The effect of dietary fructose (20% of carbohydrate calories, 45-65 g day-1 for 4 weeks) on glycaemic control, serum lipid, lipoprotein and apoprotein A-I and A-II concentrations and on insulin sensitivity was studied in 10 type 2 diabetic patients. The study was done in a randomized, double-blind fashion with crystalline fructose or placebo administered evenly during 4 meals or snacks per day. The patients were hospitalized throughout the study periods. The fasting plasma glucose concentration decreased during the fructose (from 10.7 +/- 1.4 mmol l-1 to 8.0 +/- 0.8 mmol l-1, P < 0.02) and the control diet (from 10.1 +/- 0.9 mmol l-1 to 8.0 +/- 0.7 mmol l-1, P < 0.05). The mean diurnal blood glucose concentration also fell both during the fructose (from 10.8 +/- 0.5 mmol l-1 to 8.4 +/- 0.3 mmol l-1, P < 0.001) and the control diet (from 10.3 +/- 0.3 mmol l-1 to 8.8 +/- 0.9 mmol l-1, P < 0.01). The HbA1 concentration improved (P < 0.02) only during the fructose diet. Insulin sensitivity increased by 34% (P < 0.05) during the fructose diet, but remained unchanged during the control period. Serum insulin, triglyceride, apoprotein A-I and A-II concentrations, body weight, blood pressure and blood lactate remained unchanged during both diets. In conclusion, substitution of moderate amounts of fructose for complex carbohydrates can improve glycaemic control and insulin sensitivity in patients with type 2 diabetes.