Background: To update the findings of relative risk associated with smoking for all-cause mortality and that for lung cancer by considering longitudinal changes in smoking status during follow-up.

Methods: Data from the JPHC study of 98,747 middle-aged Japanese adults, which started in 1990–1993, were analyzed. The information on smoking status was obtained from three questionnaire surveys (baseline, the 5th year, and the 10th year after the start of follow-up). A Poisson regression model was used to investigate the impact of smoking on mortality from all causes and lung cancer using two approaches. Model 1 used information only from baseline, while model 2 used the updated smoking status from all three surveys.

Results: During the 15-year follow-up, 10,702 all-cause deaths (including 870 lung cancer cases) were identified. We compared the results obtained from two models. The relative risks associated with former smokers versus never smokers were 1.42 (95% confidence interval [CI], 1.31–1.54) among men and 1.46 (95% CI, 1.23–1.73) among women for all-cause mortality and 2.98 (95% CI, 2.09–4.24) among men and 1.83 (95% CI, 0.92–3.64) among women for lung cancer mortality, as determined using model 2. All of these were higher than the relative risks obtained from model 1. In addition, former smokers who had quit smoking due to disease during follow-up had a higher mortality risk than continuous smokers did in this study.

Conclusions: The relative risks of all-cause mortality and mortality due to lung cancer among former smokers be higher than previously documented based on updated smoking status data from repeated surveys.