This exploratory double-blind, randomised, 20-week study evaluated the mechanism of action of metformin-glibenclamide combination tablets (Glucovance) vs. metformin and glibenclamide in 50 type 2 diabetes patients inadequately controlled by diet and exercise. A glycaemic target of HbA1C 7.0% was used. Final HbA(1C), fasting glucose and post-oral glucose tolerance test (OGTT) glucose were similar between groups, although average doses of metformin and glibenclamide from combination tablets (708 and 3.5 mg) were lower than monotherapy doses (1500 and 6.6 mg). Second-phase insulin during a hyperglycaemic clamp increased by 93% with combination tablets, 36% with metformin and 46% with glibenclamide. The insulin response post-OGTT was more rapid with the combination tablets vs. glibenclamide. First-phase insulin responses improved modestly in all groups, possibly due to reduced glucotoxicity. Changes in insulin sensitivity were minor. Larger beta-cell responses between combination tablets and glibenclamide may reflect more rapid glibenclamide absorption.