Rapid Responses to:

EDUCATION AND DEBATE: Joanna Moncrieff and Irving Kirsch Efficacy of antidepressants in adults BMJ 2005; 331: 155-157 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Voltaire

brian norton   (15 July 2005)

Give me an alternative

david baker   (16 July 2005)

Throwing out the baby with the bathwater

Sten Thelander   (18 July 2005)

"Categorisation"

Andrew Cowie   (21 July 2005)

Efficacy of Antidepressants in Adults

David Wheatley   (21 July 2005)

Are the antidepressants really antidepressants?

James Paul Pandarakalam   (21 July 2005)

Efficacy of antidepressants in adults: a partial account of a complex problem

Stephen Pilling, Professor Sir David Goldberg, Professor Nicol Ferrier, Professor Andre Tylee, Mrs Carol Paton, Ms Rachel Burbeck, Dr Tim Kendall   (22 July 2005)

From medicalizing to psychologizing misery

ABDI SANATI   (22 July 2005)

Efficacy of antidepressants in adults - methodological issues

Tamar D Wohlfarth, Barbara J van Zwieten, Jitschak G Storosum, Wim van den Brink   (22 July 2005)

Curvilinear relationship between severity of depression and response to antidepressants

Phil Harrison-Read   (22 July 2005)

antidepressants efficacy - where are we now?

Dr.Suleman Shabuddin Virani   (25 July 2005)

Complementary treatment may be helpful

Edmond V O`Flaherty   (26 July 2005)

Leaflet not part of a campaign

Philip W Timms   (27 July 2005)

Small differences, big effects ?

Miles Jefferson   (27 July 2005)

Fallacy in The Hamilton Rating Scale

David H. Marjot   (29 July 2005)

Blissless ignorance

Abraham Petrus Cutajar   (29 July 2005)

Tom Cruise, Psychiatry, Niels Bohr and, the Broken Mirror

Dr. Avinash R. Patwardhan   (30 July 2005)

Antidepressants work: the question is how well in the real world.

Ian M Anderson   (30 July 2005)

The future of SSRIs

Steven Taylor   (31 July 2005)

Use of antidepressants in childhood depression

Jonathan M Goldin   (5 August 2005)

Authors Response

Joanna Moncrieff, Irving Kirsch   (16 August 2005)

RE: Efficacy of antidepressants in adults - Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005;331(7509):155-7

Frederic M. Quitkin, Jonathan W. Stewart, Patrick J. McGrath, Donald F. Klein   (9 September 2005)

