Clinical and cost-effectiveness of electroconvulsive therapy for depressive illness, schizophrenia,

Document technical information

Format pdf
Size 731.2 kB
First found Jun 9, 2017

Document content analysis

Language
English
Type
not defined
Concepts
no text concepts found

Persons

John Hanson
John Hanson

wikipedia, lookup

Leonardo da Vinci
Leonardo da Vinci

wikipedia, lookup

Organizations

Places

Transcript

Health Technology Assessment 2005; Vol. 9: No. 9
Electroconvulsive therapy for depressive illness, schizophrenia, catatonia and mania
Feedback
The HTA Programme and the authors would like to know
your views about this report.
The Correspondence Page on the HTA website
(http://www.ncchta.org) is a convenient way to publish
your comments. If you prefer, you can send your comments
to the address below, telling us whether you would like
us to transfer them to the website.
We look forward to hearing from you.
Health Technology Assessment 2005; Vol. 9: No. 9
Clinical and cost-effectiveness of
electroconvulsive therapy for
depressive illness, schizophrenia,
catatonia and mania: systematic
reviews and economic modelling
studies
J Greenhalgh, C Knight, D Hind,
C Beverley and S Walters
March 2005
The National Coordinating Centre for Health Technology Assessment,
Mailpoint 728, Boldrewood,
University of Southampton,
Southampton, SO16 7PX, UK.
Fax: +44 (0) 23 8059 5639
Email: [email protected]
http://www.ncchta.org
Health Technology Assessment
NHS R&D HTA Programme
ISSN 1366-5278
HTA
HTA
How to obtain copies of this and other HTA Programme reports.
An electronic version of this publication, in Adobe Acrobat format, is available for downloading free of
charge for personal use from the HTA website (http://www.hta.ac.uk). A fully searchable CD-ROM is
also available (see below).
Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public and
private sector purchasers from our Despatch Agents.
Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is
£2 per monograph and for the rest of the world £3 per monograph.
You can order HTA monographs from our Despatch Agents:
– fax (with credit card or official purchase order)
– post (with credit card or official purchase order or cheque)
– phone during office hours (credit card only).
Additionally the HTA website allows you either to pay securely by credit card or to print out your
order and then post or fax it.
Contact details are as follows:
HTA Despatch
c/o Direct Mail Works Ltd
4 Oakwood Business Centre
Downley, HAVANT PO9 2NP, UK
Email: [email protected]
Tel: 02392 492 000
Fax: 02392 478 555
Fax from outside the UK: +44 2392 478 555
NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of
£100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300
per volume. Please see our website for details. Subscriptions can only be purchased for the current or
forthcoming volume.
Payment methods
Paying by cheque
If you pay by cheque, the cheque must be in pounds sterling, made payable to Direct Mail Works Ltd
and drawn on a bank with a UK address.
Paying by credit card
The following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard,
Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email.
Paying by official purchase order
You can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK.
We cannot at present accept purchase orders from commercial companies or from outside the UK.
How do I get a copy of HTA on CD?
Please use the form on the HTA website (www.hta.ac.uk/htacd.htm). Or contact Direct Mail Works (see
contact details above) by email, post, fax or phone. HTA on CD is currently free of charge worldwide.
The website also provides information about the HTA Programme and lists the membership of the various
committees.
Clinical and cost-effectiveness of
electroconvulsive therapy for
depressive illness, schizophrenia,
catatonia and mania: systematic
reviews and economic modelling
studies
J Greenhalgh,1*† C Knight,2 D Hind,2 C Beverley2
and S Walters2
1
2
Nuffield Institute for Health, University of Leeds, UK
The School of Health and Related Research (ScHARR),
University of Sheffield, UK
* Corresponding author
† Current affiliation: Health Care Practice R&D Unit, University of Salford, UK
Declared competing interests of authors: none
Published March 2005
This report should be referenced as follows:
Greenhalgh J, Knight C, Hind D, Beverley C, Walters S. Clinical and cost-effectiveness of
electroconvulsive therapy for depressive illness, schizophrenia, catatonia and mania:
systematic reviews and economic modelling studies. Health Technol Assess 2005;9(9).
Health Technology Assessment is indexed and abstracted in Index Medicus/MEDLINE,
Excerpta Medica/EMBASE and Science Citation Index Expanded (SciSearch®) and Current
Contents®/Clinical Medicine.
NHS R&D HTA Programme
T
he research findings from the NHS R&D Health Technology Assessment (HTA) Programme directly
influence key decision-making bodies such as the National Institute for Clinical Excellence (NICE)
and the National Screening Committee (NSC) who rely on HTA outputs to help raise standards of care.
HTA findings also help to improve the quality of the service in the NHS indirectly in that they form a key
component of the ‘National Knowledge Service’ that is being developed to improve the evidence of
clinical practice throughout the NHS.
The HTA Programme was set up in 1993. Its role is to ensure that high-quality research information on
the costs, effectiveness and broader impact of health technologies is produced in the most efficient way
for those who use, manage and provide care in the NHS. ‘Health technologies’ are broadly defined to
include all interventions used to promote health, prevent and treat disease, and improve rehabilitation
and long-term care, rather than settings of care.
The HTA programme commissions research only on topics where it has identified key gaps in the
evidence needed by the NHS. Suggestions for topics are actively sought from people working in the
NHS, the public, consumer groups and professional bodies such as Royal Colleges and NHS Trusts.
Research suggestions are carefully considered by panels of independent experts (including consumers)
whose advice results in a ranked list of recommended research priorities. The HTA Programme then
commissions the research team best suited to undertake the work, in the manner most appropriate to find
the relevant answers. Some projects may take only months, others need several years to answer the
research questions adequately. They may involve synthesising existing evidence or designing a trial to
produce new evidence where none currently exists.
Additionally, through its Technology Assessment Report (TAR) call-off contract, the HTA Programme is
able to commission bespoke reports, principally for NICE, but also for other policy customers, such as a
National Clinical Director. TARs bring together evidence on key aspects of the use of specific
technologies and usually have to be completed within a limited time period.
Criteria for inclusion in the HTA monograph series
Reports are published in the HTA monograph series if (1) they have resulted from work commissioned
for the HTA Programme, and (2) they are of a sufficiently high scientific quality as assessed by the referees
and editors.
Reviews in Health Technology Assessment are termed ‘systematic’ when the account of the search,
appraisal and synthesis methods (to minimise biases and random errors) would, in theory, permit the
replication of the review by others.
The research reported in this monograph was commissioned and funded by the HTA Programme on
behalf of NICE as project number 01/48/01. The authors have been wholly responsible for all data
collection, analysis and interpretation and for writing up their work. The HTA editors and publisher
have tried to ensure the accuracy of the authors’ report and would like to thank the referees for their
constructive comments on the draft document. However, they do not accept liability for damages or
losses arising from material published in this report.
The views expressed in this publication are those of the authors and not necessarily those of the HTA
Programme, NICE or the Department of Health.
Editor-in-Chief:
Series Editors:
Managing Editors:
Professor Tom Walley
Dr Peter Davidson, Professor John Gabbay, Dr Chris Hyde,
Dr Ruairidh Milne, Dr Rob Riemsma and Dr Ken Stein
Sally Bailey and Caroline Ciupek
ISSN 1366-5278
© Queen’s Printer and Controller of HMSO 2005
This monograph may be freely reproduced for the purposes of private research and study and may be included in professional journals provided
that suitable acknowledgement is made and the reproduction is not associated with any form of advertising.
Applications for commercial reproduction should be addressed to NCCHTA, Mailpoint 728, Boldrewood, University of Southampton,
Southampton, SO16 7PX, UK.
Published by Gray Publishing, Tunbridge Wells, Kent, on behalf of NCCHTA.
Printed on acid-free paper in the UK by St Edmundsbury Press Ltd, Bury St Edmunds, Suffolk.
T
Health Technology Assessment 2005; Vol. 9: No. 9
Abstract
Clinical and cost-effectiveness of electroconvulsive therapy for
depressive illness, schizophrenia, catatonia and mania:
systematic reviews and economic modelling studies
J Greenhalgh,1*† C Knight,2 D Hind,2 C Beverley2 and S Walters2
1
2
Nuffield Institute for Health, University of Leeds, UK
The School of Health and Related Research (ScHARR), University of Sheffield, UK
* Corresponding author
† Current affiliation: Health Care Practice R&D Unit, University of Salford, UK
Objectives: To establish the clinical effectiveness and
cost-effectiveness of electroconvulsive therapy (ECT)
for depressive illness, schizophrenia, catatonia and
mania.
Data sources: Electronic bibliographic databases. The
reference lists of relevant articles and health services
research-related resources were consulted via the
Internet.
Review methods: Identified studies were examined to
ascertain whether they met the inclusion criteria for
the review. The study quality of relevant articles was
assessed using standard checklists and data were
abstracted using standardised forms into a database.
Where relevant, results from studies were pooled for
meta-analysis. Two economic models were developed
primarily based on evidence from the clinical
effectiveness analysis and limited quality of life
studies.
Results: Two good-quality systematic reviews of
randomised evidence of the efficacy and safety of ECT
in people with depression, schizophrenia, catatonia and
mania were identified. Four systematic reviews on nonrandomised evidence were also identified, although
only one of these could be described as good quality.
There was no randomised evidence of the effectiveness
of ECT in specific subgroups including older people,
children and adolescents, people with catatonia and
women with postpartum exacerbations of depression
or schizophrenia. The economic modelling results for
depression did not demonstrate that any of the
scenarios had a clear economic benefit over the
others, mainly because of the uncertainty surrounding
the clinical effectiveness of the different treatments
and the quality of life utility gains. Sensitivity analysis
surrounding the cost of ECT and the quality of life
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
utility values had little effect on the overall results.
The results of the model for schizophrenia adapted
to include ECT suggest that clozapine is a cost-effective
treatment compared with ECT. For patients who
fail to respond to clozapine, ECT treatment may
be preferred to the comparative treatment of
haloperidol/chlorpromazine.
Conclusions: Real ECT is probably more effective than
sham ECT, but as stimulus parameters have an
important influence on efficacy, low-dose unilateral
ECT is no more effective than sham ECT. ECT is
probably more effective than pharmacotherapy in the
short term and limited evidence suggests that
ECT is more effective than repetitive transcranial
magnetic stimulation. Tricyclic antidepressants (TCAs)
may improve the antidepressant effect of ECT during
the course of treatment. Continuation
pharmacotherapy with TCAs combined with lithium in
people who have responded to ECT reduces the rate
of relapses. Overall, gains in the efficacy of the
intervention depending on the stimulus parameters of
ECT are achieved only at the expense of an increased
risk of cognitive side-effects. Limited evidence suggests
these effects do not last beyond 6 months, but there is
no evidence examining the longer term cognitive
effects of ECT. There is little evidence of the long-term
efficacy of ECT. ECT either combined with
antipsychotic medication or as a monotherapy is not
more effective than antipsychotic medication in people
with schizophrenia. More research is needed to
examine the long-term efficacy of ECT and the
effectiveness of post-ECT pharmacotherapy, the shortterm and longer term cognitive side-effects of ECT, and
the impact of ECT on suicide and all-cause mortality.
Further work is needed to examine the information
iii
Abstract
needs of people deciding whether to accept ECT and
how their decision-making can be facilitated. More
research is also needed on the mechanism of action of
ECT. Finally, the quality of reporting of trials in this area
would be vastly improved by strict adherence to the
Consolidated Standards of Reporting Trials
iv
recommendations. Economic analysis may identify
areas in which research would be best targeted by
identifying parameters where reducing the level of
uncertainty would have the most effect in helping to
make the decision on whether ECT is a cost-effective
treatment.
Health Technology Assessment 2005; Vol. 9: No. 9
Contents
List of abbreviations ..................................
vii
ix
8 Conclusions ................................................
Clinical effectiveness ..................................
Cost-effectiveness .......................................
81
81
81
Executive summary ....................................
1 Aim of the review ......................................
1
Acknowledgements ....................................
83
2 Background ................................................
Description of the underlying health
problem ......................................................
Current service provision ...........................
3
References ..................................................
85
3
4
Appendix 1 Electronic bibliographic
databases searched .....................................
95
3 Effectiveness ...............................................
Methods for reviewing effectiveness ..........
Results ........................................................
Conclusions and discussion ........................
11
11
13
47
Appendix 2 Other sources
consulted ....................................................
97
4 Economic analysis ......................................
Introduction ...............................................
Economic modelling of ECT for
depressive illness, schizophrenia,
catatonia and mania ...................................
49
49
Appendix 3 Search strategies used in
the major electronic bibliographic
databases ....................................................
99
49
5 Implications for other parties ....................
67
Appendix 5 Descriptions of included
studies ......................................................... 105
6 Factors relevant to the NHS .....................
69
Appendix 6 Results of meta-analyses ........ 141
7 Discussion ...................................................
Summary of main results and discussion ...
Assumptions, limitations and
uncertainties ...............................................
Need for further research ..........................
71
71
Health Technology Assessment reports
published to date ....................................... 157
78
78
Health Technology Assessment
Programme ................................................ 167
Appendix 4 Methodological search
filters used in Ovid MEDLINE .................. 103
v
Health Technology Assessment 2005; Vol. 9: No. 9
List of abbreviations
GAF
Global Assessment of Functioning
scale
Beck Depression Inventory
GDS
Geriatric Depression Scale
BFCRS
Bush–Francis Catatonia Rating
Scale
GRSD
Global Rating Scale for
Depression
BGT
Bender Gestalt Test
HAD
BNF
British National Formulary
Hospital Anxiety and Depression
scale
BPRS
Brief Psychiatric Rating Scale
HMIC
Health Management Information
Consortium
C.ATP
continuation therapy with
antipsychotic drugs
HRSD
Hamilton Rating Scale for
Depression
C.MAOI
continuation therapy with
monoamine oxidase inhibitors
ICD-10
International Classification of
Diseases-10
C.placebo
continuation therapy with
placebo
ITT
intention to treat
C.SSRI
continuation therapy with
selective serotonin reuptake
inhibitors
M
male
MADRS
Montgomery and Asberg
Depression Rating Scale
APA
American Psychiatric Association
BBB
Blood–brain barrier
BDI
C.TCA
continuation therapy with tricyclic
antidepressants
MAOI
monoamine oxidase inhibitor
CCTR
Cochrane Controlled Trials
Register
MDD
major depressive disorder
MEP
muscular-evoked potential
CDSR
Cochrane Database of Systematic
Reviews
MMPI
Minnesota Multiphasic
Personality Inventory
CGI
Clinical Global Impression
MMSE
Mini Mental State Examination
CI
confidence interval
MRC
Medical Research Council
CODS
Cronholme and Ottoson
Depression Scale
MRI
magnetic resonance imaging
NAA
N-acetylaspartate
CRD
Centre for Reviews and
Dissemination
NHB
net health benefit
CT
computed tomography
NHS EED
CVD
cardiovascular disease
NHS Economic Evaluation
Database
DARE
Database of Abstracts of Reviews
of Effectiveness
NICE
National Institute for Clinical
Excellence
DSM
Diagnostic and Statistical Manual
NMB
net monetary benefit
ECT
electroconvulsive therapy
NMS
neuroleptic malignant syndrome
F
female
continued
vii
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
List of abbreviations
List of abbreviations continued
NNH
number needed to harm
SCI
Science Citation Index
NNT
number needed to treat
SMD
standardised mean difference
NRS
Nurses’ Rating Scale
SNRI
NSF
National Service Framework
serotonin and norepinephrine
reuptake inhibitor
SSCI
Social Sciences Citation Index
SSRI
selective serotonin reuptake
inhibitors
SURE
Service User Research Enterprise
TCA
tricyclic antidepressant
UKU
Udvlag for Kliniske
Undersøgelser
OHE HEED Office of Health Economics
Health Economic Evaluations
Database
PANSS
Positive and Negative Symptoms
Scale
PIRS
Psychological Impairments Scale
PSE
Present State Examination
PSQI
Pittsburg Sleep Quality Index
VAS
visual analogue scale
QALY
quality-adjusted life-year
VBR
ventricular:brain ratio
RCP
Royal College of Psychiatrists
WAIS
Weschler Adult Intelligence Scale
RCT
randomised controlled trial
WBIS
RR
relative risk
Weschler–Bellevue Intelligence
Scale
rTMS
repetitive transcranial magnetic
stimulation
WMD
weighted mean difference
WMS
Weschler Memory Scale
All abbreviations that have been used in this report are listed here unless the abbreviation is well known (e.g. NHS), or
it has been used only once, or it is a non-standard abbreviation used only in figures/tables/appendices in which case
the abbreviation is defined in the figure legend or at the end of the table.
viii
Health Technology Assessment 2005; Vol. 9: No. 9
Executive summary
Objective
The aim of this review is to establish the clinical
effectiveness and cost-effectiveness of
electroconvulsive therapy (ECT) for depressive
illness, schizophrenia, catatonia and mania.
Background
ECT has been available for use since the 1930s.
It involves passing an electric current through a
person’s brain after they have been given a
general anaesthetic and muscle relaxants, to
produce a convulsion. There is a complex
interplay between the stimulus parameters of ECT,
including position of electrodes, dosage and
waveform of electricity, and its efficacy.
ECT is rarely used as a first line therapy, except in
an emergency where the person’s life is at risk as a
result of refusal to eat or drink or in cases of
attempted suicide. Current guidelines indicate
that ECT has a role in the treatment of people
with depression and in certain subgroups of
people with schizophrenia, catatonia and mania.
In England between January and March 1999
there were 16,482 administrations of ECT to 2835
patients, 85% of which were in an inpatient
setting. There were important variations in the
rates of administration of ECT by gender, age and
health region. Women received ECT more
frequently than men and the rates of
administration for both genders increased with
age. In England, rates of administration of ECT
are highest in the North West and lowest in
London.
Methods
Seventeen electronic bibliographic databases were
searched, covering biomedical, health-related,
science, social science and grey literature. In
addition, the reference lists of relevant articles were
checked and 40 health services research-related
resources were consulted via the Internet. These
included health technology assessment
organisations, guideline-producing bodies, generic
research and trials registers, and specialist
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
psychiatric sites. All abstracts were examined to
ascertain whether they met the inclusion criteria for
the review. The study quality of relevant articles was
assessed using standard checklists and data were
abstracted by two people using standardised forms
in a Microsoft Access database. Where relevant,
results from studies were pooled for meta-analysis.
Results and conclusions
Number and quality of studies
Two good-quality systematic reviews of
randomised evidence of the efficacy and safety of
ECT in people with depression, schizophrenia,
catatonia and mania were identified. Four
systematic reviews on non-randomised evidence
were also identified, although only one of these
could be described as good quality. There was no
randomised evidence of the effectiveness of ECT
in specific subgroups including older people,
children and adolescents, people with catatonia
and women with postpartum exacerbations of
depression or schizophrenia.
Summary of benefits/direction of
evidence
In people with depression, real ECT is probably
more effective than sham ECT, but stimulus
parameters have an important influence on
efficacy, low-dose unilateral ECT is no more
effective than sham ECT. ECT is probably more
effective than pharmacotherapy in the short term,
but the evidence on which this assertion is based
was of variable quality and inadequate doses of
pharmacotherapy were used. Limited evidence
suggests that ECT is more effective than repetitive
transcranial magnetic stimulation (rTMS). Limited
data suggest that tricyclic antidepressants (TCAs)
may improve the antidepressant effect of ECT
during the course of ECT, and that continuation
pharmacotherapy with TCAs combined with
lithium in people who have responded to ECT
reduces the rate of relapses. Overall, gains in the
efficacy of the intervention depending on the
stimulus parameters of ECT are achieved only at
the expense of an increased risk of cognitive sideeffects. Limited evidence suggests these effects do
not last beyond 6 months, but there is no evidence
examining the longer term cognitive effects of
ix
Executive summary
ECT. There is little evidence of the long-term
efficacy of ECT. There was much less evidence
regarding the efficacy of ECT in schizophrenia
and mania, and no randomised evidence of the
effectiveness of ECT in catatonia. ECT either
combined with antipsychotic medication or as a
monotherapy is not more effective than
antipsychotic medication in people with
schizophrenia. The evidence did not allow any
firm conclusions to be drawn regarding the
efficacy of ECT in people with mania or catatonia,
older people, younger people and women with
psychiatric problems, or the impact of ECT on
all-cause mortality. There was limited nonrandomised evidence regarding the impact of
patient acceptability and choice on the outcomes
of ECT, and this produced mixed results.
Cost-effectiveness
No previous analysis has been undertaken on the
cost-effectiveness of ECT in depression or
schizophrenia. Two economic models were
developed primarily based on evidence from the
clinical effectiveness analysis and limited quality of
life studies.
Depression
The economic model for depression was based on
a severely depressed population requiring
hospitalisation. As clinical opinion differs to
whether ECT should be used only as a last resort
treatment or whether it could be used earlier in
the treatment hierarchy, the model was
constructed to allow the evaluation of the costeffectiveness of ECT being provided as a first,
second or third line therapy.
Different scenarios that incorporated ECT as a
treatment were compared with a pharmacological
only treatment. The economic modelling results
did not demonstrate that any of the scenarios had
a clear economic benefit over the others. The
main reason for this was the uncertainty
surrounding the clinical effectiveness of the
different treatments and the quality of life utility
gains. Sensitivity analysis surrounding the cost of
ECT and the quality of life utility values had little
effect on the overall results.
x
Further economic analysis, such as expected value
of perfect information, may be able to identify
areas in which research would be best targeted by
identifying parameters where reducing the level of
uncertainty would have the most effect in helping
to make the decision on whether ECT is a costeffective treatment in the hospitalised severely
depressed population.
Schizophrenia
The main schizophrenic population for which
ECT is indicated in the guidelines of the American
Psychiatric Association and the Royal College of
Psychiatrists is patients resistant to
pharmacotherapy. Therefore, the economic model
constructed for schizophrenia was based on a
pharmacological model previously constructed
which was the only cost–utility study identified in
the treatment of schizophrenia. This model
analysed the cost-effectiveness of clozapine
compared with haloperidol/chlorpromazine
treatment in treatment-resistant schizophrenia.
The model was adapted to incorporate an ECT
arm to the decision tree analysis. The results of
the adapted model including ECT suggest that
clozapine is a cost-effective treatment compared
with ECT. For patients who fail to respond to
clozapine, ECT treatment may be preferred to the
comparative treatment of haloperidol/
chlorpromazine. However, the clinical evidence
underpinning the ECT assumptions in the model
is weak and the results should be interpreted with
caution.
Recommendations for
further research
Clinical effectiveness
There is a need for further, high-quality
randomised controlled trials (RCTs) of the use of
ECT in specific subgroups that are most likely to
receive this treatment. These include older people
with depression, women with postpartum
exacerbation of depression or schizophrenia and
people with catatonia. There is also a lack of good
quality randomised evidence of the effectiveness of
ECT in people with mania and people who are
resistant to pharmacotherapy in schizophrenia and
depression.
There is currently no randomised evidence
comparing ECT with, or in addition to newer
antipsychotic drugs (e.g. clozapine and
risperidone) and antidepressants (e.g. venlafaxine)
that are currently used in clinical practice. Further
work is needed in these areas. More research is
also needed to compare ECT with rTMS,
especially in people with schizophrenia. Again,
there is a need for further, high-quality RCTs
comparing the use of ECT with these treatments.
More research is needed to examine the long-term
efficacy of ECT and the effectiveness of post-ECT
pharmacotherapy. There is only limited evidence
regarding the efficacy of supplementing ECT with
Health Technology Assessment 2005; Vol. 9: No. 9
pharmacotherapy in people with depression and
the continuation of pharmacotherapy following
successful response to ECT to prevent relapses. In
most trials, the aftercare of people receiving ECT
was not randomised and people were rarely
followed up beyond the course of ECT. Future
work in the area requires longer follow-up periods.
Further work is also needed to develop ways of
incorporating patients’ perspectives on the impact
of ECT into future RCTs. Consideration should be
given to the use of both quantitative and
qualitative methods. The outcome measures used
should reflect both clinical and patient
perspectives on the impact of ECT.
term. There is also a need for longer term followup within RCTs to explore the impact of ECT on
suicide and all-cause mortality.
There is also little good-quality quantitative
evidence of the short-term and longer term
cognitive side-effects of ECT. Cognitive
functioning should be measured using wellvalidated instruments, and methods need to be
developed that also reflect patients’ concerns
regarding personal memory loss. These
instruments should be incorporated into trial
design at the outset, and hypotheses set and
results interpreted using a well-developed theory
or set of theories from cognitive psychology.
Again, longer term follow-up is needed as memory
losses may only become apparent in the longer
Finally, the quality of reporting of trials in this
area would be vastly improved by strict adherence
to the Consolidated Standards of Reporting Trials
(CONSORT) recommendations.
Further work is needed to examine the
information needs of people deciding whether to
accept ECT and how their decision-making can be
facilitated. The influence of these choices on the
perceived efficacy of ECT also requires further
exploration.
Despite over 50 years of research into ECT, there
is still no agreement on the mechanism of action
of ECT. More research is needed in this area.
Cost-effectiveness
Further economic analysis, such as expected value
of perfect information, may identify areas in which
research would be best targeted by identifying
parameters where reducing the level of
uncertainty would have the most effect in helping
to make the decision on whether ECT is a costeffective treatment.
xi
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 1
Aim of the review
he aim of this review is to establish the clinical
effectiveness and cost-effectiveness of
electroconvulsive therapy (ECT) for depressive
illness, schizophrenia, catatonia and mania.
The specific areas addressed by this review
are:
ECT has been available for use since the 1930s.
The therapy involves the passage of an electric
current through a person’s brain while they are
under a general anaesthetic and have been given a
muscle relaxant. This normally produces a
convulsion. A course of ECT usually consists of six
to 12 treatments given twice a week. ECT is
indicated for severely depressed patients, but also
has a role in the management of those with
schizophrenia, mania and catatonia, often when
drug therapy has proved ineffective or is not
suitable.
●
T
●
●
●
●
●
There is considerable variation in the use of ECT
within the UK and current opinion is divided
between those who consider ECT to be the most
effective treatment within psychiatry and
completely safe1 and those who consider that ECT
is probably ineffective and almost certainly causes
brain damage.2
●
●
the effectiveness of ECT for people with
depression, schizophrenia, mania and catatonia
the effectiveness of ECT in specific subgroups of
people, including older people, pregnant
women, and children and adolescents
the impact of ECT stimulus parameters
(including dosage, frequency of electricity,
number of treatments and electrode placement)
and technique of administration on the
effectiveness of ECT
the duration of the effects of ECT
the use of ECT as a maintenance therapy,
emergency therapy and the role of concomitant
therapy in the overall effectiveness of ECT
the setting in which ECT is administered and its
impact on the clinical effectiveness and costeffectiveness of ECT
the costs of additional infrastructure and
training required for the optimal delivery of
ECT
patient acceptability and choice in ECT and
how these may affect outcomes.
1
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 2
Background
Description of the underlying
health problem
●
●
Schizophrenia
Schizophrenia is a major psychotic disorder. It is
characterised by a constellation of symptoms and
signs that have been present for a significant
length of time during the past month with some
signs of the disorder persisting for at least
6 months.3 The symptoms and signs of
schizophrenia have been conceptualised as falling
into three categories: positive, negative and
disorganised. Positive symptoms include
hallucinations and delusions; negative symptoms
include loss of initiative, interest in others or sense
of enjoyment, blunted emotions and limited
speech; and disorganised symptoms include
disorganised speech and behaviour and poor
attention. The Diagnostic and Statistical Manual
of Mental Disorders (DSM-IV)4 describes four
subtypes of schizophrenia that are defined by the
predominant symptoms at the most recent
evaluation. These subtypes include paranoid type
characterised by delusions or auditory
hallucinations; disorganised type in which
disorganised speech, behaviour and blunted affect
predominate; catatonic type characterised by
immobility, excitability and mutism; and
undifferentiated type, which is a non-specific
category in which none of the other subtype signs
and symptoms are prominent.
Depression
●
According to DSM-IV,4 mild depression is defined
as five or six symptoms and only minor
impairment in occupational functioning or usual
social activities or relationships with others. Severe
depression is classified as either with or without
psychotic features; without psychotic features it is
defined as several symptoms in excess of those
required to make a diagnosis and marked
impairment in functioning; with psychotic features
also includes delusions or hallucinations.
Moderate depression is defined as symptoms or
functional impairment between ‘mild’ and ‘severe’.
Mania
Manic symptoms are considered to be part of
bipolar disorder. The DSM-IV4 minimum criterion
for bipolar affective disorder is a single episode of
mania or mixed disorder (both episodes of mania
and major depression occur). The DSM-IV4
criteria for mania are:
●
●
The DSM-IV4 criteria for a major depressive
syndrome are that at least five key symptoms
should be present during the same 2-week period
and one should be depressed mood or loss of
interest or pleasure. The key symptoms are:
●
●
●
●
●
●
depressed mood most of the day nearly every
day
markedly diminished interest or pleasure in all
or almost all activities most of the day, every day
significant weight loss or weight gain when not
dieting
insomnia or hypersomnia nearly every day
psychomotor agitation or retardation every day
(observable by others)
fatigue or loss of energy nearly every day
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
feelings of worthlessness or excessive
inappropriate guilt nearly every day
diminished ability to think or concentrate or
indecisiveness nearly every day
recurrent thoughts of death (not just fear of
dying), recurrent suicidal ideation or suicide
attempt or specific plan.
●
●
●
distinct period of elation, irritability or mood
disturbances lasting for at least 1 week (or for
any period of hospitalisation)
three of the following:
– inflated self-esteem
– decreased need for sleep
– increased talkativeness or pressure of speech
– flight of ideas or racing thoughts
– distractibility
– increase in goal-directed activity (e.g. social,
at work) or psychomotor agitation
– indiscreet behaviour with poor judgement
(sexual, financial)
symptoms that do not meet the criteria for a
mixed episode (fulfils criteria for both mania
and major depression)
marked impairment in occupational or social
function
not due to drug abuse (or other medication) or
a physical illness.
3
Background
According to DSM-IV,4 bipolar affective disorder
may be mild, moderate or severe, and severe
forms may be with or without psychotic features.
Bipolar disorder may also be associated with
catatonic features or have a postpartum onset.
DSM-IV4 also describes the long-term clinical
course of bipolar disorder, which may be with or
without full interepisode recovery, with a seasonal
pattern or with rapid cycling (four or more
affective episodes per year).
Catatonia
Catatonia is a condition that is associated with
both schizophrenia and affective disorders. It is
characterised by marked changes in muscle tone
or activity, which may alternate between extremes
of a deficit of movement (catatonic stupor) and
excessive movement (catatonic excitement). The
International Classification of Diseases (ICD-10)5
diagnostic criteria for catatonic schizophrenia state
that one or more of the following symptoms must
be present:
●
●
●
●
●
●
●
stupor (marked decrease in reactivity to the
environment and in spontaneous movements
and activity) or mutism
excitement (apparently purposeless motor
activity, not influenced by external stimuli
posturing (voluntary assumption and
maintenance of inappropriate or bizarre
postures)
negativism (an apparently motiveless resistance
to all instructions or attempts to be moved, or
movement in the opposite direction)
rigidity (maintenance of a rigid posture against
the efforts to be moved)
waxy flexibility (maintenance of limbs and body
in externally imposed positions)
other symptoms such as command automatism
(automatic compliance with instructions) and
preservation of words and phrases.
Although catatonia is most often thought to be
associated with schizophrenia, recent studies have
also found that it is associated with mania.6
Epidemiology
4
In 2000, the prevalence of depressive episode in
England, Wales and Scotland was 26 per 1000.7
Depression is more common in women than in
men. The age standardised prevalence of
depression treated in general practice in England
between 1994 and 1998 was 24.9 per 1000 in men
and 61.4 per 1000 in women.8 The agestandardised prevalence of treated depression in
Wales between 1994 and 1998 was 24.0 per 1000
in men and 57.4 per 1000 in women. These
figures may under-represent the true prevalence of
depression since it is estimated that on a typical
GP’s list, over 100 patients suffer from depression
but half go unrecognised.9
The lifetime prevalence of schizophrenia is 1%
and the incidence of first onset schizophrenia is
approximately 1 per 10,000 population per year.10
The age-standardised prevalence of schizophrenia
treated in general practice in England between
1994 and 1998 was 2.0 per 1000 in men and 1.7
per 1000 in women.8 In Wales, the age-standardised
prevalence of treated schizophrenia was 1.9 per
1000 in men and 1.3 per 1000 in women.
Standardised mortality rates in schizophrenia are
five times higher than those for the rest of the
population; 10–15% of people with the disorder
eventually commit suicide.10
Current service provision
Description of intervention
ECT has been available for use since the 1930s.
The practice of ECT has undergone a number of
modifications since its introduction, with the use of
general anaesthesia and muscle relaxants. The
current practice of ECT involves the passage of
electricity through a person’s brain while they are
under a general anaesthetic and have been given a
muscle relaxant. This normally produces a
convulsion. It was initially believed that the
production of a generalised seizure was both
necessary and sufficient for the antidepressant
effect of ECT as subconvulsive stimuli were without
therapeutic benefit. Later, it was demonstrated that
generalised seizures of adequate duration could be
reliably produced that lack therapeutic effect in
depression.11,12 Thus, the role of seizures in the
therapeutic efficacy of ECT is still open to debate
and there is currently no universally accepted
theory to explain the mechanism of action for ECT.
Current opinion on ECT ranges between those
who consider ECT to be the most effective
treatment within psychiatry and completely safe1
and those who consider that ECT is probably
ineffective and almost certainly causes brain
damage.2 ECT is a complex intervention and its
efficacy and safety are affected by a number of
parameters, including the placement of electrodes,
dosage and waveform of the electrical stimulus,
and the frequency at which ECT is administered.
Patient populations
Overall indications for ECT
Current guidelines from the American Psychiatric
Health Technology Assessment 2005; Vol. 9: No. 9
Association (APA)13 and the Royal College of
Psychiatrists (RCP)14 on the patient populations
for whom ECT is indicated are summarised below.
The APA13 guidelines recommend that ECT
should primarily be used where there is need for a
rapid response because of the severity of a
psychiatric condition, where the risks of other
treatments outweigh the risks of ECT, where there
is a history of poor medication response or a good
response to ECT, or where the patient requests it.
Secondary indications are in cases of treatment
resistance or adverse side-effects.
A survey of psychiatrists in the North West of
England indicated that 93% of respondents were
in favour of the use of ECT for appropriate
patient populations.15 The balance of opinion
favoured the use of ECT at some point in only
three conditions: depressive psychosis,
schizoaffective disorder and depression with
dementia.
The second phase of an audit of the use of ECT in
Scotland16 between 1997 and 1998 found that
85% of the people who received ECT suffered
from depressive illness, whereas only 7.8% were
diagnosed with schizophrenia, 2% a manic illness
and 1% a neurotic (anxiety) illness. These figures
were also similar during the third phase of the
audit that took place between 1998 and 1999
(87%, 6.3%, 3% and 1.5%, respectively). Among all
those who received ECT during 1997 to 1998 in
Scotland,16 the most common reason for receiving
ECT was resistance to antidepressant medications
(55%), followed by a previous good response to
ECT (39%), severe retardation (38%), being too
distressed to await response to medication (38%),
resistance to other drugs (27%) and suicidal
ideation (27%). In only 6% of cases was ECT used
as an emergency, life-saving treatment.
ECT in depressive illness
For depressive illness, first line treatment in the
acute phase is the use of antidepressant
medication.17 The APA guidelines indicate that
the effectiveness of antidepressant medications is
generally comparable,17 although a recent metaanalysis18 suggests that serotonin and
norepinephrine reuptake inhibitors (SNRIs, e.g.
venlafaxine) are more effective than selective
serotonin reuptake inhibitors (SSRIs) (e.g.
fluoxetine) or tricyclic antidepressants (TCAs) (e.g.
imipramine). A meta-analysis of 36 open and
double-blind trials suggested that 29–46% of
depressed patients failed to respond fully to
antidepressant treatment of adequate dose or
duration.19 The minimum dose of TCAs known to
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
be effective is 100 mg day–1 20 and treatment
resistance has been defined as failure to respond
to a trial of more than one antidepressant drug in
a dose equivalent to 250–300 mg of imipramine
given for a duration of 6–8 weeks each.21 The
APA22 advises that ECT should be considered only
for patients with major depression with a high
degree of symptom severity, for cases in which
psychotic symptoms or catatonia are present, or
for cases in which there is an urgent need for
response, such as patients who are suicidal or
refusing food. The RCP14 suggests that ECT may
be particularly effective in depressive illness with
psychotic features or in patients who have not
been responsive to antidepressant drug treatment.
However, studies have shown that response rates
following ECT for depressive illness are lower
(50%) in people who previously received
adequate antidepressant medication than in those
people who received inadequate treatment
(86%).23
A survey of psychiatrists in the North West of
England15 found that the most common indication
for the use of ECT in depressive illness was in
cases of refusal to eat or drink (89% agreed that it
was the treatment of choice), followed by patients
who were responsive in the past to ECT but not to
drugs (85%) or had a high suicidal risk (67%).
ECT was considered the treatment of choice for
psychotic depression by 61% of respondents, for
depression not responsive to antidepressant
medication by 53% and for depression with severe
agitation by 52%.
Repetitive transcranial magnetic stimulation
(rTMS) was developed in the 1980s and has been
reported to have an antidepressant effect, but data
on efficacy and optimal stimulation parameters
are still conflicting.24 The technique involves the
induction of a current in the brain using a
magnetic field. The stimulus is a magnetic field
that is generated by a current passing through a
coil of copper wire that is encased in plastic and
held over the patient’s head. rTMS involves the
administration of trains of stimuli to the same area
of the brain several times per second. The number
of stimuli per second, the strength of stimulus, the
duration of the train of stimulation, the interval
between trains, the total number of trains and the
total number of stimuli in a given session are
stimulus parameters that can be varied. The
adverse effects associated with rTMS are its
potential to induce a seizure, muscle tensions,
headaches, ringing in the ears and memory
problems. It is not currently used in clinical
practice.
5
Background
ECT in schizophrenia
For schizophrenia, first line treatment is with
antipsychotic medication.3 There are two main
types of antipsychotic medication. Typical
antipsychotics include chlorpromazine and
haloperidol, which have both shown to be more
effective than placebo in the treatment of
schizophrenia,25,26 but can produce a range of
unwanted side-effects including sedation, dry
mouth, tachycardia and extrapyramidal symptoms
(medication-induced parkinsonism). Atypical
antipsychotics such as clozapine have been shown
to be more effective than typical antipsychotics27
and have fewer extrapyramidal side-effects, but
cause potentially fatal agranulocytosis in about 1%
of patients.3 Adequate doses of typical
antipsychotic medication are considered to be the
equivalent of 300–600 mg of chlorpormazine a
day.3 The APA3 recommends that ECT could be
used when patients are treatment resistant or in a
catatonic state and when the psychotic symptoms
in the current episode have an abrupt or a recent
onset.13 Similarly, the RCP14 advises that the
practical usefulness of ECT in schizophrenia is
limited to acute catatonic states, schizoaffective
disorders, acute paranoid syndromes and people
with type I schizophrenia who are either
intolerant or unresponsive to a dose of a
neuroleptic equivalent to 500 mg of
chlorpromazine daily.
ECT in mania
In mania, lithium and divalproex are first line
treatments.28 The RCP14 recommends that ECT
may, in occasional circumstances, be used for people
with severe mania, or in less disturbed people with
mania who have a slow or inadequate response to
medication, and may be a safe alternative to highdose neuroleptics. The APA guidelines28 reserve
ECT as a sixth line treatment for euphoric or mixed
mania if residual symptoms are still severe following
treatment trials with lithium, divalproex with the
addition of benzodiazepines, atypical antipsychotics
or carbemazepine,28 as a fifth line treatment for
psychotic mania and almost the last resort for rapid
cycling mania. Some clinicians believe that ECT
needs to be administered more frequently to people
with mania to achieve a therapeutic effect (Birkett P,
Clinical Lecturer in Psychiatry, University of
Sheffield: personal communication, 2002).
Although there is no clear agreement on this, the
RCP guidelines recommend that this should be
considered.14
6
ECT in catatonia
First line treatment of catatonia is usually with
benzodiazepines (e.g. Lorazepam)29 and the APA28
and RCP14 guidelines recommend that catatonia is
an indication for the use of ECT in people with
schizophrenia or mania.
ECT in other subgroups
Other subgroups for which ECT is indicated as a
treatment option include older people with
depression, psychiatric illness associated with
pregnancy and the puerperium, and children and
adolescents with psychiatric problems, although it
is rarely used in the latter population.14
Stimulus parameters and
administration of ECT
Frequency and schedules
Although schedules of treatment vary, ECT is
commonly administered twice weekly in the UK,15
but three times a week in the USA.14 The courses
range from four to 12 treatments.30 Less
commonly, it is given fortnightly or monthly as
continuation ECT or maintenance ECT, to prevent
relapse of symptoms.
Electrode placement
ECT can be administered by placing electrodes on
both sides of the head (bilateral placement) or on
one side of the head (unilateral), either on the
dominant side of the brain or on the nondominant side. Unilateral ECT was introduced to
reduce the cognitive side-effects associated with
ECT, but also has a lower antidepressant effect.31
The RCP14 recommends that unilateral ECT
should be used where the speed of response is less
important or where minimising cognitive sideeffects is especially important, or where there has
been a good previous response to ECT. They
advise that bilateral ECT should be used where
speed and completeness of response have priority,
where unilateral ECT has failed, where previous
use of bilateral ECT has produced a good
response with no memory impairment or where
determining cerebral dominance is difficult. A
recent survey of psychiatrists in the North West
of England found that 57% usually used
bilateral ECT, 22% used unilateral and 16% used
either.15
Stimulus
Early ECT machines delivered an alternating sinewave stimulus at mains frequency and constant
voltage. Modern machines deliver a constant
current, variable frequency, brief pulse stimulus.
Both efficacy and cognitive side-effects are related
to the amount of electricity passed through the
brain. Modern machines use less electrical energy,
with the aim of maintaining therapeutic efficacy
and reducing cognitive side-effects.
Health Technology Assessment 2005; Vol. 9: No. 9
Seizure threshold
This refers to the minimum electrical stimulus
required to elicit a generalised seizure. It has been
shown to vary 40-fold between individuals, and to
increase over the course of ECT.12 Factors that raise
seizure threshold, and make it more difficult to
elicit seizures, include the use of benzodiazepine
anxiolytics and hypnotic drugs, anticonvulsant
medication, anaesthetic drugs, older age, male
gender, dehydration, low oxygen saturation of the
blood, and electrical parameters that raise
impedance such as poor contact between
electrodes and the scalp. The APA13 recommends
that ECT doses should be tailored to the
individual. The individual’s seizure threshold
should be determined using empirical titration and
ECT should be delivered at a moderately
suprathreshold dose, optimally at 50% above
seizure threshold.32
Seizure duration
In clinical practice, generalised motor seizures
less than 15 seconds long are considered
inadequate. Seizures of 25–30 seconds in duration
are aimed for, and monitored either by
electroencephalography or by observing and
timing motor convulsions in extremities or in a
forearm isolated from muscle relaxants by an
inflated blood-pressure cuff.13
Equipment and staffing
Both the RCP14 and the APA13 recommend that
the minimum requirement for ECT facilities is
three rooms: a quiet, comfortable waiting area, a
treatment room, and a recovery area of sufficient
size to accommodate the rate and number of
patients treated per session (possibly up to six
patients lying on trolleys). They advise that
rooms should contain the necessary equipment
to monitor patients and treat them in an
emergency. The staffing levels advised are two
trained nurses, plus four untrained nurses, an
anaesthetist, a psychiatrist and an operating
department assistant.32 The machines currently
recommended for use by both the APA13
and the RCP14 are Mecta SR2 and JR2,
Thymatron-DGx and Ectron series 5A Ectonus
machines.
Information and consent
The RCP guidelines14 highlight that under
common law in England, valid consent is required
from all patients, whether informal or detained
under the Mental Health Act, before ECT may be
given, except where statute specifically overrides
it. This consent must be given freely and be based
on an understanding:
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
●
●
●
●
●
●
of the purpose and nature of the treatment
of the likely risks and effects of treatment,
including its likely success
of the alternatives to the treatment
of the likely consequences of not receiving it
that consent can be withdrawn at any time
that new consent is required for further
treatment.
Where a patient does give consent, the RCP14
advises that this should be for a specific number of
treatments and be in the form of a written
document that is also signed by the doctor. Where
an informal patient refuses to give consent,
alternatives must be discussed, but if there are
strong grounds for the use of ECT the RCP14
recommends considering whether the person
should be detained. In the case of detained
patients refusing treatment, the commission must
be asked to issue a certificate in the prescribed
form to allow treatment to go ahead. Where a
patient is incapable of giving consent, the RCP
advises that guidance from the relevant Mental
Health Act should be followed. Under common
law, ECT may be given if the treatment is ‘in the
patient’s best interest’ after a second opinion has
been obtained.
In a recent survey of the use of ECT in England
between January and March 1999,33 75% of
people receiving ECT in the survey were not
formally detained under the Mental Health Act.
All of these informal patients consented to
treatment, with 1.4% being treated as an
emergency. Of the 709 people who were formally
detained, 29% consented to ECT, 12% were
treated as an emergency and 59% did not consent
to treatment but were treated after a second
opinion was gained.
Current service provision in England
and Wales
A recent survey of ECT use in England33 reported
that between January and March 1999 there were
16,482 administrations of ECT to 2835 patients.
Eighty five per cent of all administrations were in
an inpatient setting. The average number of
administrations per patient was 5.6, ranging from
4.8 in the Trent region to 6.6 in London.
The survey33 revealed important variations in the
rates of administration of ECT by gender, age and
health region. In the population as a whole, 5.8
people per 100,000 underwent ECT. The rate was
significantly higher in females (7.7 per 100,000
females) than for males (3.8 per 100,000). For
both genders, the rate increased with age, with
7
Background
15.1 per 100,000 population aged 65 and over
undergoing ECT. The highest rate of ECT use was
in the North West (7.1 per 100,000 population)
and the lowest was in London (3.7 per 100,000
population). The survey did not provide any
information regarding the diagnoses of those who
received ECT.
A survey of the use of ECT in Wales during 199634
found similar increases in the rate of ECT
administration with age. The age-specific rates of
administration of ECT to people aged 20–34,
34–64 and 65 and over were 7.7, 13.2 and 25.5
per 100,000 population, respectively.
A survey of the use of ECT in young people
during 199635 found that the rate of
administration to people under 18 was 0.02 per
100,000 total population per year. The agespecific rate of administration of ECT to people
aged 16 or 17 (0.62 per 100,000 age-specific
population per year) was over six times greater
than for those aged between 12 and 15 years (0.10
per 100,000 age-specific population).
An important question is whether these variations
in the use of ECT are the result in variations in
the need for ECT (e.g. as a result of variations in
the prevalence of depression) or the result of
differences in preferences for the use of ECT on
behalf of psychiatrists. Although observations of
variations in the prevalence of the underlying
disorder do not imply a causal relationship
between variations in the prevalence of a
condition and a treatment, they provide some
insight into this issue. With regard to variations by
region, between 1994 and 1998 the pattern in the
prevalence of treated depression in men and
women was similar to the use of ECT. The
prevalence of treated depression in men and
women was highest in the North West (30.4 per
1000 and 70.3 per 1000, respectively) and lowest
in North Thames (18.8 per 1000 and 46.5 per
1000, respectively) and South Thames (20.6 per
1000 and 49.7 per 1000, respectively). As
discussed earlier, the prevalence of depression is
also higher in women than in men.
8
Without statistical testing it is not possible to draw
definitive conclusions regarding trends in the
prevalence of treated depression with age in
England in men and women. In men, the
prevalence of depression in England increases with
age until 55–64 years, then drops between 65 and
74, then increases again between 75–84 and 85plus years of age. In women, the prevalence of
depression in England increases with age until
45–54 years, drops between 55–64 and 65–74 to
comparable levels with people aged 35–44,
increases again at 75–84 years and drops at
85-plus years to comparable levels with people
aged 35–44 years.
Since 1985, the use of ECT in England has been
decreasing.8 The estimated 65,930 administrations
in 1998/99 compares with 105,466 reported
administrations in 1990/91 and 137,940 in
1985.8
Training and the quality of ECT services
The RCP first issued guidance on the
administration of ECT in 1977.36 In 1981,
Pippard and Ellam37 conducted an audit against
those standards and visited about half of the ECT
clinics in the UK (180 clinics). They found that the
quality of the centres overall was low, with some
centres using obsolete machines, and the
training provision for junior doctors was generally
poor. In response to these findings, the RCP
issued revised guidance on the administration of
ECT in the form of its first ECT handbook in
1989. In 1992, Pippard38 conducted a second
audit of ECT practice in the UK against the 1989
standards, visiting 35 NHS and five private ECT
clinics in the old North East Thames and East
Anglia regions. Although improvements had been
made since 1981 in the standard of ECT facilities
and some aspects of practice, a significant
number of clinics were still failing to meet the
1989 standards. Again, the training of junior
doctors in the practice of ECT and the use of
modern ECT machines were areas in which a large
number of clinics did not meet the 1989
standards.
As a result of Pippard’s findings, the RCP
established a working group on ECT to revise and
broaden the guidelines to include both the
structures and process of ECT practice. The
guidelines were disseminated through the
publication of a revised edition of the handbook in
1995,32 along with a training video and a series of
training courses run by the RCP. A third audit
against these guidelines conducted by Duffett and
Lelliot34 took place between 1995 and 1996. They
visited all 33 NHS clinics and five private clinics in
the North East Thames and East Anglia regions,
and 17 NHS clinics in Wales. They also conducted
a postal survey of the 165 ECT clinics in England
that were not visited. Two-thirds of those who
responded were at Senior House Officer (SHO)
level. Around the same time Hillam and
colleagues39 conducted a postal survey of the
experiences of psychiatry trainees at the Royal
Health Technology Assessment 2005; Vol. 9: No. 9
Free Hospital in 1990 (n = 51) and in 1995
(n = 34).
Duffett and Lelliot34 found that despite
improvements in some aspects of care, only onethird of the clinics rated met the college
guidelines. Fifty-nine per cent of all clinics had
ECT machines of the type recommended by the
college, but 7% were still using machines
considered to be outdated in 1989. Only 16% of
consultants attended their ECT clinic weekly and
only 6% had sessional time for ECT practice.
Duffett and Lelliot34 report that the training of
junior doctors was still of a low quality. Only onethird of clinics had clear policies to guide junior
doctors to administer ECT effectively. In a survey
of junior doctors, Duffett and Lelliot40 found that
only half of respondents had been supervised by
an experienced psychiatrist on their first
administration of ECT; a similar finding was also
reported by Hillam and colleagues.39 Duffett and
Lelliot40 found that 45% of respondents lacked
knowledge about one or more basic issues relating
