This is a retrospective analysis of medical records in 4 Belgian diabetes centres of 3 cohorts of patients with type 2 diabetes, with data available, respectively, after 3 months < or =163 patients exposed), 6 months (n=77) and 9 months (n=28) with exenatide therapy. This analysis mainly focuses on the 3 and 6 month cohorts. The mean HbA1 level at baseline averaged 9% and decreased by -1.3% and -1.4% at 3 and 6 months, respectively (-1.5% at 9 months). Neither the duration of diabetes nor initial body weight did influence the metabolic response. The decrease in HbA(1c) at 6 months was greater in patients with higher baseline HbA(1c):-0.5%, -1.4% and -2.4% for a baseline HbA(1c) level <8%, 8-10% and >10%, respectively. At 6 months, the composite criterion of a reduction of HbA(1c) by >1% or a final level <7% was reached by 69% of the cohort. Body weight decreased continuously over time, with a mean reduction of -2.1 kg at 3 months and -3.0 kg at 6 months (-4.9 kg at 9 months). The greater the baseline body weight, the greater the weight loss at final evaluation. Minor nausea and more rarely vomiting were observed, essentially during the first months of exenatide treatment. According to this observational study in routine practice, exenatide may be a valuable alternative to insulin for intensification of treatment of patients with type 2 diabetes after failure of oral drug combination, independently of baseline HbA(1c), body weight and duration of diabetes.