Voltaire 15 July 2005
brian norton, staff-grade psychiatrist hebburn health centre, hebburn, south tyneside, ne31 2sp Send response to journal: Re: Voltaire	'It is the job of the physician to amuse the patient, whilst nature takes its course', said Voltaire. Or something like that. Competing interests: None declared
Give me an alternative 16 July 2005
david baker, GP Poole, Dorset Send response to journal: Re: Give me an alternative	When the next patient sits down & tells me they are depressed (almost certainly Monday morning), what will I do? I could offer them nothing, or referral to a psychologist - which amounts to the same thing, with a waiting list of over a year, access to which is screened by a psychiatrist. I could offer referral to a counsellor - only 2-3 months wait, screened by the psychologists, with rejections involving more delay & tricky explanations. Or I can treat them with an SSRI, & whether it works via a placebo effect or by altering brain biochemistry, I don't care. No delay, & it works. Patient happy, me happy. Patient still unhappy sent to a psychiatrist. If you want to change that paradigm, you've got a long way to go. David Competing interests: None declared
Throwing out the baby with the bathwater 18 July 2005
Sten Thelander, locum consultant Sahlgrenska University Hospital, 42250 Gothenburg Send response to journal: Re: Throwing out the baby with the bathwater	The opinions of Drs Moncrieff and Kirsch are well known by those following the psychiatric and psychological literature. Using biased and selective references they paint a picture were greedy drug companies and naive or corrupt researchers collaborate in misinforming the public, the politicians and the ordinary clinician. Where are they right? 1. For mild and moderate depression drug treatment and specific psychotherapies are not particularly effective. 2. Using only published trials will overestimate the efficacy of antidepressive drugs, and most other treatments as well. Where are they biased? 1. They contradict themselves by first, rightly, critizising the Hamilton Depression Rating Scale, but still rely on this when comparing active drugs and placebo. They seem unaware of the many studies using core depressive items from this scale which demonstrate a much larger difference between active drug and placebo. 2. They quote a meta-analysis not showing an increased antidepressive effect with increased severity, but failed to refer to a recent publication showing such a relationship (1). 3. They ignore the probable antidepressant effect of atropine used as active placebo (2). Surprisingly, this possibility is not even mentioned in the Cochrane review of studies using active placebo. 4. They ignore the many studies comparing psychotherapy and drug therapy, with or without a placebo comparison, which rarely show differences in effectiveness. Possibly this is so because they do not want to draw the conclusion that psychotherapy might also be no better than placebo. 5. They ignore the imaging studies showing differences in areas affected by placebo and antidepressive drugs. 6. They ignore the overwhelming evidence for a substantial reduction in relapses and recurrences with continuing antidepressive drug treatment compared to placebo. 7. Both NICE and the authors use a poorly validated definition of severe depression. The cutoff at 20 was originally used in the large NIMH-study comparing imipramine, cognitive therapy, interpersonal psychotherapy and psychological placebo. The choice of 20 as the cutoff was purely pragmatic. Most drug trials on severe depression use 25 as the cutoff, a level of severity that few patients in psychotherapy studies reach and rather few too in the studies comparing antidepressants and placebo. What should be done to make real progress? 1. Linking the subgrouping of depression into mild, moderate and severe to scores on HDRS, MADRS and other scales, maybe especially the core items identified by Per Bech (3, 4). 2. More studies comparing antidepressant with psychotherapy in more severely depressed patients. The recent study by DeRubeis and Hollon failed to resolve the question of comparative efficacy (5). 3. More studies on switch strategies, that is studies switching from unsuccessful pharmacotherapy to psychotherapy and vice versa. 1. Khan A, Brodhead AE, Kolts RL, Brown WA. Severity of depressive symptoms and response to antidepressants and placebo in antidepressant trials. J Psychiatr Res 2005;39(2):145-50. 2. Salamone JD. A critique of recent studies on placebo effects of antidepressants: importance of research on active placebos. Psychopharmacology (Berl) 2000;152(1):1-6. 3. Bech P. The Bech-Rafaelsen Melancholia Scale (MES) in clinical trials of therapies in depressive disorders: a 20-year review of its use as outcome measure. Acta Psychiatr Scand 2002;106(4):252-64. 4. Licht, R. W, Qvitzau, S, Allerup, P, Bech, P.Validation of the Bech- Rafaelsen Melancholia Scale and the Hamilton Depression Scale in patients with major depression; is the total score a valid measure of illness severity? Acta Psychiatr Scand 2005;111(2):144-9. 5. DeRubeis RJ, Hollon SD, Amsterdam JD, Shelton RC, Young PR, Salomon RM, et al. Cognitive therapy vs medications in the treatment of moderate to severe depression. Arch Gen Psychiatry 2005;62(4):409-16. Competing interests: None declared
"Categorisation" 21 July 2005
Andrew Cowie, GP Dundee DD1 5LA Send response to journal: Re: "Categorisation"	Apologies for my lack of expertise in statistics, but the declaration that a small average change is "not clinically significant" surely misses the point that a number of patients will have little or no change, and some will have a big, clinically significant change. Give a 40% clinically helpful placebo response and (say) a 70% clinically helpful response to antidepressants surely we are only discussing 30% of patients showing a response. That would give an Number Needed to Treat of 3.3. Even if we have a placebo response of 40% and drug response of 50%, that gives an NNT of 10, which is considerable better than many of the other treatments we use on a regular basis. This would give a very small average shift across the board, but would still be used as a useful treatment. It's a little like the excellent recent analysis of hypertension treatment, that showed reducing the systolic BP of a hypertensive 40yo male to optimal levels would "on average" prolong his life by only 12 days. You're asking the wrong question, I think. Competing interests: None declared
Efficacy of Antidepressants in Adults 21 July 2005
David Wheatley, MD FRCPsych Surrey KT2 5HP Send response to journal: Re: Efficacy of Antidepressants in Adults	If the implications of the paper by J.Moncrieff and I.Kirsch (1) are accepted at face value, then this strikes despair into my heart and anguish into my mind (and that doubtless of my patients also). At one fell swoop, a major linchpin in the edifice of psychopharmacological practice, is swept away; anti-depressants (A-Ds) do not work! But this is not true, as all practising psychiatrists know. To make a pronouncement such as this, only one criterion should be used: the absolute number of patients cured of major depression, after 6-months treatment in adequate dosage. And what do I mean by cure? I can do no better than quote the master himself, Max Hamilton, who told me in his lifetime, that a patient achieving a total score of 6 or less on the Hamilton depression rating scale (HAM-D, was no longer suffering from clinical depression. A modification of this criterion that has been used in many trials recently, is to divide subjects into “responders” and “non-responders”, response being defined as: “final HAM-D score of <10 or a reduction of at least 10 from baseline”. This is a method that I have used myself (2) and there are a wealth of reports using this, or similar measures, that demonstrate the effectiveness of specific A-Ds in comparison to placebo, as outlined in a recent book from the CINP (Collegium Internationale Neuro -Psychopharmacologicum), that extensively reviews the work undertaken over the past 50 years (3). Anyone who may care to check this, will soon be convinced that it is not all an illusion and that antidepressant drugs are a rightful treatment for the right indications. David Wheatley MD, FRCPsych. 1. Moncrief J. & Kirsch I. Efficacy of antidepressants in adults. BMJ 2005; 331: 155-157. 2. Wheatley D. LI 160, an extract of St.John’s Wort, versus amitriptyline in mildly to moderately depressed outpatients – a controlled 6-week clinical trial. Pharmacopsychiat. 1997; 39: 77-80. 3. Ban T E, Healy D & Shorter E. Reflections on Twentieth- century Psychopharmacology. Budapest; Animula, 2004. Competing interests: None declared
Are the antidepressants really antidepressants? 21 July 2005
James Paul Pandarakalam, consultant psychiatrist, 5 Borough Partnership NHS Trust St Helens North CMHT, Peasley Cross Resource Centre, St Helens, Merseyside WA 9 3DA Send response to journal: Re: Are the antidepressants really antidepressants?	Antidepressants have not shown convincingly any superior effect than placebo or to affect depression and suicide rates and so their uses warrant further review is an interesting finding 1. The word antidepressant itself is a misnomer. They are probably only anti stress agents or mood lifting agents. But the wording has given the wrong message to patients as well as to mental health professionals that these drugs would cure depression. Consequently psychological aspects of depression are most often over looked ; depression becomes perpetuated. The pharmaceutical industries have made a fortune out of this misunderstanding but the inapplicable designation has enhanced the placebo effect of the drug! Like Chlorpromazine, antidepressants have other functions; the tricyclics are now used as analgesics and some of them are sometimes good hypnotics and anxiolytics. Antidepressants have a role in the treatment of depression but their role is not well defined. To use a medical analogy, antipyrexial agents have a place in the symptomatic relief of pneumonia but they are not to be equated to the antibiotic treatment. Likewise, antidepressants are helpful in treating the neurotransmitter problem of depression and they lift the mood, enhancing one’s own mental resources to fight against depression. But they seem not to cure depression on their own as it is generally claimed. One cannot trivialise the neurotransmitter derangement ; depression is as serious as cancer when patients are actively suicidal. Between 20 to 40% depressed patients respond only minimally to monotherapy 2. Only 50% of patients not responding to a primary antidepressant improve when another antidepressant is introduced 3. In other words, a clinician has the formidable task of looking after a good number of treatment resistant depressed patients. This situation