective: The objective of the present work was the
investigation of the impact of admission and treatment of
patients in the intensive care unit (ICU) on their relatives,
by measuring their acute psychological reactions. More
explicitly, to disclose if ICU treatment acts as a stressor of
a potentially traumatic nature to the patient’s relatives.
Material and methods: The sample comprised of 48 subjects (24 males/24 females) who were close relatives of
patients treated in the ICU of Agios Savvas and KAT
General Hospitals for various reasons over a six-month
period. Demographic data regarding both the patients and
their relatives were collected. Information on the physical
condition of patients (type of operation, APACHE II Score,
assisted ventilation etc.) was obtained from their medical
report. The instruments applied to assess the psychological impact of ICU admission on the patient’s relatives were
the Center for Epidemiological Studies Depression Scale
(CES–D), the Spielberger’s State-Trait Anxiety Inventory
(STAI) and the Impact of Event Scale (IES), which is a
measure of intrusion and avoidance. Subjects were
assessed one week±one day after their relative’s admission to the ICU.
Results: Patients were admitted to the ICU for various
reasons (injuries, 11; head injury, 12; postoperative care,
8; various emergencies, 17). Mean age of patients was
48.6±19.5 years (min, 17; max, 81); mean APACHE II
Score upon admission was 12.8±5.0 (min, 2; max, 24).
Mean age of relatives was 40.5±12.3 years (min, 21; max,
73). They were first degree relatives; 70% were married
and had children. Mean scores±SD on the psychometric
scales were as follows: CES–D=29.8±8.9, STAI—
State=60.4±13.2, STAI—Trait=39.8±9.0, IES—Total=
26.15±7.9, IES—Intrusion=16.0±4.7, IES—Avoidance=
10.1±4.8. These scores indicate the presence of severe
depressive symptoms, high levels of subjective tension
and apprehension, and moderate levels of intrusion and
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
avoidance. Females scored significantly higher than males
on all scales (Mean ±SD and t-test 2-tail sig: CES–D=
32.8±7.4 vs 26.8±9.2 P=0.02; STAI—State= 65.5±10.8
vs 55.3±13.5 P=0.006; STAI—Trait=42.6±8.6 vs
36.8±8.7 P=0.025; IES—Intrusion=17.6±3.6 vs 14.5±5.2
P=0.02; IES—Avoidance=11.1±5.0 vs 9.0±4.5 P=0.13).
Spearman correlation coefficients were also computed
between the different scales that were found to be highly
correlated (level of sig. >0.01). Stepwise multiple linear
regression of CES–D upon various independent variables
(i.e. age of patient, ICU diagnosis, APACHE II Score, age
and gender of relative, degree of kinship, relative’s socioeconomic and marital status, and the score on STAI–Trait
inventory) showed that the severity of the depressive
symptomatology depends significantly solely on the relatives STAI–Trait score.
Conclusions: The main findings of the present study are:
a) treatment in the ICU acts as a high impact stressor for
the patient’s relatives, equivalent to other traumatic events
of an exceptionally threatening nature; b) females are significantly more prone than males to develop a severe
anxiety-depressive reaction; they also experience more
intense intrusive symptoms; and c) individuals with higher
pre-exposure anxiety levels, particularly females, are more
vulnerable to the development of an acute stress reactiontype symptomatology.
Severity of illness, critical events, organ failure assessment and ICU outcome
M Silva Alvaro, RE Nap, V Fidler and D Reis Miranda
Health Services Research Unit, University Hospital Groningen, P.O. Box 30.001, 9700 RB Groningen, The Netherlands
Background: The use of general outcome prediction
models in the ICU remains controversial: 1) none of the
existing systems is able to precisely predict individual
outcome; 2) based on data collected within the first 24 h
after admission, their prediction power is insensitive to
alternative courses of care (and of ways-of-working) in the
Objectives: To evaluate the importance of intermediate
outcomes of care upon the final outcome of patient care in
the ICU; to evaluate whether the consideration of intermediate outcomes of care do increase the predictive power
of SAPS II score.
SOFA day 1
CrEv day 1
SOFA day 3
CrEv day 3
SOFA day 5
Design: Prospective, multi-center and multinational study.
