The causes of death and associated risk factors are compared in young and old diabetic patients attending a retinopathy clinic. Mortality in those diagnosed under and over the age of 30 years is also examined in order to compare insulin-dependent with non-insulin-dependent patients. A defined cohort attending the Hammersmith Hospital Retinopathy Clinic was followed for an average of 11 years. Main outcome measures were standardized mortality ratios (SMRs) in different age/sex groups and relative hazard rates (RHRs) for possible risk factors related to mortality. The patients were divided into those aged under and over the age of 60 years at attendance at the clinic. The RHRs were smaller in the elderly for plasma urea [1.015 versus 1.107 (p < 0.01)]. Attenuation of risk was also suggested for systolic blood pressure (RHR of 1.005 in the elderly versus 1.015 in the younger patients) and for the effects of smoking [RHR of 1.17 (elderly patients) and 1.35 (younger patients)]. In those diagnosed under the age of 30 years, there were very high SMRs for renal disease, cerebrovascular disease (men only), ischemic heart disease, other heart disease, and respiratory disease (men only), but increased SMRs were also demonstrated in those diagnosed over the age of 30 years. The risk factors associated with poor survival were similar for those diagnosed over and under the age of 30 years: poor diabetic control, high systolic and diastolic blood pressure, and increased plasma urea. In conclusion, there was no evidence that blood sugar control or diastolic blood pressure were less important in older age groups. Plasma urea, systolic pressure, and being on insulin were less useful as predictors of mortality in the elderly, but were still important in patients diagnosed over the age of 30 years.