has caused an alarming economic disability with serious implications in the bio- psychosocial spectrum and also therapeutic pessimism. The chemicals involved in depression are not mapped out completely and the missing links are probably responsible for the confusion surrounding the antidepressant therapy in depression. New techniques of fluroscence histochemistry- immunohistochemical and in situ hybridization have permitted the elucidation of chemically defined neural circuits providing tools for burgeoning field of neurochemical pathology. Consequently, novel biological concepts of depression have germinated recently and they involve other receptor systems or intracellular targets. Branching out of the current monoamine cycle based antidepressants, the possibility of non-monoaminergic antidepressants are being explored. During recent years many new potentially relevant brain transmitters and proteins have been identified. There are a number of neuropeptide based approaches to develop novel antidepressants. They are substance P antagonists, vasopressin antagonists, melanocortin-concentrating hormone antagonists and corticotrophin-releasing factor receptor antagonists. Various recent findings show that there are changes in neurotrophins or corticotrophin releasing hormones associated with depression. New interventions to stop the stress hormone cascade before it gets rolling are being thought about; an input to psychopharmacology from the endocrinology . The CRF blockers obstruct release of stress hormones that may contribute to depression , creating a cushion or buffer in the brain interrupting the biochemical signals of stress response. But these drugs in the pipeline are also running the risk of getting baptised as antidepressants. It is a matter of speculation whether the so called would be modern antidepressants are going to achieve the desirable qualities of model antidepressants; only time will tell whether the “light at the end of the tunnel” is true or false. Medical scientists are blindly concentrating on new chemical entities while innovating research and development aimed at producing very expensive and colourful pills to combat depression. The government, pharmaceutical industries and the medical profession, through the acceptance of only organic, genetic and biological aetiology have adopted a form of tunnel vision. Whatever theories of psychotherapies or drug therapy one may espouse, “Human beings are the most important therapeutic agent for human beings.” However, these are exciting times in psychiatry because of the emergence of new medications with improved effectiveness and cleaner side effect profiles and well researched guidelines for good practice. References: 1.Moncrieff Joanna, Kirsch Irvine. Efficacy of antidepressants in adults. BMJ 2005: 155-157, doi : 10.1136/bmj,331. 7509.155 2. Shelton RC (1999) Treatment options of Refractory depression. J. Clinical Psychiatry 60 (suppl 4):57-61 3. Depression Guideline Panel(1993). Depression in primary care, Vol 2 Treatment of major depression. Clinical practice guideline, No 5.pp71- 86.Department of Health and Human Services. Rockville, MD. Competing interests: None declared
Efficacy of antidepressants in adults: a partial account of a complex problem 22 July 2005
Stephen Pilling, Joint Director National Collaborating Centre for Mental Health, University College London, London WC1 7HB, Professor Sir David Goldberg, Professor Nicol Ferrier, Professor Andre Tylee, Mrs Carol Paton, Ms Rachel Burbeck, Dr Tim Kendall Send response to journal: Re: Efficacy of antidepressants in adults: a partial account of a complex problem	Dear Editor As authors of the NICE Depression guideline1 we must respond to Moncrieff and Kirsch’s critique of the guideline and its evidence base.2;3 They raise a number of important points but we consider they use them inappropriately in their critique of the guideline. In addition, there are several inaccuracies in their account. Starting with these inaccuracies: • They erroneously state that the guideline recommends SSRIs as first -line treatment for moderate and severe depression. In fact, whilst SSRIs are recommended for first-line treatment of moderate depression, for severe depression the combination of SSRIs with psychological treatment is recommended. • They assert that, ‘The conclusion that the drugs had clinically important benefits was based on analysis of response and remission rates'. This is incorrect. We extracted efficacy data in 3 forms: response, remission and mean endpoint. We consistently found that response gave the largest effect sizes and mean endpoint the smallest. We had read the articles by Kirsch on this topic with great interest.4;5 We therefore took a conservative position and based our conclusions on mean endpoint data and this lead to our recommendation to avoid antidepressants for the routine treatment of mild depression. • They state that we deemed the benefits of SSRIs not to be clinically significant. Again this is inaccurate. We found the difference to be modest in many cases but, by an a priori definition, still of clinical importance. • In the summary points they state that the evidence for long term outcomes for depression has ‘not been convincingly shown’. However, recent meta-analyses, reviewed in the guideline, have demonstrated long-term benefits associated with continued antidepressant use compared with placebo.6 In addition to these inaccuracies, Moncrieff and Kirsch present only a partial account of the issues. For example, they cite evidence that SSRIs may increase suicidality in adults implying that we ignored these data,7 8 but they ignore other evidence which suggests that the data is more equivocal than they acknowledge.9 They also challenge our claim ‘that the superiority of antidepressants over placebo correlates positively with the severity of depression’. We assume they deduce this from our recommendation of antidepressants for moderate and severe depression, but not for milder symptoms. Again, this is a complex area and the subject of ongoing debate which we address in our introduction to the review.3 We contend that the weight of evidence supports our original view particularly in the light of other authors10 and more recently published findings 11. They also criticise the lack of a justification for an a priori level of clinical significance. We agree that the reasoning behind this was not described in the guideline although it was the subject of considerable debate. After careful discussion (crucially, including patient members) of existing conventions, the structure and content of the scales, and Cohen’s estimates of effect sizes, we adopted a standardised mean difference of 0.5 (i.e., a medium effect size) as our guide to clinical significance. Underlying their critique is an assumption that psychological treatments are to be preferred as a matter of course over pharmacological treatments simply on grounds of less harm is based on a naive and unsubstantiated assumption that psychological treatments can do no harm. They are also relatively expensive and scarce compared with drug treatments. Depression remains a major public health problem. The National Collaborating Centre for Mental Health was tasked by NICE with producing a pragmatic guideline for the treatment of depression in the NHS. We are convinced that we achieved this based on the best evidence available. We believe Moncrieff and Kirsch have failed, not least because of an inaccurate and partial reading of the guideline, to undermine the evidence base for antidepressants and to establish their case for a reconsideration of the NICE recommendations for their use. However, the challenge is clear. If we are to develop more effective guidelines for the treatment of depression based on a significantly improved conceptualisation of depression, then the quality of that evidence base for all interventions needs to improve significantly. Reference List (1) NICE. Treatment and management of depression in primary and secondary care. 2004. NICE. (2) Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005; 331(7509):155-157. (3) National Collaborating Centre for Mental Health. Depression: Management of depression in primary and secondary care. London: The British Psychological Society; 2005. (4) Kirsch, Scoboria A, Moore TJ. Antidepressants and placebos: Secrets, revelations, and unanswered questions. Prevention and Treatment 2002; 5:Article 33. (5) Kirsch I, Moore T, Scoboria A, Nicholls S. The emperor's new drugs: an analysis of antidepressant medication data submitted to the US Food and Drug Administration. Prev Treat 2002; 5. (6) Geddes J, Camey S, Davies C, Furukawa T, Kupfer D, Frank E et al. Relapse prevention with antidepressant drug treatment in depressive disorders: a systematic review. Lancet 2003; 361(9358):653-661. (7) Gunnel D, Ashby D. Antidepressants and suicide: what is the balance of benefit and harm. BMJ 2004; 329:34-38. (8) Fergusson D, Doucette S, Glass K, Shapiro S, Healy D, Herbert P et al. Association between suicide attempts and selective serotonin reuptake inhibitors: systematic review of randomised controlled trials. BMJ 2005; 330:396-399. (9) Cipriani A, Barbui C, Geddes J. Suicide, depression, and antidepressants. BMJ 2005; 330(7488):373-374. (10) Angst, J. (1993). "Severity of depression and benzodiazepine co-medication in relationship to efficacy of antidepressants in acute trials. A meta-analysis of moclobemide trials. Human Psychopharmacol 8: 401- 407 (11) Kahn A, Brodhead AE, Kolts RL, Brown WA. Severity of depressive symptoms and response to antidepressants and placebo in antidepressant trials. J Psychiatr Res 2005; 39(2):145-150. Competing interests: The National Collaborating Centre for Mental Health receives funding from NICE for the production of clinical practice guidleines.
From medicalizing to psychologizing misery 22 July 2005
ABDI SANATI, STAFF GRADE SOUTH LONDON & MAUDSLEY NHS TRUST, SE58BB Send response to journal: Re: From medicalizing to psychologizing misery	In the recent years there have been many debates regarding the selective serotonin reuptake inhibitors(SSRIs). Drs Moncrieff and Kirsch(1) question the efficiency of the SSRIs and criticise the National Institute of Clinical Excellence(NICE) guidelines on treatment of depression. In its guidelines the NICE agrees there are many problems inherent in the concept of depression, including classification(mild to severe based on number of symptoms). One of the main obstacles in reaching consensus in managing depression is depression as a social construct. In an era which there is less support from a person's social circle(family and friend,etc), people resort to the healthcare system more frequently. Entitlement to the benefits of the system depends on having a specific diagnosis(e.g. depression). This leads to medicalizing human misery and life in general. Consequently there is an increase in the heterogeneity of patients labelled "depressed", rendering the outcome-studies' results either invalid or insignificant. I believe it is the psychiatrist's task to distinguish clinical depression and misery and that is where art of psychiatry lies. I also agree with Simon Hatcher(2) in "when criticising the evidence, why stop at antidepressants"(commentary on the above article by Drs Moncrieff and Kirsch). Focusing the criticism only on physical treatments could lead to moving from medicalizing to psychologizing human misery and life, which could be equally dangerous. This issue is nothing new and have baan discussed by Viktor Frankl in his excellent work, Man's Search for Meaning. References: 1- Moncrieff J, Kirsch I. Efficiency of antidepressants in adults. BMJ 2005;331:155-7 2- Hatcher S. Commentary: Why stop at antidepressants. BMJ 2005;331:158 Competing interests: None declared