to the administration of ECT. Hillam and
colleagues39 report that 86% of their sample felt
confident in their administration of ECT, but onefifth admitted to distress or unease when
administering ECT.
Although improvements have been made in the
practice of ECT during the 20 years since the RCP
first issued guidance, there are still many areas of
ECT practice that would benefit from further
improvement. In particular, the training of junior
doctors in the administration of ECT is still an
area of concern.
Current mental health policy in England
and Wales
As a recent survey of ECT use in England8 has
shown, the majority (85%) of all administrations of
ECT were within an inpatient setting. In contrast,
much of recent government policy on the care and
treatment of people with mental health problems
has focused on providing more care in community
settings. The National Service Framework (NSF)
for Mental Health41 advises that people with
short-term severe mental health problems,
including severe depression, can be managed in
primary care through treatment with drugs and
psychological therapies. The NSF41 recommends
that people with recurrent or severe and enduring
mental illness, including schizophrenia and
bipolar affective disorders, who have complex
needs requiring continuing care of specialist
mental health services working with other
agencies, can also manage well with this support
while living in the community.
9
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 3
Effectiveness
Methods for reviewing
effectiveness
Search strategy: clinical effectiveness
The search aimed to identify all references
relating to the clinical effectiveness and costeffectiveness of ECT for depression,
schizophrenia, catatonia and mania.
Sources searched
Seventeen electronic bibliographic databases were
searched, covering biomedical, health-related,
science, social science and grey literature. A list of
databases is provided in Appendix 1. This
includes the Cochrane Schizophrenia Group Trials
Register, which was searched on behalf of the
review team by the Group’s Trials Search
Co-ordinator.
In addition, the reference lists of relevant articles
were checked and 40 health services researchrelated resources were consulted via the Internet.
These included health technology assessment
organisations, guideline-producing bodies, generic
research and trials registers, and specialist
psychiatric sites. A list of these additional sources
is given in Appendix 2. Finally, citation searches of
key papers were undertaken using the Science
Citation Index (SCI) citation facility and the
reference lists of included studies were checked for
additional studies.
Search terms
A combination of free-text and thesaurus terms
was used. ‘Population’ terms (e.g. depression,
schizophrenia, catatonia, bipolar disorder, mania,
mood disorders, adjustment disorders, psychotic
disorders, mental disorders) were combined with
‘intervention’ terms (e.g. electroconvulsive therapy,
electro convulsive therapy, electroshock therapy,
electro shock therapy). Copies of the search
strategies used in the major databases are
included in Appendix 3.
Search restrictions
No date or language restrictions were applied.
Where necessary (e.g. in the larger databases, such
as MEDLINE), searches were restricted to the
highest quality of evidence, namely, practice
guidelines, systematic reviews and randomised
controlled trials (RCTs), using methodological
filters (Appendix 4). These were supplemented by
strategies designed to pick up other outcomes,
such as patient acceptability, side-effects and staff
training (Appendix 4).
Search strategy: cost-effectiveness
In addition to the searches conducted above,
searches were conducted in the NHS Economic
Evaluation (NHS EED) and Office of Health
Economics Health Economics Evaluations
Database (OHE HEED) to identify specifically
cost-effectiveness literature (Appendix 3).
Methodological search filters designed to retrieve
economic evaluations and quality of life studies
(Appendix 4) were also applied to the MEDLINE
and EMBASE search strategies.
There were no company submissions.
Inclusion and exclusion criteria
Populations
Papers were included in the review if they studied
the following populations: depressive illness (both
unipolar and bipolar), schizophrenia and
schizoaffective disorder, catatonia and mania. A
further aim was to explore the clinical
effectiveness of ECT in particular subgroups
including people who are resistant to
pharmacotherapy, older people (defined as aged
65 years and over), younger people (defined as
aged 18 years or under), and women with
disorders associated with pregnancy and the
puerperium. Papers were excluded if they
included populations with more than one
diagnosis (e.g. depression and schizophrenia) and
did not stratify randomisation by disease type or
report results separately for each diagnosis.
Interventions
Papers were included in the review if they
examined the effectiveness or cost-effectiveness of
ECT either as a monotherapy or in conjunction
with other appropriate pharmacological or
psychological treatment, at all doses and frequency
of administration, by any technique, in all settings
11
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Effectiveness
and administered by any health professional. The
review also included studies investigating the
efficacy of adjunctive and continuation or
maintenance ECT or pharmacotherapy and
interventions that aimed to improve patient
knowledge about ECT.
Comparators
Papers were included if they compared ECT with
any pharmacological or non-pharmacological
treatment including sham ECT, psychotherapy or
rTMS. Studies that compared one or more type of
pharmacotherapy post-ECT were also included.
Outcomes
Studies were included if they assessed outcomes
relating to the efficacy, safety and acceptability of
ECT. The primary indicators of the efficacy of
ECT were clinically meaningful benefits in
symptoms and/or quality of life as measured by a
validated rating scale or clinical opinion.
Secondary indicators were the speed of response
to ECT, premature withdrawals by the decision of
either the participant, the clinician in charge of
their care or the researcher, discharges from
hospital and relapses. The primary indicators of
the safety of ECT were adverse events including
both objective and subjective reports of memory
loss (anterograde, retrograde and subjective
reports of memory loss) and all-cause and causespecific mortality (including suicide). All these
outcomes were considered immediately after the
course of ECT, at 6 months and 12 months or
longer. The primary indicators of acceptability
were patients’ choice of treatment and their views
and experiences of ECT from either
questionnaires or interviews.
Study methodology
Published papers were included in the review
according to the accepted hierarchy of evidence.
In the first instance papers were only included if
they were systematic reviews, RCTs or economic
evaluations. Where no RCT evidence was
available, non-randomised comparator studies
(e.g. non-randomised trials, controlled cohort
studies and case–control studies) were included in
the review. Where no evidence from nonrandomised comparator studies was available,
non-randomised, non-comparator studies (e.g.
case series, case reports, non-controlled cohort
studies) were included in the review.
12
Language
Any studies not available in English were excluded
as the timescale of the review precluded time for
translation.
Quality assessment and data
extraction strategy
Quality assessment and selection of studies
All the abstracts identified by the searches were
entered into a reference manager database and
reviewed by the relevant author to assess their
relevance to the review’s objectives in terms of the
clinical effectiveness and cost-effectiveness of ECT.
All potentially relevant papers were ordered and
assessed by the relevant author to determine
whether they met the study’s inclusion criteria in
terms of the populations, interventions, outcomes
and study quality.
The assessment of study quality was not conducted
blindly and used the following guidelines.
●
●
●
●
●
Systematic reviews were assessed according to
the users’ guides to evidence-based practice.42
RCTs were assessed with respect to
randomisation procedures, blinding, handling
of withdrawals and dropouts, guided by Jadad’s
scoring system43 and the Cochrane
Collaboration Handbook.44
Non-randomised studies using quantitative
data, such as case–control, cohort, case series
and case reports, were assessed with respect to
validity using guidelines from the Centre for
Health Evidence based on the users’ guides to
evidence-based medicine.45
Qualitative evidence was assessed using the
standards proposed by Popay and colleagues.46
The quality of the economic literature was
assessed according to the guidelines for authors
and peer reviewers of economic submissions to
the British Medical Journal.47
Data extraction and analysis
Two reviewers (JG and DH) extracted data on
clinical effectiveness using three separate, standard
abstraction forms for systematic reviews (JG), RCTs
(DH and JG) and non-randomised evidence (JG).
This procedure was not conducted blind to the
authorship of the study.
Where the reviewers were satisfied that the
populations, interventions and outcomes between
trials were sufficiently similar, results were pooled
in a meta-analysis.
Clinically meaningful improvement in symptoms
was abstracted using both binary and continuous
data. For dichotomous data the number of
responders or relapsers in each treatment arm, as
defined by the trialists, was compared. Other
binary outcomes were the numbers of
discontinuations, relapses and deaths. Those
Health Technology Assessment 2005; Vol. 9: No. 9
leaving the trial early were assigned to the worse
outcome and this was tested using a sensitivity
analysis. If the definition of responders or relapsers
used by the trialists was not clear, a clinically
meaningful cut-off was decided by an independent
clinician who was blind to the trial authors, the
intervention, numbers achieving each outcome in
each arm and number in each arm. Where trials
used different methods to define responders [e.g.
clinical opinion versus scores on the Hamilton
Rating Scale for Depression (HRSD)], this was
tested using sensitivity analysis. The data were
deemed unusable if the numbers of people
meeting responder or relapse criteria were not
specified separately in each group, or dropouts
were not accounted for on a treatment group
basis. Relative risks and confidence intervals were
calculated using the random effects method of
DerSimonian and Laird.48 All analyses were by ITT.
For continuous data group means and standard
deviations at baseline, immediately after ECT and
at 6 months’ follow-up were recorded. The data
were deemed unusable if:
●
●
●
●
no standard deviations or standard errors
and/or means were reported
the instrument used had not been published in
a peer-reviewed journal, as non-validated
outcome measures are a serious threat to the
validity of meta-analyses.49
baseline and follow-up data were based on
different samples (e.g. baseline data included all
participants but follow-up data only included
the completer sample)
at least 50% of the sample were lost to followup.
For studies reporting continuous outcome data all
measured using the same scale or instrument (e.g.
HRSD) the summary statistic used was the
weighted mean difference (WMD). Again, a
random effects model with the DerSimonian and
Laird method48 was used.
For studies reporting continuous outcome data
when different scales or instruments were used to
measure the effect (e.g. HRSD, Beck Depression
Inventory (BDI)] the summary statistic used was
the standardised mean difference (SMD). It was
assumed that these instruments were all measuring
the same underlying trait of ‘depression’. Again, a
random effects model with the DerSimonian and
Laird method48 was used.
All analyses were carried out in RevMan v4.0
(http://www.cochrane.de/cochrane/revman.htm).
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Heterogeneity was examined both graphically and
with a formal statistical test of heterogeneity. If the
confidence intervals for the results of each study
(typically represented by horizontal lines) do not
overlap, it suggests that the differences are likely
to be statistically significant. A formal statistical
test of homogeneity was also used to examine
whether the observed variation in study results is
compatible with the variation expected by chance
alone. The more significant the results of the test
(the smaller the p-value), the more likely it is that
the observed differences were not due to chance
alone.
Results
Quantity of research available
The searches generated 1647 references. Before
identification of the two systematic reviews (see
below), 790 references were included at the title
stage and 485 were included at the abstract stage
and ordered for review. The studies included in
the study are described below.
Two high-quality, recently completed systematic
reviews of the safety and efficacy of ECT were
identified through contacts with experts in the
field. One was completed by Tharyan and
Adams50 in 2002 on behalf of the Cochrane
Schizophrenia Group, and reviews the efficacy
and safety of ECT in schizophrenia. The authors
were contacted and gave their permission for
the review to be used in this report before its
official publication. The references of the review
were checked and no additional studies were
identified.
The second review was commissioned by the
Department of Health and reviews the safety and
efficacy of ECT in depression, schizophrenia and
mania. This review was conducted by the UK ECT
Group and permission was given to use the report
prior to publication in 2003.51 The majority of the
text from this report has been reproduced in this
review. The references of the report were checked
and one additional study was identified.52
This report is largely based on the results of these
two reviews and this has been acknowledged in the
text of the report.
A further high-quality, recently completed
systematic review of non-randomised evidence of
consumer’s views of ECT was also identified
through contact with experts in the field. This
report was also commissioned by the Department
13
Effectiveness
of Health and was conducted by Service User
Research Enterprise (SURE) at the Institute of
Psychiatry.53 The authors were contacted and gave
their permission to use the review prior to its
publication.
For interventions, neither the UK ECT Group
review nor the Cochrane Schizophrenia Group
ECT review included studies comparing ECT with
rTMS, nor did they include studies evaluating the
effectiveness of post-ECT drug therapy.
The populations, interventions and outcomes of
included studies in these three reviews were
compared with the scope of the National Institute
for Clinical Excellence (NICE) review to assess the
degree of overlap and identify areas not covered
(Table 1). There were several gaps in the coverage
between the scope of the UK ECT Group and the
Cochrane Schizophrenia Group ECT review and
the scope of the NICE review. Additional
randomised and non-randomised evidence was
identified to address these gaps.
In terms of populations, neither the UK ECT
Group review nor the Cochrane Schizophrenia
Group ECT review identified any RCTs evaluating
the efficacy of ECT specifically in older people,
people with catatonia, younger people or children,
or women during or after pregnancy. Some of the
trials did include people with catatonia and older
people and younger people, but results were not
reported separately and in the UK ECT Group
report data were too limited to perform reliable
subgroup analyses. The Cochrane Schizophrenia
TABLE 1 Overlap between the NICE scope and the six systematic reviews identified
Review
14
UK ECT
Group,
200251
Cochrane ECT
reviewers,
200250
SURE,
200253
Ray and
Walters, 199770
Walters et al.,
199971
Hawkins
et al.,
199578
Miller,
199481
Type of evidence
Randomised evidence
Non-randomised evidence
Y
Y
Y
N
N
Y
N
Y
N
Y
N
Y
Conditions
Depression
Schizophrenia
Mania
Catatonia
Y
Y
Y
N
N
Y
N
Y
?
?
?
?
Y
Y
Y
Y
N
N
N
Y
?
?
?
?
Specific subgroups
Younger people
Older people
Pregnant women
N
N
N
N
N
N
?
?
?
Y
N
N
N
N
N
N
N
Y
Comparators
Sham vs real
ECT vs pharmacotherapy
ECT vs psychotherapy
Stimulus parameters
ECT vs rTMS
Adjunctive drug therapy
Continuation ECT
Continuation pharmacotherapy
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
N
Y
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Specific outcomes
Symptom improvement
Perceived benefit
Cognitive functioning
Suicide
All-cause mortality
Brain damage
Other adverse events
Information and consent
Y
N
Y
Y
Y
Y
N
N
Y
N
Y
Y
Y
N
N
N
N
Y
Y
N
N
N
N
Y
Y
N
Y
N
Y
Y
Y
N
Y
N
N
N
N
N
N
N
N
N
N
N
Y
N
Y
N
Y, topic covered by the review; N, topic not covered; ?, insufficient detail to determine whether the topic was covered.
Health Technology Assessment 2005; Vol. 9: No. 9
Group ECT review conducted a subgroup analysis
for schizophrenia subtype, including one trial that
predominantly (although not exclusively) included
people with catatonia.
In terms of outcomes, the UK ECT Group review
and Cochrane Schizophrenia Group ECT review
did not identify any trials that explored either
quality of life or the impact of consumer choice on
the outcomes of ECT. Non-randomised studies
evaluating this topic were included in the SURE
review.
In populations with depressive illness, the
reviewers identified two RCTs comparing ECT
with rTMS54,55 and ten RCTs comparing ECT
combined with drug treatment versus ECT
combined with either placebo or a different
drug.56–65 In four of these trials,61–64 participants
continued taking pharmacotherapy following the
course of ECT and its impact on relapses was
assessed. The search also found three trials66–68
that compared different approaches to
antidepressant treatment following successful
treatment with ECT.
An additional RCT was identified that evaluated
the impact of an educational video on patient
knowledge about ECT;69 this was not included in
the SURE review.53
Owing to the lack of randomised evidence, nonrandomised evidence was examined for the
efficacy of ECT in older people, younger people,
people with catatonia and ECT during or
following pregnancy. For children and adolescents,
two systematic reviews70,71 of case series were
identified; the review published in 1999 was an
update of a previous review published in 1997 by
the same authors. One cohort study72 published
since this review was also identified. For older
people, one prospective cohort73,74 study
comparing older people who had received ECT
with those who had not and three retrospective
cohort studies.75–77 were identified. For catatonia
one systematic review78 of case reports and case
series of people with catatonia who received
ECT, published in 1995, and two prospective79,80
case series published since this date were
identified. For the use of ECT during pregnancy,
one systematic review of case series81 and case
reports published in 1994 and three case
reports82–84 published since that date were
identified.
Table 1 outlines the overlap between the NICE
scope and the six systematic reviews identified.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Table 2 provides an overview of the NICE scope
and indicates the sources of evidence used for
specific areas.
Tables of all included reviews or studies are shown
in Appendix 5. Figures of analysis are shown in
Appendix 6.
Quality of studies identified
Randomised evidence
Two systematic reviews including randomised
evidence examining the efficacy and safety of ECT
were identified.50,51 The discussion here reviews
the quality of these systematic reviews and then
describes the quality of the trials included as
reported by the authors of the reviews.
UK ECT Group review
The UK ECT Group review51 covers the efficacy of
ECT in people with depression, schizophrenia and
mania. Little information was provided in the
review regarding the characteristics of participants
in terms of the nature and severity of their
condition, medication history and previous use of
ECT. Information was provided regarding the
inclusion and exclusion criteria of the studies,
revealing considerable variation between them.
A wide range of interventions was included in the
review comparing the effectiveness of ECT with
sham ECT, inpatient care alone and
pharmacotherapy. The UK ECT Group review51
also examined the stimulus parameters of ECT
including electrode placement (bilateral versus
unilateral, Lancaster275 versus d’Elia86 placement,
frontotemporal versus temporoparietal),
dosage (high versus low), waveform (sine wave
versus brief pulse), frequency of ECT
administration (twice weekly versus three times
weekly), number of ECT treatments (number
considered medically sufficient versus medically
sufficient and two extra), number of seizures
induced per treatment (one versus two) and postECT nursing care.
The outcomes considered in the review were
improvements in symptoms following a course of
ECT and at 6 months’ follow-up, leaving the study
early, all-cause and case-specific mortality,
cognitive functioning (anterograde, retrograde,
orientation, subjective reports and overall
functioning) immediately after treatment, at the
end of an ECT course and at 6 months’ follow-up,
functional impairment and brain damage. The
reviewers also included studies examining quality
of life, but did not locate any trials using this
outcome.
15
Effectiveness
TABLE 2 NICE scope and sources of evidence used
NICE scope
Depression
Real vs sham ECT
ECT vs inpatient care
ECT vs pharmacotherapy
Unilateral vs bilateral
Unilateral: dominant vs non-dominant
Bilateral: frontotemporal vs temporoparietal
Frequency of administration
Dosage: high vs low
Waveform: sine wave vs brief pulse
Ultrabrief vs standard
No. of ECT sessions
Post-ECT nursing care
ECT vs rTMS
ECT + pharmacotherapy vs
ECT+ placebo/different pharmacotherapy
Continuation pharmacotherapy
Mania
ECT vs pharmacotherapy
ECT + pharmacotherapy vs pharmacotherapy alone
Schizophrenia
Real vs sham ECT
ECT vs pharmacotherapy
ECT + pharmacotherapy vs pharmacotherapy
ECT vs psychotherapy
Continuation ECT
Bilateral vs unilateral
Unilateral: dominant vs non-dominant
Dosage
Frequency of administration
No. of treatments
Specific outcomes
Severe adverse events
Patient acceptability and choice
Patient information
Specific subgroups
Catatonia
Children and adolescents
Older people
ECT during pregnancy
16
Source of evidence
UK ECT Group review of randomised evidence51 and NICE
reviewers’ reanalysis of trials identified by UK ECT Group
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51 and NICE
reviewers’ reanalysis of trials identified by UK ECT Group
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
NICE reviewers’ analysis of randomised evidence
NICE reviewers’ analysis of randomised evidence
NICE reviewers’ analysis of randomised evidence
UK ECT Group review of randomised evidence51
UK ECT Group review of randomised evidence51
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
Cochrane Schizophrenia Group ECT review of randomised
evidence50
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group51 and Cochrane Schizophrenia Group ECT review50
of randomised evidence
UK ECT Group review of non-randomised evidence51
SURE review of non-randomised evidence53
NICE reviewers’ analysis of randomised evidence and SURE review
of non-randomised evidence53
Cochrane Schizophrenia Group ECT review of randomised
evidence,50 Hawkins’ review of non-randomised evidence78 and
NICE reviewers’ analysis of non-randomised evidence
Rey and Walter’s reviews of non-randomised evidence70,71 and
NICE reviewers’ analysis of non-randomised evidence
NICE reviewers’ analysis of non-randomised evidence
Miller’s review of non-randomised evidence81 and NICE reviewers’
analysis of non-randomised evidence
Health Technology Assessment 2005; Vol. 9: No. 9
To assess the effectiveness of ECT, the reviewers
limited their inclusion criteria to RCTs. To provide
further information regarding the safety of ECT,
the reviewers included case–control and cohort
studies comparing participants who had received
ECT with those who had not. The inclusion criteria
were applied by two independent reviewers and a
list of excluded studies is provided. No information
is given regarding reasons for exclusion.
The search strategy used in the review was
comprehensive. The reviewers searched a large
number of electronic databases supplemented by
citation tracking of included articles and key texts
and contacting experts in the field and
manufacturers of ECT machines to identify
unpublished studies. Only published data were
included in the review; the reviewers did not
contact authors for unpublished data.
Data quality was not assessed blindly and the
reviewers did not quantify study quality with rating
scales, as they argued that the validity and
reliability of such scales are uncertain. Instead, the
study quality of RCTs was assessed according to
allocation concealment, blinding, loss to follow-up
and length of follow-up; cohort studies were
assessed on measurement bias, handling of
confounding factors, number of cases and loss to
follow-up; case–control studies were assessed on
measurement bias, handling of confounding
factors and number of cases. The reviewers
reported which paired reviewers extracted specific
sections of the data and any disagreements were
resolved by discussion.
The primary outcome of the review was
continuous data from depression rating scales such
as the HRSD.87 The reviewers did not consider
any dichotomous data of improvement in their
analysis, which may mean that important evidence
regarding the effectiveness of ECT is lost. It has
been acknowledged that health status measures
providing continuous data can be difficult to
interpret clinically85 and the clinical significance of
‘X points change’ on the HRSD has yet to be
clarified. Guidance on interpreting the HRSD
relates to its use as a discriminant instrument used
to divide people into groups based on the severity
of their depression. For example, McDowell and
Newell88 advise that a score of 7 indicates the
absence of depression, 7–17 mild depression,
18–24 moderate and 25 or above severe
depression. As a measure of change, the HRSD
has been criticised for its lack of responsiveness
owing to its multidimensional nature.89
Furthermore, many different versions of the
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
HRSD exist and such guidelines do not necessarily
translate to other versions of the scale.
It is evident from the table of studies in the UK
ECT Group report that not all trials used the
HRSD and no discussion was provided of the
extent to which unpublished symptom rating
scales have been psychometrically validated. They
did not examine whether these trials had different
results from those that did use validated scales,
and do not provide any raw data to give readers
sufficient information to decide for themselves. It
has been demonstrated that unpublished scales
are an important source of bias in systematic
reviews in psychiatry; studies using unvalidated
scales are more likely to find statistically significant
differences between treatments than validated
scales.49 Furthermore, with continuous summary
measures it is not possible to conduct an ITT
analysis, only one based on a completer sample.
The reviewers did not discuss how this may have
influenced their results.
Where appropriate, data from individual trials
were summarised by meta-analysis using a random
effects model. The reviewers calculated
standardised WMDs, which were summarised to
produce a standardised effect size according to the
method of Hedges and Olkin.90 The standardised
effect size is the difference in means divided by
the pooled study standard deviation. The pooled
study standard deviation is based on a weighted
mean of the intervention and control group
variances. The reviewers used this method to allow
information from different instruments measuring
the same construct (i.e. severity of symptoms) to
be summarised, and to take into account the
number of trial participants from each trial when
other usable data were not available. Dichotomous
and categorical data were combined to produce
estimates of odds ratios and absolute risk
differences. All estimates had confidence intervals.
The reviewers investigated heterogeneity between
studies, but did not describe how this was done.
Sensitivity analyses were conducted excluding
studies of inferior quality. Subgroup analyses were
identified a priori and included psychotic
depression, retarded depression, the effect of age,
treatment resistance, gender and severity at entry
to trial. However, no subgroup analyses were
conducted owing to limited information on these
subgroups. Publication bias was assessed using
funnel plots.
There were two important areas in the review in
which data may have been pooled inappropriately.
17
Effectiveness
In the comparison of ECT and pharmacotherapy
for depression, the reviewers pooled data from
trials comparing ECT with different classes of
antidepressants including TCAs, MAOIs and
SSRIs. Some trials also used L-tryptophan, which
in current clinical practice is not used as a first
line treatment and is used only rarely in
combination with other antidepressants. In the
comparison of real versus sham ECT for
depression, the reviewers pooled trials that used
bilateral and unilateral ECT. In a later section of
the report the reviewers provided evidence that
bilateral ECT is more effective than unilateral
ECT. No reference to this finding was made and
no justification for pooling the trials using
different electrode placements was given. The
reviewers did not provide any raw data to allow
the reader to investigate these issues.
To assess whether the conclusions drawn by the
UK ECT Group would be affected by different
methods of data analysis, further analysis of the
trials was undertaken in the following ways.
●
●
●
18
Trials comparing sham ECT with real ECT and
ECT versus pharmacotherapy were reabstracted
using dichotomous data.
Trials comparing real ECT with sham ECT were
reanalysed, with separate analyses for bilateral
ECT, unilateral ECT and trials that used both
methods.
Trials comparing ECT with pharmacotherapy
were reanalysed, with separate analyses by drug
class (e.g. SSRIs and TCAs).
electrode placement (bilateral versus unilateral),
dose (threshold versus suprathreshold), frequency
of ECT administration (three times weekly versus
five days a week) and the number of ECT
treatments (long courses versus short course).
The primary outcomes of interest were clinically
meaningful benefits in overall functioning,
hospitalisation status, changes in mental state,
behaviour, social and occupational functioning,
remission of symptoms and discharge from
hospital or care. Secondary outcomes were
premature withdrawal from the trial by the
decision of either the participant or the
researchers, and adverse events including
cognitive functioning and mortality. Each outcome
was reviewed during the ECT course, in the short
term (less than 6 weeks), medium term (6 weeks to
6 months) and long term (over 6 months).
The search strategy of the review was
comprehensive and a range of electronic databases
was searched using established search strategies
from the Cochrane Schizophrenia Group. These
searches were supplemented by citation tracking,
and the editorial board of the leading journal in
the field and first authors of all trials published
since 1980 were contacted for additional
references and unpublished trials. The
manufacturers of ECT machines were also
contacted for additional studies.
The reviewers limited their review to RCTs only.
Two reviewers independently assessed every report
identified by the electronic search for its relevance
to the review and disagreements were discussed.
Where disagreements remained unresolved, the
report was ordered and the study added to those
awaiting assessment while the authors of the study
were contacted for additional information.
Cochrane Schizophrenia Group ECT review
The Cochrane Schizophrenia Group ECT review
conducted by Tharyan and Adams50 included
people with schizophrenia, schizoaffective disorder
or chronic mental disorder (non-affective). They
identified a total of 24 studies including 1451
participants, of whom 779 were treated with ECT.
The reviewers provided a description of the
participants included in the trials in terms of
diagnoses, age, gender, whether participants were
treatment resistant and the duration of the
disorder. They also described the study setting and
length of the trials.
Study quality was assessed using guidelines in the
Cochrane Collaboration Handbook.44 Two
reviewers independently assessed the trials and
only those where the method of randomisation
was classed as concealed (A) or unclear (B) were
included. In cases of disagreement, further
clarification was sought from the author.
The review examined the effectiveness of ECT in
comparison with placebo, sham ECT,
pharmacological interventions and nonpharmacological interventions (e.g.
psychotherapy). They also assessed the
effectiveness of continuation ECT compared with
continuation pharmacotherapy. The review also
examined ECT stimulus parameters including
The Cochrane Schizophrenia Group50 used
dichotomous data of global improvement as
defined by the trialists as their primary outcome
measure of efficacy. They argued that clinicians
can better make sense of data indicating whether
someone has improved or not. Relative risks and
confidence intervals were calculated for each
outcome. They also calculated the number needed
Health Technology Assessment 2005; Vol. 9: No. 9
to treat (NNT) and number needed to harm
(NNH). All analyses were undertaken on an ITT
basis and participants who left the study early were
assigned to the least favourable outcome. The
effects of this assignment were tested in a
sensitivity analysis. For the outcome of global
improvements in functioning, the reviewers
compared the numbers who did not improve in
each arm of the trial. No information was
provided regarding how ‘no improvement’ was
defined within the various trials. Trials91 of
pharmacotherapy for depression often use the
criterion of a 50% reduction in HRSD to define
responders. Fink1 points out that trials of ECT
often use a different criterion to distinguish
responders from non-responders. There are two
important disadvantages to using dichotomous
data. First, it is difficult to know what degree of
improvement was made in those people who did
improve. Second, it is not known whether the nonresponders did not change or got worse. These
changes are not taken into account when
dichotomous data are used.
Continuous data were excluded if more than 50%
of people were lost to follow-up and data were
analysed as reported by the authors without
making any assumptions about those who were
lost to follow up. Continuous data were also
excluded if the rating scale used had not been
published in a peer-reviewed journal or if the data
did not meet a priori criteria for parametric data.
Data were combined using both fixed and random
effects models. Heterogeneity was investigated
with the Mantel–Haenszel 2 test of heterogeneity
to check whether differences in results were due to
chance alone. A significance level of 0.10 was
interpreted as evidence of heterogeneity. If
heterogeneity remained after the data were
combined using a random effects model, the data
were not pooled and results are reported and
discussed separately.
Sensitivity analyses were undertaken in all cases
where heterogeneity was detected and the effect of
including studies with high attrition rates was also
analysed. In addition, subgroup analyses were
undertake to detect any differences in outcomes
between (1) people with operationally defined
schizophrenia as opposed to those diagnosed by
clinical consensus, (2) people with varying degrees
of treatment resistance and those whose illness was
not designated as such, (3) people having
predominantly positive or negative symptoms of
schizophrenia and those without this designation,
and (4) people ill for less than 2 years and those at
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
a later stage of their illness. Publication bias was
assessed using a funnel plot.
The reviewers pooled data from different classes of
antipsychotics, including some that are no longer
used in current clinical practice. They found little
statistical heterogeneity in their analysis and
provided the current authors with raw data to
allow this issue to be explored in more detail if
necessary.
The methods used in this review were of a high
quality and the conclusions follow from the results.
The two reviews both explored the effectiveness of
ECT in people with schizophrenia. Although there
was a good degree of overlap between the two
reviews in the trials included, there were
important differences. These differences arose for
several reasons: (1) in a minority of cases, some
trials were included in one review and excluded in
another; (2) the trials were grouped differently for
analysis, particularly with respect to comparisons
with ECT and antipsychotic drugs; (3) the
different methods of analyses between the reviews
resulted in different trials providing usable data
for analysis; and (4) some trials were identified by
one review but missed by the other. The Cochrane
Schizophrenia Group provides reasons why each
study was excluded, whereas the UK ECT Group
does not, so it was not always possible to identify
why one study was included in one review but not
the other. Despite using different primary
methods to analyse the outcome data, the two
reviews drew similar conclusions regarding the
effectiveness of ECT in people with schizophrenia.
Although a large number of trials explored the
effectiveness of ECT in people with schizophrenia
and depression, both reviews reported that the
overall quality of trials is generally low. The
method of allocation was rarely described and
blinding was also inadequately explained. Often,
continuous data were only presented in graphical
form or only presented for the completer samples
and dropouts were not accounted for. There were
also significant gaps in the evidence of the efficacy
of ECT for important subgroups that are most
likely to receive ECT, such as older people and
women with postpartum depression. There was
little randomised evidence of the effectiveness of
ECT in people with mania and catatonia. There
was also little randomised evidence of the longterm efficacy or side-effects of ECT, with trials
rarely following people up beyond the course of
ECT. Furthermore, the methods used to measure
efficacy and side-effects do not adequately
19
Effectiveness
represent the views on users who receive ECT.
There were no trials exploring the impact of ECT
on quality of life. This had important implications
for the cost-effectiveness modelling within the
NICE review.
Quality of RCTs identified by the NICE
reviewers
The quality of the RCTs identified was also
generally low. Of the trials comparing ECT with
rTMS, one used concealed randomisation55 and
both were single blind.54,55 None of 13 trials
examining the efficacy of adjunctive or
continuation pharmacotherapy adequately
described the method of randomisation. Seven of
these trials were double blind,59–61,63,66–68 four
were single blind57,58,62,64 and in two it was not
clear whether the clinician or the patient was blind
to treatment allocation.56,65 One RCT examining
the impact of the educational video on patient
knowledge69 used concealed randomisation, but
was not blind and only measured knowledge at
follow-up using an instrument with no evidence to
support its psychometric properties. The second
trial was also unblinded and it was unclear
whether allocation was concealed.92
Non-randomised evidence
Owing to the gaps in the randomised evidence,
the non-randomised evidence was explored. Four
systematic reviews of non-randomised evidence
were identified that covered different aspects of
the NICE scope.
SURE review
The review conducted by SURE at the Institute of
Psychiatry53 aimed to summarise systematically
patients’ perspectives of ECT and to understand
the sources and nature of controversy about ECT
between some user and professional groups.
20
The review included all patients who had received
ECT, although little information was provided
regarding the types of participants included in the
studies and their conditions. Information was
provided on certain studies regarding the gender
and age of participants and the percentage of the
study sample who were sectioned and the
percentage who were legally compelled to have
ECT. No information was provided regarding the
stimulus parameters of ECT received by
participants included in the review, or whether
such information was reported in the original
studies. The review was more concerned with the
methods through which patients’ views were
elicited and the influence that this had on the
accuracy of such views.
Six main outcomes were considered in the review:
long-term memory loss, information and consent,
objective knowledge, felt compulsion, perceived
benefit of ECT and emotional reactions to ECT.
Long-term memory loss was defined as subjective
amnesia or gaps in memory still present at least
6 months after the course of treatment.
Information and consent was defined as the extent
to which patients felt that they had adequate or
sufficient information about ECT or were told
about the risks of ECT. Objective knowledge of
ECT was defined as how far people knew that ECT
involved the use of anaesthetic, an electric current
and a convulsion. Felt compulsion was defined as
the extent to which voluntary patients felt that
they had no choice but to have ECT. Perceived
benefit was defined as either the degree to which
consumers felt that ECT had helped them or
whether the user would agree to have ECT again.
Emotional responses were not defined in absolute
terms and included any comments indicating the
emotional tone of participants’ responses.
The review included both research studies and
testimonies. Testimonies were defined as an
individual speaking or writing directly about their
own experience of ECT. Reviewers did not restrict
inclusion of research studied by study type. The
studies included in the review used a wide range
of study designs and methods. They included
quasi-experimental studies, surveys and case
reports, and cross-sectional, retrospective and
prospective longitudinal study designs. The
studies used both quantitative and qualitative
methods.
The search strategy was described in detail and
combined searches of electronic databases for
research studies reporting patients’ views and
searches of a variety of other sources for
testimonies. The electronic sources searched
included PsycINFO, MEDLINE, Web of Science
and the King’s Fund database 1975–2001, Proquest
newspaper database, Mental Health Media
Testimony archive, searches of the Internet, e-mail
forums and chat rooms. Patient groups were also
contacted to identify unpublished patient-led
studies, local patient group newsletters, patientauthored chapters and collections of accounts of
ECT. Thirty-five research studies and an
unquantified volume of testimonies were identified.
No attempt was made by the reviewers to rank the
studies according to a hierarchy of evidence.
Instead, the reviewers describe a number of key
methodological issues that they identified as
having an influence on the ability of the studies to
Health Technology Assessment 2005; Vol. 9: No. 9
reflect patients’ views of ECT adequately and
accurately. These included the setting in which
attitudes to ECT were elicited, who conducted the
interview, the sample included in the study
(whether from clinical research studies or patientled surveys), the interval since ECT, the depth and
complexity of the questions asked, the degree to
which the questions are value laden and the
different methods (e.g. dichotomous, Likert scales
or in-depth interviews) used to quantify patients’
views.
A template was developed to analyse the research
studies and testimonies. When research studies
using a range of methodologies produced the
same results, findings were presented in terms of
‘at least X% of patients experience Y’. The degree
of variation in these percentages across the
different methodological factors discussed above
was explored to identify whether it was possible to
provide an overall percentage across the studies
and to explore the source of any variations.
The testimonies were analysed using a mixture of
content and discourse analysis. A grid was
developed with the review theme on the horizontal
axis and the testimony on the vertical axis to
illustrate the extent to which each testimony
contained each theme. Illustrative quotations of
each theme were used to allow the reader to
interpret the interpretative strength of each
theme. The inter-rater reliability of allocating
testimonies to categories in a subset of 25
testimonies was 83%.
This review53 did not conform to the traditional
methods of a systematic review because of
important differences in the focus and nature of
the review question. The review was conducted
rigorously, although the methods used to
demonstrate this rigour have not been used
previously or empirically tested. The reviewers’
conclusions follow from the results.
Reviews on younger people and children by Rey
and Walter70,71
Two systematic reviews70,71 examining the evidence
of the efficacy of ECT in younger people and
children were identified. The reviews were by the
same authors and one review71 was an update on a
previous review.70
The review included all studies examining the
effectiveness of ECT in younger people, defined as
people aged 18 years or under. The reviewers did
not identify any randomised evidence of the
effectiveness of ECT in this subgroup and did not
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
restrict inclusion criteria by study type. Studies
were only included if they provided sufficient
information on diagnosis and individual outcomes.
The outcomes of interest were not defined a priori
and appear to be governed by the content of the
studies identified. The outcomes covered in the
review were the percentage of participants with
remission or marked improvement of symptoms
immediately after ECT and at 6 months’ followup, adverse events including mortality, prolonged
seizures, subjective side-effects and cognitive
functioning.
The reviewers did not provide any information
regarding the medical and psychological databases
searched or give details of the manual searches, so
it is difficult to ascertain the comprehensiveness of
the review. Language bias was reduced as the
reviewers translated papers from other languages
into English and included them in the review. The
reviewers identified 60 reports describing 396
cases in their initial review and a further 11
reports by 1999. Information on diagnosis and
short-term outcome was available for 224 cases in
1999 and 154 out of 396 (39%) of cases in 1997.
The present authors’ searches did not identify any
studies published before 1999 that were not
included in the review.
No information was provided regarding how data
were abstracted. Two independent reviewers rated
the quality of the studies and only included those
that provided sufficient information on diagnosis
and outcome. However, other elements of study
quality were not taken into account when the
results of the papers were summarised. The
reviewers provided details of how they summarised
outcomes. Reviewers defined responders as those
who showed marked improvement or recovery
both immediately after ECT and 6 months postECT as defined by the study authors. However,
this assessment was not reported as being blind to
either the study authors or the results of treatment
and was open to some degree of subjective
interpretation. The data on efficacy were
summarised by adding case series and reports
together to produce an overall percentage of these
with a good outcome after ECT and at 6 months
by diagnosis. However, it was not clear whether
this was undertaken on an ITT basis. A qualitative
overview of data on adverse effects was
undertaken.
Overall, the quality of the studies included in the
review was poor and there were no controlled
studies. Reviewers’ quality ratings ranged from
21
Effectiveness
2 to 17 (minimum possible 0, maximum 20) with a
mean of 8.9 and a SD of 3.2. The quality of the
reporting within the studies was also poor; 43% of
studies in the 1997 review provided no diagnosis
for cases and only two reports used quantitative
measures of outcome. To examine the quality of
studies over time, the reviewers divided reports
into those published before and those published
after DSM-III in 1980. Studies published after
1980 had higher quality scores (mean 9.9, SD 2.9)
than those published before (mean 7.5, SD 3.2),
which was statistically significant at the 0.01 level
(t = 3.06, df = 58, p = 0.003).
It was difficult to ascertain whether this review
may have missed important studies owing to the
lack of information on search strategies. The
reviewers rated the quality of studies and only
included papers with sufficient information on
outcome and diagnosis. The methods of data
analysis of the efficacy of ECT are subject to some
degree of subjective interpretation and the
qualitative analysis of adverse events may be
subject to selective reporting. However, given the
poor quality of the evidence available, it is likely
that these reviews are currently the most
comprehensive available.
Hawkins and colleagues’ review of ECT in
catatonia78
One systematic review examining non-randomised
evidence of the effectiveness of somatic treatments
for people with catatonia78 was identified. This
aimed to summarise the literature on the
treatment of catatonia.
Papers were included if they provided sufficient
information to determine whether cases met DSMIV criteria for catatonia. Papers were excluded if
the clinical descriptions were likely to be due to
neuroleptic malignant syndrome (NMS). The
review included papers describing any treatment
of catatonia, although this was not defined a priori
but appeared to be governed by the content of the
studies identified. The treatments considered
included benzodiazepines, antipsychotics, ECT,
amobarbital, benztropine, ammantidine,
dontrolene, phenytonin, carbamazepine, ECT plus
other interventions (not defined) and
antipsychotics plus other interventions.
22
Only one outcome was considered by the review:
response to treatment. This was based on the
original authors’ clinical description of change in
catatonic symptoms after treatment. This response
was then retrospectively rated by the reviewers on
a three-point scale of none, partial or complete.
None was defined as no improvement or
worsening requiring a change in treatment, partial
as some improvement but incomplete requiring a
switch in treatment, and complete as resolution of
catatonic symptoms but not necessarily the
underlying pathology. However, no information is
given as to whether these ratings were made blind
to either authors or treatment type and as such the
results of the review are open to information bias.
Papers were excluded if either the treatment or
the response to treatment was inadequately
defined. The authors did not identify any
randomised evidence and inclusion was not
limited by study type.
Limited search strategies were used and only one
electronic database was searched (Paperchase)
from 1985 to 1994. Citation tracking from
included studies was used, but no attempt was
made to identify unpublished studies. The present
authors’ searches did not identify any further
studies published between these dates. The
reviewers identified 87 articles pertaining to the
treatment of catatonia and 70 (80%) met the
inclusion criteria for further analysis. The authors
provide specific reasons why certain studies were
excluded, including not meeting DSM-IV criteria
for catatonia, treatment responses not defined and
NMS suspected. In total, 270 treatment episodes
in 178 patients were included.
No information was provided regarding how the
data were abstracted or summarised. The unit of
analysis in the review was not explicitly defined,
but appears to be the treatment episode rather
than by case. The percentage of treatment
episodes having no, partial or complete response
were calculated for each treatment type. However,
it is not clear in the case of ECT whether
treatment episode implies a single administration
of ECT or a course of ECT. It was therefore
difficult to interpret the results of the review.
Given the poor description of the analysis and the
limited search strategies, the findings of this
review need to be treated with caution.
Miller’s review of ECT in pregnancy81
One systematic review of the use of ECT in
pregnancy81 was identified. This review aimed to
review case reports of the use of ECT during
pregnancy to clarify potential risks and
modifications of ECT techniques that make the
procedure safer for women.
Studies were included in the review if they reported
on the use of ECT in women during pregnancy.
Health Technology Assessment 2005; Vol. 9: No. 9
The primary outcome of interest was any adverse
events occurring as a result of ECT during
pregnancy. No randomised studies were identified
and inclusion was not limited by study type.
The review used a limited search strategy only
searching one electronic database (MEDLINE)
from 1966 to 1991. However, some reports were
identified dating back to 1942, although no
information is provided regarding how these were
identified. The present authors’ searches did not
identify any further studies not included in this
review. No information is given regarding whether
attempts were made to identify unpublished
literature. The reviewer identified 300 cases
reported in the literature.
No information was given regarding how data
were extracted and no attempt was made to rate
study quality. As such, the results of the review
may be biased owing to the risk of selective
reporting. The prevalence of adverse events in the
cases identified was outlined and no information is
provided regarding the efficacy of ECT in these
cases. It is not stated whether this information was
provided in the original studies. Given the limited
search strategies used by this review, the lack of
information about how data were extracted and
the relatively poor quality of the available
evidence, the results of this review should be
interpreted with caution.
Supplementary non-randomised evidence
identified by NICE reviewers
The authors also identified supplementary nonrandomised evidence of the efficacy of ECT in
subgroups of patients with catatonia, older people,
younger people and adolescents and its use in
pregnancy that were not included in the above
reviews.
In people with catatonia, two prospective case
series78,80 were identified. Both used a validated
instrument to measure outcomes and ECT was
used in participants who had failed to respond to
lorazepam.
For older people, one prospective cohort study73,74
comparing older people who had received ECT
with those who had not and three retrospective
cohort studies75–77 were identified. In one study75
some control over confounding variables was
attained through matching, but in two studies the
groups were different at baseline.76,77 In the
Kroessler and Fogel study,76 participants who
received ECT were medically and mentally more
ill than those who did not receive ECT. In the
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Philibert study,77 the ECT group was more likely
to be judged as suffering from psychomotor
retardation and to have had a prior course of ECT
than the pharmacotherapy group. The differences
in the Kroessler and Fogel76 study may be due to
the fact that a significant proportion of those who
did not received ECT were recruited from a
different hospital.
In adolescents an additional cohort study72 was
identified. There was a large loss to follow-up in
the ECT group, with only ten out of 20
adolescents identified as being treated with ECT
being included in the study. Although matching
allowed some control over confounding variables,
the two groups were different with regard to
diagnoses and the initial level of severity of their
diagnoses. Furthermore, participants were
interviewed a mean of 5.2 years post-ECT, leaving
considerable scope for information bias.
Finally, a further three case studies of the use of
ECT in pregnancy82–84 were identified. In all four
cases ECT was used because the women had failed
to respond to pharmacotherapy.
Overall, the quality of the systematic reviews of
non-randomised evidence is poor to moderate and
non-randomised evidence is poor. Only two of the
systematic reviews53,70 evaluated the quality of the
studies included and only one provided sufficient
detail of the search strategies used.53 In three of
the reviews70,78,81 the methods of abstracting
outcomes was open to a significant degree of
interpretation. However, the reviews are likely to
be the best evidence currently available in these
specific areas. The quality of the non-randomised
evidence included in these reviews or identified by
the present authors is poor. Most studies were
subject to confounding by baseline differences
between groups who received ECT and those that
did not, or lacked any control group at all.
Results of clinical effectiveness
Depression
One systematic review51 evaluating the efficacy of
ECT in people with depression was identified.
The results of this review and the present authors’
additional analyses are presented here.
ECT versus sham ECT
The UK ECT Group51 identified six trials
including a total of 256 patients that compared
real with sham ECT.65,93–97 In four trials the
position of the electrodes was reported: two used
unilateral placement,95,97 one bilateral96 and one
both.93 In four trials65,93,94,96 participants received
23
Effectiveness
ECT twice weekly and in the remaining two95,97 it
was administered three times weekly. Two trials
reported the waveform of ECT: one used sine
wave96 and the other brief pulse.97
Nine trials comparing real ECT with sham
ECT52,65,93–99 were identified. In four trials the
position of electrodes was reported: two used
unilateral placement,95,97 one bilateral96 and one
both.93 In four trials65,93,94,96 participants received
ECT twice weekly and in the remaining two95,97 it
was administered three times weekly.
Five trials specified the machine used to deliver
ECT: two used Duopulse Mk IV machines,93,97 two
used Ectron Mk IV machines95,96 and one used a
Transycon machine.94 Of the seven trials that
specified the dosage and waveform of ECT, none
used stimulus dosing but gave a fixed dose. Two
used sine wave at 150 V,52,97 one used sine wave
but did not specify the dosage,93 one used
chopped sine wave (dosage not specified),98 one
used 60% sine wave at 400 V,96 one used a doublesided unrectified wave at 40 J94 and only one used
brief pulse at 10 J.95
52,96
In two trials
the control arm also received at
least one real ECT. In Jagadeesh,52 participants in
the control arm received one real and five sham
ECT administrations. In Freeman,96 participants
in the control arm received two initial sham ECT
administrations and the remaining ones were real.
Efficacy at end of course The UK ECT Group51
found six trials that provided usable data on
depressive symptoms. The standardised difference
between real and simulated ECT was –0.91 [95%
confidence interval (CI) –1.27 to –0.54], indicating
a significant effect of real ECT. This result
translates to a mean difference in the HRSD score
of 9.67 (95% CI 5.72 to 13.53) in favour of real
ECT. Eighty-two per cent of patients receiving
ECT would be less depressed than the average
patient treated with sham ECT.
Four trials provided dichotomous data for analysis
of improvement at the end of an ECT
course.52,95–97 One trial used unilateral ECT95 and
the other three used bilateral ECT52,96,97 and were
analysed separately. The relative risk (RR) of a
reduction of at least a 50% in HRSD score for
unilateral ECT was 1 (95% CI 0.54 to 1.84, p = 1,
n = 32), indicating no statistically significant
difference between real and sham ECT.
24
Data from the three trials using bilateral ECT had
a relative risk of improvement, as defined by the
trialists at the end of a course, of 1.21 (95% CI
0.61 to 2.40, p = 0.6, n = 134), indicating no
statistically significant difference between real and
sham ECT. There was a significant degree of
heterogeneity within these three trials and removal
of the Freeman study96 resulted in a homogeneous
result with non-significant trend in favour of real
