Nicotine Exposure in the U.S. Population: Total Urinary Nicotine Biomarkers in NHANES 2015–2016

We characterize nicotine exposure in the U.S. population by measuring urinary nicotine and its major (cotinine, trans-3′-hydroxycotinine) and minor (nicotine 1′-oxide, cotinine N-oxide, and 1-(3-pyridyl)-1-butanol-4-carboxylic acid, nornicotine) metabolites in participants from the 2015–2016 National Health and Nutrition Examination Survey. This is one of the first U.S. population-based urinary nicotine biomarker reports using the derived total nicotine equivalents (i.e., TNEs) to characterize exposure. Serum cotinine data is used to stratify tobacco non-users with no detectable serum cotinine (−sCOT), non-users with detectable serum cotinine (+sCOT), and individuals who use tobacco (users). The molar concentration sum of cotinine and trans-3′-hydroxycotinine was calculated to derive the TNE2 for non-users. Additionally, for users, the molar concentration sum of nicotine and TNE2 was calculated to derive the TNE3, and the molar concentration sum of the minor metabolites and TNE3 was calculated to derive the TNE7. Sample-weighted summary statistics are reported. We also generated multiple linear regression models to analyze the association between biomarker concentrations and tobacco use status, after adjusting for select demographic factors. We found TNE7 is positively correlated with TNE3 and TNE2 (r = 0.99 and 0.98, respectively), and TNE3 is positively correlated with TNE2 (r = 0.98). The mean TNE2 concentration was elevated for the +sCOT compared with the −sCOT group (0.0143 [0.0120, 0.0172] µmol/g creatinine and 0.00188 [0.00172, 0.00205] µmol/g creatinine, respectively), and highest among users (33.5 [29.6, 37.9] µmol/g creatinine). Non-daily tobacco use was associated with 50% lower TNE7 concentrations (p < 0.0001) compared with daily use. In this report, we show tobacco use frequency and passive exposure to nicotine are important sources of nicotine exposure. Furthermore, this report provides more information on non-users than a serum biomarker report, which underscores the value of urinary nicotine biomarkers in extending the range of trace-level exposures that can be characterized.


Introduction
Tobacco use is the leading preventable cause of disease, disability, and death in the United States. Each year, more than 400,000 deaths are attributed to cigarette smoking and exposure to secondhand smoke (SHS) [1]. The overall cigarette smoking rate has declined from 20.9% in 2005 to 15.5% in 2016; however, in 2016, nearly 38 million American adults continued to smoke cigarettes every day or some days [2]. Moreover, there are large

Study Design
The NHANES survey has been conducted by the National Center for Health Statistics (NCHS), a division of the U.S. Centers for Disease Control and Prevention (CDC), periodically since 1971 and continuously in two-year cycles since 1999. NHANES is a program of cross-sectional studies designed to assess the health and nutritional status of non-institutionalized U.S. civilians based on data collected from questionnaires, physical ex-aminations, and biological samples [12]. The NCHS Research Ethics Review Board reviewed and approved the study (NCHS ERB Protocol #2011-17). Participants aged ≥18 provided informed written consent before taking part in the study. Participants < 18 years obtained parental permission, and documented assent for children and adolescents aged 7-17 was required, before taking part in the study. We measured nicotine and its metabolites in one-third of the spot urine samples from participants aged ≥6 years (NHANES datasets UCOT_I, COT_I; n = 3279). As laboratory examination components are carried out on a subsample of NHANES participants, NHANES 2015-2016 participants for some, but not all, ages were selected to provide urine samples for testing for nicotine metabolites. Each subsample is selected to be a nationally representative sample of the target population and has its own designated sample weight that accounts for the additional probability of selection into the subsample component. The results reported here are from a subset of these participants (n = 2281) remaining after applying eligibility criteria and discarding records with incomplete data.

