From a public health perspective, country of birth has been shown to provide insufficient information to assess the health of migrants, as substantial differences have been documented between newly arrived and more established migrants. In this study we investigate the effect of duration of residence, age at arrival, and the combination of the two on the association between country of birth and all-cause mortality in Sweden. Using population-based registers, we conducted a follow-up (1990-2008) study of individuals aged between 25-64 years. Parametrical survival models (Gompertz) were fitted to derive mortality Hazard Ratios (HRs) of all-cause mortality. Compared to Swedes, and after adjusting for time-varying socioeconomic characteristics, different patterns emerged between European and non-European migrants in favor of the latter group, which showed systematically lower HRs regardless of duration of residence and age at arrival, save those who spent less than 15 years and arrived during their adolescence.