In a TriNetX retrospective cohort of 249,237 propensity-matched pairs with type 2 diabetes who started therapy between 2018 and 2023, SGLT-2 inhibitor initiation was linked to a lower kidney failure/death composite than GLP-1 receptor agonist initiation (HR 0.87, 95% CI 0.85-0.89). Kidney failure, dialysis or transplantation fell most (HR 0.74), all-cause mortality was modestly lower (HR 0.90), and the composite cardiovascular outcome was essentially equal (HR 0.98).
The kidney signal fits what randomized trials like EMPA-REG, CANVAS, DAPA-CKD and CREDENCE showed for SGLT-2 inhibitors: reduced intraglomerular pressure and slower eGFR decline. The null cardiovascular result also matches expectations, since both classes reduce major adverse cardiac events, and GLP-1 agonists carry strong evidence for atherosclerotic protection and weight loss. The practical takeaway is complementarity rather than a winner: patients with chronic kidney disease or albuminuria have a strong case for an SGLT-2 inhibitor, while those with obesity or prior atherosclerotic events may gain more from a GLP-1 agonist, and many benefit from both. Limitations matter. Electronic health record data carry residual confounding despite matching, since prescribers may steer higher-risk patients toward one class, and baseline kidney function, albuminuria and BMI may be imperfectly captured. The mortality gap may partly reflect that confounding. This is confirmatory, large-scale real-world evidence that supports individualized selection but cannot replace head-to-head randomized trials.