(B-091) Utilization of Type 2 Diabetes Mellitus Products in the Sentinel Distributed Database, 2008-2024

Background: Postmarket safety studies of type 2 diabetes (T2D) products are planned using data from the US Food and Drug Administration’s Sentinel System, a distributed database of commercial and public insurance claims. Contemporary information on patterns of T2D product use in Sentinel is needed to determine the feasibility of complex exposure modeling for safety analyses.

Objectives: To characterize the population of T2D product initiators and duration of T2D product use from 2008-2024.

Methods: We identified initiators of 4 classes of T2D products (dipeptidyl peptidase 4 inhibitors [DPP-4i], sodium glucose co-transporter 2 inhibitors [SGLT-2i], glucagon-like peptide-1 receptor agonists [GLP-1ra], and glucose-dependent insulinotropic polypeptide [GIP]/GLP-1ra) with continuous enrollment and no drug class dispensings in the year before initiation (baseline period). All information was assessed by drug class. We collected information on demographics and baseline health characteristics, total number of exposure episodes, and episode duration from claims. We characterized cumulative exposure duration in all available data after initiation using dispensed days’ supply; overlapping days’ supplies were shifted to be non-overlapping.

Results: We identified 2,921,067 DPP-4i, 2,842,393 SGLT-2i, 3,321,029 GLP-1ra, and 374,607 GIP/GLP-1ra initiators during the study period. Initiations of SGLT-2i, GLP-1ra, and GIP/GLP-1ra increased annually following approval, while DPP-4i initiations peaked in 2017. Mean age was between 58.3 (GIP/GLP-1ra) and 65.3 (DPP-4i) years. Percent female was between 46.5% (SGLT-2i) and 62.0% (GIP/GLP-1ra). The majority of each drug class were White and non-Hispanic; race and ethnicity data were missing for 22.9-31.8%.
Over 72% of initiators in each drug class had ≥1 T2D and no type 1 diabetes diagnosis codes during baseline. The most frequent comorbidities were hypertension (>74%), hyperlipidemia (>73%), and chronic kidney disease (>44%). Obesity prevalence ranged from 29.2% (DPP-4i) to 68.2% (GIP/GLP-1ra), and 19.0-24.7% of initiators had evidence of smoking during baseline.
We observed an average of 2.0 (GIP/GLP-1ra) to 5.6 (DPP-4i) episodes per initiator. Mean cumulative exposure duration varied between 150.8 (GLP-1ra) and 611.7 (DPP-4i) days. Only 54.8% of GIP/GLP-1ra initiators attained >90 days of cumulative exposure in all available data. The proportion of non-GIP/GLP-1ra initiators attaining 1 and 2 years exposure varied from 34-47% and 17-29%, respectively. The most frequent duration for initial exposure episodes was ≤30 days.

Conclusions: Our data indicated that discontinuation or gaps in T2D product use is common, and our ability to study dose-response and conduct as-treated analyses may be limited.