SGLT2 inhibitors and cardiorenal outcomes across operationally mapped cardiovascular-kidney-metabolic syndrome stages 2–3 and stage 4: a systematic review and meta-analysis

Abstract Background Sodium-glucose cotransporter 2 (SGLT2) inhibitors reduce heart failure and kidney events across diverse cardiometabolic populations. The cardiovascular-kidney-metabolic (CKM) framework offers a clinically relevant way to view these linked conditions, but randomized evidence has not been synthesized across this continuum. We assessed cardiorenal effects across two operationally mapped CKM stage groups. Methods We searched PubMed/MEDLINE, Web of Science, Embase, and CENTRAL to February 25, 2026, for randomized controlled trials of an SGLT2 inhibitor versus placebo or usual care. Trial populations or mutually exclusive published subgroups were categorized as CKM stages 2–3 or CKM stage 4. No single primary outcome was prespecified; five outcomes were registered as main outcomes, and all-cause mortality was registered as a secondary outcome. The six time-to-event outcomes were pooled using REML random-effects models. Prespecified sensitivity analyses examined model choice, exclusion of SOLOIST-WHF—whose selected efficacy result was judged at high risk of bias—kidney endpoint definitions, leave-one-out stability, and small-study effects. Results Nineteen reports representing 11 parent randomized controlled trials (73,220 participants) were included: 18,547 in CKM stages 2–3 and 54,673 in CKM stage 4. Composite heart failure, hospitalization for heart failure, and kidney outcomes favored SGLT2 inhibitors in stages 2–3 and stage 4: HRs 0.80 (95% CI 0.68–0.93) and 0.77 (0.73–0.81; P for group difference = 0.65); 0.61 (0.48–0.77) and 0.71 (0.66–0.76; p = 0.24); and 0.61 (0.52–0.71) and 0.64 (0.54–0.75; p = 0.64), respectively. CKM stage 4 estimates for MACE and mortality were more precise. No group-difference test was significant. Sensitivity analyses were stable; an exploratory all-cause-mortality Egger test gave p = 0.025. Conclusions Heart failure and kidney estimates favored SGLT2 inhibitors in both groups. No statistically detectable between-group difference was observed; however, CKM stages 2–3 analyses relied on four estimates and were less precise, so the findings should not be interpreted as evidence of equivalent treatment effects. Systematic review registration PROSPERO CRD420261305463.