Cigarette smoking has been reported to alter relapse rate in patients with Graves' disease (GD). However, the predictive effect of smoking in GD patients after withdrawal of antithyroid drug treatment (ATDT) is still controversial. A prospective multicenter trial has previously identified smoking as an independent risk factor for relapse. Based on this study, the present paper gives a more detailed analysis of the impact of smoking on the long-term course of GD after ATDT withdrawal. To this end, 86 smokers and 177 non-smokers were followed during two years after ATDT cessation. At the end of ATDT (visit 1) and four weeks later (visit 2) smokers had significant higher TSH receptor antibody (TRAb) levels than non-smokers (10.0 IU/L+/-1.6; mean+/-SEM vs. 6.4 IU/L+/-0.9; 11.0 IU/L+/-1.8 vs. 6.8 IU/L+/-0.8, p < 0.01, respectively). During follow-up, Kaplan Meier analysis showed a significantly higher relapse rate in smokers than non-smokers. A subset of GD patients with TRAb levels >10 IU/L had the highest risk to develop relapse during follow-up. Among them, smokers more often relapsed than non-smokers irrespective of TRAb levels, p < 0.01. Thus, in smokers with TRAb levels > or =10 IU/L the predictive values of a positive and negative test for relapse was 68% and 73%, respectively (specificity 95%). In conclusion, we identified two effects by which smoking alters the course of GD. First, smoking is implicated to elevate TRAb levels and therefore increase the risk for relapse during follow-up. Second, smoking is an independent risk factor to worsen the clinical course of both, GD patients with low and high immunological risk to experience relapse after a successful outcome of ATDT. Thus, our data suggest that smoking has modifying immunological consequences and an adverse impact on the course of GD after withdrawal of ATDT. Therefore, patients should be encouraged to stop smoking.