We first studied 583 men and women of European, Japanese or Native Hawaiian ancestry who were long-term smokers of >10 cigarettes day. Participants were randomly selected among members of the Multiethnic Cohort Study (88%) or controls of several completed population-based case-control studies (12%) living on Oahu, Hawaii (6
). Other inclusion criteria included having no previous history of invasive cancer, having both parents of Japanese or European ethnicity, or of any amount of Native Hawaiian ancestry, and smoking at least 10 cigarettes per day. These individuals were re-contacted for this study and instructed on how to record their food consumption for three days, as well as to collect a 12-hour, overnight urine sample at the end of those 3 days. A blood sample was then collected and a short questionnaire (including tobacco use during the previous three days) administered. The overall target sample size for this study was 100 in each sex and ethnic group. A total of 596 participants completed all aspects of the study, corresponding to a participation rate of 64.4%. Eight subjects were excluded for reporting to smoke fewer than the required 10 cigarettes per day during data collection, and 5 were excluded for missing covariate.
Subjects were genotyped for the synonymous variant rs1051730 in exon 5 of CHRNA3
and the non-synonymous variant rs16969968 in CHRNA5
, the candidate SNPs the most likely to be causal in the 15q region associated with lung cancer in the three published GWAS (3
). DNA was extracted from lymphocytes and the two variants were genotyped with the TaqMan allele discrimination assay (Applied Biosystems, Foster City, CA). The genotype frequencies were consistent with Hardy Weinberg equilibrium in each ethnic group (p>0.05). Concordance rate across the ~10% blinded duplicate samples genotyped with the study samples was 100%. The genotyping call rate was 99%. The frequency of the T allele for rs1051730 was 0.34 in European Americans, 0.19 in Native Hawaiians and 0.03 in Japanese Americans. The corresponding frequencies of the A allele for rs16969968 were 0.34, 0.20 and 0.03.
To access total nicotine, and hence tobacco smoke exposure, the urinary concentration of the sum of nicotine and 5 nicotine metabolites was determined. The sum of these 6 compounds, which account for 75−95% of the nicotine dose is referred to as nicotine equivalents and is considered the most comprehensive measure of exposure to nicotine (11
). Specifically we measured the molar concentrations of total nicotine (free plus nicotine N
-glucuronide), total cotinine (free plus cotinine N
-glucuronide), total trans
-3′-hydroxycotinine (3-HC) (free plus 3-HC N
- and O
-glucuronide), as well as a metabolite of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) and its detoxification product NNAL-glucuronide (NNAL-Gluc), in the 12-hour urine. Nicotine metabolites were measured by gas chromatography-mass spectrometry (GC-MS), and nicotine equivalents calculated as the sum of the molarity of total nicotine, total cotinine and total 3-HC. NNK metabolites were measured by GC with nitrosamine selective detection (GC-TEA) (9
). Based on 65 blind duplicate pairs analyzed with the study samples for total nicotine and total cotinine, and 6 pairs for total 3HC, the intraclass correlation coefficient was 0.98, 0.96 and 0.62, respectively. NNK metabolites were determined as described with slight modifications (12
We sought to replicate findings from the Hawaii study in two studies conducted at the University of Minnesota (UMN) that collected first morning urine. UMN Study 1 (UMN-1) included 99 participants (only 2 of which were non-white) in a clinical trial [the Tobacco Reduction Intervention Program (TRIP) Study] which recruited smokers of more than 14 cigarette/day and aged 18−70 years, who were interested in reducing cigarette use in the next 30 days (13
). UMN Study 2 (UMN-2) included 137 smokers (118 European Americans, 15 African Americans and 4 Asian/Pacific Islanders). Smokers of 10−40 “light” cigarettes (0.7−1.0 mg nicotine/cigarette)/day, aged 18−70, who were interested in quitting smoking were recruited via advertisement. To be eligible for both UMN studies, smokers had to: a) be in good physical health; b) have no contraindication to nicotine replacement therapy; c) be in good psychiatric health; d) not be using other tobacco products; and e) not be pregnant or nursing. In both studies, the average number of cigarettes per day was computed from a daily diary and a first morning urine sample was collected from each subject. In these studies, the frequency of the rs16969968 A allele was 0.35 in European Americans and 0.10 in African Americans. The genotype frequencies were consistent with Hardy Weinberg equilibrium (African Americans, p>0.05; Caucasians, p=0.01). Concordance rates across the ~15% blinded duplicate samples genotyped were 97−100% and the genotyping call rate was 100%.
A square-root transformation was used for nicotine equivalents to achieve normal distribution. The sum of NNK metabolites was log-transformed. Least-square means for these two variables were computed for each genotype using the general linear model (GLM) procedure in SAS 9.0 (SAS, Inc. Cary, NC) adjusting for age, sex, race/ethnicity, cigarettes per day and body mass index, and further for nicotine equivalents. In the pooled-analysis, an adjustment variable for study was included. Tests of interaction were conducted between genotype and study to identify modifying effects, but were not statistically significant and, therefore, were not included in the final models. Tests for significance were two-tailed with an alpha level of 0.05.