Efficacy of antidepressants in adults - methodological issues 22 July 2005
Tamar D Wohlfarth, clinical assessor Medicines Evaluation Board of the Netherlands, Kalvermarkt 53, P.O. Box 16229, 2500 BE, Den Haag, T, Barbara J van Zwieten, Jitschak G Storosum, Wim van den Brink Send response to journal: Re: Efficacy of antidepressants in adults - methodological issues	Montcrieff and Kirsch(1) question the efficacy of antidepressants, specifically of SSRIs. Based on the results of the NICE meta-analysis of placebo controlled SSRI studies, they note that the difference between active and placebo arms in mean improvement of symptoms from baseline is small and not likely to be clinically relevant. Responders and remitters analyses do suggest clinically relevant results but these are discounted as flawed. They show that a clinically relevant result in terms of responders might be achieved in spite of a small difference in mean improvement between SSRIs and placebo, if improvement scores are normally distributed and response is defined in terms of improvement on the outcome measure that is similar to the mean improvement that was achieved in the study. We challenge the validity of both their assumptions. Rather than assuming that response scores are normally distributed, we propose that they are more likely to have a bimodal distribution, as some patients respond well to treatment while others do not. This could explain why mean response over all patients, those who do respond well and those who do not, is low. The effect of those who respond is presumably diluted by the lack of effect in the non-responding patients. In order to examine the nature of the distribution of response scores, we analysed a large data set of 32 placebo controlled trials(2). The distribution of the response scores in the TCA arms over all the studies, approaches a bimodal form. Furthermore, Kolmogorov-Smirnov test indicates a significant deviation from normality (Z=2.97; P<0.0001). The second assumption made by Moncrieff and Kirsch, namely that the definition of response is similar to the mean reduction in symptoms from baseline, is also not necessarily correct. The results of studies indicate that the difference between active compound and placebo in the percentage of responders/remitters is maintained regardless of the definition of responders that is used. In summary, the argument that mean response is a valid indicator of treatment efficacy while responders/remitters analysis is not, rests on two assumption that are not necessarily and not likely to be correct. If the assumption that some patients respond to treatment while others do not is correct, then analysing responders/remitters is more appropriate than calculating mean response. Results of clinical trials that examine the effect of antidepressants in terms of responders/remitters indicate that the effect is moderate (15-20%) but not negligible, neither from a clinical nor from a public health perspective. In the CHMP guidelines for developing of medicinal products in depression, both endpoints are mentioned as necessary. 1. Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005;331:155-157. 2. Storosum JG, Elferink AJA, van Zwieten BJ, van den Brink W, Gersons BPR, Broekmans AW: Short-term efficacy of tricyclic antidepressants revisited: a meta-analytic study. Eur Neuropsychopharmacol 2001; 11:173–179. Competing interests: None declared
Curvilinear relationship between severity of depression and response to antidepressants 22 July 2005
Phil Harrison-Read, consultant psychiatrist Department of Psychiatry, Royal Free Hospital, Pond Street, London NW3 2QG O Send response to journal: Re: Curvilinear relationship between severity of depression and response to antidepressants	Editor- Moncrieff and Kirsch remind us that the efficacy of antidepressants measured in clinical trials is often quite small, and sometimes of doubtful clinical relevance (1), a warning even given in textbooks nowadays (2). However, as we are also reminded by Hatcher in his commentary (3), the majority of randomised controlled trials (RCTs) involving antidepressants are of short duration (6 weeks or less), whereas clinical experience and some clinical trials suggest that the full therapeutic effect of antidepressant drugs takes much longer to develop fully. Although there are measurable benefits on 'core' affective components of the depressive syndrome within the first few days of treatment with antidepressants (4), improvements in patients' habitual cognitive responses of pessimism and futility seem to lag behind, and the full impact and effectiveness of antidepressant treatment may not be felt until after patients' mood and capacity for pleasure have been restored. Moncrieff and Kirsch's critical appraisal of the efficacy as opposed to the effectiveness of antidepressant drugs in adults is welcome, but some of their arguments do not seem convincing. For example the authors cast doubt on the validity of a defined categorical outcome such as 'improvement'. This type of outcome is oten used in RCTs of antidepressants on the grounds that it is clinically relevant, and positive treatment effects are frequently described. The authors argue that small between-group differences in mean scores on the Hamilton rating scale for depression (HRSD) can in theory translate into large differences in improvement rates which may exaggerate the true clinical impact of treatment. However this argument relies on the assumption that changes in depression rating scores due to drug treatment are normally distibuted, and the criterion response for improvement is close to the group mean response. In practice, there is often marked heterogeneity in patients' responses to antidepressants, with some showing little or no improvement, and others showing a marked reduction in depression scores (4). Under these circumstances , depression rating scores may not be normally distributed, and it is easy to demonstrate that statistically significant and clinically relevant between-group differences in depression rating scores following treatment can occur even in the absence of a group treatment effect on categorical outcomes such as 'improvement'(2). In other words, the relative merits of categorical and continuous outcome data have to be judged separately for each RCT under consideration, and meta-analyses must allow for heterogeneity in the distribution of data when studies are pooled. An important potential source of heterogeneity in the response to antidepressants derives from the type and severity of the depressive disorders being treated. In particular, a consistent finding from many sources over more than 40 years is that both mild (HDRS scores of 14 or less) (5)(6)(7) and very severe (HDRS scores of 26 or more)(4)(8)(9)(10) cases of major depression respond less well to antidepressants than cases of intermediate severity. The poorer response to antidepressants of patients with very severe depression is seen regardless of the presence of psychotic symptoms or marked cognitive impairment (10). A curvilinear relationship between severity of major depression and response to antidepressants may be difficult to demonstrate robustly in meta-analyses when there are insufficient numbers of patients across the full range of illness severity, which is often the case. Including an excess of mild cases of major depression in RCTs of antidepressant drug efficacy may result in underestimation of their value, a possibility acknowledged but minimised by Moncrieff and Kirsch. However including cases of major depression which are very severe, as in the meta- analysis by Kirsch et al.(11) will also dilute the effect size of antidepressant treatments, a point which is ignored by Moncrieff and Kirsch. Moncrieff and Kirsch's dismissal of the "possibility that patients in the mid-range of severity show a greater antidepressant response" because it "would not be expected from a simple biological effect" is puzzling. Biological effects, especially those pertaining to depressive illness, are not inherently likely to be "simple". Moncrieff and Kirsch's alternative explanation, that the selective benefit of antidepressants on moderate depression reflects "some methodological artefact" perhaps partly reflects excessive scepticism on their part regarding the value of antidepressant drugs. Although there may need to be a correction to the over-valuation of antidepressant drug treatment by some doctors who admittedly may have few alternative therapeutic measures at their disposal, the positive experience of antidepressants of large numbers of patients and clinicians over half a century cannot be explained away so easily. (1) Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005;331:155-7. (2) Harrison-Read P, Tyrer P. The clinical principles underlying drug treatment in psychiatric practice. In: King DJ, ed. Seminars in Clinical Pychopharmacology. Second edition. London:Gaskell, 2004:92-138. (3) Hatcher S. Why stop at antidepressants? BMJ 2005;331:158. (4) Katz MM, Koslow SH, Maas JW, Frazer A, Bowden CL, Casper R, Croughan J, Kocsis J, Redmond Jr.E. The timing, specificity and clinical prediction of tricyclic drug effects in depression. Psychological Medicine 1987;17:297-309. (5) Stewart JA, Quitkin FM, Liebowitz MR, McGrath PJ, Harrison WM, Klein DF. Efficacy of desipramine in depressed outpatients. Archives General Psychiatry 1983;40:202-7. (6) Quitkin FM, Rabkin JG, Ross D, McGrath PJ. Duration of antidepressant drug treatment. Archives General Psychiatry 1984;41:238-45. (7) Paykel ES, Hollyman JA, Freeling P, Sedgwick P. Predictors of therapeutic benefits from amitriptyline in mild depression: a general practice placebo-controlled trial. Journal Affective Disorders 1988;14:83- 95. (8) Medical Research Council Clinical Psychiatry Committee. Medical Research Council trial of the treatment of depressive illness. BMj 1965;1:881-6. (9) Abou-Saleh MT, Coppen A. Classification of depression and response to antidepressive therapies. British Journal Psychiatry 1983;143:601-3. (10) Kocsis JH, Croughan J, Katz MM, Butler TP, Secunda S, Bowden CL, Davis JM. Response to treatment with antidepressants of patients with severe or moderate nonpsychotic depression and of patients with psychotic depression. American Journal Psychiatry 1990;147:621-4. (11) Kirsch I, Moore TJ, Scoboria A, Nicholls SS. The emperor's new drugs: an analysis of antidepressant medication data submitted to the US Food and Drug Administratio. Prevention and Treatment 2002;5.www.journals.apa.org/prevention/volume5/pre0050023a.html (accessed 18th July 2005). Competing interests: None declared
