Estimating smoking-attributable lung cancer mortality in Chinese adults from 2000 to 2020: a comparison of three methods

Abstract

Background Smoking is a significant public health concern in China and a leading cause of lung cancer deaths among adults. This study aims to employ three methods to estimate smoking-attributable lung cancer mortality among Chinese adults from 2000 to 2020. Methods Population attributable fractions (PAFs) of lung cancer deaths caused by smoking were estimated using lagged smoking prevalence, Peto-Lopez, and dose–response relationship methods, separately. Smoking exposure was obtained from national tobacco surveys in China, and relative risks (RR) were derived from a meta-analysis of state-of-the-art studies among the Chinese population. Finally, we estimated the sex- and age-stratified smoking-attributable lung cancer deaths in Chinese population in 2000, 2005, 2010, 2015, and 2020. Results The PAFs estimated using 5- and 10-year lagged smoking prevalence method (45–47%) and Peto-Lopez method (46–47%) were similar, while PAFs calculated using the dose–response method were highest (47–58%). The PAFs were consistently higher in males than in females. Age-specific PAFs estimated by lagged smoking prevalence method (54–60%) and the Peto-Lopez method (57–61%) in males were similar and relatively stable, with slight decreases in older populations, while the dose–response relationship-based PAFs increased with age and fluctuated by year. By using the above methods, smoking-attributable lung cancer deaths were estimated to be 134,100, 134,600, 136,600, and 155,300 in 2000 increasing to 310,300, 301,100, 306,000, and 314,700 in 2020, respectively. Conclusion The estimation from dose–response methods could better reflect the smoking effect, however, high-quality data and accurate estimation of parameters are necessary.