To compare the impact of dietary fish oil supplementation (FO, 22 ml daily, containing 4.6 g of n-3 (omega-3) fatty acids, equalling 14.4 mmol) on carbohydrate and lipid metabolism with that of conventional lipostatic therapy (Gemfibrozil (G), 900 mg daily, equalling 3.6 mmol) on hyperlipidemic non-insulin dependent diabetes mellitus (NIDDM), 10 patients were selected for a randomized, short-time, cross-over study. Each patient was treated for a duration of 2 weeks, with an individual washout period of 8 weeks. Metabolic variables and intravenous glucose tolerances (1.2 mmol/kg body weight, t = 30 min) were determined on days 1 and 15 of each treatment period. Plasma lipid concentrations were identical at baseline, but were reduced more markedly following G as against FO exposure (% change vs. baseline: total cholesterol (chol), - 13**/-6* (G vs FO: p = 0.05); total triglycerides (TG), -39**/-18** (p < 0.05); APO B, -17**/- 10* (N.S.); LDL-chol, -15**/0 (p < 0.02); VLDL-chol, -50***/- 34*** (N.S.); VLDL-TG, -44***/-27** (N.S.); (p vs. baseline: * < 0.05, ** < 0.01, *** < 0.001). Total-HDL, HDL2, HDL3 and APO A were not influenced by either FO or G. Neither FO nor G induced a change in intravenous glucose tolerances or associated basal and incremental concentrations of insulin and C-peptide. We concluded, based on short-time applications, that (a) neither treatment affected the carbohydrate metabolism in patients with NIDDM, and (b) a greater hypolipidemic efficacy had to be assigned to Gemfibrozil than to fish oil. It would therefore appear that Gemfibrozil acts as a useful lipostatic pharmacologic compound, whilst fish oil could serve as a potential ingredient of a prudent cardio-protective diet which favours the low plasma triglyceride concentrations found in NIDDM patients.