CrEv day 5
1.0302–1.0526 0.0000
1.6787–5.7809 0.0003
95% CI
36.4±16.3 1.0480 1.0376–1.0586 0.0000
1.2611 1.1753–1.3532 0.0000
2.7951 2.0528–3.8056 0.0000
36.4±16.3 1.0379 1.0253–1.0525 0.0000
1.3086 1.2047–1.4215 0.0000
2.7118 1.7977–4.0908 0.0000
Setting: 47 ICU’s from 9 European countries.
Methods: All consecutive admissions were enrolled
during a four-month period.
Patient data: admission data; first day SAPS II score;
hourly registration of Critical Events (CrEv) defined as the
duration (in hours) of out-of-range measurements of four
parameters (heart rate, blood pressure, urine output and
oxygen saturation); Sequential Organ Failure Score
(SOFA) at admission and then every 24 h; ICU outcome.
Measurements: Final outcome: ICU mortality. Intermediate
outcomes: SOFA and CrEv, expressed as the percentage
of time spent in CrEv (single or combined). SOFA computations included total daily score and Delta-SOFA [1].
Analysis: variables included for explaining final outcome
(the dependent variable) on Day 3 through Day 10: Day 1
SAPS II score; intermediate outcome variables (SOFA
and CrEv) on Day 1, Day 3 and Day 5.
Results: Data on 1633 patients were analysed: median age
of 69 years; median SAPS II score of 35 and ICU mortality
rate of 14%. Including the variables indicated in a logistic
regression, three models (Table) could be constructed.
Conclusion: Confirming previous studies, the predictive
power of first day SAPS II score decreases over time. The
inclusion of intermediate outcomes contributes, significantly, to explain ICU mortality. This study strongly suggests the importance of accurate control of processes of
care (and the way of working) in the ICU: showing that the
incidence and time spent in out-of-range measurements
are clearly associated to the final outcome of ICU-patients.
Moreno R et al.: The use of maximum SOFA score to quantify
organ dysfunction/failure in intensive care. Results of a prospective, multi centre study. Intensive Care Med 1999, 25:686–696.
An alternative, and more sensitive, approach to detecting differences in outcome in sepsis
RS Wax, WT Linde-Zwirble, M Griffin, MR Pinsky and DC Angus
Department of Anesthesiology and Critical Care Medicine, University of Pittsburgh School of Medicine; Pittsburgh, PA 15261,
When comparing a characteristic (e.g. outcome) between
two groups, tests of continuous (as opposed to categorical) data that assume parametric (as opposed to nonparametric) distributions are the most powerful. Currently,
we measure outcome differences in sepsis trials in two
ways. Typically, we compare mortality rates at a given
time-point using categorical, parametric tests (e.g. χ2 or
Fisher’s Exact test of differences in mortality at day 28) or
we compare survival times using categorical, non-parametric tests (e.g. the Log-rank test to compare KaplanMeier curves). But survival after sepsis decreases
exponentially [1]. Thus, survival could be described by
exponential curves, which can be compared using continuous, parametric tests, such as the Cox’s F-test. We
therefore used this approach in a cohort of septic patients
to determine sample size requirements in comparison to
traditional approaches.
Patients: 1102 patients with severe sepsis enrolled in a
US multi-center trial.
Sub-groups: We divided patients into those with and
without septic shock to select two groups with a difference in survival (10–15%) typical for many power calculations in sepsis trials.
Statistical procedures: For each sub-group, we plotted the
survival to day 28 and fit distributions with exponential
curves using the maximum likelihood procedure. Curves
were then compared using the Cox’s F-test. We also compared differences in outcome using the Fisher’s Exact test
(for day-28 mortality) and the log-rank test. Statistical significance was assumed at P<0.05.
Sampling procedure: After comparing tests on the entire
sample, we then drew progressively smaller random
samples of the cohort and repeated the test comparisons
to determine the point at which statistical significance was
lost for each test.
Results: Patients with shock had a higher mortality than
those without shock (see Table). This difference was statistically significant by Fisher’s Exact and Log-rank tests
until sample size fell below 500. Survival in all sub-groups
was modeled by exponential curves with excellent fit (R2
>0.98). Comparing these curves by Cox’s F-test, statistical significance was maintained with a much smaller
sample size (see Fig.).