Chemical Analysis
Nicotine biomarkers were measured by one of two separate isotope dilution liquid chromatography tandem mass spectrometry (LC-MS/MS) methods. We measured NIC and its six metabolites in urine samples with a total COT concentration of ≥20 µg/L ("high samples") [13], and for urine samples with a total COT < 20 µg/L ("low samples"); only COT and HCT were measured [14]. The limit-of-detection (LOD) for NIC and its minor metabolites ranged from 1.38-10.5 µg/L, whereas the LOD for COT and HCT was determined to be 0.030 µg/L for both metabolites. Measurements below the LOD were substituted with the quotient of the LOD divided by the square root of two [15].
Briefly, urine aliquots were fortified with a labeled internal standard mixture and then incubated with beta-glucuronidase enzyme to hydrolyze the conjugated analytes. Samples were extracted and the nicotine biomarkers were measured by high-performance LC-MS/MS using electrospray ionization for high samples or ultra-high-performance LC-MS/MS using atmospheric-pressure chemical ionization for low samples. We monitored one quantitation transition, one confirmation transition, and one corresponding internal standard transition for each analyte quantified. Analyte concentrations were derived from the ratios of native-to-labeled compounds in the sample by comparing to a standard curve. Reported results met the accuracy and precision specifications of the quality control and quality assurance programs of CDC's National Center for Environmental Health, Division of Laboratory Sciences [16].

Data Attrition and General Description of Dataset
Scheme 1 provides a summary of the data attrition process. Briefly, a total of 3321 participants were examined, of which 42 had provided no laboratory results. An additional 308 participants with missing urinary COT and HCT results and 330 with missing serum COT results were excluded from further data analysis. Participants with missing demographics information (n = 310) were also excluded, leaving a total of 2290 participant for additional attrition steps.
To distinguish non-users from those who use tobacco, we used a serum COT threshold of >10 µg/L, which has been identified as consistent with the active use of combusted cigarette products [17]. Among samples with serum COT ≤ 10 µg/L, those with serum COT less than or equal to the reported LOD (0.015 µg/L) were categorized as non-users with undetectable serum COT ("−sCOT"), and those with serum COT within 0.015 < x ≤ 10 µg/L were categorized as non-users with detectable serum COT ("+sCOT") [18]. Responses from the NHANES questionnaire set, "Smoking-Recent Tobacco Use" (SMQRTU_I), were used to further categorize recent (within the past five days), daily, and non-daily tobacco users. Daily users are participants with serum COT > 10 µg/L who had reported using tobacco (at least one product type) daily within the past five days. Non-daily users are participants with serum COT > 10 µg/L who had reported using tobacco (any one product, or a combi-Scheme 1. Data attrition and general description of non-user and user populations, in NHANES 2015-2016 (n = 2281). COT = cotinine; HCT = trans-3′-hydroxycotinine; NRT = nicotine replacement therapy; −sCOT = non-users with undetectable serum COT; +sCOT = non-users with detectable serum COT. a : Samples with urinary creatinine concentrations outside of the 10-370 mg/dL range indicated excessively diluted or concentrated (in vivo) urine samples. b : Users of NRT products were excluded from the analysis if participants indicated "yes" to the NHANES question SMQ863, within the SMQRTU_I questionnaire set. c : Seven participants under the age of 17 were excluded due to the small sample size of this age group. For the user population, the steps taken to categorize daily and non-daily users have been provided in text.
To distinguish non-users from those who use tobacco, we used a serum COT threshold of >10 µg/L, which has been identified as consistent with the active use of combusted cigarette products [17]. Among samples with serum COT ≤ 10 µg/L, those with serum COT less than or equal to the reported LOD (0.015 µg/L) were categorized as non-users with undetectable serum COT ("−sCOT"), and those with serum COT within 0.015 < x ≤ 10 µg/L were categorized as non-users with detectable serum COT ("+sCOT") [18]. Responses from the NHANES questionnaire set, "Smoking-Recent Tobacco Use" (SMQRTU_I), were used to further categorize recent (within the past five days), daily, and non-daily tobacco users. Daily users are participants with serum COT > 10 µg/L who had reported using tobacco (at least one product type) daily within the past five days. Nondaily users are participants with serum COT > 10 µg/L who had reported using tobacco (any one product, or a combination of multiple products) for at least one day and up to four days, within the past five days. Within the SMQRTU_I dataset, the following NHANES questions for product usage frequency were used to categorize recent daily and non-daily users-SMQ710 (cigarettes), SMQ740 (pipes), SMQ770 (cigars), SMQ845 (hookah/water pipes), SMQ849 (e-cigarettes), SMQ800 (chewing tobacco), SMQ817 (snuff), SMQ857 (snus) and SMQ861 (dissolvables). Scheme 1. Data attrition and general description of non-user and user populations, in NHANES 2015-2016 (n = 2281). COT = cotinine; HCT = trans-3 -hydroxycotinine; NRT = nicotine replacement therapy; −sCOT = non-users with undetectable serum COT; +sCOT = non-users with detectable serum COT. a : Samples with urinary creatinine concentrations outside of the 10-370 mg/dL range indicated excessively diluted or concentrated (in vivo) urine samples. b : Users of NRT products were excluded from the analysis if participants indicated "yes" to the NHANES question SMQ863, within the SMQRTU_I questionnaire set. c : Seven participants under the age of 17 were excluded due to the small sample size of this age group. For the user population, the steps taken to categorize daily and non-daily users have been provided in text. Table 1 shows the sample sizes and sample-weighted distributions for demographic groups stratified by tobacco use status for the 2281 participants included in this study. Self-reported information on sex, age, race/Hispanic origin, and education was collected by interview. Race/Hispanic origin was categorized as "non-Hispanic White", "non-Hispanic Black", "Hispanic" (participants identifying as "other Hispanic" or "Mexican American"), and "Other/Multiracial" (participants identifying as "non-Hispanic Asian", "other race", or "multiracial"). Age, in years, was divided into 18-29, 30-44, 45-59, and ≥60 for non-users and users; age categories of 6-11 and 12-17 were included for non-users only. Education attainment was defined based on the highest level of education completed, and categorized as "less than high school", "high school graduate", "some college (no degree)", and "Bachelor's degree or above". The weighted urinary COT detection rates among each population sub-group were calculated as the percentage of measured analyte concentrations at or above the LOD. The COT detection rates are indicated in the superscript for each population sub-group. −sCOT = non-users with undetectable serum COT; +sCOT = non-users with detectable serum COT. Detection rates were calculated using weighted, urinary cotinine concentrations-1 : 96%; 2 : 94%; 3 : 99%; 4 : 100%. *: Age categories of 6-11 and 12-17 were included for non-users only.

