Résumé : Method: A first meta-analysis (2283 primary care patients were followed for an average 8 weeks) showed a superior efficiency of antidepressants to placebo: relative risk of a response or a remission of 1.24 [1.11 to 1, 38] for tricyclics and 1.28 [1.15 to 1.43] for selective serotonin reuptake inhibitors. A second meta-analysis (718 patients in outpatient psychiatric) showed that efficiency increased with the initial severity, and did similarly in both classes. The efficacy endpoint selected by the National Institute for Clinical Excellence was reached for “very severe” depression. The other two randomized clinical trials which were undertaken showed no significant difference vs placebo.
Discussion: Antidepressants are only effective in general practice beyond a threshold of “very severe” depression, which still increases if we include unpublished data, but also in terms of age, patient's choice and monitoring treatment accompanying the drug. The benefit-risk ratio must take into account the adverse effects (the risk of suicide increases in young but has no attested impact in adults). The recommendations to continue these treatments 16 to 20 weeks after consolidation are based on only one extended trial.
Conclusion: Studies of long duration and without conflicts of interest should be systematically conducted and published regardless of their results.