Who is missing from SGLT-2 inhibitor trials? Implications for drug safety and generalizability: a systematic review and meta-analysis

SGLT-2 inhibitors improve type 2 diabetes and heart failure outcomes, yet representation of key demographic and clinical subgroups remains unclear. We aimed to quantify their representation in phase 3/4 trials. We searched CENTRAL, MEDLINE, and Embase (inception–5 May 2025) for trials (≥300 participants) reporting subgroup enrollment. We estimated pooled prevalence using random-effects meta-analyses and calculated exploratory Participation-to-Prevalence Ratios (PPRs) against real-world U.S. disease burden. Across 91 RCTs, type 2 diabetes prevalence was: females 43.8% (95% CI 42.0%–45.7%), eGFR < 60 32.5% (13.1%–55.6%), older adults 31.8% (25.7%–38.2%), and Black participants 4.9% (4.3%–5.6%). In heart failure, prevalence was: older adults 68.7% (59.9%–76.8%), females 34.1% (27.8%–40.7%), and Black participants 7.6% (5.7%–9.7%). Exploratory PPRs indicated underrepresentation of Black participants (0.3 diabetes; 0.5 heart failure), older adults in diabetes (0.7), and females in heart failure (0.7). Data for liver disease and children/adolescents were insufficient. Prevalence was low for Black participants in both populations, older adults in diabetes trials, and females in heart failure trials. These disparities, alongside the exclusion of children and adolescents, may limit safety data generalizability, underscoring the need for inclusive recruitment. www.crd.york.ac.uk/prospero identifier is CRD42024561154.