ECT (RR = 1.64, 95% CI 0.92 to 2.49, p = 0.1,
n = 84). The control arm of this trial only received
two sham ECTs; the rest were real ECTs. A further
remaining trial52 also included one real ECT
treatment in the control arm along with five sham
ECT treatments. Removal of this trial,52 leaving
one trial only, suggests that real bilateral ECT is
more effective than sham ECT (RR = 1.98, 95%
CI 1.05 to 3.73, p = 0.03, n = 70).
Discontinuations by end of treatment Discontinuations
occurred in both groups and three trials provided
usable data.93,94,97 The odds ratio for the
comparison was 0.80 (95% CI 0.30 to 2.40), which
indicates no difference between treatment groups.
The risk difference was –0.003 (95% CI –0.06 to
0.06).
Efficacy at 6-month follow-up Only one study97
reported depression rating scores at 6 months
following the end of ECT. This study reports a
two-point difference in final HRSD score in favour
of the sham group.
Adverse events: mortality No deaths occurred in
these trials.
Adverse events: cognitive functioning One trial97
provided data on cognitive functioning as an
immediate consequence of ECT, at the end of a
course of treatment and at 6 months’ follow-up.
A meta-analysis was not conducted because of
limited data and the UK ECT Group reviewers
describe the results as reported by the trial author
with a warning of the risk of bias due to selective
reporting.
Immediately after ECT, patients treated with real
ECT were more able to retrieve remote memories
than those treated with real ECT (retrograde
memory), but also had more word recognition
errors than those treated with sham ECT
(anterograde memory). The differential in
anterograde memory deficits between the two
groups was still present at the end of the course of
ECT and those treated with real ECT reported
more subjective memory complaints. At 6 months’
follow-up the authors reported no statistically
significant differences between those treated with
real or with sham ECT on measures of subjective
Health Technology Assessment 2005; Vol. 9: No. 9
memory complaints, new learning and remote
memories.
ECT versus inpatient care alone
The UK ECT Group identified one trial that
compared ECT with inpatient care alone100 and
included 139 patients. The mean decrease in
depression scores was 3.6 points greater on the
HRSD Scale in the ECT group. There was one
suicide in the ECT group and one death due to
other causes in the control arm.
ECT versus pharmacotherapy
Although these trials provide an estimation of the
relative efficacy of ECT compared with drug
therapy, most trials did not include sham ECT in
the control arm. As such, any difference may not
be due to the electrical stimulus and induction of
a seizure alone, but could be due to other
components of the ECT procedure, including
anaesthesia and nursing care.
Eighteen trials containing 1144 patients that were
included in the analysis65,99–115 were identified.
Bilateral ECT was used in five trials102,103,107,110,111
and unilateral in two.108,112 ECT was administered
twice a week in four studies65,107–109 and three
times a week in five studies.99,102,103,112,113 In five
trials65,103,107,110 participants were treated with
TCAs at doses between 75 and 150 mg of
imipramine or 150 mg of amitriptyline.99
L-Tryptophan was used in two trials at doses of
3 g108 and 6–8 g.109 The remaining trials used
paroxetine 40–50 mg,112 lithium 800 g,111
phenelzine 15–45 mg, either imipramine 50 g or
phenelzine 15 mg100 or a TCA or an MAOI.113
Only four studies102,107,111,112 required participants
to have failed to respond to at least one trial of
antidepressant drugs for inclusion in the study.
Treatment was continued for a range of durations.
Three studies65,111,113 reported the end of
treatment at 3 weeks, one at 3–5 weeks,102 four
trials reported 4 weeks,99,108,109,112 one at
5 weeks,107 one at 12 weeks110 and one at
approximately 2–4 weeks.100 Only three of the 18
trials identified used sham ECT in the
pharmacotherapy arm.65,105,115
Treatment with ECT led to a significantly greater
decrease in depressive symptoms than drug
treatment (standardised effective size –0.80, 95%
CI –1.29 to –0.29). This translates to a mean
difference of 5.2 (95% CI 1.37 to 8.87) in the
HRSD in favour of ECT.
Our own analysis compared ECT to each drug
class separately. One trial compared right
unilateral ECT with an SSRI (paroxetine
40–50 mg) in people with treatment-resistant
depression. The criterion for clinical improvement
in the trial was a reduction of at least 50% in
baseline HRSD scores. The relative risk of being a
responder was 3.14 (95% CI 1.39 to 7.11, n = 43,
p = 0.006) in favour of ECT.
Fourteen trials compared ECT with a TCA; in one
trial the TCA was combined with an MAOI102 and
in another it was combined with lithium111 in
people with treatment-resistant depression. Six
trials including 394 participants provided
dichotomous data for analysis.100,103,104,107,113,115
The criteria used to define responders varied
between trials. Two trials107 defined responders
using different criteria specified a priori based on
scores from quantitative outcome measures, while
the remaining four100,104,113,115 were based on
clinical opinion of improvement. To explore
whether the heterogeneity in defining responders
influences outcomes, the relative risk of being
both a responder and non-responder was
calculated and the trials were analysed separately
and together.
Pooled analysis of all six trials showed that people
treated with ECT were statistically significantly
more likely to be defined as a responder by the
trialists (RR = 1.42, 95% CI 1.17 to 1.72,
p = 0.0004) and also statistically significantly less
likely to be defined as a non-responder (RR =
0.47, 95% CI 0.31 to 0.69, p = 0.0002).
The UK ECT Group identified 18 trials containing
1144 patients that were included in the
analysis.65,99–115 Only published data are reported
by the reviewers. Not all studies provided usable
data.
Analysing the two trials103,107 based on a
quantitative assessment of improvement separately
resulted in no difference in the likelihood of being
defined as a responder between ECT and TCAs
(RR = 1.23, 95% CI 0.90 to 1.67, p = 0.58,
n = 38). Analysis of heterogeneous data from the
four trials100,104,113,115 based on clinical opinion
gave a relative risk of improvement of 1.63 (95%
CI 1.21 to 2.20, p = 0.001, n = 346) in favour of
ECT.
Efficacy at the end of treatment The UK ECT
Group51 found 13 trials that contributed sufficient
data to be included in the pooled analysis.
Discontinuations by end of treatment The UK ECT
Group51 found that discontinuations commonly
occurred in both groups. The odds ratio for the
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
25
Effectiveness
different types of treatment was 0.34 (95 CI 0.06 to
0.90), indicating a significant difference between
treatment groups. The risk difference was 0.03
(95% CI –0.0.09 to 0.03) in favour of ECT.
placement was described used non-dominant or
right unilateral placement. Three types of
unilateral placement were used: d’Elia and
Lancaster, and in one trial117 Raotma placement.
Depression at 6-month follow-up The UK ECT
Group51 found a single study that reported
depression rating at 6 months’ follow-up. The
results showed a lower score in the ECT group of
five points on the HRSD.
The trials rarely defined the course, duration and
frequency of treatment and those that did
demonstrated a significant degree of heterogeneity
in the methods used.
51
Adverse events: mortality The UK ECT Group
found one trial that reported a death in each arm
of the trial.100
Adverse events: cognitive functioning The UK ECT
Group51 reported that the data on cognitive
functioning were heterogeneous owing to the
different tests of cognitive functioning used. The
reviewers reported the data on cognitive functioning
as reported by the authors of the trials, with the
warning of a risk of bias due to selective reporting.
The UK ECT Group51 reported that no trials were
identified that provided data on orientation, new
learning or subjective distress as an immediate
consequence of ECT treatment. Three randomised
trials99,105,110 measured cognitive functioning at
the end of a course of ECT, comparing patients
treated with drugs with those treated with ECT.
One assessed retrograde memory,105 but only
reported within-group results of tests, which are
difficult to interpret. With regard to anterograde
memory, McDonald and colleagues99 reported no
statistically significant difference between patients
treated with ECT and those treated with drug
therapy on the Weschler–Bellevue Intelligence
Scale (WBIS). Bagadia and colleagues105 only
reported within-group results of tests, which are
difficult to interpret. Finally, Gangadhar116
reported that more patients treated with ECT
complained of loss of memory than those treated
with drug therapy. No trials reported on cognitive
function at longer than 6 months.
Unilateral versus bilateral ECT
The UK ECT Group51 identified 28 trials31,93,117–141
using 1408 patients; in 21 of these, data were
available to calculate effect size.
26
Various electrode placements were used for both
unilateral and bilateral ECT. Two studies129,142
reported bitemporal placement, two used
bifrontal122,139 and one reported
bifrontotemporal.132 In three trials93,125,139 either
dominant or non-dominant unilateral placements
were reported and the remaining studies where
Efficacy at end of course The UK ECT Group51
reports that the standardised effect size between
the two types of electrode placement was –0.32
(95% CI 0.46 to –0.20), which is a significant result
in favour of bilateral ECT. This translates to a 3.58
(95% CI 2.24 to 5.15) change in depression score
in favour of bilateral ECT. A test of heterogeneity
produced no effects of publication year on this
outcome. Removal of trials by Sackeim that may
have included different populations from the rest
of the studies shifted the point estimate of effect
size, but it remained statistically significantly in
favour of bilateral ECT.
Discontinuations The UK ECT Group51 found two
trials130,137 that reported events for this outcome,
so it was not possible to summarise results.
Numbers were similar for each group.
Adverse events: mortality The UK ECT Group51
found two trials130,140 that reported mortality, but
the data were unusable.
Adverse events: cognitive functioning The UK ECT
Group51 reported that the data on cognitive
functioning were heterogeneous owing to the
different tests of cognitive functioning used. The
reviewers reported the data on cognitive
functioning as reported by the authors of the
trials, with the warning of a risk of bias owing to
selective reporting. Their findings are as follows.
(a) As an immediate consequence of ECT
● Orientation
Six studies found that the time to recovery of
orientation was longer for patients treated with
bilateral ECT than for those treated with
unilateral ECT.31,127,129,134,135,138 Similarly,
Fleminger and colleagues139 found greater
impairment of orientation among those treated
with bilateral compared with either nondominant or dominant unilateral ECT when
assessed at 10 minutes post-treatment.
●
Retrograde memory
Two studies assessed retrograde memory
postictally127,135 and both found greater
Health Technology Assessment 2005; Vol. 9: No. 9
impairment in people treated with bilateral
ECT. Sackeim and colleagues135 found no
difference in retrograde memory between those
treated with bilateral ECT and high-dose
unilateral ECT. Levy119 reported data during
the course of ECT with results given after the
sixth session of ECT. No difference was found
on tests of recent personal events between
groups, although there was a significant
deterioration in memory for general events
among patients treated with bilateral ECT
compared with non-dominant unilateral ECT.
●
found that where differences did occur, those
receiving bilateral ECT fared worse than those
receiving unilateral ECT. Two studies showed no
differences in new learning tasks between
bilateral and non-dominant unilateral
ECT,128,138 while one study found that those
treated with bilateral ECT performed better
than those treated with dominant unilateral
ECT.138
Three studies described cognitive testing
subsequent to testing at the end of the course of
treatment: Bidder and colleagues140 at 10 days,
Letemendia and colleagues128 and Halliday and
colleagues138 at 3 months. In addition, Fraser
and Glass134 reported results at 3 weeks. Bidder
and colleagues140 and Fraser and Glass134 found
no statistically significant differences between
bilateral and unilateral ECT. At 3 months
following treatment, Halliday and colleagues138
reported that patients who had been treated
with non-dominant unilateral ECT performed
better on digit span and delayed non-verbal
learning compared with those treated with
bilateral ECT, with no difference between
dominant unilateral and bilateral ECT.
Letemendia and colleagues128 found no
difference between non-dominant unilateral
and bitemporal ECT on four tests of verbal and
non-verbal functioning when patients were
tested 3 months post-ECT.
Anterograde memory
One study140 reported repeated testing of new
learning and episodic memory during the course
of ECT and found that bilateral ECT resulted in
greater impairment of new learning at 36 hours
post-ECT than unilateral ECT.
The Weschler Memory Scale (WMS) was used in
six studies to compare pre-ECT memory
functioning with functioning at a point during
the course of ECT.119,122,128,133,134,139 Overall,
these studies show that where there are
differences in cognitive functioning, people
treated with bilateral ECT fared worse than
those treated with unilateral ECT. People who
received high-dose or dominant unilateral ECT
also reported greater memory impairment than
those receiving low-dose or non-dominant
unilateral ECT.
●
Overall cognitive functioning
Three studies31,130,135 reported results from
overall cognitive testing in the week following
the end of the randomised phase of treatment.
These studies showed that where differences did
occur, people who received bilateral ECT
showed greater cognitive impairment than
those receiving unilateral ECT. Two studies
reported results at later stages: Sackeim and
colleagues31 at 2 months and Heshe130 at
3 months. They showed no statistically
significant differences between bilateral and
unilateral ECT.
●
Subjective reports
Two studies120,142 described subjective report of
cognitive functioning at the end of the course of
ECT. Horne and colleagues142 found that
patients treated with bilateral ECT described
more subjective impairment of memory than
those treated with non-dominant unilateral
ECT, including a subjectively greater
impairment of recall of the events surrounding
their admission. Weiner and colleagues120
however, reported no statistically significant
Subjective distress
One study144 reported on subjective complaints
of memory impairment during the course of
ECT. In this study, patients receiving bilateral
ECT reported both greater post-treatment
confusion and memory problems than the
group receiving non-dominant unilateral ECT.
Complaints of cognitive side-effects were
essentially non-existent among patients treated
with non-dominant unilateral ECT.
(b) At the end of a course of ECT
● Retrograde memory
Four studies31,120,140,145 reported results from
testing of retrograde memory within a week of
end of a course of ECT. All these studies
reported greater impairment among patients
treated with bilateral ECT.
●
●
Anterograde memory
Seven studies118,120,128,133,135,138,140 reported
results from tests assessing anterograde memory
within 7 days of the end of the randomised
phase of treatment. Five studies120,133,135,140,146
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
27
Effectiveness
difference in global self-rating of memory
function 2–3 days post-ECT.
(c) Long-term (>6 months)
Two studies120,140 reported long-term data. Weiner
and colleagues120 found no between-group
difference in anterograde memory function at
6 months, with scores returning to at least
pretreatment levels. Long-term personal memory
was more impaired in the bilateral than in the
non-dominant unilateral ECT group, although
there were no differences in recall of famous
events or faces. At 1 year following treatment,
Bidder and colleagues140 found no difference
between bilateral and unilateral ECT-treated
patients on assessment of verbal memory and both
groups had improved since they were tested at
10 days post-treatment.
(b) At the end of a course of ECT
Four randomised trials125,138,139,147 measured
cognitive functioning at the end of a course of
ECT, comparing ECT applied to the dominant
hemisphere with ECT applied to the nondominant hemisphere, but in one the results were
difficult to interpret.147 The remaining three all
showed that people treated with unilateral ECT to
the dominant hemisphere did worse than those
treated with unilateral ECT to the non-dominant
hemisphere on tests of new learning (anterograde
memory).125,138,139 The trials did not assess
retrograde memory or subjective distress and none
of the trials assessed cognitive functioning at
6 months.
Unilateral electrode placement
The UK ECT Group51 found five
trials125,138,139,147,148 that provided data on 174
patients. The trials provided limited descriptive
information concerning the parameters of ECT
administration. Their analyses of these trials are
detailed below.
Bilateral electrode placement
Efficacy: end of course and 6 months The
standardised effect size between the two types of
electrode placement was 0.387 (95% CI –0.09 to
0.87), which does not represent a significant effect.
The result translates to a 3.87-point (95% CI –0.90
to 8.70) non-significant change on the HRSD in
favour of unilateral dominant rather than nondominant electrode placement. No studies
reported depression ratings at 6 months.
There were no discontinuations or deaths reported
in these trials.
Adverse events: cognitive functioning
(a) As an immediate consequence of ECT
Three randomised trials138,139,148 measured
cognitive functioning immediately after ECT. It was
not possible to perform a meta-analysis because of
the interstudy variations in the measures used.
●
28
anterograde memory,148 but the reviewers found
the results difficult to interpret.
Orientation
Two studies assessed recovery following
ECT138,139 and both reported that patients
treated with unilateral ECT to the nondominant hemisphere recovered orientation
more quickly than those treated with unilateral
ECT to the dominant hemisphere.
No trials measured retrograde memory or
subjective distress and one trial measured
The UK ECT Group51 found two trials128,149 that
compared frontotemporal and temporoparietal
bilateral electrode placement and included results
for 100 patients. Participants received brief-pulse
ECT three times a week in both trials. In one
trial128 the dose was give at seizure threshold and
in the other149 it was given at 1.5 times this value.
In the latter trial patients were treated until
remission or up to a maximum of 12 weeks. The
UK ECT Group’s51 analyses of these trials are
described below.
Efficacy at end of course The pooled standardised
effects analysis showed no difference between the
two forms of bilateral placement, with a result of
–0.01 (95% CI –0.75 to 0.74). This translates to a
mean change in HRSD score of 0.05 (95% CI
–4.34 to 4.28).
Depression rating at 6-month follow-up One study128
reported follow-up data for efficacy scores. The
final scores of the two groups had a difference of
two points in favour of temporoparietal
positioning.
Adverse events: mortality No deaths occurred in
either of the trials.
Adverse events: cognitive functioning
(a) As an immediate consequence of ECT
No trials were identified describing results for
orientation, cognitive change or subjective
distress.
(b) At the end of a course of ECT
Two randomised trials128,149 measured cognitive
functioning at the end of a course of ECT. It was
Health Technology Assessment 2005; Vol. 9: No. 9
not possible to perform a meta-analysis because of
the interstudy variations in the measures used.
Bailine and colleagues149 reported that patients
treated with bitemporal ECT had lower Mini
Mental State Examination (MMSE) scores after
treatment than patients treated with bifrontal ECT.
Letemendia and colleagues128 reported that there
were no statistically significant differences on
several cognitive measures between patients treated
with bitemporal ECT at 7 days or 3 months after
treatment.
(b) At the end of a course of ECT
Four randomised trials151,152,154,155 measured
cognitive functioning at the end of a course of
ECT. The reviewers were not able to conduct a
meta-analysis because of variations between
studies in the measures used.
●
Retrograde memory
Two studies151,152 reported that patients treated
with ECT three times weekly did worse than
those treated twice weekly on tests of personal
memory. No statistically significant differences
were apparent 1 month after the course of
treatment.
●
Anterograde memory
Kellner and colleagues155 reported no
statistically significant difference in WMS scores
between patients treated with ECT once weekly
and those treated three times weekly. Lerer and
colleagues151 reported a poorer performance on
anterograde and retrograde immediate and
delayed facial recognition and digits backwards
in patients treated three times weekly compared
with those treated twice weekly. No statistically
significant differences were apparent 1 month
after the course of treatment.
●
Overall cognitive functioning
Kellner and colleagues155 reported no
statistically significant difference in MMSE
scores between patients treated with ECT once
weekly and those treated three times weekly.
Lerer and colleagues151 reported a greater
deterioration in overall function in patients
treated three times weekly than in those treated
twice weekly. No statistically significant
differences were apparent 1 month after the
course of treatment. Vieweg and Shawcross154
reported no statistically significant differences
in MMSE scores between patients treated three
times weekly and those treated twice weekly.
No trials were identified that provided data on
cognitive function at 6 months or more
post-ECT.
Frequency of ECT
The UK ECT Group51 identified six trials
containing results for 210 patients.150–155 Two
trials153,155 compared once-weekly with three timesweekly ECT, while the remaining four150–152,154
compared twice-weekly and three times-weekly
administrations. Their analyses are set out below.
Efficacy at end of course The reviewers analysed the
four trials comparing ECT twice weekly versus
three times weekly separately and together with
the two that reported once versus three times
weekly. The standardised effect size was –0.30
(95% CI –0.76 to 0.20) for twice versus three times
weekly and 0.83 (95% CI –0.39 to 1.89) for once
versus three times weekly. When all the results
were combined there was no significant difference
between the two regimens, with a mean change in
depression score of 0.40 (95% CI –5.26 to 6.30) in
favour of more frequent ECT administration.
Discontinuation Two trials151,154 reported
discontinuations and they were equivalent for both
groups.
Depression rating at 6-month follow-up No data were
available.
Adverse events: mortality One trial150 reported a
death due to suicide. No analysis was possible
based on the limited data.
Adverse events: cognitive functioning
(a) As an immediate consequence of ECT
One randomised trial151 measured cognitive
functioning immediately after ECT. Lerer and
colleagues151 reported no difference in time to
reorientation in patients treated three times
weekly compared with those treated twice weekly.
No trials provided data on retrograde memory,
anterograde memory or subjective distress.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
No trials reported data on subjective distress or
any aspect of cognitive functioning at 6 months.
Dose of electrical stimulus
The UK ECT Group51 identified seven trials
containing results for 342 patients.31,135,153,156–159
The actual doses used in the trials varied and the
reviewers classed the dose as ‘higher’ and ‘lower’ for
the purposes of analysis. In the McCall trial,158
lower dose was defined as 2.5 times the convulsive
threshold and higher dose was 403 mC for 2
seconds, Janakiramaiah and colleagues153 used
threshold for lower dose and 240 mC for higher
dose, Sackeim and colleagues135 used either 50% or
29
Effectiveness
150% above threshold as lower dose and 500%
above threshold as higher dose, McCall and
colleagues157 had lower dose as 2.5 times the
convulsive threshold compared with a higher dose
of 403 mC for 2 seconds, and Warren and Tissera159
used 7–10 J for lower and 40–555 J for higher dose.
The UK ECT Group51 analyses are detailed below.
people treated with high-dose or low-dose
energy pulses on tests of new learning.158,159
Two studies found worse scores on tests of
new learning in those treated with high-dose
pulse compared with low or moderate
doses.31,135
●
Efficacy Six trials provided usable data for
analysis.31,135,153,157–159 The results indicated a
standardised treatment effect of 0.73 (95% CI 0.41
to 1.08) or mean change in HRSD score of 5.24
(95% CI 2.94 to 7.75) in favour of the higher dose
group. No trials provided information on
depression ratings at 6 months.
Discontinuations One trial157 reported events for
discontinuations, with similar numbers in each
arm, so analysis was not possible.
Adverse events: mortality No deaths were reported
in these trials.
Adverse events: cognitive functioning
(a) As an immediate consequence of ECT
● Orientation
Three trials found that people treated with
high-dose unilateral ECT took longer to recover
than those treated with low-dose unilateral
ECT.31,135,158
●
Anterograde memory
Sackeim135 reported that patients treated with
high-dose unilateral ECT had worse scores on
some measures of new learning than patients
treated with low- or moderate-dose unilateral
ECT.
No trials provided data on retrograde memory
or subjective distress.
(b) At the end of a course of ECT
Five randomised trials measured cognitive
functioning at the end of a course of
ECT.31,135,157–159 The reviewers did not conduct a
meta-analysis because of the interstudy variations
in the measures used.
●
●
30
Retrograde memory
None of the studies found any differences
between people treated with high-dose or lowdose energy pulses on tests of personal,
autobiographical, subjective or overall
memory.31,135,157–159
Anterograde memory
Two studies found no differences between
Overall cognitive functioning
Sackeim and colleagues31 reported no
statistically significant differences between highand low-dose bilateral and unilateral ECT on
total MMSE score. McCall and colleagues158
reported that patients treated with fixed highdose unilateral ECT performed worse on the
MMSE than patients treated with titrated,
moderate-dose unilateral ECT.
Stimulus waveform
The UK ECT Group51 found eight trials containing
results for 296 patients.120,126,129,131,159–162 Five
trials provided data for a meta-analysis which
compared brief-pulse and sine-wave ECT for
electrical stimulation.129,131,159,160,162 The UK ECT
Group’s51 analyses of these trials are described
below.
Efficacy at end of course: depression rating The
standardised effect size was 0.62 (95% CI –0.31 to
1.54) in favour of sine wave. This translates to a
mean change in HRSD score of 4.21 (95% CI
–2.08 to 10.5). The trials did not provide any data
on depression ratings at 6 months or
discontinuations.
Adverse events: mortality No deaths occurred in the
trials.
Adverse events: cognitive functioning
(a) As an immediate consequence of ECT
Two trials measured cognitive functioning
immediately after ECT, comparing sinusoidal with
brief pulse.126,129
●
Orientation
Valentine and colleagues129 reported that
patients receiving brief-pulse ECT began
breathing, recovered consciousness and became
orientated more quickly than patients receiving
sinusoidal ECT.
●
Retrograde memory
Daniel and Crovitz126 reported no statistically
significant difference between brief-pulse and
sine-wave ECT on several measures of
perceptual learning and autobiographical
memory. Valentine and colleagues129 reported
that patients receiving brief-pulse ECT had
Health Technology Assessment 2005; Vol. 9: No. 9
better recall of word associations learned shortly
before the treatments than did patients
receiving sinusoidal ECT. No trials were
identified that reported data for anterograde
memory or subjective distress as an immediate
consequence of ECT.
(b) At the end of a course of ECT
Two randomised trials120,159 measured cognitive
functioning at the end of a course of ECT,
comparing sinusoidal with brief-pulse ECT. The
reviewers did not conduct a meta-analysis because
of the interstudy variations in the measures used.
The results are described as reported by the
authors, with the consequent risk of bias due to
selective reporting.
●
●
●
Retrograde memory
Warren and Tissera159 reported no statistically
significant difference between pulse (high and
low energy) and sine-wave ECT on logical
memories, verbal recognition, facial recognition
or a measure of remote memories. Weiner and
colleagues120 reported no statistically significant
difference in overall self-rating memory (which
seemed to improve in all groups) between
patients treated with pulse and sine-wave ECT,
but found that patients treated with sine-wave
ECT received more electrical energy and
performed worse on measures of retrograde
memory.
Anterograde memory
Warren and Tissera159 reported no statistically
significant difference between pulse (high and
low energy) and sine-wave ECT on digit span.
Weiner and colleagues120 reported that patients
treated with sine-wave ECT performed worse on
measures of anterograde memory, including
verbal paired associations, paragraph recall,
facial recognition and complex figure
reproduction.
Overall cognitive functioning
Weiner and colleagues120 reported no
statistically significant difference on a
neuropsychological test battery between patients
treated with pulse and sine-wave ECT.
No data were available on subjective distress at
the end of a course of ECT.
(c) At 6 months
Weiner and colleagues120 reported no statistically
significant difference at 6 months post-treatment
in overall self-rating memory between patients
treated with pulse and sine-wave ECT.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Ultrabrief ECT versus standard ECT
The UK ECT Group51 found one study163 that
considered the use of ultrabrief stimulus for ECT
administration. The study contained 72 patients
but did not report efficacy scores. On cognitive
testing, 1 week after last treatment, all patients
showed improvement on tests of immediate
reproduction and delayed reproduction, with no
significant difference between the groups. There
was no significant difference between the two
groups on tests of subjective memory change or
overall on tests of forgetting. No discontinuations
or mortality were reported.
Number of ECT sessions
The UK ECT Group51 identified a single trial164
that considered this comparison. Twenty-six
patients were recruited and available for
assessment; no depression rating scores were given
but none of the patients discontinued treatment or
died. No usable data on cognitive functioning
were reported.
Number of seizures per treatment session
The UK ECT Group51 found a single study165 that
compared the induction of one seizure per
treatment session with the induction of two
seizures per treatment session. The study
contained data on 29 patients with no deaths or
discontinuations. The change in HADS score was
greater in the multiple monitored ECT group by
4.1 points. Charts recorded significantly greater
post-treatment confusion among patients treated
with multiple monitored ECT. No other tests of
cognitive functioning were performed.
Extra sessions of ECT
The UK ECT Group51 identified one trial166 that
considered the effect of performing two additional
ECT sessions above what they classified as
medically sufficient. Seventy-five patients were
recruited; four patients in the sufficient group and
five in the extra group refused ECT at some point.
No usable efficacy data or cognitive data were
reported and no deaths occurred.
Post-ECT nursing care
The UK ECT Group51 found one study167 that
compared usual nursing care post-ECT with a
procedure in which patients were taken to a small,
dimly lit room for 2–4 hours, where ambient noise
was minimised but nurses monitored patients as
usual (rest). The trial recruited 19 patients, none
of whom discontinued treatment or died. No
measure of efficacy of the techniques was made.
They found that there were statistically
significantly fewer subjective memory complaints
31
Effectiveness
in the rest group then in the ward group. There
were no other significant differences on cognitive
testing.
ECT versus rTMS
Two RCTs were identified that evaluated the
efficacy of rTMS with ECT in people with
depression, including 63 participants.54,55 One
trial compared ECT alone with rTMS,54 while the
other compared ECT with ECT plus rTMS.55 One
trial specifically included people with medicationresistant depression.55 Both trials used unilateral
ECT placement and only one described the
frequency of administration, which was three times
per week.55 The rTMS methods different between
the two studies. In Pridmore,55 a Magtism Super
Rapid Stimulator was used, with a Magstim
70-mm double coil, at an intensity of 100%, a
frequency of 20 Hz and a train length of 2
seconds. The number of trains was 30, with an
intertrain interval of 20 seconds. In Grunhaus,54
the motor threshold was determined daily by
electromyography and stimulus intensity was the
lowest machine power output that would provide
five of ten stimulations with a muscular-evoked
potential (MEP) of at least 50 V. Electrodes were
placed over the left dorsolateral prefrontal cortex.
During stimulation the coil was held with the
handle towards the back of the head. rTMS was
administered five times a week for 4 weeks (for a
total of 20 stimulations).
Efficacy: depression at end of course Only one trial54
provided usable data on 40 participants for
analysis. The efficacy of the treatment was
measured using continuous data from the HRSD.
The WMD between ECT and rTMS was 6.8 (95%
CI 1.41 to 12.19, n = 40), which was statistically
significant at the 0.01 level in favour of ECT.
Thus, people treated with ECT fared, on average,
6.8 points better on the HRSD than people
receiving rTMS. Efficacy was also measured as a
dichotomous variable, with responders defined as
those whose scores at the end of the course were
greater than or equal to 60 on the Global
Assessment of Functioning (GAF) scale and had
decreased by at least 50% on the HRSD from
baseline, but the data were unusable. There were
no discontinuations or deaths reported in this trial.
Adverse events: side-effects The two trials only
reported data on subjective side-effects.
32
Grunhaus and colleagues54 (ECT versus rTMS)
found that five patients in the rTMS group
complained of mild headache, which responded to
analgesics. In one patient and only during one of
the treatment sessions an MEP discharge was
noted 20 ms after each magnetic pulse.
Pridmore55 (ECT versus ECT + rTMS) used a sixitem subjective side-effects questionnaire derived
from a report on the side-effects of ECT
(Gomez277). Over the 2-week study period the
ECT-only stream scored 56 positive responses on
the side-effects questionnaire, while the ECT plus
rTMS stream scored a little over half of that
number. None of the observed differences in
proportions of patients having side-effects were
statistically significant. The main symptoms were
‘memory problems’, ‘headache’ and ‘muscle
pains’; these scored most complaints in both
streams. Memory problems were twice as common
in the ECT-only stream. Because of the small
sample, the possibility that these results are due to
the play of chance cannot be excluded.
ECT plus pharmacotherapy versus ECT plus
placebo/different pharmacotherapy
The present review identified 11 trials56–66 that
compared ECT combined with pharmacotherapy
versus ECT combined with either placebo or a
different type of pharmacotherapy. Two trials
compared unilateral ECT combined with
L-tryptophan versus unilateral ECT and
placebo.59,60 Three trials compared ECT
combined with imipramine versus ECT combined
with placebo.61,62,65 In one study the dosage of
imipramine ranged from 25 to 50 mg,61 in
another the dosage was 25 mg three times daily62
and in the third it ranged from 150 to 220 mg.65
Imlah and colleagues62 also had an arm in the
trial where ECT was combined with phenelzine
(15 mg three times daily). None of the trials
reported any details of electrode placement.
Lauritzen and colleagues64 had two arms in the
trial that were separately randomised to receive
either bilateral then unilateral ECT combined with
paroxetine (30 mg) or placebo (Group A), or
bilateral then unilateral ECT combined with
paroxetine (30 mg) or imipramine (150 mg). Kay
and colleagues63 compared ECT combined with
either amitriptyline (50–150 mg) or diazepam
(4–12 mg). Mayur and colleagues56 compared
unilateral ECT combined with continuation of the
antidepressants (either TCAs or SSRIs, dose or
type not defined) that participants were taking on
entry to the trial versus ECT alone. Arfwidsson
and colleagues58 compared bilateral ECT
combined with chlorpromazine (50–150 mg)
versus bilateral ECT combined with placebo.
Shiah and colleagues57 compared either unilateral
or bilateral ECT combined with pindolol (7.5 mg)
with ECT and placebo. Coppen and colleagues66
Health Technology Assessment 2005; Vol. 9: No. 9
compared ECT and lithium (plasma levels
between 0.8 and 1.2 mmol l–1) continuation
therapy with ECT and placebo. In five trials, the
length of ECT treatment was determined by a
clinical decision on response to ECT,58,60–62,64
while Shiah and colleagues57 fixed the number of
treatments at six in each arm. In the remaining
four trials, the length of ECT treatment was
unclear.56,59,63,66 In five of the trials, participants
continued to take the pharmacotherapy they had
been randomised to after ECT treatment and were
followed up at 3 months,63 6 months61,62,64 or
1 year66 to assess the impact of post-ECT
pharmacotherapy on relapse rates.
Efficacy: depression rating at end of course Three
trials provided dichotomous data on global
improvement,57,58,60 but were analysed separately
owing to the different types of drugs in the
comparison. Shiah and colleagues57 defined
responders as those scoring less than 12 on the
29-item version of the HRSD, whereas Arfwidsson
and colleagues58 and d’Elia and colleagues60
defined improvement according to clinical
opinion. One trial provided dichotomous data on
relapses at end of ECT course based on clinical
opinion.63
In the Arfwidsson trial58 there was a nonsignificant trend for people treated with ECT plus
chlorpromazine to be more likely to have
improved than people treated with ECT and
placebo (RR = 1.13, 95% CI 0.88 to 1.46, n = 52).
Shiah and colleagues57 also found a nonsignificant trend for people treated with pindolol
to have responded after six ECTs compared with
those treated with placebo (RR = 10.8, 95% CI
0.66 to 177.33, p = 0.1, n = 20). There was also
no difference in the likelihood of being a
responder in the d’Elia trial60 when ECT was
combined with either L-tryptophan and placebo
(RR = 0.96, 95% CI 0.83 to 1.12, p = 0.6, n = 61).
Kay and colleagues63 found that those treated with
ECT plus diazepam were more likely to have
relapsed at the end of ECT course than those
treated with ECT plus amitriptyline (RR = 0.55,
95% CI 0.33 to 0.90, p = 0.02, n = 132).
Three trials provided continuous data on
completer samples for analysis and all used the
HRSD; Mayur56 and Lauritzen64 used the 17-item
version and Shiah57 used the 29-item version. All
trials were analysed separately owing to the
different drugs involved in the comparisons.
Lauritzen and colleagues64 found no statistically
significant differences in scores on the HRSD
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
between those treated with ECT plus paroxetine
and those treated with ECT plus placebo at the
end of the course of ECT. The WMD was –0.30
(95% CI –0.301 to 2.4, n = 35, p = 0.83) in favour
of paroxetine. The WMD between paroxetine plus
ECT and imipramine plus ECT was –3.00 (95% CI
–5.65 to 0.33, n = 52), which is a statistically
significant difference at the 0.05 level in favour of
imipramine.
Mayur and colleagues56 found no statistically
significant differences in HRSD scores between
ECT combined with antidepressants and ECT
alone at 6 weeks’ follow-up (WMD = 1.7, 95% CI
–5.54 to 8.94, p = 0.6, n = 22).
Shiah and colleagues57 found statistically
significantly lower scores in participants treated
with ECT plus pindolol compared with
participants treated with ECT plus placebo after
six ECTs (WMD = –9.10, 95% CI –16.08 to –2.12,
p = 0.01, n = 15).
Efficacy: prevention of relapses Only one trial
provided usable data for analysis regarding the
efficacy of continuing to take pharmacotherapy
following the course of ECT in preventing
relapses.62 There was a statistically non-significant
trend for those treated with imipramine to have a
reduced risk of experiencing a relapse (RR = 0.83,
95% CI 0.58 to 1.19, p = 0.32, n = 100). However,
if those who withdrew from the trial or were lost to
follow-up were not allocated the worst outcomes
and removed from the nominator in the analysis
then those who continued to take imipramine were
statistically significantly less likely to experience a
relapse at 6 months (RR = 0.33, 95% CI 0.16 to
0.71, p = 0.005). There were no statistically
significant differences in the likelihood of
relapsing between those treated with TCAs and
MAOIs (RR = 0.80, 95% CI 0.52 to 1.24, p = 0.3,
n = 100).
Coppen and colleagues66 compared the mean
number of weeks spent depressed during the
following 6 months. They found a statistically
significant different in the number of weeks spent
depressed during the 6 months after ECT between
those taking lithium and those taking placebo, in
favour of lithium. The WMD was 0.90 (95% CI
0.29 to 1.51, p = 0.004).
In the study by Arfwidsson and colleagues,58
chlorpormazine was discontinued at the end of the
ECT course and patients were followed up at
3 months. They found that those who received
chlorpromazine in addition to ECT were not
33
Effectiveness
statistically significantly less likely to experience a
relapse at this time compared with those who
received ECT plus diazepam (RR = 1.17, 95% CI
0.76 to 1.79, p = 0.48, n = 57).61
Adverse effects Two studies explored adverse effects
using the UKU scale of adverse drug reactions and
the Columbia side-effect checklist.56,64 Lauritzen
and colleagues64 found only minor differences
between the treatment groups on the Udvlag for
Kliniske Undersøgelser (UKU) scale. Paroxetine
was associated with increased frequency of
dreaming periods at night according to
assessments after 2 months, but not after
6 months. Imipramine was associated with
complaints of constipation, although these only
reached significance at 3 months.
Mayur and colleagues56 found no significant
differences between groups in the mean number of
side-effects at the 2- or the 4-week stage of the
acute phase as measured by the Columbia checklist.
The antidepressant group had significantly higher
mean ratings on the anticholinergic subscale of
UKU. There were no significant differences in any
other UKU subscale. No patient had significant
arrhythmias. There were no intolerable
anticholinergic side-effects among patients with
tricyclic drugs and ECT warranting discontinuation
of the drug during the ECT course.
Continuation pharmacotherapy
The present group identified a further two
double-blind trials67,68 that compared different
approaches to antidepressant treatment following
successful treatment with ECT. In these trials,
participants had to have responded to ECT and
were then randomised to different
pharmacotherapies. Grunhaus and colleagues67
defined responders as those with an HRSD
(17-item version) score of less than or equal to 10
that was maintained for a week. Sackeim and
colleagues68 defined responders as those who had
a decrease of at least 60% on the HRSD (17-item
version) from baseline.
34
Sackeim and colleagues68 compared continuation
with nortriptyline (25 mg) alone versus
nortriptyline plus lithium (300 mg) versus placebo.
Grunhaus and colleagues67 compared fluoxetine
(20 mg day–1 combined with melatonin (5 mg)
with fluoxetine (20 mg) and placebo. Sackeim used
either bilateral or unilateral ECT and Grunhaus
used unilateral ECT that was switched to bilateral
if a response was not achieved within six
treatments. In the Sackeim trial,68 ECT was
administered three times weekly for a duration
determined on clinical grounds. In both the
Grunhaus67 and Sackeim trials, seizure threshold
was determined either using the method of
limits67 or by empirical titration;68 the stimulus
was delivered at 2.5 times threshold in Grunhaus67
and at 1.5 times threshold in Sackeim.68
Efficacy: relapses Two trials67,68 provided usable
data for analysis on relapses within 6 months. All
trials were analysed separately owing to the
different classes of drugs compared. Withdrawals
were assigned to the worst outcome (relapse).
The results of the Sackeim trial68 showed that
there was a non-statistically significant trend for
those treated with nortriptyline to have a reduced
risk of relapse compared with those treated with
placebo (RR = 0.73, 95% CI 0.53 to 1.01,
p = 0.06, n = 56). Those treated with
nortriptyline plus lithium had a statistically
significant reduced risk of relapse at 6 months
compared with those treated with placebo (RR =
0.58, 95% CI 0.39 to 0.86, p = 0.007, n = 57).
However, the absolute rate of relapses across the
sample was still high, with 61% of the 73
participants followed up relapsing.
In the Grunhaus trial,67 there was no statistically
significant difference in the likelihood of
experiencing a relapse in those treated with
fluoxetine combined with melatonin compared
with those treated with fluoxetine alone (RR =
067, 95% CI 0.29 to 1.52), p = 0.3, n = 40).
Adverse events Grunhaus and colleagues67 found
no significant differences between the
fluoxetine–melatonin and fluoxetine–placebo
group in cognitive functioning measured by the
MMSE or sleep quality measured by the Pittsburgh
Sleep Quality Index (PSQI). Sackeim and
colleagues68 found no statistically significant
differences in the mean number of clinically
significant side-effects per patient between the
three treatment groups (F = 0.13, p = 0.88).
Mania
The UK ECT Group51 found very little
randomised evidence regarding the effectiveness
of ECT in people with mania.
ECT versus pharmacotherapy
The UK ECT Group51 found one trial168 that
compared these treatment regimens. Forty-four
patients were initially recruited and 34 completed
the investigation, the remainder being lost to
follow-up. No efficacy scores were available. No
final cognitive testing results were reported.
Health Technology Assessment 2005; Vol. 9: No. 9
ECT plus pharmacotherapy versus
pharmacotherapy alone
The UK ECT Group51 found one trial169 that
included 30 patients who completed the trial and
had results for the mania rating scale. The
decrease in these scores was greater in the
combined therapy group by a factor of two. No
cognitive testing was reported.
Schizophrenia
Two systematic reviews evaluated the effectiveness
of ECT in schizophrenia.50,51 The results are
reproduced here.
Real versus sham ECT
Efficacy immediately after a course of ECT The UK
ECT Group51 included six trials containing 140
patients that compared sham ECT with real
ECT.170–175 All participants had been diagnosed as
having schizophrenia, apart from one small trial174
where the duration of symptoms was less than
2 months and was characterised as
schizophreniform. Two small trials171,175
predominantly included people with catatonia, but
the UK ECT Group51 reports that they were too
small to conduct reliable subgroup analyses.
The primary outcome measure used by the UK
ECT Group51 was continuous data on change in
symptoms from baseline to post-ECT treatment.
Four of these trials170,172–174 provided usable
continuous data for analysis. They found
significant heterogeneity within the trials and two
trials were not included in the analysis. The
standardised effect size of real compared with
sham ECT was –0.22 (95% CI –1.7 to 1.27) in
favour of real ECT. The result is not statistically
significant and represents a mean change in the
Brief Psychiatric Rating Scale (BPRS) score of 0.10
(95% CI 0.56 to 0.75) in favour of real ECT.
The UK ECT Group also identified two trials176,177
that compared ECT with placebo, or inpatient
care alone. A meta-analysis could not be
conducted since one trial did not report efficacy
ratings. In the other trial, the ECT group had a
3.5-point advantage on the Menninger
Health–Sickness Scale at the end of treatment.
The Cochrane Schizophrenia Group ECT review50
identified 12 trials comparing ECT with sham
ECT.105,170,172–176,178–182 They report that all but
two175,179 also used additional antipsychotic drugs
(chlorpromazine, haloperidol or trifluoperazine)
and one105 used additional chlorpromazine only
for people given sham ECT, while participants
allocated to ECT were given placebo. They also
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
identified two trials177,183 that compared ECT plus
placebo with placebo. They analysed the trials
together (562 participants, 294 treated with ECT).
The primary outcome measure of efficacy used by
the Cochrane Schizophrenia Group ECT review50
was dichotomous data on clinical global
improvement, classified as the number who had
not improved in each treatment group, as defined
by the trialist.
The Cochrane Schizophrenia Group ECT review50
reported that nine trials provided usable data for
analysis. Their analysis indicated that treatment
with ECT was significantly more likely to result in
clinical global improvement, at the end of the
course than with placebo/sham ECT (RRfixed =
0.77, CI 0.6 to 0.9; NNT 7, 95% CI 4 to 25,
n = 400), but the data were heterogeneous (2
13.46, df = 8, p = 0.097). Using a random effects
model made little difference. One trial173 was
clearly statistically outlying. Removal of this good
study resulted in a homogeneous result (eight
RCTs, RRfixed = 0.83, 95% CI 0.7 to 1.01,
n = 380). Removal of the study180 containing
people with treatment-resistant illnesses decreased
the heterogeneity (eight RCTs, RRfixed = 0.74,
95% CI 0.6 to 0.9, 2 10.97, df = 7,
p = 0.14, n = 370).
The Cochrane Schizophrenia Group ECT review50
reported one trial174 that showed that the benefit
of ECT on global improvement in the short to
medium term was equivocal (RR = 0.71, 95% CI
0.3 to 1.8, n = 30).
Other outcomes The Cochrane Schizophrenia
Group ECT review50 also explored a number of
other outcomes relating symptoms and overall
functioning, including short- and long-term
relapses, scores on the BPRS, and behaviour and
social functioning. Their results are summarised
below.
Relapses and discharge from hospital The Cochrane
Schizophrenia Group ECT review50 found that
results from two trials170,178 suggested that ECT
resulted in fewer relapses in the short term than
sham ECT (RRfixed = 0.26, 95% CI 0.03 to 2.2,
n = 47) and a greater likelihood of being
discharged from hospital (RRfixed = 0.59, 95% CI
0.34 to 1.01, n = 98),176 although the data on
which these outcomes are based are limited. There
was no evidence that this early advantage for ECT
is maintained over the medium to long term, as
assessed by other measures of symptomatic
improvement over a 6-month and 2-year follow-up
35
Effectiveness
period, although the trend favoured ECT. Again,
however, the data on which these results are based
were sparse.
Leaving the study early The UK ECT Group51
included one trial170 that reported
discontinuations and these were similar in each
group. The Cochrane Schizophrenia Group ECT
review50 found homogeneous data from the 14
trials comparing ECT with sham ECT which did
not suggest that people treated with ECT dropped
out of treatment earlier than those treated with
sham ECT (RRfixed = 0.71, 95% CI 0.33 to 1.52,
n = 495).
Efficacy at 6 months The UK ECT Group51
identified two trials170,174 that reported efficacy
scores at 6 months post-treatment, but these did
not provide sufficient data for analysis. One
trial170 indicated a 5-point greater efficacy score in
the ECT group than in the sham group, but the
other174 showed a 1.5 greater improvement in the
sham group over time. The Cochrane
Schizophrenia Group ECT review50 reported that
no data were available for the effects of ECT
versus sham ECT in the medium to long term.
Adverse events: cognitive functioning The UK ECT
Group51 did not find any data in the included
trials on cognitive functioning. The Cochrane
Schizophrenia Group ECT review50 found very
limited data from one trial184 on cognitive
functioning. This indicated that visual memory
declined after ECT compared with sham ECT
(one RCT, WMD = –14.0, 95% CI –23 to –5,
n = 24); the results of verbal memory tests were
equivocal.
Adverse effects: mortality There were no deaths in
the trials included by the UK ECT Group.51 The
Cochrane Schizophrenia Group ECT review50
identified one trial176 that reported on mortality
over a 3-year follow-up. No deaths were discovered
(n = 98).
36
ECT versus antipsychotic drugs
The UK ECT Group51 separated the analysis into
trials comparing ECT combined with
pharmacotherapy versus pharmacotherapy and
trials comparing ECT alone with pharmacotherapy.
The Cochrane Schizophrenia Group ECT review50
included at least one trial in their analysis of ECT
versus pharmacotherapy185 that had been classed
by the UK ECT Group as a combination of ECT
and antipsychotics versus ECT. It appears that the
Cochrane Schizophrenia Group ECT review50
analysed all trials that compared ECT with
antipsychotics together and completed a separate
subanalysis of ECT in combination with
antipsychotic drugs. For this analysis they included
five173,174,178,182,184 of the eight trials that
contributed data on clinical global improvement
in the comparison of ECT and sham/placebo ECT
plus antipsychotics against sham ECT plus
antipsychotics. The UK ECT Group51 included
three of these trials176,186,187 in their analysis of
ECT alone versus pharmacotherapy. Only one173
of these trials had been included in the ECT
versus sham ECT analysis in the UK ECT Group
review.51
ECT alone versus pharmacotherapy
The UK ECT Group51 included four
trials171,176,186,187 containing 163 patients. One
trial187 used sham ECT and drug placebos to blind
study participants to treatment allocation.
Treatment lasted for between 3 weeks171 and
1 year.176 Two trials176,187 provided sufficient data
for analysis. The standardised effect size of 0.26
shows a non-significant difference between the two
treatment groups (95% CI –0.92 to 1.42). This
translates as a mean change in efficacy score of 1.8
(95% CI –6.35 to 9.84) in favour of
pharmacotherapy.
ECT in combination with antipsychotics versus
pharmacotherapy, plus or minus sham
ECT/placebo
The UK ECT Group51 identified three trials
containing 147 patients.171,185,188 One trial182
included participants as young as 15 years, but
results are not reported separately for this
subgroup. Comparable doses of neuroleptic
medication were administered in both arms of
these studies, except for the Ungvari trial185 where
the pharmacotherapy group received a higher
dose of haloperidol than the ECT group. There
was a positive trend associated with treatment with
ECT and pharmacotherapy compared with ECT
alone. The standardised effect size is 0.43 (95% CI
–0.62 to 1.48) and the mean change in efficacy
score 2.04 (95% CI –2.92 to 6.96) in favour of
combined therapy.
The Cochrane Schizophrenia Group ECT review50
included eight trials105,176,177,183,185,186,188,189 that
compared ECT directly with antipsychotic drugs.
They report that four of these105,186,188,189 used
chlorpromazine as the comparator drug, Small183
compared ECT with thiothixine, May176 with
trifluoperazine and Naidoo177 used reserpine,
a drug that pre-dated chlorpromazine. Ungvari
and Petho185 compared ECT plus low-dose
haloperidol with very high-dose haloperidol, while
Health Technology Assessment 2005; Vol. 9: No. 9
Janakiramiah182 compared ECT in two groups of
people treated with low- and high-dose
chlorpromazine with two other groups given the
two strengths of the drug without ECT.
The Cochrane Schizophrenia Group ECT review50
reported that there was some variability in the
doses of antipsychotics used in these trials, as well
as in the trials of ECT versus sham ECT that used
concurrent antipsychotics. Taylor and
Fleminger173 and Brandon and colleagues178 used
doses of antipsychotics that were lower than those
used in the other trials and lower than those
currently recommended for acute-phase treatment
in people with schizophrenia.
The Cochrane Schizophrenia Group ECT review50
reported that when ECT is directly compared with
antipsychotic drug treatment, the pooled
dichotomous results strongly favour the
medication group (three RCTs, RRfixed = 2.18,
95% CI 1.3 to 3.6, n = 175). Homogeneous data
also favoured antipsychotic drugs over ECT with
regard to numbers discharged after treatment (two
RCTs, RRfixed = 1.98, 95% CI 0.97 to 4, n = 135).
The Cochrane Schizophrenia Group ECT review50
identified very limited data indicating that people
treated with ECT are less likely to relapse than
those treated with antipsychotics (one RCT,
RRfixed = 0.33, 95% CI 0.1 to 0.9, n = 32).
Continuous measures of global improvement from
one trial favoured ECT in the short term, although
the results were equivocal in the long term.
To evaluate whether the addition of ECT is
beneficial to those being treated with antipsychotic
drugs, the Cochrane Schizophrenia Group ECT
review50 analysed five of the eight trials that
contributed data on clinical global improvement
in the comparison of ECT and sham placebo ECT
plus antipsychotics against sham ECT plus
antipsychotics (see above). Their analysis of
heterogeneous data from the first five studies
results in a non-significant trend favouring the
ECT and antipsychotic combination (RRrandom =
0.74, 95% CI 0.4 to 1.3, n = 165).
Efficacy at 6 months The UK ECT Group51 found
no usable data relating to the efficacy of ECT
compared with antipsychotics at 6 months’ followup. The Cochrane Schizophrenia Group ECT
review50 found only one study176 reporting on the
long-term outcome of ECT compared with
antipsychotic, and the results were equivocal.
Discontinuations/leaving the study early The UK ECT
Group51 did not find any usable data relating to
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
discontinuations in studies comparing combined
ECT and pharmacotherapy with ECT alone. No
discontinuations occurred in studies comparing
ECT alone with pharmacotherapy. The Cochrane
Schizophrenia Group ECT review50 found no
differences in numbers leaving the study early in
the trials that compared ECT to treatment with
antipsychotics (seven RCTs, RRfixed = 0.99, 95%
CI 0.8 to 1.3, n = 419). They report that similar
numbers remained in the trial by May176 5 years
after treatment with ECT or antipsychotics,
although by this time 73% of the people in both
arms had been lost to follow-up.
Adverse effects: mortality The UK ECT Group51
found no deaths reported in any of the trials
comparing ECT with pharmacotherapy, either
alone or in combination with antipsychotic drugs.
The Cochrane Schizophrenia Group ECT review50
found that one patient who had not received ECT
died within the 3-year follow-up by May176 (one
RCT, RR = 0.63, 95% CI 0.03 to 15, n = 149).
Adverse affects: cognitive functioning The UK ECT
Group51 identified one randomised trial190 that
compared cognitive functioning of patients who
had received ECT with those who had received
chlorpromazine at the end of a course of ECT.
Only data on retrograde memory were identified
and the trial reported no difference on several
measures of retrograde memory between patients
treated with ECT and those treated with
chlorpromazine.
ECT versus psychotherapy
The Cochrane Schizophrenia Group ECT review50
reported limited data from one study comparing
ECT alone with individual psychoanalytic
psychotherapy alone,176 showing a consistent,
although non-significant, trend favouring ECT
(both short term and 2 years later) on several
outcomes. When antipsychotics were added to
psychoanalytic psychotherapy, however, a
significant advantage of the drug group over ECT