Conclusion: Taking advantage of the parametric distribution that characterizes survival after sepsis, we can apply a
test that finds statistical differences in survival with smaller
sample sizes than traditional approaches. These data
suggest that the application of exponentially-modeled sur-
Proportion of cases
Day 28 mortality (%)
No shock
No shock
Fisher’s Exact
Cox’s F
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
vival curve comparisons may be the preferred approach in
studies with small sample sizes, such as Phase II trials.
Furthermore, this approach may prove to be generally
preferable to categorical survival data comparisons, such
as day-28 mortality.
Knaus et al.: JAMA 1993, 270:1233–1241.
Evaluation of the logistic organ dysfunction system for the assessment of organ dysfunction and
PGH Metnitz*, Th Lang†, B Kabon*, H Steltzer* and JR Le Gall‡
*Department of Anesthesiology and General Intensive Care, Univ. Hospital of Vienna, Waehringer Guertel 18-20, A-1090
Vienna, Austria; †Department of Medical Statistics, Univ. of Vienna, Austria; ‡Department of Intensive Care Medicine, Saint-Louis
Hospital, Paris, France
Objective: To evaluate the performance of the Logistic
Organ Dysfunction (LOD) system for the assessment of
morbidity and mortality in multiple organ dysfunction/
failure (MOD/F) in an independent database and to evaluate the use of sequential LOD measurements for the prediction of outcome.
Design: Prospective, multicentric cohort study.
Setting: Thirteen adult medical, surgical, and mixed intensive care units (ICUs) in Austria.
vs 6 median score, P<0.05). Within organ systems, higher
levels of the severity of organ dysfunction were consistently associated with higher mortality. For the prediction
of hospital mortality, the original prognostic LOD model
did not perform well in our patients as indicated by the
goodness-of-fit statistic (C=37.2, 9 df, P<0.0001). Using
multiple logistic regression, we developed a prognostic
model (using the LOD of the first ICU day) with a satisfactory fit in our patients. The integration of further measurements during the ICU stay could not increase the
accuracy of the prediction.
Patients: A total of 2893 consecutive admissions to 13
adult medical/surgical intensive care units (ICUs) in Austria.
Outcome measure: Patient vital status at ICU- and hospital discharge.
Results: Univariate analysis showed that the LOD was
able to distinguish between survivors and nonsurvivors (2
Conclusions: The LOD system can be used to quantify
the baseline severity of organ dysfunction. Moreover, after
customization of the predictive equation, the LOD was
able to predict hospital mortality in our patients with high
precision. It thus provides a combined measure of morbidity and mortality for critically ill patients with MOD/F.
Predictive accuracy of three scoring systems in a single intensive care unit
PS Martins and S Blecher
Hospital Santa Marcelina, Rua Costa Gomes, 178 Vila Canaã, CEP 04382-130, São Paulo, Brazil
Objective: To evaluate the predictive accuracy of the
severity of three scoring systems — SAPS II, MPM II and
LODS — in a single Intensive Care Unit (ICU).
Design: Prospective cohort study.
Setting: General medical and surgical ICU in a tertiary
teaching hospital in city of São Paulo, Brazil.
Patients: 318 consecutive, unselected patients over the
period from February to June of 1999. Cardiac surgical
and burns patients were excluded.
Measurements and statistical analysis: Data required to
calculate the patient’s predicted mortality by the considered scoring systems were collected. Area under the
receiver operating characteristic curve (ROC curve) was
calculated to evaluate discrimination of each scoring
system. The goodness-of-fit statistic was used to assess
calibration of the models and the strength of the association between the predicted and observed outcome.
Main results: The mean age was 55.2±19 years. The
length of ICU stay was 10.64±13.09 days, and the mortality rate was 26.1 %. Calibration and discrimination results
and the predicted risk of death are shown in the Table.
Outcome measure: Patients vital status at ICU discharge.
Interventions: None.
Conclusions: The results showed good discrimination for
all the models.
Predicted risk
of death
ROC curve
MPM 24
LODS, MPM 0 and SAPS II showed good calibration, i.e,
there were no significant discrepancy between predicted
and observed mortality. MPM 24 showed a poor fit.
