BackgroundSecond-generation antipsychotics (SGAs) are commonly used to treat schizophrenia (SCZ), but SGAs may differ in the severity of side effects. Long-term studies are lacking, and previous observational studies have limitations, such as failure to account for confounding factors and short follow-up durations. AimsTo compare the long-term anthropometric and metabolic side effects of seven SGAs in a Chinese population, using a within-subject approach to reduce the risk of confounding. MethodWe collected longitudinal data of SGA prescriptions, concomitant medications, fasting blood glucose (BG), lipid profiles, and BMI in a cohort of 767 patients with SCZ, with follow-up lasting up to 18.7 years (median ∼6.2 years). A total of 192,152 prescription records were retrieved, with 27,723 metabolic measures analysed. Linear mixed models were used to estimate the effects of SGA on BG, lipid profiles and BMI. Besides studying the effects of SGA medications (as binary predictors), we also investigated the effects of SGA dosage on metabolic profiles. ResultsConsidering SGA medications as binary predictors, clozapine and olanzapine were associated with the most substantial worsening of lipid profiles and BMI. A significant increase in BG was observed with clozapine only. Amisulpride, paliperidone and quetiapine were associated with worsened lipid profiles and increased BMI. Conversely, aripiprazole was associated with significant improvement in lipid profiles but a small increase in BMI.When SGA dosage was considered, the model showed consistent results overall. At the minimum effective dose, clozapine was associated with the most severe metabolic side effects, followed by olanzapine. Risperidone and aripiprazole showed the least metabolic side effects, with aripiprazole being significantly associated with lower lipids. ConclusionsThis study clarified the long-term and dose-dependent effects of different SGAs on anthropometric and metabolic parameters in Chinese SCZ patients. Our findings may inform clinicians and SCZ patients of SGA choices.
Read full abstract