BACKGROUND: A large earthquake can cause extreme stress and may adversely affect cognitive function in humans. We aimed to examine a possible association between psychological distress and incident dementia after the 2004 Niigata-Chuetsu earthquake in Japan. METHODS: This is a retrospective cohort study followed participants for 10-12 years. Subjects were 6,012 residents in 2005, 5,424 in 2006, and 5,687 in 2007 (age ≥40 years) living in Ojiya city who participated in the annual health check examinations after the 2004 Niigata-Chuetsu earthquake. Psychological distress was assessed using the Kessler Psychological Distress Scale (K10), and individuals with a K10 score ≥10 were considered to have psychological distress. Incident dementia cases were identified from a long-term care insurance database of the local government during the follow-up period. We evaluated hazard ratios (HRs) of psychological distress for incident dementia in each year, unadjusted and adjusted for covariates, including sex, age, occupation, BMI, and property damage of residential area. RESULTS: The average age of the subjects was 64.6 years in 2005, 64.6 in 2006, and 65.2 in 2007. Adjusted HRs were significantly higher (HR = 1.20-1.66) in the psychological distress group than in the reference group in each year. In particular, adjusted HR was high (HR = 2.89) in those with psychological distress in all three years (2005-2007). CONCLUSION: Psychological distress, especially persistent distress, is a risk factor for incident dementia in victims of large disasters.