Statistical Analysis
NHANES recruited participants through a multistage, probability sampling design involving selection of primary sampling units in counties, households in the counties, and sample patients in selected households [19]. Using this dataset, we calculated nationally representative summary statistics with appropriate variance estimates and investigated the associations of select demographic factors on nicotine exposure levels by applying survey sample weights (NHANES Subsample A Weight, WTSA2YR) and using Taylor series linearization for variance estimation. We used this estimation approach as it was implemented in the SURVEYFREQ, SURVEYMEANS, and SURVEYREG subroutines of the SAS ® statistical software application version 9.4 (SAS Institute, Cary, NC, USA). An evaluation of statistical reliability was performed to ensure all proportions followed NCHS Data Presentation Standards [20].
The Pearson correlation coefficients (r) and their p-values were computed between COT, HCT, NIC, TNE2, TNE3, and TNE7, where statistical significance was set to α ≤ 0.05. Pearson correlation coefficients were calculated from the log-transformed (base 10) biomarker and TNE concentrations without using sample-weights or adjusting the data for creatinine.
Creatinine concentration data were used to normalize the concentrations of nicotine exposure biomarkers to account for urine volume variability and the variability in concentrations of endogenous and exogenous chemicals [21]. Summary statistics, including sample-weighted geometric means (GM) of biomarkers and TNEs, along with their 95% confidence intervals (CI), are reported as a ratio of creatinine (µg/g creatinine, or µmol/g creatinine) in the main tables and volume-weighted concentrations (µg/L, or µmol/L) in the Supplementary Tables S1 and S2.
Sample-weighted multiple linear regression models stratified by tobacco use status were fitted to data from the NHANES 2015-2016 cycle, where the dependent factors were the creatinine-unadjusted concentrations of COT, HCT, TNE2, TNE3, and TNE7, and the independent factors included both continuous (creatinine, g/L) and categorical types (i.e., sex, age, race/Hispanic origin, education attainment, extent of passive nicotine exposure among non-users, and tobacco use frequency). Because the distribution of biomarker measurements was highly right-skewed-which would have adversely affected hypothesis testing-urinary creatinine, COT, HCT, and TNEs concentration data were log-transformed (base 10) to enable evaluation of the statistical significance of regression coefficients. We report the exponentiated coefficients from these models along with their 95% CIs and p-values, where statistical significance was set to α ≤ 0.05. The exponentiated coefficients represent the proportional change of biomarker concentration associated with an independent categorical or continuous predictor. To interpret the categorical factors in the model, the associated percentage difference in biomarker concentration was calculated as the exponentiated coefficient minus 1 and then multiplied by 100.
For the regression models, we accounted for urinary dilution by including urinary creatinine as an independent factor. Among the other independent factors included in the model, we used the following categories as reference groups: males for sex, "45-59" years for age group, "non-Hispanic White" for race/Hispanic origin, "Bachelor's degree or above" for education attainment group, "+sCOT" for non-user sub-group and "daily user" for user sub-group. Regarding our regression analyses on the education attainment group, to ensure that none of the younger participants in the main analysis were misclassified due to being "too young" to have attained their highest degree at the time of the survey, we performed a sub-analysis of adults aged 25 or older in separate sample-weighted log-linear regression models after stratifying for tobacco use status (Supplementary Tables S5 and S6). We found no difference in significance of education attainment between our main analyses and the sub-analyses. We also performed pairwise comparisons of least-square means from the regressions, among different demographic groups, for both the non-user and user populations (Supplementary Tables S3 and S4). To correct for multiple comparisons, we adjusted the p-values from pairwise comparisons by the Bonferroni method.