antidepressants efficacy - where are we now? 25 July 2005
Dr.Suleman Shabuddin Virani, psychiatry resident,Sir J J Group of Hospitals and Grant Medical College. Byculla,Mumbai,400008, INDIA Send response to journal: Re: antidepressants efficacy - where are we now?	There are total of 12 published randomised clinical trials in entire world literature for evaluating efficacy of antidepressants in children.Each of these trials failed to find any significant benefit of medication over inert placebos.only 4 of RCTs claimed significant differances between drugs and placebo and these did so only on clinician rated measures,not patient rated measures. 3 of the trials did not report mean std deviations,leaving 9 nfor meta anlysis.when these 9 studies are combined,the placebo response is 87% of drug response.It is 75% of SSRI response and 97% of TCA response.Thus meta analysis shows that TCA have no significant effent on depression in children.The efeect of SSRI is statistically significant,but it is not clinically significant.Overall the efficacy of antidepressants in children is weaker than in adults. 1.These results are drawn from studies with design flaws typically favour the study drug. 2.They frequently exclude placebo responders before random assignment. 3.They rely on ratings by clinicians who may have vested interest in outcome. 4.They are likely to be unblinded by medicine side effects. 5.Results are drawn from published literature - which is subject to publication bias,meaning many studies with negative results do not get published.adding unpublished studies,most of which have negative results will surely shrink differances between antidepressants and placebo even further. Antidepressants came into psychiatry scene in 1950,followed in 1980 by SSRI,new class of drugs promoted as more efeective,less likely to cause side effects than older versions,despite HIGH COST,SSRI gradually became treatment of choicegiven by GP`s and psychiatrists on both sides of atlantic,for mild,moderate and severe depression. Around same time in Britain,THE DEFEAT DEPRESSION CAMPAIGN organised by Royal college of psychiatrists and Royal College of physicians and sponsored in no small part by DRUG COMPANIES,was seeking the help of media to destigmatise depression and erge people to go to their doctors for treatment. Rise in popularity of antidepressants coincided with fall of prescriptions for BENZODIAZEPINES,once they were found to be addictive.It can also partially be accounted for by PSYCHIATRY`S WISH to improve its status through a closer alliance with the rest of medicine.The arrival of drugs helped psychiatrist profession to assert itself as a speciality.It got them away from ASYLUMS and allowed the setting up of OUT PATIENT facilities,where people could walk away with a prescription,just as they could in other areas of medicine.The notion of specific drugs for specific psychiatric disorder helped to bolster the case that psychiatry was no different from other medical specialities. Genetic influence on depression have been shown to be weaker than initial ones. Biochemical theories are as yet unproven the biological markers for depression remain alusive.Most antidepressants interact with SEROTONIN OF NORADRENALINE or BOTH.The potentiation of either of these neurotransmitter systems has been shownto stimulate the other system,which makes the details of pharmacodyanamics of each drug difficult to translate into a predictor of efficacy. Priliminary evidence shows that patients who respond to cognitive bahaviour therapy show similar biological changes to those who respond to medicines. ONE study has calculated that only one depressed person in 20 in Britain at present benefits from antidepressants.This is mainly bcoz the drugs are never prescribed,but also they are not taken at allor if taken,not for long enough or bcoz the dose is inadequete.Just under 10 years on,despite confident claims by Royal College of Psychiatrists that 6 -7 of every 10 people will get better with antidepressants within 6 -8 weeks,the safet and efficacy of SSRI and the rarity of withdrawl symptoms are all being strongly questioned. Enhanced treatment programs in primary care in AMERICA led to increased prescribing and patient compliance,but which failed to help those with mild depression and had short term but not LONG TERM positive effects on those with major depression. In the UK,educational initiatives with GP`S that have improved recognotion of depression and increased prescribing of antidepressants have not been found to improve the OUTCOME for patients. One not very large study concluded that ,if people were taken off their drugs earlier then 6 months,they were more likely to RELAPSE.But these were probably DISCONTINUATION SYMTOMS and all this study shows how difficult it can be to take people off antidepressants,rather than that they need to be on them any specific length of time.so the evidence on which the 6 months recommendation is based is not strong at all. GLAXOSMITHKLINE has been forced to admit that its best selling drug ,one antidepressant - PAROXETINE, SSRI can cause severe withdrawl symptoms when stopped.What can we say about other significant side effects like - tics, agitation,muscle cramps,parkinsonism,memory loss,blanking out. Canadian research reported that most guidelines on clinical practice are written by experts with UNDISCLOSED LINKS with drug company.They surveyed 192 authors of 44 clinical guidelines and of just over half who responded,87% had admitted to financial links with at least one drug company.over half had been paid to conduct research,a third had acted as a consultant or had been an employee and two thirds had been paid for giving talks. Most of depression cases have significant stressors- life events,and most,if not all, symtoms resolves spontaneously once some time has passed away. On the whole,the new drugs will keep on coming and prescriptions will keep on flowing no matter whether the therapeutic effects benefit the patient or placebo effect does so.THE need of hour is unbiased researchon the subject like-- 1. one on one comparison of antidepressant with psychotherapy. 2. efficacy of antidepressants in SEVERE depression with classification of cases into mild,moderate and severe cases. Competing interests: None declared
Complementary treatment may be helpful 26 July 2005
Edmond V O`Flaherty, G.P. Dublin Send response to journal: Re: Complementary treatment may be helpful	I use SSRIs on a large number of patients. Since the use of benzodiazepines has gone out of fashion it has become necessary to use them for those suffering from anxiety without depression in many cases. Much to my surprise I have found complementary treatment very useful and often it helps to get patients off SSRIs or SNRIs (as well as reducing the amount of antipsychotics prescribed by their psychiatrists for those with schizophrenia or bipolar disorder).The late Dr David Horrobin did a lot of research into e.g. fish oil in depression as did Dr Joseph Hibbeln at the National Institutes of Health in U.S.I started a website at www.omega3.20megsfree.com on fish oil which has now extended its range considerably.The section of it on Alternative Mental Health,which is largely copied from the site of the Pfeiffer Center in Chicago (www.hriptc.org) has benefitted many,although I am of the complementary rather than alternative school myself. I get emails from all over the world thanking me for the help the site has given the writers.Even though nutritional treatment of psychiatric illness makes up no more than 5-10% of my work I have found that it is by far the most useful and satisfying work that I do.It also saves the country a fortune in mental health costs,as hospitalisation is greatly reduced. Competing interests: None declared
Leaflet not part of a campaign 27 July 2005
Philip W Timms, Consultant psychiatrist START team, Masters House, Kennington, SE11 4TH Send response to journal: Re: Leaflet not part of a campaign	Sir As the chair of the editorial board that produces the “Help is at Hand” series of leaflets for the Royal College of Psychiatrists, I am concerned about your use of an image of our leaflet on depression (Efficacy of antidepressants in adults, 16th July 2005). The caption implies both that this leaflet is part of a campaign to increase awareness of depression and, less directly, that it is advocating the use of antidepressants above other forms of intervention. This is not the case. The Royal College of Psychiatrists ran a “Defeat Depression” campaign several years ago, but this leaflet pre-dated the campaign and has continued long after it. It is one of a series of leaflets about mental disorders. They are intended to provide information that will help people to find out more about their condition and make more informed choices about their treatment. In the text of the leaflet itself, roughly equivalent space is given to self-help, psychotherapy and antidepressants. Competing interests: See text
Small differences, big effects ? 27 July 2005
Miles Jefferson, Consultant Psychiatrist Manchester Royal Infirmary, Oxford Road, Manchester, M13 9WL Send response to journal: Re: Small differences, big effects ?	Editor - Moncrieff and Kirsch criticize how antidepressant trial data has been analyzed [1]. They point out that continuous variables should not be used to categorize. This is correct. Unfortunately, they then proceed to make exactly the same mistake themselves. This undermines their argument and invalidates their conclusion that antidepressants are ineffective. “Depression outcome” has two continuous dimensions. Firstly, the measurement itself, e.g., Hamilton Depression Rating Scale (HDRS). Secondly, the position in time of each recording. When the depression measure itself is categorized (by taking an arbitrary cut-off) small differences, such as 1.7 points on the HDRS, can misleadingly appear significant. Moncrieff and Kirsch present this as the main support for their argument. Unfortunately, they (conveniently?) ignore that depression severity varies continuously over time. When an arbitrary cut-off “end-point” in time is chosen (usually 8 weeks in antidepressant trials) small differences, such as 1.7 points on the HDRS, can lead to much larger effects (see figure)*. Miles Jefferson Consultant Psychiatrist Manchester Royal Infirmary, Oxford Road, Manchester, M13 9WL Miles.Jefferson@mhsc.nhs.uk [1] Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005; 331: 155-7. Competing interests: None declared
