Panoramic Smoking Burden and Genetic Susceptibility in Relation to All-Cause and Cause-Specific Mortality: a Prospective Study in UK Biobank.

Qiying Song,Tao Zhou,Dianjianyi Sun,Hao Ma,Xiang Li,Yoriko Heianza,Lu Qi

DOI: https://doi.org/10.1111/add.15711

2021-01-01

Addiction

Abstract:Background and aims Various smoking behaviors, including smoking initiation, age of initiation, heaviness of smoking and smoking cessation, have been individually related to the risk of mortality; however, no study has assessed these smoking behaviors jointly in relation to mortality. Our study aimed to measure prospectively the association of panoramic smoking burden (PSB), generated from the four aforementioned smoking behaviors, with all-cause and cause-specific mortality, and measure whether such associations are modified by genetic variations. Design Prospective cohort study. Setting UK Biobank. Participants A total of 360 937 participants aged between 37 and 73 years were enrolled in 2006-10 and -followed-up to 2018. Measurements The exposure was PSB, constructed based on four smoking behaviors including smoking initiation, age of initiation, heaviness of smoking and smoking cessation in a weighted method. A genetically determined PSB was also constructed with smoking-associated single nucleotide polymorphisms (SNPs) and categorized into tertiles. The primary outcomes were all-cause and cause-specific mortality. Findings We identified 15 968 deaths [9022 from cancer and 5092 from cardiovascular disease (CVD)] over a median of 11.36 years' follow-up. For all-cause mortality, compared with participants with the PSB of zero, the hazard ratios of participants who had a PSB of one, two, three and four were 1.23 [95% confidence intervals (CI) = 1.18-1.29), 1.66 (95% CI = 1.59-1.75), 3.33 (95% CI = 3.17-3.51) and 5.76 (95% CI = 4.66-7.13), respectively. Among participants within each genetic risk category, low and intermediate PSB were associated with 45-58% reduced risk of all-cause death compared with high PSB. Analysis of population-attributable risk percentage indicated that 21.9, 19.1 and 24.7% of all-cause-, cancer- and CVD-specific death could have been avoided if all ever smokers initiated smoking after age 18 years, smoked < 20 cigarettes/day and quit smoking. Conclusions The panoramic smoking burden, based on smoking initiation, age of initiation, heaviness of smoking and smoking cessation, appears to be associated with all-cause and cause-specific mortality in a gradient manner with increasing panoramic smoking burden independent of other traditional and genetic risk factors.