Objectives To estimate excess mortality in patients with an alcohol use disorder and to calculate the population-based impact attributable to this disorder, with emphasis on differences according to gender and age. Methods We carried out a longitudinal study of a dynamic retrospective cohort in which 7,109 patients (76.5% men) aged 18-64 years who started medical day treatment between 1997 and 2006 were recruited. Deaths were monitored until the end of 2006 using a national mortality register. Standardized mortality ratios (SMR) by age and gender and population attributable risks (PAR) by age group were estimated. Results There were 557 deaths (83% were men). The total excess mortality was 8 times higher than in the general population aged 18-64 years, especially in the group aged 25-34 years old (SMR: 11.2 [95% CI: 7.8-16.0] in men and 24 [IC95%: 11.5-50.4] in women). Significant differences by gender were found in all the variables analyzed. In Barcelona, alcohol use disorder was estimated to cause 73 deaths per year, representing 3.3% of the total annual deaths in this city in persons aged 18-64 years. This percentage was substantially higher in the groups aged 25-34 years (PAR: 19.4% [95% CI: 19.2-19.6]) and 18-24 years (PAR: 11.1% [95% CI: 10.7-11.4]). Conclusions Women and young men with an alcohol use disorder have a higher risk of early mortality. Early detection of problematic alcohol consumption and selective and well-indicated prevention programs should be improved.