HbA1c Outcomes in Patients Treated With Canagliflozin Versus Sitagliptin in US Health Plans

Abstract

Purpose: Clinical trial evidence supports greater glycemic control with canagliflozin than with sitagliptin. The objective of this study was to provide real-world evidence comparing outcomes in routine clinical practice among patients initiating each medication.

Methods: With the use of a health care administrative database, patients initiating canagliflozin were compared with patients initiating sitagliptin (first prescription fill as index date). Baseline (6 months before index date) demographic and clinical (eg, comorbidities and diabetes-related complications) characteristics were compared, and propensity score matching was used to control for baseline differences between cohorts. Outcomes included change in glycosylated hemoglobin (HbA_1c) and persistence with medication over a 9-month period after index date.

Findings: Before matching, the canagliflozin cohort (N = 3993) was younger than the sitagliptin cohort (N = 12,153) and was composed of fewer women and Medicare Advantage enrollees, with lower mean baseline comorbidity scores (all p < 0.001). Before matching, the canagliflozin cohort (valid n = 1482) had a significantly (p < 0.001) higher baseline HbA_1c (8.60) than the sitagliptin cohort (valid n = 3697; HbA_1c, 8.32). After matching (n = 1472 per cohort), patients were well balanced on baseline characteristics, and HbA_1c values were not significantly different (p = 0.634) between the cohorts. Patients initiating canagliflozin had greater reductions in HbA_1c than patients in the sitagliptin cohort (-0.93% versus -0.57%, respectively; p = 0.004), with similar mean (median) time from index date to follow-up HbA_1c of 185.4 (199.0) and 184.3 (190.5) days, respectively (p = 0.802). Only 29.8% of canagliflozin patients discontinued during follow-up compared with 41.5% of sitagliptin patients (p < 0.001); the average days of persistence on index therapy was longer for canagliflozin patients (152 days) than for sitagliptin patients (139 days; p < 0.001).

Implications: In this observational study, patients initiating canagliflozin had greater reduction in HbA_1c and longer persistence with medication than did patients who initiated sitagliptin, over a 9-month period. Better understanding of antihyperglycemic treatment, HbA_1c results, and differences among patients in demographic/clinical characteristics as well as persistence with treatment will inform optimal diabetes treatment choice in routine practice.

Keywords: HbA(1c); canagliflozin; diabetes; glycosylated hemoglobin; sitagliptin.