Antipsychotic maintenance treatment for patients with schizophrenia has been demonstrated to be the single most important modifiable factor to prevent unplanned readmissions. Effectiveness studies have indicated different risk for drug discontinuation between current antipsychotics.

To evaluate time until discontinuation of antipsychotic treatment, and specifically to investigate if differences between the prescribed antipsychotics could be detected.

396 patients with schizophrenia were included in an open cohort study and followed through treatment until all antipsychotics prescribed at inclusion were discontinued, predictors for time to discontinuation were analysed with univariate and multivariate Cox survival analyses with time until discontinuation as the dependent variable and antipsychotic monotherapy as the predictor variable. The analysis was controlled for common confounders.

65.7% of the patients were men, mean age was 42.4 years, and 12.9% were first-episode patients. 287 were prescribed antipsychotic monotherapy. In the multivariate Cox analysis with time to all-cause discontinuation only clozapine was significant different from olanzapine, Adjusted Hazard Rate (AHR) 0.17 (0.07,0.45) (p = 0.0003), this was also the case for the prediction of time to clinician-decided discontinuation, AHR (0.20 (0.06,0.70) (p = 0.012). In the analysis with time to patient-decided discontinuation as the dependent variable also Second Generation Antipsychotic Long-Acting Injectables (LAI) (AHR 0.26 (0.09,0.77) (p = 0.015) and First Generation Antipsychotic (LAI) (AHR 0.35 (0.16,0.80) (p = 0.013) had significant lower risk for discontinuation.

Clinicians and patients discontinued clozapine at a lower rate than olanzapine, patients also discontinued LAI formulations at a lower rate than olanzapine.