Objectives  To determine predictors of glycemic response to exenatide and to assess change in glycosylated hemoglobin (A1C) and whether a correlation exists between weight loss and glycemic response.

Design  Retrospective observational cohort study.

Setting  United States in 2009.

Patients  100 adult patients with type 2 diabetes prescribed exenatide.

Intervention  Retrospective chart review of patients to collect demographic data, weight, serum creatinine, diabetes education, and concurrent diabetes medications.

Main outcome measures  Patients were categorized as responders or nonresponders based on change in A1C. Responders had an A1C decrease of 0.5% or more and nonresponders had an A1C decrease of less than 0.5% from baseline to post–exenatide initiation. Demographic data for each cohort were analyzed.

Results  100 patients met inclusion criteria (61 responders and 39 nonresponders). Responders had a mean A1C decrease of 1.57%, whereas nonresponders had a mean A1C increase of 0.23% (P < 0.001). Post hoc linear regression analysis revealed that baseline A1C was a predictor of response to exenatide (P < 0.001). Binary logistic regression analysis demonstrated that no other variables were predictors of response to exenatide (P > 0.05 for all). No correlation was found between weight loss and exenatide and glycemic response (P = 0.99).

Conclusion  Our data indicate that patients with a higher baseline A1C are more likely to have a glycemic response to exenatide than patients with a lower baseline A1C.