was seen in the short term (WMD = –5.0, 95% CI
–0.54 to –9.46, n = 90), with a continuing trend
2 years later.
Unilateral versus bilateral ECT
The UK ECT Group51 included two trials191,192
containing 147 patients. Adolescents were
included in one trial191 and unmodified ECT was
used in the other.192 Neuroleptic medication was
not coadministered in either of these studies. The
Cochrane Schizophrenia Group ECT review50
identified an additional trial184 that involved trial
arms that compared unilateral with bilateral ECT
37
Effectiveness
for people who had also been given concurrent
haloperidol.
Efficacy The UK ECT Group51 found that the
standardised effect size between the two electrode
placements was 0.03 (95% CI –0.91 to 1.03), with
a mean change in efficacy score of 0.32 (95% CI
–10.56 to 11.99), indicating no difference between
the two electrode placements. The Cochrane
Schizophrenia Group ECT review50 found neither
unilateral nor bilateral ECT to be superior in
terms of global improvement (two RCTs,184,191 RR
= 0.79, not improved at end of course of ECT,
95% CI 0.5 to 1.4, n = 78). They report that none
of the three trials reported long-term efficacy data.
Discontinuations/leaving the study early None of the
three studies reported data on discontinuations.
Adverse events: mortality No deaths were reported
in the trials.
Adverse events: cognitive functioning The UK ECT
Group51 reports that one randomised trial192
measured cognitive functioning at the end of a
course of ECT, comparing patients treated with
unilateral ECT with those treated with bilateral
ECT. The trial reported no difference on learning
tasks between patients treated with unilateral ECT
and those treated with bilateral ECT. Four patients
treated with bilateral ECT complained of
subjective forgetfulness compared with one of
those treated with unilateral ECT.
Unilateral placement
The Cochrane Schizophrenia Group ECT review50
identified one trial191 that compared the effect of
dominant and non-dominant electrode
placements on schizophrenic patients. BPRS
scores were available for pretreatment and posttreatment. The change in scores was greatest in
the non-dominant group by more than two points.
No deaths were reported in this trial.
Dose of ECT
Both reviews50,51 identified one trial193 of 67
participants. In this study, people with treatmentresistant schizophrenia were administered variable
numbers of ECT at stimulus intensities just above
the seizure threshold (T) twice the seizure
threshold (2T) or four times the threshold (4T).
End-point average scores for global impression
(GAF), mental state (BPRS) and cognitive function
(MMSE) were not extractable.
38
Efficacy The Cochrane Schizophrenia Group ECT
review50 reported that the three stimulus doses did
not differ in numbers improved at the end of the
course of ECT (~50% in each group). In the
subgroup of people given ECT who met criteria
for remission (n = 22; 34% of sample), those given
ECT at twice the threshold required fewer doses of
ECT to attain remission than those given
threshold doses (WMD = 6.1, 95% CI 2.4 to 10).
Similarly, those given 4T required fewer
treatments than those treated at threshold doses
(WMD = 9.4, 95% CI 6.3 to 12.5). Treatment at
4T was non-significantly superior to treatment at
2T in reducing the number of treatments required
to achieve remission (WMD = 3.23, 95% CI 0.8 to
5.6). Similarly, those treated at 2T and 4T
required fewer days to attain remission than those
given threshold stimuli, but those treated at 4T
required on average fewer days of treatment than
those given ECT at 2T (WMD = 9.4, 95% CI 2.1
to 16.8).
Leaving the study early The Cochrane
Schizophrenia Group ECT review50 reported that
only five out of 67 people left this study before
completion, with no clear trend favouring any one
group.
Adverse events: cognitive functioning The UK ECT
Group51 reported that there were no significant
differences between the groups on scores on the
MMSE at the end of a course of ECT.
Frequency of administration
Both reviews50,51 identified only one study194
comparing unilateral ECT given three versus five
times a week, which included only ten participants.
This trial had usable data for cognitive
functioning only. Average end-point scores on the
MMSE indicated no significant advantage for the
less frequent treatments, and no one developed
clinical evidence of cognitive impairment.
Number of ECT treatments
The Cochrane Schizophrenia Group ECT review50
reported limited data from one trial195 that
showed a significant advantage for 20 treatments
over 12 treatments in numbers globally improved
at the end of the ECT course (RRfixed = 2.53, 95%
CI 1.1 to 5.7, n = 43). No one was taking
concurrent antipsychotics. This trial was excluded
by the UK ECT Group.51
Continuation ECT
Both reviews50,51 identified one trial196 that
compared continuation ECT alone with
antipsychotics, with continuation ECT added to
antipsychotics, for people with treatment-resistant
schizophrenia.
Health Technology Assessment 2005; Vol. 9: No. 9
Efficacy The Cochrane Schizophrenia Group ECT
review50 reported that when continuation ECT was
compared with antipsychotics at the end of the
6-month trial, results for overall functioning as
measured on the GAF scale were equivocal (one
RCT, WMD = –1.24, 95% CI –6.4 to 3.9, n = 30).
However, when continuation ECT was added to
antipsychotic drugs, the combination was clearly
superior to the use of antipsychotics alone (WMD
= 19.1, 95% CI 9.7 to 28.5, n = 30) or
continuation ECT alone (WMD = –20.3, 95% CI
–11.5 to –29.1, n = 30). Similarly, at 6 months,
continuation ECT was no better than treatment
with antipsychotic drugs in reducing BPRS scores,
although the combination of continuation ECT
and antipsychotics was superior to continuation
ECT alone (WMD = 18.6, 95% CI 8.6 to 27.6,
n = 30) or antipsychotics alone (WMD = –19.8,
95% CI –10.3 to 29.2, n = 30).
Relapses The Cochrane Schizophrenia Group ECT
review50 reported that equal numbers (14 and 15)
of people on continuation ECT alone or
antipsychotics alone relapsed over the 6-month
trial period. The addition of continuation ECT to
antipsychotic drugs, however, was clearly beneficial
in reducing relapses compared with antipsychotics
alone or continuation ECT alone (RRfixed = 0.43,
95% CI 0.23 to 0.81, n = 30, NNT = 2, 95% CI
1.5 to 2.5).
Leaving the study early The Cochrane
Schizophrenia Group ECT review50 reported that
few people (six out of 45) left the study early, with
no clear pattern emerging to suggest a trend in
favour of any of the three comparisons.
Adverse effects: mortality No death occurred in this
trial.
Adverse effects: cognitive functioning The Cochrane
Schizophrenia Group ECT review50 reported that
no significant differences were seen in cognitive
impairment scores between those treated for
6 months with continuation ECT or antipsychotics.
Continuation ECT added to antipsychotics
resulted in non-significant trends favouring
antipsychotic drugs used alone and the
combination versus continuation ECT used alone.
Specific outcomes not covered by the
randomised evidence
The randomised evidence reviewed by the UK
ECT Group,51 the Cochrane Schizophrenia Group
ECT review50 and the current authors did not
address two key areas of outcome: (1) long-term
adverse effects of ECT including suicide, all-cause
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
mortality and brain damage, and (2) consumers’
views and experiences of ECT and whether these
experiences influenced the outcomes of ECT.
Therefore, sources that reviewed the nonrandomised evidence for these outcomes were
identified.
Severe adverse events
The UK ECT Group51 included cohort studies and
case–control studies that compared people with
depression, schizophrenia and mania who had
received ECT at some point during their care with
those who had not. The reviewers examined
evidence on five key outcomes: all-cause mortality,
suicide, cerebral haemorrhage, functional
impairment and structural brain damage. Their
findings are set out below.
The search strategy used by the UK ECT Group51
to locate non-randomised studies could not be
comprehensive before 1966 owing to limitations in
time and resources. For earlier studies, they used
the review of ECT and mortality (particularly
suicide risk) by Prudic and Sackeim.197 These early
studies provided the main evidence that ECT
reduces mortality. Prudic and Sackeim described
six studies comparing the suicide rates in the preECT and ECT eras. The results were variable, with
four studies reporting some evidence of reduced
suicide and mortality rates following the
introduction of ECT. They also identified six
studies comparing suicide rates in the era before
the introduction of psychotropic drugs (ECT
alone) with those after the introduction of drugs.
Four of these studies claimed that the rate
increased following the introduction of drugs. As
identified by Prudic and Sackeim,197 all these
historical comparison studies are methodologically
unreliable because of the lack of control for other
confounders between the cohorts. Their results are
reproduced below.
All-cause mortality
The UK ECT Group51 reports that all the studies
described in the review of severe adverse events
suffer from the major methodological shortcoming
of patient selection. For example, ECT may not
have been used for medically ill patients, which may
explain any observed lower mortality. Conversely,
patients selected for ECT may have been very
severely ill or suicidal, or both, and therefore any
failure to find a difference may be because ECT
has reduced suicide in a high-risk group.
Five non-randomised cohort studies compared
mortality rates in patients contemporaneously
treated with ECT with those not treated with ECT.
39
Effectiveness
Babigian and Guttmacher198 compared mortality
rates in depressed patients receiving treatment
with ECT during their first hospitalisation with
patients who did not receive ECT. All-cause
mortality at up to the 20-year follow-up was
significantly lower in the ECT-treated group; this
difference remained following age standardisation.
199
Avery and Winokur reported a 3-year follow-up
of 519 consecutively admitted patients with
depression. In the ECT-treated group, the
mortality rates were 0.7% at 1 year and 2.2% at
3 years. In the groups with adequate treatment
with antidepressant drugs, the corresponding
figures were 1.4% and 2.8%. In patients who were
treated with neither drugs nor ECT, the mortality
rates were 10% and 11.4%.
Tsaung and colleagues200 followed up 74 (out of
85 consecutive admissions) patients with a
diagnosis of schizoaffective disorder. Seventeen
(34%) of the patients treated with ECT were
deceased at follow-up compared with two (8%) of
the patients who did not receive ECT.
Black and colleagues201 reported a follow-up of
1076 patients with primary affective disorders,
carefully controlling for medical co-morbidity and
length of follow-up, and attempting to control for
other important confounders. They found no
differences in total mortality between patients
treated with ECT, antidepressants or no adequate
treatment.
Philibert and colleagues77 reported a follow-up of
192 patients aged over 65 years and compared
mortality in those treated with ECT with those
treated who did not receive ECT. ECT-treated
patients had lower mortality rates than those who
only received drug therapy. Adjustment for
confounding was attempted, but was incomplete.
Cause-specific mortality
Suicide Tsaung and colleagues200 followed up 74
(out of 85 consecutive admissions) patients with a
diagnosis of schizoaffective disorder. None of the
patients treated with ECT killed themselves,
compared with three of the patients who did not
receive ECT.
Avery and Winokur199 reported a 3-year follow-up
of 519 consecutively admitted patients with
depression. There were eight deaths due to
suicide, too few to allow meaningful comparisons.
40
Babigian and Guttmacher198 compared mortality
rates in depressed patients receiving treatment
with ECT during their first hospitalisation with
patients who did not receive ECT. Cardiovascular
deaths (RR = 0.75) and accidental deaths (RR =
0.18) were significantly lower in the ECT-treated
group. There were no differences in the suicide
rates, but numbers were small.
Black and colleagues201 reported a follow-up of
1076 patients with primary affective disorders,
carefully controlling for medical co-morbidity and
length of follow-up, and attempting to control for
other important confounders. They found no
differences in suicide rates between patients
treated with ECT, antidepressants or no adequate
treatment, but the numbers were very small and
the study had very limited power to detect a
moderate, but important effect.
The UK ECT group51 found two case–control
studies202,203 comparing the rates of use of ECT in
patients who committed suicide with control
patients who did not commit suicide.
Sharma202 compared the use of ECT in 45
inpatients who committed suicide with a matched
group of inpatients who did not commit suicide.
Eight patients in the ECT group killed themselves
compared with four in the control group.
Brådvik and Berglund203 compared the last
treatment received by 89 patients with severe
depression admitted to Lund Hospital in Sweden
between 1956 and 1969 who committed suicide by
1984 with a matched control group who did not
commit suicide. There was no difference in the
rates of ECT between the two groups.
No studies were identified that examined cerebral
haemorrhage or functional impairment.
Brain scanning and ECT
Computerised tomography Nasrallah and
colleagues204 measured ventricular:brain ratios
(VBRs) with X-ray computed tomography (CT) for
young patients with mania compared with agematched controls. A high VBR reflects loss of
brain mass. Patients had a higher VBR than
controls and the effect was not associated with a
history of exposure to ECT.
Kolbeinsson and colleagues205 compared VBR and
cortical atrophy (another measure that reflects loss
of brain mass) measured with X-ray CT for agematched unipolar and bipolar patients with or
without a history of exposure to ECT. Both patient
groups showed increased VBR and cortical
atrophy compared with controls, with a trend
Health Technology Assessment 2005; Vol. 9: No. 9
towards a larger effect in the ECT group.
However, there was no confirmatory correlation
between lifetime ECT exposure and VBR. Instead,
the measures were strongly correlated with age
within all groups. This is a common finding in
groups with an average age over 40 years.
Calloway and colleagues,206 in a technically similar
X-ray CT study, compared much older patients
with and without previous ECT using judgements
or regional atrophy. They also found more
evidence of atrophy in the ECT group, specifically
in the frontal areas, but again there was no
correlation with the total number of ECT
treatments.
Magnetic resonance imaging Magnetic resonance
imaging (MRI) is a more sensitive neuroimaging
technique than CT. Hickie and colleagues207
investigated a cohort of elderly patients. They
measured the extent of a pathology not specific
for depression but associated with it: the density of
subcortical hyperintensities. They showed a strong
association between age and the severity of these
lesions in white matter, but no association with
previous ECT emerged as being significant from
their analysis.
Experimental investigation of acute effects of
ECT on brain images
Several studies used within-subject measures
before and after ECT. Mander and colleagues208
measured the T1 relaxation time in patients
receiving ECT with MRI. The T1 reflects tissue
fluidity or the extent to which the protons in water
are free to distribute in a magnetic field. It was
anticipated that it might measure changes relevant
to the permeability of the blood–brain barrier
(BBB). Patients were studied before and at varying
times after ECT. There was a small global increase
in T1 with a peak at around 6 hours, after which
measures returned to normal. Controls receiving
anaesthesia for other reasons showed falls rather
than increases in T1. The results were interpreted
in relation to other work suggesting that transient
reductions in the BBB are probably secondary to
the increased systemic blood pressure observed
during seizures.
Ende and colleagues209 studied patients during
ECT and measured N-acetylaspartate (NAA) in the
hippocampus. Brain injury would predict reduced
NAA levels and evidence of effects on episodic
memory would localise the most likely target as
the hippocampus. This investigation therefore
tested a specific hypothesis about the integrity of
cellular elements in this key structure. No change
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
was observed in NAA during ECT, although the
sensitivity of the method was confirmed by the
detection of an unexplained rise in cholinecontaining compounds.
Patient acceptability and choice
The reviewers identified one good quality
systematic review of non-randomised evidence
relating to users’ views and experiences of ECT,
conducted by SURE at the Institute of Psychiatry.53
The results of this review are summarised below.
Persistent memory loss
Twenty studies made reference to long-term
memory loss in the abstract, but six did not report
memory loss at 6 months and seven did not
provide raw numbers of the percentage or number
of people experiencing the side-effect. Only seven
papers210–216 provided usable information on
long-term memory loss.
There was no difference in the rates of people
reporting persistent memory loss at 6 months
between clinical studies and those carried out in
collaboration with patients. As a lower limit, at
least 28.1% of patients experience persistent
memory loss as a result of ECT. It is difficult to
differentiate memory loss caused by ECT, memory
loss due to depression and the maintenance of
depression as a result of memory loss.
The testimonies revealed that the types of memory
loss that are important to people who receive ECT,
such as autobiographical memories, are not those
captured in neuropsychological tests used in RCTs.
The reviewers suggested that this may explain the
different conclusions between patient reports,
where memory loss is a key issue, and RCTs, where
only a significant minority of people are reported
as having persistent memory loss. The testimonies
also revealed complex and important emotional
reactions to memory loss following ECT.
Information and consent
Sixteen papers included information on
information or consent, but only 12211,213,215,216–224
provided usable data and only one study asked
whether people had been told about the risks of
ECT. Four studies219,220,225,226 included
information on objective knowledge of ECT.
There were no important differences between
clinical studies and those carried out in
collaboration with patients in the rates of people
reporting that they had received adequate
information before receiving ECT. At least 50% of
users felt that they had been given inadequate
41
Effectiveness
information before receiving ECT. Between 7 and
16% were judged to have full knowledge that ECT
involved an anaesthetic, the passing of an electric
current and the induction of a convulsion. The
testimonies revealed that side-effects such as
memory loss were the main area where patients
felt that they had not received sufficient
information.
Felt compulsion
Seven studies asked about felt
compulsion.211,213,215,216,219,225,227 There were no
important differences between clinical studies and
those carried out in collaboration with patients in
the rates of people reporting that they had felt
they had no choice but to have ECT. Between onequarter and one-third of people who sign a
consent form for ECT do so under pressure or in
the belief that they cannot refuse.
Perceived benefit
Sixteen research studies211,213,215,217,220–223,225–232
asked about the perceived benefit of ECT. These
included two main types of questions: perceived
helpfulness and whether the user would agree to
ECT again.
Fewer respondents from patient-led surveys
reported feeling that ECT had helped or that they
would have it again compared with the
respondents in clinically-led research.
Methodological variables such as the interval since
ECT, the setting in which the views were elicited,
and the number and complexity of questions used
to measure perceived benefit all had important
influences on perceived benefit of ECT within
both clinical and patient-led surveys. The
reviewers concluded that studies that interview
patients immediately after ECT are more likely to
overestimate the degree of perceived benefit of
ECT, especially if the interview is conducted within
a hospital setting by a clinician using brief
interviews.
42
The review also explored the relationship between
legal compulsion to have ECT and satisfaction
with ECT. One clinical study reported no
difference in satisfaction between patients who
were legally compelled to have ECT and those
who consented.222 However, this study was based
on a small sample of people whose legal status
with respect to ECT was very different from the
national average. Patient-led studies included
more representative samples of patients who were
legally compelled to have ECT. One of these
studies reported a negative association between
legal compulsion and satisfaction,216 and the other
did not state whether there was an association
between legal compulsion and satisfaction. None
of the studies analysed the relationship between
legal compulsion and the perceived benefit of
ECT.
The testimonies revealed that the perceived
benefit of ECT from the patient’s perspective was
much more complex and divergent from clinical
conceptualisations of benefit that underlie the
construction of symptom scales used in RCTs.
Many of the issues raised by patients, such as lying
about the success of treatment in order to avoid
further ECT and wishing to take legal action
against clinicians, are not addressed in clinical
research. Patients’ views were heterogeneous, with
some reporting that it was a life-saving treatment
and others feeling violated and not helped by the
treatment. Furthermore, patients made trade-offs
between the side-effects and benefits of ECT, and
for some, the way in which ECT was given was a
more important issue than whether or not the
treatment had helped. Thus, the review concluded
that there is no single, unidimensional patient
voice regarding the perceived benefit of ECT, but
those opposed to ECT cannot be seen as a small
vocal minority.
Interventions to improve patient knowledge
about ECT
In addition, two RCTs69,92 that assessed the impact
of a video on knowledge about ECT were
identified in the present review. One of these
trials92 was included in the SURE review. A pooled
analysis of knowledge scores in the two trials
revealed significant statistical heterogeneity and
the results are therefore reported separately.
In the trial by Westreich and colleagues,69
participants were psychiatric inpatients who had
received ECT in the past and the intervention was
delivered during the consent procedure for a
further treatment of ECT. One group was
randomised to watch a video (n = 11) in addition
to receiving a written consent form, while the
other group received the written consent form
only (n = 7). Postconsent knowledge was assessed
using an instrument with no assessment of its
psychometric properties. There was no statistically
significant difference between the two groups in
the mean number of items answered correctly
(WMD = –0.81, 95% CI –1.86 to 0.24, p = 0.13,
n = 18).
In the trial, by Battersby and colleagues,92 the
intervention was delivered to a group of
psychiatric inpatients who were not about to have
Health Technology Assessment 2005; Vol. 9: No. 9
ECT and it was not clear how many had
personally experienced ECT in the past. One
group was randomised to watch the video
(n = 40), while the other group did not watch the
video (n = 40). Knowledge was assessed before
and after the video using an instrument with
limited assessment of its psychometric properties.
There was no statistically significant difference
between the two groups in the mean knowledge
score after watching (or not watching) the video
(WMD = 1.28, 95% CI –2.3 to 2.79, p = 0.1,
n = 69).
Efficacy of ECT in specific subgroups
The UK ECT Group51 reported that data were too
limited to undertake reliable subgroup analyses.
The Cochrane Schizophrenia Group ECT review50
reports the following subgroup analyses in their
review of ECT in schizophrenia.
Diagnostic criteria
When studies that used diagnostic criteria to
diagnose schizophrenia were evaluated separately,
a modest but non-significant advantage of ECT
over sham ECT in the numbers improved at the
end of the course of treatment was maintained
from heterogeneous data from five trials (RRrandom
= 0.72, 95% CI 0.4 to 1.3, n = 165). A significant
advantage for ECT for this outcome was more
evident when the three trials that did not use
operational definitions of schizophrenia176,177,179
were analysed separately (RRfixed = 0.74, 95% CI
0.6 to 0.98, n = 205). The degree of overlap in
the confidence intervals of these comparisons,
however, indicates that the rigour with which the
diagnosis of schizophrenia was made did not
significantly affect the outcome with ECT.
Duration of illness
The Cochrane Schizophrenia Group ECT review50
acknowledges that the power of the review to
detect a differential response to ECT for those
with a short duration of illness (<2 years) as
opposed to those with chronic schizophrenia was
very limited. Six trials restricted inclusion to
participants with durations of illness less than
2 years.170,172,174,181,191,194 Two of these170,172
provided the data used in the comparison of
mental state assessment. This demonstrated a
significant advantage for an ECT/antipsychotic
drug combination over sham ECT and
antipsychotics in both the rate of clinical
improvement and the degree of improvement at
the end of the course and in the short term. The
participants in the trial by Sarkar and
colleagues174 were acutely ill, with onset of
symptoms less than 2 months before the start of
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
treatment. This trial found that the combination
of ECT and antipsychotics provided no additional
benefit to treatment with antipsychotics (and sham
ECT) in terms of the numbers improved at the
end of the course of ECT, or in the short to
medium term. The trials by Brill and co-authors179
and Miller175 included people with chronic
schizophrenia. ECT alone did not result in greater
clinical improvement than sham ECT by the end
of treatment in these trials. Chanpattana and
colleagues193,196 included participants who had
been ill for between 3 and 30 years, and duration
of illness did not significantly alter outcome. The
remainder of the selected trials were
heterogeneous for illness duration, thus
preventing their inclusion in the evaluation of the
effect of this variable on ECT response.
Catatonia
The Cochrane Schizophrenia Group ECT review50
found that ECT did not have significant beneficial
effects in people with chronic catatonic
schizophrenia, who comprised the participants in
the trial by Miller,175 although this finding could
equally be attributed to chronicity rather than the
subtype of schizophrenia. However, they found
that ECT did result in significant clinical
improvement by the end of the course for those
people diagnosed as having paranoid
schizophrenia in the study by Taylor and
Fleminger173 (RRfixed = 0.74, 95% CI 0.6 to 0.91,
n = 20). It was not possible to separate the
influence of the duration of illness from the
symptom profile of the participants in the selected
trials to assess whether ECT has differential effects
on positive or negative symptoms. The trials that
favoured ECT170,173,176,178,189,193,196 reported a
beneficial effect on positive symptoms. These trials
included participants with varying durations of
illness. The trial by Chanpattana and colleagues196
on people with treatment-resistant schizophrenia
provided data on symptom clusters on BPRS, in
those responding to ECT before randomisation to
continuation treatments. These data indicate
significant reductions in positive and negative
symptoms, as well as depressive and aggressive
symptoms.
The present analysis also identified one review78 of
270 treatment episodes in 178 cases treated for
catatonia. Of these cases, 55 episodes involved the
use of ECT and five involved the use of ECT in
combination with another drug. In the 55
episodes, 47 (85%) resulted in a complete
resolution of symptoms in response to ECT, 73 out
of 104 (70%) episodes involving treatment with
benzodiazepines (70%) had a complete resolution,
43
Effectiveness
TABLE 3 Summary of efficacy of ECT in children and adolescents70,71
Diagnosis
Depression total
Major depression
Psychotic depression
Manic episode
Bipolar disorder
Schizophrenia
Schizoaffective disorder
Catatonia
Responders immediately
post-ECT71
n
N
%
n
N
%
58
33
25
22
54
17
4
21
87
52
35
28
70
41
6
29
67
64
71
79
71
42
67
72
13
11
2
8
17
1
–
1
18
14
4
10
24
10
–
2
72
79
50
80
71
10
–
50
57 out of 72 (79%) treatment episodes
demonstrated a complete resolution in response to
lorazepam and three out of 40 (7.5%) had a
complete response to antipsychotics.
Since publication of this review78 in 1995, two
prospective case series studies79,80 have reported
on eight cases who failed to respond to lorazepam
and who were subsequently treated with ECT with
varying lengths of treatment. One study did not
provide details of ECT electrode placement,79
while the other used bilateral ECT.80 Both studies
used the Bush–Francis Catatonia Rating Scale
(BFCRs) to evaluate outcomes. In Bush,79 four out
of five cases offered ECT showed a remission of
symptoms, while in Malur80 two out of three cases
showed a full remission of symptoms. No data on
adverse effects were recorded.
Children and adolescents
The authors identified two systematic reviews of
non-randomised evidence70,71 and one
case–control study72 published since the review
evaluating the efficacy of ECT in children and
adolescents. The cases included in the 1999 review
had the following diagnoses: major depression
(n = 52), psychotic depression (n = 35), manic
depression (n = 28), schizophrenia (n = 41),
schizoaffective disorder (n = 6), catatonia
(n = 29), neuroleptic malignant syndrome (n = 4)
and other disorders (n = 29).
44
Responders 6 months
post-ECT70
Information on prior treatment was available for
57 patients: 20 had previously received a course of
both antipsychotic and antidepressants, five had
received antidepressants alone and 15 had
received antipsychotics alone. Information on
gender was provided in 118 cases and 55 (47%)
were female. Information on age was provided in
98 cases; the mean was 15.4 years and the
youngest patient was 7 years old.
Information on electrode placement in the
systematic review was provided for 61 patients:
23 (38%) had unilateral ECT, 29 (48%) had
bilateral ECT and nine (15%) had both.
Information on the number of ECTs administered
was available for 95 patients and the mean was
9.6 with a range of 1–23. Thirty-eight patients
had received electroencephalographic monitoring
and no studies mentioned the use of stimulus
dosing.
Efficacy
The systematic review presents data comparing
the relative efficacy of ECT immediately post-ECT
and at 6 months’ follow-up in adolescents with
different diagnoses (Table 3), although no
information is given regarding whether this
analysis is on an ITT basis. It is therefore difficult
to draw reliable conclusions from the review,
although the results suggest that ECT is more
effective in adolescents with depression, mania
and catatonia than in schizophrenia.
In the case–control study,72 all participants
receiving ECT showed recovery immediately after
ECT, although six had relapsed by the time of
follow-up (mean 5.2 years).
Adverse events: mortality
The 1997 review by Rey and Walter70 included all
396 cases in their analysis of adverse events. They
identified no deaths in adolescents with
depression, schizophrenia, catatonia or mania who
received ECT. One death occurred in a case with
NMS due to cardiac failure. One person from the
case–control72 study had committed suicide since
receiving ECT.
Adverse events: post-ECT seizures
The review70 reported post-ECT seizures in 15
cases.
Health Technology Assessment 2005; Vol. 9: No. 9
Adverse effects: cognitive functioning
The review70 found few studies that assessed
cognitive functioning systematically, as children
were “too sick” to undergo psychometric testing.
Those studies that did formally assess cognitive
functioning after ECT were conducted in the
1940s and 1950s, when the techniques used to
administer ECT are not generalisable to current
practice and results were not reported
systematically. Cohen and colleagues72 found no
significant differences on the MMSE, the WMS
and the California Verbal Learning Test at a mean
5.2 years’ follow-up.
Adverse effects: subjective side-effects
The review70 found that, overall, the most
common complaint was headaches, reported in 16
out of 396 cases. Subjective memory loss was
described by nine patients, manic symptoms in
seven, disinhibition in two and hemifacial flushing
in one. The review found that more recent studies
reported a higher percentage of side-effects. One
study reported mild side-effects in seven out of
nine (78%) of patients, while another reported
headaches in the entire group (n = 11). Another
study included in the review reported mild,
transient side-effects following 28% of ECTs,
including headache (15%), confusion (5%),
agitation (3%), hypomanic symptoms (2%),
subjective memory loss (2%) and vomiting (1%).
Cohen and colleagues72 found that six patients
who received ECT reported having subjective
memory impairment.
Older people
There was no randomised evidence of the efficacy
of ECT in people older than 65 years. In
searching for non-randomised evidence the
reviewers limited the inclusion criteria to studies
whose populations were all aged 65 or over. One
prospective73,74 and three retrospective
case–control studies75–77 were identified that
compared older people who had been treated with
ECT and those who had not.
scores (F = 3.56, df 6,65, p = 0.004, r2 = 0.25)
and that the other covariates, with the exception
of baseline GDS scores, did not. A similar result
was obtained for the BDI scores at discharge.
Admission and discharge scores on the GDS were
not statistically significantly different between the
two groups. When changes in scores on the GDS
from baseline to discharge were analysed, those
treated with ECT (mean 10.8, SD 7.5) showed a
statistically significantly greater improvement
(p = 0.002) than those who did not receive ECT
(mean 4.2, SD 6). A similar result was also
obtained for change in BDI scores. Finally, 36 out
of 46 patients (75%) treated with ECT showed
major improvement over baseline level as rated by
a physician compared with 23 out of 55 (42%) who
did not receive ECT.
Philibert and colleagues77 compared physicianrated global improvement at discharge between
those who had received ECT and those who had
not. In the ECT group 43 out of 108 (40%) made
complete recovery, 60 out of 108 (56%) had
improved and five out of 108 (5%) had not
improved. In the non-ECT group, 16 out of 84
(19%) had made a complete recovery, 56 out of 84
(66%) had improved and 12 out of 84 (14%) had
not improved. The differences in the numbers
who completely recovered were statistically
significant (p < 0.05).
Manly and colleagues75 also compared physicianrated outcome, although it is not clear when this
outcome was measured. In the ECT group, 30 out
of 39 (77%) had a good outcome compared with
13 out of 39 (33%) in the pharmacotherapy group
(p = 0.001). In the ECT group, nine out of 39
(23%) had a moderate outcome compared with 22
out of 39 (56%) in the pharmacotherapy group
(p = 0.003). None of the ECT group had a poor
outcome, while four out of 39 in the
pharmacotherapy group had a good outcome
(p = 0.06).
Improvement at end of course of ECT
Three studies provided information on symptom
improvement following treatment with ECT
compared with pharmacotherapy.74,75,77
However, physician- or patient-rated outcomes
were not made blind to treatment in any of the
studies and results must be interpreted with
caution. In two studies some effort was made to
control for confounding variables.
Rubin and colleagues73 conducted an analysis of
covariance using Geriatric Depression Scale (GDS)
scores at discharge from hospital as the dependent
variable and ECT, gender, psychotic symptoms,
cognitive dysfunction and baseline GDS scores as
covariates, and found that the presence or absence
of ECT had a statistically significant effect on GDS
Relapses and rehospitalisation
One study76 provided data on relapses and
rehospitalisation. At follow-up, 29 out of 37 (78%)
in the ECT group had a recurrence, compared
with eight out of 28 (29%) in the non-ECT group,
and 17 out of 37 (46%) in the ECT group were
rehospitalised, compared with four out of 28
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
45
Effectiveness
(14%) in the non-ECT group. Following treatment,
19 out of 37 (51%) in the ECT group were in a
nursing home compared with 13 out of 28 (46%)
in the non-ECT group. The statistical significance
of these differences was not reported.
Adverse effects: mortality and survival
Two studies76,77 provided data on mortality and
survival and reported conflicting results. Kroessler
and Fogel76 followed up 65 participants for 3
years, of whom 37 had received ECT. They found
that 27 out of 37 (73%) in the ECT group were
alive at 1 year, compared with 27 out of 28 (96%)
in the non-ECT group, and eight out of 37 (22%)
in the ECT group were alive at the end-point of
the study, compared with 17 out of 28 (61%) for
the non-ECT group. In terms of mortality, ten out
of 37 (27%) in the ECT group were dead at 1 year
compared with one out of 28 (4%) in the non-ECT
group. At 3 years’ follow-up, 18 out of 37 (49%) in
the ECT group were dead at 3 years, compared
with nine out of 27 (33%) in the non-ECT group.
The statistical significance of these differences was
not reported. In contrast, Philibert and
colleagues77 reported that those who received ECT
at some point during their care in hospital were
statistically significantly more likely to be alive at
follow-up than those who received
pharmacotherapy, with only 45 out of 84 (53%) in
the non-ECT group and 68 out of 108 (63%) in
the ECT group alive at follow-up (p < 0.05).
However, in the Kroessler study,76 participants who
received ECT were medically and mentally more
ill than those who did not receive ECT. In the
Philibert study,77 the ECT group was more likely
to be judged as suffering from psychomotor
retardation and to have had a prior course of ECT
than the pharmacotherapy group.
Adverse effects: other
Two studies74,75 reported data on a range of
adverse effects following ECT. Manly and
colleagues75 compared a number and types of
complications reported in case notes between
those who had received ECT (n = 39) and those
who had not (n = 39), including cardiovascular
disease (CVD), confusion/neurological symptoms,
gastrointestinal, pulmonary and metabolic
complications, and falls. The pharmacotherapy
group experienced statistically significantly more
CVD (p = 0.013) and gastrointestinal
complications (p = 0.027), but there were no other
differences between the two groups.
46
Rubin and colleagues74 reported MMSE scores at
admission and discharge for groups who did or
did not receive ECT, but results were not on an
ITT basis. The results indicate similar scores
between the two groups.
The use of ECT in pregnancy
The authors identified one review81 of case reports
and case series on the use of ECT during
pregnancy and three further studies82–84 reporting
on four cases published since the review. In two
cases ECT was administered during the third
trimester, in one case during the second trimester
and in one case during the first trimester. The
review81 identified reports of 300 cases of the use
of ECT during pregnancy, published between
1942 and 1991. Of these cases, 14 (4.7%) used
ECT during the first trimester, in 36 (12%) cases
the use of ECT began in the second trimester and
in 31 (10.3%) in the third. In the remaining 219
(73%) of cases, the timing of ECT with respect to
stage of pregnancy was not reported. In 44 cases
(14.7%) unmodified ECT was used and 21 (7%)
reported that modified ECT was used. In the
remaining 235 cases (78%) the method of ECT
was not reported. The number of ECTs per
patient ranged from one to 35. In 89 cases (30%)
there was some follow-up of offspring after birth,
with the length of follow-up ranging from 2 months
to 19 years.
Efficacy
The review81 provides no information on the
efficacy of ECT during pregnancy. In the three out
of four of the cases82,83 reported subsequently,
improvement in symptoms as judged by clinical
opinion was observed, which was still evident at
1-year follow-up. All of the women gave birth to
healthy babies. In the remaining case,84 no clinical
improvement was observed and no information is
provided regarding the health of the baby.
Adverse effects
The review81 provides details of the prevalence of
complications when ECT was used during
pregnancy. Complications were noted in 28 cases
(9.3%) and these are summarised below.
Foetal cardiac arrhythmia Five cases reported
transient self-limiting disturbances in foetal
cardiac rhythm including irregular foetal heart
rate postictally (three cases), foetal bradycardia
during the tonic phase (one case) or postictally,
and reduced variability of foetal heart rate (one
case). In all cases the babies were born healthy.
Vaginal bleeding Five cases of known or suspected
vaginal bleeding related to ECT were reported. In
one case the bleeding was the result of mild
Health Technology Assessment 2005; Vol. 9: No. 9
abruptio placentae, but in the other four cases the
source of bleeding was not identified. No adverse
effects on the babies were reported in any of these
cases. In the subsequent studies,83 one case of
vaginal bleeding was reported, which led to
miscarriage (see below).
premature labour occurred 6 days post-ECT, which
subsided following hydration and ritodrine
hydrochloride tocolytic therapy. In the other
case,84 premature labour occurred immediately
after the first ECT and was treated successfully
with indomethacin and ritodrine.
Uterine contractions In two cases uterine
contractions began shortly after ECT, but neither
resulted in premature labour. In the subsequent
reports,82 in one case uterine contractions were
reported following the second, third and sixth
ECT treatments. Contractions following the
second and sixth treatments were self-limiting,
whereas those following the third treatment
required tocolytic therapy. In another case,
premature labour was reported on day 6 post-ECT,
which subsided following hydration and ritodrine
hydrochloride tocolytic therapy.
Miscarriage Five cases of miscarriage were
reported. In subsequent reports, one case of
miscarriage was reported.83
Abdominal pain Three cases of abdominal pain
were reported following ECT and of unknown
aetiology, and healthy babies were born in all
cases.
Premature labour Four cases of premature labour
after women had ECT were reported. In
subsequent reports,82,84 premature labour was
reported in a further two cases. In one case,82
Still birth and neonatal death Three cases of
stillbirth or neonatal death were reported.
Respiratory distress One case of the baby having
difficulty breathing at birth was reported.
Teratogenicity Five cases of congenital anomalies in
offspring of mother who received ECT have been
reported. The anomalies included hypertelorism,
optic atrophy, anencephaly, club foot and
pulmonary cysts. Four cases of developmental
delay or mental retardation have been reported.
Conclusions and discussion
The conclusions and a discussion of the
effectiveness review are considered in Chapter 7.
47
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 4
Economic analysis
Introduction
There were no sponsor submissions to NICE to be
evaluated. Therefore, economic models were
constructed based on the review of published
evidence to estimate whether ECT is a costeffective treatment for depression and
schizophrenia. No economic models were
constructed for mania or catatonia owing to the
lack of published data on these specific depression
subgroups. An attempt to estimate the cost per
quality-adjusted life-year (QALY) has been made
using published data on health state utilities.
Search strategy
Searches were undertaken to identify any
economic studies relating to ECT, as reported in
Chapter 3. No papers were identified in the
economics search. The economics search was then
extended to relate to any treatment undertaken in
treating depression, schizophrenia, mania and
catatonia, and any data relating to ECT that could
be used in an economic model were identified.
Overview of economic literature review
and economic evidence
There was no literature concerned with the costeffectiveness of ECT to review. This resulted in the
need to build an economic model based on the
authors’ perceived view of how ECT is used in the
UK, through dialogue with advisors on what are
the comparator treatments to ECT.
Economic modelling of ECT for
depressive illness, schizophrenia,
catatonia and mania
Modelling depressive illness
Introduction
It is commonplace today to see cost-effective
modelling techniques regularly used in deciding
whether a treatment is deemed to be superior or
otherwise to any other. Although not widespread,
cost-effective modelling has been used in the area
of depression, comparing one pharmacological
treatment with another. However, to the authors’
knowledge no one has attempted to evaluate the
cost-effectiveness of ECT.
ECT and antidepressant therapy are the primary
treatments available to patients suffering from
depressive illness. For mild to moderate
depression drug therapy is usually the first line of
treatment in the UK. ECT is primarily only
administered for patients suffering from severe
depression and is usually administered on an
inpatient basis. Even for patients suffering from
severe depression and requiring hospitalisation,
antidepressant therapy is still seen as the first line
treatment, with ECT only being administered to
patients deemed as being resistant to drug therapy
or those who have previously been treated
successfully with ECT.233 However, some people1
support the view that ECT could be seen as a first
line treatment for severe depression.
Methodology
As the literature search produced no economic
analysis on ECT within depression, a
mathematical model was constructed using data
from the clinical effectiveness evidence review and
other relevant studies to derive clinical outcomes
for ECT and its comparators. Health utility scores
were adapted from relevant studies and
incorporated in the model. As ECT is primarily
provided on an inpatient basis for severely
depressed patients the analysis concentrated on
comparing inpatient ECT with other inpatient
treatments for severe depression. Input from Dr
Paul Birkett (Clinical Lecturer, Honorary
Consultant Psychiatrist, University of Sheffield)
was sought for help in constructing the model.
The pharmacoeconomic model used for the costeffective analysis is based on a decision tree model
incorporating Monte Carlo simulation techniques
that determine the movement through the states
depending on the treatment that the patient
receives. The model attempts to evaluate the costeffectiveness of ECT for adult patients suffering
from a major depressive disorder (MDD) who
require hospitalisation. The model attributes
quality of life utility scores to each health state and
determines the movement through the states.
The health states in question are:
●
state 1: severely depressed receiving inpatient
treatment
49
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Economic analysis
●
●
●
state 2: receiving maintenance/continuation
therapy following successful antidepressant
therapy
state 3: receiving longer term psychotherapy
having failed to respond to acute antidepressant
therapy
state 4: failing to respond to maintenance
therapy and returning to a moderately
depressed state.
Figure 1 shows the structure of the decision model.
The model uses a 12-month time horizon, as valid
data for longer periods are not readily available
and hence discounting has not been undertaken.
The time unit used in this model is a week. For
each week throughout the year the model
determines whether the patient is severely
depressed and receiving acute treatment; has
successfully completed acute treatment, is no
longer severely depressed and is receiving
maintenance/continuation therapy; is receiving
longer term psychotherapy; or is in a relapsed
state following successful treatment. Each state has
a quality of life utility score attached to it and
incorporates a relevant cost.
As opinion differs as to whether ECT should be
undertaken as a final option when all else has failed
or should be provided higher up the treatment
hierarchy, the model has been constructed to allow
the evaluation of cost-effectiveness of ECT
provided as a first, second or third line (defined as
treatment-resistant) treatment.
50
ECT can be provided using either bilateral or
unilateral placement of electrodes on the head.
Bilateral ECT is generally more efficacious, but
also results in more side-effects. A randomised
trial by Sackeim and colleagues135 found that
unilateral ECT delivered with high stimulus
intensity relative to seizure threshold is equivalent
in efficacy to a criterion standard form of bilateral
ECT, yet retains important advantages with respect
to cognitive adverse effects. Patients who fail to
respond to unilateral ECT are frequently moved to
bilateral treatment. Therefore, the approach that
has been taken in the model is to group ECT as
one treatment and by varying the efficacy,
outcomes and cost in the sensitivity analysis
incorporate the different approaches used in
providing ECT. The main comparative treatments
to ECT analysed here are the three main classes of
antidepressants used in the UK: TCAs, SSRIs and
SNRIs. Augmentation of a pharmacological
intervention with lithium is also considered in the
analysis.
Following successful therapy, patients are usually
treated on maintenance/continuation therapy to
help to prevent relapse. Following successful ECT,
maintenance ECT can also be provided, normally
on an outpatient basis. The comparative treatments
that are used for maintenance/continuation
therapy that the model addresses are TCA,
lithium, ECT and no therapy.
The model shows that three different phases of
treatment are allowed before a final treatment of
psychotherapy is used on non-responders. During
each treatment episode there is a probability that
the patient could have an adverse event or be
deemed as not responding to the treatment and so
move to the next treatment phase before
completing the current treatment phase. After
completion of a treatment phase there is a
probability that the treatment is successful and the
patient is discharged. Patients who are deemed not
to have responded to treatment move to the next
treatment phase. The probability of successful
treatment and leaving the treatment early owing
to an adverse event or not responding to
treatment is related to the type of treatment
received and at which phase of the process the
treatment was administered.
Following successful treatment, patients may be
given continuation therapy to help to prevent
relapse.
Parameter values used in the model are based on
data from the clinical effectiveness element of the
review for ECT for depressive illness,
schizophrenia, catatonia and mania, together with
literature searches on the economic evaluation of
depression. Analysis of the literature produced
different definitions of what constituted ‘successful
treatment’. For the model, therapeutic success has
been quantified as a 50% decrease in the HRSD or
other depression scoring system as used in other
economic evaluations in depression.234–236
Caveat
The model has only used monotherapy
pharmacological treatments as comparators to
ECT, although combination treatments are
sometimes used in the treatment of depression.
However, there is very little quality research on
the success or otherwise of these treatments or
on combining drug therapies. The model
makes no assumptions about previous
depressive episodes and previous treatment
received.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Discontinue
treatment
Treatment 1
FIGURE 1 Structure of the decision model
MDD
episode
State 1
Discontinue
treatment
Treatment 2
Non-responder
Responder
Discontinue
treatment
Treatment 3
Non-responder
Responder
Maintenance therapy
State 2
State 4
Continued care
State 3
Moderately depressed
Non-responder
Responder
Relapse
Health Technology Assessment 2005; Vol. 9: No. 9
51
Economic analysis
Assumptions and probabilities
Efficacy
A meta-analysis of ECT efficacy undertaken by
Janicak and colleagues237 in 1985237 showed that
ECT was approximately 20% more effective than
TCAs in the treatment of depressed patients.
Although the analysis looked at studies from the
1960s, no comparative study has ever found a
medication regimen to be more effective than
ECT in the treatment of major depression.238 An
RCT by Prudic and colleagues in 1990239
compared ECT in patients who were defined as
treatment resistant and those that were not. They
found that the success rate (>60% reduction in
HRSD score) was 86.2% and 50% for nontreatment-resistant and treatment-resistant
patients, respectively. An RCT by Folkerts and
colleagues in 1997112 comparing ECT with an
SSRI in treatment-resistant depression (defined as
failing at least two previous antidepressant trials)
showed that 71% of patients fulfilled the response
criteria of a 50% decrease in the HRSD score
compared with 29% for the SSRI.
The clinical effectiveness review concludes that
based on trials of ECT versus pharmacological
treatment, the people treated with ECT were 42%
more likely to be defined as responders than those
treated with a TCA (RR = 1.42, 95% CI 1.17 to
1.72, p = 0.0004). A meta-analysis of randomised
trials by Einarson and colleagues240 found that the
average successful treatment rate for TCA
treatment was 58.2%. Applying a relative risk of
1.42 to this figure results in an expected success
rate for ECT of 82.6%, which is very close to the
success rate that Prudic and colleagues239 found
for ECT.
The model default assumption for clinical success
for the treatment of major depressed patients
undertaking ECT was taken from the Prudic study,
with first and second line therapy for ECT having
an 86.2% success rate and the third line therapy
having a 50% success rate.
The failure to complete treatment rates for ECT,
derived from Burke and colleagues,241 suggests
that between 18 and 35% of ECT patients do not
complete the treatment. For the model it has been
assumed that these figures are the 95% CI and the
mean has been calculated as the midpoint.
52
The assumptions regarding the successful
treatment rates and dropout/failure to complete
treatment rates for the different classes of
antidepressant drugs are taken from Doyle,242
Freeman and colleagues243 and Einarson and
colleagues240 which, in turn, were all based on a
meta-analysis of randomised trials comparing
TCAs, SSRIs and SNRIs undertaken by Einarson
and colleagues.18 It has been assumed that each
treatment’s failure to complete treatment rate is
independent of the line of therapy. The efficacy
rates for the pharmacological treatments are from
trials undertaken in an inpatient setting on
patients who had an HRSD score of at least 15 or
a Montgomery and Asberg Depression Rating
Scale (MADRS) score of at least 18. The measure
of success is the percentage of patients who
achieved a 50% reduction in their score. The
failure to complete treatment rates are a
combination of lack of efficacy and patients
experiencing adverse events. For patients who are
deemed treatment resistant, lithium augmentation
is seen as an effective pharmacological
intervention. A meta-analysis by Bauer and
Dopfmer244 of placebo-controlled studies of
lithium augmentation in treatment-resistant
depression concluded that lithium augmentation,
usually an SSRI with lithium, “should be the first
choice treatment procedure for depressed patients
who fail to respond to antidepressant
monotherapy”. The results of this paper were used
as the successful treatment rates for the third line
pharmacological therapy. The failure to complete
treatment rates for this third line therapy are
assumed to be the same as those for an SSRI
intervention.
The model assumes that when primary
pharmacological treatment fails, a second line
treatment would have the same success rate, as it
would have been as the primary treatment. This
assumption may not be true and it could be
viewed as favouring the less effective treatments
when the more effective treatments are given as
back-up. For a given population of depressed
patients there would be a proportion who would
respond well to treatment irrespective of whether
that treatment was an SSRI or a TCA.
Consider the following simplified example in
which it is assumed that there are only two
treatments, treatment A with a success rate
of 60% and treatment B with a success rate of
50%, and for simplicity both have a failure to
complete treatment rate of zero. The overall
successful treatment rate (after both treatments
had been administered) could vary from 60%
(success rate of treatment A) to 100% depending
on the proportion of patients who would have
responded to either treatment. Given that
the sum of the success rates of treatment A
and treatment B is greater than 100%, implicitly
Health Technology Assessment 2005; Vol. 9: No. 9
A only
30%
A&B
30%
B only
20%
Neither
A nor B
20%
FIGURE 2 Venn diagram of treatment success. Successfully
treated by treatment A as a single therapy = A% = 60%;
successfully treated by treatment B as a single therapy = B% =
50%; successfully treated by A and B = X%, where 10% ≤
X ≥ 50%; successfully treated by A or B = A% + B% – X%.
there must be at least a 10% overlap in which
patients would have responded to either
treatment. If the overlap rate were only 10% then
the overall treatment success following both
treatments would be 100%. If the assumption is
that the success rate is the same for the treatment
regardless of whether it is given as a first or
second line therapy, then with a population of
1000 people, 800 (80%) will be successfully
treated after both treatments have been given
(1000 × 0.6) + (1000 – 1000 × 0.6) × 0.5.
Figure 2 shows a Venn diagram that represents the
above example. The square box represents the
population, while the circles represent the success
rates for treatments A and B. The area where the
circles overlap represents the proportion of
patients who would have responded to treatment
A and also responded to treatment B. The area
outside the circles represents the proportion of
patients who would not respond to either
treatment A or treatment B.
If it is assumed that the success rate for a second
line treatment is the same as for a first line
treatment then in the example shown in Figure 2
X must equal 30% to give the overall success
rate of both treatments as 80%. However, if
the proportion of patients who would respond
to both treatment A and treatment B were
40% (X) then the overall success rate following
both treatments would be 70%. This would
be equivalent to assuming that the success rate
for the second line treatment B is half that if it
were given as a first line treatment in this
example.
Therefore, the assumption in the model that
treatments given as a second line therapy have the
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
same success rate as if they were given as a first
line therapy has implications on the assumed
proportion of patients who would have responded
to either treatment.
Patients requiring third line therapies are deemed
“treatment resistant” and thus lithium
augmentation has been assumed as the preferred
third line pharmacological therapy.
Table 4 summarises the model’s default values for
clinical success for each treatment when used as a
first, second or third line therapy, together with
each treatment’s dropout rates. Table 5 summarises
the model default values for failure to complete
treatment rates.
The final longer term treatment of psychotherapy
has been assumed to be an 8-week treatment in
which patients are assumed to make a moderate
improvement. More detailed assumptions about
this treatment can be found in the quality of life
and cost sections.
