To investigate the relationship between measures of social deprivation and mortality in adults with diabetes, data from 2104 randomly selected adults (> 16 years of age) with Type 1 and Type 2 diabetes mellitus from 8 hospital out-patient departments were analysed. A total of 38% of subjects had Type 1 (diagnosed before the age of 36 years and treated with insulin), 55% were male and 85% Caucasian. During a follow-up period (mean (SD) of 8.4 (0.9) years), 293 (14%) of the subjects died, the most commonly recorded cause of death being cardiovascular disease. Duration adjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated separately for Type 1 and Type 2 subjects. The mortality rates for men were higher than for women (Type 1: OR 1.27, CI 0.61-2.62; Type 2: OR 1.79, CI 1.27-2.52); were higher for those of lower vs higher social class (Type 1: OR 1.34, CI 0.61-2.96; Type 2: OR 2.0, CI 1.41-2.85); and were higher for those who left school before 16 years of age compared to those who left school at or after 16 years of age (Type 1: OR 3.98, CI 1.96-8.06; Type 2: OR 2.86, CI 1.93-4.25). Subjects who were unemployed had a higher mortality rate than those employed at the time of the study (Type 1: OR 3.10, CI 1.67-5.79; Type 2: OR 2.88, CI 2.12-3.91) and those living in council housing had a greater mortality than those who were living in other types of housing (Type 1: OR 2.57, CI 1.35-4.91, Type 2: OR 2.76, CI 2.05-3.73). Also for both Type 1 and Type 2 subjects mortality was significantly higher in those subjects who had a least one diabetic complication at baseline and reported one or more hospital admissions in the previous year and in Type 2 subjects with poor glycaemic control. After adjusting for duration of diabetes, hospital admissions, and the presence of diabetic complications, being unemployed, male, in poor glycaemic control (Type 2 only), and less educated were significant risk factors for mortality (p<0.001). These results suggest that there are important indicators of social deprivation which predict mortality over and above diabetic health status itself. Locally targeted action will be required if these inequalities in health experienced by people with diabetes are to be reduced.