That some diabetics improve on high carbohydrate diets is an established fact (1, 2). The selection of suitable patients for this type of diet has, until the present, been one of trial and error. Recently MacBryde (3) has made an attempt to select cases on the basis of insulin sensitivity. He studied the response of a small group of diabetics to a standard test dose of insulin and concluded that they fell into two groups, the relatively insulin-resistant and the relatively insulinsensitive. The resistant group gained tolerance on high carbohydrate diets while the sensitive group did not. In addition, the patients showing insulin-resistance were usually older, frequently obese, often had vascular hypertension and showed little tendency to acidosis. The relatively sensitive group were usually younger, often thin, had as a rule low blood pressure and were more prone to develop acidosis and coma. Although the insulin requirement of the resistant group was larger, the sensitive group were looked upon as more serious, judged by their tendency to acidosis. On the basis of a somewhat different test, Himsworth (4) classified diabetics in a similar manner, but concluded from his experimental findings and clinical data that insulin-sensitive patients tolerated high carbohydrate diets better than did the insulin-insensitive.Since these two investigators appear to have come to diametrically opposite conclusions in regard to the relationship between insulin sensitivity and response to high carbohydrate diets, it was felt worth while to study this problem further. The purpose of the present investigation has been to study a relatively large group of diabetics with respect to their blood sugar response to a standard test of insulin, and to correlate, if possible, insulin sensitivity with their clinical characteristics and responses to diets of variable carbohydrate content.
METHOD OF STUDYFifty of a total of 197 patients attending the adult Diabetic Clinic of the Strong Memorial Hospital were chosen for this study. Each had previously had a complete physical examination, blood count, urinalysis, and Wassermann reaction. They represented a fair crosssection of the total clinic population and were arbitrarily selected from amongst those who had attended the clinic for at least five months and who had been most cooperative and regular in their attendance. None was suffering from infection at the time of the studies. Cases in which the diagnosis of diabetes mellitus was at all questionable were subjected to a sugar tolerance test and were accepted only if they had typical diabetic responses (Cases 7, 25, 40, 45).Each patient was subjected to an "insulin tolerance test" and classified as relatively insulin-sensitive or relatively insulin-resistant. His past record in our clinic and on any admission to the hospital was then studied and analyzed. These studies constitute the basis for this report. The patients had previously been followed from 5 to 123 months, an average of 42 months each, and had usually been seen at monthly interval...