METHODS

Patients with severe adult or juvenile diabetes, without renal or hepatic disease, served as subjects. Six studies were performed in five patients on a metabolic ward. They received diets kept at constant caloric levels designed to maintain body weight. Insulin dosages were also kept constant whenever possible. Two diets were compared:(1) a basal formula, containing 15 per cent protein calories, 40 per cent fat calories as corn oil, and 45 per cent carbohydrate calories, fed as a liquid diet;(2) a low-fat, high-carbohydrate diet, containing 15 per cent protein calories, 85 per cent carbohydrate calories, and virtually no fat, fed as a liquid diet with rice and matzoh (unleavened bread) supplements, for periods ranging from two to ten weeks. The subjects received supplementary multivitamins, iodized salt, and ferrous gluconate.

Samples of venous blood were obtained after an overnight fast prior to the administration of insulin. Glucose was measured in filtrates of whole blood17 using the enzymatic reagent Glucostat (glucose oxidase, horseradish peroxidases, phosphate buffer, and o-dianisidine; Worthington Biochemical Corp., Freehold, New Jersey), as described by Saifer and Gerstenfeld. 18 Plasma nonesterified fatty acids were determined by the single extraction method of Dole. 18 Measurements of total cholesterol20 and lipid phosphorus21 were made on aliquots of serum.