BACKGROUND: The article analyses effects of health during young adulthood, lifestyles during adulthood, and working conditions on variations in general health among men aged 55. MATERIAL AND METHODS: Data were obtained from interviews in 2001 in a nation-wide sample of men born in 1946 (N = 380). Additive indices were constructed for health problems at age 55, health problems at age 25, smoking habits, physical exercise and weight problems during adulthood, and physical and psychological working conditions and working time. The indices were analysed by correlation and multivariate regression. RESULTS: Significant bivariate correlations were found between health problems at age 55 and all the explanatory variables. In the multivariate regression analysis, only health at age 25 and physical and psychological working conditions had significant effects. INTERPRETATION: The analysis indicates the relevance of a life-course perspective on health. Health problems as young adults and physical/ psychological working conditions appeared as important explanations for health variations among men in their mid-fifties. An unexpected finding was that variations in lifestyle had practically no effect when adjustments for previous health and working conditions were made. Sources of error are discussed.