The pharmacokinetics and dose-exposure relationship of an extended-release formulation of metformin (ER-metformin) was investigated in a randomized, single-dose, four-period crossover study in 24 healthy male volunteers. During each study period, subjects received a randomly assigned dose containing 1000, 1500, 2000 or 2500 mg metformin. Blood samples were drawn 0-72h after dosing for pharmacokinetic and dose-proportionality assessment. Although several pairwise comparisons between dose groups were significant (p < 0.05) with respect to dose-normalized C-max, AUC(0-72h), and AUC(infinity), the magnitude of the difference across the dose range was <20% for AUC(0-72h) and AUC(infinity), and was less than or equal to30% for C-max. The results indicate a consistent and predictable increase in metformin exposure with an extended-release formulation of metformin over 1000 to 2500 mg. Copyright (C) 2004 John Wiley Sons, Ltd.