The effect of smoking on breast cancer risk has been null in large, well-conducted cohort studies. In a previous study, we stratified the population into active smokers, passive smokers, and never-active never-passive smokers and modeled early life cigarette smoke exposures as causal and later life cigarette smoke exposure as preventive. We observed a complex association between cigarette smoke and breast cancer risk. Using a similar design and population, this study did not confirm the earlier result. Neither ever-active smoking (adjusted odds ratio (OR) = 0.72, 95% confidence intervals (CI) 0.55-0.95) nor ever-passive smoking (adjusted OR = 0.85, 95% CI 0.63-1.1) were strongly associated with breast cancer risk compared with never-active never-passive smoking. No patterns of effects were observed in subgroup analyses.