BACKGROUND: The aims of the present study are to analyse the association between marital status at age 24, 29, 34, and 39 years and subsequent mortality in a cohort of men born in 1953 (sensitive period); to study the impact of number of years married, number of years divorced/widowed, and number of marital break-ups on mortality (cumulative effect), and to examine whether these effects were independent of marital status at age 39 (proximity effect). METHODS: Prospective birth cohort study with follow-up of mortality from 1992 to 2002. Participants were 10891 men born within the metropolitan area of Copenhagen, Denmark. Marital status in 1992 as well as start and termination of all previous marital status events from 1968 to 1992 were retrieved from the Danish Civil Registration System. MAIN OUTCOME MEASURES: Were hazard ratios (HR) for all-cause mortality from age 40 to 49 years. RESULTS: We found a strong protective effect of being married compared with never being married or divorced/widowed at every age. The association increased in strength with increasing age. Number of years divorced was associated with increased mortality risk in a dose-dependent manner at age 34 and 39 years. One or more marital break-ups was associated with higher mortality, whereas increasing number of years married was associated with lower mortality. Inclusion of current marital status attenuated the strength of the associations but most of them remained statistically significant. CONCLUSIONS: Marital status and cumulated marital periods, especially cumulated periods divorced/widowed are strong independent predictors of mortality among younger males.