According to this, MPM0 can be used as a predictor of
death at patients admission, SAPS II as a predictor of
death in the first 24 h of admission, and LODS as end
point in studies of organ dysfunction in our ICU.
Scoring systems need to be evaluated in terms of predictive accuracy within a single institution before applying
them to make quality of care assessments.
The predictive value of serial acute physiology score (APS) and simplified acute physiology score
II (SAPS II) in post-operative liver transplant patients
Y Arabi, R Goraj, R McIntyre and A Al-Abdulkareem
Departments of Intensive Care and Hepatobiliary Sciences, MC 1425, PO Box 22490, King Fahad National Guard Hospital,
Riyadh, 11426, Saudi Arabia
Purpose: To evaluate the predictive value of daily acute
physiology score (APS) of the APACHE II system and simplified acute physiology score II (SAPS II) in postoperative
liver transplant patients.
Methods: Records of 20 consecutive postoperative liver
transplant patients were reviewed. Daily APS and SAPS II
scores were calculated for the first three days. Outcome
at hospital discharge was documented. Values were compared using paired t-test and Kruskal-Wallis test.
Results: APS and SAPS II were not significantly different
between survivors and nonsurvivors on day 1 and 2 but
were different on day 3 (see Table). APS and SAPS II
decreased from day 1 to 2 in both survivors and nonsur-
vivors (-29% vs -31%, P=NS for APS, -24% vs–21%,
P=NS for SAPS II). However, while APS and SAPS II continued to decrease from day 2 to 3 in survivors it
increased in nonsurvivors (–15% vs +8%, P=0.03 for
APS, –7% vs +5%, P=0.26).
Conclusions: In the postoperative liver transplant patients,
the following are concluded: 1) in contrast to the progressive decrease in APS and SAPS II in survivors, these
scores increased in nonsurvivors between day 2 and 3
after an initial drop; 2) day 3 APS and SAPS II scores
better differentiated survivors from nonsurvivors than day 1
and 2 scores; 3) a mortality prediction system based on
day 3 score or on serial scores is likely to be more accurate
than the traditional systems using the first 24 h data.
Day 1
Day 2
Survivors (n=16)
Nonsurvivors (n=4)
Day 3
Day 1
Day 2
Day 3
P<0.05 compared to previous day, b: P<0.05 for survivors compared to nonsurvivors.
TISS 76 and TISS 28: outcome discrimination and correlation with length of ICU/hospital stay in
303 consecutive patients of a medical intensive care unit
U Janssens, C Graf, J Graf and P Hanrath
Medical Clinic I, RWTH Aachen, Pauwelsstr. 30, 52057 Aachen, Germany
Objectives: TISS 76 (T76) and the simplified version
TISS 28 (T28) quantify nursing workload and allow evaluation of severity of illness and, therefore, may be applied for
outcome discrimination. The objective of this study was to
compare T76 and T28 with regard to outcome discrimination in patients of a medical intensive care unit (ICU) and
correlation with ICU length of stay (LOS) and hospital
Methods: All patients who stayed >24 h in ICU were
included between 11/97 and 2/98. T76, T28 were collected daily. Discrimination power for survivors (S) and
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
nonsurvivors (NS) was assessed by the area under the
Receiver Operating Characteristic (AUROC) curve. Relation of both scores on day 1 as well as last measured
score before ICU discharge (T76x, T28x) with ICU/hospital LOS and ICU/hospital mortality was assessed with
Pearson’s correlation and logistic regression.
Results: 303 patients (216 male [71.3%], 62±12 years,
ICU LOS 3.7±4.7 days, hospital LOS 15.1±13.2 days,
SAPS II 26±13) were studied. ICU mortality was 7.6%,
hospital mortality was 14.5%. AUROC for T76 day 1 was
0.79±0.04 for T28 0.76±0.05. Correlation between
T76/T28 day 1 and ICU/hospital LOS was only weak.
T76x and T28x did not correlate with hospital LOS. Risk of
death was associated with T76 (odds ratio 1.1, 95% CI
1.04 to 1.18) but not with T28 or T76x/T28x.
Conclusion: Outcome discrimination with T76 and T28
on day 1 was reliable. T76 was slightly superior to T28.