Fallacy in The Hamilton Rating Scale 29 July 2005
David H. Marjot, Consultant Psychiatrist. rrey. KT13 8NF Send response to journal: Re: Fallacy in The Hamilton Rating Scale	Sir, The article uses the scores of the Hamilton Rating Scale for Depression to criticise the belief that anti-depressants have some efficacy. In my opinion there is a fatal flaw at the heart of the Hamilton Scale. Hamilton put together behaviours that he regarded as symptoms so as to describe a syndrome that he calls DEPRESSION. He then faced the difficulty of quantifying the qualitative. We are or rather were a nautical nation. It is easy to use the metaphor such as 'depth' to conceive of the intensity of depression. There are no units of measurement for depression such as fathoms or metres but not deterred he constructed a scale using the symbols that when used normally stand for natural numbers such as 0,1,2,3,4. This metaphorical scale now stands for the metaphorical depth. However 0,1,2,3,4 do not represent numbers but are symbols in a code whereby 0 = absent,1 = mild, 2 = moderate, 3 = severe and 4 = incapacitating.You cannot add up these code symbols as if they are natural numbers!You can creat as valid a code with 4 = absent, 2 = mild 1 = moderate, 0 = severe and 3 = incapacitating. This second code is just as valid as the other but you can see now how ludicrous it is in this case to try and add the symbols 01234 as if they were numbers. In the days when one sent information by Telegraph and paid by the word it would be economic to construct a code whereby the code 'number'for the intensity of each symptom would be sent always in the same order on the telegraph message. However you would be very foolish to 'add up' all those code symbols in the Hmilton Rating Scale to try give a score. A profound loss of information but also a grave lapse in logic. I fear the authors of this paper have been on a bit of a wild goose chase in so far as they used The Hamilton Rating Scale in their argument. Competing interests: None declared
Blissless ignorance 29 July 2005
Abraham Petrus Cutajar, Free-lance Gp and sas care of the elderly North Devon. EX39 5EH Send response to journal: Re: Blissless ignorance	Where are they that clobbered others for under-diagnosing and under- treating "eminently treatable" depression? How can statisticians draw opposite conclusions from the same pool of data as the authors and responders have? Do they dice? Confused? Nay, depressed by my uselessness! Competing interests: I precribe/d anti-depressants
Tom Cruise, Psychiatry, Niels Bohr and, the Broken Mirror 30 July 2005
Dr. Avinash R. Patwardhan, Stress Management Consultant 5610 Lee Hwy, Arlington VA USA 22207 Send response to journal: Re: Tom Cruise, Psychiatry, Niels Bohr and, the Broken Mirror	Dear Editor: Recently WebMDHealth published an article about Tom Cruise's widely publicized comments against antidepressants, quoting this article . That article observes that since the interview, psychiatrists across the U.S. have passionately defended the drugs. American Psychiatric Association in a press release accused Tom Cruise of irresponsible behavior. I am intrigued that an army of medical pundits in the society has to rise to defend itself and its science against criticism of a layman. “The opposite of a fact is falsehood, but the opposite of one profound truth may very well be another profound truth”, said Niels Bohr. The profound truth that Tom Cruise might be alluding to is that the fear of unknown drives the hell out of us and the terror of knowledge drives that hell back into us. Such is tragedy of man. This unsettling truth, otherwise subdued in familiar circumstances, becomes glaringly obvious at the cutting edge, the frontiers of science and technology . In the person of Tom Cruise, our society overwhelmed by Future Shock, seems to mouth its trepidations. On the other hand, debate involving scholars like Dr. Moncrieff and Dr. Regier etc. points towards the other profound truth that even science is inherently biased, subjective and, is colored by adjacent milieu. Rigor and peer review does not change it. Statistical science is to truth what democracy is to politics- hopelessly inefficient though the best among the hopeless. Scores, ‘p value’, distributions, formalism of meta-analysis etc. are subject to individuals and their subjectivity. Postmodernism broke the mirror that until recently reflected a neat and sensible view of world. Unless the medical/scientists community start seeing the art in science , it is unlikely that likes of Cruise will see and trust science in it- while humanity helplessly waits the dawn of the age after postmodernism. Competing interests: None declared
Antidepressants work: the question is how well in the real world. 30 July 2005
Ian M Anderson, Senior Lecturer in Psychiatry University of Manchester M13 9PT Send response to journal: Re: Antidepressants work: the question is how well in the real world.	Sir, The debate aired yet again by Moncrieff and Kirsch (1) is an important part of the process of reaching a balanced view about the efficacy of antidepressants and their place in the treatment of depression. It is important not to accept either therapeutic nihilism, or a naively over-optimistic view of antidepressant efficacy. The debate, as usually presented, misses the point, that efficacy trials are poorly designed to quantify effects in usual clinical practice. Moncrieff and Kirsch make two main claims, first that the effect of antidepressants is so small that it is clinically meaningless and second, that the small effect is due to methodological confounds rather than a pharmacological effect. The second point is difficult to sustain given one of the author’s systematic reviews where a difference was seen between antidepressants and ‘active’ placebo (2), the robust effect of antidepressants in relapse prevention (3) which is difficult to explain away as a placebo effect, and the finding that some antidepressants are more effective than others (4). A dispassionate assessment of the available data leaves little doubt that antidepressants do have an effect. The core of the debate is about size of effect. Antidepressant efficacy studies are beset with methodological problems including rating scale inflation to recruit patients, regression to the mean, unrepresentative patient samples, limited duration of treatment (5). These studies, if randomised and blinded, can answer whether antidepressants work, but they cannot answer how well they work in the real world. The common practice of extrapolation from available efficacy studies to give size of effect and numbers needed to treat is simply misguided. We now need to develop the methodology, and have the will and resources, to examine how useful antidepressants really are, and for which patients, rather than continuing to pontificate on the basis of inadequate data. 1 Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005; 331, 155-157 2 Moncrieff J, Wessely S, Hardy R. Meta-analysis of trials comparing antidepressants with active placebos. Br J Psychiatry 1998; 172, 227-231. 3 Geddes JR., Carney SM, Davies C, Furukawa T, Kupfer D, Frank E et al Relapse prevention with antidepressant drug treatment in depressive disorders: a systematic review. Lancet 2003; 361, 653-661. 4 Smith D, Dempster C, Glanville J, Freemantle N, Anderson I. The efficacy and tolerability of venlafaxine compared with selective serotonin reuptake inhibitors and other antidepressants: a meta-analysis. Br J Psychiatry 2002; 180, 396-404 5 Anderson I M, Haddad P Clinical trials of antidepressant medications are producing meaningless results (In Debate) Br J Psychiatry 2003; 183:102-104. Competing interests: The author has published reviews of antidepressant efficacy and has received honoraria, support to attend conferences and grants from pharmaceutical companies marketing antidepressants
The future of SSRIs 31 July 2005
Steven Taylor, Professor Dept of Psychiatry, Univ of British Columbia, Vancouver, BC, Canada, V6T 2A1 Send response to journal: Re: The future of SSRIs	To the Editor: In their recent BMJ article, Moncrieff and Kirsch [1] questioned the conclusions from the National Institute for Health and Clinical Excellence (NICE) about the efficacy of selective serotonin reuptake inhibitors (SSRIs) in the treatment of depression. The NICE recommendations were that SSRIs are a first-line treatment for moderate or severe depression. In their review of the evidence, Moncrieff and Krisch concluded that SSRIs have no clinically meaningful advantage over placebo, and have not been shown to affect the long-term outcome of either placebo or suicide rates. These conclusions have been presented several times in the past by Kirsch and colleagues [e.g., 2]. What has been lacking in these discussions, is a detailed consideration of the implications of these findings. If Kirsch and colleagues are correct, then there are many important implications bearing examination. I would like to briefly offer four implications. First, regardless of whether one agrees with the conclusions of Kirsch and colleagues, the controversy over the efficacy of SSRIs would never have arisen if these medications were highly powerful treatments for depression. Clearly, these drugs are not satisfactory and there is a pressing need to develop more potent treatments. Second, the evidence generally suggests that antidepressant medications have much the same efficacy as psychological therapies (cognitive-behaviour therapy and interpersonal therapy) for outpatients suffering from major depression [e.g., 3]. If SSRIs and another antidepressant medications are simply placeboes, then the same can be said of the psychological therapies—at least when short-term efficacy is considered. Over the longer term, psychological therapies tend to have a lower relapse rate than antidepressant medications [3]. Does this mean that psychotherapies are simply a more enduring form of placebo? The key to better understanding the treatment of depression may require a better understanding of the nature of the placebo effect. Third, it is possible that SSRIs (and other depression treatments) are more effective for some forms of depression than others. This idea has been the subject of research for many years, and so far the results have been disappointing. As Moncrieff and Kirsch point out, there is no persuasive evidence that SSRIs are selectively more effective in severe depression. Similarly, in our research and the work conducted by others, it has been found that tricyclic antidepressants are not differentially