The hypothesis was that smokers might have more aggressive types of breast cancer because of either delayed diagnosis or higher grade and hence have a worse prognosis. A cohort of breast cancer patients completed a lifestyle questionnaire at the time of diagnosis, including whether they were current smokers, ex-smokers or lifelong non-smokers. Ex-smokers were asked when they had stopped. The participants were 166 women with stage I/II invasive breast cancer diagnosed between October 1984 and March 1987. Participants were divided into three groups: current smokers, ex-smokers and non-smokers. Survival curves were produced by using Cox proportional hazards analysis, with outcome variables for overall and breast cancer-specific survival together with distant relapse-free survival. Smoking was the third most important predictor of distant relapse-free, breast cancer-specific and overall survival after stage and age at diagnosis. These results suggest that smokers are not only more likely to die of other diseases, but also have a higher mortality from breast cancer, compared with those with the disease who have never smoked. The best prognosis, however, was found in those who had given up smoking.