Letters

Naturalistic treatment study of depression in general practice

Clinical management is important in treatment of depression

Antidepressants are overrated

Author's reply

Clinical management is important in treatment of depression

EDITORThe striking finding of Malt et al's study in patients with depression is not the superiority of combined pharmacotherapy and psychotherapy but the 47% complete response rate to placebo and psychological support, which was present from about 12 weeks onwards.¹ Unfortunately, little information is given about the quantity and quality of the psychological support provided, although reference is made to the clinical management protocol used in the National Institute of Mental Health's treatment of depression study, in which patients were seen regularly for 20-30 minute sessions of supportive therapy. ^{2 3} In that study, at 16 weeks 42% of patients in the imipramine-clinical management group had recovered compared with 21% in the placebo-clinical management group.³

The difference between the studies is unlikely to be due to severity of illness as in both studies the initial rating indicated illnesses of moderate severity (Malt et al's study, mean initial Montgomery Åsberg depression rating scale=26.7; National Institute of Mental Health study, mean initial Hamilton rating scale for depression=19.5). However, the recovery criteria in Malt et al's study were less stringentnamely, a 50% reduction in the score on the Montgomery Åsberg depression rating scale and improvement to at least only mild illness remaining. To compare the results usefully with those of other studies, data needed to be presented on the proportion of patients achieving full remission (a score  6).

The authors state that the findings challenge current guidelines that claim equality of effect between drug and psychological treatment in mild to moderate depression. These guidelines, however, have been based on controlled trials of formal psychotherapies rather than clinical management. In the National Institute of Mental Health study, cognitive therapy and interpersonal therapy were both superior to placebo-clinical management.³ What the findings do indirectly confirm is the crucial importance of good clinical management in enhancing the effectiveness of drug treatment. This has been shown in controlled trials showing the superiority of treatment based on guidelines compared with usual treatment of depression in primary care. ^{4 5}

Peter L Cornwall, senior lecturer in community psychiatry. 
University Department of Psychiatry, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP p.l.cornwall{at}ncl.ac.uk

Antidepressants are overrated

EDITORThe most interesting aspect of the results of the trial by Malt et al is that the difference between antidepressants and placebo was not impressive and was of doubtful clinical relevance.¹ At most it consisted of a difference of 3 points on a scale with a maximum of 60 points and a pretreatment average value of 27. In most subjects whose depression was classified as severe or major depression the difference was smaller and not significant.

At least part of this difference probably represents the non-specific effects of taking an active drug as opposed to an inert placebo. It has been found that both researchers and patients can distinguish antidepressants, including selective serotonin reuptake inhibitors, from inert placebos. ^{2 3} This "unblinding" effect may be important, since trials comparing antidepressants with active placebos have generally found smaller effects.⁴

Side effects are obviously an important way in which the identity of a drug or placebo is revealed. Unfortunatelyespecially since the doses of drugs used were higher than those used routinelyMalt et al provide no information on this outcome except to say that side effects were "few and rather insignificant." But patients may be able to identify that they are taking an active drug without necessarily reporting side effects. The authors did not investigate whether patients could identify which group they were in, but they do report that none of the general practitioners could identify the patient-drug combination correctly. It is not clear, however, that this was assessed systematically, since by chance some of the general practitioners would be expected to guess correctly.

These reservations about the importance of Malt et al's findings are consistent with a recent meta-analysis which found that differences between antidepressants and placebo were less noticeable than has traditionally been assumed.⁵ Doctors must be cautious about recommending routine drug treatment for mild and moderate depression in general practice on ethical as well as scientific grounds. It is clearly in the interests of the pharmaceutical industry to promote the notion that drug treatment can resolve human unhappiness and the problems of living. Professionals have a duty to consider the wider implications of such a scenario.

Joanna Moncrieff, specialist registrar in psychiatry. 
Department of Psychological Medicine, Chelsea and Westminster Hospital, London SW10 9NG joannamoncrieff{at}compuserve.com

1.	Malt U, Robak HO, Madsbu H-P, Bakke O, Loeb M. The Norwegian naturalistic treatment study of depression in general practice (NORDEP)1: randomised double blind study. BMJ 1999; 318: 1180-1184. (1 May.)
2.	White K, Kando J, Park T, Waternaux C, Brown WA. Side effects and the "blindability" of clinical drug trials. Am J Psychiatry 1992; 149: 1730-1761[Abstract/Free Full Text].
3.	Kranzler HR, Burleson JA, Korner P, del Boca FK, Bohn MJ, Brown J, et al. Placebo controlled trial of fluoxetine as an adjunct to relapse prevention in alcoholics. Am J Psychiatry 1995; 152: 391-397[Abstract/Free Full Text].
4.	Moncrieff JM, Wessely S, Hardy R. Meta-analysis of trials comparing antidepressants with active placebos. Br J Psychiatry 1998; 172: 227-231[Abstract/Free Full Text].
5.	Bollini P, Pampallona S, Tibaldi G, Kupelnick B, Munizza C. Effectiveness of antidepressants. Meta-analysis of dose-effect relationships in randomised clinical trials. Br J Psychiatry 1999; 174: 297-303[Abstract/Free Full Text].

Author's reply

EDITORCornwall asks for the proportion of patients achieving what he calls full remission (score on the Montgomery Åsberg depression rating scale <7) in our study to compare it with Elkin et al's study. After 16 weeks 38/87 (44%) taking sertraline (odds ratio 0.45; 95% confidence interval 0.23 to 0.89) and 31/84 (37%) taking mianserin (0.60; 0.30 to 1.20) had a total score of <7 compared with 23/89 (26%) taking placebo.

Our results are comparable to those of Elkin et al's study.¹ Nevertheless, we think that emphasising just a score of <7 at a given time addresses only part of the picture. Because of the serious consequences of depression, differences in the proportion of subjects having final scores in the range of 7-11 with a given treatment are also clinically important. The difference in clinical disability between patients with a score of, say, 8 or 18 may be that of being able to work or not.

Moncrieff suggests that the additional benefit of drugs in our study is due to non-specific effects of active treatment. This argument has less validity in primary care. The studies cited by Moncrieff et al are conducted with depressed outpatients, not primary care attenders.² In our experience, primary care attenders have strikingly fewer side effects than do psychiatric outpatients, whatever the drug given. In our study the general practitioners were required to explore possible side effects systematically. This method provides a higher prevalence of side effects than just recording spontaneously reported side effects. The mean number of baseline-corrected side effects during the study was 7.11, 6.51, and 6.45 after 8 weeks of treatment with sertraline, mianserin, or placebo respectively, and 3.16, 3.09, and 3.02 after 24 weeks of treatment.

Unipolar and bipolar depressions are the leading causes of the world's total disability adjusted life years. When several features of this study (statistical power 0.95, 70% completing 16 weeks of treatment, no exclusion of patients who responded to placebo, measurement of serum drug concentrations to control for compliance, reliable raters, psychological support as it is provided in general practice) are considered, the superiority of combined drug-psychological treatment of depression in primary care is remarkable. If this treatment was adopted worldwide a considerable reduction in disability adjusted life years should be achieved.

We agree, however, that good clinical management is the key to therapeutic success. The high response rate to good clinical management and placebo is remarkable indeed. More research on identifying predictors of response to different treatments is obviously needed.

Ulrik Fredrik Malt, professor of psychiatry. 
Department of Psychosomatic and Behavioural Medicine, University of Oslo, N-0027 Oslo, Norway u.f.malt{at}psykiatri.uio.no