Beyond cigarettes per day. A genome-wide association study of the biomarker carbon monoxide

Abstract

Rationale: The CHRNA5-CHRNA3-CHRNB4 locus is associated with self-reported smoking behavior and also harbors the strongest genetic associations with chronic obstructive pulmonary disease (COPD) and lung cancer. Because the associations with lung disease remain after adjustment for self-reported smoking behaviors, it has been asserted that CHRNA5-CHRNA3-CHRNB4 variants increase COPD and lung cancer susceptibility independently of their effects on smoking.

Objectives: To compare the genetic associations of exhaled carbon monoxide (CO), a biomarker of current cigarette exposure, with self-reported smoking behaviors.

Methods: A total of 1,521 European American and 247 African American current smokers recruited into smoking cessation studies were assessed for CO at intake before smoking cessation. DNA samples were genotyped using the Illumina Omni2.5 microarray. Genetic associations with CO and smoking behaviors (cigarettes smoked per day, Fagerstrom test for nicotine dependence) were studied.

Measurements and main results: Variants in the CHRNA5-CHRNA3-CHRNB4 locus, including rs16969968, a nonsynonymous variant in CHRNA5, are genomewide association study-significantly associated with CO (β = 2.66; 95% confidence interval [CI], 1.74-3.58; P = 1.65 × 10(-8)), and this association remains strong after adjusting for smoking behavior (β = 2.18; 95% CI, 1.32-3.04; P = 7.47 × 10(-7)). The correlation between CO and cigarettes per day is statistically significantly lower (z = 3.43; P = 6.07 × 10(-4)) in African Americans (r = 0.14; 95% CI, 0.02-0.26; P = 0.003) than in European-Americans (r = 0.36; 95% CI, 0.31-0.40; P = 0.0001).

Conclusions: Exhaled CO, a biomarker that is simple to measure, captures aspects of cigarette smoke exposure in current smokers beyond the number of cigarettes smoked per day. Behavioral measures of smoking are therefore insufficient indices of cigarette smoke exposure, suggesting that genetic associations with COPD or lung cancer that persist after adjusting for self-reported smoking behavior may still reflect genetic effects on smoking exposure.

Keywords: chronic obstructive pulmonary disease; lung cancer; nicotine; nicotinic receptor; smoking.

Figure 1.

Distribution of mean baseline exhaled carbon monoxide (CO) among current smokers stratified by ancestry group.

Figure 2.

Genomewide association scan among European Americans for genes associated with exhaled carbon monoxide (CO). (A) The –log₁₀ of the P value is plotted for each single nucleotide polymorphism (SNP) in chromosomal order. The spacing between SNPs on the plot is based on physical map length. (B) Q-Q plot for genomewide association scan among European Americans for genes associated with exhaled CO.

Figure 3.

Mean baseline exhaled carbon monoxide (CO) among current smokers divided by categorical cigarettes per day (CPD) and rs16969968 genotype among European Americans. The boxplot represents the interquartile range divided by a line indicating the median; whisker lines extend to values within 1.5× the interquartile range; further outliers are marked with circles.

Figure 4.

Mean baseline exhaled carbon monoxide (CO) among current smokers divided by categorical cigarettes per day and ancestry group. The boxplot represents the interquartile range divided by a line indicating the median; whisker lines extend to values within 1.5× the interquartile range; further outliers are marked with circles.