Senior Research Specialist Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, United States
Background: The association between sodium-glucose cotransporter 2 (SGLT2) inhibitors and diabetic retinopathy (DR) remains largely unexplored in clinical practice, with existing evidence only based on one post hoc analysis from the EMPA-REG OUTCOME trial. Within EMPRISE, a study that aims to assess the effectiveness and safety of empagliflozin using routine clinical data, we investigated two DR outcomes among adults with type 2 diabetes (T2D) in clinical practice.
Objectives: To compare the risk of non-proliferative DR (NPDR) onset and the risk of DR progression in patients with T2D initiating empagliflozin or DPP4 inhibitors (DPP4i).
Methods: Using data from Medicare-fee-for-service and 2 U.S. commercial claims databases (08/2014-09/2019),we compared the risk of NPDR (defined as incident NPDR diagnosis) in a cohort of T2D patients without history of any DR initiating empagliflozin or DPP4i (cohort 1); and the risk of DR progression (defined as incident proliferative diabetic retinopathy or vitreous hemorrhage diagnosis, initiation of intravitreal anti-VEGF injection, or initiation of photocoagulation) among T2D initiators of empagliflozin or DPP4i with history of NPDR, but no diagnosis or treatment for advanced DR prior to cohort entry (cohort 2). For each cohort, follow-up started one day after cohort entry and continued until the first occurrence of outcome, index drug discontinuation (after 60 days grace period elapsed) or switching, insurance disenrollment, death, or end of study period. Separately in each database, propensity scores (PS) were estimated based on 143 baseline covariates and used to 1:1 PS-match initiators of empagliflozin or DPP4i. For each cohort, dataset-specific hazard ratio (HR) and rate differences (RD)/1,000 PY with their 95% CI were estimated and pooled across the three databases.
Results: In cohort 1, we identified 34,262 1:1 PS matched patient pairs without history of any DR with mean age of 66 years. In cohort 2, we identified 7,839 1:1 PSM patient pairs with mean age of 67 years. Over 52% of the population in both cohorts was male. Compared to DPP4i, empagliflozin initiation was associated with a similar risk of NPDR onset [HR: 1.05 (0.95, 1.16); RD: 1.52 (-1.61, 4.65)] (cohort 1) and with a decreased risk of DR progression [HR: 0.77 (0.62, 0.95); RD: -9.49 (-16.90, -2.08)] (cohort 2), over a mean follow-up of ~8 months on treatment.
Conclusions: In a broad population of patients with T2D as treated in routine care, empagliflozin was associated with a lower risk of DR progression, compared with DPP4i, in line with direction and magnitude of a similarly defined DR outcome in the EMPA-REG OUTCOME trial.