Correlation of Nicotine Biomarkers and TNEs
Correlation plots were generated to determine the strength of associations between nicotine and its major metabolites and the TNEs (Figure 1). Among users, TNE7 positively correlated with TNE3 and TNE2 (r = 0.99 and 0.98, respectively), and TNE3 positively correlated with TNE2 (r = 0.98). COT and HCT concentrations were strongly correlated in both user and non-user population sub-groups, with a higher degree of correlation recorded among the non-users (r = 0.94 vs. r = 0.81). TNE7 and TNE3 correlated very well with COT and HCT (r ≥ 0.91), though a slightly stronger correlation was found with COT (r = 0.94 and 0.93, respectively). TNE2 was strongly correlated to COT and HCT (r ≥ 0.92) within either population sub-groups, where we found a slightly higher degree of correlation between HCT and TNE2 (r = 0.97 and 0.99 among individuals who use tobacco and non-users, respectively). for multiple comparisons, we adjusted the p-values from pairwise comparisons by the Bonferroni method.

Correlation of Nicotine Biomarkers and TNEs
Correlation plots were generated to determine the strength of associations between nicotine and its major metabolites and the TNEs (Figure 1). Among users, TNE7 positively correlated with TNE3 and TNE2 (r = 0.99 and 0.98, respectively), and TNE3 positively correlated with TNE2 (r = 0.98). COT and HCT concentrations were strongly correlated in both user and non-user population sub-groups, with a higher degree of correlation recorded among the non-users (r = 0.94 vs. r = 0.81). TNE7 and TNE3 correlated very well with COT and HCT (r ≥ 0.91), though a slightly stronger correlation was found with COT (r = 0.94 and 0.93, respectively). TNE2 was strongly correlated to COT and HCT (r ≥ 0.92) within either population sub-groups, where we found a slightly higher degree of correlation between HCT and TNE2 (r = 0.97 and 0.99 among individuals who use tobacco and non-users, respectively).

Factors Influencing Nicotine Exposure in the U.S. Population
Results in this sub-section are from multiple linear regressions of logarithmic COT, HCT, or TNEs on non-user and user sub-groups, controlled for urinary creatinine and the demographic factors sex, age, race/Hispanic origin, and education attainment. The results presented in Tables 4 and 5 were obtained after including all participants aged ≥6 years for non-users and ≥18 years for individuals who use tobacco, respectively. In all regression models, urinary creatinine had a small but statistically significant (p < 0.0001) association with the individual biomarker and TNE concentrations.