more effective than cognitive-behavioural therapies in severe depression [4]. In other words, there is no convincing evidence of a treatment-by- severity interaction in the treatment of depression. It is possible that other forms of classification may reveal specific indications for SSRIs, but the findings so far have not been encouraging. In a twin study of the heritability of depressive symptoms, we recently reported that some symptoms (particularly the somatic symptoms such as fatigue) were heritable, whereas other symptoms (e.g., psychological symptoms such as low self-esteem) were not heritable [5]. It could be speculated that SSRIs would be more effective than placebo for the more heritable symptoms, and no better than placebo for the non-heritable symptoms. However, if this was the case, then one would expect that SSRIs would show an overall superiority over placebo on measures based on the sum of severity scores for depressive symptoms. The research does not support this conclusion [1], and so the search continues for specific indications for SSRIs. It may be time for researchers, clinicians, and the pharmaceutical industry to consider the possibility that there may be no forms of depression for which SSRIs are particularly efficacious. It may be time to look for more effective treatments. Finally, the evidence suggests that SSRIs are clearly more effective than placebo in the treatment of anxiety disorders. For example, in a meta -analysis of treatments for posttraumatic stress disorder, we found that SSRIs (and cognitive-behavioural therapies) were clearly more effective than placebo [6]. In another meta-analysis, we found similar results for social phobia [7]. These findings, along with the conclusions of Moncrieff and Kirsch, suggest that SSRIs may be more useful for anxiety disorders rather than as treatments for mood disorders. Steven Taylor, Ph.D., Professor, Department of Psychiatry, University of British Columbia, Vancouver, BC, Canada, V6T 2A1, E-mail: taylor@unixg.ubc.ca. References 1. Moncrieff, J., & Kirsch, I. (2005). Efficacy of antidepressants in adults. British Medical Journal, 331, 155-157. 2. Kirsch, I., & Sapirstein, G. (1998). Listening to Prozac but hearing placebo: A meta-analysis of antidepressant medication. Prevention & Treatment, 1. 3. Hollon, S. D., DeRubeis, R. J., Shelton, R. C., Amsterdam, J. D., Salomon, R. M., O’Reardon, J. P., et al. (2005). Prevention of relapse following cognitive therapy vs medications in moderate to severe depression. Archives of General Psychiatry, 62, 417-422. 4. McLean, P., & Taylor, S. (1992). Severity of unipolar depression and choice of treatment. Behaviour Research and Therapy, 30, 443-451. 5. Jang, K. L., Livesley, W. J., Taylor, S., Stein, M. B., & Moon, E. C. (2004). Heritability of individual depressive symptoms. Journal of Affective Disorders, 80, 125-133. 6. van Etten, M., & Taylor, S. (1998). Comparative efficacy of treatments for posttraumatic stress disorder: A meta-analysis. Clinical Psychology and Psychotherapy, 5, 126-145. 7. Fedoroff, I. C., & Taylor, S. (2001). Psychological and pharmacological treatments for social phobia: A meta-analysis. Journal of Clinical Psychopharmacology, 21, 311-324. Competing interests: None declared
Use of antidepressants in childhood depression 5 August 2005
Jonathan M Goldin, Consultant Child and Adolescent Psychiatrist Great Ormond Street Hospital for Children, London WC1N 3JH Send response to journal: Re: Use of antidepressants in childhood depression	Editor, I was most interested to read Moncrieff and Kirsch's paper on the efficacy of antidepressants in adults(1). During my years of clinical practice in child psychiatry I have never been convinced of the effectiveness of antidepressants to treat childhood depression. In some cases they seem to help, but one cannot be sure to what extent the improvement would have occurred with the passage of time and other changing factors. In some quarters, unfortunately, it seems almost heretical to question the value of antidepressants and I think this is at least partly due to the influence of "big pharma" who fund outcome research and who clearly have a massive interest in promoting such medications (2). In my experience, patients and their families do tend to prefer talking therapies to medication in treating children with low mood and other features of depression. However, this inevitably raises resource issues and it is far easier to reach for the prescription pad than grapple with complexity and address underlying issues. (Of course it is possible to usefully combine psychological and pharmacological approaches). There has been an increasing amount of evidence casting doubt on the widespread prescription of antidepressants to children (3) and I welcome the freer debate that now seems to be opening up. (1) Moncrieff and Kirsch Efficacy of antidepressants in adults BMJ July 05 331:155-57 (2) Ferner The influence of big pharma BMJ April 2005 330:855-56 (3) Jureidini et al Efficacy and safety of antidepressants for children and adolescents BMJ April 04 328: 879-883 Competing interests: None declared
Authors Response 16 August 2005
Joanna Moncrieff, Senior Lecturer Department of Mental Health Sciences, University College London, W1N 8AA, Irving Kirsch Send response to journal: Re: Authors Response	We are glad that we have generated so much discussion. Firstly we want to respond to the accusation by the authors of the NICE guidelines that our paper contained “inaccuracies” (Pilling et al, rapid responses). In response to the first substantial point, the authors contend that they based their conclusions on mean endpoint data. We quote from the full guideline itself: “There is evidence to suggest that there is a statistically significant difference favouring SSRIs over placebo on reducing depression symptoms as measured by the HRSD but the size of this difference is unlikely to be of clinical significance.” P. 186-7. In the conclusions to this section entitled “Clinical Summary” they again state that “there appears to be no difference between SSRIs and placebo on mean endpoint or change scores.” P 189 (1) Therefore if NICE’s recommendations are based on endpoint data, as they say they are in their response, they should have recommended that SSRIs are not used at all. In response to their point about long-term outcomes of depression, and to the other authors that raised a similar point, firstly we were not talking about whether long-term antidepressant treatment is effective, but were making the point that the long-term benefit of short term treatment has not been established. This is because most trials are short term, and other population based measures of outcome do not indicate observable benefits. Secondly however, we suggest that results of trials of long-term antidepressant treatment are potentially even more flawed than trials of short term treatment. These trials are discontinuation trials in which one group of people are withdrawn to placebo treatment and compared with patients who continue their antidepressant. The outcome in the placebo group probably reflects the problems of withdrawal rather than the benefits of treatment. Since these trials were conducted prior to recent widespread publicity about antidepressant withdrawal reactions, there was no consideration of this possibility and it is highly likely that symptoms of withdrawal and anxiety associated with it was classified as relapse in some cases. This is supported by the observation that “relapse” rates in the placebo group are the same regardless of the duration of prior treatment whether it is 3 weeks or 4 years (2). There are few true continuation trials that look at the outcome of people continued on drug treatment or placebo after a short term trial. Such that exist show no difference between antidepressants and placebo in terms of relapses (3). In response to the point about severity of depression, we reiterate that the NICE data did not demonstrate a gradient. The Ansgt et al (4) meta-analysis cited also did not show convincing evidence of a gradient and other research we cited suggests that antidepressants are not very effective in more severe conditions. We did acknowledge that some research had shown such a gradient (5) and we thank those authors who drew our attention to the fact that the same group have reproduced this finding (6). We still feel the evidence so far is inconclusive. Professor Taylor (rapid responses) cites further evidence against a relationship of efficacy and severity for tricyclic antidepressants as well as SSRIs. Several respondents (Wohlfarth et al, rapid responses, Harrison-Read, rapid responses, Cowie, rapid responses) suggest that if response to antidepressants was bimodal then our critique of the categorisation of continuous data would not hold. We illustrated our point using Normal data because Normal distributions are common and many antidepressant trials use parametric statistics that assume Normal distributions. However, a similar conclusion could be reached using bimodal distributions. As an example take the hypothetical graph displayed in our article. In the drug condition add one point to each score above the cut-off for classifying a patient as a responder. Now you have a bimodal distribution, with a trough going all the way down to zero. In this case, the mean difference between drug and placebo will be 2 points, which will be the exact advantage for each patient given the active drug (given the assumption made to generate the graph). Nevertheless the response rates will remain 33% and 50%. It is interesting to hear about the data from Wohlfarth et al (rapid responses), which could be very useful. However, what would be helpful would be graphs comparable to the one we demonstrated using real data of both drug and placebo score distributions. On this point, one of the reasons continuous measures have been favoured is because there is no clear distinction between depression and normality. Defining “clinical depression” is not straightforward, despite Sanati’s (rapid responses) and others’ aspirations. Harrison-Read (rapid responses) suggests that failure to demonstrate larger differences may be due to heterogeneity of patients entered into trials. While this is not in doubt, decades of research have failed to identify a homogenous group that respond to antidepressants. It seems simpler to conclude that antidepressants don’t work than to continue with the endless search for the responsive subgroup. There is nothing in our paper which suggests that drug treatment should be replaced wholesale by psychotherapy. Instead we feel that the whole area of depression needs radical rethinking and that people need to be helped to reconceptualise their feelings in ways which empower them rather than making them passive recipients of medical treatment. This point of view has been put