ASPO Abstracts
Racial and place-based differences in smoking population attributable risk for lung cancer
Category: Cancer Health Disparities
Conference Year: 2020
Abstract Body:
Purpose: To estimate attributable risks for smoking in lung cancer among blacks and whites in a large prospective
cohort.
Methods: Lung cancer is the leading cause of cancer-related death worldwide. Racial differences in both lung
cancer incidence and smoking patterns exist, but few studies have examined racial differences in the smoking
attributable risk for lung cancer. We conducted a prospective study among 41,038 African American and 18,062
white low-income adults age 40-79 at baseline in the Southern Community Cohort Study (SCCS) followed for up
to 15 years. We estimated the population attributable risk (PAR) percent of lung cancer associated with smoking
reported at baseline (classified as ever/never; current or former versus never; <10, 10-19, 20+ pack-years versus
never) using multivariable Cox proportional hazard models with robust standard errors by race, race-sex, and
geographic urban/rural residence groups.
Results: Overall, 83% of lung cancer cases (N=1,039 of 1,251 incident lung cancers) among SCCS participants
were attributable to smoking, but the PAR varied by race (PARs of 80% in blacks versus 88% in whites). The
higher overall PAR in whites was mainly due to their higher prevalence of 20+ pack-year smokers, with PARs
among such heavy smokers of 76% in whites versus 46% in blacks. Lighter smokers (< 10 pack-years) accounted
for 3% of cases among whites but 10% among blacks, with moderate smoking (10-19 pack-years) accounting for
9% of cases among whites and 24% among blacks. Despite lower prevalence of smokers among women than
men, little differences in smoking PARs were observed between males and females (PARs of 78%, 78%, 88% and
87% respectively among black males, black females, white males and white females), but PARs tended to be
higher among urban (PAR 85%) than rural (PAR 76%) residents. Higher PARs for urban residents were
maintained across all race-sex groups except for white females.
Conclusions: Racially-specific and place-based population attributable risks were observed. These PARs provide
a tool for prioritizing intervention targets to address inequities in lung cancer.
Keywords: disparities, smoking, lung cancer, population attributable risk