Correlation with ICU/hospital LOS was only weak. Risk of
death was significantly associated with T76 on day 1. Last
measured score before ICU discharge did not correlate
with hospital LOS or mortality.
Severity evaluation in acute pancreatitis: the role of SOFA score and general severity scores
R Matos, R Moreno and T Fevereiro
UCIP, Hospital de St António dos Capuchos, Alameda de St. António dos Capuchos, 1150 Lisboa, Portugal
Introduction: Severity stratification in acute pancreatitis
has long been a subject of debate. The availability of
instruments specific for this pathologic condition lead
some intensivists to argue for their use in this condition.
However, to the best of our knowledge, no published
study competed all these scores with general severity
scores and organ failure scores on the same cohort. The
objective of this work is to compare six disease-specific
scores with two general severity scores (APACHE II and
SAPS II) and one organ failure score (Sequential Organ
Failure Assessment [SOFA] score) in patients admitted
with acute pancreatitis to a mixed medical-surgical ICU.
Material and methods: We analysed all the patients discharged from the UCI from July 1 1991 to November 30
1999 with a diagnosis of acute pancreatitis. Basic demographic and clinical data were registered, as were
outcome at ICU and hospital discharge as well as
APACHE II, SAPS II, SOFA score (at admission, 24 h, 48
h and maximum during ICU stay), admission Ramson
score, Ramson score at 48 h, Imrie score, Osborn score,
Blamey score, Balthasar score, collected according to the
original descriptions. Raw data necessary for the computation of the scores has been registered prospectively,
using a proprietary computerised system.
The discriminative power of the scores was evaluated
through the use of the area under the Receiver Operating
Characteristics (ROC) curve. Two-sample student T-test
was used for the comparison of survivors and nonsurvivors.
The outcome measure used was vital status at hospital discharge.
Results: During the study period, 49 patients were discharged with a diagnosis of acute pancreatitis. Biliar tract
disease (n=26) and alcoholism (n=8) where the most
Mean ±SD
Survivors mean ±SD Non-survivors mean ±SD
Area under ROC
curve (SE)
SAPS II score
SOFA admission
SOFA 24 h
SOFA 48 h
SOFA maximum
Ramson 0 score
Ramson 48 score
Imrie score
Osborn score
Blamey score
Balthasar score
common aetiologies. In 13 patients no aetiology could be
found. Mean age was 53.3±16.6 years, lower in survivors
than in nonsurvivors (45.3±16.0 vs 63.1±11.9, P<0.001).
Median (interquartile range) length of stay in the ICU was
8.09 days (4.0 to 22.1 days), similar in survivors and nonsurvivors (P=NS). Mortality in the ICU was 34.7% (17
patients) with a corresponding hospital mortality of 44.9%
(22 patients). Mean ±standard deviation in the global population, in survivors and in nonsurvivors are presented in
the Table, together with the area under ROC curve±standard error for all the scores.
Conclusions: In this cohort of patients with acute pancreatitis, general severity scores and SOFA score presented
a better discriminative capability than disease-specific
scores, and the results support their use in risk stratification in these patients. If we take into account the lower
work needed for the computation of the SOFA score,
when compared to APACHE II and SAPS II, our results
support the use of SOFA score, both at admission and
latter during the ICU stay, for risk stratification in patients
with acute pancreatitis.
Therapeutic Intervention Scoring System (TISS) — a method for calculating costs in the intensive
care unit (ICU) and intermediate care unit (IMCU)
I Kaufmann and J Briegel
Department of Anesthesiology, Hospital of the University of Munich, Marchioninistraße 15, D-81377 Munich, Germany
The Therapeutic Intervention Scoring System (TISS) quantifies type and number of intensive care treatments. This
system, therefore, indicates the work load of intensive care
and may be used for calculating costs in the ICU. In the
present study, TISS scoring was prospectively performed
daily on 653 patients in a 14-bed multidisciplinary ICU of
a 1400-bed university hospital. In addition, we scored 952
patients after major surgical procedures treated in a 4-bed
IMCU affiliated to a 24-h recovery room. Besides TISS, we
registered age, diagnosis, and length of stay in the unit in
each patient. Data were entered from January 1 1998 to
December 31 1998. A retrospective cost analysis was
performed. The cost of each TISS point was derived from
the division of the costs for clinical support services, consumption and staff for all patients in 1998 and the sum of
TISS points of all patients in 1998. To determine the cost
of each patient, the TISS point price was multiplied with
the sum of daily and total TISS points of the patients.
