• Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD, USA, fangyigu@gmail.com.
• Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD, USA.
• School of Public Health, University of Michigan, Ann Arbor, MI, USA and.
• Cancer Epidemiology Research Program, Vanderbilt University School of Medicine, Nashville, TN, USA.

BACKGROUND: Identifying circulating metabolites related to cigarette smoking may provide insight into the biological mechanisms of smoking-related diseases and the nature of addiction. However, previous studies are limited, generally small, and have largely targeted a priori metabolites.

METHODS: We examined associations between cigarette smoking and metabolites using an untargeted metabolomics approach in 892 men and women from four studies including participants from Italy, USA, China and Finland. We examined associations between individual log-transformed metabolites and two key smoking phenotypes (current smoking status and cigarettes per day [cig/day]) using linear regression. Fixed-effect meta-analysis was used to combine results across studies. Strict Bonferroni thresholds were used as our significance criteria. We further examined associated metabolites with other metrics of smoking behaviuor (current versus former, former versus never, smoking duration and years since quitting) in the US study.

RESULTS: We identified a total of 25 metabolites associated with smoking behaviours; 24 were associated with current smoking status and eight with cig/day. In addition to three well-established nicotine metabolites (cotinine, hydroxycotinine, cotinine N-oxide), we found an additional 12 xenobiotic metabolites involved in benzoatic (e.g. 3-ethylphenylsulphate) or xanthine metabolism (e.g. 1-methylurate), three amino acids (o-cresol sulphate, serotonin, indolepropionate), two lipids (scyllo-inositol, pregnenolone sulphate), four vitamins or cofactors [e.g. bilirubin (Z,Z)], and one carbohydrate (oxalate).

CONCLUSIONS: We identified associations between cigarette smoking and a diverse range of metabolites. Our findings, with further validation in future studies, have implications regarding aetiology and study design of smoking-related diseases.