Patient characteristics, drug adherence patterns, and hypoglycemia costs for patients with type 2 diabetes mellitus newly initiated on exenatide or insulin glargine

Abstract

Objective: Examine real-world effectiveness and hypoglycemia cost burden in patients with type 2 diabetes newly initiated on exenatide or insulin glargine.

Design and methods: Retrospective cohort study describing patient characteristics, drug adherence patterns, and 1-year hypoglycemia rates with associated costs using an administrative claims database. Adult subjects with type 2 diabetes had an initial claim for exenatide or insulin glargine between May 1, 2005 and June 30, 2007, and had continuous eligibility for >or= 6 months pre- and >or= 12 months post-initiation.

Results: The exenatide cohort (n = 3262) was 53 +/- 10 years (+/-SD); 54% female. The insulin glargine cohort (n = 3038) was 56 +/- 12 years; 41% female. The mean Deyo-Charlson comorbidity index score was 1.45 for exenatide versus 1.82 for insulin glargine (p < 0.001). Baseline OAD use rates for exenatide and insulin glargine, respectively, were 77% versus 69% metformin; 47% versus 65% sulfonylurea; 50% versus 49% thiazolidinedione; 56% versus 60% multiple OAD. For patients with two or more pharmacy claims for exenatide or insulin glargine, the 12-month medication possession ratio (MPR) was 68 +/- 29% for exenatide and 58 +/- 28% for insulin glargine (p < 0.001). MPR >or= 80% was higher for exenatide (p < 0.001) and fewer patients discontinued therapy (p < 0.001). The probability of a hypoglycemic event was significantly lower for exenatide (p < 0.005), resulting in lower associated annual costs.

Conclusions: This study provides the first real-world observational comparison of type 2 diabetes patients newly initiated on exenatide or insulin glargine. Exenatide patients had a lower comorbidity burden, better drug adherence, and a lower rate of hypoglycemic events with associated costs. Retrospective database analyses examine medical care utilization in large populations using a relatively inexpensive and expedient approach. However, data are only representative of a commercial health-care plan with limited information on multiple variables usually collected during clinical trials.