Our findings should be interpreted with a few caveats. COPD cases in this study were identified by self-reports of diagnosis of emphysema, chronic bronchitis, or both; these cases are highly likely to represent truly diseased individuals, but a degree of non-differential misclassification may be present [33]. TTFC was only assessed once (six years post-baseline), but this assessment was at an age when dependency is likely to be highly stable; study participants had, on average, smoked for 50 years and over one-third of our sample initiated this lifelong behavior prior to age 17. The effects of misclassification of COPD and TTFC on our estimates, if any, would be to attenuate the true association, but we detected a robust significant increased risk of COPD with shortened TTFC. PLCO questionnaires queried both emphysema and chronic bronchitis, but did not specifically use the term COPD; future studies should ask participants about emphysema, chronic bronchitis, and COPD, which may capture a greater portion of individuals with disease. Although we included participants in the screening and control arms of PLCO there was no evidence of differential detection, with COPD identified in 19% of participants in each arm.
