Antidepressants for depression in physically ill people

Background There is an increased risk of depression in people with a physical illness. Depression is associated with reduced treatment adherence, poor prognosis, increased disability and higher mortality in many physical illnesses. Antidepressants are effective in the treatment of depression in physically healthy populations, but there is less clarity regarding their use in physically ill patients. This review updates Gill's Cochrane review (2000), which found that antidepressants were effective for depression in physical illness. Since Gill there have been a number of larger trials assessing the efficacy of antidepressants in this context. Objectives To determine the efficacy of antidepressants in the treatment of depression in patients with a physical illness. Search strategy Electronic searches of the Cochrane Depression, Anxiety and Neurosis Review Group (CCDAN) trial registers were conducted together with supplementary searches of The Cochrane Central Register of Controlled Trials (CENTRAL) and the standard bibliographic databases, MEDLINE, EMBASE and PsycINFO. Reference lists of included studies were scanned and trials registers were searched to identify additional unpublished data. Last searches were run in December 2009. Selection criteria Randomised controlled trials comparing the efficacy of antidepressants and placebo in the treatment of depression in adults with a physical illness. Depression included diagnoses of Major Depression, Adjustment Disorder and Dysthymia based on standardised criteria. Data collection and analysis The primary outcome was efficacy 6-8 weeks after randomisation. Data were also extracted at three additional time-points (4-5 weeks, 9-18 weeks, > 18 weeks). Acceptability and tolerability were assessed by comparing the number of drop-outs and adverse events. Odds ratios with 95% confidence intervals were calculated for dichotomous data (response to treatment). Standardised mean differences with 95% CI were calculated for continuous data (mean depression score). Data were pooled using a random effects model. Main results Fifty-one studies including 3603 participants were included in the review. Forty-four studies including 3372 participants contributed data towards the efficacy analyses. Pooled efficacy data for the primary outcome provided an OR of 2.33, CI 1.80-3.00,