Duration of treatment
Folkerts and colleagues112 found that ECT is
considered to be quicker than pharmacological
interventions in achieving a positive treatment
response. Pharmacological treatments are usually
continued for 6 weeks before the full effectiveness
is achieved.245 Therefore, the model defaults for
TABLE 4 Clinical success for pharmacological and ECT
interventions in major depression
Treatment Clinical success
Mean
First line
58.2
58.6
62.3
82.6
58.2
58.6
62.3
82.6
27.0
50.0
TCA
SSRI
SNRI
ECT
Second line TCA
SSRI
SNRI
ECT
Third line
Lithium augmentation
ECT
95% CI
43.0 to 73.5
48.2 to 69.0
49.7 to 74.9
52.1 to 98.8
43.0 to 73.5
48.2 to 69.0
49.7 to 74.9
52.1 to 98.8
9.8 to 44.2
30.0 to 70.0
TABLE 5 Failure to complete treatment rates
Treatment
Average
TCA
SSRI
SNRI
Lithium augmentation
ECT
29.9
25.8
20.7
25.8
26.5
95% CI
22.7 to 37.1
20.3 to 31.3
15.3 to 26.1
20.3 to 31.3
18.0 to 35.0
53
Economic analysis
TABLE 6 Maintenance therapy relapse assumptions
Maintenance therapy
Following pharmacological intervention
SSRI
No therapy
13%
46%
Following ECT
ECT
Lithium + TCA
TCA only
No therapy
33%
32%
56%
72%
the duration of treatments within each phase of
the model are:
●
●
6 weeks for pharmacological treatments,
dropouts averaging 2 weeks of treatment
4 weeks for ECT, dropouts averaging 1 week of
treatment.
Continuation/maintenance therapy
As relapse rates following successful treatment in
major depression are high, up to 80% within a
year,246 the common practice is to provide
maintenance or continuation therapy to help to
prevent relapse. A study by Hirschfeld in 2001247
showed that approximately one-third to half of all
patients will relapse within a year following
pharmacological therapy if medication is not
continued. An RCT by Sackeim and colleagues68
showed that a combination of lithium and a TCA
had the greatest effect in reducing the number of
relapses following successful ECT in medicationresistant patients.
Caveat
The survival rates from Sackeim and
colleagues68 were to 24 weeks only. In the
model the survival times were extended to
48 weeks. This assumption may not be valid.
However, most relapses occur in the first
10 weeks of treatment.
Costs and treatment dosage
The cost for each pharmacological therapy was
taken from the British National Formulary, 42nd
edition, September 2001 (BNF42).249 The doses of
SSRIs and TCAs were taken from Hirschfeld’s
study of clinical trials of SSRIs and TCAs
conducted on severely depressed patients
receiving inpatient treatment.250 The dosage for
venlafaxine (SNRI) was taken from Einarson and
colleagues’ pharmacoeconomic analysis of
venlafaxine.251
The Kaplan–Meier survival curves in these studies
were translated into the model to serve as default
assumptions for relapse rates following successful
depression treatment.
The number of ECT treatments was based on the
UK practice of two treatments per week and with
average treatment duration of 4 weeks; an average
of eight ECT treatments is given per therapy. The
cost of ECT was taken from Montgomery and
colleagues’ study,252 which had a 1994 cost of
£2055 for six sessions. The estimated cost for ECT
was increased from 1994 to 2001 values using the
Hospital and Community Health Services inflation
index from the Unit costs of health and social care.253
A pharmacoeconomic model by Hatziandreu in
1994 looking at the maintenance treatment of
recurrent depression listed the resource utilisation
and costs of maintenance treatment for patients
with major depression.254 This comprised blood,
thyroid and liver tests, and visits to the GP,
psychiatrist and psychiatric nurse. This resource
pattern was adopted for the maintenance resource
use for this model, with the costs increased to
2001 values.
The model default values for relapse prevention
for each type of maintenance/continuation therapy
are shown in Table 6.
Tables 7 and 8 summarise the default dosage and
cost estimates for each acute treatment and
maintenance therapy.
Continuation/maintenance ECT has been shown to
be an effective treatment in preventing relapse in
patients successfully treated with ECT. Swoboda
and colleagues248 found that, for patients with
affective and schizoaffective disorders following
successful ECT, 33% of patients who received
continuation/maintenance ECT relapsed (defined
as being readmitted to hospital), while 67%
patients who had not received continuation/
maintenance relapsed after 12 months. No studies
were found that analysed maintenance ECT for
non-schizoaffective patients; therefore, an
assumption has been made that continuation ECT
is as effective for depressive patients as for patients
with affective and schizoaffective disorders.
54
Relapse rate at 48th week
Health Technology Assessment 2005; Vol. 9: No. 9
TABLE 7 Cost of acute treatment for major depression
Acute therapy
Drug
Dosage
Unit cost
Hospital costs
per day
Cost per
weeka
TCA
Clomipramine
(non-proprietary)
Paroxetine (Seroxat®)
Venlafaxine (Efexor®)
150 mg day–1
£0.26
£171
£1198.82
30 mg day–1
300 mg day–1
Two sessions
per week
800 mg Lithium +
30 mg paroxetine
£1.04
£2.86
£2475 per
six treatments
£1.12
£171
£171
£1204.27
£1216.99
£171
£171
£2022.00
£1204.84
SSRI
SNRI
ECT
Lithium augmentation
a
Lithium + SSRI
Weekly cost equals 7 days at the inpatient costs per day of £171 plus 7 days at the unit treatment cost. ECT weekly dose
is two treatments per week (£825).
TABLE 8 Cost of continuation/maintenance therapy for major depression
Acute therapy
Drug
Dosage
Unit cost
Hospital costs
per yeara
Cost per
weekb
TCA
SSRI
Lithium + TCA
Nortriptyline
Nefazodone (Dutonin®)
Lithium + nortriptyline
50 mg day–1
412 mg per day
600 mg lithium +
50 mg TCA day–1
Average two
per month
£0.46
£0.62
£0.54
£260
£260
£260
£5.24
£9.33
£8.78
£2475 per
six treatments
Included
£190.4
ECT
a
b
Based on tests and visits to the GP, psychiatrist and psychiatric nurse as stated in Hatziandreu.254
Based on 24 treatments per year divided by 52 weeks.
The cost of continued care therapy (state 3) is
based on the daily cost of maintaining a nursing
home placement with psychiatric provision at a
cost of £993253 per week for an average of
8 weeks. This cost averages out at £6951 per
patient who fails to respond to acute treatment.
For patients who relapse from maintenance therapy
it has been assumed that they continue to take
medication (equivalent of 20 mg of fluoxetine per
day) and attend an outpatient visit once per month
(£131). This averages out at £32.05 per week.
Caveat
The costs for continued care therapy (state 3)
and maintenance relapse (state 4) are not
based on any research but are estimates made
by the authors. The model uses them as a cost
offset in that the cost of treating patients in
trying to prevent them reaching state 3 is
offset by the cost savings of not having to treat
them in state 3. The higher the costs of
treating patients in states 3 and 4 the higher
the potential savings will be.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Quality of life utility estimates
In order to estimate QALYs, information is
needed on the utility values that can be assigned
to different health states. Utility values are defined
on a 0–1 scale, where 1 represents perfect health
while 0 represents death. The sources for this
information were primarily two independent
studies in which utility values for severe
depression, moderate depression, mild depression
and depression in remission were estimated.255,256
Other studies have derived utility values for
depressed patients receiving different
pharmacological treatments and their estimates
were also included in the modelling exercise where
appropriate.236,254
The utility values from the study by Bennett and
colleagues256 were elicited using the McSad health
states classification system. Values were obtained
from 105 patients who had experienced at least
one episode of major, unipolar depression in the
previous 2 years, but who were currently in
remission. The health state descriptions
referred to untreated depression. The mean
utility values for each health state are shown in
Table 9.
55
Economic analysis
TABLE 9 Mean utility values for depression states256
Mean utility
95% CI
0.09
0.32
0.59
0.79
0.05 to 0.13
0.29 to 0.34
0.55 to 0.62
0.74 to 0.83
Severe depression
Moderate depression
Mild depression
Depression in remission
TABLE 10 Mean utility values for depression states255
State
Mean utility (SD)
Severe depression, untreated
0.30 (0.28)
Moderate depression
Nefazodone
Fluoxetine
Imipramine
0.63 (0.23)
0.63 (0.19)
0.55 (0.03)
Mild depression
Nefazodone
Fluoxetine
Imipramine
0.73 (0.21)
0.70 (0.20)
0.64 (0.20)
Depression in remission
Nefazodone
Fluoxetine
Imipramine
0.83 (0.13)
0.80 (0.15)
0.72 (0.17)
The utility values from the Revicki and Wood
study255 were elicited through the administration
of standard gamble questions to 70 patients with
major depressive disorder or dysthymia. Unlike
the Bennett study,256 the health state descriptions
that were evaluated included descriptions of the
side-effects of drug treatment. Three different
drugs were considered: nefazodone (SSRI),
fluoxetine (SSRI) and imipramine (TCA). The
mean utility values and standard deviations for
each health state are shown in Table 10.
The utility values from the Revicki study255 have
very large standard deviations. This reduces the
confidence that there is any significant difference
both between the treatments within each level of
severity of depression and between the different
severity levels. With this in mind, it was decided to
use the Bennett256 utility values as the model
defaults. Results using the Revicki study255 utility
values in the model are presented in the sensitivity
analysis, later in this chapter.
In the model it is assumed that patients admitted
to hospital are classed as having severe depression.
This would translate to a high HRSD score,
probably over 20.
The default model parameter values for QALY
utility estimates were taken from Bennett and
colleagues256 and translate to the health states
within the model. They are shown in
Table 11.
Non-responders (state 3) receive intensive
psychotherapy and on completion of treatment are
deemed to have improved to a depression level
similar to mild depression. Patients who relapse
from maintenance therapy (state 4) do not revert
to being severely depressed, but require treatment
to maintain a quality of life equivalent to
moderate depression.
The default scenario is that the QALY utility
scores are the same for all patients regardless of
which treatment they have received. This
assumption may not be true as side-effects
following treatments such as ECT may result in
memory loss and hence a lower QALY utility
score. Variation in the QALY assumptions is
analysed in the sensitivity analysis section.
Caveat
QALY utilities appear low for severely
depressed patients, but reflect what a
debilitating illness depression can be. The
assignment of QALYs to states 3 and 4 is not
based on any research, but is the authors’
decision.
TABLE 11 Quality of life utility assumptions
56
State
Definition
1
2
3
4
Severely depressed, receiving inpatient treatment
Responded to treatment, receiving maintenance therapy
Non-responder
Relapsed from maintenance therapy
Mean utility
95% CI
0.09
0.79
0.59
0.32
0.05 to 0.13
0.74 to 0.83
0.55 to 0.62
0.29 to 0.34
Health Technology Assessment 2005; Vol. 9: No. 9
TABLE 12 Summary of model scenarios
Scenario
1
2
3
4
5
6
7
8
Strategy
First treatment
Second treatment
Third treatment
SNRI
ECT
ECT
SNRI
ECT
SNRI
SNRI
SNRI
SSRI
SSRI
SSRI
ECT
SSRI
SSRI
ECT
SSRI
Lithium augmentation
Lithium augmentation
Lithium augmentation
Lithium augmentation
Lithium augmentation
ECT
Lithium augmentation
ECT
Suicide risks
Evidence from the review of clinical effectiveness
tends to support the view that there is no
significant difference in suicide rate between
patients treated with ECT and those treated with
pharmacological treatment. A suicide rate of
0.85% per depressive episode is widely quoted and
has been used in other economic evaluations.236
The assumption used in the model is that the
longer the patient remains a non-responder the
greater the chance of their committing suicide.
Once the patient has failed the third line therapy
they are assumed to receive psychotherapy (state
3). After this point is reached the chance of suicide
is reduced to zero. Therefore the assumption is
that patients who fail to respond to treatment or
are not receiving treatment have a risk of suicide.
The 0.85% suicide rate per depressive episode was
converted into a weekly chance by assuming an
arbitrary average duration per depressive episode
(13 weeks). This assumption favours the
treatments with higher efficacy and shorter
duration to success. Sensitivity analysis performed
on this variable is reported.
Summary of scenarios
Table 12 shows a summary of the treatment
therapies that were combined to form the eight
scenarios that were analysed by the model.
Results
A Monte Carlo simulation approach was taken by
varying the inputs for the successful treatment
rates, failure to complete therapy rates, quality of
life utility values and treatment costs. Values were
selected at random from within the 95% CI, based
on a normal distribution (Tables 4 and 5). For all
costs, a pseudo-confidence interval was generated
using a standard deviation of 15%. This generated
a 60% range in cost that was considered suitable to
reflect fluctuations in cost that may occur.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Combining the different treatments available into
first, second and third treatment therapies can
generate a number of different treatment
strategies. Table 13 shows the results from the 3000
Monte Carlo simulation runs of different
treatment strategies.
Scenario 4 has the cheapest average total cost per
patient at £10,592, while scenario 2 is the most
expensive with an average total treatment cost of
£15,354. Scenario 5 generates the most QALYs
(0.539), while scenario 3 generates the fewest with
only 0.424 QALYs. However, it should be noted
that when considering the 95% CIs for both the
average costs and QALYs, there is a high degree of
overlap between the scenarios. Scenario 1 was
considered as the pharmacological treatment
comparator as it is the best in terms of cost per
QALY. This is mainly due to both the SNRI
success and SNRI failure to complete treatment
rates, which have the highest and lowest mean
value, respectively. However, it should be noted
that owing to the range of values the parameters
can take, the 95% CIs do overlap with other
pharmacological treatments (not shown).
Scenarios 2, 3 and 5 represent the results of having
ECT as the primary strategy. The only difference
between the strategies is the maintenance therapy
provided to the patients treated with ECT.
Scenario 2 provides maintenance ECT, while
scenario 3 provides lithium plus TCA combination
as the maintenance therapy and scenario 5
assumes that an SSRI is an effective maintenance
treatment to prevent relapse.
Scenarios 4 and 7 show the results of having ECT
as the second line therapy. The only difference
between the strategies is the maintenance therapy
provided to the patients treated with ECT.
Scenario 4 has lithium and TCA as the
maintenance therapy for patients successfully
57
Economic analysis
TABLE 13 Treatment scenario results
Scenario
Strategy
First
treatment
Second
treatment
Third
treatment
Maintenance therapy
Average total
cost/patient
(95% CI)
QALYs
(95% CI)
1
SNRI
SSRI
Lithium
augmentation
SSRI following all
three treatments
£11,400
(£9349 to £13,718)
0.490
(0.453 to 0.526)
2
ECT
SSRI
Lithium
augmentation
SSRI following two
treatments. Maintenance
ECT following ECT
£15,354
(£13,445 to £17,361)
0.458
(0.422 to 0.493)
3
ECT
SSRI
Lithium
augmentation
SSRI following two
treatments. Lithium +
TCA following ECT
£10,997
(£9080 to £13,045)
0.424
(0.389 to 0.459)
4
SNRI
ECT
Lithium
augmentation
SSRI following two
treatments. Lithium +
TCA following ECT
£10,592
(£8874 to £12,435)
0.470
(0.431 to 0.508)
5
ECT
SSRI
Lithium
augmentation
SSRI following all three
treatments
£11,022
(£9016 to £13,069)
0.539
(0.498 to 0.579)
6
SNRI
SSRI
ECT
SSRI following two
treatments. Lithium +
TCA following ECT
£13,939
(£11,161 to £17,049)
0.489
(0.452 to 0.524)
7
SNRI
ECT
Lithium
augmentation
SSRI following two
treatments. Maintenance
ECT following ECT
£12,591
(£10,678 to £14,497)
0.486
(0.449 to 0.522)
8
SNRI
SSRI
ECT
SSRI following two
treatments. Maintenance
ECT following ECT
£14,548
(£11,680 to £17,717)
0.494
(0.459 to 0.529)
treated with ECT, while scenario 7 provides
maintenance ECT.
Scenarios 6 and 8 show results of having ECT as
the third line therapy. Again the only difference
between the strategies is the maintenance therapy
provided to the patients treated with ECT.
Scenario 6 has lithium and TCA as the
maintenance therapy for patients successfully
treated with ECT, while scenario 8 provides
maintenance ECT.
Net benefit
When comparing the cost-effectiveness of two or
more treatments a consideration of the
incremental net benefit of one treatment over
another is required. The net benefit of the
treatments combines the health gain and financial
consequences. The net benefit can be presented in
monetary terms as the net monetary benefit
(NMB) or in health outcome terms as the net
health benefit (NHB):
58
NMB = E – C
NHB = E – C/
where is the amount that one is prepared to pay
to gain one unit of health benefit (also called the
societal value), in this case a QALY, E is the effect
(or health outcome) and C is the cost.
For example, if the societal value of a QALY (the
amount that one is prepared to pay to gain 1
QALY) is £30,000 then for a treatment that
provides 2.0 QALYs for a cost of £15,000 the net
benefit is:
£30,000 × 2.0 – £15,000 = £45,000
That is, £45,000 is the average NMB of
introducing this treatment.
The incremental net benefit (NMB and NHB) of
one treatment (T1) over another (T0) is
represented by the formulae:
(QALYs T1 – QALYs T0) – (Cost T1 – Cost T0)
or .∆E – ∆C
(QALYs T1 – QALYs T0) – ((Cost T1 – Cost T0)/)
or ∆E – ∆C/
Health Technology Assessment 2005; Vol. 9: No. 9
TABLE 14 Results of average NMB analysis for the scenarios
Scenario
Average NMB
1
2
3
4
5
6
7
8
£3330
–£1614
£1422
£3508
£5148
£731
£1989
£272
where is the societal value of a QALY, ∆E is the
difference in effect and ∆C is the difference in
cost.
The difference between two average net monetary
benefits has a useful property in that:
—
—
NM B1 – NM B0 =
=
=
=
(.E1 – C1) – (.E0 – C0)
(E1 – E0) – (C1 – C0)
.∆E
– ∆C
—
∆NM B
Therefore, one can formulate the average net
benefit for each scenario. The scenario with the
highest net benefit is the preferred option and
there is no need to worry about an appropriate
comparator, as there would be with cost-effective
ratios.
Table 14 shows the average NMB for each of the
treatment strategies assuming that the societal
value of a QALY is £30,000.257
Table 14 shows that scenario 5 would be the
preferred strategy as it has the highest average
net benefit. If scenario 5 did not exist as a
realistic option then scenario 4 would be the
preferred strategy. Scenario 2 is the only
scenario with a negative average net benefit.
The 95% CIs for the average net benefit for
each scenario have a high degree of overlap.
This shows that one cannot be certain of the
rank order of the average net benefit of the
scenarios.
Caveat
The average net benefit analyses were
undertaken on the mean cost and mean
QALYs of each scenario only. Table 14 shows
that there is a high level of overlap in the
confidence intervals.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
95% CI
Rank order
£720 to £5723
–£3866 to £498
–£771 to £3586
£1134 to £5624
£2660 to £7602
–£2643 to £3917
–£311 to £4275
–£3325 to £3482
3
8
5
2
1
6
4
7
Sensitivity analysis
This section of the report attempts to evaluate the
robustness of the model assumptions and show
which variables require further information to
increase confidence in the results.
QALY sensitivity analysis
The default quality of life utility scores used in the
model were derived from the Bennett study.256
However, another study by Revicki and Wood255
presented significantly different QALY scores,
especially for severely depressed patients. Table 15
shows the results of the costs and QALYs for each
scenario based on the Revicki QALY utility
estimates following 3000 runs of the model. The
costs should be very similar to the results in
Table 13, as these assumptions have not altered.
The QALYs gained by each scenario have decreased
owing to the reduction in QALY utility between
severely depressed and the other depression levels.
As with the scenarios based on the default
assumptions, there is a high degree of overlap
between each scenario’s cost and QALY results.
An NMB analysis between each of the eight
scenarios using the Revicki QALY assumptions is
shown in Table 16. Again, it has been assumed that
the willingness to pay for one QALY is £30,000.
The results in Table 16 show that scenarios 1 and 4
have changed places in the preferred strategy order.
Sensitivity of the cost of ECT
The assumption for the cost of ECT is based on a
paper from 1994 and increased for inflation. The
following analysis reports the effect on the eight
scenario results of decreasing the average cost of
ECT by 25% while keeping all the other
assumptions at their default values. Table 17 shows
the cost and QALYs for each of the eight
scenarios. All of the scenarios that have ECT
included as a treatment have reduced their
average cost. This reduction in cost varies between
59
Economic analysis
TABLE 15 Scenario results based on Revicki QALYs255
Scenario
Cost (95% CI)
1
2
3
4
5
6
7
8
QALY (95% CI)
£11,325 (£9204 to £13,647)
£15,329 (£13,452 to £17,291)
£11,205 (£9206 to £13,405)
£10,613 (£8913 to £12,450)
£10,965 (£8978 to £13,065)
£13,946 (£11,201 to £17,061)
£12,597 (£10,751 to £14,587)
£14,550 (£11,736 to £17,704)
0.346 (0.311 to 0.381)
0.297 (0.261 to 0.333)
0.261 (0.225 to 0.296)
0.314 (0.278 to 0.353)
0.378 (0.338 to 0.419)
0.341 (0.305 to 0.377)
0.329 (0.293 to 0.365)
0.344 (0.309 to 0.381)
TABLE 16 Average NMB results for each strategy using Revicki QALYs255
Scenario
Average NMB
1
2
3
4
5
6
7
8
–£945
–£6419
–£3375
–£1193
£375
–£3716
–£2727
–£4230
95% CI
–£3441 to £1461
–£8656 to –£4207
–£5581 to –£1206
–£3454 to £996
–£2160 to £2700
–£7217 to –£707
–£4868 to –£411
–£7647 to –£1035
Rank order
2
8
5
3
1
6
4
7
TABLE 17 Scenario results based on reduction of 25% in ECT cost
Scenario
1
2
3
4
5
6
7
8
Cost (95% CI)
QALY (95% CI)
£11,349 (£9191 to £13,699)
£12,747 (£11,104 to £14,552)
£9739 (£7962 to £11,710)
£9871 (£8184 to £11,684)
£9518 (£7661 to £11,485)
£13,568 (£10,876 to £16,760)
£11,296 (£9595 to £13,063)
£13,990 (£11,167 to £17,169)
0.490 (0.453 to 0.525)
0.458 (0.424 to 0.492)
0.421 (0.388 to 0.456)
0.470 (0.432 to 0.509)
0.538 (0.499 to 0.580)
0.490 (0.453 to 0.526)
0.486 (0.449 to 0.523)
0.494 (0.457 to 0.531)
the scenarios depending on whether ECT is
prescribed as a first line therapy and whether
maintenance ECT is also given. The confidence
intervals of the cost and QALYs still have a high
level of overlap between the scenarios.
Analysis of the average NMB does not produce
anything surprising. Table 18 shows that although
the actual NMBs have changed from the scenarios
with the default ECT costs, the preferred strategy
order remains the same. Again, there is a large
amount of overlap in the confidence intervals of
the different scenarios.
60
Sensitivity analysis was also performed on the
cost assumptions of treatment for continued
care (state 3) and cost of patients who fail to
respond to maintenance therapy (state 4), but this
made little difference to the overall scenario
results.
Sensitivity analysis was also performed on the
model assumptions of suicide rates. The average
duration per depressive episode was altered to
increase and decrease the suicide rate. These
changes had little effect on the overall results.
Conclusions
The model described here is the first known
attempt at modelling the cost-effectiveness
of ECT in a depressed population. Evidence from
published trials was used where possible, but it is
Health Technology Assessment 2005; Vol. 9: No. 9
TABLE 18 Average NMB for each of the treatment strategies based on a 25% reduction in ECT cost
Scenario
Average NMB
1
2
3
4
5
6
7
8
£3399
£67
£2035
£3763
£5792
£888
£2827
£628
accepted that a few assumptions were made based
on the authors’ limited knowledge of the area,
owing to a lack of available data. The model
appears to suggest that ECT treatment provided
as a second line therapy (scenario 4) would
be the preferred strategy as the average NMB is
greater than that of the pharmacological only
treatment (scenario 1), assuming a £30,000
willingness to pay threshold. However, this cannot
be stated with any great confidence as the
sensitivity analysis around the QALYs changes the
preferred strategy order. The main drawbacks in
terms of cost-effectiveness of using ECT as a
therapy are its higher costs and its higher rate of
relapse than the pharmacological treatments.
However, on the plus side there is evidence that
ECT has a high success rate of treatment both for
treatment-resistant and non-treatment-resistant
patients.
The economic modelling does not demonstrate
that any of the available scenarios has a clear
economic benefit over the other available
options. Specifically, if ECT should be used, the
mode does not indicate whether it should be a
first, second, or third line treatment. The main
reason for this is that there is a lot of uncertainty
around the values of the main parameters,
efficacy, and failure to complete treatment and
quality of life measures. This may be due in part
to the lack of RCTs concerned with ECT in the
severely depressed. However, it could also be the
nature of depressive illness. The clinical evidence
produced by this review suggests that ECT is an
effective treatment for depression for some
people, whereas for others it could even have a
detrimental effect.
Further research
The economic modelling undertaken for
depression showed a need for more robust
information on the effectiveness of treatment for
depressed patients. There is a lack of studies that
have attempted to estimate the quality of life of
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
95% CI
£697 to £5755
–£2133 to £2112
–£255 to £4186
£1521 to £5892
£3330 to £8146
–£2590 to £3900
£567 to £5019
–£3086 to £3816
Rank order
3
8
5
2
1
6
4
7
patients suffering from depression and there are
currently no studies that have tried to estimate the
quality of life of depressed patients who have been
treated with ECT.
Further economic analysis, such as expected value
of perfect information, may be useful in
identifying key parameters where further research
would reduce the uncertainty of the costeffectiveness estimate.
Modelling schizophrenia
Introduction
The main schizophrenic population for which
ECT is indicated in the APA and RCP guidelines
comprises patients resistant to
pharmacotherapy.3,32 Therefore, the model
structure concentrated on the use of ECT in
treatment-resistant schizophrenia. All the
economic analysis concentrated on
pharmacological intervention in the treatment of
schizophrenia. One cost–utility study was
identified that analysed treatment-resistant
schizophrenia: a Canadian study by Oh and
colleagues258 that centred on treating treatmentresistant schizophrenia with clozapine. This was a
decision tree model that compared clozapine with
a standard treatment using chlorpromazine or
haloperidol. Oh and co-workers obtained clinical
outcomes from a random effect, single-arm metaanalysis and utility weights were evaluated in a
cohort of patients by using a standard gamble
technique. As no cost-effectiveness study
incorporating ECT in the treatment of
schizophrenia existed and this was the only
cost–utility study that analysed treatment–resistant
schizophrenia, it was decided to use the
framework of Oh’s model and incorporate an ECT
arm to the decision tree by acquiring clinical
outcomes and other information on ECT in
treatment-resistant schizophrenia from other
appropriate studies. This would allow analysis of
whether ECT was a cost-effective treatment
compared with both clozapine, the standard
61
Economic analysis
treatment for patients who are treatment resistant,
and chlorpromazine, a neuroleptic which, as stated
by Thornley and colleagues,26 “remains the
benchmark treatment for patients with
schizophrenia”.
Methodology
Oh’s model is a cost–utility analysis that compares
the costs and quality-adjusted outcomes of
hospitalised treatment-resistant schizophrenia
with moderate symptomatology. Costs and
outcomes were evaluated over a time-frame of
1-year. Figure 3 shows the decision tree framework
with the added treatment arm of ECT.
The clinical outcomes for the pharmacological
interventions were obtained from the metaanalysis in Oh’s study. This meta-analysis was
conducted in 1995 and the search concentrated on
all RCTs involving clozapine, haloperidol and
chlorpromazine compared with placebo or active
therapy in treatment-resistant schizophrenia. For
ECT the clinical success outcome was based on a
study by Chanpattana and colleagues,196 which
was the only study in the clinical effectiveness
review that had both clinical outcomes and a
treatment-resistant population. The authors state
that research on the use of ECT in treatmentresistant schizophrenia has been characterised by a
variety of methodological limitations. There have
been no randomised single-blind studies
contrasting the efficacy of ECT and neuroleptic
treatment with neuroleptic treatment alone in
patients with treatment-resistant schizophrenia.
However, they conclude that the literature suggests
that ECT is effective in the treatment of
schizophrenia, and that ECT with a neuroleptic
appears to be more effective than either ECT
alone or neuroleptic treatment alone.
Chanpattana and co-workers196 conclude that
combined ECT and neuroleptic therapy effectively
reduced psychotic symptoms in 57% of treatmentresistant patients with schizophrenia.
The failure to complete treatment rates for ECT
were derived from Burke and colleagues241 and
suggest that between 18 and 35% of ECT patients
do not complete the treatment. For the model it
was assumed that these figures are the 95% CI and
the mean was calculated as the midpoint.
Table 19 shows the event rates for the three
comparators in the treatment of treatmentresistant schizophrenia.
62
Quality of life utility scores in the Oh study258 were
obtained through interviews with seven patients
with schizophrenia using the standard gamble
technique and a rating scale. Standardised patient
profiles were developed based on the average
Positive and Negative Symptoms Scale (PANSS)
score in each of three PANSS subscales (positive,
negative and general psychopathology) from
clinical trials used in their meta-analysis. It should
be noted that with only seven patients in the study
the confidence intervals for each estimate of quality
of life in each state overlap. Therefore, it could be
argued that there is no difference in quality of life
between the states. The robustness of these
assumptions is examined in the sensitivity analysis.
Caveat
The Oh paper258 was the only study that
incorporated utility scores for patients
suffering from treatment-resistant
schizophrenia. These patients were described
as having only moderate symptomatology.
These utility scores are higher than those used
in the depression illness model and the
variation between severities of illness is smaller.
It is unknown to the authors whether this is a
real reflection of the difference in quality of
life between patients with depression and those
with schizophrenia.
The resultant utility scores from the Oh study258
are shown in Table 20.
It was assumed that the utility scores of patients on
clozapine are applicable to patients following ECT.
The robustness of all the assumptions used in the
model was investigated in the sensitivity analysis.
Table 21 shows the dosage and cost assumptions
for each of the comparable treatments for
treatment-resistant schizophrenia.
The pharmacological treatment costs were taken
from the BNF42249 and dosages from Oh and
colleagues.258 The ECT cost is based on the study
by Montgomery and colleagues252 which estimated
that the cost of ECT in 1994 was £2055 for six
sessions. The estimated cost for ECT has been
increased from 1994 to 2001 values using the
Hospital and Community Health Services inflation
index from the Unit Costs for Health and Social
Care 2001.253 ECT incorporates a neuroleptic, as
combined ECT and neuroleptic treatment appears
to be more effective than either ECT alone or
neuroleptic alone.259,260 The neuroleptic chosen is
flupenthixol as this was the neuroleptic of choice
in the Chanpattana study.196
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
ECT
Discontinue
Continue
Haloperidol
Failure
Move to standard treatment
Discontinue
FIGURE 3 One-year treatment-resistant schizophrenia treatment model
Moderate schizophrenia
Inpatient setting
Clozapine
Continue
Moderate
p Success
Mild
Moderate
p Success
Mild
Moderate
p Success
Mild
Moderate
p Success
Mild
Hospital
p Discharge
Discharge
Hospital
p Discharge
Discharge
Hospital
p Discharge
Discharge
Hospital
p Discharge
Discharge
Hospital
p Discharge
Discharge
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
No relapse
Relapse
Health Technology Assessment 2005; Vol. 9: No. 9
63
Economic analysis
TABLE 19 Event probabilities
TABLE 21 Dosage and cost estimates
Variable
Estimate
(95% CI)
Success rate
Clozapine
ECT + neuroleptic
Chlorpromazine/haloperidol
0.65 (0.04 to 1.0)
0.57 (0.48 to 0.67)
0.04 (0.01 to 0.08)
Discontinue rate
Clozapine
ECT + neuroleptic
Chlorpromazine/haloperidol
0.05 (0.02 to 0.09)
0.26 (0.18 to 0.35)
0.05 (0.02 to 0.09)
Discharge if symptoms improve
0.81 (0 to 1)
Relapse within 1 year
Clozapine
ECT + neuroleptic
Chlorpromazine/haloperidol
0.16 (0 to 1) within
48 weeks
0.40 within
10 weeks
0.16 (0 to 1) within
48 weeks
Treatment
Dose
Cost
Clozapine
500 mg day–1
£9.78 per dose
Blood test
One per week
£25 per test
(18 weeks), one
per fortnight thereafter
ECT acute
Two sessions per
week for 4 weeks
£2475 per
six sessions
Flupenthixol
12 mg day–1
£0.60 per dose
ECT
maintenance
One session
per fortnight
£212.12 per
session
Flupenthixol
12 mg day–1
£0.60 per dose
Haloperidol
–1
£0.43 per dose
20 mg day
Hospital costs
£171 per day
At-home costs
£275 per year
TABLE 22 Cost-effectiveness results
TABLE 20 Quality of life utility estimates
Description
Average
utility rating
95% CI
0.82
0.76 to 0.88
Mild symptoms: community
Clozapine
0.91
Chlorpromazine
0.86
0.86 to 0.96
0.77 to 0.95
●
Mild symptoms: hospitalised patient
Clozapine
0.87
Chlorpromazine
0.84
0.82 to 0.92
0.75 to 0.93
●
Moderate symptoms:
hospitalised patient
Results
Table 22 shows the results from the decision model
assuming the central values for each parameter.
The results suggest that clozapine is the most costeffective treatment for patients with treatmentresistant schizophrenia since clozapine dominates
the other two strategies as it is cheaper and
generates more QALYs. ECT dominates the
chlorpromazine/haloperidol strategy. The results
show that ECT may be cost-effective compared
with the standard treatment of chlorpromazine/
haloperidol. These results suggest that ECT for
treatment-resistant schizophrenia may be a costeffective treatment for patients who do not
respond to clozapine.
64
Sensitivity analysis
The sensitivity of the model assumptions was
examined by undertaking a threshold analysis to
determine:
Treatment
Average
cost
QALYs
Clozapine
ECT
Chlorpromazine/haloperidol
£34,787
£55,267
£58,265
0.863
0.842
0.820
the parameter values for which ECT would be
the preferred strategy in the treatment of
treatment-resistant schizophrenia
the parameter values for which ECT would not
be the least preferred strategy in the treatment
of treatment-resistant schizophrenia.
Results of the analysis are shown in Tables 23
and 24.
Threshold analysis showed that ECT could not
become the cheapest treatment per QALY by just
altering any one of the ECT variable assumptions.
Even reducing the cost of ECT to zero on its own
would not alter the results sufficiently without also
reducing the cost of inpatient care from £171 to
£42. Altering the quality of life utility estimates
did not change the results sufficiently to make
ECT the preferred option, even if it was assumed
that the QALYs of patients following ECT were
higher than those for clozapine. For ECT to
become the preferred treatment strategy the one
variable that could realistically vary sufficiently to
change the results would be the probability of
clozapine success. The central default value is
0.65, or 65%. If this value were to fall below 21%
then ECT would become the preferred option, as
Health Technology Assessment 2005; Vol. 9: No. 9
TABLE 23 Threshold analysis for treatment-resistant schizophrenia: ECT as the preferred strategy
Variable
Baseline value
(95% CI)
Threshold
value
Direction of effect
Cost of clozapine
£9.78
£72.80
If the cost of clozapine rises above £72.80 then ECT would
be the preferred strategy. This would require over a sevenfold increase in cost
Adverse events for
clozapine
0.5 (0.02 to 0.09)
0.837
If the adverse events rate for clozapine rises above 83.7%
then ECT would be the preferred strategy. This is well
above its 95% CI
Probability of
clozapine success
0.65 (0.04 to 1.0)
0.21
If the probability of clozapine success falls below 21% then
ECT would be the preferred strategy. The 95% CI for this
variable is large, although 0.21 is towards the lower end
TABLE 24 Threshold analysis for treatment-resistant schizophrenia: ECT as the least preferred option
Variable
Baseline value
(95% CI)
Threshold
value
Direction of effect
Cost of ECT
£2475
£5900
If the cost of ECT rises to £5900 then ECT would be the
least preferred strategy
Adverse events for
ECT
0.26 (0.18 to 0.35)
0.87
If the adverse events rate for ECT rises above 87% then
ECT would be the least preferred strategy. This is well
above its 95% CI
Probability of ECT
success
0.57 (0.48 to 0.67)
0.26
If the probability of ECT success falls below 26% then ECT
would be the least preferred strategy
the cost per QALY of clozapine would increase
beyond £65,672. The 95% CIs for the probability
of clozapine success vary from 4 to 100% based on
the meta-analysis undertaken by Oh and
colleagues,258 and 21% lies within these limits.
Conclusions and recommendations
The cost-effective analysis using the model
presented here shows that clozapine for treatmentresistant schizophrenia is a cost-effective alternative
compared with ECT or chlorpromazine/haloperidol
treatment. The results of the model showed that
ECT was a cost-effective option compared with
chlorpromazine/haloperidol treatment. However,
the model shown here is based on limited data
owing to a lack of research in this area and cannot
be considered as robust. These results suggest that
ECT for treatment-resistant schizophrenia could
be effective in patients who do not respond well to
clozapine.
65
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 5
Implications for other parties
mplications for other parties are discussed
in Chapter 7.
I
67
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 6
Factors relevant to the NHS
CT is an intervention that has been used in
the NHS since its formation in 1948. Since
1985, the use of ECT in England has been
decreasing.8 The estimated 65,930 administrations
in 1999 compares with 105,466 reported
administrations in 1990/91 and 137,940 in 1985.8
E
Most administrations of ECT are provided on an
inpatient basis. In contrast, current government
policies such as the NSF on mental health41 advise
that the care and treatment of people with
psychiatric illness should be provided in
community settings.
69
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 7
Discussion
Summary of main results and
discussion
Depressive illness
Real versus sham ECT
The efficacy of real versus sham ECT is unclear.
The UK ECT Group51 found that in the short
term, real ECT is more effective than sham ECT
when data from all six trials were pooled. The
pooled effect size from the UK ECT Group review
was –0.91 (95% CI –1.27 to –0.54). An effect size
of 0.9 indicates that about 82% of patients treated
with ECT would be less depressed at the end of
treatment than the average patient treated with
sham ECT. The average size of the difference
between real and sham ECT on the HRSD was
9.7 points.
The present analysis of limited data from one trial
suggests that unilateral ECT is not more effective
than sham ECT (RR = 1, 95% CI 0.54 to 1.84).
Heterogeneous, dichotomous data from three
trials suggested that real bilateral ECT was also
not more effective than sham ECT (RR = 1.21,
95% CI 0.61 to 2.40) and homogeneous data from
two trials also suggested that real bilateral ECT
was not more effective than sham ECT (RR =
1.51, 95% CI 0.94 to 2.49). However, removal of
the trial52 that included one real ECT treatment in
the control group, leaving one trial,97 suggests that
real bilateral ECT is more effective than sham
ECT (RR = 1.98, 95% CI 1.05 to 3.73).
Only four out of nine trials provided sufficient
dichotomous data for this analysis, compared with
six out of seven providing sufficient continuous
data for analysis by the UK ECT Group. This may
explain, in part, the differences in the results of
the analyses. A further explanation for the
difference is the impact of stimulus parameters on
the effectiveness of ECT. When all the results are
pooled (as in the UK ECT Group analysis51), real
ECT is more effective than sham ECT. However,
when the results are analysed separately, real
unilateral ECT is not more effective than sham
ECT and it is unclear whether bilateral ECT is
more effective than sham ECT.
These trials also varied in other aspects of the
stimulus parameters used, such as the machine
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
used to administer the stimulus, the number of
ECT treatments administered, the dosage and the
waveform of the stimulus. Most of the trials were
conducted during the 1970s and 1980s, and in all
cases, the methods used to administer ECT do not
conform to current guidelines set by the RCP14 or
the APA.13 Five trials specified the machine used to
deliver ECT and none was of the type
recommended by current guidelines.13,14 Two used
Duopulse Mk IV machines,93,97 two used Ectron
Mk IV machines95,96 and one used a Transycon
machine.94 Of the seven trials that specified the
dosage and waveform of ECT, none used stimulus
dosing; rather, they gave a fixed dose. Two used
sine wave at 150 V,52,97 one used sine wave but did
not specify the dose,93 one used chopped sine wave
(dosage not specified),98 one used 60% sine wave at
400 V,96 one used a double-sided unrectified wave
at 40 J94 and only one used brief pulse at 10 J.95
Seizure threshold has been shown to vary 40-fold
between individuals, and to increase over the course
of ECT.12 Thus, it is possible that the dosages used
in these trials were below the minimum necessary to
induce a seizure of therapeutic efficacy, which is
likely to explain why unilateral ECT was not found
to be more effective than sham ECT.95 It has
subsequently been shown that the stimulus dose
needs to be increased to between five and six times
higher than seizure threshold for unilateral ECT to
equal bilateral ECT in efficacy.135
ECT versus antidepressant pharmacotherapy
Overall, the data suggest that ECT is more
effective than pharmacotherapy in the short term,
but the data on which this assertion is based are
subject to important flaws. The UK ECT Group51
found that ECT is more effective than drug
therapy in the short-term treatment of depression
(17 RCTs, 1136 participants). The pooled effect
size from the UK ECT Group review51 was –0.75
(95% CI –1.28 to –0.20), which indicates that
about 77% of patients treated with ECT would be
less depressed at the end of treatment than the
average patient treated with drug therapy. The
average size of the difference on the HRSD was
5.2 points. The present analysis of limited data
from one trial suggests that ECT is more effective
than SSRIs in the short term (RR = 3.41, 95% CI
1.39 to 7.11). The pooled analysis of data from six
trials suggests that ECT was also more effective
71
Discussion
than TCAs in the short term (RR = 1.42, 95% CI
1.17 to 1.72).
The analysis conducted by the UK ECT Group51
pooled data from trials comparing ECT with a
number of different antidepressant drugs
including SSRIs, TCAs, MAOIs and L-tryptophan.
The last two are not considered first line
treatments for depression in current clinical
practice.22 The present separate analysis of ECT in
comparison to SSRIs and TCAs found ECT to be
superior in both cases. However, the results of this
analysis need to be interpreted with some degree
of caution. Only one trial112 compared right
unilateral ECT with an SSRI (paroxetine). It was
unclear how participants were randomised or
whether the outcomes were rated blindly, but in
other respects the trial was of a reasonable quality.
The criterion for a response was defined a priori
(reduction of 50% on HRSD) and is similar to that
used to define response in trials of antidepressants.
Stimulus dosing was used and the dosage of
paroxetine (50 mg) was therapeutically adequate.
The quality of reporting in the 14 trials was largely
inadequate and only six trials (43%) provided data
for analysis. Thus, a large amount of data was
unusable, with consequent loss of power in the
analyses. Overall, the trials that provided data for
analysis were of low quality. Only one115 of the six
trials that contributed data for analysis used blinded
clinicians to rate outcomes; the remaining
five100,103,104,107,113 were not blind or the blinding
was not clear. This is of particular importance when
the method of judging responders is considered.
Two trials103,107 defined responders using different
criteria specified a priori based on scores from
quantitative outcome measures, while the remaining
four100,104,113,115 were based on clinical opinion of
improvement. Analysing the two trials based on a
quantitative assessment of improvement separately
results in no difference in the likelihood of being
defined as a responder between ECT and TCAs
(RR = 1.23, 95% CI 0.90 to 1.67, p = 0.58, n = 38).
However, the number of people included in this
analysis is very small and thus there is a low power
to detect any differences between ECT and TCAs.
Analysis of heterogeneous data from the four trials
based on clinical opinion gives a relative risk of
1.63 (95% CI 1.21 to 2.20, p = 0.001, n = 346) in
favour of ECT. This suggests that the method used
to define responders may have an important
influence on judgements of the efficacy of ECT
relative to antidepressant medication.
72
A further issue that may influence the relative
efficacy of ECT in comparison to pharmacotherapy
is the dosage of drugs used. Of the 15 trials that
compared ECT with either TCAs or SSRIs, one112
used a fully adequate therapeutic dose of SSRI
(50 mg paroxetine), but none used a fully
adequate dose up to 300 mg or equivalent of
imipramine. Two trials used 250 mg,106,113 one
used 220 g65 one used 200 mg100 and four used
150 mg.103,105,107,110 One trial102 used 100 mg, the
minimum dose shown to be therapeutically
effective, while two trials used doses below this
levels.99,104 Two trials did not state the dosage of
TCA used.101,115 Although most trials used a dose
of TCA above the minimally therapeutic dose,
none compared ECT to a dose of TCAs that would
normally be administered before ECT would be
considered in the case of treatment resistance.
It is also important to consider the extent to which
trial findings can be generalised to usual clinical
practice in terms of the characteristics of
participants included in the study and the ways in
which the interventions are delivered. In 15
studies the dosage of the ECT stimulus was not
specified and in 17 studies the type of ECT
machine used was not specified. It is therefore
very difficult to assess the extent to which the
administration of ECT used in these trials is
similar to current clinical practice. Of three trials
that did specify the stimulus dose used, one105
used a fixed dose of 110 V of alternating current,
whereas the other two used stimulus dosing at 2.5
times112 or 60 mC103 above seizure threshold. One
trial112 used an ECT machine that is in line with
current standards.13,14
Trials examining the efficacy of ECT have been
criticised for rarely reporting the number of
people who were initially screened before
inclusion in the trial, making it impossible to
assess whether the results apply to all or only a
fraction of patients seen in usual clinical
practice.260 A recent study has shown the ECT was
less effective in a ‘real-life’ heterogeneous patient
sample compared with homogeneous patient
samples used in RCTs.261 None of the trials
comparing ECT with pharmacotherapy provided
any information regarding the number of people
initially screened before entry into the trial.
Important parameters that influence current
clinical decisions regarding the use of ECT are
the severity of depression and treatment
resistance. Treatment resistance has been shown
to have an important impact on the efficacy of
ECT. Those who received an adequate dose of
antidepressant medication were less likely to
respond to ECT than those who had not received
an adequate dose of antidepressants.23
Health Technology Assessment 2005; Vol. 9: No. 9
In terms of inclusion criteria, three trials did not
specify inclusion criteria and eight did not use
explicit diagnostic criteria to diagnose or assess
the severity of depression.65,99–101,104,109,113,114 Of
these, five stated that the severity of depression
was severe enough to indicate the use of
ECT.99,104,109,113,114 The remaining three did not
state the severity of depression.65,100,101 Six trials
used explicit diagnostic criteria. Two used ICD-105
criteria for major depression,110,112 one111 used
DSM-III,262 one used DSM-IV,103 one used the
Feighner264 criteria102 and one107 used the criteria
specified by Klein.265 Four trials specified the
severity of depression for inclusion according to
the HRSD, with two103,105 specifying scores on the
17-item HRSD of less than 17, one111 specifying a
score of less than 20 and one specifying scores of
less than 22 on the 21-item HRSD.112
Four trials explicitly included people who were
treatment resistant to antidepressants.102,107,111,112
Two did not define treatment resistance.102,107 One
study111 defined treatment resistance as failure to
respond to a full course of TCAs, defined as at
least 150 mg of anitryptaline for at least 4 weeks
and failure of HRSD to drop by 40% or at least to
fall by 20 points. The other112 defined treatment
resistance as failure to respond to at least two
different antidepressants (including at least one
TCA) at a dosage of at least 100 g imipramine or
equivalent and no improvement for a total period
of 8 weeks. These definitions are both different,
and are different to that proposed by Nierenberg
and Amsterdam, defined as failure to respond to a
trial of more than one antidepressant drug in a
dose equivalent to 250–300 mg of imipramine
given for a duration of 6–8 weeks each.21 A further
five trials99,100,103,105,110 indicated that a certain
percentage of participants in the trial had been
treated with antidepressants during the current
episode, but did not state the dosages or type of
drugs used, or for how long the drugs had been
administered. None of the trials included people
for whom ECT was indicated as an emergency.
This suggests that nine trials included
participants who had severe depression and four
included people who were treatment resistant,
although none of the participants met the criteria
for treatment resistance specified by Nierenberg.21
None of the trials reported data separately for
older people.
Only one trial105 out of 18 administered ECT on
an outpatient basis; in the rest ECT was
administered on an inpatient basis. This is similar
to current clinical practice, where the majority of
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
ECTs are administered on an inpatient basis.33 In
contrast, current government policies such as the
NSF on mental health41 advise that the care and
treatment of people with psychiatric illness should
be provided in community settings.
ECT versus rTMS
Limited data from one trial including 40
participants indicated that ECT is significantly
more effective than rTMS in the short term. The
WMD was 6.8 points (95% CI 1.41 to 12.19) on
the HRSD in favour of ECT.
This treatment is not currently used in routine
clinical practice.
Adjunctive pharmacotherapy
Limited data from two separate trials suggest that
the efficacy of ECT may be improved by the
concomitant use of TCAs during the ECT course
(WMD = –3.00, 95% CI –5.65 to 0.35, n = 52;
RR = 0.55, 95% CI 0.33 to 0.90, n = 132) and
that the addition of pindolol may increase the
speed but not the extent of response to ECT.
Limited data suggest that continuing to take TCAs
following ECT does not reduce the risk of relapses
at 6 months (RR = 0.80, 95% CI 0.55 to 1.15,
p = 0.23, n = 100).
Not any of the participants in the 11 included
trials56–66 were specifically selected because they
had treatment-resistant depression. However,
many of the participants in the trials had
previously been treated with pharmacotherapy for
the current episode and had received ECT in the
past. In the Shiah trial,57 nine out of 35 (26%)
were treatment resistant. In Arfwidsson,58 42% of
participants had received antidepressant
medication during the current episode, in d’Elia60
39% had received antidepressants, and in
Lauritzen64 90% in the paroxetine group and 76%
in the placebo group had received antidepressants
during the current episode. The inferior response
of paroxetine-treated patients in group A and
imipramine patients in group B in this trial64
could reflect the fact that participants had failed to
respond to the same class of antidepressant
medication before ECT.266 Mayur and colleagues56
report that only half of the participants in either
group had received an adequate drug trial before
participation in the study. Depression was
diagnosed according to standardised criteria in
three trials, with Lauritzen using DSM-IIIR64 and
Shiah57 and Mayur56 using DSM-IV. The
remaining six trials did not use standardised
criteria to diagnose depression in their inclusion
criteria.
73
Discussion
Only one trial62 provided usable data to assess the
efficacy of continuation pharmacotherapy on
relapses and the interpretation of the results was
heavily influenced by the inclusion of those who
withdrew from the trial in the analysis. Eighteen
people dropped out from the treatment group
compared with seven in the control and assuming
that they all relapsed, this resulted in a nonstatistically significant difference in relapse rates
between the treatment arms. Given the high rates
of dropout in this study, these results should be
interpreted with caution.
Continuation pharmacotherapy
Limited data suggest that continuation
pharmacotherapy with tricyclic antidepressants
does not reduce the relapse rate in those who have
successfully responded to ECT (RR = 0.73, 95%
CI 0.53 to 1.01, p = 0.06, n = 56).68 However,
when TCAs were augmented with lithium there
was a statistically significant reduction on the rate
of relapses compared with placebo (RR = 0.58,
95% CI 0.39 to 0.86).
Electrode placement
In the short term, bilateral ECT is more effective
than unilateral ECT (27 RCTs, 1367 participants).
The pooled effect size from the UK ECT Group
review was –0.29 (95% CI –0.43 to –0.15), which
indicates that about 62% of patients treated with
bilateral ECT would be less depressed at the end
of treatment than the average patient treated with
unilateral ECT. The average size of the difference
on the HRSD was 3.4 points.
Dosage and frequency of administration
Higher dose ECT was more effective than lower
dose ECT (seven RCTs, 342 patients). The pooled
effect size from the UK ECT Group review51 was
–0.73 (95% CI –0.41 to –1.08), which indicates
that about 77% of patients treated with higher
dose ECT would be less depressed at the end of
treatment than the average patient treated with
lower dose ECT. The average size of the difference
on the HRSD was 5.2 points. Although the trials
differed in the precise doses used, there was a
consistent benefit for higher dose treatment.
There was no different in effectiveness between
twice weekly and three times weekly ECT (six
RCTs, 210 patients).
74
removal of one outlying trial resulted in no
difference between the two interventions on their
primary outcome measure of global improvement.
The UK ECT Group51 found that real ECT was no
more effective than sham ECT.
ECT versus antipsychotic drugs
The Cochrane Schizophrenia Group ECT review50
found that ECT alone was less effective than
antipsychotic medication. When ECT was added to
antipsychotic medication, there was no clear
difference between those treated with ECT in
addition to antipsychotic and those treated with
antipsychotics alone. Limited data from one trial
suggested an advantage of ECT antipsychotic
combination, but only in relation to mental state
as measured by the BPRS. The UK ECT Group51
found no advantage of ECT over antipsychotic
medication either alone or in combination with
antipsychotic medication.
ECT versus psychotherapy
The Cochrane Schizophrenia Group ECT review50
found limited evidence from one trial that ECT is
more effective than psychotherapy in both the
short and longer term, but that adding medication
to psychotherapy reverses the trend. There were
no trials comparing ECT with family therapy or
other psychosocial interventions.
Continuation ECT
Both reviews50,51 found limited evidence from one
trial to support the efficacy of maintenance ECT
added to antipsychotic medication in a population
who were medication resistant but who had
responded to a course of ECT by strict criteria.
The Cochrane Schizophrenia Group ECT review50
suggested that the NNT to prevent a relapse in
this population was two (95% CI 1.5 to 2.5).
Electrode placement
Neither review50,51 found evidence for a difference
between unilateral and bilateral ECT.
Schizophrenia
Dosage and frequency
The Cochrane Schizophrenia Group ECT review50
found limited data from one trial that suggested
that higher doses resulted in a faster rate of
improvement, but had no impact on the extent of
improvement compared with lower doses. No
conclusions can be drawn from the limited
evidence on the impact of the frequency of ECT.
Real versus sham ECT
The Cochrane Schizophrenia Group ECT review50
found a non-significant trend towards real ECT
being more effective than sham ECT. There was
considerable heterogeneity in the trials and
Generalisability of the trial evidence in
schizophrenia
The Cochrane Schizophrenia Group ECT review50
reported that there was considerable variation
Health Technology Assessment 2005; Vol. 9: No. 9
between trials in the clinical and demographic
profile of the participants, criteria used to
establish the diagnosis of schizophrenia and
methods of administering ECT. The APA3
recommends that ECT could be used when
patients are treatment resistant or in a catatonic
state and when the psychotic symptoms in the
current episode have an abrupt or a recent
onset.13 Similarly, the RCP32 advises that the
practical usefulness of ECT in schizophrenia is
limited to acute catatonic states, schizoaffective
disorders, acute paranoid syndromes and
people with type I schizophrenia who are either
intolerant of or unresponsive to a dose of a
neuroleptic equivalent to 500 mg of
chlorpromazine daily.
The Cochrane Schizophrenia Group ECT review50
found that the diagnosis of schizophrenia was
established using operationally defined criteria in
13 of the 24 trials, while the remainder diagnosed
the disorder by clinical consensus. Diagnostic
criteria used included ICD-9, ICD-10, DSM-IIIR,
DSM-IV, Feighner’s criteria, Present State
Examination (PSE) and CATEGO Research
Diagnostic Criteria, and the Chinese Medical
Council Clinical Diagnostic Criteria. Ungvari and
Petho185 classified participants based on the
classification of Leonhard267 into systematic and
unsystematic schizophrenia, a classification similar
to the process and reactive or non-process
classification of Langfeldt.268 Two trials included
people with homogeneous clinical subtypes of
schizophrenia, namely chronic catatonic
schizophrenia175 and paranoid schizophrenia.173
One trial174 included only young males with
schizophreniform disorder (a diagnosis made when
the symptoms of schizophrenia have been present
