Metformin efficacy and safety in type 2 diabetes has not been studied in the elderly patient under conditions pertaining to clinical practice. We have therefore, studied prospectively over two months 24 patients aged between 70-88 years with 115% mean ideal body wt and glycemic control greater than 10 mol.l-1 fasting and/or greater than 14 mol.l-1 postprandial, and/or an HbA1 value greater than 10%. At entry into the study, patients were on metformin + sulfonylurea (n = 15), sulfonylurea (n = 3), metformin (n = 2), insulin (n = 1) or diet alone (n = 3). They received metformin as the sole therapy when possible (sulfonylureas were discontinued in 9 cases) at a dosage of either 850 mg or 1,700 mg/day dependent on creatinine clearance values of 30-60 ml.min-1 (n = 11) and greater than 60 ml.min-1 (n = 13), respectively. Compared to pretreatment values, glycemic control at 1 and 2 months were unchanged, with metformin blood levels remaining within expected values for both dosage groups. Blood lactate levels remained unchanged in the high dosage group but were reduced (p less than 0.05) on 850 mg/day probably because of the reduction of metformin dosage in patients already treated prior to the study. It is concluded that provided the dosage is adjusted to renal function, the metabolic tolerance of metformin therapy is satisfactory in the elderly type 2 diabetic patient. With regard to efficacy, longer observations are needed to know whether this therapy may offer on a long-term basis equivalent glycemic control compared to other treatment modalities.