The relationship between serum uric acid (SUA) and risk of coronary heart disease (CHD) mortality remains controversial, particularly in diabetic subjects. The aim of the present study is to evaluate whether SUA independently predicts CHD mortality in non-insulin-dependent elderly people from the general population and to investigate the interactions between SUA and other risk factors. Five hundred and eighty-one subjects aged >/=65 years with non-insulin-dependent diabetes mellitus were prospectively studied in the frame of the CArdiovascular STudy in the ELderly (CASTEL). Historical and clinical data, blood tests and 12-year fatal events were recorded. SUA as a continuous item was divided into tertiles and, for each tertile, adjusted relative risk (RR) with 95% confidence intervals (CI) was derived from multivariate Cox analysis. CHD mortality was predicted by SUA in a J-shaped manner. Mortality rate was 7.9% (RR 1.28, CI 1.05-1.72), 6.0% (reference tertile) and 12.1% (RR 1.76, CI 1.18-2.27) in the increasing tertiles of SUA, respectively, without any difference between genders. In diabetic elderly subjects, SUA independently predicts the risk of CHD mortality in a J-shaped manner.