Predictors of hospitalisation and recovery following full antipsychotic discontinuation in first episode psychosis. A naturalistic retrospective cohort study.

Whilst antipsychotic medication reduces risk of relapse following a first episode of psychosis (FEP), some individuals can discontinue medication and remain relapse free. We aimed to identify patient and service-specific factors which influence clinical outcome following antipsychotic discontinuation. The outcomes 'admission to hospital' and 'remaining free from psychotic symptoms', both within one year from discontinuation, were explored retrospectively in an established naturalistic cohort of 354 patients with FEP. Logistic regression analysis was used to explore influence of routinely available baseline and treatment course variables on these outcomes. Seventy-seven individuals (22 %) fully discontinued antipsychotic treatment within a year, at mean 102 days from initiation. Only antipsychotic type had significant association with discontinuation; aripiprazole was discontinued more than olanzapine (p = 0.028). Seventeen individuals required admission to hospital; significantly associated with prior admission at first illness onset (p = 0.004), and prior legal detention to hospital (p = 0.001). Admission was less likely in those discontinuing aripiprazole vs olanzapine (p = 0.044). Twenty-four patients remained psychosis symptom free and were most significantly likely to have received clinician support in discontinuation; this group had no association with either initial duration of untreated psychosis or prior duration of antipsychotic treatment. Future studies exploring outcomes following antipsychotic discontinuation require consistency of choice of outcome measures and sample stratification by vulnerability factors including severity of first illness episode, whether remaining symptom free after first episode, which medication switched from and baseline functioning. The impact and nature of clinician support to discontinue requires further exploration alongside its association with abruptness of discontinuation.