As the life expectancy of people with intellectual disability (ID) increases, it is becoming necessary to understand factors affecting survival. However, predictors that are typically assessed among healthy people have not been examined. Predictors of all-cause mortality, including blood, urine, anthropometry, and nutritional indices, were examined among institutionalized people with ID. This retrospective cohort study involved 316 participants (191 males, 125 females; mean age, 36.5 plus or minus 10.5 years) at a public facility for people with ID in Ibaraki Prefecture, Japan. During the follow-up from the examination day in 1984-1992 through December 31, 2007 (mean follow-up, 18.6 years), 44 deaths occurred. Mean age at death was 47.1 plus or minus 10.0 years (range, 22.3-65.3 years). Early deaths within three years (n = 4) were treated as censored cases. Cox proportional hazard regression models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) of all-cause mortality. Sex- and age-adjusted analysis (p less than 0.15) revealed positive associations with mortality for high serum cholesterol, high thymol turbidity test (TTT), and glucosuria and negative associations with mortality for high serum albumin, high uric acid, high potassium, high calcium, and high systolic blood pressure. Multivariate analysis revealed that male sex (HR, 4.11; 95% CI, 1.59-10.59), high serum cholesterol (1.01; 1.00-1.02), high serum TTT (1.21; 1.03-1.41), and epilepsy significantly increased the mortality risk. The results indicate that the predictors of life expectancy for people with ID included both factors that are shared with healthy people (male sex, high serum cholesterol) and factors specific to people with disabilities (high serum TTT and epilepsy). (Contains 3 tables.)