Concordance Between Trial Efficacy and Real-World Effectiveness of Sodium-Glucose Cotransporter 2 Inhibitors in Heart Failure With Reduced Ejection Fraction: A Systematic Review and Meta-Analysis

Randomized trials established the efficacy of sodium-glucose cotransporter 2 (SGLT2) inhibitors in heart failure with reduced ejection fraction (HFrEF). We assessed whether their effectiveness in routine practice is concordant with trial efficacy, appraising and pooling each evidence stream separately. PubMed/MEDLINE, Cochrane Central Register of Controlled Trials (CENTRAL), and ClinicalTrials.gov were searched from January 1, 2015 to July 2026 for randomized trials and comparative observational studies of SGLT2 inhibitors vs. placebo or nonuse in adults with left ventricular ejection fraction ≤40% reporting mortality or heart failure hospitalization. Risk of bias was assessed with the Cochrane Risk of Bias 2 and the Newcastle-Ottawa Scale. Adjusted all-cause mortality was pooled by random-effects meta-analysis with the Hartung-Knapp adjustment; other outcomes were synthesized narratively. Trial and observational estimates were never combined, and certainty was graded with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria. Eleven studies were included (two trials and nine observational). For all-cause mortality, the pooled observational HR (0.75, 95% CI 0.71-0.79; I² = 0%) indicated a larger apparent benefit than the pooled trial estimate (HR 0.87, 0.77-0.98). In one registry, the association attenuated to borderline nonsignificance after adjustment for background therapy. For heart failure hospitalization, a consistent trial benefit (HR 0.69, 0.62-0.78) contrasted with attenuated or null real-world estimates (HR 0.84-1.03). Real-world mortality associations with SGLT2 inhibitors in HFrEF are directionally concordant with trial efficacy, but their larger magnitude most plausibly reflects residual confounding, including co-prescription of other guideline-directed therapy. Real-world hospitalization estimates diverged from trials, likely reflecting outcome ascertainment. With very low-certainty observational evidence, the trial estimate remains the best guide to causal benefit.