Efficacy and tolerability of pharmacological interventions for schizophrenia non-responsive to prior treatment: a systematic review and network meta-analysis.

Background: Treatment-resistant schizophrenia (TRS) poses significant challenges for both clinicians and patients. This systematic review and network meta-analysis (NMA) aimed to compare the efficacy and tolerability of all available pharmacotherapy options. Methods: We systematically searched MEDLINE, Cochrane Central, Embase, PsycINFO, ClinicalTrials.gov, WHO trials registry, and FDA website through March 2025 for randomised controlled trials (RCTs) comparing pharmacological treatments for TRS. NMA estimated pooled effects, with the primary outcome being overall symptom change. Secondary outcomes included treatment response, individual symptom domains, discontinuation, adverse events, quality of life, and functioning. Effect sizes were reported as standardized mean differences (SMDs) for continuous outcomes and odds ratios (ORs) for dichotomous outcomes, with 95% confidence intervals (CIs). Meta-regression and sensitivity analyses explored variability in findings. Findings: 150 RCTs with 11,375 patients examined 78 drug options or placebo. Clozapine showed superior efficacy for overall symptoms compared to haloperidol, chlorpromazine, quetiapine, and sulpiride (SMDs 0.35 to 1.00). It slightly outperformed olanzapine for positive symptoms (SMD 0.19; 95% CI 0.00 to 0.37) and risperidone for response rates (OR 0.64; 95% CI 0.41 to 1.01). Clozapine combinations with amisulpride, duloxetine, memantine, mirtazapine, topiramate, and ziprasidone improved overall symptoms more than clozapine monotherapy (SMDs -1.53 to -0.51). In a similar vein, clozapine combinations with amisulpride, lamotrigine, and topiramate reduced positive symptoms more than monotherapy (SMDs -1.13 to -0.54), while with duloxetine, memantine, and ziprasidone negative symptoms (SMDs -1.98 to -0.99). Some antipsychotic combinations may outperform monotherapy, but data on non-clozapine combinations were limited. Higher baseline severity was associated with higher clozapine efficacy. Confidence in most estimates was low or very low. Interpretation: Clozapine remains the gold standard, outperforming several antipsychotics, while specific combinations may offer added benefits but require careful risk-benefit evaluation. Networks sparsity increases the likelihood of chance findings for estimates based on single studies. These results emphasise the need for personalised treatment, further research comparing non-clozapine antipsychotic combinations to high-dose clozapine monotherapy, and studies on long-term outcomes. Funding: None.