Long-acting Injectable Antipsychotics and Risk of Treatment Failure in a Population-based Cohort

Background: Individuals with psychotic disorders are known to be at risk of hospitalization, non-adherence to treatment and involvement with the criminal justice system. Second generation long-acting injectable antipsychotics (SG-LAIAs) have been shown to reduce risk of these outcomes in observational studies and pragmatic clinical trials.

Objectives: To determine the risk of treatment failure, defined as treatment discontinuation, psychiatric hospitalization, suicide and incarceration, in patients treated with SG-LAIAs versus patients treated with other antipsychotics.

Methods: We used linked administrative health, registry and justice data from the Manitoba Centre for Health Policy (MCHP) in Manitoba, Canada. We formed a cohort of new SG-LAIA users between February 2005 and March 2020 who had no previous SG-LAIA dispensation in a 365-day look back period. SG-LAIA users were categorized as incident new user (no prior-year antipsychotic dispensation) or prevalent new user (prior-year oral or first-generation long-acting injectable antipsychotic dispensation). We created an exposure set of eligible antipsychotic user comparators for each SG-LAIA new user, matched on calendar time, prior duration of antipsychotic use, and number of antipsychotics used in the look back period. Each SG-LAIA user was matched 1:1 with the comparator from the exposure set with the closest time-conditional propensity score. We used Cox proportional hazards regression adjusted for time-varying antipsychotic polypharmacy during follow up to estimate risk of treatment failure in SG-LAIA users versus matched comparators. Secondary outcomes were treatment discontinuation, psychiatric hospitalization, and incarceration.

Results: The cohort consisted of 343 matched pairs, of which 156 pairs were prevalent new users and 187 were incident new users. Overall, SG-LAIA use was associated with lower risk of treatment failure versus matched antipsychotic comparators (adjusted hazard ratio 0.68, 95% confidence interval 0.59-0.77). SG-LAIAs reduced risk of incarceration (aHR 0.67, 95% CI 0.50-0.90) and treatment discontinuation (aHR 0.62, 95% CI 0.53-0.72) but not psychiatric hospitalization (aHR 1.07, 95% CI 0.89-1.29).

Conclusions: SG-LAIAs reduced the risk of treatment failure, incarceration, and treatment discontinuation by approximately 30% versus matched antipsychotics but had no significant impact on psychiatric hospitalization.
Acknowledgements: Results and conclusions are those of the authors; no official endorsement by Manitoba Health or Manitoba Centre for Health Policy is intended or should be inferred.