A prospective, stratified, randomized 3-year clinical trial was conducted on the effect of rigorous versus conventional glucose control on peripheral nerve function in 33 insulin-treated diabetic patients with a duration of diabetes of less than 2 years. The goals for conventional glucose control were the mean of fasting and 80-minute postprandial plasma glucose of 150 mg/dl for non-insulin-dependent diabetes and 200 mg/dl for insulin-dependent diabetes. The goal of rigorous glucose control was an approximation of nondiabetic glucose control. No significant difference in glucose control or peripheral nerve function was observed between the rigorously and the conventionally controlled groups. Eight patients in the conventional-control group spontaneously achieved glucose control in the range that was the objective for the rigorous-control group, and five patients in the rigorous-control group never achieved the desired glucose control. In the remaining 20 patients, with similar baseline glucose control and peripheral nerve function characteristics, observed over a median of 2 years, improved blood glucose control (P less than 0.01) was not associated with any significant improvement in peripheral nerve function. Nevertheless, a significant (P less than 0.05) correlation was found between the degree of abnormal nerve function at entry into the study and change in nerve function during the study. If control of hyperglycemia benefits peripheral nerve function of diabetic patients, its demonstration may require a closer approximation of normoglycemia, a larger difference in glucose control between the two study groups, a longer duration of treatment, and the use of patients with more advanced peripheral nerve function abnormalities than those in this study.