for less than the 6 months required for the
diagnosis of schizophrenia. If the symptoms persist
beyond 6 months this provisional diagnosis is
changed to schizophrenia). One trial189 included
12 people with unspecified psychosis among the
40 participants in the trial. None of the included
trials studied people with schizoaffective disorder,
which is one of the few indications for which
clinicians currently use ECT, according to a
recent survey.15
The Cochrane Schizophrenia Group ECT review50
found little homogeneity between trials in the
duration of the disorder, with seven trials
stipulating a duration of less than 2 years, of
which Abrams194 included participants with onset
of disorder less than 3 months and Sarkar and
colleagues174 less than 2 months. Seven trials
included participants who had been ill for more
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
than 2 years and two of these trials175,177 included
individuals with chronic illness hospitalised for
10 years or more, with the former including some
individuals who had been treated with leucotomy
as well. Seven trials included people with varying
duration of the disorder ranging from 1 month to
32 years. From the reports of Bagadia and
colleagues105 and Baker,195 it was unclear for how
long the participants had been ill.
In terms of past history of response to
antipsychotic drugs, the Cochrane Schizophrenia
Group ECT review50 found three trials180,196,269
that specifically included people with treatmentresistant schizophrenia that fulfilled modified
criteria for treatment-resistant schizophrenia.270
A further three trials172,173,195 also included
participants who had failed to respond to
antipsychotics, although it is uncertain how many
would meet stringent criteria for treatment
resistance. The review50 also reports that other
trials included people with varying degrees of
non-response to conventional antipsychotics,
although Abrams,194 Sarkar and colleagues174 and
possibly Ungvari and Petho185 included people
who were acutely ill and hence unlikely to be
resistant to treatment. One trial175 predominantly
included people with catatonia and one included
only people with paranoid schizophrenia.173
The Cochrane Schizophrenia Group ECT review50
also found considerable variation in the quality of
reporting of details of the administration of ECT.
Thirteen of the trials described that ECT was
modified, while seven appear to have used
unmodified ECT. It was unclear from three reports
whether ECT was modified.
The Cochrane Schizophrenia Group ECT review50
reported that five trials173,180,184,196,269 stated that
brief-pulse ECT devices were used; the remainder
appear to have used sine-wave machines. The
review50 found that the quality of reporting on
electrode placement, frequency and duration of
ECT administration was generally adequate in the
selected trials. With the exception of five studies
out of the 24, little information was provided in the
trial reports on methods used to ensure the
adequacy of treatments with ECT. Two studies196,269
titrated individual thresholds for participants and
monitored seizures with the cuff method and EEG
recordings. Two studies180,184 used suprathreshold
stimuli and monitored motor and electrical seizure
activity as above. One study174 used sine-wave
stimuli at settings sufficient to ensure seizures of
25 seconds or more, monitored by the cuff
method.
75
Discussion
Thus, it appears that many of the included trials
did not deliver ECT in line with currently
recommended standards,13,14 with reference to the
use of stimulus dosing and brief-pulse stimuli.
Mania
The UK ECT Group review51 found very limited
evidence of the efficacy of ECT in mania. They
were unable to draw any firm conclusions on the
use of ECT in this group.
Catatonia
Limited subgroup analyses by the Cochrane
Schizophrenia Group ECT review50 suggested that
ECT had no significant benefits in people with
catatonia. The poor quality of randomised
evidence does not allow firm conclusions to be
drawn regarding the relative efficacy of ECT in
this group.
Children and adolescents
The use of ECT in adolescents and children is
rare. This explains, in part, why there are no RCTs
of the efficacy of ECT in this group. The nonrandomised evidence did not allow firm
conclusions to be drawn regarding the efficacy of
ECT compared with other treatments. It suggests
that ECT is probably more effective in adolescents
or children with depression, mania or catatonia
than in schizophrenia. Studies rarely studied or
reported information on adverse events.
Older people
The UK ECT Group51 could not conduct reliable
subgroup analyses of the use of ECT in this group
and older people were not well represented in the
RCTs. The trials reviewed by the present group,
comparing real versus sham ECT and ECT versus
antidepressant medication, did not report results
separately for older people. Non-randomised
evidence of the use of ECT in older people with
depression was subject to difficulties with
confounding variables and information bias. It did
not provide consistent results, making it difficult
to draw any firm conclusions regarding the
efficacy of ECT in this group.
Pregnancy
76
There was no randomised evidence relating to
the use of ECT during or after pregnancy. At
the time of writing, non-randomised evidence
provides limited information on the rate of
complications only and suggests that the rate of
complications tends to be relatively low at around
1%. However, these figures should be interpreted
with caution because of the poor reporting in the
studies.
Long-term efficacy of ECT
Very few of the trials included in the UK ECT
Group review51 and the Cochrane Schizophrenia
Group ECT review50 assessed the efficacy of ECT
beyond the end of the course of ECT. It is
therefore not possible to determine for how long
the short-term benefits of ECT are maintained.
Evidence from the SURE review53 suggests that
there is a negative relationship between the length
of time since ECT and satisfaction with outcome,
such that satisfaction with treatment is reduced in
the longer term.
Adverse events: mortality
Trials in the UK ECT Group review51 did not
suggest that there was in increased risk of death
due to ECT. The short-term nature of the
trials meant that they did not provide any
evidence of the long-term impact ECT on
mortality rates. In the trials included by the
Cochrane Schizophrenia Group ECT review50
on schizophrenia, none of the 779 participants
died during or immediately after a course of ECT.
The non-randomised evidence from the
UK ECT Group review51 produced inconsistent
results and did not provide clear evidence
that ECT either increased or decreased death
rates.
Adverse events: cognitive functioning
The Cochrane Schizophrenia Group ECT review50
found limited evidence to suggest that greater
cognitive impairment occurs at the end of a course
of ECT than for antipsychotics in people with
schizophrenia. The UK ECT Group51 found it
difficult to summarise the data on the cognitive
effects of ECT. They included trials that measured
different aspects of cognitive functioning often
using instruments that had not been
psychometrically validated. Parallel forms of the
tests were rarely used and there was little
consistency in the types of instrument used across
studies. The trials also varied in the stimulus
parameters of ECT or did not report them,
making it very difficult to compare results across
studies.
Cognitive testing was often used in the trials in an
ad hoc way and as a result lacked a consistent
theoretical underpinning to predict and interpret
findings. Owing to the small sample sizes of many
trials, confounding factors were dealt with
inadequately and between-group comparisons
rarely controlled for baseline differences in
cognitive functioning. The analyses often used
multiple testing without provision with a high risk
of type I errors.
Health Technology Assessment 2005; Vol. 9: No. 9
As a result, no quantitative summary of the
findings on cognitive functioning could be
performed. Despite this, a number of conclusions
could be drawn from the findings of the UK ECT
Group review.51 Cognitive impairments following
ECT mostly reflect memory impairment. On the
whole, bilateral ECT resulted in greater memory
impairment than unilateral ECT, a higher dose of
ECT produced more impairment than lower doses
and administration of ECT three times a week
resulted in greater memory impairment than twice
a week.
This suggests that the stimulus parameters of
ECT have an important impact on cognitive
impairment following ECT. Limited evidence from
the UK ECT Group51 suggests that these
impairments do not last beyond 6 months. The
trials included in the UK ECT Group review51
rarely measured cognitive functioning beyond the
course of ECT and no trials assessed cognitive
functioning 12 months post-ECT.
Evidence from the SURE review53 suggested that a
significant proportion of people who receive ECT
report memory loss that persists for longer than
6 months. This review53 suggested a number
of reasons for the mismatch between patient
and clinical perspectives on memory loss as a
result of ECT. First, objective tests used in
RCTs rarely capture the type of memory
problems that occur most frequently in the
subjective reports of participants, such as
personal autobiographical memories. Second,
patients and clinicians interpret memory loss
in different ways. Patients see memory loss
as an important side-effect of ECT, whereas
clinicians may attribute memory loss to other
factors such as age and the symptoms of
depression. The extent to which trials attempted
to handle or discuss the interactions between
impairments in cognitive functioning as a
direct result of ECT, improvements in cognitive
functioning as a result of improvements in
depression and decreases in cognitive
functioning as a result of age is unclear from
the evidence presented in the UK ECT
Group51 or Cochrane Schizophrenia Group
ECT review.50
Adverse effects: brain damage
The UK ECT Group review51 found no evidence
from structural brain imaging studies that ECT
causes brain damage. Where moderate
abnormalities were detected at higher rates, they
were likely to be due to clinical factors such as
severity of illness.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Patient acceptability in choice
None of the RCTs included in the reviews
conducted by the Cochrane Schizophrenia
Group50 and the UK ECT Group51 explored the
impact of patient acceptability or choice on the
outcomes of ECT. The rate of discontinuations was
generally similar between ECT and other
comparison interventions.
Evidence from the SURE review53 suggests that
estimates of the perceived benefit of ECT are
influenced by the timing and methods used to
obtain this information. Studies that interview
consumers shortly after ECT are likely to
overestimate the perceived benefit, especially if
the interviews are conducted by a clinician in a
hospital setting using a brief set of questions. The
SURE review53 also suggests that perceived benefit
of ECT from the patient’s perspective is not
unidimensional but complex. Patients make tradeoffs between the benefits and risks of ECT that are
not the same for each individual. The review53
also argues that the patient’s perspective of the
perceived benefit of ECT are not adequately
captured by clinical measures assessing signs and
symptoms. A reduction in severity of symptoms on
a depression rating scale is not the same as
subjective relief from depression. These clinical
measures may also be subject to bias in people
reporting that they are better in order to avoid
further ECT, although this has not been explored
systematically.
The SURE review53 found mixed results
regarding the relationship between patient
choice and satisfaction with ECT. The review did
not find any studies that explored the impact
of patient choice on the perceived benefit
of ECT.
Patient information and consent
Evidence from the SURE review53 suggests that at
least 50% of users feel that they have inadequate
information before ECT and between 7 and
16% were judged to have adequate objective
knowledge about the procedure of ECT. They also
found that between one-quarter and one-third of
people who sign a consent form to ECT do so
under pressure, in the belief that they cannot
refuse.
Limited data from one small69 and one larger
trial92 suggested that patient information videos
do not improve patient knowledge of ECT.
In both trials, there were no statistically significant
differences between the two groups in either
the number of questions correctly answered69
77
Discussion
or mean knowledge score following the
intervention. However, the results of these trials
should be interpreted with caution. The
sample size in one trial was small and included
no baseline assessment of knowledge,69 and in
both trials69,92 the instrument used to measure
knowledge had not been psychometrically
tested.
Assumptions, limitations and
uncertainties
Comprehensiveness of the review
The present searches of the randomised evidence
and those included in the three good systematic
reviews were exhaustive and the authors are
confident that they have not missed any
important RCTs of ECT. They are less certain that
their searches of the non-randomised literature
were as comprehensive. They did not review
evidence concerning the different types of
anaesthesia or the impact of pretreatment with
caffeine on the efficacy of ECT. They also
did not examine adjunctive or post-treatments
that aimed to reduce the cognitive side-effects
of ECT.
Cost effectiveness modelling for
schizophrenia
This report includes the first attempt, to the
authors’ knowledge, of modelling the costeffectiveness of ECT for schizophrenia. The
robustness of the model was constrained by a lack
of data in this field. As such, the conclusions
should be interpreted with caution.
Need for further research
Clinical effectiveness
This review highlighted many areas where there is
a need for further research into the clinical
effectiveness and cost-effectiveness of ECT.
78
There is no good quality randomised evidence of
the effectiveness of ECT in specific subgroups that
are most likely to receive ECT. These included
older people with depression, women with
postpartum exacerbations of depression or
schizophrenia and people with catatonia. There is
also a lack of good quality randomised evidence of
the effectiveness of ECT in people with mania and
people who are resistant to pharmacotherapy in
schizophrenia and depression. There is a need for
further, high-quality RCTs of the use of ECT in
these populations.
There is currently no randomised evidence
comparing ECT with, or in addition to, newer
antipsychotic drugs (e.g. clozapine and
risperidone) and antidepressants (e.g. venlafaxine)
that are currently used in clinical practice. Further
work is needed in these areas. More research is
also needed to compare ECT with rTMS,
especially in people with schizophrenia. Again,
there is a need for further, high-quality RCTs
comparing the use of ECT with these
treatments.
More research is needed to examine the
long-term efficacy of ECT and the effectiveness
of post-ECT pharmacotherapy. There is only
limited evidence regarding the efficacy of
supplementing ECT with pharmacotherapy
in people with depression and the continuation
of pharmacotherapy following successful
response to ECT to prevent relapses. In most
trials, the aftercare of people receiving ECT was
not randomised and people were rarely
followed up beyond the course of ECT. Future
work in the area requires longer follow-up
periods. Further work is also needed to develop
ways of incorporating patients’ perspectives
on the impact of ECT into future RCTs.
Consideration should be given to the use of
both quantitative and qualitative methods.
The outcome measures used should reflect both
clinical and patient perspectives on the impact
of ECT.
There is also little good quality quantitative
evidence of the short-term and longer term
cognitive side-effects of ECT. Cognitive
functioning should be measured using wellvalidated instruments and methods need to be
developed that also reflect patients’ concerns
regarding personal memory loss. These
instruments should be incorporated into trial
design at the outset, and hypotheses set and
results interpreted using a well-developed
theory or set of theories from cognitive
psychology. Again, longer term follow-up is
needed as memory losses may only become
apparent in the longer term. There is also a need
for longer term follow-up within RCTs to explore
the impact of ECT on suicide and all-cause
mortality.
Further work is needed to examine the
information needs of people deciding whether to
accept ECT and how their decision-making can be
facilitated. The influence of these choices on the
perceived efficacy of ECT also requires further
exploration.
Health Technology Assessment 2005; Vol. 9: No. 9
Despite over 50 years of research into ECT, there
is still no agreement on the mechanism of action
of ECT. More research is needed in this area.
Finally, the quality of reporting of trials in this
area would be vastly improved by strict adherence
to the Consolidated Standards of Reporting Trials
(CONSORT) recommendations.
Cost-effectiveness
Further economic analysis, such as expected value
of perfect information, may identify areas in which
research would be best targeted by identifying
parameters where reducing the level of
uncertainty would have the most effect in helping
to make the decision on whether ECT is a costeffective treatment.
79
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Chapter 8
Conclusions
Clinical effectiveness
The results of this review largely relate to the use of
ECT within well-developed health services. In
people with depression, real ECT is probably more
effective than sham ECT, but stimulus parameters
have an important influence on efficacy; low-dose
unilateral ECT is no more effective than sham
ECT. ECT is probably more effective than
pharmacotherapy in the short term, but the
evidence on which this assertion is based was of
variable quality and inadequate doses of
pharmacotherapy were used. Limited evidence
suggests that ECT is more effective than rTMS.
Limited data suggest TCAs may improve the
antidepressant effect of ECT during the course of
ECT, and that continuation pharmacotherapy with
TCAs combined with lithium in people who have
responded to ECT reduces the rate of relapses.
Overall, gains in the efficacy of the intervention
depending on the stimulus parameters of ECT are
achieved only at the expense of an increased risk of
cognitive side-effects. Limited evidence suggests
that these effects do not last beyond 6 months, but
there is no evidence examining the longer term
cognitive effects of ECT. There is little evidence of
the long-term efficacy of ECT, much less evidence
regarding the efficacy of ECT in schizophrenia and
mania, and no randomised evidence of the
effectiveness of ECT in catatonia. ECT either
combined with antipsychotic medication or as a
monotherapy is not more effective than
antipsychotic medication in people with
schizophrenia. The evidence did not allow any firm
conclusions to be drawn regarding the efficacy of
ECT in people with mania or catatonia, older
people, younger people and women, or the impact
of ECT on all-cause mortality. There was limited
non-randomised evidence regarding the impact of
patient acceptability and choice on the outcomes of
ECT, and this produced mixed results.
Cost-effectiveness
Depression
No previous analysis has been undertaken on the
cost-effectiveness of ECT in depression. The
model described here attempted to reflect the
possible treatment protocols that could be used in
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
treating severely depressed patients who require
hospitalisation through devising different
treatment scenarios. Different treatment scenarios,
which are based on ECT being provided as a first,
second, or third line therapy, have been compared
with a pharmacological-only therapy.
The results from the model are not conclusive
regarding the cost-effectiveness of ECT. Based on
the default assumptions the economic modelling
results suggest that ECT provided as a second line
therapy is the preferred treatment strategy.
However, the confidence intervals around the results
are large, primarily because of the large confidence
intervals around the inputs due to a lack of good
quality clinical evidence. The clinical evidence
seems to suggest that ECT is an effective treatment,
although there is no evidence of ongoing
antidepressant action beyond the duration of the
course of treatment. ECT needs to be followed by
pharmacological treatment or maintenance ECT to
maintain improvement, and the limited evidence
seems to suggest that the relapse rates of patients
following ECT even with maintenance therapy are
higher than the relapse rates of patients who have
received pharmacological therapy. This is reflected
in the model, which suggests that if an effective
treatment could be found that reduces the relapse
rates of patients following ECT, ECT would become
a cost-effective treatment in hospitalised, severely
depressed people.
Schizophrenia
No previous analysis has been undertaken of the
cost-effectiveness of ECT in schizophrenia. The
economic model constructed for schizophrenia was
based on a pharmacological model constructed by
Oh and colleagues258 which was the only cost–utility
study identified in the treatment of schizophrenia.
This model analysed the cost-effectiveness of
clozapine compared with haloperidol/
chlorpromazine treatment in treatment-resistant
schizophrenia. The results of the adapted model
including ECT suggest that clozapine is a costeffective treatment compared with ECT. However,
for patients who fail to respond to clozapine, ECT
may be preferred to the comparative treatment of
haloperidol/chlorpromazine. However, the clinical
evidence underpinning the ECT assumptions in the
model is weak.
81
Health Technology Assessment 2005; Vol. 9: No. 9
Acknowledgements
P
aul Birkett, Clinical Lecturer in Psychiatry,
Charlie Brooker, Professor of Psychiatry,
both in Sheffield, and Simon Gilbody, Lecturer
in Clinical Psychiatry, and David Cottrell,
Professor of Child and Adolescent Psychiatry,
provided clinical advice and guidance. Clive
Adams gave clinical and methodological guidance
and provided the group with the raw data from
the recently updated Cochrane Group Review of
ECT in schizophrenia. The UK ECT group
provided the group with a copy of their report
before publication. Suzy Paisley provided guidance
on literature searching and proof-reading, and in
the production of the report. Suzy Paisley, Ron
Akehurst and Jim Chilcott (ScHARR), Dr Niall
Moore, Consultant Psychiatrist, Bristol, Dr
Douglas Gee, Consultant Psychiatrist,
Humberside, Sarah Garner and Tina Eberstein
(NICE) provided comments on the initial draft of
the report.
The ScHARR Technology Assessment Group
(ScHARR-TAG) synthesises research on the
effectiveness and cost-effectiveness of healthcare
interventions for the NHS R&D Health
Technology Assessment Programme on behalf of a
range of policy makers, including the National
Institute of Clinical Excellence. ScHARR-TAG is
part of a wider collaboration of six units from
other regions. The other units are: Southampton
Health Technology Assessment Centre (SHTAC),
University of Southampton; Aberdeen Health
Technology Assessment Group (Aberdeen HTA
Group), University of Aberdeen; Liverpool
Reviews & Implementation Group (LRiG),
University of Liverpool; Peninsular Technology
Assessment Group (PenTAG), University of Exeter;
NHS Centre for Reviews and Dissemination,
University of York; and West Midlands Health
Technology Assessment Collaboration
(WMHTAC), University of Birmingham.
About ScHARR
Contributions of authors
The School of Health and Related Research
(ScHARR) is one of the four Schools that
comprise the Faculty of Medicine at the
University of Sheffield. ScHARR brings together a
wide range of medical- and health-related
disciplines including public health, general
practice, mental health, epidemiology, health
economics, management sciences, medical
statistics, operational research and information
science. It includes the Sheffield unit of the
Trent Institute for Health Services Research,
which is funded by NHS R&D to facilitate
high-quality health services research and capacity
development.
Joanne Greenhalgh and Daniel Hind carried out
the review of clinical effectiveness. Chris Knight
carried out the review of cost-effectiveness.
Catherine Beverley carried out the electronic
searches. Stephen Walters provided statistical
advice. Joanne Greenhalgh is responsible for the
report as lead author.
This report was commissioned by the NHS R&D
HTA programme. All responsibility for the content
of the report remains with the authors. None of
the authors has any financial interests in the
companies producing or marketing ECT
machines.
83
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
References
1.
Fink M. Convulsive therapy: a review of the first
55 years. J Affect Disord 2001;63:1–15.
2.
Sterling P. ECT damage is easy to find if you look
for it. Nature 2000;403:242.
3.
American Psychiatric Association. Practice
guideline for the treatment of patients with
schizophrenia. Am J Psychiatry 1997;
154(4 Suppl):1–63.
4.
American Psychiatric Association. Diagnostic and
statistical manual of mental disorders. 4th ed.
Washington, DC: APA; 2000.
16.
Fergusson G, Cullen L, Freeman C, Hendry J.
Electroconvulsive therapy in Scottish clinical
practice: a national audit of demographics,
standards and outcome. Journal of ECT
2004;20:166–73.
17.
Adler DA, Gorelick DA, Rummans TA, Bell C,
Greene JA, Shapiro HL, et al. Practice guideline
for the treatment of patients with major depressive
disorder (Revision). Am J Psychiatry 2000;
157(4 Suppl.):1–45.
18.
Einarson TR, Arikian S, Casciano J, Doyle JJ.
Comparison of extended-release venlafaxine,
selective serotonin reuptake inhibitors, and
tricyclic antidepressants in the treatment of
depression: a meta-analysis of randomized
controlled trials. Clin Ther 1999;21:296–308.
5.
World Health Organization. International
classification of diseases. 10th revision. Geneva:
WHO; 1992.
6.
Kruger S, Braunig P. Catatonia in affective
disorder: new findings and a review of the
literature. CNS Spectr 2000;5(7):48–53.
19.
Singleton N, Bumpstead R, O’Brien M, Lee A,
Meltzer HY. Psychiatric morbidity among adults living
in private households, 2000: summary report. London:
Office for National Statistics; 2002.
Fava M, Davidian KG. Definition and
epidemiology of treatment resistant depression.
Psychiatr Clin North Am 1996;19:179–200.
20.
Blashki TG, Mowbry R, Davies B. A controlled
trial of amitryptaline in general practice. BMJ
1971;i:133–8.
21.
Nierenberg AA, Amsterdam JD. Treatment
resistant depression: definition and treatment
approaches. J Clin Psychiatry 1990;
51(Suppl 6):39–47.
22.
American Psychiatric Association. Practice
guidelines for the treatment of patients with major
depressive disorder. Am J Psychiatry 2000;
157(4 Suppl):1–45.
23.
Sackeim HA, Prudic J, Devanand DP, Decina P,
Kerr B, Malitz S. The impact of medication
resistance and continuation pharmacotherapy on
relapse following response to electroconvulsive
therapy in major depression. J Clin Psychopharmacol
1990;10:96–104.
24.
Loo C, Mitchell P, Sachdev P MB, Parker G,
Gandevia S. A double blind controlled
investigation on transcranial magnetic stimulation
for the treatment of resistant major depression.
Am J Psychiatry 1999;156:946–8.
25.
Joy C, Adams CE, Laurie SM. Haloperidol vs
placebo for schizophrenia. In The Cochrane Library
(Issue 2). Oxford: Update Software; 2001.
26.
Thornley B, Adams CE, Awad G. Chlorpromazine
versus placebo for schizophrenia. In The
Cochrane Library (Issue 4). Oxford: Update
Software; 2001.
7.
8.
Office for National Statistics. Key health statistics
from general practice 1998. MB6 No. 2. London:
ONS; 2000.
9.
Hale A. ABC of mental health: depression. BMJ
1997;315:43–6.
10.
Clinical Standards Advisory Group Committee on
Schizophrenia. Schizophrenia, Vol. 1. London:
HMSO; 1995.
11.
Devanand DP, Verma AK, Tirumalasetti F,
Sackeim HA. Absence of cognitive impairment
after more than 100 lifetime ECT treatments.
Am J Psychiatry 1991;148:929–32.
12.
Sackeim HA, Devanand DP, Prudic J. Stimulus
intensity, seizure threshold, and seizure duration:
impact on the efficacy and safety of
electroconvulsive therapy. Psychiatr Clin North Am
1991;14:803–44.
13.
American Psychiatric Association. The practice of
electroconvulsive therapy: recommendations for
treatment, training and privileging. 2nd ed.
Washington, DC: APA; 2001.
14.
Freeman C. The ECT handbook. London: Royal
College of Psychiatrists; 1995.
15.
Benbow SM, Tench D, Darvill SP. Electroconvulsive
therapy practice in north-west England. Psychiatric
bulletin 1998;22:226–9.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
85
References
27.
28.
Jadad AR, Moore A, Carroll D, Jenkinson C,
Reynolds DJM, Gavaghan DJ, et al. Assessing the
quality of reports of randomised clinical trials: is
blinding necessary? Control Clin Trials 1996;17:1–12.
44.
Clarke M, Oxman AD. Cochrane Collaboration
handbook. In The Cochrane Library (Issue 3).
Oxford: Update Software; 2002.
45.
Levine M, Walter S, Lee H, Haines T, Holbrook A,
Moyer V. Users’ guides to the medical literature.
IV. How to use an article about harm. EvidenceBased Medicine Working Group. JAMA 1994;
271:1615–19.
46.
Popay J, Rogers A, Williams G. Rationale and
standards in the systematic review of qualitative
literature in health services research. Qualitative
Health Research 1998;8:341–51.
47.
Drummond MF, Jefferson TO. Guidelines for
authors and peer reviewers of economic
submissions to the BMJ. The BMJ Economic
Evaluation Working Party. BMJ 1996;313:275–83.
48.
DerSimonian R, Laird N. Meta-analysis in clinical
trials. Control Clin Trials 1986;7:177–88.
49.
Duffett R, Lelliott P. Auditing electroconvulsive
therapy: the third cycle. Br J Psychiatry 1998;
172:401–5.
Marshall M, Lockwood A, Bradley C, Adams C,
Joy C, Fenton M. Unpublished rating scales: a
major source of bias in randomised controlled
trials of treatments for schizophrenia. Br J
Psychiatry 2000;176:249–52.
50.
Duffett R, Hill P, Lelliot P. Use of electroconvulsive
therapy in young people. Br J Psychiatry 1999;
175:228–30.
Tharyan P, Adams CE. Electroconvulsive therapy
for schizophrenia. Cochrane Database of Systematic
Reviews (2): CD000076, 2002.
51.
Royal College of Psychiatrists. The Royal College
of Psychiatrists’ memorandum on the use of
electroconvulsive therapy. Br J Psychiatry 1977;
131:261–72.
UK ECT Group. Efficacy and safety of
electroconvulsive therapy in depressive disorders:
a systematic review and meta-analysis. The Lancet
2003;361:799–808.
52.
Jagadeesh HN, Gangadhar BN, Janakiramaiah N,
Subbakrishna DK, Jain S. Time dependent
therapeutic effects of single electroconvulsive
therapy (ECT) in endogenous depression. J Affect
Disord 1992;24:291–5.
53.
Rose D, Fleischmann P, Wykes T, Leese M,
Bindman J. Patients’ perspectives on
electroconvulsive therapy: systematic review. BMJ
2003;326(7403):1363.
54.
Grunhaus L, Dannon PN, Schreiber S,
Dolberg OH, Amiaz R, Ziv R, et al. Repetitive
transcranial magnetic stimulation is as effective as
electroconvulsive therapy in the treatment of
nondelusional major depressive disorder: an open
study. Biol Psychiatry 2000;47:314–24.
55.
Pridmore S. Substitution of rapid transcranial
magnetic stimulation treatments for
electroconvulsive therapy treatments in a course of
electroconvulsive therapy. Depress Anxiety 2000;
12:118–23.
Sachs GS, Printz DJ, Kahn DA, Carpenter D,
Docherty JP. The Expert Consensus Guideline
Series: Medication treatment of bipolar disorder
2000 [review]. Postgrad Med 2000; (Special Issue):
1–104.
Ungvari GS, Leung SK, Wai KT, Ng FS.
The pharmacological treatment of catatonia: an
overview. Eur Arch Psychiatry Clin Neurosci 2001;
251(Suppl 1):31–4.
30.
Weiner R. Treatment optimisation with ECT.
Psychopharmacol Bull 1994;30:313–20.
32.
33.
34.
35.
36.
Evidence Based Medicine Working Group. JAMA
1994;272:1367–71.
43.
29.
31.
86
Wahlbeck K, Cheine M, Essali MA. Clozapine
versus typical neuroleptic medication in
schizophrenia. In The Cochrane Library (Issue 4).
Chichester: John Wiley, 2004.
Sackeim HA, Prudic J, Devanand DP, Kiersky JE,
Fitzsimons L, Moody BJ, et al. Effects of stimulus
intensity and electrode placement on the efficacy
and cognitive effects of electroconvulsive therapy.
N Engl J Med 1993;328:839–46.
Freeman CP. The ECT handbook: the second report of
the Royal College of Psychiatrists’ Special Committee on
ECT. London: Royal College of Psychiatrists; 1995.
Department of Health. Electro-convulsive therapy:
survey covering the period January 1999 to March
1999, England. London: Department of Health;
2002.
37.
Pippard J, Ellam L. Electroconvulsive therapy in
Great Britain: a report to the college. Br J
Psychiatry 1981;139:563–8.
38.
Pippard J. Audit of electroconvulsive treatment in
two National Health Service regions. Br J Psychiatry
1992;160:621–37.
39.
Hillam J, Thompsell A, Tobiansky R.
Administration of ECT by trainee psychiatrists.
An audit of supervision, adequacy of tuition and
attitudes. Psychiatric Bulletin 1997;21:217–20.
40.
Duffett R, Lelliot P. Junior doctors’ training in the
theory and the practice of electroconvulsive
therapy. Psychiatric Bulletin 1997;21:563–5.
41.
Department of Health. National framework for mental
health. London: Department of Health; 1999.
42.
Oxman AD, Cook DJ, Guyatt GH. Users’ guides to
the medical literature. VI. How to use an overview.
Health Technology Assessment 2005; Vol. 9: No. 9
56.
57.
Mayur PM, Gangadhar BN, Subbakrishna DK,
Janakiramaiah N. Discontinuation of
antidepressant drugs during electroconvulsive
therapy: a controlled study. J Affect Disord 2000;
58:37–41.
Shiah IS, Yatham LN, Srisurapanont M, Lam RW,
Tam EM, Zis AP. Does the addition of pindolol
accelerate the response to electroconvulsive
therapy in patients with major depression? A
double-blind, placebo-controlled pilot study. J Clin
Psychopharmacol 2000;20:373–8.
58.
Arfwidsson L, Arn L, Beskow J, d’Elia G, Laurell B,
Ottosson JO, et al. Chlorpromazine and the antidepressive efficacy of electroconvulsive therapy.
Acta Psychiatr Scand 1973;49:580–7.
59.
Kirkegaard C, Moller SE, Bjorum N. Addition of
L-tryptophan to electroconvulsive treatment in
endogenous depression. A double-blind study. Acta
Psychiatr Scand 1978;58:457–62.
60.
d’Elia G, Lehmann J, Raotma H. Evaluation of the
combination of tryptophan and ECT in the
treatment of depression. I. Clinical analysis. Acta
Psychiatr Scand 1977;56:303–18.
61.
Seager CP, Bird RL. Imipramine with electrical
treatment in depression: a controlled trial. Journal
of Mental Science 1962;108:704–7.
62.
Imlah NW, Ryan E, Harrington JA. The influence
of antidepressant drugs on the response to
electroconvulsive therapy and on subsequent
relapse rates. Neuropharmacology 1965;4:438–42.
63.
64.
Kay DW, Fahy T, Garside RF. A seven-month
double-blind trial of amitriptyline and diazepam
in ECT-treated depressed patients. Br J Psychiatry
1970;117:667–71.
Lauritzen L, Odgaard K, Clemmesen L, Lunde M,
Ohrstrom J, Black C, et al. Relapse prevention by
means of paroxetine in ECT-treated patients with
major depression: a comparison with imipramine
and placebo in medium-term continuation
therapy. Acta Psychiatr Scand 1996;94:241–51.
65.
Wilson IC, Vernon JT, Guin T. A controlled study
of treatments of depression. Journal of
Neuropsychiatry 1963;4:331–7.
66.
Coppen A, Abou-Saleh MT, Milln P, Bailey J,
Metcalfe M, Burns BH, et al. Lithium continuation
therapy following electroconvulsive therapy. Br J
Psychiatry 1981;139:284–7.
67.
68.
Grunhaus L, Hirschman S, Dolberg OT,
Schreiber S, Dannon PN. Coadministration of
melatonin and fluoxetine does not improve the
3-month outcome following ECT. J ECT 2001;
17:124–8.
Sackeim HA, Haskett RF, Mulsant BH, Thase ME,
Mann JJ, Pettinati HM, et al. Continuation
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
pharmacotherapy in the prevention of relapse
following electroconvulsive therapy: a randomized
controlled trial. JAMA 2001;285:1299–307.
69.
Westreich L, Levine S, Ginsburg P, Wilets I.
Patient knowledge about electroconvulsive
therapy: effect of an informational video.
Convulsive Therapy 1995;11:32–7.
70.
Rey JM, Walter G. Half a century of ECT use in
young people [review]. Am J Psychiatry 1997;
154:595–602.
71.
Walter G, Rey JM, Mitchell PB. Practitioner review:
Electroconvulsive therapy in adolescents [review].
J Child Psychol Psychiatry 1999;40:325–34.
72.
Cohen D, Taieb O, Flament M, Benoit N,
Chevret S, Corcos M, et al. Absence of cognitive
impairment at long-term follow-up in adolescents
treated with ECT for severe mood disorder. Am J
Psychiatry 2000;157:460–2.
73.
Rubin EH, Kinscherf DA, Wehrman SA. Response
to treatment of depression in the old and very old.
J Geriatr Psychiatry Neurol 1991;4:65–70.
74.
Rubin EH, Kinscherf DA, Figiel GS, Zorumski CF.
The nature and time course of cognitive side
effects during electroconvulsive therapy in the
elderly. J Geriatr Psychiatry Neurol 1993;6:78–83.
75.
Manly DT, Oakley SP Jr, Bloch RM.
Electroconvulsive therapy in old-old patients. Am J
Geriatr Psychiatry 2000;8:232–6.
76.
Kroessler D, Fogel BS. Electroconvulsive therapy
for major depression in the old. Am J Geriatr
Psychiatry 1993;1:30–7.
77.
Philibert RA, Richards L, Lynch CF, Winokur G.
Effect of ECT on mortality and clinical outcome in
geriatric unipolar depression. J Clin Psychiatry
1995;56:390–4.
78.
Hawkins JM, Archer KJ, Strakowski SM, Keck PE.
Somatic treatment of catatonia. Int J Psychiatry Med
1995;25:345–69.
79.
Bush G, Fink M, Petrides G, Dowling F, Francis A.
Catatonia 2. Treatment with lorazepam and
electroconvulsive therapy. Acta Psychiatr Scand
1996;93:137–43.
80.
Malur C, Pasol E, Francis A. ECT for prolonged
catatonia. J ECT 2001;17:55–9.
81.
Miller LJ. Use of electroconvulsive therapy during
pregnancy. Hospital and Community Psychiatry 1994;
45:444–50.
82.
Bhatia SC, Baldwin SA, Bhatia SK.
Electroconvulsive therapy during the third
trimester of pregnancy. J ECT 1999;15:270–4.
83.
Echevarria Moreno M, Martin Munoz J, Sanchez
Valderrabanos J, Vazquez Gutierrez T.
87
References
Electroconvulsive therapy in the first trimester of
pregnancy. J ECT 1998;14:251–4.
84.
Polster DS, Wisner KL. ECT-induced premature
labor: a case report. J Clin Psychiatry 1999;60:53–4.
85.
Greenfield S, Nelson EC. Recent developments
and future issues in the use of health status
assessment measures in clinical settings. Med Care
1992;30(5 Suppl):MS23–41.
86.
87.
Hamilton M. A rating scale for depression.
J Neurol Neurosurg Psychiatry 1960;23:56–62.
88.
McDowell I, Newell C. Measuring health: a guide to
rating scales and questionnaires. 2nd ed. Oxford:
Oxford University Press; 1996.
89.
Faries D, Herrera J, Rayamajhi J, DeBrota D,
Demitrack MPWZ. The responsiveness of the
Hamilton Depression Scale. J Psychiatr Res 2000;
34:3–10.
90.
Hedges LV, Olkin I. Statistical methods for metaanalysis. London: Academic Press, 1985.
91.
Einarson TR, Addis A, Mittmann N, Iskedjian M.
Meta-analysis of venlafaxine, SSRIs and TCAs in
the treatment of major depressive disorder. Can J
Clin Pharmacol 1998;5:205–16.
92.
Battersby M, Ben Tovim D, Eden J.
Electroconvulsive therapy: a study of attitudes and
attitude change after seeing an educational video.
Aust NZ J Psychiatry 1993;27:613–19.
93.
Gregory S, Shawcross CR, Gill D. The Nottingham
ECT Study. A double-blind comparison of
bilateral, unilateral and simulated ECT in
depressive illness. Br J Psychiatry 1985;146:520–4.
94.
West ED. Electric convulsion therapy in
depression: a double-blind controlled trial. BMJ
1981;282:355–7.
95.
96.
88
d’Elia G, Frederiksen SO, Raotma H, Widepalm K.
Comparison of fronto-frontal and temporoparietal unilateral ECT. Acta Psychiatr Scand 1977;
56:233–9.
Lambourn J, Gill D. A controlled comparison of
simulated and real ECT. Br J Psychiatry 1978;
133:514–19.
Freeman CP, Basson JV, Crighton A. Double-blind
controlled trial of electroconvulsive therapy
(E.C.T.) and simulated E.C.T. in depressive illness.
Lancet 1978;i:738–40.
97.
Johnstone EC, Deakin JF, Lawler P, Frith CD,
Stevens M, McPherson K, et al. The Northwick
Park electroconvulsive therapy trial. Lancet 1980;
ii:1317–20.
98.
Brandon S, Cowley P, McDonald C, Neville P,
Palmer R, Wellstood-Eason S. Electroconvulsive
therapy: results in depressive illness from the
Leicestershire trial. BMJ 1984;288:22–5.
99.
McDonald IM, Perkins M, Marjerrison G,
Podilsky M. A controlled comparison of
amitriptyline and electroconvulsive therapy in the
treatment of depression. Am J Psychiatry
1966;122:1427–31.
100. Shepherd M. Clinical trial of the treatment of
depressive illness. BMJ 1965;i:881–6.
101. Kendrick DC, Parboosingh R, Post F. A synonym
learning test for use with elderly psychiatric
subjects: a validation study. British Journal of Social
and Clinical Psychology 1965;4:63–71.
102. Davidson J, McLeod M, Law-Yone B, Linnoila M.
A comparison of electroconvulsive therapy and
combined phenelzine–amitriptyline in refractory
depression. Arch Gen Psychiatry 1978;35:639–42.
103. Janakiramaiah N, Gangadhar BN, Naga
Venkatesha Murthy PJ, Harish MG, Subbakrishna
DK, Vedamurthachar A. Antidepressant efficacy of
Sudarshan Kriya yoga (SKY) in melancholia: a
randomized comparison with electroconvulsive
therapy (ECT) and imipramine. J Affect Disord
2000;57:255–9.
104. Bruce EM, Crone N, Fitzpatrick G. A comparative
trial of ECT and Tofranil. Am J Psychiatry 1960;
117:76.
105. Bagadia VN, Shah LP, Pradhan PV, Doshi J,
Abhyankar RR. Evaluation of cognitive effects of
ECT: preliminary observations. Indian Journal of
Psychiatry 1981;23:324–9.
106. Hutchinson J, Smedberg D. Treatment of
depression: a comparative study of ECT and six
drugs. Br J Psychiatry 1963;109:536–8.
107. Steiner M, Radwan M, Elizur A. Failure of
L-triiodothyronine (T3) to potentiate tricyclic
antidepressant response. Current Therapeutic
Research 1978;23:655–9.
108. MacSweeney DA. Treatment of unipolar
depression [Letter]. Lancet 1975;ii:510–11.
109. Herrington RN, Bruce A, Johnstone EC.
Comparative trial of L-tryptophan and E.C.T. in
severe depressive illness. Lancet 1974;ii:731–4.
110. Gangadhar BN, Kapur RL, Kalyanasundaram S.
Comparison of electroconvulsive therapy with
imipramine in endogenous depression: a double
blind study. Br J Psychiatry 1982;141:367–71.
111. Dinan TG, Barry S. A comparison of
electroconvulsive therapy with a combined lithium
and tricyclic combination among depressed
tricyclic nonresponders. Acta Psychiatr Scand 1989;
80:97–100.
112. Folkerts HW, Michael N, Tolle R, Schonauer K,
Mucke S, Schulze-Monking H. Electroconvulsive
therapy vs. paroxetine in treatment-resistant
depression – a randomized study. Acta Psychiatr
Scand 1997;96:334–42.
Health Technology Assessment 2005; Vol. 9: No. 9
113. Greenblatt M, Grosser G. Differential response of
hospitalised depressed patients to somatic therapy.
Am J Psychiatry 1964;120:935–43.
129. Valentine M, Keddie KM, Dunne D. A comparison
of techniques in electro-convulsive therapy. Br J
Psychiatry 1968;114:989–96.
114. Stanley W, Fleming H. A clinical comparison of
phenelzine and electro-convulsive therapy in the
treatment of depressive illness. Journal of Mental
Science 1962;108:708–10.
130. Heshe J, Roder E, Theilgaard A. Unilateral and
bilateral ECT. A psychiatric and psychological
study of therapeutic effect and side effects. Acta
Psychiatr Scand 1978; Suppl 275:1–180.
115. Robin A, Harris JA. A controlled comparison of
imipramine and electroplexy. Br J Psychiatry 1962;
108:217–19.
131. Carney MW, Rogan PA, Sebastian J, Sheffield B.
A controlled comparative trial of unilateral and
bilateral sinusoidal and pulse E.C.T. in
endogenous depression. Physicians Drug Manual
1976;7(9–12):77–9.
116. Gangadhar BN. Side effects of somatic therapies
in depression: a double blind comparison of ECT
and imipramine. National Institute of Mental Health
and Neurosciences Journal 1985;3:13–16.
117. Abrams R, Swartz CM, Vedak C. Antidepressant
effects of high-dose right unilateral
electroconvulsive therapy. Arch Gen Psychiatry 1991;
48:746–8.
118. d’Elia G. Comparison of electroconvulsive therapy
with unilateral and bilateral stimulation. II:
Efficacy in endogenous depression. Acta Psychiatr
Scand 1970;215:30–43.
119. Levy R. The clinical evaluation of unilateral
electroconvulsive therapy. Br J Psychiatry 1968;
114:459–63.
120. Weiner RD, Rogers HJ, Davidson JR, Squire LR.
Effects of stimulus parameters on cognitive side
effects. Ann N Y Acad Sci 1986;462:315–25.
121. Abrams R, DeVito RA. Clinical efficacy of
unilateral ECT. Diseases of the Nervous System 1969;
30:262–3.
122. Martin RW. Clinical evaluation of unilateral EST.
Am J Psychiatry 1965;121:1087–90.
123. Papakostas Y, Stefanis C, Sinouri A, Trikkas G.
Increases in prolactin levels following bilateral and
unilateral ECT. Am J Psychiatry 1984;141:1623–4.
124. Krystal AD, Weiner RD, Coffey CE, Smith P. EEG
evidence of more ‘intense’ seizure activity with
bilateral ECT. Biol Psychiatry 1992;31:617–21.
125. Costello CG, Belton GP. The amnesic and
therapeutic effects of bilateral and unilateral ECT.
Br J Psychiatry 1970;116:69–78.
126. Daniel WF, Crovitz HF. Disorientation during
electro convulsive therapy. Technical, theoretical
and neuropsychological issues. Ann N Y Acad Sci
1986;462:293–306.
127. Malitz S, Sackeim HA, Decina P, Kanzler M,
Kerr B. The efficacy of electroconvulsive therapy:
dose–response interactions with modality. Ann N Y
Acad Sci 1986;462:56–64.
128. Letemendia FJ, Delva NJ, Rodenburg M,
Lawson JS, Inglis J, Waldron JJ, et al. Therapeutic
advantage of bifrontal electrode placement in
ECT. Psychol Med 1993;23:349–60.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
132. Sackeim HA, Decina P, Kanzler M, Kerr B,
Malitz S. Effects of electrode placement on the
efficacy of titrated, low-dose ECT. Am J Psychiatry
1987;144:1449–55.
133. Stromgren LS, Juul-Jensen P. EEG in unilateral
and bilateral electroconvulsive therapy. Acta
Psychiatr Scand 1975;51:340–60.
134. Fraser RM, Glass IB. Unilateral and bilateral ECT
in elderly patients. A comparative study. Acta
Psychiatr Scand 1980;62:13–31.
135. Sackeim HA, Prudic J, Devanand DP, Nobler MS,
Lisanby SH, Peyser S, et al. A prospective,
randomized, double-blind comparison of bilateral
and right unilateral electroconvulsive therapy at
different stimulus intensities. Arch Gen Psychiatry
2000;57:425–34.
136. Taylor MA, Abrams R. Short-term cognitive effects
of unilateral and bilateral ECT. Br J Psychiatry
1985;146:308–11.
137. Abrams R, Taylor MA, Faber R, Ts’o TO,
Williams RA, Almy G. Bilateral versus unilateral
electroconvulsive therapy: efficacy in melancholia.
Am J Psychiatry 1983;140:463–5.
138. Halliday AM, Davison K, Browne MW, Kreeger LC.
A comparison of the effects on depression and
memory of bilateral E.C.T. and unilateral E.C.T. to
the dominant and non-dominant hemispheres.
Br J Psychiatry 1968;114:997–1012.
139. Fleminger JJ, de Horne DJ, Nair NP, Nott PN.
Differential effect of unilateral and bilateral ECT.
Am J Psychiatry 1970;127:430–6.
140. Bidder TG, Strain JJ, Brunschwig L. Bilateral and
unilateral ECT: follow-up study and critique. Am J
Psychiatry 1970;127:737–45.
141. Abrams R, Taylor MA. Electroconvulsive therapy
and the diencephalon: a preliminary report. Compr
Psychiatry 1974;15:233–6.
142. Horne RL, Pettinati HM, Sugerman AA, Varga E.
Comparing bilateral to unilateral electroconvulsive
therapy in a randomized study with EEG
monitoring. Arch Gen Psychiatry 1985;42:1087–92.
143. Sackeim HA. Affective and cognitive consequences of
ECT. National Research Register; 2000.
89
References
144. Sackeim HA, Ross FR, Hopkins N, Calev L,
Devanand DP. Subjective side effects acutely
following ECT: associations with treatment
modality and clinical response. Convulsive Therapy
1987;3:100–10.
157. McCall WV, Farah BA, Reboussin D, Colenda CC.
Comparison of the efficacy of titrated, moderatedose and fixed, high-dose right unilateral ECT in
elderly patients. Am J Geriatr Psychiatry 1995;
3:317–24.
145. Lisanby SH, Maddox JH, Prudic J, Devanand DP,
Sackeim HA. The effects of electroconvulsive
therapy on memory of autobiographical and
public events. Arch Gen Psychiatry 2000;57:
581–90.
158. McCall WV, Reboussin DM, Weiner RD, Sackeim
HA. Titrated moderately suprathreshold vs fixed
high-dose right unilateral electroconvulsive
therapy: acute antidepressant and cognitive
effects. Arch Gen Psychiatry 2000;57:438–44.
146. d’Elia G. Comparison of ECT with unilateral and
bilateral stimulation. II. Retrograde amnesia. Acta
Psychiatr Scand 1970;Suppl 215:61–75.
159. Warren EW, Tissera S. Memory test performance
under three different waveforms of ECT for
depression. Br J Psychiatry 1984;144:370–5.
147. Kronfol Z, Hamsher KD, Digre K, Waziri R.
Depression and hemispheric functions: changes
associated with electro convulsive therapy. Br J
Psychiatry 1978;132:560–7.
160. Weaver LA, Ives JO, Williams R, Nies A.
A comparison of standard alternating current and
low-energy brief-pulse electrotherapy. Biol
Psychiatry 1977;12:525–43.
148. Cohen BD, Penick SB, Tarter RE. Antidepressant
effects of unilateral electric convulsive shock
therapy. Arch Gen Psychiatry 1974;31:673–5.
161. Scott AI, Rodger CR, Stocks RH, Shering AP.
Is old-fashioned electroconvulsive therapy more
efficacious? A randomised comparative study of
bilateral brief-pulse and bilateral sine-wave
treatments. Br J Psychiatry 1992;160:360–4.
149. Bailine SH, Rifkin A, Kayne E, Selzer JA, VitalHerne J, Blieka M, et al. Comparison of bifrontal
and bitemporal ECT for major depression. Am J
Psychiatry 2000;157:121–3.
150. Gangadhar BN, Janakiramaiah N,
Subbakrishna DK, Praveen J, Reddy AK. Twice
versus thrice weekly ECT in melancholia: a
double-blind prospective comparison. J Affect
Disord 1993;27:273–8.
151. Lerer B, Shapira B, Calev A, Tubi N, Drexler H,
Kindler S, et al. Antidepressant and cognitive
effects of twice- versus three-times-weekly ECT.
Am J Psychiatry 1995;152:564–70.
152. Shapira B, Tubi N, Drexler H, Lidsky D, Calev A,
Lerer B. Cost and benefit in the choice of ECT
schedule: twice versus three times weekly ECT.
Br J Psychiatry 1998;172:44–8.
153. Janakiramaiah N, Motreja S, Gangadhar BN,
Subbakrishna DK, Parameshwara G. Once vs.
three times weekly ECT in melancholia:
a randomized controlled trial. Acta Psychiatr Scand
1998;98:316–20.
154. Vieweg R, Shawcross CR. A trial to determine any
difference between two and three times a week
ECT in the rate of recovery from depression.
J Mental Health 1998;7:403–9.
155. Kellner CH, Monroe J, Pritchett J, Jarrel MP,
Bernstein HJ, Burns CM. Weekly ECT in geriatric
depression. Convulsive Therapy 1992;8:245–52.
90
156. Krystal AD, Weiner RD, Gassert D, McCall WV,
Coffey CE, Silbert T et al. The relative ability of
three ictal EEG frequency bands to differentiate
ECT seizures on the basis of electrode placement,
stimulus intensity, and therapeutic response.
Convulsive Therapy 1996;12:13–24.
162. Andrade C, Gangadhar BN, Subbakrishna DK,
Channabasavann ASM. A double blind comparison
of sinusoidal wave and brief pulse electro
convulsive therapy in endogenous depression.
Convulsive Therapy 1988;4:297–305.
163. Cronholm B, Ottosson JO. Ultrabrief stimulus
techniques in electroconvulsive therapy. II.
Comparative studies of therapeutic effects and
memory disturbances in treatment of endogenous
depression with the Elther ES electroshock
apparatus and Siemens Konvulsator III. Journal of
Nervous and Mental Disease 1963;137:268–76.
164. Bachar E, Lerer B, Shapira B. Increment in
reminiscing after ECT: possible connections to
neuropsychologic changes. J ECT 1999;15:165–6.
165. Roemer RA, Dubin WR, Jaffe R, Lipschutz L,
Sharon D. An efficacy study of single- versus
double-seizure induction with ECT in major
depression. J Clin Psychiatry 1990;51:473–8.
166. Barton JL, Mehta S, Snaith R. The prophylactic
value of extra ECT in depressive illness. Acta
Psychiatr Scand 1973;49:386–92.
167. Suedfeld P, Ramirez CE, Fleming JA, Remick RA.
Reduction of post-ECT memory compliants
through brief, partial restricted environmental
stimulation. Biol Psychiatry 1989;13:693–700.
168. Small JG, Klapper MH, Kellams JJ, Miller MJ,
Milstein V, Sharpley PH, et al. Electroconvulsive
treatment compared with lithium in the
management of manic states. Arch Gen Psychiatry
1988;45:727–32.
169. Sikdar S, Kulhara P, Avasthi A, Singh H.
Combined chlorpromazine and electroconvulsive
therapy in mania. Br J Psychiatry 1994;164:806–10.
Health Technology Assessment 2005; Vol. 9: No. 9
170. Abraham KR, Kulhara P. The efficacy of
electroconvulsive therapy in the treatment of
schizophrenia. A comparative study. Br J Psychiatry
1987;151:152–5.
171. Gambill JA, Wilson IC. Activation of chronic
withdrawn schizophrenics. Diseases of the Nervous
System 1966;27:615–17.
172. Agarwal AKWGC. TI: role of ECT phenothiazine
combination in schizophrenia. Indian J Psychiatry
1985;27:233–6.
173. Taylor P, Fleminger JJ. ECT for schizophrenia.
Lancet 1980;i:1380–2.
174. Sarkar P, Andrade C, Kapur B, Das P,
Sivaramakrishna Y, Harihar C, et al. An
exploratory evaluation of ECT in haloperidoltreated DSM-IIIR schizophreniform disorder.
Convulsive Therapy 1994;10:271–8.
175. Miller DH. A comparison between unidirectional
current nonconvulsive electrical stimulation given
with Reiter’s machine, standard alternating
current electroshock (Cerletti method), and
pentothal in chronic schizophrenia. Am J Psychiatry
1953;109:617–20.
176. May PR. Treatment of schizophrenia. New York:
Science House; 1968.
177. Naidoo D. The effects of reserpine (Serpasil) on
the chronic disturbed schizophrenic: a
comparative study of rauwolfia alkaloids and
electroconvulsive therapy. Journal of Nervous and
Mental Disease 1956;123:1–13.
184. Sarita E, Janakiramaiah N, Gangadhar BN,
Subbakrishna DK, Rao K. Efficacy of combined
ECT after two weeks of neuroleptics in
schizophrenia: a double blind controlled study.
National Institute of Mental Health and Neurosciences
Journal 1998;16:243–51.
185. Ungvari G, Petho B. High-dose haloperidol
therapy: its effectiveness and a comparison with
electroconvulsive treatment. J Psychiatric Treatment
and Evaluation 1982;4:279–83.
186. Baker A. Physical treatment for schizophrenia. Br J
Psychiatry 1958;1958:860–4.
187. Bagadia VN, Abhyankar RR, Doshi JM. A double
blind controlled study of ECT vs chlorpromazine
in schizophrenia. J Assoc Physicians India 1983;
31:637–40.
188. Janakiramaiah N, Channabasavanna SM,
Narasimha MNS. ECT/chlorpromazine
combination versus chlorpromazine alone in
acutely schizophrenic patients. Acta Psychiatr Scand
1982;66:464–70.
189. Wu D, She CW, Liu CZ, Cho WL, Quon M, Liu SY,
AI SC. Using BPRS and serial numbers and
picture recall to test the effectiveness of ECT
versus chlorpromazine versus chlorpromazine
alone in the treatment of schizophrenia: 40 cases,
single blind observations. Chinese J Nervous and
Mental Disorders 1989;15:26–8.
190. Thorpe JG, Baker AA. The effects of physical
treatment on some psychological functions. J
Mental Science 1958;104:865–9.
178. Brandon S, Cowley P, McDonald C, Neville P,
Palmer R, Wellstood-Eason S. Leicester ECT trial:
results in schizophrenia. Br J Psychiatry 1985;
146:177–83.
191. Doongaji DR, Jeste DV, Saoji NJ, Kane PV,
Ravindranath S. Unilateral versus bilateral ECT in
schizophrenia. Br J Psychiatry 1973;123:73–9.
179. Eiduson S, Brill NG, Crumpton E. The effect of
electroconvulsive therapy on spinal fluid
constituents. Br J Psychiatry 1958;104:692–8.
192. Bagadia VN, Abhyankar RR, Pradhan PV, Shah LP.
Re-evaluation of ECT in schizophrenia: right
temporoparietal versus bitemporal electrode
placement. Convulsive Therapy 1988;4:215–20.
180. Goswami U. Efficacy of electro convulsive therapy
in treatment resistant schizophrenia syndrome:
a double blind study. In Proceedings of the 53rd
Annual Conference of the Indian Psychiatry Society,
Pune, 9–11 January 2001.
193. Chanpattana W, Chakrabhand ML, Buppanharun
W, Sackeim HA. Effects of stimulus intensity on the
efficacy of bilateral ECT in schizophrenia:
a preliminary study. Biol Psychiatry 2000;48:222–8.
181. Janakiramiah N. ECT-chlorpromazine
combination compared with chlorpromazine only
in schizophrenia. Indian J Psychiatry 1981;
23:230–3.
194. Abrams R. Daily administration of unilateral ECT.
Am J Psychiatry 1967;124:384–6.
195. Baker JC. ECT in schizophrenia. J Mental Science
1960;106:1506–11.
182. Janakiramaiah N, Channabasavann ASM,
Narasimha-Murphy N. ECT/chlorpromazine
combination versus chlopromazine alone in
acutely ill schizophrenia patients. Acta Psychiatr
Scand 1982;66:464–70.
196. Chanpattana W, Chakrabhand ML, Sackeim HA,
Kitaroonchai W, Kongsakon R, Techakasem P, et al.
Continuation ECT in treatment-resistant
schizophrenia: a controlled study. J ECT 1999;
15:178–92.
183. Small JG. ECT combined with neuroleptics in the
treatment of schizophrenia. Psychopharmacol Bull
1982;18:34–5.
197. Prudic J, Sackeim HA. Electroconvulsive therapy
and suicide risk [review]. J Clin Psychiatry 1999;
60 Suppl 2:104–10.
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
91
References
198. Babigian HM, Guttmacher LB. Epidemiological
considerations in electroconvulsive therapy. Arch
Gen Psychiatry 1984;41:661–72.
213. Faulkner A. Knowing our own minds: a survey of how
people in emotional distress take control of their lives.
London: Mental Health Foundation; 1997.
199. Avery D, Winokur G. Mortality in depressed
patients treated with electroconvulsive therapy and
antidepressants. Arch Gen Psychiatry 1976;
33:1029–37.
214. Pettinati HM, Rosenberg J. Memory self-ratings
before and after electroconvulsive therapy:
depression-versus ECT induced. Biol Psychiatry
1984;19:539–48.
200. Tsuang MT, Dempsey GM, Fleming JA. Can ECT
prevent premature death and suicide in
‘schizoaffective’ patients? J Affect Disord 1979;
1:167–71.
215. Philpot M. Users’ views of ECT in two mental
health trusts. Maudsley Hospital: 2002.
201. Black DW, Winokur G, Monandoss E, Woolson RF,
Nasrallah A. Does treatment influence mortality in
depressives? Ann Clin Psychiatry 1989;1:165–73.
217. ECT Anonymous. Questionnaire results to March 1999.
West Yorks: ECT Anonymous; 1999.
202. Sharma V. Retrospective controlled study of
inpatient ECT: does it prevent suicide? J Affect
Disord 1999;56:183–7.
218. Freeman CP, Weeks D, Kendell RE. ECT. II:
Patients who complain. Br J Psychiatry 1980;
137:17–25.
203. Brådvik L, Berglund M. Treatment and suicide in
severe depression: a case–control study of
antidepressant therapy at last contact before
suicide. J ECT 2000;16:399–408.
219. Malcolm K. Patients’ perceptions and knowledge
of electroconvulsive therapy. Psychiatric Bulletin
1989;13:161–5.
204. Nasrallah HA, McCalley WM, Pfohl B. Clinical
significance of large cerebral ventricles in manic
males. Psychiatry Res 1984;13:151–6.
205. Kolbeinsson H, Arnaldsson OPH, Skulason S.
Computed tomographic scans in ECT patients.
Acta Psychiatr Scand 1986;132:560–7.
206. Calloway SP, Dolan RJ, Jacoby RJ, Levy R. ECT
and cerebral atrophy: a computed tomographic
study. Acta Psychiatr Scand 1981;64:442–5.
207. Hickie I, Scott E, Mitchell P, Wilhelm K, Austin MP,
Bennett B. Subcortical hyperintensities on
magnetic resonance imaging: clinical correlates
and prognostic significance in patients with severe
depression. Biol Psychiatry 1995;37:151–60.
208. Mander AJ, Whitfield A, Kean DM, Smith MA,
Douglas RH, Kendell RE. Cerebral and brain stem
changes after ECT revealed by nuclear magnetic
resonance imaging. Br J Psychiatry 1987;151:69–71.
209. Ende G, Braus DF, Walter S, Weber-Fahr W,
Henn FA. The hippocampus in patients treated
with electroconvulsive therapy: a proton magnetic
resonance spectroscopic imaging study. Arch Gen
Psychiatry 2000;57:937–43.
210. Freeman CP, Weeks D, Kendell RE. ECT: patients
who complain. Br J Psychiatry 1980;137:17–25.
211. Kerr R, McGrath J, O’Kearnery, Price J. ECT:
misconceptions and attitudes. Aust N Z J Psychiatry
1982;16:43–9.
92
216. UKAN. ECT Survey: the national experience.
Sheffield; UKAN; 1995.
212. Squire LR, Slater PC. Electroconvulsive therapy
and complaints of memory dysfunction: a
prospective three-year follow-up study. Br J
Psychiatry 1983;142:1–8.
220. Hughes J, Barraclough B, Reeve W. Are patients
shocked by ECT? J R Soc Med 1981;74:283–5.
221. Sestoft D, Pedersen OL, Bendsen B, Hancke B,
Larsen A, Mikkelsen S, et al. The effect of
electroconvulsive therapy on patients’ attitude to
treatment of depression. Nord J Psychiatry 1998;
52:31–7.
222. Wheeldon TJ, Robertson C, Eagles JM, Reid IC.
The views and outcomes of consenting and nonconsenting patients receiving ECT. Psychol Med
