Mortality according to glucose tolerance was studied to determine the prognosis of impaired glucose tolerance. In 1962-65, 228,833 subjects were screened for glycosuria. Of 2477 with glycosuria, 2180 were given oral glucose tolerance tests and grouped according to normal tolerance, impaired glucose tolerance, or diabetes by World Health Organization criteria. Among subjects at least 25 years old with normal tolerance, impaired glucose tolerance, or diabetes, age-sex-adjusted mortality through 1983 was 39 plus/minus 2, 49 plus/minus 4, and 71 plus/minus 4 deaths/1000 person-years (plus/minus standard error) for all causes p less than .001 for difference in 3 groups), and 24 plus/minus 2, 25 plus/minus 3, and 40 plus/minus 3 for vascular causes (cardiovascular, cerebrovascular, or renal disease) (p less than .001). 206 men with abnormal tolerance by local, but not World Health Organization, criteria were randomly assigned to diet with tolbutamide, diet only, or no treatment, which was continued through 1975. Age-adjusted all-cause mortality through 1983 did not differ significantly among treatment groups (34 plus/minus 9, 52 plus/minus 10, 45 plus/minus 19), but vascular mortality was 10 plus/minus 8, 31 plus/minus 8, and 38 plus/minus 19 in those assigned to tolbutamide, diet only, or no treatment (p less than .05). Thus compared with persons with normal tolerance, diabetic subjects had higher all-cause and vascular mortality, and those with impaired glucose tolerance had higher all-cause but similar vascular mortality. Treatment of abnormal glucose tolerance apparently reduced vascular but not total mortality.
