Aim: To test the hypothesis that glycaemic control with exenatide added to thiazolidinediones (TZDs) with or without metformin was superior to placebo.
Methods: A 26‐week, multi‐country (Canada, Mexico, Romania, South Africa and the USA), randomized, double‐blind, placebo‐controlled study compared exenatide twice‐daily vs. placebo in 165 subjects suboptimally controlled with TZDs with or without metformin [HbA1c 8.2% (s.d. 0.9), fasting serum glucose 9.1 (2.6) mmol/l, body weight 93.9 (17.8) kg, diabetes duration 6.4 (4.3) years]. After a 2‐week, single‐blind, lead‐in period, subjects were randomly assigned (2 : 1) to add exenatide or placebo to current regimens. The primary endpoint was HbA1c change at endpoint (Week 26 or last‐observation‐carried‐forward).
Results: Only 8 subjects were treated with concomitant TZD alone. Exenatide reduced HbA1c significantly more than placebo [−0.84% (s.e. 0.20) vs. −0.10% (0.23), treatment difference −0.74% (0.16), p < 0.001)]. Mean reductions in body weight were similar in both treatments at endpoint [exenatide, −1.4 (s.e. 0.6) kg vs. placebo, −0.8 (0.7) kg, p = 0.176)]. Nearly 71% of subjects had both a reduction in HbA1c and body weight with exenatide compared with 54% with placebo. The most common adverse events (exenatide vs. placebo) were nausea (12% vs. 2%, p = 0.037), vomiting (8% vs. 0%, p = 0.031) and headache (4% vs. 4%). Confirmed (blood glucose <3.0 mmol/l) minor hypoglycaemia was experienced by 4 and 2% of subjects treated with exenatide and placebo, respectively. Incidence of hypoglycaemia was not significantly different between groups.
Conclusions: Exenatide added to TZDs alone or in combination with metformin significantly improved glycaemic control as determined by significant improvement in HbA1c without associated hypoglycaemia.