The p values reported above were based on the adjusted analyses correcting for smoking exposure. The results from the unadjusted association analyses for COPD status were highly significant (Bergen 2×10−4 and 4×10−4; ICGN 7.46×10−7 and 1.40×10−6; NETT/NAS, 2.0×10−5 and 2.5×10−4 and combined p values of 1.86×10−12 and 6.6×10−11 for rs8034191 and rs1051730 respectively; Table S3). Although the adjustments for smoking exposure attenuated the p values, they still remained highly significant (Table 2). In the Norwegian discovery cohort, a significant genotype-by-environment interaction (P = 0.002, Table 3) was observed with a substantially higher risk of COPD in current smokers carrying the rs8034191 C allele (OR = 2.00) than in former smokers (OR = 1.10). In the overall population, the C allele of rs8034191 was estimated to have a population attributable risk of 12.2% for COPD. This risk was 14.3% in current smokers and 3.1% in former smokers. The p values were attenuated in the ICGN family-based population following adjustment for age, sex, pack-years of smoking and center but remained highly significant (Table 2). We identified a SNP by pack-years interaction (p
