Aims: To assess the diagnostic accuracy of four undiagnosed Type 2 diabetes mellitus risk scores accounting for erectile dysfunction status. Methods: This was a population-based cross-sectional study. Type 2 diabetes was defined according to a oral glucose tolerance test and self-reported physician diagnosis. Erectile dysfunction was defined according to the answer to the question, ‘Have you had difficulties obtaining an erection in the last 6 months?’ (yes/no). The risk scores used were the FINDRISC, LA-FINDRISC, American Diabetes Association score and the Peruvian Risk Score. A Poisson regression model was fitted to assess the association between Type 2 diabetes and erectile dysfunction. The area under the receiver-operating characteristic curve was estimated overall and by erectile dysfunction status. Results: A total of 799 men with a mean (sd) age of 48.6 (10.7) years were included in the study. The overall prevalence of Type 2 diabetes was 9.3%. Compared with healthy men, men with Type 2 diabetes had 2.71 (95% CI 1.57–4.66) higher chances of having erectile dysfunction. Having excluded men aware of Type 2 diabetes status (N=38), the area under the receiver-operating characteristic curve of three of the risk scores (not the American Diabetes Association score) improved among those who had erectile dysfunction in comparison with those who did not; for example, the area under the receiver-operating characteristic curve of the LA-FINDRISC score was 89.6 (95% CI 78.7–99.9) in men with erectile dysfunction and 76.5 (95% CI 68.5–84.4) overall. Conclusions: In a population-based study, erectile dysfunction was more common in men with Type 2 diabetes than in the otherwise healthy men. Screening for erectile dysfunction before screening for Type 2 diabetes seems to improve the accuracy of well-known risk scores for undiagnosed Type 2 diabetes. / Antonio Bernabe-Ortiz is a Research Training Fellow in Public Health and Tropical Medicine (103994/Z/14/Z), funded by the Wellcome Trust. / Revisión por pares