It has been shown that smokers are more likely than non-smokers to develop Crohn's disease. In order to examine the influence of smoking on the course of Crohn's disease, as measured by the risk of surgery, the need for surgery was assessed in patients with Crohn's disease. The course of Crohn's disease was evaluated as to smoking status by life table analysis. A tertiary hospital in Essen, Germany. The study involved 346 patients with Crohn's disease admitted to hospital between 1989 and 1992. Only patients residing in the City of Essen were included. Of the 346 patients, 144 smokers and 143 non-smokers were included in the analysis. Overall, 73% of smokers and 39% of non-smokers required one or more operations. The corresponding relative risks (RR) and 95% confidence intervals were for one operation and more than one operation or no operation at all (RR 1.0) 3.9 (2.2–6.9) and 10.8 (5.3–22.1), respectively. There were significant differences as to recurrence rates (defined as further surgery after first surgery for control of disease) between smokers and non-smokers. For smokers the 5− and 10-year recurrence rates were 43% and 64%. For non-smokers corresponding recurrence rates were 26% and 33%. For 5− and 10-year recurrence rates the RR estimates for smokers versus non-smokers were 3.1 (1.7–5.8) and 6.7 (2.7–6.8). When stratified by gender, the increased risk for recurrence was obvious in both sexes. For the number of cigarettes smoked a dose-response effect was obvious in women. These data suggest that the course of Crohn's disease is less favourable in smokers than in non-smokers.