Table 3.
Hazard ratios of lung cancer per unit increase in 3-year average PM2.5 (μg/m3), NO2 (ppb), warm-season ozone (ppb), and PR (mBq/m3) exposures in single- and multi-pollutant models in Medicare cohort from 2004 to 2016.
| Model | PM2.5 | NO2 | Warm- season ozone |
PR |
|---|---|---|---|---|
| Single-Pollutant | ||||
| Full cohort a | 1.015 (1.012, 1.018) | 1.012 (1.011, 1.012) | 0.997 (0.996, 0.998) | 1.016 (1.010, 1.022) |
| Low-exposure cohort b | 1.007 (1.001, 1.012) | 1.014 (1.013, 1.016) | 1.008 (0.999, 1.016) | 1.134 (1.059, 1.214) |
| Multi-pollutant a | ||||
| Full cohort a | 1.008 (1.005, 1.011) | 1.013 (1.012, 1.013) | 0.991 (0.990, 0.992) | 1.005 (0.999, 1.012) |
| Low-exposure cohort b | 1.004 (0.998, 1.010) | 1.016 (1.015, 1.017) | 1.005 (0.996, 1.013) | 1.125 (1.046, 1.210) |
Analysis conducted in full cohort (PYs = 95,280,525, lung cancer events = 166,860).
Analysis restricted to person-years with exposure levels ≤ 10 μg/m3 for PM2.5 (PYs = 55,771,344, lung cancer events = 94,429), ≤ 30 ppb for NO2 (PYs = 84,120,835, lung cancer events = 145,449), ≤ 35 ppb for warm-season ozone (PYs = 7,584,079, lung cancer events = 13,344), and ≤ 8 mBq/m3 for PR (PYs = 2,951,296, lung cancer events = 4,620). All models were adjusted for sex, race (Black/White/Other), Medicaid eligibility (yes/no), temperature, population density, % Black population, % American Indian and Alaska Native population, % Asian population, % Two or More Races population, % Native Hawaiian and Other Pacific Islander population, % Hispanic population, % population using automobile to transport, % population receiving less than high school education, % population above 65 years of age living below the poverty line, median household income, % population living in rented houses or apartments, distance to the nearest hospital, % of Medicare enrollees having at least one ambulatory visit to a primary care clinician in a year, % of diabetic Medicare enrollees aged 65–75 having hemoglobin A1c test in a year, number of hospitals, number of medical doctors, number of hospital beds, NDVI, BMI, and smoking rate. The multi-pollutant models were additionally adjusted for other pollutants.