Background: There is uncertainty about the incidence of breakthrough psychosis in treatment adherent patients, and the role that factors, such as cumulative antipsychotic exposure, play in this phenomenon.
Methods: In a nationwide cohort of individuals treated for schizophrenia-spectrum disorders in Finland between 1 January 1996 and 31 December 2015, 'Breakthrough Psychosis on Antipsychotic Maintenance Medication' (BAMM) was defined as hospitalization for psychosis despite ongoing continuous treatment with long-acting injectable antipsychotics (LAIs) or oral antipsychotics (OAPs) for ⩾8 weeks. Incidence rates, survival curves, and risk factors were presented.
Results: In a cohort of 16 031 continuous LAI treatment episodes with virtually assured adherence [median duration = 441 days, interquartile range (IQR) = 155-1277], BAMM incidence was 31.5%. For 42 867 OAPs treatment episodes (median duration = 483 days, IQR = 167-1491), for whom adherence was modeled by the PRE2DUP method, BAMM incidence was 31.1%. Factors related to illness instability at treatment onset were associated with BAMM, although median time to BAMM was 291 days (IQR = 121-876) for LAIs and 344 days (IQR = 142-989) for OAPs, and 27.4% (N = 1386) of the BAMM events in the LAI, and 32.9% (N = 4378) in the OAP group occurred despite >1 year since last hospitalization at treatment onset. Cumulative antipsychotic exposure was not a consistent risk factor.
Conclusion: BAMM was relatively common even when adherence was confirmed with LAIs. Illness instability at treatment onset accounted for most cases, but relapse after years of continuous treatment was still prevalent. There was insufficient evidence to support causality between cumulative antipsychotic exposure and BAMM. Future research needs to address the role of symptom severity and neurobiology in BAMM.
Keywords: Adherence; long-acting injectable; relapse; schizophrenia.