Emulating a Randomized Clinical Trial (TECOS) Using Real-World Data to Assess the Effect of Sitagliptin on Cardiovascular Outcomes in Type 2 Diabetes

OBJECTIVES: We assessed the potential of using a large US claims database to emulate the TECOS - Trial Evaluating Cardiovascular Outcomes with Sitagliptin clinical trial in a type-2 diabetic population aged ≥50, assessing the effects of adding sitagliptin to usual care on major cardiovascular events (MACE) using second-generation sulfonylureas as proxy for placebo.

METHODS: Oracle’s US claims collects data from over 160 payers including commercial, Medicare Advantage, and Medicaid data across all states including 150million patients. Inclusion/exclusion criteria from TECOS were adapted for population selection. First prescription date of sitagliptin or 2G-sulfonylurea was considered as cohort entry. Patients were excluded if they qualified in both exposure categories. Baseline characteristics including cardiovascular and diabetes-related complications were determined at index-date while study outcomes were determined at 12 months after index-date. 1:1 Propensity score matching was performed. Standardized mean differences were estimated pre- and post-matching. Descriptive analysis and Cox proportional hazards model were performed.

RESULTS: We identified 47,859 individuals of which 13,621 were sitagliptin and 34,238 2G-sulfonylureas users. Mean age in the treated group was 63.9(8.6) and 64.1(8.7) in the reference group, of which were followed up for a mean duration of 1614 and 1566 days, respectively. 13,614 pairs of exposed and non-exposed patients were identified, who were balanced on 74 baseline covariates. During the first year of follow-up among matched patients followed for at least one year, 44.1% of sitagliptin and 44.0% 2G-sulfonylureas patients experienced a MACE event. For the matched participants, the multivariable-adjusted Hazard Ratios for MACE events related to the use of sitagliptin was 1.00(95%CI, 0.97 1.04;p=0.981).

CONCLUSIONS: This study did not identify any statistical differences in the risk of cardiovascular outcomes associated to sitagliptin compared to 2G-sulfonylureas, confirming findings from the TECOS trial. This proves the possibility of complementing drug development and regulatory approval with real-world evidence studies.