Comparative effectiveness of a glucagon-like peptide-1 receptor agonist versus a dipeptidyl peptide-4 inhibitor on kidney and cardiovascular outcomes in routine clinical practice.

Whether glucagon-like peptide-1 receptor agonists (GLP1-RA) reduce detrimental kidney outcomes is uncertain. In secondary analyses, trials have shown consistent reductions in macroalbuminuria, but inconclusive results about kidney function decline. To help clarify this, we conducted a cohort study to compare kidney and cardiovascular outcomes in individuals who started GLP1-RA or dipeptidyl peptidase-4 inhibitors (DPP4i) (reduces degradation of endogenous GLP1). The primary outcome was a composite of sustained doubling of creatinine, kidney failure or kidney death. The secondary outcomes were three-point major adverse cardiovascular events (MACE) and its individual components. Propensity score weighted Cox regression was used to balance 53 confounders. A total of 19,766 individuals were included, of whom 5,699 initiated GLP1-RA, and were followed a median 2.9 years. Mean age was 63 years, 26.2% had atherosclerotic cardiovascular disease and 16.0% had an estimated glomerular filtration rate (eGFR) under 60 ml/min/1.73m2. The adjusted hazard ratio for GLP1-RA vs. DPP4i was 0.72 (95% confidence interval 0.53-0.98) for the composite kidney outcome and 0.85 (0.73-0.99) for MACE, with absolute five-year risk reductions of 0.8% (0.1%-1.5%) and 1.6% (0.2%-2.9%), respectively. Hazard ratios were 0.79 (0.60-1.05) for cardiovascular death, 0.86 (0.68-1.09) for myocardial infarction and 0.74 (0.59-0.93) for stroke. Results were consistent within subgroups, including age, sex, eGFR and baseline metformin use. Thus, in our analysis of patients from routine clinical practice, use of GLP1-RA was associated with a lower risk of kidney outcomes compared with DPP4i. The reductions in both kidney outcomes and MACE were similar in magnitude to those reported in large cardiovascular outcome trials.