OBJECTIVES: While sodium glucose cotransporter-2 inhibitors (SGLT2is) provide cardiovascular benefit when used in patients with type 2 diabetes (T2D) at increased cardiovascular risk, real-world effectiveness of SGLT2is as first-line treatment for T2D has not been widely studied. We compared the effectiveness of first-line treatment with an SGLT2i vs metformin in an all-comer population with T2D eligible for both therapies.

METHODS: Treatment-naïve T2D patients newly prescribed an SGLT2i or metformin as monotherapy between 2016 and 2022 were identified in the Truveta platform, which aggregates and normalizes de-identified EHR data from a group of US health care systems (HCS). We excluded patients not receiving regular care at the HCS and those with HIV, ESRD, transplant, history of gestational diabetes, missing sex, or missing age. Patients were followed until outcome occurrence, initiation of the comparator drug, loss to follow-up, or end of the study period (12/31/2022). We used 1:1 nearest neighbor propensity score matching to balance populations. The time to a composite cardiovascular outcome (hospitalization for heart failure, myocardial infarction, or ischemic stroke) was compared using a matched and adjusted Cox proportional hazards model. For the subset of patients with available baseline and 12-month A1c values, we compared change in 12-month A1c using propensity score matched and adjusted linear regression.

RESULTS: Of 135,729 patients meeting our study criteria, 12,848 remained after 1:1 propensity score matching. Cardiovascular outcomes did not differ significantly between those on an SGLT2i and metformin (hazard ratio: 1.025 [95% CI: 0.901, 1.166]). Among those with baseline and 12-month A1c values available (n = 5,472), SGLT2i use was associated with a smaller absolute decrease in A1c by 0.25% (0.19% - 0.32%).

CONCLUSIONS: In a real-world EHR dataset, patients initiated on an SGLT2i had similar risk of cardiovascular events to those initiated on metformin, but experienced a smaller 12-month reduction in A1c.