Mortality related to diabetes and blood glucose levels in a community study.

A prospective 29-year population study of diabetes in Oxford, MA, is the basis for a case-control analysis of diabetes mortality rates and their validity. Two case groups were considered, one consisting of previously known diabetics at the study's inception in 1946 and the other defined by hyperglycemia above 140 mg/dl. Mortality rates were significantly higher for diabetics than for age- and sex-matched controls, and disproportionately higher for women largely due to the advantageous mortality experience of nondiabetic women compared to nondiabetic men. Estimated survival experience showed longevity to be shorter by some 10 to 15 years for the female and six to nine years for the male diabetic. Mortality rates were also significantly higher for the hyperglycemic group than for matched controls. Although deaths from cardiovascular causes were predominant, they proved significantly higher for women with known diabetes only. Substantial underreporting of diabetes on the death certificates persisted in every decade of the study. The proportion of prevalent and incident cases of diabetes during the 19-year period of the study, whose diagnosis failed to appear anywhere on the death certificate, was found to be at or above 50%. Among subjects with death certificates that did not list diabetes, 4.4 to 14.1% were known to have diabetes, based on the underlying composition of the sample. Prevalent cases of diabetes had a higher reporting rate than incident cases, suggesting that duration of diabetes may also be a factor in such rates.