難治思覺失調｜一定要換 Clozapine 嗎？Lancet 統合分析

Background: Clozapine is recommended by national and international guidelines for people with treatment-resistant schizophrenia. However, available meta-analyses of randomised controlled trials have not shown superior efficacy of clozapine when compared with other second-generation antipsychotics, with heterogeneity identified between the original studies. We aimed to use individual patient data (IPD) to account for potential reasons of variability and to synthesise an adjusted estimate for the difference in efficacy between clozapine and other second-generation antipsychotics for treatment-resistant schizophrenia. Methods: In this systematic review and IPD meta-analysis, we searched the Cochrane Schizophrenia Group's Study-Based Register from inception to Jan 24, 2024, and previous reviews for blinded randomised controlled trials comparing clozapine with other second-generation antipsychotics in participants with treatment-resistant schizophrenia. Trials were eligible if they included patients with treatment-resistant schizophrenia and had a duration of at least 6 weeks. IPD were requested from trial investigators. The primary outcome was change in overall schizophrenia symptoms as measured by the Positive and Negative Syndrome Scale (PANSS) between clozapine and other second-generation antipsychotics after 6-8 weeks of treatment. The effect size measure for the primary outcome was mean difference with 95% credible interval (CrI). We fitted a Bayesian random-effects IPD meta-regression model that included duration of illness, baseline severity, and sex as potential prognostic factors or treatment effect modifiers. Confidence in the evidence was assessed using Grading of Recommendations, Assessment, Development, and Evaluation (GRADE). People with lived experience of mental illness were involved in this study. This study is registered with PROSPERO, CRD42021254986. Findings: We screened 13,876 references and included 19 studies with data for 1,599 participants; IPD were available for 12 of 19 trials (n=1,052; mean age 37.67 years [SD 11.24; range 10-66]; 348 [33.08%] women and 704 [66.92%] men). Data on ethnicity were not available. The estimated mean difference in change from baseline PANSS total score was -0.64 points (95% CrI -3.97 to 2.63; τ=2.68) in favour of other second-generation antipsychotics. Shorter duration of illness and higher baseline severity were prognostic factors associated with a larger reduction in symptoms, but neither those factors nor sex were found to modify the relative effect between clozapine and other second-generation antipsychotics. The confidence in the evidence was graded as very low. Interpretation: This IPD meta-analysis found a small and uncertain advantage of other second-generation antipsychotics, mainly olanzapine and risperidone, and so did not provide evidence for superior efficacy of clozapine compared with other second-generation antipsychotics in treatment-resistant schizophrenia. It is limited by unavailability of IPD for some studies, uncaptured sources of variance, and uncertainty due to premature study discontinuation. Given the side-effects of clozapine, the observed uncertainty regarding clozapine's superiority warrants prudent use and further research.