Although outdoor air pollution including particulate matter (PM) was classified as carcinogenic to humans based on accumulating epidemiological evidence, these findings were suggested mostly from low-dose environments in North America and Europe. We aimed to examine the association of long-term exposure to PM ≤ 10 and 2.5 μm in diameter (PM10 and PM2.5) and nitrogen dioxide (NO2) with lung cancer incidence using a population-based cohort in the Seoul Metropolitan Area (SMA), South Korea. Our study included 83,478 people residing in the SMA and followed up for 2007–2015 from the National Health Insurance Service-National Sample Cohort. This cohort was constructed based on the National Health Insurance database that contains sociodemographic and medical information under universal health coverage. Individual long-term concentrations of PM10, PM2.5, and NO2 were estimated at people's district-level and annually-updated residential addresses for the previous 5 years by using previously-validated prediction models. We applied a time-dependent Cox proportional hazards model and estimated hazard ratios (HRs) per 10 μg/m3 and 10 ppb increases in PM and NO2, respectively, after adjusting for individual characteristics. During 9 years of follow-up, 489 lung cancer new cases occurred (714,012 person-year). The adjusted HRs for PM10 were greater than 1 but statistically non-significant (HR = 1.15; 95% CI = 0.88–1.52). We also did not find associations for PM2.5 and NO2. Despite null associations for the total population, our subgroup analysis suggested associations with PM in family members with cancer (PM10: HR = 2.59, 95% CI = 1.26–5.32; PM2.5: 5.55, 1.09–27.91) and in those who have smoked more than 1 pack per day (1.77, 0.96–3.25; 3.81, 1.00–14.44) or for less than 20 years (2.81; 1.13–7.07; 2.02, 0.21–19.23). Our study based on a highly urbanized population exposed to relatively high air pollution provides no evidence of the association between PM and lung cancer incidence in the total population but indicates the potential susceptibility in heavy smokers for relatively short periods and family members of cancer patients. Future studies should re-examine the association using improved exposure assessment and extended population.