1999;29:221–3.
223. Szuba MP, Baxter LR, Liston E, Roy-Byrne P.
Patients’ and family perspective of
electroconvulsive therapy – correlation with
outcome. Convulsive Therapy 1991;7:175–83.
224. Baxter LR, Roy-Byrne P, Liston E, Fairbanks L.
The experience of electroconvulsive therapy in the
1980s: a prospective study of knowledge, opinions
and attitudes of California electroconvulsive
therapy patients in the Berkley Years. Convulsive
Therapy 1986;4:179–89.
225. Benbow SM. Patients views on electroconvulsive
therapy on completion of a course of treatment.
Convulsive Therapy 1988;4:146–52.
226. Riordan DM, Barron P, Bowden MF. ECT:
a patient-friendly procedure? Psychiatr Bull
1993;17:531–3.
227. Freeman CPL, Kendell RE. Patients’ experience of
an attitude to ECT. In Palmer RL, editor.
Electroconvulsive therapy: an appraisal. Oxford:
Oxford University Press; 1981. pp. 270–87.
228. Pettinati HM, Tamburello TA, Ruetsch CR,
Kaplan FN. Patient attitudes toward
Health Technology Assessment 2005; Vol. 9: No. 9
electroconvulsive therapy. Psychopharmacol Bull
1994;30:471–5.
229. Johnstone L. Adverse psychological effects of ECT.
J Mental Health 1999;8:69–85.
230. Rogers A, Pilgrim D. Service users’ views of
psychiatric treatments. Sociology of Health and
Illness 1993;15:612–31.
231. Aperia B. Hormone pattern and post treatment
attitudes in patients with major depressive
disorder given electroconvulsive therapy. Acta
Psychiatr Scand 2002;73:271–4.
major depressive disorder in the United Kingdom.
Pharmacoeconomics 2000;18:143–8.
244. Bauer M, Dopfmer S. Lithium augmentation in
treatment-resistant depression: meta-analysis of
placebo-controlled studies. J Clin Psychopharmacol
1999;19:427–34.
245. Clerc GE, Ruimy P, Verdeau-Pailles J. A doubleblind comparison of venlafaxine and fluoxetine in
patients hospitalized for major depression and
melancholia. Int Clin Psychopharmacol
1994;9:139–43.
232. Bernstein H, Beale MD, Kellner CH. Patients’
attitudes to ECT after treatment. Psychiatric Annals
1998;28:524–7.
246. Byrne SE, Rothschild AJ. Loss of antidepressant
efficacy during maintenance therapy: possible
mechanisms and treatments. J Clin Psychiatry 1998;
59:279–88.
233. Moore P, Landolt HP, Seifritz E, Clark C, Bhatti T,
Kelsoe J, et al. Clinical and physiological
consequences of rapid tryptophan depletion
[review]. Neuropsychopharmacology 2000;23:601–22.
247. Hirschfeld RMA. Clinical importance of long-term
antidepressant treatment. Br J Psychiatry 2001;
179:s4–8.
234. Dardennes RM, Lafuma A, Fagnani F, Pribil C,
Bisserbe JC, Berdeaux G. Economic assessment of
a maintenance treatment strategy in prevention of
recurrent depressive disorder. Value Health 2000;
3:40–7.
235. Kamlet MS, Paul N, Greenhouse J, Kupfer DJ,
Frank E, Wade M. Cost utility analysis of
maintenance treatment for recurrent depression.
Control Clin Trials 1995;16:17–40.
236. Revicki DA, Brown RE, Palmer W, Bakish D,
Rosser W, Anton SF, et al. Modelling the cost
effectiveness of antidepressant treatment in
primary care. Pharmacoeconomics 1995;8:524–40.
237. Janicak PG, Davis JM, Gibbons RD, Ericksen S,
Chang S, Gallagher P. Efficacy of ECT: a metaanalysis. Am J Psychiatry 1985;142:297–302.
238. Sackeim HA. Continuation therapy following ECT
– directions for future research. Psychopharmacol
Bull 1994;30:501–21.
239. Prudic J, Sackeim HA, Devanand DP. Medication
resistance and clinical response to electroconvulsive
therapy. Psychiatr Res 1990;31:287–96.
240. Einarson TR, Arikian S, Sweeney S. A model to
evaluate the cost-effectiveness of oral therapies in
the management of patients with major depressive
disorders. Clin Ther 1995;17:136–53.
241. Burke WI, Rubin EH, Zorumski CF, Wetzel RD.
The safety of ECT in geriatric psychiatry. J AM
Geriatr Soc 1987;35:516–21.
242. Doyle JJC. A multinational pharmacoeconomic
evaluation of acute major depressive disorder
(MDD): a comparison of cost-effectiveness between
venlafaxine, SSRIs and TCAs. Value Health 2001;
4:16–31.
243. Freeman H, Arikian S, Lenox SA.
Pharmacoeconomic analysis of antidepressants for
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
248. Swoboda E, Conca A, Konig P, Waanders R,
Hansen M. Maintenance electroconvulsive therapy
in affective and schizoaffective disorder.
Neuropsychobiology 2001;43:23–8.
249. Royal Pharmaceutical Society of Great Britain.
BNF42. WeBNF No. 42; 2002.
250. Hirschfeld RM. Efficacy of SSRIs and newer
antidepressants in severe depression: comparison
with TCAs [review]. J Clin Psychiatry 1999;
60:326–35.
251. Einarson TR, Addis A, Iskedjian M.
Pharmacoeconomic analysis of venlafaxine in the
treatment of major depressive disorder.
Pharmacoeconomics 1997;12:286–96.
252. Montgomery SA, Brown RE, Clark M. Economic
analysis of treating depression with nefazodone v.
imipramine. Br J Psychiatry 1996;168:768–71.
253. Netten A, Rees T, Harrison G. Unit costs of health
and social care. Canterbury: PSSRU; 2001.
254. Hatziandreu EJ, Brown RE, Revicki DA, Turner R,
Martindale J, Levine S, Siegel JE. Cost utility of
maintenance treatment of recurrent depression
with sertraline versus episodic treatment with
dothiepin. Pharmacoeconomics 1994;5(3):249–68.
255. Revicki DA, Wood M. Patient-assigned health
state utilities for depression-related outcomes:
differences by depression severity and antidepressant medications. J Affect Disord 1998;
48:25–36.
256. Bennett KJ, Torrance GW, Boyle MH, Guscott R.
Cost–utility analysis in depression: the McSad
utility measure for depression health states.
Psychiatr Serv 2000;51:1171–6.
257. Raftery J. NICE: faster access to modern
treatments? Analysis of guidance on health
technologies. BMJ 2001;323:1300–3.
93
References
258. Oh PI, Iskedjian M, Addis A. Pharmacoeconomic
evaluation of clozapine in treatment-resistant
schizophrenia: a cost–utility analysis.
Pharmacoeconomics 2001;8:199–206.
259. Royal College of Psychiatrists. The practical
administration of electroconvulsive therapy. London:
Gaskell; 1989.
260. Fink M, Sackeim HA. Convulsive therapy in
schizophrenia? Schizophr Bull 1996;22:27–39.
261. McCall WV, Dunn AG, Kellner CH. Recent
advances in the science of ECT: can the findings
be generalized? J ECT 2000;16:323–6.
262. Heikman P, Katila H, Sarna S, Wahlbeck K,
Kuoppasalmi K. Differential response to right
unilateral ECT in depressed patients: impact of
co-morbidity and severity of illness. BMC Psychiatry
2002;2(2).
263. American Psychiatric Association. Diagnostic and
Statistical Manual of Mental Disorders. 3rd ed.
Washington, DC: APA; 1987.
264. Feighner JP, Robins E, Guze SB, Woofruff RA,
Winokur G, Munoz RA. Diagnostic criteria for use
in psychiatric research. Arch Gen Psychiatry
1972;26:57–63.
265. Klein DF. Endomorphic depression. Arch Gen
Psychiatry 1974;31:447–54.
266. Flint A. The impact of treatment resistance on
depressive relapse following electroconvulsive
therapy [letter; comment]. Acta Psychiatr Scand
1997;96:405–6.
94
267. Leonhard K. The classification of endeogenous
psychoses. New York: Irvington; 1979.
268. Langfeldt G. Diagnosis and prognosis of
schizophrenia. Proceedings of the Royal Society of
Medicine 1960;53:1047–52.
269. Chanpattana W. Maintenance ECT in treatmentresistant schizophrenia. J Med Assoc Thai 2000;
83:657–62.
270. Kane JM, Honigfeld G, Singer J Meltzer H,
Clozaril Collaborative Study Group. Clozapine for
the treatment of treatment resistant schizophrenia:
a double blind comparison with chlorpormazine.
Arch Gen Psychiatry 1988;45:789–96.
271. Fink M. Electroshock restoring the mind. New York:
Oxford University Press, 1999.
272. Abrams R. Electroconvulsive therapy. New York:
Oxford University Press, 2000.
273. Spitzer RL, Endicott J, Robins E. Research
Diagnostic Criteria: Rationale and reliability. Arch
Gen Psychiatry 1978;35:773–82.
274. Lancaster NP, Steinert RR, Frost I. Unilateral
electroconvulsive therapy. J Mental Sci
1958;104:221–7.
275. Foulds GA, Caine TM. The assessment of some
symptoms and signs of depression in women.
J Mental Sci 1959;105:182.
276. Gomez J. The subjective side-effects of ECT. Br J
Psychiatry 1975;127:609–11.
Health Technology Assessment 2005; Vol. 9: No. 9
Appendix 1
Electronic bibliographic databases searched
Biological Abstracts
Health Technology Assessment (HTA) Database
CINAHL
MEDLINE
Cochrane Controlled Trials Register (CCTR)
NHS Economic Evaluations Database (NHS EED)
Cochrane Database of Systematic Reviews (CDSR)
Office of Health Economics Health Economic
Evaluations Database (OHE HEED)
Cochrane Schizophrenia Group Trials Register
Database of Abstracts of Reviews of Effectiveness
(DARE)
PreMEDLINE
PsycINFO
EBM Reviews
Science Citation Index (SCI)
EMBASE
Social Sciences Citation Index (SSCI)
Health Management Information Consortium
(HMIC)
95
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Appendix 2
Other sources consulted
Agency for Healthcare Research and Quality
(AHRQ)
National Assembly for Wales
Aggressive Research Intelligence Facility (ARIF)
National Coordinating Centre for Health
Technology Assessment (NCCHTA)
AltaVista
National Guideline Clearinghouse (NGC)
Association of British Health Care Industries
National Research Register (NRR)
Bandolier
Organising Medical Networked Information
(OMNI)
Canadian Co-ordinating Centre for Health
Technology Assessment (CCOHTA)
Research Findings Register (ReFeR)
CenterWatch Trials Register
Royal College of Anaesthetists
Centre for Health Economics, University of York
Royal College of Nursing
Copernic
Royal College of Psychiatrists
Current Controlled Trials (CCT)
ScHARR Library Catalogue
Current Research in Britain (CRiB)
Schizophrenia Association of Great Britain
Dantec Electronics Ltd
Scottish InterCollegiate Guideline Network
(SIGN)
Department of Health
Ectron Ltd
eGuidelines
Health Evidence Bulletins, Wales
Index to Theses
International Network of Agencies for Health
Technology Assessment (INAHTA) Clearinghouse
Medical Research Council (MRC) Funded Projects
Database
Mental Health Foundation
The Association of Anaesthetists of Great Britain
and Ireland
The Mental Health Act Commission
Trent Working Group on Acute Purchasing
Turning Research into Practice (TRIP) Database
Wessex Development and Evaluation Committee
(DEC) Reports
West Midlands Development and Evaluation
Services (DES) Reports
World Health Organization (WHO)
MIND
97
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Appendix 3
Search strategies used in the major electronic
bibliographic databases
Biological abstracts
1985–2001
SilverPlatter WebSPIRS
Search undertaken December 2001
#1
#2
#3
electroconvulsive therap* or electro
convulsive therap* or electroshock therap*
or electro shock therap* or ect
depression or schizophreni* or catatoni* or
bipolar disorder* or mania or manic or
mood disorder* or mental disorder*
#1 and #2
CDSR and CCTR
2001 Issue 4
The Cochrane Library, Update Software (CD-ROM
version)
Search undertaken December 2001
#1
#2
#3
#4
#5
#6
#7
#8
#9
#10
#11
#12
#13
#14
#15
#16
#17
#18
#19
#20
ELECTROCONVULSIVE-THERAPY*:ME
ELECTRIC-STIMULATION*:ME
ELECTRIC-STIMULATIONTHERAPY*:ME
((ELECTRO NEXT CONVULSIVE) NEXT
THERAP*)
(ELECTROCONVULSIVE THERAP*)
(ELECTRO NEXT SHOCK) NEXT
THERAP*)
(ELECTROSHOCK NEXT THERAP*)
(ELECTRIC* NEXT STIMULATION)
#1 OR '2 OR #3 OR #4 OR #5 OR #6
OR #7 OR #8
DEPRESSION*:ME
SCHIZOPHRENIA*:ME
SCHIZOPHRENI*
CATATONIA*:ME
CATATONI*
BIPOLAR-DISORDER*:ME
(MANIA OR MANIC)
MOOD-DISORDERS*:ME
ADJUSTMENT-DISORDERS*:ME
PSYCHOTIC-DISORDERS*:ME
AFFECTIVE-SYMPTOMS*:ME
#21
#22
#23
MENTAL-DISORDERS:ME
#10 OR #11 OR #12 OR #13 OR #14
OR #15 OR #16 OR #17 OR #18 OR
#19 OR #20 OR #21
#9 AND #22
CINAHL
1982–2001
Ovid Biomed
Search undertaken December 2001
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
electroconvulsive therapy/
electro convulsive therap$.tw
electroconvulsive therap$.tw
electro shock therap$.tw
electroshock therap$.tw
ect.tw
or/1-6
exp depression/
exp schizohrenia/
schizophreni$.tw
catatoni$.tw
exp affective disorders, psychotic/
(mania or manic).tw
exp affective disorders/
exp adjustment disorders/
exp mental disorders/
or/8-16
7 and 17
Centre for Reviews and
Dissemination (CRD) databases
(NHS DARE, EED and HTA)
CRD website – complete databases
Search undertaken December 2001
(electro convulsive therapy or electroconvulsive
therapy or electroshock therapy or electro shock
therapy or electrical stimulation)/All fields AND
(depression or schizophrenia or catatonia or
bipolar disorder or mania or manic or mood
disorders or mental disorders)/All fields
99
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Appendix 3
EMBASE
MEDLINE
1980–2001
SilverPlatter WebSPIRS
Search undertaken December 2001
1966–2001
Ovid Biomed
Search undertaken December 2001
#1
#2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
#3
#4
#5
#6
#7
#8
#9
#10
#11
#12
#13
#14
#15
'electroconvulsive-therapy' / all subheadings
electroconvulsive therap* or electro
convulsive therap*
electroshock therap* or electro shock
therap*
ect
#1 or #2 or #3 or #4
explode 'affective-neurosis' / all subheadings
depression
schizophreni*
explode 'schizophrenia-' / all subheadings
catatoni*
'catatonia-' / all subheadings
explode 'manic-depressive-psychosis' / all
subheadings
mania or manic
#6 or #7 or #8 or #9 or #10 or #11 or
#12 or #13
#5 and #14
OHE HEED
CD-ROM version
Search undertaken December 2001
Search terms
●
ect or electroconvulsive or electro convulsive or
electroshock or electro shock
Fields searched
●
●
●
●
●
●
Abstract
All data
Article title
Book title
Keywords
Technology assessed
HMIC
1980–2001
SilverPlatter WinSPIRS
Search undertaken December 2001
#1
#2
#3
#4
100
ect
electroconvulsive therap*
electro convulsive therap*
#1 or #2 or #3
electroconvulsive therapy/
electro convulsive therap$.tw
electroconvulsive therap$.tw
electro shock therap$.tw
electroshock therap$.tw
exp electric stimulation/
electric$ stimulation.tw
or/1-7
depression/
exp schizophrenia/
schizophreni$.tw
catatonia/
catatoni$.tw
exp bipolar disorder/
(mania or manic).tw
exp mood disorders/
adjustment disorders/
psychotic disorders/
affective symptoms/
mental disorders/
or/9-20
8 and 21
PsycINFO
1967–2001
SilverPlatter WebSPIRS
Search undertaken December 2001
#1
#2
#3
#4
#5
#6
#7
#8
#9
'electroconvulsive-shock-therapy' in de
electroconvulsive therap* or electro
convulsive therap*
electroshock therap* or electro shock
therap*
ect
#1 or #2 or #3 or #4
explode 'mental-disorders' in de
schizophreni* or catatoni* or bipolar
disorder* or mania or manic or depression
#6 or #7
#5 and #8
SCI and SSCI
1981–2001
Web of Science
Search undertaken December 2001
Health Technology Assessment 2005; Vol. 9: No. 9
Title=(ect or electroconvulsive therapy or electro
convulsive therapy or electroshock therapy or
electro shock therapy) and (depression or
schizophreni* or catatoni* or bipolar disorder* or
mania or manic or mood disorder* or mental
disorder*); DocType=All document types;
Languages=All languages; Databases=SCIEXPANDED, SSCI; Timespan=All Years
101
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment 2005; Vol. 9: No. 9
Appendix 4
Methodological search filters used in Ovid MEDLINE
Guidelines
1
2
3
4
5
guideline.pt
practice guideline.pt
exp guidelines/
health planning guidelines/
or/1-4
Systematic reviews
1
2
3
4
5
6
7
8
9
10
11
12
13
meta-analysis/
exp review literature/
(meta-analy$ or meta analy$ or metaanaly$).tw
meta analysis.pt
review academic.pt
review literature.pt
letter.pt
review of reported cases.pt
historical article.pt
review multicase.pt
or/1-6
or/7-10
11 not 12
Randomised controlled trials
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
randomized controlled trial.pt
controlled clinical trial.pt
randomized controlled trials/
random allocation/
double blind method/
or/1-5
clinical trial.pt
exp clinical trials/
((clin$ adj25 trial$)).ti, ab
((singl$ or doubl$ or trebl$ or tripl$) adj25
(blind$ or mask$)).ti, ab
placebos/
placebos.ti, ab
random.ti, ab
research design/
or/7-14
comparative study/
exp evaluation studies/
follow up studies/
(control$ or prospectiv$ or volunteer$)).ti, ab
prospective studies/
21 or/16-20
22 6 or 15 or 21
Economic evaluations
1
2
3
4
5
6
7
8
9
10
11
12
13
economics/
exp “costs and cost analysis”/
economic value of life/
exp economics, hospital/
exp economics, medical/
economics, nursing/
economics, pharmaceutical/
exp models, economic/
exp “fees and charges”/
exp budgets/
ec.fs
(cost or costs or costed or costly or costing$).tw
(economic$ or pharmacoeconomic$ or price$
or pricing).tw
14 or/1-13
Quality of life
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
exp quality of life/
quality of life.tw
life quality.tw
hql.tw
(sf 36 or sf36 or sf thirtysix or sf thirty six or
short form 36 or short form thirty six or short
form thirtysix or shortform 36).tw
qol.tw
(euroqol or eq5d or eq 5d).tw
qaly$.tw
quality adjusted life year$.tw
hye$.tw
health$ year$ equivalent$.tw
health utilit$.tw
hui.tw
quality of wellbeing$.tw
quality of well being.tw
qwb.tw
(qald$ or qale$ or qtime$).tw
disability adjusted life year$.tw
daly$.tw
(hamilton depression rating scale or hdrs-17
or ham-d).tw
hopkin$ symptom checklist score$.tw
103
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Appendix 4
22 chronic disease score4.tw
23 (montgomery asberg depression rating scale or
madrs).tw
24 brief psychiatric rating scale.tw
25 "kiddie schedule for affective disorders and
schizophrenia".tw
26 clinical global impression.tw
27 (symptom free days or sfd).tw
28 social functioning scale.tw
29 depression recurrence rate$.tw
30 mini-mental state examination.tw
31 retrograde memory test$.tw
32 anterograde memory test$.tw
33 or/1-32
Side-effects
1
2
3
4
5
6
7
8
9
10
11
12
13
ae.fs
ct.fs
co.fs
((side or adverse or unintended or unwanted)
adj2 (effect$ or event$)).tw
harm$.tw
complication$.tw
contraindication$.tw
exp suicide/
exp memory disorders/
exp cognition disorders/
memory loss$.tw
cognitive$ impairment$.tw
or/1-12
Patient acceptability
1
2
3
4
5
6
7
104
exp patient acceptance of health care/
patient$ acceptabil$.tw
patient$ complian$.tw
patient$ choice$.tw
patient$ preference$.tw
patient$ knowledge$.tw
or/1-6
Staff training
1
2
3
4
5
6
7
8
(staff adj3 train$).tw
(staff adj3 supervision$).tw
exp inservice training/
audit$.tw
exp medical audit/
nursing audit/
exp management audit/
or/1-7
Health Technology Assessment 2005; Vol. 9: No. 9
Appendix 5
Descriptions of included studies
105
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
106
Inclusion/exclusion criteria
Search strategies
Study types: effectiveness: RCTs; safety:
case–control and cohort studies
Outcomes: Primary: symptoms: change on
a continuous scale at the end of a course of
ECT and at 6 months’ follow-up; mortality:
all cause and cause specific, including
suicide; cognitive functioning: orientation,
retrograde memory, anterograde memory,
subjective distress immediately after ECT, at
end of treatment and at 6 months; quality of
life; duration of hospital admission;
functional impairment; structural brain
changes
Study quality rating: no. RCTs:
allocation concealment, blinding, loss
to follow up and length of follow-up;
cohort studies: measurement bias,
handling of confounding factors,
number of cases and loss to
follow-up; case–control studies:
measurement bias, handling of
confounding factors and number of
cases
Blinded assessment: no
Data quality
Other: checked reference lists of the
Methods: two independent
ECT guidelines issued by the APA
reviewers
(2001) and RCP (1995);
handsearched specialist textbooks on
ECT by Fink271 and Abrams;272
citation tracking of included studies
and systematic reviews; contacted
experts in the field and
manufacturers of ECT machines for
unpublished studies
Electronic databases: CDSR,
DARE, Cochrane Collaboration
Depression Anxiety and Neurosis
(Group) Controlled Trial Register
(CCDANCTR), Cochrane
Comparators: no ECT, sham ECT,
Schizophrenia Group Register
pharmacotherapy and/or psychotherapy
(CSGCTR), February 2001; Biological
Abstracts, CINAHL, EMBASE,
Populations: same diagnosis of depression, LILACS, MEDLINE, PsycINFO and
schizophrenia and mania, according to
SIGLE, March 2001; CCTR, June
explicit criteria
2001
UK ECT Group, Interventions: ECT: electrode placement
(bilateral vs unilateral), dosage, waveform,
200351
frequency of administration, number of
ECT sessions
Study
TABLE 25 Systematic reviews of the clinical effectiveness and safety of ECT in depression, schizophrenia and mania
Confidence intervals: yes
continued
Dichotomous data: odds ratios and
absolute risk differences
Continuous data: standardised effect
size
Publication bias: assessed by funnel
plots
Subgroup analyses: defined a priori
psychotic depression, retarded
depression, age, treatment resistance,
gender and severity at entry to the
trial, but not conducted owing to
limited data
Sensitivity analysis: conducted
excluding studies of inferior quality
Heterogeneity: conducted but
methods not reported
Meta-analysis: full random effects
model
Data synthesis methods
Appendix 5
Interventions: ECT (modified or
unmodified), electrode placement (bilateral
vs unilateral), dosage, waveform, frequency
of administration, number of ECT sessions
Tharyan and
Adams, 200250
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Study types: all relevant RCTs with quality
rating A or B according to the Cochrane
Handbook
Outcomes: Primary: clinically meaningful
benefits in overall functioning,
hospitalisation status, changes in mental
state, behaviour, social and occupational
functioning, remission of symptoms in the
short term (<6 weeks), medium term
(6 weeks to 6 months) and long term
(>6 months). Secondary: premature
withdrawal from trial by decision of either
researcher or investigators and adverse
events such as cognitive side-effects and
mortality. Continuous data excluded if
>50% of people were lost to follow-up or
if the instrument had not been published in
a peer-reviewed journal. Also excluded
from analysis if did not report means and
standard deviations, or did not meet a priori
criteria for normal distribution
Populations: people with schizophrenia,
schizoaffective disorder or chronic mental
disorder (non-affective)
Comparators: placebo, sham ECT,
pharmacological interventions, nonpharmacological interventions
Inclusion/exclusion criteria
Study
Blinded assessment: not reported
Data quality
Meta-analysis: fixed then random
effects
Data synthesis methods
Confidence intervals: yes
Dichotomous data: relative risk,
NNT, NNH
Continuous data: pooled WMD
Publication bias: funnel plot
Study quality rating: Cochrane
Collaboration Handbook categories A Heterogeneity: Mantel–Haenszel
and B
test, significance level <0.10 =
evidence of heterogeneity. If remained
Methods: two independent
following use of random effects model,
Other: citations of included studies
reviewers
results were not pooled and sensitivity
were checked for additional trials and
analysis was undertaken
the first author of each trial published
since 1980 was contacted for
Sensitivity analysis: conducted where
additional references and unpublished
there was evidence of heterogeneity
trials, manufacturers of ECT
and to test the effect of including
machines and the editorial board of
studies with high attrition rates
the journal Convulsive Therapy were
contacted for additional studies
Subgroup analyses: defined a priori
and tested for method of
schizophrenia diagnosis, symptom
profile, duration of illness and trial size
Electronic databases: Biological
Abstracts, 1966–1996, EMBASE,
1980–1996, MEDLINE, 1966–2001,
PsycLIT, 1974–1996, Cochrane
Schizophrenia Group Register up to
2001
Search strategies
TABLE 25 Systematic reviews of the clinical effectiveness and safety of ECT in depression, schizophrenia and mania (cont’d)
Health Technology Assessment 2005; Vol. 9: No. 9
107
108
Interventions/populations: research
participants and all patients who provided
testimonies that they had received ECT
SURE, 200353
Search strategies
Inclusion/exclusion criteria
Electronic databases: medical and
psychological database (names not
stated) up to March 1996
Search strategies
Study types: included if data on diagnosis
and individual outcomes were provided, in
all languages, all study types
Data synthesis methods
Meta-analysis: no
Other methods: data on outcome
summarised by adding case series and
reports together to produce an overall
percentage of those with a good
outcome after ECT (ITT) and at
6 months (not ITT) by diagnosis.
Qualitative overview of data on
adverse effects
Study quality rating: yes
Data synthesis methods
Blinded assessment: no
Data quality
Research studies: when research
studies using a range of methodologies
Study quality rating: no
produced the same results, findings
were presented in terms of ‘at least
Methods: described a number of key X% of patients’ experience Y’
methodological issues identified
influencing the ability of the studies
Testimonies: analysed using a mixture
to reflect adequately and accurately
of content and discourse analysis; interpatients’ views of ECT. These include rater reliability of allocating testimonies
the setting in which attitudes to ECT to categories was 83%; illustrative
were elicited, who conducted the
quotations presented to represent
interview; the source of the sample
themes
included in the study; the interval
since ECT; the depth and complexity
of the questions asked, the degree to
which the questions were value laden
and the different scales
Blinded assessment: not reported
Data quality
Other: manual searches to identify
Methods: two independent raters
Outcomes: response to treatment defined studies that assessed the
rated study quality on several
by reviewers as those who showed marked effectiveness of ECT in people under variables to obtain a quality score
improvement or recovery both immediately the age of 18
after ECT and 6 months post-ECT as
defined by the study authors, adverse
events including cognitive functioning,
seizures and subjective side-effects
Rey and Walter, Intervention: ECT
199770, Walter
Population: people aged ≤ 18 years who
et al., 199971
received ECT
Study
TABLE 27 Systematic reviews of non-randomised evidence: children and adolescents
Electronic databases: psycINFO,
MEDLINE, Web of Science and the
King’s fund database, 1975–2001,
Proquest newspaper database,
Outcomes: long-term memory loss,
Mental Health Media Testimony
information and consent, felt compulsion to archive, searches of the Internet,
have ECT, perceived benefit and satisfaction e-mail forums and chat rooms
with ECT
Other: handsearches and contacting
Study types: all forms of evidence from
patient groups to identify unpublished
research studies and patient testimonies in literature
which patients’ views about ECT were
ascertained directly
Inclusion/exclusion criteria
Study
TABLE 26 Systematic reviews of non-randomised evidence: patient acceptability and choice
Appendix 5
Other: citation tracking from
included studies
Populations: studies providing sufficient
detail to determine whether cases met
DSM-IV criteria for catatonia. Papers were
excluded if clinical descriptions were likely
to be due to NMS or if the treatment and
response were not clearly described
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Intervention: ECT
Miller, 199481
Study types: all
Outcomes: physiological effects of ECT
during pregnancy, risk of ECT
Population: pregnant women
Inclusion/exclusion criteria
Study
Other: not reported
Electronic databases: MEDLINE,
dates not reported
Search strategies
TABLE 29 Systematic review of non-randomised evidence: use of ECT in pregnancy
Study types: all study types, written in
English
Outcomes: response to treatment based
on original authors’ clinical description of
change in catatonic symptoms after
treatment. Response was then rated
retrospectively by reviewers on a threepoint scale: none, partial and complete
Electronic databases: paperchase
medical literature search system,
1985–1994
Interventions: any intervention to treat
catatonia, including ECT combined with
pharmacotherapy and ECT alone
Hawkins et al.,
199578
Search strategies
Inclusion/exclusion criteria
Study
TABLE 28 Systematic reviews of non-randomised evidence: catatonia
Meta-analysis: no
Other methods: results summarised
in terms of the percentage of cases
reporting each complication
Blinded assessment: no
Study quality rating: no
Data synthesis methods
Other methods: descriptive statistics
of percentage of cases with each
outcome by treatment type
Study quality rating: no
Data quality
Meta-analysis: no
Data synthesis methods
Blinded assessment: no
Data quality
Health Technology Assessment 2005; Vol. 9: No. 9
109
110
West, 198194
Allocation: B, unclear
Gregory et al.,
198593
Blinding: double-blind
Allocation: B, unclear
Blinding: double-blind
Methods
Study
Interventions
History: all patients given 50 mg
amitryptaline at night during the
study. All had depression severe
enough to warrant ECT and all
had suicidal ideas. 16 had
previously had unipolar illness
and two had bipolar illness. No
information on previous ECTs
Gender: real ECT: 6 M, 5F;
sham ECT: 7M, 4F
Comparator: sham ECT: received
anaesthesia as treatment group but no
electricity twice a week for 3 weeks
Number and follow-up
continued
Length of follow-up:
3 weeks: until end of
treatment
n completed: 22
N randomised: 25
Continuous: MADRS,
N randomised: 69
HRSD, PIRS, PSE (unusable,
graph or mean change scores n completed: 48
only, no mean or SD)
Length of follow-up: until
end of ECT
Dichotomous: none
Outcomes
Inclusion: met Feighner criteria Comparison: real ECT vs sham ECT Continuous: BDI, NRS,
for primary affective
psychiatrist’s rating
disorders264
ECT: bilateral anterior placement ECT
using double-sided unrectified
Dichotomous: none
Age: real ECT: mean (SD), 52
waveform of 40 J from a Transycon
(11.1) (range 35–78) years;
machine twice weekly for 3 weeks,
sham ECT: 53.3 (22.9) (26–82)
receiving a total of six treatments
years
Inclusion: met MRC (1965)
Comparison: real ECT vs sham ECT
criteria for depression of
>1 month in duration and were ECT: either unilateral or bilateral ECT
right-handed
at waveform 1 of the duopulse Mk IV
machine twice weekly, number of
Exclusion: severe physical
treatments determined by clinical team
illness or had already received
in charge of the patient’s care. Right
ECT for current episode of
unilateral ECT in the tempoparietal
illness
position; bilateral in the
bifrontotemporal position. Monitored
Age: not specified
using the cuff method and length of fits
timed with a stopwatch
Gender: not specified
Comparator: sham ECT twice weekly
History: not specified
as treatment group but with no
electricity; number of ECTs
determined by clinical team in charge
of the patient’s care
Participants
TABLE 30 RCTs of real versus sham ECT: depression
Appendix 5
Methods
Jagadeesh et al., Allocation: B, unclear
199252
Blinding: double-blind
Study
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
History: first episode for 5/12
patients in real ECT and 2/12 in
sham ECT group. Mean duration
of current episode 2.91 months
in real ECT and 1.92 months in
sham ECT group. Mean initial
HRSD score was 26.83 for real
ECT and 26.17 for sham ECT
Gender: real ECT: 5 M, 7 F;
sham ECT: 5 M, 7F
Age: six real ECT: mean (SD)
39.92 (8.39) years; five sham
ECT: 31.92 (6.34) (range 22–52)
years
Exclusion: organic factors
contraindicating ECT, current
suicide attempt or suicide score
>3 on HRSD
Inclusion: aged between 20 and
60 years, diagnosis of major
depression endogenous subtype
on Research Diagnostic
Criteria.273 Present depressive
episode untreated with ECT,
antidepressants or
antipsychotics, informed consent
Participants
TABLE 30 RCTs of real versus sham ECT: depression (cont’d)
Continuous: HRSD, GRSD
Outcomes
Comparator: one real + five sham
received initial real ECT as for
treatment group, plus five sham ECT
where received anaesthesia but no
electricity
Dichotomous: responder:
ECT: six real ECT: bifrontotemporal
a score of ≤ 2 on GRSD at
bilateral ECT, sine wave 120–150 V for end of treatment
0.5–0.8 s three times per week for 2
weeks. Seizure monitored using the
cuff method
Comparison: six real vs one real +
five sham ECT
Interventions
continued
Length of follow-up:
2 weeks
n completed: 23
N randomised: 24
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
111
112
History: all ECT group
inpatients, two sham ECT group
outpatients. Eight real ECT and
six sham ECT had previous
failed courses of antidepressants.
11/16 in real ECT and 10/16 in
sham ECT group had received
at least one course of ECT in
the past. Mean HRSD score was
25 for real ECT and 27 for sham
ECT
Gender: real ECT: 7 M, 9 F;
sham ECT: 7 M, 9 F
Age: real ECT: mean 54.4
(range 36–69) years; sham ECT:
mean 53.4 (37–66) years
Exclusion: another psychiatric
or organic disorder or received
ECT within the previous
3 months
Inclusion: right-handed,
diagnosis of depressive illness
referred for ECT
Allocation: C, quasirandomised
Lambourn and
Gill, 197895
Blinding: double-blind
Participants
Methods
Study
TABLE 30 RCTs of real versus sham ECT: depression (cont’d)
Comparator: sham ECT three times
a week for 2 weeks, received
anaesthesia but no electricity
ECT: unilateral right tempoparietal274
brief-pulse ECT at 10 J from Ectron
Duopulse Mk IV three times a week
for 2 weeks
Comparison: real ECT vs sham ECT
Interventions
Dichotomous: individual
data presented, a priori
decision by reviewer of
a 50% reduction
on HRSD
Continuous: HRSD (15
item) (unusable, mean
change only reported)
Outcomes
continued
Length of follow-up:
2 weeks
n completed: 26
N randomised: 32
Number and follow-up
Appendix 5
Allocation: B, unclear
Freeman et al.,
197896
Blinding: double-blind
Methods
Study
History: 50% of real ECT and
60% of sham ECT had received
ECT before, and 14 real and 14
sham ECT had one or more
previous episodes of depression.
Seven real and 11 sham ECT
were taking some sort of
antidepressant medication. 25%
in each group had had previous
manic illness
Gender: real ECT: 6 M, 14 F;
sham ECT: 5 M, 15 F
Age: real ECT: mean 51 years;
sham ECT: mean 50.5 years
Exclusion: depression
secondary to other psychiatric
illnesses such as schizophrenia,
major or progressive physical
illness, organic brain disease or
received ECT in past 6 months
Inclusion: inpatients, aged
20–78 years, clinical diagnosis of
depression and a minimum
score of 15 on both the BDI and
HRSD
Participants
TABLE 30 RCTs of real versus sham ECT: depression (cont’d)
Continuous: HRSD,
Wakefield Scale, BDI, VAS
(unusable, graph only)
Outcomes
Comparator: sham ECT: first two
treatments were sham ECTs where
patients received anaesthesia but no
electric current; remaining ECTs were
real, as above
ECT: bilateral twice a week with
bidirectional 60% sine-wave current of
400 V for a peak of 1.5 s from Ectron Dichotomous: clinical
Mk IV machine. Number of ECTs
judgement of a ‘satisfactory
titrated against treatment outcome,
response’
and number ranged from three to 12
ECTs
Comparison: real ECT vs sham ECT
Interventions
continued
Length of follow-up: not
specified, but outcome
measurement occurred after
last ECT
n completed: 38
N randomised: 40
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
113
114
Methods
Johnstone et al., Allocation: B, unclear
198097
Blinding: double-blind
Study
History: 46 had definite
previous episodes of depressive
illness and seven had definite
previous episodes of mania.
15 patients had received ECT
for a previous episode (21%).
49 patients had had
antidepressant prescribed for
the index episode before the
trial
Gender: 18 M, 52 F
Comparison: real ECT vs sham ECT
Interventions
Outcomes
Continuous: HRSD, HAD
(then the ‘Leeds Scale’),
ECT: eight treatments of twice-weekly memory tests, Bunney and
bifrontal ECT using Duopulse
Hamburg NRS (unusable,
waveform 1 at 150 V for 3 s over
graph only)
4 weeks. Confirmation that a
convulsion had taken place was done
Dichotomous: HRSD score
using the cuff method
below or above median of 17
for final rating: above is a
Exclusion: poor anaesthetic risk Comparator: sham ECT: received
‘good outcome’ and below is
anaesthesia and muscle relaxants, but
a ‘poor outcome’
Age: mean 49.4 years
no electricity was passed
Inclusion: aged 30–69 years,
met MRC criteria for depressive
illness, Feighner criteria for
primary depressive illness,
Newcastle criteria for
endogenous depressive illness,
Newcastle criteria for predicting
a good outcome to ECT
Participants
TABLE 30 RCTs of real versus sham ECT: depression (cont’d)
continued
Length of follow-up:
4 weeks, 1 month and
6 months, but after the end
of ECT care was not
randomised
n completed: 62
N randomised: 70
Number and follow-up
Appendix 5
Allocation: A, concealed
Brandon et al.,
198498
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Comparison: real ECT vs sham ECT
Interventions
History: mean number of
previous admissions was 2.6 in
the real ECT group and 2.5 in
the sham ECT group. 36% in
the real and 48% in the sham
ECT group were judged to have
received an adequate course of
antidepressants before the trial.
55% in the real and 65% in the
sham ECT group had received
ECT before
ECT: bilateral using chopped sinewave current from Ectron Mk IV
machine on setting 1 twice a week for
4 weeks. Received a maximum of eight
but clinician could withdraw patient if
deterioration occurred. Patient
carefully observed to ensure fit took
Age: real ECT: mean 55.4 years; place
sham ECT: mean: 53 years
Comparator: sham ECT: received
Gender: real ECT: 21 M, 32 F; ECT procedure as for control group,
sham ECT: 13 M, 29 F
but without electricity
Inclusion: all patients
prescribed for inpatient ECT
(n = 219). 186 interviewed and
48 refused treatment; of
remaining patients, 95 had
depression and 43 had no
depressive diagnoses. Total of
138 entered trial
Participants
Continuous: HRSD
(unusable, graphs only)
Outcomes
Length of follow-up: until
end of treatment
n completed: 77
N randomised: 95
Number and follow-up
M, male; F, Female; PIRS, Psychological Impairments Scale; NRS, Nurses’ Rating Scale; GRSD, Global Rating Scale for Depression; VAS, visual analogue scale; HAD, Hospital Anxiety
and Depression scale.
Blinding: double-blind
Methods
Study
TABLE 30 RCTs of real versus sham ECT: depression (cont’d)
Health Technology Assessment 2005; Vol. 9: No. 9
115
116
Continuous: HRSD
(unusable, graph only)
Outcomes
N randomised: 30
Number and follow-up
continued
n completed: 30
ECT: bilateral, six treatments Dichotomous: unclear, no a
over 3 weeks, other stimulus priori definition, independent Length of follow-up:
clinician unclear
3 weeks
parameters not specified
(n = 15)
Comparison: ECT vs TCA
+ lithium
Interventions
Comparator: TCA +
lithium: remained on
History: all had failed to respond to a full course prestudy dose of TCA with
of TCAs, defined as ≥ 150 mg of anitryptaline for lithium added initially at a
≥ 4 weeks and failure of HRSD to drop by 40%, dose of 600 or 800 mg and
or at least to fall by 20 points. 11 had previously dose adjusted to obtain
received ECT and 28 had a previous history of
serum lithium between 0.5
depression. Mean duration of current episode
and 0.7 mEq l–1
was 6.1 months in lithium group and 7.7 in ECT
group
Gender: 10 M, 20 F
Age: 29–77 years
Exclusion: not on any other medication
Inclusion: fulfilled DSM-II for major depression;
score >20 on HDRS; Newcastle endogenicity
score >5
Allocation: B,
unclear
Dinan and
Barry, 1989111
Blinding: clinician
Participants
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression
Appendix 5
Allocation: B,
unclear
Folkerts et al.,
1997112
Blinding: unclear
Methods
Study
ECT: right unilateral at 2.5
suprathreshold, brief pulse
(1 ms, 0.9 A) performed with
a Thymatron-DGx three
times per week. Mean
number of ECTs received:
7.2 (n = 21)
Comparison: ECT vs SSRI
Interventions
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
History: all treatment resistant with a mean of
4–5 previous antidepressant trials. Baseline
HRSD was 31.1 in ECT group and 32.6 in
paroxetine group. Current episode lasted for a
mean of 59.8 weeks in ECT group and 75.2 in
paroxetine group
Gender: 18 M, 21 F, gender of dropout not
specified
Exclusion: Major depressive disorder with
psychotic features, pronounced suicidal
tendencies, severe physical illness or history of
Comparator: paroxetine
substance abuse; previous paroxetine or ECT for (SSRI): starting dose 20 mg
current episode; aged >80 years
day–1, 40 mg within 7 days
with a maximum of 50 mg.
Mean end dose 44 mg day–1
Age: ECT group: mean (SD) 47.6 (14.7) years;
paroxetine group: 52.3 (15.7) years
Inclusion: fulfil ICD-10 criteria for major
depression; score of ≥ 22 on the HDRS 21-item
version; relative therapy resistance, defined as at
least two different antidepressants (including at
least one TCA) at a dosage of ≥ 100 g
imipramine or equivalent and no improvement
for a total period of 8 weeks
Participants
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
Dichotomous: responder
defined as reduction of
≥ 50% on HRSD 21-item
version
Continuous: HRSD
(21 item)
Outcomes
continued
Length of follow-up: until
end of ECT or 4 weeks
n completed: 39
N randomised: 43
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
117
118
Bruce et al.,
1960104
Allocation: B,
unclear
Herrington
et al., 1974109
History: mean duration of current episode:
ECT: 4.1 months; L-tryptophan: 6.1 months;
mean number of previews episodes ECT: 2.4;
L-tryptophan 1.6
Gender: ECT: 6 M, 15 F; L-tryptophan: 7 M,
15 F
Age: ECT: 54.8 years; L-tryptophan: 52.7 years
Inclusion: physically healthy adults aged 25–69
years with a primary diagnosis of depression.
The severity of their illness was such that
immediate admission to hospital and ECT were
considered appropriate
Participants
Outcomes
History: no data
Number and follow-up
Comparator: L-tryptophan,
6 g day–1 for first 2 weeks
and 8 g day–1 for the second
2 weeks. Option for crossover if no success after 2
weeks
continued
Length of follow-up:
1 month and 3 months
n completed: 49
N randomised: 50
Continuous: MRC
N randomised: 40
depression scale, HRSD, BDI,
Taylor Manifest Anxiety Scale n completed: 38
ECT: administered twice a
(unusable, graphs only)
week, for a total of six to
Length of follow-up:
eight treatments. Option for Dichotomous: clinical
6 months
cross-over if no success after opinion of response, not
2 weeks
defined
L-tryptophan
Comparison: ECT vs
Interventions
Inclusion: suffering from depression, considered Comparison: ECT vs TCA
Continuous: none
to be sufficiently ill to require ECT. 49/50 had
ECT: average 6.1 treatments Dichotomous: clinical
endogenous depression (no details on the
Blinding: not blind remaining patient)
in first month
opinion as responder (not
defined)
Comparator: imipramine
Exclusion: not recorded
(Tofranil, TCA) rising to 75
Age: no data
mg t.d.s. or less if patient
was responding well
Gender: no data
Allocation: B,
unclear
Blinding: none
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
Appendix 5
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Greenblatt and
Grosser,
1964113
Allocation: A,
concealed
Steiner et al.,
1978107
Inclusion: met criteria for endogenomorphic
depression as defined by Klein265
Participants
Blinding: unclear
Allocation: B,
unclear
ECT: three times per week
Comparison: ECT vs TCA
vs MAOI
Comparator: (1)
imipramine 150 g plus
placebo for 5 weeks;
(2) imipramine 150 mg plus
L-triiodothyronine
ECT: bilateral twice a week
until improvement was
noticed, but no more than
ten treatments allowed.
Waveform, dosage and
machine not specified
Comparison: ECT vs TCA
+ placebo vs TCA +
L-triiodothyronine
Interventions
History: not recorded
Comparator: imipramine
(Tofranil, TCA) 200 mg +
optional 50 mg, phenelzine
Exclusion: patients with severe organic brain
(Nardil, MAOI) 60 mg +
syndromes, chronic alcoholism or drug addiction optional 15 mg, or
isocarboxazid (Marplan,
Age: M 46.8 years, F 45.4 years
MAOI) 40 mg + optional
10 mg
Gender: 90 M, 191 F
Inclusion: all patients admitted with a
symptomatology of severe depression, regardless
of dynamics or specific diagnostic category. The
major diagnostic categories comprised
psychoneurosis, manic depression, involution,
schizophrenic reactions, schizoaffective type and
a mixed category of character
History: mean number of previous episodes of
depression 2.4, and a family history of
depression in four patients. All had been
currently depressed for 6 weeks and had been
unsuccessfully treated in an outpatient treatment
trial (definition not specified)
Gender: all F
Age: mean 55.5 (range 30–60) years
Blinding: not blind Exclusion: known endocrine or cardiovascular
disorders, CNS disorders including brain trauma
or convulsive disorders, drug addiction or mental
deficiency and treated with ECT at any time in
the past 6 months
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
n completed: 12
N randomised: 12
Number and follow-up
N randomised: 281
continued
Dichotomous: clinician
n completed: 281
opinion of ‘marked
improvement’: the patient is Length of follow-up: until
practically symptom free and end of treatment
capable of functioning in the
community
Continuous: none
Dichotomous: responder
Length of follow-up:
defined as moderate or
5 weeks
marked improvement on
CGI and a total score on
HRSD of ≤ 10 (50%
reduction in HRSD also gives
same result)
Continuous: personal data
inventory, CGI, HRSD, SideEffect Symptom Scale
Outcomes
Health Technology Assessment 2005; Vol. 9: No. 9
119
120
History: past history of affective illness:
ECT: depression 3, mania 3, both 1;
imipramine: depression 2, mania 1. Family
history of affective illness: ECT: 0; imipramine: 3.
Duration of illness: ECT: <3 months 10,
>3 months 6; imipramine: <3 months 7,
>3 months 9
Gender: ECT 9 M, 7 F; imipramine: 5 M, 11 F
Age: ECT: mean (SD) 46.06 (11.80) years,
imipramine: 42.19 (12.66) years
Exclusion: patients treated with any
psychopharmacological agents except for
benzodiazepines in the past month, those who
had received ECT for the current depressive
episode and patients who had major physical
illnesses
Inclusion: fulfilled criteria for major depressive
episode according to ICD-10 as judged
independently by two psychiatrists. Two had
bipolar depression, the others had either single
or recurrent major depression
Allocation: B,
unclear
Gangadhar
et al., 1982110
Blinding: doubleblind
Participants
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
Outcomes
Number and follow-up
Comparator: Imipramine
(TCA) (25 mg) three per day
week 1, six per day weeks
2–11 and three per day
during week 12
continued
Continuous: HRSD, social
N randomised: 32
dysfunction and organic brain
ECT: modified bilateral using dysfunction battery, siden completed: 24
150–250 mg of thiopentone, effects checklist (unusable,
20–30 mg of succinylcholine medians, no SD)
Length of follow-up:
and 0.65 mg of atropine. Six
6–12 months
ECTs on alternate days for
Dichotomous: none
the first 2 weeks and one
ECT each week in the next
2 weeks. Three maintenance
ECTs were administered in
the next 8 weeks during
weeks 6, 8 and 12 of the trial
period
Comparison: ECT vs TCA
Interventions
Appendix 5
Janakiramaiah
et al., 2000103
Allocation: B,
unclear
McDonald
et al., 196699
History: none (new admissions)
Gender: 11 M, 19 F
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
History: duration of current episode:
mean (SD) ECT: 4.8 (3.3) months;
imipramine: 5.4 (3.5) months; yoga: 3.8
(2.8) months. Recurrent: ECT: 3;
imipramine 2; yoga: 4
Gender: ECT: 6 M, 9 F; imipramine:
10 M, 5 F; yoga: 9 M, 6 F
Age: ECT: mean (SD) 36.7 (2.5) years;
imipramine: 43.4 (11.9) years; Yoga: 36.0
(7.8) years
Comparator: imipramine (Tofranil,
TCA) 150 mg once daily for 4
weeks, no other psychotropic
drugs; or Sudarshan Kriya yoga for
45 minutes; 6 days a week, mean
(SD) number of sessions 20.3 (2.8)
ECT: bilateral, three times a week.
The stimulus was set 60 mC above
threshold (determined on the first
and seventh ECT). Mean (SD)
number of ECT sessions 8.9 (3.3).
Seizures of 25 s on EEG or 15 s on
motor were ensured in all sessions
Comparison: ECT vs TCA vs yoga
Dichotomous: remitters
defined as HRSD 17-item
score <8
Continuous: BDI, HRSD
(17 item)
Comparator: amitriptyline (TCA)
20 mg i.m. for 3 days, 50 mg orally
for the remainder of the 1-month
trial period. Sham ECT was
delivered while the patient was
unconscious through injection of
thiopental sodium (Pentothal,
barbiturate)
Exclusion: organic complications to
contraindicate drugs or ECT,
antidepressants in the past 2 weeks,
unable to speak English
Age: 20–65 years
Outcomes
Comparison: ECT vs TCA vs sham Continuous: MMPI, WBIS,
ECT
BGT, unvalidated depression
scale (unusable, no SD)
ECT: electrode placement unclear;
minimum of eight treatments, three Dichotomous: none
times a week
Interventions
Inclusion: all new admissions eligible
Participants
Inclusion: patients with DSM-IV
melancholic depression who were never
treated for the current episode, medically
Blinding: not blind fit, HRSD ≥ 17
Allocation: B,
unclear
Blinding: doubleblind
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
continued
Length of follow-up:
4 weeks
n completed: 45
N randomised: 45
Length of follow-up: until
end of treatment
n completed: 30
N randomised: 30
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
121
122
Shepherd,
1965100
Blinding: unclear
Allocation: B,
unclear
History: number rated severely ill: ECT:
35/65; imipramine: 27/63; phenelzine:
20/61; placebo: 16/65
Gender: completers: ECT: 24 M, 42 F;
imipramine: 22 M, 41 F; phenelzine: 18 M,
43 F; placebo: 17 M, 44 F
Age: ECT: mean 55.4 years; imipramine:
54.8 years; phenelzine: 54.7 years;
placebo: 56.3 years
Exclusion: treatment during past 6
months with either ECT or adequate trial
of pharmacotherapy, depression
secondary to other psychiatric illness such
as schizophrenia or an obsessional state,
physical disease such as malignancy,
organic cerebral disease
Inclusion: aged 40–69 years, previous
duration of illness <18 months,
depressive illness
History: not specified
Comparator: 50 mg of
imipramine, 15 mg of phenelzine or
15 mg of placebo, with two tablets
on day 1, three on day 2, four on
days 3–28, four on days 29–56, two
on days 57–84 and one on days 85
and 112
ECT: four to eight treatments
within first 3.5 weeks of trial,
according to physician’s judgement
Comparison: ECT vs TCA vs
MAOI vs placebo
Dichotomous: clinical
opinion of wholly or almost
without symptoms
Continuous: physician’s
rating on 15 symptoms
(unvalidated) (unusable, no
SD)
Comparator: imipramine: mean
dose 150 g in the first and last
thirds of the study, 220 g in the
middle third of the study
Age: 40–59 years
Gender: all F
Continuous: HRSD,
MMPI-D (unusable, graph or
mean change only reported)
Outcomes
ECT: two treatments per week for Dichotomous: none
a total of six treatments, electrode
placement, dosage waveform and
machine not specified
Comparison: ECT + TCA vs ECT
+ placebo vs sham ECT +
imipramine
Interventions
Exclusion: schizophrenia and organic
brain disorder
Inclusion: all women aged 40–59 years
admitted to a psychiatric hospital with
depressive symptoms
Allocation: B,
unclear
Wilson et al.,
196365
Blinding: unclear
Participants
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
continued
Length of follow-up:
4, 8, 12 and 24 weeks and
immediately postdischarge
n completed: 250
N randomised: 269
Length of follow-up:
5 weeks
n completed: 22
N randomised: 24
Number and follow-up
Appendix 5
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Kendrick et al.,
1965101
MacSweeney,
1975108
Allocation: B,
unclear
Stanley and
Fleming,
1962114
Blinding: unclear
Allocation: B,
unclear
Blinding: not blind
Allocation: B,
unclear
Blinding: clinician
Methods
Study
History: not reported
Gender: 32 M, 34 W
Age: elderly, but age not reported
Exclusion: not reported
Inclusion: elderly patients admitted to
Bethem Royal Hospital suffering from
affective disorder
History: not reported
Continuous: nine
‘depressive scales’ found to
be valid by Foulds and
Caine275 (unusable, no SD)
Outcomes
Comparison: ECT vs L-tryptophan Continuous: BDI (unusable,
no SD)
ECT: unilateral ECT administered
twice weekly
Dichotomous: none
Comparator: phenelzine (MAOI)
ECT: three times a week; total
number of treatments determined
by response, usually six to eight
Comparison: ECT vs MAOI
Interventions
Comparison: ECT vs TCA + TCA Continuous: Mill Hill
Vocabulary Scale, Raven’s
ECT: not reported
Coloured Progressive
Matrices, WAIS, Synonym
Comparator: imipramine and
Learning Test, Ing’s Paired
Trofranil
Associate Learning Test, Digit
Copying Test (unusable, no
symptom scales reported)
Age: ECT (completers): mean 57.2 years;
pharmacotherapy (completers): 54.8 years Comparator: 3 g of L-tryptophan
and 1 g of nicotinamide daily
Gender: ECT: 3M, 11 F;
pharmacotherapy: 3 M, 10 F
Exclusion: not reported
Inclusion: not reported
History: acute admissions
Gender: all F
Age: ECT: mean 43.8 years; phenelzine
51.3 years
Exclusion: not reported
Inclusion: patient suffering from
depression and where ECT was normally
indicated
Participants
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
continued
Length of follow-up: not
reported
n completed: 68
N randomised: 69
Length of follow-up:
28 days
n completed: 25
N randomised: 27
Length of follow-up:
1 month
n completed: 38
N randomised: 47
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
123
124
Bagadia et al.,
1981105
Blinding: doubleblind
History: not reported
Gender: both, numbers not reported
Age: actual age of participants not
reported
with an initial dose of two tablets a
day increased to six tablets a day,
up to 150 mg. Placebo was calcium
lactate 300 mg
ECT: bilateral with stimulus of
110 V a.c. for approx. 0.5 s. One
Exclusion: treatment with antidepressant person received eight ECTs, the
or antipsychotic drugs within the previous others received six ECTs. Three
3 weeks, with ECT or insulin therapy
ECTs were given in the first week
within the previous 8 weeks, organic brain and two in the following week.
syndrome, convulsive disorder and
physical illness
Comparator: imipramine 25 mg
Comparison: ECT + placebo vs
TCA + sham ECT
Inclusion: aged 18–65 years, clear
depression of non-organic cause, score of
≥ 16 on HRSD (17-item version), score of
≥ 12 on BDI
Comparison: ECT vs TCA +
MAOI
Allocation: B,
unclear
Inclusion: unipolar depression or
depression secondary to anxiety or
character disorder as defined by the
Feighner criteria264 and therapy resistant
(no definition given)
Interventions
ECT: bilateral, minimum of four
and maximum of ten, three times
per week, with the mean number
Age: ECT mean 40.7 years;
of ECTs received 5.4. Dosage,
pharmacotherapy 41.5 years
waveform and machine not
Gender: ECT: 2M, 7 F; pharmacotherapy: specified
3 M, 5 F
Comparator: combination of
MAOI (phenelzine) and TCA
History: all were treatment resistant to
(amitryptaline): initiated with
conventional psychotropic drugs in
amitryptaline up to 100 mg for 5–7
clinically adequate doses. Baseline mean
HRSD scores were 26.5 in ECT group and days with addition of 15 mg of
phenelzine up to a maximum of
22.8 in pharmacotherapy group. The
45 mg for a minimum of 3 weeks.
pharmacotherapy group had a greater
Mean daily dose was 34 mg of
mean number of previous illnesses (2.5)
MAOI and 71 mg of TCA
than the ECT group (1.1)
Allocation: A,
concealed
Davidson et al.,
1978102
Participants
Blinding: clinician
Methods
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
Length of follow-up:
unclear: 3–5 weeks
n completed: 17
N randomised: 19
Number and follow-up
continued
Continuous: HRSD, BDI,
N randomised: 35
BPRS, Clinical Global
Assessment, cognitive test
n completed: 20
battery (unusable, HRSD not
Length of follow-up: until
reported)
end of ECT course
Dichotomous: none
Continuous: HRSD, BDI,
State Trait Anxiety (mean
and SE)
Outcomes
Appendix 5
Methods
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Blinding: clinician
Allocation: B,
unclear
History: not specified
Gender: not specified
Age: not specified
Exclusion: not specified
Inclusion: not specified
History: not specified
Gender: all F
Age: not specified
Exclusion: not specified
Inclusion: not specified
Participants
Continuous: Immobility
Index, Clinical Item score,
HRSD, Behaviour Score
(unusable; not reported)
Comparator: imipramine (TCA) + Dichotomous: clinical
anaesthesia biweekly
opinion of marked or
moderate improvement
ECT: ECT plus placebo biweekly
Comparison: ECT+ placebo vs
TCA + sham ECT
Comparator: imipramine up to
250 mg day–1; parstelin one tablet
t.d.s.; amitryptaline up to 75 mg
t.d.s; pheniprazine 12 mg day–1;
phenelzine 15 mg t.d.s.;
chloroprothixene 120 mg day–1 up
to 180 mg day–1. 25 people in each
group, except for imipramine
(n = 50)
Dichotomous: none
Continuous: unvalidated
depression scale (unusable,
no SD)
Comparison: ECT vs TCA vs
MAOI
ECT: no description given
Outcomes
Interventions
CGI, Clinical Global Impression; MMPI, Minnesota Multiphasic Personality Inventory; BGT, Bender Gestalt Test; WAIS, Weschler Adult Intelligence Scale.
Robin and
Harris, 1962115
Hutchinson and Allocation: A,
Smedberg,
concealed
1963106
Blinding: patient
Study
TABLE 31 RCTs of ECT versus phamacotherapy: depression (cont’d)
Length of follow-up:
3 weeks
n completed: 31
N randomised: 31
Length of follow-up:
3 weeks
n completed: 0
N randomised: 200
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
125
126
Allocation: A,
concealed
Pridmore,
200055
Blinding: patient
Grunhaus et al., Allocation: B,
unclear
200054
Blinding: clinician
Methods
Study
ECT: non-dominant hemisphere unilateral;
three times a week for 2 weeks; number of
treatments and dosage according to agebased protocol in instruction manual
(percentage of 504 mC equivalent to the
patient’s age)
Comparison: ECT vs ECT + rTMS
Interventions
ECT: non-dominant unilateral, switched to
bilateral electrode placement if no
improvement. Waveform brief-pulse
bidirectional current. Mean number of
treatments 9.6 (range 7–14)
Comparison: ECT vs rTMS
Comparator: rTMS: Motor threshold
Age: ECT: mean (SD) 63.6 (15.0) years; ECT determined daily by electromyographic
+ rTMS: 58.4 (15.7) years
method, placement of the electrode over the
left dorsolateral prefrontal cortex. During
Gender: ECT: F 14, M 6; RTMS: F 12, M 8.
stimulation the coil was held with the handle
towards the back of the head. Administered
History: duration of episode: ECT: mean
five times a week for 4 weeks (for a total of
(SD) 6.9 (7.9) months; rTMS: 8.3 (7.4)
20 stimulations)
months; previous episodes: ECT: 2.4 (3.05)
months; rTMS 2.3 (2.85) months; previous
ECT: ECT: 9/20; rTMS: 14/20
Exclusion: additional DSM-IV axis I diagnoses
Inclusion: aged 18+ years, DSM-IV diagnosis
of MDD, 17-item HRSD score of ≥ 18, no
personal or first degree relative history of
seizure, no medical, neurological or
neurosurgical disorder that would preclude
the administration of ECT or rTMS
Comparator: rTMS (Magstim Super Rapid
Gender: ECT alone: 5 M, 6 F; ECT + rTMS: stimulator) and Magstim 70 mm double coil;
6 M, 5 F
intensity 100%, frequency 20 Hz, train length
2 s, number of trains 30, intertrain interval
History: not recorded
20 s
Age: ECT alone: median 48 (range 25–70)
years; ECT + rTMS: median 46 (26–58)
years
Exclusion: not recorded
Inclusion: medication-resistant major
depressive episode, diagnosis of MDD
(DSM-IV)
Participants
TABLE 32 RCTs of ECT versus rTMS: depression
Number
HRSD, BPRS, GAF, GRSD,
PSQI
N: 40
Clinical response defined as N: 23
MADRS of ≤ 12 and HRDS of
≤ 8; VAS one-item scale,
GAF; side-effects: six-item
subjective side-effects
questionnaire derived from
Gomez276
Outcomes
Appendix 5
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Shiah et al.,
200057
Allocation: B,
unclear
Mayur et al.,
200056
Blinding: unclear
Allocation: B,
unclear
Blinding: unclear
Methods
Study
Comparison: ECT + TCA/SSRI vs
ECT + placebo
Interventions
Continuous: HRSD (17
item), MADRS, UKU
subscales 1–3
Outcomes
Comparison: ECT + pindolol vs
ECT + placebo
Comparator: TCA (n = 26),
fluoxetine (SSRI) (n = 4)
History: five in Pindolol and four in
placebo group were treatment resistant
Gender: completers: 5 M, 10 F
ECT: Stimulus delivered at just
above threshold for bilateral ECT
Exclusion: other DSM-IV axis I diagnoses, and three times above threshold for
past alcohol or substance abuse,
unilateral ECT, three times a week
contraindications to the use of -blockers, for 2 weeks
received fluoxetine within 5 weeks or
MAOIs within 2 weeks
Comparator: pindolol: 2.5 mg
orally, three times a day; placebo:
Age: completers: ECT plus pindolol:
2.5 mg orally, three times a day
mean (SD) 50 (9.3) years; ECT plus
placebo: 45.8 (6.3) years
Inclusion: people routinely referred for
ECT because of treatment-resistant
depression, depression characterised by
psychotic features or acute suicidality
History: previously on antidepressant
drugs with or without psychotropics;
previous ECT use unclear. Group 1:
episode number: mean (SD) 2.7 (1.2);
mean episode duration: 4.3 (2.5) months.
