BACKGROUND: A trial emulation approach evaluated kidney outcomes for addition of three diabetes medications among Veterans, including older Veterans under-represented in trials.
METHODS: A retrospective cohort of Veterans with diabetes who were new users adding glucagon-like peptide 1 receptor agonists (GLP1RA), Sodium-glucose cotransporter-2 inhibitors (SGLT2i), or Dipeptidyl peptidase-4 inhibitors (DPP4i). Each group was followed until a composite of death or kidney event defined as a 50% decline in estimated glomerular filtration rate (eGFR), kidney failure (transplant or dialysis) or censored for non-persistence of drug class, loss of follow-up, and study end. The primary analyses used propensity score weights to compare the hazard of the composite outcome. The secondary analysis considered cause-specific hazard of kidney events with death as censoring. An Aalen Johansen plot displayed individual outcomes of kidney event or death as a competing risk. Subgroup analyses evaluated Veterans by duration of use, age, eGFR and urine protein.
RESULTS: After propensity score weighting, the cohort included 42,684 GLP1RA, 44,198 SGLT2i and 44,213 DPP4i episodes. Median age was 69 years, diabetes duration 10.6 (6.8, 15.2) years, and eGFR was 75 [58, 92] ml/min per 1.73 m2. Event rates per 1000 person-years were 34.6 (33.3, 36.0) vs 29.0 (27.8, 30.3) vs. 47.4 (46.0, 48.9) for GLP1RA, SGLT2i and DPP4i respectively. There was a protective association for composite of kidney event and death for GLP1RA, Hazard Ratio (HR) 0.70 (0.67, 0.74) and SGLT2i users [HR 0.65 (0.61, 0.68)] versus DPP4i. Results were similar for of the cause-specific hazard of kidney event alone GLP1RA [HR 0.72 (0.65, 0.80)] and SGLT2i [HR 0.48 (0.43, 0.55), and for subgroups.
CONCLUSION: GLP1RA and SGLT2i as add-on diabetes treatment was associated with lower composite and individual outcomes of kidney event and death. The association was similar across subgroups highlighting the value in primary prevention among older populations.
