We performed sensitivity analyses to assess whether educational level modified the relationship of current or former smoking to each end point, using both stratified analyses and interaction terms. To assess whether the lack of fully updated smoking information in some cohorts or differences in the follow-up periods might have biased the observed associations, we compared the results from the CPS II Nutrition cohort according to the time when smoking status was recorded: at baseline, 2 years before the end of follow-up, or at the most recent update through 2005.
