A Perspective article published early online in the New England Journal of Medicine comments on the recent proposal by the FDA to enhance the 'black box' warning on antidepressive drugs marketed in the US. The new text adds warnings on the risk of suicidality (suicidal thoughts or behaviour) in young adults during the first few weeks of treatment. It also includes a warning that depression itself carries a risk of suicide - an inclusion that the authors of the Perspective note is unique, as no other black box warnings also refer to a risk from the disease being treated. They note that clinicians have long appreciated that there is an increase in suicide risk in the early stages of the treatment of depression, and discuss the efforts undertaken by the FDA to control for this in the analysis presented to the advisory committee that proposed the warning.

This huge analysis included data on 99,839 participants who had enrolled in 372 randomised clinical trials of antidepressants conducted by 12 pharmaceutical companies during the past two decades. For the primary analysis, only trials in psychiatric disorders were included, however this included the majority: 39,729 assigned to investigational drug, 27,164 to placebo, and 10,489 to an active comparator. There were only eight suicides reported in this set of data, but 501 had suicidal thoughts or behaviour: 243 while receiving an investigational drug, 194 while receiving placebo, and 64 while receiving an active comparator. There was no significant difference between the groups overall, however, the risk for patients 18 to 24 years of age was elevated, albeit not significantly (odds ratio, 1.55; 95% CI, 0.91 to 2.70). The authors of the Perspective note that although this was not statistically significant, the threshold for safety is lower than that for efficacy; there is also evidence of a trend when data from younger groups is included.

Nevertheless, the authors suggest, the data set used has significant deficiencies when used for this purpose: for example, most of the studies were assessing short term efficacy, not long-term safety. Suicidality reports come from adverse effect reporting, leading to a high risk of ascertainment bias - for example patients reporting one adverse effect are more likely to be asked whether they have had others. The greatest difficulty in interpretation, however, may come from confounding by indication - i.e. prescribing for patients at risk of suicidality because of their disease. Analysis of the data from trial in non-psychiatric indications showed no increase in suicidality.

Finally, the authors comment that the new warning has the potential to confuse, and may lead to a reduction in the use of these drugs when they are appropriate. They note that the risk from untreated depression is much greater than that from the drugs; however doctors in the US must now warn patients of the possibility of increased suicidality, and should follow them carefully for the first few weeks of treatment.

Link: http://content.nejm.org/cgi/content/full/NEJMp078015v1
 

Antidepressive drugs may not help in bipolar disorder?

Source: New England Journal of Medicine 2007; 356: 1711-22, 1771-3

A controlled trial published in the NEJM (New England Journal of Medicine) found that adding an antidepressive drug to a mood stabiliser in patients with bipolar disorder did not improve depressive symptoms - although neither did it increase the risk of switch to mania. The authors comment that depressive symptoms are a major cause of disability in these patients, but that it is uncertain whether antidepressive drugs are beneficial or not in reducing depressive symptoms. There is also a concern that these drugs may trigger a switch to mania.

This trial was intended to clarify the effects of antidepressive drugs in patients with bipolar disorder under routine practice conditions. It involved patients taking part in a clinical study programme sponsored by the US National Institute of Mental Health: Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) is intended to study the effectiveness of treatments for bipolar disorder in the community, with a large patient population who can be assigned to different treatments. For the current study, a group of patients from the programme were randomised to treatment with one of two antidepressive drugs (bupropion m/r or paroxetine) or placebo, in addition to their mood stabilising therapy. Primary outcome was durable recovery, defined as euthymia* for at least eight weeks; secondary outcomes included rates of mania emerging during treatment. Treatment duration was up to 26 weeks.

STEP-BD includes 4360 patients, but only 366 were randomised in this study - 179 to antidepressive (paroxetine n=93, bupropion n=86) and 187 to placebo. Mean duration of treatment was just over twelve weeks (active 88 weeks, placebo 84 weeks). There were no significant differences between the groups in primary outcome. Neither were there significant differences for any of the secondary outcomes, including switch to mania, transient remission, or medication withdrawal due to adverse events. The authors therefore conclude that treatment of patients with bipolar disorder with an antidepressive drug in addition to a mood stabiliser has no benefits, however it also has no impact on the risk of switch to mania.

An accompanying editorial discusses the study. The author notes that there are significant differences in opinion between the US, where this study was carried out, and Europe. The general consensus amongst European experts considers antidepressives to be useful and unlikely to induce mania in bipolar disorder. He notes a number of caveats on the trial, and suggests that it is unlikely to change currently held beliefs on the place of these drugs in the condition. Treatment of these patients should continue to be based on individual patient's clinical history and response to treatment.

Link: http://content.nejm.org/cgi/content/full/356/17/1771