OBJECTIVE A prospective cohort study was conducted to assess the excess mortality among victims of arsenic poisoning who had ingested "Morinaga dry-milk" that was contaminated with arsenic compounds in 1955. METHODS We identified and enrolled 5,064 individuals who had ingested contaminated Morinaga dry-milk when they were aged two years or younger, in 1982 (mean age: 27.4 years) and they were followed until 2004 (mean length of follow-up: 22.3 years). The death certificates of subjects who died between 1982 and 2004 were examined. The risk of death was assessed by the ratio of the observed number (O) to the expected number of deaths (E), calculated from the mortality rate among Osaka residents. RESULTS The O/E ratio for all causes of death was 1.3 (O = 211, P < 0.01). Significant excess mortality was observed for diseases of the central nervous system (O/E = 5.4) or circulatory system (O/E = 1.4), external causes (O/E = 1.4) and traffic accidents (O/E = 2.0). Excess mortality from all causes appeared 0 to 4 years after study enrollment (O/E = 2.1, P < 0.01), and then the O/E ratio decreased to unity (O/E = 1.2) beyond 10 years after study enrollment. The 352 males who were unemployed at the time of enrollment in the study showed significantly elevated risk of death from all causes (O/E = 3.3), death from disease of the central nervous system (O/E = 36.7), circulatory system (O/E = 3.7) or respiratory system (O/E = 5.7), and death from external causes (O/E = 3.4). CONCLUSIONS This prospective cohort study showed that excess mortality from all causes among the victims of arsenic poisoning from ingestion of arsenic-contaminated "Morinaga dry-milk" in 1955 decreased to unity when they reached middle age.