Objectives To quantify the relation between prenatal famine exposure and adult mortality, taking into account mediating effects of intermediary life conditions. Design Historical follow-up study. Setting The Dutch famine (Hunger Winter) of 1944–1945 which occurred towards the end of WWII in occupied Netherlands. Study population From 408,015 Dutch male births born 1944–1947, examined for military service at age 18, we selected for follow-up all men born at the time of the famine in six affected cities in the Western Netherlands (n = 25,283), and a sample of unexposed time (n = 10,667) and place (n = 9087) controls. These men were traced and followed for mortality through the national population and death record systems. Outcome measure All-cause mortality between ages 18 and 63 years using Cox proportional hazards models adjusted for intermediary life conditions. Results An increase in mortality was seen after famine exposure in early gestation (HR 1.12; 95% confidence interval (CI): 1.01–1.24) but not late gestation (HR 1.04; 95% CI: 0.96–1.13). Among intermediary life conditions at age 18 years, educational level was inversely associated with mortality and mortality was elevated in men with fathers with manual versus non-manual occupations (HR 1.08; CI: 1.02–1.16) and in men who were declared unfit for military service (HR 1.44; CI: 1.31–1.58). Associations of intermediate factors with mortality were independent of famine exposure in early life and associations between prenatal famine exposure and adult mortality were independent of social class and education at age 18. Conclusions Timing of exposure in relation to the stage of pregnancy may be of critical importance for later health outcomes independent of intermediary life conditions.