Discontinuation of antipsychotics predisposes patients with remitted/stable first-episode psychosis (FEP) to a higher risk of relapse, but it remains unclear how long discontinuation increases the relapse rate in these patients compared with maintenance.

This meta-analysis of randomized controlled trials (RCTs) compared relapse rates in FEP patients between antipsychotic treatment discontinuation and maintenance groups at 1, 2, 3, 6, 9, 12 (primary), and 18–24 months. The risk ratio (RR) and numbers needed to treat/harm (NNT/NNH) were calculated using a random-effects model.

Ten RCTs were identified (n = 776; mean study duration, 18.6 ± 6.0 months). The antipsychotics were discontinued abruptly in four RCTs (which reported data only at 12 months) and after tapering off gradually over several months (mean length, 3 months) in six RCTs. Compared with the discontinuation group, the maintenance group experienced significantly fewer relapses at all time points except 1 month [RR (NNT): 2 months, 0.49 (13); 3 months, 0.46 (9); 6 months, 0.55 (6); 9 months, 0.48 (3); 12 months, 0.47 (3); and 18–24 months, 0.57 (4)]. The maintenance group was associated with higher discontinuation due to adverse events (RR, 2.61; NNH, not significant).

Maintaining antipsychotic treatment prevented relapse for up to 24 months in FEP patients. Discontinuation of antipsychotics for ⩾2 months significantly increased the risk of relapse. However, 45.7% of patients who discontinued antipsychotics for 12 months (39.4% after 18–24 months) did not experience a relapse.