Group 2: episode number: 3.1 (1.5); mean
episode duration: 5.3 (3.4) months. 17/30
(56%) had an adequate drug trial
Dichotomous: responder
defined as HRSD (29 item)
score of ≤ 12 after sixth
treatment
Continuous: HRSD (29
item), CGI
ECT: non-dominant d’Elia unilateral
Dichotomous: relapses
ECT;118 three times a week;
Age: group 1: mean (SD) 33.8 (8.0) years; machine waveform; dosage 30 mC defined as HRSD > 7
group 2:34.6 (11.9) years
upwards in steps to threshold
stimulus dose (at least 25 s of EEG
Gender: group 1: 6 M, 9 F; group 2: 8 M, seizure) (n = 15)
7F
Exclusion: neurological and cardiological
disorders
Inclusion: DSM-IV major depression
Participants
TABLE 33 RCTs of ECT plus pharmacotherapy versus ECT plus placebo/pharmacotherapy only: depression
continued
Length of follow-up:
2 weeks
n completed: 15
N randomised: 20
Length of follow-up:
2 weeks
n completed: 30
N randomised: 30
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
127
128
History: previous treatment: 40/61
antidepressants in previous periods; 24/61
antidepressants in present period. 24/61
had had previous ECT courses. Duration
of present period: 0.5–6.5 months
Gender: ECT + placebo: 12 M, 18 F;
ECT + L-tryptophan: 11 M, 20 F
Age: ECT + placebo: mean (SD) 46.1
(12.7) years; ECT + L-tryptophan 48.3
(12.4) years
Exclusion: patients aged >65 years,
somatic disease that could have a relation
to the depressive period, pregnant
patients or patients given ECT in the past
3 months
Inclusion: symptomatically, all syndromes
with a global, pervasive depression of
mood as central symptom, with one or
more concomitant symptoms, such as
psychomotor retardation, anxiety, sleep
disturbance, depressive ideas, suicidal
tendencies and diurnal rhythm with
amelioration of symptoms in the evening.
Aetiologically, endogenous symptoms.
Severity severe enough that ECT
considered the treatment of choice by
doctor responsible
Allocation: B,
unclear
D’Elia et al.,
197760
Blinding: doubleblind
Participants
Methods
Study
Outcomes
Comparator: ECT as above plus:
L-tryptophan, class unknown,
dosage 6 g daily, initiated at least
1 day before first ECT and
terminated 4 days after last ECT
ECT: unilateral stimulation on the
non-dominant hemisphere.
Number of treatments: individual:
mean (SD) ECT + placebo 6.1
(2.1); ECT + L-tryptophan: 6.3
(2.5). Frequency not clear, may be
available from d’Elia.118 Machine
waveform not clear (n = 30)
N randomised: 61
Number and follow-up
continued
n completed: 57
Dichotomous: clinical
opinion of recovered and
Length of follow-up:
much improved
1 month
(responders), and slight
improvement and unchanged
(non-responders)
Comparison: ECT + L-tryptophan Continuous: CODS, NRS,
vs ECT + placebo
HRSD (unusable, no SD)
Interventions
TABLE 33 RCTs of ECT plus pharmacotherapy versus ECT plus placebo/pharmacotherapy only: depression (cont’d)
Appendix 5
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Comparator: L-tryptophan in
isotonic saline, class unknown,
dosage 1 ml kg–1 body weight of a
10 mg ml–1 solution, length of time
taken unknown, change in dosage
unknown
Dichotomous: none
Comparison: ECT + L-tryptophan Continuous: HRSD (17
vs ECT + placebo
item) (unusable, graph only)
ECT: unilateral (side unclear);
number of treatments unclear;
twice a week; machine waveform
unclear, dosage unclear
N randomised: 57
Number and follow-up
Length of follow-up: until
end of ECT course
n completed: 20
N randomised: 20
n completed: 57
Dichotomous: clinical
opinion of recovered or
Length of follow-up:
much improved (responder), 4–5 days after end of
or slightly improved or
treatment
resistant (non-responder)
Continuous: CODS
(unusable, no SD)
Outcomes
C.ATP, continuation therapy with antipsychotic drugs; C.placebo, continuation therapy with placebo; CODS, Cronholme and Ottoson Depression Scale.
History: previous treatment not
recorded. ECT use in the past not
recorded. Duration of illness not
recorded. Prognostic factors: at least two
of the following four criteria: a phasic
course, changes in psychomotor activity,
exacerbation of the symptoms during
morning hours and unfounded changes in
self-esteem. Treatment resistance not
recorded
Gender: both groups: 3 M; 7 F
Age: both groups: mean 63 years
Exclusion: not recorded
Inclusion: not recorded
Allocation: B,
unclear
Blinding: doubleblind
ECT: bifrontotemporal electrodes,
threshold stimulation with
Age: chlorpromazine: mean 45.7 (range
unidirectional stimuli. Initially three
19–64) years; placebo: 47.5 (22–63) years times a week, later two or one
treatment(s) determined by clinical
Gender: chlorpromazine: 11 M, 17 F ;
effect
placebo: 14 M, 14 F
Comparator: chlorpromazine
History: 24/57 had received ECT
50–150 mg for 32 days with
previously; 31/57 had received
augmented daily dose 106 mg
antidepressant medication during the
current episode
Kirkegaard
et al., 197859
Exclusion: aged ≥ 65 years
Blinding: patient
Comparison: ECT + C.ATP vs
ECT + C.placebo
Inclusion: endogenous or mixed
endogenous depression
Allocation: B,
unclear
Interventions
Arfwidsson
et al., 197358
Participants
Methods
Study
TABLE 33 RCTs of ECT plus pharmacotherapy versus ECT plus placebo/pharmacotherapy only: depression (cont’d)
Health Technology Assessment 2005; Vol. 9: No. 9
129
130
Kay et al.,
197063
Allocation: B,
unclear
Imlah et al.,
196562
Blinding: doubleblind
Allocation: B,
unclear
Blinding: unclear
Methods
Study
History: mostly inpatients, none with
ECT over the past 6 months, no
restriction on prior drug therapy
Gender: 48 M; 84 F. Gender differences
between groups were non-significant
Age: overall age range 20–75 years, with
>50% 40–59. Age differences between
groups were ‘non-significant’
Exclusion: organic brain disease,
schizophrenia or subnormality
Inclusion: affective disorders
History: 54% had duration of illness
<6 months, 26% 6–12 months and 20%
>12 months
Gender: 53 M, 97 F
Age: 32% aged <40, 63% aged 40–60
and 5% aged >60 years
Inclusion: suffering from depressive
illness of sufficient degree to warrant use
of ECT
Participants
Comparator: amitriptyline (TCA):
25 mg, three tablets at start (two to
six tablets) daily at doctor’s
discretion, 1-month trial. Diazepam
(benzodiazepine): 2 mg, three
tablets at start (two to six tablets)
daily at doctor’s discretion,
1-month trial
ECT: no details; 1-month trial
Comparison: ECT + TCA vs ECT
+ diazepam
Comparator: placebo: one tablet;
imipramine: 25 mg t.d.s.;
phenelzine: 15 mg t.d.s.
ECT: twice weekly, discontinued
when two observers agreed that
the patient had reached a maximal
response and discontinued after 12
weeks in those who had residual
symptoms
Comparison: ECT + C.MAOI vs
ECT + C.TCA vs ECT +
C.placebo
Interventions
TABLE 34 RCTs of ECT plus pharmacotherapy/placebo versus continuation pharmacotherapy: depression
n completed: 111
Dichotomous: clinical
opinion of relapse (not
defined)
continued
Dichotomous: clinical failure Length of follow-up:
defined as removal from trial 3 months
owing to relapse,
unsatisfactory progress, sideeffects, taking an overdose
Continuous: Mood rating,
N randomised: 132
HRSD, BDI, Lubin (unusable,
no SD)
n completed: 53
Length of follow-up:
6 months
N randomised: 150
Number and follow-up
Continuous: none
Outcomes
Appendix 5
Methods
Blinding: doubleblind
Seager and Bird, Allocation: B,
196261
unclear
Study
Age: ECT + imipramine: mean 47.9
(range 28–71) years; ECT + placebo: 49
(30–70) years
Inclusion: inpatients suffering from a
depressive illness of moderate to severe
intensity, with retardation or agitation,
feelings of hopelessness and pessimism,
warranting electrical treatment
Participants
Comparator: imipramine: 25 mg
t.d.s. for 3 days increased to 50 mg
for hospital and first month after
treatment, then reduced to 25 mg;
placebo: identical in appearance
ECT: modified ECT twice a week
using an Ecton machine (1-s
duration shock), number of
treatments based on clinical
opinion; no information on
electrode placement
Comparison: ECT + C.TCA vs
ECT + C.placebo
Interventions
TABLE 34 RCTs of ECT plus pharmacotherapy/placebo versus continuation pharmacotherapy: depression (cont’d)
n completed: 28
Dichotomous: clinical
opinion of a satisfactory
response or a relapse (not
defined)
continued
Length of follow-up:
6 months
N randomised: 43
Number and follow-up
Continuous: none
Outcomes
Health Technology Assessment 2005; Vol. 9: No. 9
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
131
132
Allocation: B,
unclear
Lauritzen et al.,
199664
Group B:
Age: paroxetine: mean (SD) 55.9 (12.7) years; imipramine:
63.3 (11.5) years
Gender: paroxetine: 3 M, 24 F; imipramine: 9 M, 16 F
History: number of previous depressive episodes:
paroxetine 2.9; imipramine: 2.4. Bipolar/unipolar:
paroxetine: 7/20; imipramine: 2/23. Mean duration of
current episode: paroxetine: mean (SD) 17.2 (13.5) weeks;
imipramine: 12.8 (8.3) weeks. Received treatment for
current episode: paroxetine: 92%; imipramine 84%
Group A:
Age: paroxetine: mean (SD) 71.4 (8.5) years; placebo:
73.0 (8.5) years
Gender: paroxetine: 7 M, 11 F; placebo: 4 M, 13 F
History: number of previous depressive episodes:
paroxetine: 2.1; placebo: 3.8. Bipolar/unipolar: paroxetine:
7/11; placebo: 4/13. Mean duration of current episode:
paroxetine: mean (SD) 19.1 (9.5) weeks; placebo: 22.4
(24.9) weeks. Received treatment for current episode:
paroxetine: 90%; placebo: 76%
Comparator:
group A: paroxetine
(30 mg day–1) or placebo
group B: paroxetine
(30 mg day–1) or imipramine
(150 mg day–1)
ECT: EEG-monitored ECT was Dichotomous: no
data
applied, three sessions per
week, total number of sessions
decided by the treating clinician.
Bilateral placement for the first
three sessions; thereafter, nondominant ECT. Stimulation levels
adjusted by patient over sessions
Exclusion: severe cardiovascular disease within the
preceding 6 months, including intraventricular conduction
abnormalities, severe unstabilised somatic diseases,
untreated glaucoma, dementia, schizophrenia, chronic
alcohol/drug abuse, treatment with irreversible MAOIs
within the preceding 14 days, pregnancy/nursing mothers,
epilepsy, prophylactic lithium treatment
Group A:
N randomised: 35
Continuous:
HRSD, Newcastle
scale, Melancholia
scale
All:
Length of follow-up:
6 months
n completed: 45
Group B:
N randomised: 52
n completed: 33
Number and follow-up
Outcomes
Comparison: group A: ECT +
C.SSRI vs ECT + C.placebo;
group B: ECT + C.SSRI vs ECT
+ C.TCA
Interventions
Inclusion: major depressive episode in accordance with
DSM-III-R, HRSD score of ≥ 18, age ≥ 18 years, ability to
understand oral and written information about the trial and
giving informed consent
Participants
C.MAOI, continuation therapy with monoamine oxidase inhibitors; C.TCA, continuation therapy with tricyclic antidepressants; C.SSRI, continuation therapy with selective serotonin
reuptake inhibitors.
Blinding:
patient
Methods
Study
TABLE 34 RCTs of ECT plus pharmacotherapy/placebo versus continuation pharmacotherapy: depression (cont’d)
Appendix 5
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Blinding: doubleblind
Grunhaus et al., Allocation: B,
200167
unclear
Exclusion: not recorded
Blinding: doubleblind
Comparison: C.SSRI vs C.SSRI + melatonin
ECT: started on unilateral but switched to
bilateral if not achieved decrease of 30% in
baseline HRSD scores by sixth treatment.
Age: fluoxetine + melatonin: mean Seizure threshold determined by method of
(SD) 61.1 (10.7) years; fluoxetine + limits and second treatment delivered at
placebo: 59.6 (14.1) years
2.5 times threshold; at following sessions
electrical parameters were set to deliver
History: duration of illness:
seizures of > 25 s
fluoxetine + melatonin: mean (SD)
of 6.6 (8.3 months); fluoxetine +
Comparator: fluoxetine + melatonin: 7 days
placebo: 8.7 (7.6 months). Referred post-ECT 20 mg fluoxetine daily plus 5 mg
to ECT because of medication
slow-release melatonin 3 h before bedtime.
resistance, presence of delusions or Following 3 months received 20–40 mg
hallucinations and/or very severe
fluoxetine plus 5 or 10 mg melatonin.
depressive illness
Fluoxetine + placebo: 7 days post-ECT
20 mg fluoxetine daily plus 5 mg placebo 3 h
before bedtime. Following 3 months received
20–40 mg fluoxetine plus 5 or 10 mg placebo
Inclusion: successfully responded
to a course of ECT (post-HRSD
17-item ≤ 10 maintained for
1 week)
History: for 12 patients this was
the first episode of depression. No
history of mania. Number of
previous episodes: placebo: mean
(SD) 2.2 (0.5); lithium: 1.6 (0.4)
Age: placebo: mean (SD) 54.0 (2.8) Comparator: lithium carbonate (Priadel,
years; lithium: 56.2 (3.0) years
Delandale: antimanic drugs). Lithium plasma
maintained throughout between 0.8 and
Gender: placebo: 8 M, 12 F;
1.2 mmol l–1
lithium: 6 M, 12 F
ECT: not described
Inclusion: MDD with HRSD scores Comparison: continuation lithium vs
of ≥ 16
C.placebo
Allocation: B,
unclear
Interventions
Coppen et al.,
198166
Participants
Methods
Study
TABLE 35 RCTs comparing continuation pharmacotherapy only
Dichotomous: relapse
defined as return of five
or more DSM-IV
symptoms of major
depression and HRSD of
≥ 16
Continuous: HRSD,
BPRS, GRSD, MMSE,
PSQI
Dichotomous: none
Continuous: HRSD
(unusable graph only), no
weeks with depression
Outcomes
continued
Length of follow-up:
3 months
n completed: 35
N randomised: 39
Length of follow-up:
1 year
n completed: 38
N randomised: 38
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
133
134
Allocation: B,
unclear
Sackeim et al.,
200168
Blinding: doubleblind
Methods
Study
Interventions
Outcomes
Number and follow-up
History: psychotic: placebo 44.8%;
nortriptyline + placebo: 37.0%; nortriptyline
+ lithium: 42.9%. Medication resistant:
placebo: 48.3%; nortriptyline + placebo:
44.4%; nortriptyline + lithium: 50.0%
Gender: placebo: 31.0% M, 69.0% F;
nortriptyline + placebo: 29.5% M, 70.4% F;
nortriptyline + lithium: 39.3% M, 60.7% F
Inclusion: ECT remitters (improvement of
>60% reduction in HRSD score) randomised
to three continuation pharmacotherapy
groups, stratified by classification of the index
episode as psychotic depression, medicationresistant non-psychotic depression and nonpsychotic depression without medication
resistance
Comparison: C.TCA vs C.TCA +
lithium vs C.placebo
Continuous: HRSD, CGI, N randomised: 84
GAF
n completed: 73
Dichotomous: relapse
ECT: based on clinical judgement:
defined as mean HRSD
Length of follow-up:
either unilateral or bilateral ECT
24 weeks
using the d’Elia or bifrontotemporal (continuous rater and
placement respectively. Three times study psychiatrist) of ≥ 16
that was maintained for
per week. Seizure threshold
≥ 1 week
calculated at first treatment using
Exclusion: history of bipolar disorder,
empirical titration; minimal duration
schizophrenia, schizoaffective disorder, non- 20 s of motor/25 s EEG. Length of
mood disorder psychosis, neurological illness, ECT course determined on clinical
alcohol or drug abuse within the past year,
grounds
ECT within the past 6 months, or severe
medical illness that markedly increased the
Comparator: nortriptyline (TCA)
risks of ECT. Patients with medical
25 mg; lithium (antimanic) 300 mg;
contraindications to nortriptyline or lithium
oral doses adjusted to maintain
plasma levels at 17–125 ng ml–1
(nortriptyline) and 0.7 mEq l–1
Age: placebo: mean (SD) 55.8 (13.6) years;
(lithium)
nortriptyline + placebo: 57.2 (19.8) years;
nortriptyline + lithium: 59.2 (18.3) years
Participants
TABLE 35 RCTs comparing continuation pharmacotherapy only (cont’d)
Appendix 5
Methods
Inclusion: not reported
Participants
Comparison: video vs no video
Interventions
Blinding: not blind
Westreich et al., Allocation: A,
concealed
199569
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
History: number of past ECT courses: video: mean
(SD) 2.57 (3.95), no video: 1.00 (1.34). Score on BPRS:
video: 34.71 (7.32); no video: 40.00 (5.04)
Gender: not reported
Age: video: median 63 years; no video: 65 years
Exclusion: non-English speaking
Inclusion: drawn from geropsychiatric inpatient unit
and two general psychiatry inpatient units, English
speaking
Continuous: MMSE and
BPRS as measures of
illness severity, eight item
knowledge questionnaire
Dichotomous: none
Continuous: knowledge,
behavioural intent, fear
Outcomes
Written consent alone: received written consent
form only before giving consent to ECT
Video + written consent: received information video
on ECT and written consent form before giving
Dichotomous: none
consent to ECT
ECT: no ECT involved
Comparison: video vs no video
Exclusion: acute or chronic brain disorder, dysfunction ECT: no ECT involved
Blinding: not blind or distress to limit participation. Patients about to have
Video: watched a video of a psychiatrist interviewing a
ECT were excluded
depressed elderly inpatient before receiving ECT.
Age: not reported
Interspersed were segments of her receiving ECT, a
post-ECT interview and her leaving hospital well.
Gender: not reported
Psychiatrists discussed ECT itself, its indications and
side-effects. No person was interviewed who
History: outpatients and admission to psychiatric ward expressed dissatisfaction with ECT or had a negative
with spectrum of diagnoses of psychotic, neurotic and outcome with ECT
personality disorders
No video: usual care, did not watch a video
Battersby et al., Allocation: B,
199392
unclear
Study
TABLE 36 RCTs of patient information videos
Health Technology Assessment 2005; Vol. 9: No. 9
135
136
Kroessler and
Fogel, 199376
All patients who received ECT at Rhode Island
Hospital between 1974 and 1983 who were
aged >80 years when admitted and who had a
discharge diagnosis of MDD according to either
DSM-II or ICD-9 or 8, and were treated with
ECT or pharmacotherapy. Some patients from
the pharmacotherapy group were recruited from
another hospital
Design: cohort (retrospective)
Quality assessment: no
control of confounding factors,
unblinded outcome assessment
Quality assessment: some
control of confounding by
matching and blinding:
Comparison: patients aged ≥ 75 years treated
comparison treatments and
pharmacologically, computer matched by age,
length of follow-up not reported gender and discharge diagnosis
Outcomes
Pharmacotherapy: TCAs (n = 20),
benzodiazepines (n = 15), trazodone (n = 6),
neuroleptics (n = 5), chloral hydrate (n = 2),
lithium carbonate (n = 2), maprotiline (n = 1),
carbamazepine (n = 1) and nomifensine (n = 1)
ECT: mean number of ECTs received was 7.9
(SD 2.9). No information on electrode
placement, dosage or waveform used. Two
patients had only two ECTSs, one patient
withdrew consent and one developed congestive
heart failure and died before treatment could be
continued
continued
Mortality, survival,
recurrence of depression,
rehospitalisation,
additional ECT and
residence following
hospitalisation
Response to treatment
(good, moderate, poor),
complications including
falls, CVD, confusion,
gastrointestinal,
pulmonary, metabolic and
Pharmacotherapy: no information provided on total complications
drugs received by the pharmacology group
ECT: administered two or three times per week
using a brief-pulse device (Mecta SRI). 19 patients
received bilateral ECT, nine right unilateral, nine
both bilateral and unilateral, and in two patients
it was not noted
ECT: patients aged ≥ 75 years, diagnosed with
major depression, who had received ECT
between 1987 and 1993; 3 M, 36 F
Design: cohort (retrospective)
Manly et al.,
200075
Interventions
Participants
Methods
Study
TABLE 37 Non-randomised evidence of efficacy of ECT in older people with depression
Appendix 5
Rubin et al.,
199173, 199374
Design: cohort (retrospective)
Philibert et al.,
199577
All patients aged >65 years and
admitted to hospital meeting the
DSM-III criteria for unipolar
depression between 1980 and
1987, identified by computerised
search
Participants
All patients with a major affective
disorder (either unipolar or
Quality assessment: some
bipolar), without other psychiatric
control over confounding
diagnoses and without possible or
variables using statistical analyses probable dementia admitted to an
and exclusions, unblinded
inpatient unit for people aged
outcome assessment but loss to >65 years
follow-up reported
Design: cohort (prospective)
Quality assessment: no
control of confounding factors,
unblinded outcome assessment
Methods
Study
TABLE 37 Non-randomised evidence of efficacy of ECT in older people with depression (cont’d)
Outcomes
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Pharmacotherapy: both the nonECT group and the ECT group
received pharmacotherapy, and the
type and dose of treatment
including TCAs, antipsychotics,
lithium and antianxiety agents, were
determined by the treating
physician
ECT: three times per week at a
GDS, BDI, MMSE and
moderately suprathreshold dose
length of stay
using a Mecta SRI brief-pulse
device. 36 patients received
bilateral ECT, six received unilateral
ECT using the d’Elia placement and
six received both. Seizures were
monitored using
electroencephalography. The mean
(SD) number of treatments was 9.3
(3)
Pharmacotherapy: no information
provided on treatment received by
those not receiving ECT
ECT: mean (SD) number of ECTs
Global improvement and
10.7 (4.1). ECT administered three all-cause mortality
times per week; both unilateral and
bilateral ECT was used, but no
information is provided on the
numbers receiving either treatment
Interventions
Lost to follow-up: 7/48
in ECT group; 8/55 in
control group
N: 103
Lost to follow-up:
unclear
N: 192
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
137
138
Design: case–control
(retrospective)
Cohen et al.,
200072
Participants
20 adolescents treated with ECT
for a mood disorder before the age
of 19 in three adolescent units and
Quality assessment: large loss three adult clinics in Paris between
to follow-up, no control of
1987 and 1996; only ten were
confounding variables, unblinded included in the study (six women,
outcome assessment
four men). Five had major
depression with psychotic features,
three had manic depression with
psychotic features and two had
mixed depression with psychotic
features. Ten matched controls had
never received ECT
Methods
Study
TABLE 38 Non-randomised evidence: children and adolescents
Outcomes
Number and follow-up
ECT: bilateral ECT between 2 and Clinical judgement of
N: 30
9 years before interviews. Received improvement, relapses
a mean of 9.8 ECTs
and various cognitive test Test to follow-up: 10
including MMSE, WMS
Comparison: no information on
and California Verbal
treatment received
Learning Test. Perceptions
of the adequacy of ECT
information and of the
perceived benefit of ECT
Interventions
Appendix 5
Malur et al.,
200180
Design: case series
(prospective)
Bush et al.,
199679
Design: case series
(prospective)
Quality assessment: no
control group, unblinded
assessment of outcome, loss to
follow-up
Methods
Study
TABLE 39 Non-randomised evidence: catatonia
Comparison: none
ECT: in five patients the symptoms
of catatonia resolved 2 days before
treatment; two patients were
withdrawn. 21 patients received a
full trial of lorazepam for up to
5 days. 16/21 had signs of catatonia
relieved and 11 of these had a full
resolution of catatonic symptoms.
The five non-responders were
treated with ECT; one refused
consent
Interventions
BFCRS scores
Outcomes
BFCRS
ECT: case 1: lorazepam max.
12 mg day–1 for 5.5 weeks,
resulting in a BFCRS score of 15,
followed by 15 bilateral ECTs over
a 6-week period; case 2: lorazepam
max. 4 mg day–1 for 10 weeks,
resulting in a BFCRS score of 10,
followed by 14 bilateral ECTs over
5 weeks; case 3: lorazepam max.
Case 2: aged 26 years, female,
systemic lupus erythematosus, four 16 mg day–1 for 3 weeks, resulting
signs of catatonia with a duration of in a BFCRS score of 10, followed by
22 bilateral ECTs over 3 months. All
14 weeks, BFCRS score 14,
ECTs were administered using a
respiratory acidosis
Thymatron DG device with
bidirectional brief-pulse square
Case 3: aged 39 years, male,
current three times per week.
hypertension, schizoaffective
Initial stimulus intensity was 50% in
disorder and mild mental
retardation, four signs of catatonia case 1, 20% in case 2 and 40% in
case 3
with a duration of 10 weeks,
BFCRS score 16, definite NMS,
acute respiratory insufficiency
Case 1: aged 24 years, female, no
known medical or psychiatric
history, seven catatonic signs with a
duration of 14 weeks before ECT,
BFCRS score of 19, probable NMS,
respiratory acidosis and cardiac
asystole
Those treated with ECT were
those who failed to respond to
lorazepam (5/28): 3/5 had mania,
three were women, two were men
and the mean duration of catatonia
was 11 days (SD 12.1)
Participants
N: 3
N=5
4 treated, 1 refused
consent
Number and follow-up
Health Technology Assessment 2005; Vol. 9: No. 9
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
139
140
Aged 25 years, 8 weeks of gestation.
ECT: bilateral ECT with sine wave
Diagnosed with severe depression with
of 2.5-s duration at an intensity of
psychotic symptoms. Initially treated with
0.7 A for nine treatments
levopromazine (25 mg i.m.) and haloperidol
(5 mg) then changed to amitryptaline (75 mg),
haloperidol (10 mg) and carbemazepine
(1,200 mg). Treatment with amitryptaline and
carbemazepine was stopped when a second
pregnancy test was positive
Aged 29 years, white, in week 23 of
pregnancy. History of paranoid schizophrenia
and depression. Current episode became
catatonic and suicidal. Did not respond to
resperidone (3 mg b.d.), loxapine (75 mg
b.d.), lorazepam (1 mg t.d.s.) or nortriptyline
(50 mg)
Design: case report
(prospective)
Design: case report
Polster and
Wisner, 199984
Outcomes
ECT: unilateral ECT, pulse width
Clinical improvement,
1.2 ms, frequency 50 Hz, current
adverse events
0.6 A and seizure length 89 s for
eight treatments followed by
bilateral ECT three times per week
for 3.5 weeks
Clinical opinion of
efficacy, adverse events
ECT: case 1: bilateral ECT three
Clinical opinion on
times per week for six treatments
efficacy, complications
in delivery room; case 2: bilateral
ECT, five treatments, one on day 1,
two on day 2 and two on day 3
Moreno et al.,
199883
Case 2: aged 23 years, white gravida at
27 weeks of gestation. Pregnancy
complicated by generalised anxiety disorder
with panic attacks and depression resulting in
weight loss and an episode of threatened
abortion. Failed to respond to desipramine
400 mg/day–1, oxazepam 15 mg q.d.s. and
tryptophan 1 g every bedtime
Case 1: aged 26 years, white primagravida
at 35 weeks of gestation, uncomplicated
pregnancy. Current episode treated with
desipramine (150 g day–1) and lorazepam
(0.5 mg t.d.s.)
Design: case series
(prospective)
Interventions
Bhatia et al.,
199982
Participants
Methods
Study
TABLE 40 Non-randomised evidence: pregnancy
N: 1
N: 1
N: 2
Number
Appendix 5
Health Technology Assessment 2005; Vol. 9: No. 9
Appendix 6
Results of meta-analyses
141
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
142
FIGURE 4b
45
49
Total (95% CI)
Total events: 30 (real ECT), 17 (sham ECT)
Test for heterogeneity: 2 = 1.64, df = 1 (p = 0.20)
Test for overall effect: Z = 1.63 (p = 0.10)
Sham ECT
n/N
9/33
8/12
Real ECT
n/N
20/37
10/12
Johnstone et al.97
Jagadeesh et al.52
Study
or subcategory
Comparison:
Outcome:
01 Real bilateral ECT vs sham ECT
02 Improvement excluding Freeman et al. (1978)96
65
69
Total (95% CI)
Total events: 46 (real ECT), 37 (sham ECT)
Test for heterogeneity: 2 = 15.42, df = 2 (p = 0.0004)
Test for overall effect: Z = 0.55 (p = 0.59)
FIGURE 4a
20/20
9/33
8/12
Sham ECT
n/N
16/20
20/37
10/12
Real ECT
n/N
01 Real bilateral ECT vs sham ECT
01 Improvement (all studies)
Freeman et al.96
Johnstone et al.97
Jagadeesh et al.52
Study
or subcategory
Comparison:
Outcome:
0.2
0.5
0.2
0.5
Favours treatment
0.1
1
1
5
2
Favours control
5
Favours control
2
RR (random)
95% CI
Favours treatment
0.1
RR (random)
95% CI
10
10
1.98 (1.05 to 3.73)
1.25 (0.78 to 2.01)
1.51 (0.92 to 2.49)
100.00
RR (random)
95% CI
1.21 (0.61 to 2.40)
0.80 (0.64 to 1.00)
1.98 (1.05 to 3.73)
1.25 (0.78 to 2.01)
RR (random)
95% CI
41.44
58.56
Weight
%
100.00
37.66
29.36
32.98
Weight
%
Appendix 6
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
FIGURE 5
Total (95% CI)
Total events: 9 (unilateral ECT), 9 (sham ECT)
Test for heterogeneity: not applicable
Test for overall effect: Z = 0.00 (p = 1.00)
9/16
16
9/16
16
Unilateral ECT
n/N
Sham ECT
n/N
33
37
02 Real unilateral ECT vs sham ECT
01 Improvement (clinical opinion)
Lambourn and Gill95
Study
or subcategory
Comparison:
Outcome:
FIGURE 4c
Total (95% CI)
Total events: 20 (real ECT), 9 (sham ECT)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.12 (p = 0.03)
9/33
Sham ECT
n/N
20/37
Real ECT
n/N
0.2
0.5
0.2
0.5
Favours treatment
0.1
2
1
5
2
Favours control
5
Favours control
RR (random)
95% CI
1
RR (random)
95% CI
Favours treatment
0.1
01 Real bilateral ECT vs sham ECT
03 Improvement: all trials that did not give real bilateral ECT to the control arm
Johnstone et al.97
Study
or subcategory
Comparison:
Outcome:
10
10
100.00
100.00
1.00 (0.54 to 1.84)
1.00 (0.54 to 1.84)
RR (random)
95% CI
1.98 (1.05 to 3.73)
100.00
Weight
%
1.98 (1.05 to 3.73)
RR (random)
95% CI
100.00
Weight
%
Health Technology Assessment 2005; Vol. 9: No. 9
143
144
21/23
12/15
46/63
41/74
3/4
14/15
Bruce et al.104
Robin and Harris115
Greenblatt and Grosser113
Shepherd100
Steiner et al.107
Janakiramaiah et al.103
FIGURE 6a
194
Total (95% CI)
Total events: 137 (ECT), 99 (TCAs)
Test for heterogeneity: 2 = 6.79, df = 5 (p = 0.24)
Test for overall effect: Z = 3.47 (p = 0.0005)
ECT
n/N
03 ECT vs TCAs
01 Marked or moderate improvement (all studies)
Study
or subcategory
Comparison:
Outcome:
200
16/27
3/16
36/73
30/65
3/4
11/15
TCAs
n/N
0.2
0.5
Favours treatment
0.1
1
5
Favours control
2
RR (random)
95% CI
10
100.00
21.27
3.09
27.27
21.65
5.16
21.56
Weight
%
1.40 (1.16 to 1.69)
1.54 (1.10 to 2.16)
4.27 (1.49 to 12.20)
1.48 (1.12 to 1.95)
1.20 (0.86 to 1.67)
1.00 (0.45 to 2.23)
1.27 (0.91 to 1.78)
RR (random)
95% CI
Appendix 6
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
190
194
Total (95% CI)
Total events: 29 (ECT), 65 (TCAs)
Test for heterogeneity: 2 = 4.75, df = 5 (p = 0.45)
Test for overall effect: Z = 3.77 (p = 0.0002)
FIGURE 6b
11/27
7/16
19/63
23/65
1/4
4/15
2/23
2/15
6/63
17/74
1/4
1/15
Bruce et al.104
Robin and Harris115
Greenblatt and Grosser113
Shepherd100
Steiner et al.107
Janakiramaiah et al.103
TCAs
n/N
ECT
n/N
03 ECT vs TCAs
02 No improvement (all studies)
Study
or subcategory
Comparison:
Outcome:
0.2
0.5
Favours treatment
0.1
1
5
Favours control
2
RR (random)
95% CI
10
100.00
8.02
7.97
21.83
55.78
2.73
3.67
Weight
%
0.47 (0.31 to 0.69)
0.21 (0.05 to 0.87)
0.30 (0.07 to 1.24)
0.32 (0.14 to 0.74)
0.65 (0.38 to 1.10)
1.00 (0.09 to 11.03)
0.25 (0.03 to 1.98)
RR (random)
95% CI
Health Technology Assessment 2005; Vol. 9: No. 9
145
146
19
19
Total (95% CI)
Total events: 17 (ECT), 14 (TCAs)
Test for heterogeneity: 2 = 0.30, df = 1 (p = 0.58)
Test for overall effect: Z = 1.31 (p = 0.19)
FIGURE 6d
3/4
11/15
3/4
14/15
Steiner et al.107
Janakiramaiah et al.103
TCAs
n/N
ECT
n/N
Study
or subcategory
Comparison:
Outcome:
03 ECT vs TCAs
04 Improvement (quantitative definition)
181
165
Total (95% CI)
Total events: 122 (ECT), 85 (TCAs)
Test for heterogeneity: 2 = 7.44, df = 3 (p = 0.06)
Test for overall effect: Z = 3.22 (p = 0.001)
FIGURE 6c
16/27
3/16
36/73
30/65
21/23
12/15
48/53
41/74
Bruce et al.104
Robin and Harris115
Greenblatt and Grosser113
Shepherd100
TCAs
n/N
ECT
n/N
03 ECT vs TCAs
03 Improvement (clinical opinion only)
Study
or subcategory
Comparison:
Outcome:
0.2
0.5
0.2
0.5
Favours treatment
0.1
1
1
5
2
Favours control
5
Favours control
2
RR (random)
95% CI
Favours treatment
0.1
RR (random)
95% CI
10
10
100.00
14.82
85.18
1.23 (0.90 to 1.67)
1.00 (0.45 to 2.23)
1.27 (0.91 to 1.78)
RR (random)
95% CI
1.63 (1.21 to 2.20)
100.00
Weight
%
1.54 (1.10 to 2.16)
4.27 (1.49 to 12.20)
1.84 (1.43 to 2.35)
1.20 (0.86 to 1.67)
RR (random)
95% CI
28.98
6.87
34.88
29.27
Weight
%
Appendix 6
20
Grunhaus et al.54
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
24/28
FIGURE 8a
28
Total (95% CI)
Total events: 24 (ECT + chlorpromazine), 22 (ECT + placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 0.94 (p = 0.35)
Arfwidsson et al.58
29
22/29
ECT + placebo
n/N
ECT + chlorpromazine
n/N
10.40 (7.54)
rTMS
Mean (SD)
Study
or subcategory
20
20
N
05 ECT plus chlorpromazine vs ECT plus placebo
01 Improvement at end of ECT course (clinical opinion)
17.20 (9.72)
ECT
Mean (SD)
Comparison:
Outcome:
FIGURE 7
20
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.47 (p = 0.01)
N
04 ECT vs rTMS
01 Improvement (HRSD)
Study
or subcategory
Comparison:
Outcome:
–5
0.2
0.5
Favours treatment
0.1
0
1
5
2
Favours control
5
Favours control
RR (random)
95% CI
Favours treatment
–10
WMD (random)
95% CI
10
10
1.13 (0.88 to 1.46)
1.13 (0.88 to 1.46)
100.00
RR (random)
95% CI
6.80 (1.41 to 12.19)
6.80 (1.41 to 12.19)
WMD (random)
95% CI
100.00
Weight
%
100.00
100.00
Weight
%
Health Technology Assessment 2005; Vol. 9: No. 9
147
148
4/9
ECT + pindolol
n/N
06 ECT + pindolol vs ECT + placebo
01 Improvement at end of ECT course
FIGURE 9a
9
Total (95% CI)
Total events: 4 (ECT + pindolol), 0 (ECT + placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 1.67 (p = 0.10)
Shiah et al.57
Study
or subcategory
Comparison:
Outcome:
FIGURE 8b
28
Total (95% CI)
Total events: 18 (ECT + chlorpromazine), 16 (ECT + diazepam)
Test for heterogeneity: not applicable
Test for overall effect: Z = 0.70 (p = 0.48)
18/28
ECT + chlorpromazine
n/N
05 ECT plus chlorpromazine vs ECT plus placebo
02 Relapse at three months
Arfwidsson et al.58
Study
or subcategory
Comparison:
Outcome:
11
0/11
ECT + placebo
n/N
29
16/29
ECT + diazepam
n/N
0.2
0.5
0.2
0.5
Favours treatment
0.1
1
2
1
5
2
Favours control
5
Favours control
RR (random)
95% CI
Favours treatment
0.1
RR (random)
95% CI
10
10
100.00
100.00
Weight
%
100.00
100.00
Weight
%
10.80 (0.66 to 177.36)
10.80 (0.66 to 177.36)
RR (random)
95% CI
1.17 (0.76 to 1.79)
1.17 (0.76 to 1.79)
RR (random)
95% CI
Appendix 6
8
Shiah et al.57
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
29/31
FIGURE 10
31
Total (95% CI)
Total events: 29 (ECT + L-tryptophan), 27 (ECT + placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 0.50 (p = 0.62)
d’Elia et al.60
30
27/30
ECT + placebo
n/N
ECT + L-tryptophan
n/N
Study
or subcategory
20.90 (5.20)
ECT + placebo
Mean (SD)
07 ECT + L-tryptophan vs ECT + placebo
01 Improvement at end of ECT course (clinical opinion)
7
7
N
Comparison:
Outcome:
FIGURE 9b
11.80 (8.40)
ECT + pindolol
Mean (SD)
8
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.56 (p = 0.01)
N
06 ECT + pindolol vs ECT + placebo
02 Improvement at end of ECT course (HRSD)
Study
or subcategory
Comparison:
Outcome:
–5
0.2
0.5
Favours treatment
0.1
0
1
5
Favours control
2
5
Favours control
RR (random)
95% CI
Favours treatment
–10
WMD (random)
95% CI
10
10
1.04 (0.89 to 1.21)
1.04 (0.89 to 1.21)
100.00
100.00
RR (random)
95% CI
–9.10 (–16.08 to –2.12)
100.00
Weight
%
–9.10 (–16.08 to –2.12)
WMD (random)
95% CI
100.00
Weight
%
Health Technology Assessment 2005; Vol. 9: No. 9
149
150
21
Folkerts et al.112
12.50 (3.90)
ECT
Mean (SD)
22
21
Total (95% CI)
Total events: 15 (ECT), 5 (paroxetine)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.75 (p = 0.006)
FIGURE 11b
5/22
15/21
Folkerts et al.112
Paroxetine
n/N
23.00 (10.40)
Paroxetine
Mean (SD)
ECT
n/N
09 ECT vs SSRIs
02 Improvement (50% reduction in HRSD)
22
22
N
Study
or subcategory
Comparison:
Outcome:
FIGURE 11a
21
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 4.42 (p = 0.00001)
N
09 ECT vs SSRIs
01 Improvement (HRSD scores)
Study
or subcategory
Comparison:
Outcome:
–5
0.2
0.5
Favours treatment
0.1
0
1
5
2
Favours control
5
Favours control
RR (random)
95% CI
Favours treatment
–10
WMD (random)
95% CI
10
10
100.00
100.00
3.14 (1.39 to 7.11)
3.14 (1.39 to 7.11)
RR (random)
95% CI
–10.50 (–15.15 to –5.85)
100.00
Weight
%
–10.50 (–15.15 to –5.85)
WMD (random)
95% CI
100.00
Weight
%
Appendix 6
15/59
ECT + amitryptaline
n/N
10 ECT + TCA vs ECT + diazepam
01 Relapse by end of ECT course
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
18
Lauritzen et al.64
FIGURE 13
8.90 (4.70)
ECT + paroxetine
Mean (SD)
18
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 0.22 (p = 0.83)
N
11 ECT + SSRI vs ECT + placebo
01 Improvement at end of ECT course (HRSD)
Study
or subcategory
Comparison:
Outcome:
FIGURE 12
59
Total (95% CI)
Total events: 15 (ECT + amitryptaline), 34 (ECT + diazepam)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.37 (p = 0.02)
Kay et al.63
Study
or subcategory
Comparison:
Outcome:
17
17
N
9.20 (3.40)
ECT + placebo
Mean (SD)
73
34/73
ECT + diazepam
n/N
0.2
0.5
–5
Favours treatment
–10
1
2
0
5
Favours control
5
Favours control
WMD (random)
95% CI
Favours treatment
0.1
RR (random)
95% CI
10
10
100.00
100.00
–0.30 (–3.01 to 2.41)
–0.30 (–3.01 to 2.41)
WMD (random)
95% CI
0.55 (0.33 to 0.90)
100.00
Weight
%
0.55 (0.33 to 0.90)
RR (random)
95% CI
100.00
Weight
%
Health Technology Assessment 2005; Vol. 9: No. 9
151
152
25
Lauritzen et al.64
FIGURE 15
18
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.90 (p = 0.004)
1.50 (0.80)
18
20
20
2.40 (1.10)
ECT + placebo
Mean (SD)
Coppen et al.66
N
ECT + lithium
Mean (SD)
N
–5
–5
Favours treatment
–10
0
5
Favours control
5
Favours control
WMD (fixed)
95% CI
0
WMD (random)
95% CI
Favours treatment
–10
9.60 (5.60)
ECT + paroxetine
Mean (SD)
Study
or subcategory
27
27
N
13 ECT + lithium vs ECT + placebo
01 Number of weeks spent depressed during 6 months following ECT
6.60 (4.10)
ECT + imipramine
Mean (SD)
Comparison:
Outcome:
FIGURE 14
25
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.22 (p = 0.03)
N
12 ECT + TCA vs ECT + SSRI
01 Improvement at end of ECT course (HRSD)
Study
or subcategory
Comparison:
Outcome:
10
10
–0.90 (–1.51 to –0.29)
–0.90 (–1.51 to –0.29)
100.00
100.00
WMD (fixed)
95% CI
–3.00 (–5.65 to –0.35)
100.00
Weight
%
–3.00 (–5.65 to –0.35)
WMD (random)
95% CI
100.00
Weight
%
Appendix 6
24/50
Imlah et al.62
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
7/50
Imlah et al.62
FIGURE 16b
50
Total (95% CI)
Total events: 7 (ECT + imipramine), 21 (ECT + placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.83 (p = 0.005)
ECT + imipramine
n/N
14 ECT + TCA vs ECT + placebo
02 Relapse at 6 months (non-ITT)
Study
or subcategory
Comparison:
Outcome:
FIGURE 16a
50
Total (95% CI)
Total events: 24 (ECT + imipramine), 30 (ECT + placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 1.19 (p = 0.23)
ECT + imipramine
n/N
14 ECT + TCA vs ECT + placebo
01 Relapse at 6 months (ITT)
Study
or subcategory
Comparison:
Outcome:
50
21/50
ECT + placebo
n/N
50
30/50
ECT + placebo
n/N
0.2
0.5
0.2
0.5
Favours treatment
0.1
1
2
1
5
2
Favours control
5
Favours control
RR (random)
95% CI
Favours treatment
0.1
RR (random)
95% CI
10
10
0.33 (0.16 to 0.71)
0.33 (0.16 to 0.71)
100.00
100.00
RR (random)
95% CI
0.80 (0.55 to 1.15)
100.00
Weight
%
0.80 (0.55 to 1.15)
RR (random)
95% CI
100.00
Weight
%
Health Technology Assessment 2005; Vol. 9: No. 9
153
154
14/28
Nortriptyline + lithium
n/N
FIGURE 18
28
Total (95% CI)
Total events: 14 (nortriptyline + lithium), 25 (placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 2.68 (p = 0.007)
Sackeim et al.68
Study
or subcategory
Comparison:
Outcome:
29
25/29
Placebo
n/N
16 Continuation TCA + lithium vs continuation placebo
01 Relapse at 6 months
29
27
Total (95% CI)
Total events: 17 (nortriptyline), 25 (placebo)
Test for heterogeneity: not applicable
Test for overall effect: Z = 1.90 (p = 0.06)
FIGURE 17
25/29
17/27
Sackeim et al.68
Placebo
n/N
Nortriptyline
n/N
15 Continuation TCA vs continuation placebo
01 Relapse at 6 months
Study
or subcategory
Comparison:
Outcome:
0.2
0.5
0.2
0.5
Favours treatment
0.1
1
2
1
5
2
Favours control
5
Favours control
RR (random)
95% CI
Favours treatment
0.1
RR (random)
95% CI
10
10
100.00
100.00
0.58 (0.39 to 0.86)
0.58 (0.39 to 0.86)
RR (random)
95% CI
0.73 (0.53 to 1.01)
100.00
Weight
%
0.73 (0.53 to 1.01)
RR (random)
95% CI
100.00
Weight
%
Appendix 6
20
6/20
20
Grunhaus et al.67
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
11
Westreich et al.69
FIGURE 20
6.00 (1.41)
Information video
Mean (SD)
11
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 1.51 (p = 0.13)
N
18 Information video in patients about to have ECT
01 Knowledge of ECT
Study
or subcategory
Comparison:
Outcome:
FIGURE 19
Total (95% CI)
Total events: 6 (SSRI + melatonin), 9 (SSRI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 0.96 (p = 0.34)
9/20
SSRI + melatonin
n/N
Study
or subcategory
7
7
N
SSRI
n/N
17 Continuation SSRI + melatonin vs continuation SSRI
01 Relapse at 3 months
Comparison:
Outcome:
0.5
–5
1
2
0
5
Favours control
5
Favours control
WMD (random)
95% CI
Favours treatment
–10
No information video
Mean (SD)
6.81 (0.87)
0.2
Favours treatment
0.1
RR (random)
95% CI
10
10
100.00
100.00
–0.81 (–1.86 to 0.24)
–0.81 (–1.86 to 0.24)
WMD (random)
95% CI
0.67 (0.29 to 1.52)
100.00
Weight
%
0.67 (0.29 to 1.52)
RR (random)
95% CI
100.00
Weight
%
Health Technology Assessment 2005; Vol. 9: No. 9
155
156
35
Battersby et al.92
FIGURE 21
35
Total (95% CI)
Test for heterogeneity: not applicable
Test for overall effect: Z = 1.66 (p = 0.10)
N
12.63 (2.98)
Information video
Mean (SD)
19 Information video in general psychiatric patients
01 Knowledge about ECT
Study
or subcategory
Comparison:
Outcome:
34
34
N
11.35 (3.39)
–5
0
5
Favours control
WMD (random)
95% CI
Favours treatment
–10
No information video
Mean (SD)
10
1.28 (–0.23 to 2.79)
1.28 (–0.23 to 2.79)
0.00
WMD (random)
95% CI
0.00
Weight
%
Appendix 6
Health Technology Assessment 2005; Vol. 9: No. 9
Health Technology Assessment
Programme
Members
Chair,
Professor Tom Walley,
Director, NHS HTA Programme,
Department of Pharmacology &
Therapeutics,
University of Liverpool
Members
Prioritisation Strategy Group
Professor Bruce Campbell,
Consultant Vascular & General
Surgeon, Royal Devon & Exeter
Hospital
Dr John Reynolds, Clinical
Director, Acute General
Medicine SDU, Radcliffe
Hospital, Oxford
Professor Shah Ebrahim,
Professor in Epidemiology
of Ageing, University of
Bristol
Dr Ron Zimmern, Director,
Public Health Genetics Unit,
Strangeways Research
Laboratories, Cambridge
HTA Commissioning Board
Programme Director,
Professor Tom Walley,
Director, NHS HTA Programme,
Department of Pharmacology &
Therapeutics,
University of Liverpool
Professor John Brazier, Director
of Health Economics,
Sheffield Health Economics
Group, School of Health &
Related Research,
University of Sheffield
Chair,
Professor Shah Ebrahim,
Professor in Epidemiology of
Ageing, Department of Social
Medicine, University of Bristol
Dr Andrew Briggs, Public
Health Career Scientist, Health
Economics Research Centre,
University of Oxford
Deputy Chair,
Professor Jenny Hewison,
Professor of Health Care
Psychology, Academic Unit of
Psychiatry and Behavioural
Sciences, University of Leeds
School of Medicine
Dr Jeffrey Aronson
Reader in Clinical
Pharmacology, Department of
Clinical Pharmacology,
Radcliffe Infirmary, Oxford
Professor Ann Bowling,
Professor of Health Services
Research, Primary Care and
Population Studies,
University College London
Professor Andrew Bradbury,
Professor of Vascular Surgery,
Department of Vascular Surgery,
Birmingham Heartlands
Hospital
Professor Nicky Cullum,
Director of Centre for Evidence
Based Nursing, Department of
Health Sciences, University of
York
Dr Andrew Farmer, Senior
Lecturer in General Practice,
Department of Primary Health
Care, University of Oxford
Professor Fiona J Gilbert,
Professor of Radiology,
Department of Radiology,
University of Aberdeen
Professor Adrian Grant,
Director, Health Services
Research Unit, University of
Aberdeen
Professor F D Richard Hobbs,
Professor of Primary Care &
General Practice, Department of
Primary Care & General
Practice, University of
Birmingham
Professor Peter Jones, Head of
Department, University
Department of Psychiatry,
University of Cambridge
Professor Sallie Lamb, Research
Professor in Physiotherapy/CoDirector, Interdisciplinary
Research Centre in Health,
Coventry University
Professor Julian Little,
Professor of Epidemiology,
Department of Medicine and
Therapeutics, University of
Aberdeen
Professor Stuart Logan,
Director of Health & Social
Care Research, The Peninsula
Medical School, Universities of
Exeter & Plymouth
Professor Tim Peters, Professor
of Primary Care Health Services
Research, Division of Primary
Health Care, University of
Bristol
Professor Ian Roberts, Professor
of Epidemiology & Public
Health, Intervention Research
Unit, London School of
Hygiene and Tropical Medicine
Professor Mark Sculpher,
Professor of Health Economics,
Centre for Health Economics,
Institute for Research in the
Social Services, University of York
Professor Martin Severs,
Professor in Elderly Health
Care, Portsmouth Institute of
Medicine
Dr Jonathan Shapiro, Senior
Fellow, Health Services
Management Centre,
Birmingham
Ms Kate Thomas,
Deputy Director,
Medical Care Research Unit,
University of Sheffield
Professor Simon G Thompson,
Director, MRC Biostatistics
Unit, Institute of Public Health,
Cambridge
Ms Sue Ziebland,
Senior Research Fellow,
Cancer Research UK,
University of Oxford
Professor Peter Sandercock,
Professor of Medical Neurology,
Department of Clinical
Neurosciences, University of
Edinburgh
167
Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.ncchta.org)
© Queen’s Printer and Controller of HMSO 2005. All rights reserved.
Health Technology Assessment Programme
Members
Diagnostic Technologies & Screening Panel
Chair,
Dr Ron Zimmern, Director of
the Public Health Genetics Unit,
Strangeways Research
Laboratories, Cambridge
Ms Norma Armston,
Freelance Consumer Advocate,
Bolton
Professor Max Bachmann
Professor Health
Care Interfaces,
Department of Health
Policy and Practice,
University of East Anglia
Professor Rudy Bilous
Professor of Clinical Medicine &
Consultant Physician,
The Academic Centre,
South Tees Hospitals
NHS Trust
Dr Paul Cockcroft,
Consultant Medical
Microbiologist/Laboratory
Director, Public Health
Laboratory, St Mary’s Hospital,
Portsmouth
Members
Chair,
Dr John Reynolds, Clinical
Director, Acute General
Medicine SDU, Oxford
Radcliffe Hospital
Professor Tony Avery,
Professor of Primary Health
Care, University of Nottingham
Professor Adrian K Dixon,
Professor of Radiology,
Addenbrooke’s Hospital,
Cambridge
Dr David Elliman,
Consultant in Community
Child Health, London
Professor Glyn Elwyn,
Primary Medical Care
Research Group,
Swansea Clinical School,
University of Wales
Swansea
Dr John Fielding,
Consultant Radiologist,
Radiology Department,
Royal Shrewsbury Hospital
Mr Tam Fry, Honorary
Chairman, Child Growth
Foundation, London
Dr Margaret Somerville,
Director of Public Health,
Teignbridge Primary Care Trust
Dr Edmund Jessop,
Medical Adviser,
National Specialist
Commissioning Advisory Group
(NSCAG), Department of
Health, London
Professor Lindsay Wilson
Turnbull, Scientific Director,
Centre for MR Investigations &
YCR Professor of Radiology,
University of Hull
Dr Jennifer J Kurinczuk,
Consultant Clinical
Epidemiologist,
National Perinatal
Epidemiology Unit,
Oxford
Dr Susanne M Ludgate, Medical
Director, Medical Devices
Agency, London
Dr Karen N Foster, Clinical
Lecturer, Dept of General
Practice & Primary Care,
University of Aberdeen
Dr William Rosenberg, Senior
Lecturer and Consultant in
Medicine, University of
Southampton
Professor Antony J Franks,
Deputy Medical Director,
The Leeds Teaching Hospitals
NHS Trust
Dr Susan Schonfield, CPHM
Specialised Services
Commissioning, Croydon
Primary Care Trust
Professor Martin J Whittle,
Head of Division of
Reproductive & Child Health,
University of Birmingham
Dr Dennis Wright, Consultant
Biochemist & Clinical Director,
Pathology & The Kennedy
Galton Centre, Northwick Park
& St Mark’s Hospitals,
Harrow
Pharmaceuticals Panel
Dr Christopher Cates, GP and
Cochrane Editor, Bushey Health
Centre
Mrs Sharon Hart, Managing
Editor, Drug & Therapeutics
Bulletin, London
Professor Imti Choonara,
Professor in Child Health,
University of Nottingham,
Derbyshire Children’s Hospital
Dr Christine Hine, Consultant in
Public Health Medicine,
Bristol South & West Primary
Care Trust
Mr Charles Dobson, Special
Projects Adviser, Department of
Health
Professor Stan Kaye,
Professor of Medical Oncology,
Consultant in Medical
Oncology/Drug Development,
The Royal Marsden Hospital
Professor Stirling Bryan,
Professor of Health Economics,
Health Services
Management Centre,
University of Birmingham
Dr Robin Ferner, Consultant
Physician and Director, West
Midlands Centre for Adverse
Drug Reactions, City Hospital
NHS Trust, Birmingham
Mr Peter Cardy, Chief
Executive, Macmillan Cancer
Relief, London
Dr Karen A Fitzgerald,
Pharmaceutical Adviser, Bro Taf
Health Authority, Cardiff
Ms Barbara Meredith,
Project Manager Clinical
Guidelines, Patient Involvement
Unit, NICE
Professor Jan Scott,
Professor of Psychological
Treatments,
Institute of Psychiatry,
University of London
Mrs Katrina Simister, New
Products Manager, National
Prescribing Centre, Liverpool
Dr Richard Tiner, Medical
Director, Association of the
British Pharmaceutical Industry
Dr Helen Williams,
Consultant Microbiologist,
Norfolk & Norwich University
Hospital NHS Trust
Dr Frances Rotblat, CPMP
Delegate, Medicines Control
Agency, London
168
Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.ncchta.org)
Health Technology Assessment 2005; Vol. 9: No. 9
Members
Chair,
Professor Bruce Campbell,
Consultant Vascular and
General Surgeon, Royal Devon
& Exeter Hospital
Dr Mahmood Adil, Head of
Clinical Support & Health
Protection, Directorate of
Health and Social Care (North),
Department of Health,
Manchester
Dr Aileen Clarke,
Reader in Health Services
Research, Public Health &
Policy Research Unit,
Barts & the London School of
Medicine & Dentistry,
Institute of Community Health
Sciences, Queen Mary,
University of London
Therapeutic Procedures Panel
Mr Matthew William Cooke,
Senior Clinical Lecturer and
Honorary Consultant,
Emergency Department,
University of Warwick, Coventry
& Warwickshire NHS Trust,
Division of Health in the
Community, Centre for Primary
Health Care Studies, Coventry
Dr Carl E Counsell, Senior
Lecturer in Neurology,
University of Aberdeen
Dr Keith Dodd, Consultant
Paediatrician, Derbyshire
Children’s Hospital
Professor Gene Feder, Professor
of Primary Care R&D, Barts &
the London, Queen Mary’s
School of Medicine and
Dentistry, University of London
Professor Paul Gregg,
Professor of Orthopaedic
Surgical Science, Department of
Orthopaedic Surgery,
South Tees Hospital NHS Trust
Ms Bec Hanley, Freelance
Consumer Advocate,
Hurstpierpoint
Ms Maryann L. Hardy,
Lecturer,
Division of Radiography,
University of Bradford
Professor Alan Horwich,
Director of Clinical R&D, The
Institute of Cancer Research,
London
Dr Phillip Leech, Principal
Medical Officer for Primary
Care, Department of Health,
London
Dr Simon de Lusignan,
Senior Lecturer, Primary Care
Informatics, Department of
Community Health Sciences,
St George’s Hospital Medical
School, London
Professor James Neilson,
Professor of Obstetrics and
Gynaecology, Dept of Obstetrics
and Gynaecology,
University of Liverpool,
Liverpool Women’s Hospital
Dr John C Pounsford,
Consultant Physician, North
Bristol NHS Trust
Dr Vimal Sharma,
Consultant Psychiatrist & Hon
Snr Lecturer,
Mental Health Resource Centre,
Victoria Central Hospital,
Wirrall
Dr L David Smith, Consultant
Cardiologist, Royal Devon &
Exeter Hospital
Professor Norman Waugh,
Professor of Public Health,
University of Aberdeen
Dr Mike McGovern, Senior
Medical Officer, Heart Team,
Department of Health, London
169
Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.ncchta.org)
Health Technology Assessment Programme
Members
Professor Douglas Altman,
Director of CSM & Cancer
Research UK Med Stat Gp,
Centre for Statistics in
Medicine, University of Oxford,
Institute of Health Sciences,
Headington, Oxford
Professor John Bond,
Director, Centre for Health
Services Research,
University of Newcastle upon
Tyne, School of Population &
Health Sciences,
Newcastle upon Tyne
Mr Shaun Brogan,
Chief Executive, Ridgeway
Primary Care Group, Aylesbury
Mrs Stella Burnside OBE,
Chief Executive,
Office of the Chief Executive.
Trust Headquarters,
Altnagelvin Hospitals Health &
Social Services Trust,
Altnagelvin Area Hospital,
Londonderry
Ms Tracy Bury,
Project Manager, World
Confederation for Physical
Therapy, London
Mr John A Cairns,
Professor of Health Economics,
Health Economics Research
Unit, University of Aberdeen
Professor Iain T Cameron,
Professor of Obstetrics and
Gynaecology and Head of the
School of Medicine,
University of Southampton
Dr Christine Clark,
Medical Writer & Consultant
Pharmacist, Rossendale
Professor Collette Mary Clifford,
Professor of Nursing & Head of
Research, School of Health
Sciences, University of
Birmingham, Edgbaston,
Birmingham
Professor Barry Cookson,
Director,
Laboratory of Healthcare
Associated Infection,
Health Protection Agency,
London
Professor Howard Stephen Cuckle,
Professor of Reproductive
Epidemiology, Department of
Paediatrics, Obstetrics &
Gynaecology, University of
Leeds
Expert Advisory Network
Professor Nicky Cullum,
Director of Centre for Evidence
Based Nursing, University of York
Dr Katherine Darton,
Information Unit, MIND – The
Mental Health Charity, London
Professor Carol Dezateux,
Professor of Paediatric
Epidemiology, London
Mr John Dunning,
Consultant Cardiothoracic
Surgeon, Cardiothoracic
Surgical Unit, Papworth
Hospital NHS Trust, Cambridge
Mr Jonothan Earnshaw,
Consultant Vascular Surgeon,
Gloucestershire Royal Hospital,
Gloucester
Professor Martin Eccles,
Professor of Clinical
Effectiveness, Centre for Health
Services Research, University of
Newcastle upon Tyne
Professor Pam Enderby,
Professor of Community
Rehabilitation, Institute of
General Practice and Primary
Care, University of Sheffield
Mr Leonard R Fenwick,
Chief Executive, Newcastle
upon Tyne Hospitals NHS Trust
Professor David Field,
Professor of Neonatal Medicine,
Child Health, The Leicester
Royal Infirmary NHS Trust
Mrs Gillian Fletcher,
Antenatal Teacher & Tutor and
President, National Childbirth
Trust, Henfield
Professor Jayne Franklyn,
Professor of Medicine,
Department of Medicine,
University of Birmingham,
Queen Elizabeth Hospital,
Edgbaston, Birmingham
Ms Grace Gibbs,
Deputy Chief Executive,
Director for Nursing, Midwifery
& Clinical Support Servs,
West Middlesex University
Hospital, Isleworth
Dr Neville Goodman,
Consultant Anaesthetist,
Southmead Hospital, Bristol
Professor Alastair Gray,
Professor of Health Economics,
Department of Public Health,
University of Oxford
Professor Robert E Hawkins,
CRC Professor and Director of
Medical Oncology, Christie CRC
Research Centre, Christie
Hospital NHS Trust, Manchester
Professor F D Richard Hobbs,
Professor of Primary Care &
General Practice, Department of
Primary Care & General
Practice, University of
Birmingham
Professor Allen Hutchinson,
Director of Public Health &
Deputy Dean of ScHARR,
Department of Public Health,
University of Sheffield
Dr Duncan Keeley,
General Practitioner (Dr Burch
& Ptnrs), The Health Centre,
Thame
Dr Donna Lamping,
Research Degrees Programme
Director & Reader in Psychology,
Health Services Research Unit,
London School of Hygiene and
Tropical Medicine, London
Mr George Levvy,
Chief Executive, Motor
Neurone Disease Association,
Northampton
Professor James Lindesay,
Professor of Psychiatry for the
Elderly, University of Leicester,
Leicester General Hospital
Professor Rajan Madhok,
Medical Director & Director of
Public Health, Directorate of
Clinical Strategy & Public
Health, North & East Yorkshire
& Northern Lincolnshire Health
Authority, York
Professor David Mant,
Professor of General Practice,
Department of Primary Care,
University of Oxford
Professor Alexander Markham,
Director, Molecular Medicine
Unit, St James’s University
Hospital, Leeds
Dr Chris McCall,
General Practitioner,
The Hadleigh Practice,
Castle Mullen
Dr Andrew Mortimore,
Consultant in Public Health
Medicine, Southampton City
Primary Care Trust
Dr Sue Moss,
Associate Director, Cancer
Screening Evaluation Unit,
Institute of Cancer Research,
Sutton
Professor Jon Nicholl,
Director of Medical Care
Research Unit, School of Health
and Related Research,
University of Sheffield
Mrs Julietta Patnick,
National Co-ordinator, NHS
Cancer Screening Programmes,
Sheffield
Professor Robert Peveler,
Professor of Liaison Psychiatry,
University Mental Health
Group, Royal South Hants
Hospital, Southampton
Professor Chris Price,
Visiting Chair – Oxford,
Clinical Research, Bayer
Diagnostics Europe,
Cirencester
Ms Marianne Rigge,
Director, College of Health,
London
Dr Eamonn Sheridan,
Consultant in Clinical Genetics,
Genetics Department,
St James’s University Hospital,
Leeds
Dr Ken Stein,
Senior Clinical Lecturer in
Public Health, Director,
Peninsula Technology
Assessment Group,
University of Exeter
Professor Sarah Stewart-Brown,
Director HSRU/Honorary
Consultant in PH Medicine,
Department of Public Health,
University of Oxford
Professor Ala Szczepura,
Professor of Health Service
Research, Centre for Health
Services Studies, University of
Warwick
Professor Alistair McGuire,
Professor of Health Economics,
London School of Economics
Dr Ross Taylor,
Senior Lecturer,
Department of General Practice
and Primary Care,
University of Aberdeen
Dr Peter Moore,
Freelance Science Writer,
Ashtead
Mrs Joan Webster,
Consumer member, HTA –
Expert Advisory Network
170
Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.ncchta.org)
HTA
How to obtain copies of this and other HTA Programme reports.
An electronic version of this publication, in Adobe Acrobat format, is available for downloading free of
charge for personal use from the HTA website (http://www.ncchta.org). A fully searchable CD-ROM is
also available (see below).
Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public and
private sector purchasers from our Despatch Agents, York Publishing Services.
Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is
£2 per monograph and for the rest of the world £3 per monograph.
You can order HTA monographs from our Despatch Agents, York Publishing Services by:
– fax (with credit card or official purchase order)
– post (with credit card or official purchase order or cheque)
– phone during office hours (credit card only).
Additionally the HTA website allows you either to pay securely by credit card or to print out your
order and then post or fax it.
Contact details are as follows:
York Publishing Services
PO Box 642
YORK YO31 7WX
UK
Email: [email protected]
Tel: 0870 1616662
Fax: 0870 1616663
Fax from outside the UK: +44 1904 430868
NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of
£100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300
per volume. Please contact York Publishing Services at the address above. Subscriptions can only be
purchased for the current or forthcoming volume.
Payment methods
Paying by cheque
If you pay by cheque, the cheque must be in pounds sterling, made payable to York Publishing
Distribution and drawn on a bank with a UK address.
Paying by credit card
The following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard,
Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email.
Paying by official purchase order
You can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK.
We cannot at present accept purchase orders from commercial companies or from outside the UK.
How do I get a copy of HTA on CD?
Please use the form on the HTA website (www.ncchta.org/htacd.htm). Or contact York Publishing
Services (see contact details above) by email, post, fax or phone. HTA on CD is currently free of charge
worldwide.
The website also provides information about the HTA Programme and lists the membership of the various
committees.
Health Technology Assessment 2005; Vol. 9: No. 9
Electroconvulsive therapy for depressive illness, schizophrenia, catatonia and mania
Feedback
The HTA Programme and the authors would like to know
your views about this report.
The Correspondence Page on the HTA website
(http://www.ncchta.org) is a convenient way to publish
your comments. If you prefer, you can send your comments
to the address below, telling us whether you would like
us to transfer them to the website.
We look forward to hearing from you.
Health Technology Assessment 2005; Vol. 9: No. 9
Clinical and cost-effectiveness of
electroconvulsive therapy for
depressive illness, schizophrenia,
catatonia and mania: systematic
reviews and economic modelling
studies
J Greenhalgh, C Knight, D Hind,
C Beverley and S Walters
March 2005
The National Coordinating Centre for Health Technology Assessment,
Mailpoint 728, Boldrewood,
University of Southampton,
Southampton, SO16 7PX, UK.
Fax: +44 (0) 23 8059 5639
Email: [email protected]
http://www.ncchta.org
Health Technology Assessment
NHS R&D HTA Programme
ISSN 1366-5278
HTA

Similar documents

×

Report this document