In the ICU, the total accumulated TISS score was 174 950
points. This gave an average cumulative TISS score of 268
in these patients. The median length of stay was 10 days
and the median SAPS II score was 23 points. The mortality
reached 5.05%. A cost analysis revealed a value of 74.66
DM (38.17 Euro) per TISS point, which can be used to calculate the average costs for groups of patients. The daily
cost of ICU care was 3183 DM (1627.44 Euro) per patient
in survivors and 4079 DM (2085.56 Euro] in nonsurvivors. For mechanically ventilated patients the daily costs
averaged 3404 DM (1740.44 Euro], for patients with
spontaneous breathing 2501 DM (1278.74 Euro). Daily
cost also depended on age (<60 years, 3367 DM
[1721.52 Euro]; 60–70 years, 2463 DM [1259.31
Euro]; >70 years, 3487 DM [1782.87 Euro]) and length
of stay (<24 h, 2733 DM [1397.36 Euro]; 2–4 days,
2613 DM [1336.01 Euro]; ≥5 days, 3658 DM [1870.31
Euro]). Liver transplant recipients consumed 3621 DM
(1851.39 Euro) per day, patients after lung transplantation
2971 DM (1519.05 Euro) per day. Patients after major
surgical procedures staying less than 24 h reached daily
cost of 2733 DM (1397.36 Euro).
In the IMCU, the total accumulated TISS-score was
39,765 points. This gave an average cumulative TISS
score of 42 in these patients. The median length of stay
was 1.4 days, the median SAPS II score 18 points and the
mortality rate 0%. A cost analysis revealed a value of
78.99 DM (40.39 Euro) per TISS point. Postoperative
patients in the intermediate care unit staying <24 h consumed 2562 DM (1309.93 Euro) daily.
In conclusion, in this study we could demonstrate an activity based costing methodology which measures the
patient-related costs of care in both, ICU and IMCU. The
TISS is a valuable tool when evaluating costs in the ICU
and in the IMCU. Charging patients to diagnosis related
groups allows calculation of costs and resource allocation
to different specialties. Comparing the ICU and the IMCU
there was a daily cost difference of 171 DM (87.43
Euro) per patient due to a lower workload and lower costs
of the nursing staff in the IMCU. These findings support
the potential cost saving of managing patients in an IMCU.
Managing cost in ICU — a report of an experience
CP Souza Paulo, F Leibel and MD Magalhães
ICU Hospital de Clínicas de Niterói, Rua José Higino, 214/1104–CEP: 20520-200, Rio de Janeiro, Brazil
Background: Costs have a huge impact in ICU bills and
are the subject of many articles in the medical literature.
To reduce costs without decreasing quality is a challenge.
Objectives: To demonstrate a cost management program
in a 12-bed ICU of a general hospital with 72 beds. We’ve
chosen three drugs of great impact in the bill and we’ve
Critical Care
Vol 4 Suppl 1
20th International Symposium on Intensive Care and Emergency Medicine
established a program that rationalized or replaced these
drugs and a way to control if there was any negative
impact on quality.
Methods: The period of the study was from January 1998
to July 1999. The three drugs chosen were human
albumin, omeprazole and midazolan. To decide when and
how to use albumin and omeprazole, we employed evidence-based medicine concepts; with midazolan, we
changed to diazepam, in equivalent doses, and controlled
the “ideal doses” by the Ramsay scale.
Results: We achieved a considerable reduction in global
costs, specifically US$ 20 000 for albumin, US$ 3200 for
omeprazole and US$ 2000 for midazolan, each month.
The global economy was US$ 277 000 a year, without
major mortality (compared to predicted by the APACHE II
score), increased length of stay (LOS), discomfort or difficulties inside the medical group. The benefits are maintained until now (19 months later).
Conclusion: Control costs without loss of quality is a
challenge and a duty. We proved that this is possible
without any problem. It's necessary to have clear targets
and scientific basis, as evidence-based medicine, in conducting the protocols. Other areas are being included in
this approach in our group.

Similar documents


Report this document