Objective: To study whether cognitive behavioural therapy decreases suicide attempts in people with previous suicide attempts. Design: Systematic review and meta-analysis. Setting: Randomised trials that compare cognitive behavioural therapy with treatment as usual. Participants: Patients who had engaged in any type of suicide attempt in the six months prior to trial entry resulting in presentation to clinical services. Main outcome measure: Suicide attempt. Results: We included ten trials, eight from Cochrane reviews and two from our updated searches (1241 patients, 219 of whom had at least one new suicide attempt). Cognitive behavioural therapy compared to treatment as usual reduced the risk of a new suicide attempt; risk ratio 0.47; 95% confidence interval 0.30– 0.73; p ¼ 0.0009; I2¼ 57%. Only seven suicides were reported (3 versus 4). One trial had an unusually large effect; if this trial is excluded, the risk ratio becomes 0.61 (0.46–0.80) and the heterogeneity in the results disappears (I2¼ 0%). Conclusions: Cognitive behavioural therapy reduces not only repeated self-harm but also repeated suicide attempts. It should be the preferred treatment for all patients with depression