Abstract Background: Studies have inconsistently reported an association between tobacco smoking and Hodgkin lymphoma (HL) risk. The conflicting findings may reflect etiologic heterogeneity between HL subtypes, warranting further characterization of the relationship. Methods: We collected information on tobacco-smoking habits in 586 classic HL cases and 3,187 population controls in a Danish-Swedish case-control study. HL EBV status was established for 499 cases by standard techniques. Odds ratios (OR) for an association with cigarette smoking were calculated by logistic regression for HL overall and stratified by age, sex, major histology subtypes, and tumor EBV status, adjusting for known confounders. Results: Compared with never smokers, current cigarette smokers were at an increased overall HL risk (adjusted OR, 1.57; 95% confidence interval (95% CI), 1.22-2.03). The association was strongest for EBV-positive HL (adjusted OR, 2.36; 95% CI, 1.51-3.71), but also applied to EBV-negative HL (adjusted OR, 1.43; 95% CI, 1.05-1.97; Phomogeneity EBV-pos versus EBV-neg = 0.04). The association did not vary appreciably by age, sex, or histologic subtype, the apparent EBV-related difference present in all strata. There was no evidence of a dose-response pattern, whether by age at smoking initiation, daily cigarette consumption, number of years smoking, or cumulative number of cigarettes smoked. Similar results were obtained in analyses using non-HL patients (n = 3,055) participating in the founding study as comparison group. Conclusion: The observed association between cigarette smoking and HL risk is consistent with previous findings and biologically plausible. Although not easily dismissed as an artifact, the limited evidence of a dose-response pattern renders the overall evidence of causality weak. (Cancer Epidemiol Biomarkers Prev 2007;16(8):1561–6)