The effect of dietary fructose on glycemic control in subjects with diabetes mellitus is controversial. Therefore our aim was to conduct a long-term study to examine the effects of dietary fructose on glucose tolerance and insulin sensitivity and to delineate the mechanisms for the effects observed. Six subjects with non-insulin-dependent diabetes mellitus (NIDDM) who were being treated by diet alone consumed 13% of their calories as fructose incorporated into mixed meals in place of sucrose for 3 months as inpatients on a metabolic ward. The following parameters were measured: (1) weekly fasting plasma-glucose concentrations, (2) postprandial serum glucose and insulin levels after four sugar tolerance tests, (3) basal hepatic glucose production, and (4) hepatic and whole-body insulin sensitivity determined during a hyperinsulinemic, euglycemic clamp. When modest amounts of fructose were substituted for sucrose in the diet for 3 months, basal hepatic glucose output remained unchanged (12.84 +/- 1.83 nmol/kg/min v 12.51 +/- 2.00 nmol/kg/min) as did hepatic insulin sensitivity (92% +/- 4% v 93% +/- 4% suppression) and peripheral glucose disposal (22.52 +/- 4.56 nmol/kg/min v 25.80 +/- 9.45 nmol/kg/min) to a 860 pmol/m2/min insulin infusion at euglycemia (4.8 mmol/L). Fructose feeding also did not alter fasting plasma-glucose concentrations or postprandial plasma glucose and insulin responses to oral glucose or fructose loads or to mixed meals containing either sucrose or fructose. In conclusion, substitution of physiologic amounts of sucrose by fructose for prolonged periods is unlikely to have adverse effects on glucose metabolism in diabetic subjects who are being treated with diet alone.