The relationship between 15 measures of stressful working conditions and high alcohol consumption (35 g 100% ethanol per day or more for men and 25 g or more for women) was studied, using cross-sectional data from a general population survey of 1344 males and 1494 females; the ages 25-64 years in metropolitan Stockholm in 1984. In a longitudinal component of the study, hospitalization and mortality with alcohol-related diagnosis was assessed during 1984-90, and also the association between previous experience of unemployment and high alcohol consumption. Some of the associations, expressed as age-adjusted odds ratios, were positive and some were negative when high alcohol consumption was the endpoint, but there was a clear variation by sex and social class. Generally the positive associations were stronger among male non-manual employees. Among males, there was a clear association between stressful working conditions and subsequent risk of severe medical alcohol-related problems, but the precision of the estimates was low due to low number of cases. The odds ratio was 6.18 (95% confidence interval 1.86, 20.61) for twisted working positions and 6.74 (95% confidence interval 1.67, 27.19). Previous unemployment among males was associated with increased risk for high alcohol consumption, with an odds ratio of 5.71 (95% CI 1.39, 15.97) among those who had been unemployed more than once, and 1.67 (95% CI 0.76, 3.64) among those who had been unemployed once during the previous 5 years. Those and other increased odds ratios were lower when subjects with an alcohol diagnosis at inpatient care during 1980-84 were excluded in the analyses. On the whole, our findings are not conclusive. The strong, but imprecise associations between stressful working conditions and severe alcohol problems, are however challenging, and warrants further studies, preferably with longitudinal design and repeated measurements of both working condition and alcohol habits.