Poor response to antidepressants predicts new suicidal ideas and behavior in depressed outpatients.

BACKGROUND: Only a few studies have investigated the factors associated with suicidal behavior after antidepressant treatment onset in adults. We examined the specific predictors of de novo suicidal ideas or attempts among depressed patients in the community, including subjects potentially at risk of suicidal behaviors, who initiated a new antidepressant treatment.

METHODS: A large set of GPs and psychiatrists throughout France followed-up, for 6 weeks, 4357 outpatients for whom an antidepressant drug was prescribed. Dimensions related with antidepressant-induced suicidal events, such as depression, anxiety or hopelessness, were assessed longitudinally using univariate and multivariate approaches among subjects with treatment-emergent suicide ideation or attempts.

RESULTS: New suicidal ideas were observed in 9% of patients with no suicidal ideation at baseline (n=81), while suicidal attempts were reported for 1.7% of the sample during the 6-week observation period (n=75). The onset of suicidal ideas and attempts was associated with the initial features of the patients (baseline level of anxiety, past history of suicide attempts and alcohol misuse) and the non-improvement of depression. Worsening of depressive symptoms during the follow-up increased the onset of new suicidal ideas (OR=5.67, p<0.001) and attempts (OR=2.60, p=0.002), corresponding to 67.5% and 56.5% of attributable risk respectively.

CONCLUSIONS: When the analyses are restricted to the occurrence of suicidal ideas or attempts, the link between antidepressants and suicide risk might be more adequately explained by a poor response to antidepressant treatment rather than by a direct trigger-effect. This naturalistic study is limited by the use of non-structured diagnoses and self-report outcomes.