Abstract
BACKGROUND: Antipsychotics are recommended for prevention of relapse in schizophrenia. It is unclear whether increased risk of relapse following antipsychotic discontinuation is predominantly associated with an absolute magnitude of dose reduction or rate of antipsychotic reduction. Establishing the responsible mechanism is important because prolonged withdrawal schedules have been suggested to reduce risk of relapse. STUDY DESIGN: Individual patient data from antipsychotic discontinuation studies were obtained. We estimated the occupancy of receptors over time using half-lives and median effective dose ED50 values obtained from pharmacokinetic and receptor occupancy studies. Hazard ratios for relapse events were calculated using Cox proportional hazards models to assess the influence of formulation (oral, 1-monthly, and 3-monthly injections). The change in hazard ratio over time was estimated, and the effect of time-varying covariates was calculated, including rate of occupancy reduction and absolute receptor occupancy. STUDY RESULTS: Five studies including 1388 participants with schizophrenia were identified (k = 2: oral, k = 2: 1-monthly injection, k = 1: 3-monthly injection). Withdrawal of long-acting injectable medication did not lead to a lower hazard ratio compared with withdrawal of oral medication, and this included the period immediately following randomization. Hazard ratios were not associated with the rate of decline of receptor occupancy; however, they were associated with reduced absolute occupancy in trials of long-acting injections (P = .038). CONCLUSIONS: Antipsychotic discontinuation is associated with an increased risk of psychotic relapse, related to receptor occupancy. Although relapse does not appear to be related to the rate of discontinuation, gradual discontinuation strategies may allow for easier antipsychotic reinstatement in case of symptomatic worsening.