In a prospective, single blind study, 51 adult diabetic, male outpatients at or below ideal body weight (IBW), all but four of whom were insulin-treated, were randomly assigned to a calorically defined exchange diet (EXCH) or an unmeasured diet avoiding refined sugars (UNMEAS). Fasting chemistries, weights and 48-hr dietary recalls were obtained every 3 months for 3 years. There was no difference between groups in mean body weight, mean caloric intake, percentages of carbohydrate or fat intake, day to day consistency of caloric intake, or fasting chemistries. The patients whose actual weight tended to be less than IBW consumed more calories than their theoretical calculated needs (p = 0.002) without effect on glycemic control or stability of weight. Patients with low carbohydrate intake (less than 30%) had significantly higher triglyceride levels (p = 0.015). Otherwise, variations in dietary patterns were not reflected in fasting clinical chemistries. Individual fasting serum glucose levels were not related to consistency of carbohydrate intake, caloric distribution throughout the day or composition of diets. These results suggest that actual dietary patterns of subjects following exchange diabetic diets with precise caloric prescriptions are similar to those resulting from unmeasured diets with restriction of simple sugars. Other than correlation between carbohydrate intake and triglycerides, there was no detectable effect of variations in dietary patterns upon conventional chemical metabolic levels.