forward by several authors, among them by Christopher Dowrick in his recent book “Beyond depression” (7). In response to Wheatley (rapid responses), not all practising psychiatrists do agree that antidepressants are useful. Many of them have written to us to say that they too have never believed antidepressants were effective. The suggestion of using the criterion of HRSD score of 10 or less suffers from the same problems of potential inflation of differences as using other categorical adaptations of continuous data. In response to Jefferson (rapid responses), we know of know evidence that drug placebo differences increase over time. The duration of an antidepressant trial has not been found to predict outcome in meta- analyses where this has been investigated (8). It seems just as plausible that differences decrease over time. In response to other points made by Thelander (rapid responses) we agree that the HRSD is particularly susceptible to non specific sedative effects, but the HRSD 6 also contains items on anxiety and somatic symptoms that might respond to sedative effects, and does not necessarily show antidepressants in any better light than the standard HRSD scale (9). The Cochrane active placebo meta-analysis discusses the possibility of atropine having antidepressant effects in the discussion section. In our article we mention research which suggests that many drugs not traditionally called antidepressants, including antipsychotics, barbiturates and benzodiazepines have been shown to have comparable effects to antidepressants in depressed patients. In contrast to Anderson’s (rapid responses) claims for superiority of venlafaxine, most analyses conclude that no type of drug action has been shown to be superior to any other (10). It therefore seems that effects of drugs on depression are not due to any specific pharmacological mechanism but more likely due to non specific pharmacological or psychological effects. This means effects seen in trials and in real life are likely to be due to a mixture of pharmacological effects such as sedation, which improves rating scale scores and may mask, and therefore temporarily reduce feelings of depression, and expectancy effects due to being on a drug that is experienced as being active. Results of imaging studies are not consistent, with some showing similar changes in drug and placebo responders (11). However, even if differences are demonstrated, this only shows that antidepressants are centrally active agents, and says nothing about their effects on mood. Finally, we agree with the concern voiced by Dr Virani (rapid responses) about the influence of the pharmaceutical industry on perceptions and prescribing of antidepressants. The NICE guideline notes that there was evidence of publication bias in the NICE data, which means the small effects found are likely to overestimate real differences. (1) NICE. Treatment and management of depression in primary and secondary care. 2004. NICE. (2) Viguera AC, Baldessarini RJ, Friedberg J. Discontinuing antidepressant treatment in major depression. Harv Rev Psychiatry 1998; 5:293-306. (3) Hypericum Depression Trial Study Group. Effect of Hypericum perforatum (St John's wort) in major depressive disorder: a randomized controlled trial. JAMA 2002; 287:1807-1814. (4) Angst J. Severity of depression and benzodeiazepine comedication in relationship to efficacy of antidepressants in acute trials. A meta-analysis of moclobemide trials. Human Psychopharmacol 1993;8:401-407. (5) Khan A, Leventhal RM, Khan SR, Brown WA. Severity of depression and response to antidepressants and placebo: an analysis of the Food and Drug Administration database. J Clin Psychopharmacol 2002; 22:40-45. (6) Khan A, Brodhead AE, Kolts RL, Brown WA. Severity of depressive symptoms and response to antidepressants and placebo in antidepressant trials. J Psychiatr Res 2005; 39(2):145-150. (7) Dowrick C. Beyond Depression. 2004. Oxford University Press:Oxford. (8) Moncrieff J. A comparison of antidepressant trials using active and inert placebos. Int J Methods Psychiatr Res 2003; 12:117-127. (9) Bech P, Cialdella P, Haugh MC, Birkett MA, Hours A, Boissel JP et al. Meta-analysis of randomised controlled trials of fluoxetine v. placebo and tricyclic antidepressants in the short-term treatment of major depression. Br J Psychiatry 2000; 176:421-428. (10) Freemantle N, Anderson IM, Young P. Predictive value of pharmacological activity for the relative efficacy of antidepressant drugs. Meta-regression analysis. Br J Psychiatry 2000; 177:292-302. (11) Mayberg HS, Liotti M, Brannan SK, McGinnis S, Mahurin RK, Jerabek PA et al. Reciprocal limbic-cortical function and negative mood: converging PET findings in depression and normal sadness. Am J Psychiatry 1999; 156:675-682. Competing interests: None declared
RE: Efficacy of antidepressants in adults - Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005;331(7509):155-7 9 September 2005
Frederic M. Quitkin, Director, Depression Evaluation Service New York State Psychiatric Institute, 1051 Riverside Drive, New York, NY 10032, Jonathan W. Stewart, Patrick J. McGrath, Donald F. Klein Send response to journal: Re: RE: Efficacy of antidepressants in adults - Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005;331(7509):155-7	To the Editor -- Moncrieff and Kirsch (“Efficacy of antidepressants in adults” [1]) criticize the National Institute for Health and Clinical Excellence (NICE) recommendation that SSRIs should be the first line treatment for depression, [2] concluding “given doubt about their benefits and concern about their risk, current recommendations for prescribing antidepressants should be reconsidered.” [1] We suggest that Moncrieff and Kirsch’s review is selective and biased. The authors quote Fergusson et al. [3] and Gunnell et al. [4] as supporting “concern that SSRI’s increase the risk of suicidal behavior.” While the original Fergusson et al. meta-analysis suggested increased suicide on SSRI’s, Moncrieff and Kirsch do not cite subsequent letters to the editor pointing out statistical errors in Fergusson et al. [5] or Fergusson and colleagues’ post-publication retraction. [6] Similarly, Gunnell and colleagues [4] did not find statistically significant increased self-harm on SSRIs similar to the FDA’s position [7] but counter to the Moncrieff and Kirsch conclusion partly based on the same article. Moncrieff and Kirsch criticise the “key claim in the NICE guidelines is the superiority of antidepressants over placebo correlate with severity of depression.” However, the proper test for such a relationship is heterogeneity of slope in ANCOVA, not a correlation coefficient. [8] Studies drawing different conclusions than Moncrieff and Hirsch used these or similar techniques. [9,10] Other meta-analyses supporting the NICE guidelines are not cited. [e.g., 11,12] By ignoring alternative views, Moncrieff and Hirsch appear to have selectively chosen literature that can be interpreted, and perhaps misinterpreted, as supporting what appear to be preconceptions. Such a biased rendering of the literature lends little credence to their tendentious statement that the balance of benefits to risks of antidepressant medications in adults may be “unfavorable”. REFERENCES 1. Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ 2005;331(7509):155-7. 2. National Institute for Health and Clinical Excellence (NICE). Depression:management of depression in primary and secondary care. Clinical practice guideline. No. 23. London:NICE, 2004:http://www.nice.org.uk/page.aspx?o=235213. 3. Fergusson D, Doucette S, Glass KC, Shapiro S, Healy D, Hebert P, et al. Association between suicide attempts and selective serotonin reuptake inhibitors: systematic review of randomised controlled trials. BMJ 2005;330(7488):19. 4. Gunnell D, Saperia J, Ashby D. Selective serotonin reuptake inhibitors (SSRIs) and suicide in adults: meta-analysis of drug company data from placebo controlled, randomised controlled trials submitted to the MHRA's safety review. BMJ 2005;330(7488):19. 5. Sakinofsky I, Streiner DL. Do selective serotonin reuptake inhibitors cause suicide? Let's keep it in perspective. BMJ 1149;330(7500):14. 6. Correction for Fergusson et al., BMJ 330 (7488) 396. British Medical Journal, 2005. http://bmj.bmjjournals.com/cgi/content/full/330/7492/653-a 7. US Food and Drug Administration. FDA Public Health Advisory: Worsening depression and suicidality in patients being treated with antidepressant medications. Available at http://www.fda.gov/cder/drug/antidepressants/antidepressantpha.htm. Accessed July 11 2005. 8. Klein, DF & Ross, DC. Reanalysis of the National Institute of Mental Health Treatment of Depression Collaborative Research Program General Effectiveness Report. Neuropsychopharmacology 8(3):241-252, 1993. 9. Elkin I, Gibbons RD, Shea MT, Sotsky SM, Watkins JT, Pilkonis PA, Hedeker D. Initial severity and differential treatment outcome in the National Institute of Mental Health Treatment of Depression Collaborative Research Program. J Consult Clin Psychol 1995 63:841-7. 10. Leon AC. Solomon DA. Mueller TI. Endicott J. Rice JP. Maser JD. Coryell W. Keller MB. A 20-year longitudinal observational study of somatic antidepressant treatment effectiveness. American Journal of Psychiatry 2003;160:727-33. 11. Khan A, Leventhal RM, Khan S, Brown WA. Severity of depression and response to antidepressants and placebo: an analysis of the Food and drug Administration Database. J Clin Psychopharmacol 2002;22:40-5. 12. Khan A, Brodhead AE, Kolts RL, Brown WA. Severity of depressive symptoms and response to antidepressants and placebo in antidepressant trials. J Psychiatric Res 2005;39:145-50. Competing interests: FMQ has served as a consultant for Organon, Inc., Sepracor, and Bristol-Myers, has received honoraria from Pfizer, has been reimbursed by Eli Lilly for serving as a speaker, has shares in Cyberonics, and has offered expert testimony for Sedgwick, Detert, Moran and Arnold, Sterne, Kessler, Goldstein and Fox, McKinsey and Company, and Gerson Lehman Group. JWS has served as a consultant for Organon, Inc., Shire, and Somerset, has received research support from Lilly, Pfizer, Glaxo, Organon, Inc. and Bristol-Myers, has received honoraria from Eli Lilly, Pfizer, and Organon, Inc., and has been reimbursed by Eli Lily, Pfizer, Organon, Inc. and Glaxo for serving as a speaker or on an advisory board. DFK has be reimbursed for scientific advisory to Eli Lilly, Glaxo, Pfizer, Roche Laboratories, Smithkline Beecham, Upjohn, Vela Pharmaceutical and VivoMetrics, Inc., has served on Speaker’s Bureau at Eli Lilly, Glaxo, Pfizer, Roche Laboratories, Smithkline Beecham and Upjohn, has served as a consultant to Eli Lilly, Roche Laboratories, Smithkline Beecham, and Upjohn, and is a stockholder in Forrest Laboratories and Interneuron Pharmaceuticals, Inc. PJM has no competing financial interests.