Antidepressants for depression in medical illness

To determine whether antidepressants are clinically effective and acceptable for the treatment of depression in people who also have a physical illness.

Medline, Cochrane Library Trials Register and Cochrane Depression and Neurosis Group Trials Register were all systematically searched, supplemented by hand searches of two journals and reference searching.

All relevant randomised trials comparing any antidepressant drug (as defined in the British National Formulary) with placebo or no treatment, in patients of either sex over 16, who have been diagnosed as depressed by any criterion, and have a specified physical disorder (for example cancer, myocardial infarction). "Functional" disorders where there is no generally agreed physical pathology (e.g. irritable bowel syndrome) were excluded. The main outcome measures are numbers of individuals who recover/improve at the end of the trial and, as a proxy for treatment acceptability, numbers who complete treatment.

Data was extracted independently by the reviewers onto data collection forms and differences settled by discussion.

18 studies were included, covering 838 patients with a range of physical diseases (cancer 2, diabetes 1, head injury 1, heart 1, HIV 5, lung 1, multiple sclerosis 1, renal 1, stroke 3, mixed 2). Depression was diagnosed clinically in 3 studies, otherwise by structured interview or checklist. Only 5 studies described how they performed randomisation. 1 study compared drug with no treatment, and the rest with placebo: all of the latter said they were double blind. 6 studies used SSRIs, 3 atypical antidepressants, and the remainder tricyclics. Patients treated with antidepressants were significantly more likely to improve than those given placebo (13 studies, OR 0.37, 95% CI 0.27-0.51) or no treatment (1 study, OR 3.45, 95% CI 11.1-1.10). About 4 patients would need to be treated with antidepressants to produce one recovery from depression which would not have occurred had they been given placebo (NNT 4.2, 95% CI 3.2-6.4). Most antidepressants (tricyclics and SSRIs together, 15 trials ) produced a small but significant increase in dropout (OR 1.66, 95% CI 1.14-2.40. NNH 9.8, 95% CI 5.4-42.9). The "atypical" antidepressant mianserinproduced significantly less dropout than placebo. Only 2 studies used numerical scales designed to measure effects on function and quality of life; in HIV (Karnofsky scale), drug was better than no treatment; in lung disease (Sickness Impact Profile), drug was not significantly different from placebo. Only 7 studies reported looking for changes in the physical disease. Antidepressants produced no change in immune function in HIV relative to placebo (2 studies) or no treatment (1 study). Relative to placebo, antidepressants produced no change in cardiovascular function in heart disease, in respiratory function in lung disease, or in vital signs or laboratory tests in cancer (1 study each). Nortriptyline produced worse control in diabetes. Trends towards tricyclics being more effective than SSRIs, but also more likely to produce dropout were noted, but these are based on non-randomised comparisons between trials.

The review provides evidence that antidepressants, significantly more frequently than either placebo or no treament, cause improvement in depression in patients with a wide range of physical diseases. About 4 patients would need to be treated with antidepressants to produce one recovery from depression which would not have occurred had they been given placebo (NNT 4.2, 95% CI 3.2-6.4). Antidepressants seem reasonably acceptable to patients, in that about 10 patients would need to be treated with antidepressants to produce one dropout from treatment which would not have occurred had they been given placebo (NNH 9.8, 95% CI 5.4-42.9). The evidence is consistent across the trials, apart from 2 trials in cancer, where the "atypical" antidepressant mianserin produced significantly less dropout than placebo. Trends towards tricyclics being more effective than SSRIs, but also more likely to produce dropout were noted, but these are based on non-randomised comparisons between trials. Problems with the evidence include most of the trials' use of observers, rather than patients, to decide on improvement, and concentration mainly on symptoms rather than function and quality of life. There is also a possibility of undetected negative trials. Nevertheless, the review provides evidence that use of antidepressants should at least be considered in those with both physical illness and depression. Regarding diagnosis, the existence of a cheap and readily available treatment for depression should encourage detailed assessment of persistent low mood in the physically ill.