Prevalence of diabetic retinopathy. Age, sex, and duration of diabetes.

From a list of all patients seen at the Joslin Clinic in a recent 6-month period, a random sample of 965 was selected for record review by a clinic assistant. A house physician specializing in diabetes then verified and adjusted the assistant's work as necessary. The data were analyzed both by simple cross-classification and by fitting a multiple logistic risk function. As often reported previously, we found a strong positive association between retinopathy and duration of diabetes. One of our findings which has not been so clearly reported in earlier papers is that the positive association between retinopathy and age is limited to the group with diabetes of less than 10 years' duration. We also found sex differences in retinopathy prevalence which were not large enough to rule out attributing them to chance.