The aim of this study is to investigate the association between alcohol consumption and the risk of bacterial infection and its dose-response association. Participants in the Swedish Mammography Cohort and Cohort of Swedish Men answered lifestyle questionnaires in 1997 and have since been followed in national registers. The risks of acquiring infection, intensive care unit (ICU) admission and dying due to infection were assessed with Cox regression. Among 58,078 cohort participants followed for 23 years, 23,035 participants were diagnosed with an infection and 4,030 died from infection. Alcohol consumption exhibited a J-shaped association with the risk of acquiring infection and dying due to infection: compared to consuming 5-10 g of alcohol per day, consuming < 0.5 g/day and consuming > 30 g/day were both associated with higher risk of acquiring infection, ICU admission and dying due to infection, whereas alcohol consumption between 5 and 30 g/day was not associated with acquiring infection, ICU admission or death due to infection. In conclusion, moderate alcohol consumption was not associated with infection, but both very low and high levels of consumption were associated with acquiring infection, ICU admission and death. If replicated, this suggests that reduction of alcohol consumption might reduce mortality from bacterial infections.