OBJECTIVES: To compare the risk of adverse cardiovascular outcomes among elderly patients with type 2 diabetes mellitus (T2DM) treated with SGLT2 inhibitors or incretin-based therapies (IBT) (glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors) using 2012-2016 5% national Medicare data.

METHODS: A retrospective cohort study identified elderly T2DM patients initiating SGLT2 inhibitor therapy or IBT between January 2013 and June 2016. Index date was the first date of a SGLT2 inhibitor or any IBT prescription. Patients initiating SGLT2 inhibitor therapy were matched to patients initiating an IBT on baseline demographic and clinical characteristics, and baseline medication use using 1:2 propensity score matching with a greedy match algorithm. Primary outcome was a composite of stroke, myocardial infarction (MI), and all-cause mortality. Other cardiovascular outcomes evaluated included heart failure (HF)-related hospitalization and emergency room (ER) visits. Stratified Cox proportional hazards models were used to estimate the risk of cardiovascular outcomes associated with the two therapies in the matched cohort.

RESULTS: The final study cohort consisted of 5,061 beneficiaries – 1,687 SGLT2 inhibitor users and 3,374 IBT users. The risk of composite outcome was significantly lower among SGLT2 inhibitor therapy users compared to IBT users (Hazard ratio (HR):0.778, p=0.001, 95% confidence interval (CI):0.671–0.904). When individual outcomes were considered SGLT2 inhibitor users had a lower risk of stroke (HR:0.831, p=0.028, 95%CI:0.704–0.981) and MI (HR:0.704, p=0.024, 95%CI:0.519–0.955) compared to IBT users. However, the difference in the risk of all-cause mortality was non-significant between two groups (HR:1.605, p=0.433, 95%CI:0.492–5.238). Among patients with co-morbid HF, the risk of HF-related ER visits was significantly lower for SGLT2 inhibitor therapy users (HR:0.629, p<0.001, 95%CI:0.507–0.779) compared to IBT users, but the association was not significant for HF-related hospitalizations (HR:0.770, p=0.109, 95%CI:0.560–1.060).

CONCLUSIONS: SGLT2 inhibitor therapy use was associated with a lower risk of stroke, MI, and HF-related ER visits compared to IBT use.