Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Antidepressants for bipolar depression

Systematic review

The use of antidepressants as monotherapy for depression in patients with bipolar depression is one of those areas where there is genuine uncertainty about its value. There may be benefits, but there may also be harm. The benefit would be a reduction in depressive symptoms, though this is by no means certain. The harm may come from the induction of manic episodes or mood instability. A superb systematic review [1] helps in understanding the amount of evidence and the likely benefits and harms.

Systematic review

The review used an extensive search strategy for randomised double blind studies comparing antidepressants with placebo, or with other drug treatments. Many electronic databases were searched, with the use of a huge number of individual drug names, and without exclusion of language.

Patients included in the trials had to have a current depressive or mixed depressive/manic episode and with at least one previous episode of mania or hypomania. The main outcome measures were clinical response and remission rates derived from observer-rated symptom reductions, induction of mania or hypomania, and discontinuation.


Twelve randomised, double blind trials were included, five with placebo comparators, four comparing antidepressants, and three comparing antidepressants to other types of drug. Most patients were in the placebo comparisons. Duration was 4-10 weeks, in adults, and with a majority of women.

Clinical response was found in 123/213 (58%) of patients on antidepressant, compared with 153/449 (38%) on placebo (Figure 1). Three of these studies, including the largest, had some or all patients on lithium or olanzapine. The relative benefit was 1.9 (95% confidence interval 1.5 to 2.3), and the number needed to treat to produce one patient with clinical response was 4.2 (3.2 to 6.4).

Figure 1: Clinical response in randomised trials comparing antidepressant with placebo

Switching to mania was rare, occurring in 11/287 (3.8%) of patients on antidepressant, compared with 23/492 (4.7%) on placebo (Figure 2). There was no significant difference, with a relative risk of 1.0 (0.5 to 2.0). Withdrawal was significantly lower in patients on antidepressant (32%) than on placebo (43%).

Figure 2: Switching to mania in randomised trials comparing antidepressant with placebo

There is tentative evidence that tricyclic antidepressants may be less effective than other antidepressants, but this is based on a small number of trials with few patients. The same comparisons consistently found a higher rate of switching to mania with tricyclic antidepressants (19/184; 10%) than with other antidepressants (6/186; 3%), and though this was statistically significant with a relative risk of 2.9 (1.3 to 6.7), the number of events was very small.


There are some papers that it is both a joy and a privilege to read, because they combine a good analysis of the evidence with clinical wisdom, and simultaneously give a wider perspective to the subject. This is one such, and if it is not required reading in psychiatry (and outside psychiatry), then perhaps it should be, as it has all these qualities.

It beautifully dissects how small positive or negative studies have swayed opinion. For instance, it describes studies that have been included in other (non-systematic) reviews, despite having different diagnoses and no bipolar patients at all.

The bottom line is the conclusion, that, on the basis of current evidence, it is probably overcautious and potentially not in the best interests of patients to discourage the use of antidepressants for bipolar depression.


  1. HJ Gijsman et al. Antidepressants for bipolar depression: a systematic review of randomized, controlled trials. American Journal of Psychiatry 2004 161: 1537-1547.

previous or next story