We assessed the efficacy of gemfibrozil therapy on lipid profile and glucose metabolism in a large cohort of (type 2) non-insulin-dependent diabetic patients. We enrolled 217 type 2 diabetic patients with plasma triglyceride concentrations equal to or above 2 mmol/l: 110 were randomized to gemfibrozil (600 mg twice daily) and 107 to placebo treatment in a double blind fashion. Each treatment was followed for 20 weeks. To assess postprandial glucose metabolism and insulin secretion, at time 0 and 20 weeks, a standard meal containing 12.5 g of proteins, 40.1 g of carbohydrate, 10 g of lipids was given. No differences in demographic characteristics were observed between patients randomized either to gemfibrozil or to placebo therapy. No differences were observed in total cholesterol and LDL-cholesterol concentration changes between the baseline observations and week 20 of both treatments. At variance, both treatments significantly increased HDL cholesterol. Gemfibrozil treatment significantly decreased plasma triglyceride concentration from 316+/-84 to 214+/-82 mg/dl (P < 0.001), whereas with placebo triglyceride levels increased from 318 + 93 to 380 + 217 mg/dl. No changes were observed in non-esterified fatty acid concentrations or in fasting plasma glucose concentrations, in HbA(1C) values, insulin and C-peptide concentrations. Gemfibrozil treatment: 1) significantly reduces circulating triglyceride concentration; 2) does not significantly affect cholesterol concentration; 3) does not worsen glucose metabolism.