Characteristics of meta-analyses of multi-indication drugs: a systematic review of systematic reviews

Multi-indication drugs (MIDs) are increasingly common, but their meta-analyses face methodological challenges due to heterogeneity across indications. This review aims to characterize the methodological and reporting approaches of meta-analyses evaluating MIDs. We conducted a systematic review following PRISMA guidelines, searching Ovid-Medline, Embase, and Cochrane Library for articles published from 2000 to 2025. We included meta-analyses of randomized controlled trials, with or without additional observational studies, that evaluated efficacy and/or safety of a drug across at least two distinct indications. Data were extracted and synthesized descriptively. The methodological quality of included reviews was assessed using AMSTAR 2, and key methodological practices were compared across strata defined by methodological quality, publication period, and journal quartile using Fisher’s exact tests. Of the 76 included studies, nearly half (48.7%) were published in the last five years. Most reviews (78.9%) pooled outcome data across indications, but only 32.9% performed subgroup analysis by indication, indicating substantial variability in how between-indication differences were addressed. Regarding outcomes, 44.7% focused exclusively on safety, 25.0% on efficacy, and 30.3% evaluated both. Random-effects models were predominant (51.3%), while exclusive use of fixed-effect models was less common (9.2%). Common drug classes included glucagon-like peptide-1 receptor agonists and/or sodium-glucose cotransporter-2 inhibitors (10.5%) and nonsteroidal anti-inflammatory drugs (9.2%). Only 26.3% applied formal evidence grading systems. Methodological quality was frequently limited: 76.3% of reviews were rated low or critically low on AMSTAR 2, with critical weaknesses concentrated in the assessment of publication bias (60.5% not met), protocol registration (31.6% not met), and consideration of risk of bias in interpretation (28.9% not met). Meta-analyses of MIDs are increasing but exhibit methodological variability and frequently limited methodological quality, particularly in exploring between-indication heterogeneity and assessing evidence certainty. Future studies should adopt more standardized methods to improve evidence synthesis for clinical and policy decision-making.