During initial examinations of the residents of Tecumseh, Michigan, one-hour 100-gm. glucose tolerance tests were applied to 2,983 persons sixteen years of age and older.
Distributions of one-hour blood glucose values divided by age, sex and interval since last food were unimodal with slight skewing toward higher levels. In general, means and variances of distributions for men and women of the same age tested under like circumstances were almost identical. In contrast, age and recent food independently exerted striking effects on one-hour levels. With increasing age through the seventh decade the whole one-hour glucosemia distribution shifted steadily toward higher levels, the slope of its regression on age approximating 13 mg. per 100 ml. per decade. Persons challenged within four hours of eating exhibited one-hour values averaging 25 mg. per 100 ml. below those of similar ages challenged at longer intervals after eating. The degree of depression of the one-hour value by recent food increased detectably with the total carbohydrate intake during the last four hours. No diurnal variation was noted; seasonal variation was observed only in males, whose values were slightly higher in autumn months than in other seasons.
The consistent observation of a linear age gradient challenges current diagnostic criteria for diabetes which take no account age. Instead, the present data suggest that much of what has been designated as maturity-onset diabetes results from a diminution of carbohydrate tolerance with age which is characteristic of the entire population. Accordingly, hypotheses regarding the etiology of diabetes must be rexamined.