Smoking-Attributable Mortality in China - Epidemic Stage as a Temporal Dimension in Interpreting GBD Estimates

Tobacco use remains a leading cause of preventable mortality worldwide. Estimates from the Global Burden of Disease, Injuries, and Risk Factors Study (GBD) are widely used to quantify smoking-attributable mortality. Over recent GBD cycles, substantial methodological refinements - including improved quantification of cumulative smoking exposure and updated relative risk estimation using the Burden of Proof (BOP) Framework - have strengthened the methodological robustness of smoking-attributable mortality estimates. Nevertheless, recent GBD estimates for China remain difficult to reconcile with evidence from large national cohort studies and repeated surveys. Smoking-attributable mortality is fundamentally shaped by cumulative exposure and long latency, while the stage of the tobacco epidemic reflects the evolution of smoking exposure across successive birth cohorts and calendar periods. Evidence from China suggests that smoking-attributable mortality has not yet reached its expected plateau, whereas GBD estimates imply that smoking has accounted for approximately 30% of male deaths since around 2000. These contrasting patterns reflect different implied temporal trajectories of the tobacco epidemic, rather than simple differences in numerical estimates. We propose that the epidemic stage should be considered an essential temporal dimension when interpreting smoking-attributable mortality estimates. Incorporating the epidemic stage alongside existing modeling approaches would better reflect the differences in the timing of cumulative exposure and evolution of smoking-attributable mortality, thereby improving the interpretation, comparability, and policy relevance of model-based estimates across populations.