OBJECTIVE To compare the outcome of renal replacement treatment in patients with diabetes mellitus and in non-diabetic patients with end stage renal failure. DESIGN Retrospective comparison of cases and matched controls. SETTING Renal unit, Western Infirmary, Glasgow, providing both dialysis and renal transplantation. PATIENTS 82 Diabetic patients starting renal replacement treatment between 1979 and 1988, compared with 82 matched non-diabetic controls with renal failure and 39 different matched controls undergoing renal transplantation. MAIN OUTCOME MEASURES Patient characteristics, history of smoking, prevalence of left ventricular hypertrophy and myocardial ischaemia at start of renal replacement treatment; survival of patients with renal replacement treatment and of patients and allografts with renal transplantation. RESULTS The overall survival of the diabetic patients during the treatment was 83%, 59%, and 50% at one, three, and five years. Survival was significantly poorer in the diabetic patients than the controls (p less than 0.001). Particularly adverse features for outcome at the start of treatment were increasing age (p less than 0.01) and current cigarette smoking (relative risk (95% confidence interval) 2.28 (0.93 to 4.84), p less than 0.05). Deaths were mainly from cardiac and vascular causes. The incidence of peritonitis in patients on continuous ambulatory peritoneal dialysis was the same in diabetic patients and controls (49% in each group remained free of peritonitis after one year), and the survival of renal allografts was not significantly worse in diabetic patients (p less than 0.5). CONCLUSIONS Renal replacement treatment may give good results in diabetic patients, although the outlook remains less favourable than for non-diabetic patients because of coexistent, progressive vascular disease, which is more severe in older patients.