Discussion
We measured COT, HCT, and TNEs in urine samples from a one-third subset of NHANES 2015-2016 cycle, aged ≥6. Our regression models show that tobacco use frequency and passive exposure to nicotine are important sources of nicotine exposure in the U.S. population. After controlling for the extent of passive nicotine exposure, creatinine, and other demographic factors, we find that non-Hispanic Whites tended to have higher urinary TNE2 than other race/ethnicities, with this difference reaching statistical significance for Hispanics. Furthermore, education attainment was inversely associated with urinary TNE2 levels: attaining a Bachelor's (or higher) degree was associated with less nicotine exposure than that found in people with lower educational attainment. Among individuals who use tobacco, demographic factors were also evaluated for association with nicotine exposure in the sample-weighted multiple linear regression models. After controlling for tobacco use frequency, creatinine, and other demographic factors, we find that non-Hispanic Whites had significantly higher urinary TNE7 compared with other race/ethnicities. Additionally, urinary TNE7 was higher in the 45-59 age group compared with any of the other age groups.
A key strength of our study was the use of measured concentrations of multiple, wellestablished nicotine metabolites, rather than the use of a single metabolite or product use questions alone. Another major strength of this study is the use of biochemical verification of tobacco use status. We confirmed passive nicotine exposure status using participants' serum COT concentrations, where we were able to compare the urinary COT, HCT, and TNE2 concentrations between −sCOT and +sCOT and provide reference ranges for the biomarkers among the two population sub-groups. We noted good agreement between the COT and TNE2 concentrations in our population estimates and regression analysis, which further supports the utility of either one of these biomarkers for monitoring and assessing exposure levels within the two population sub-groups. Recent active use of tobacco products was confirmed by the serum COT > 10 µg/L cutoff and responses to the product use survey questionnaire. By accounting for questionnaire responses regarding past-five-day use of single or multiple tobacco products, along with participant serum COT concentrations, we were also able to remove some degree of uncertainty in identifying non-daily users. The past-five-day responses were deemed more appropriate than responses from the past 30 days usage questionnaire for identifying non-daily users because the metabolite half-lives are relatively short. Moreover, as this study used nationally representative data, our results provide reliable measures of nicotine exposure among the U.S. population.
Within non-users, we see quantitative differences in TNE2 GMs among the two population sub-groups and some of the demographic groups. As anticipated, the TNE2 GM for the +sCOT sub-group was elevated when compared to −sCOT. Within +sCOT, higher TNE2 among those aged 6-11 years than those aged ≥12 years could result from a larger proportion of younger children being exposed to higher levels of SHS and/or SHA than youths and adults. For anyone living with one or more family members who smoke tobacco, one can expect children, who are generally spending more time within their homes, to have a greater propensity for exposure to SHS and/or SHA [22,23]. Within the same population sub-group, individuals with lower education attainment tended to have higher TNE2, whereas those with higher education attainment had lower biomarker concentrations. A possible explanation for such exposure pattern could be that people with lower education attainment may be less aware of the health hazards of smoking, SHS, and SHA, and thus have a greater propensity for exposure. It may be of interest to further stratify this population sub-group by perceived SHS or SHA exposure to note any substantial differences in biomarker concentrations and track such information across multiple NHANES cycles.
The GMs of urinary nicotine metabolites and TNEs varied by age, sex, race/Hispanic origin and education attainment among individuals who use tobacco, where the exposure patterns were generally consistent with those from previous studies. Advanced age was associated with higher nicotine exposure, and we generally noted similar patterns in exposure across the different age groups when using either TNEs or COT. An overall pattern of increasing biomarker concentration by age was also reported in other studies [24,25].
Possible explanations for such a pattern may include differences in tobacco use frequency, differences in the intensity in smoking behavior, and/or lower representation of light users among the older age groups. Creatinine adjustment has a well-known impact on sex differences, as males, on average, have~30% higher urinary creatinine concentrations than females [26]. As such, creatinine adjustment has a predictable influence on the reported GMs, as shown by the 3-15% higher nicotine metabolite and TNE concentrations among female users. When categorizing by race/Hispanic origin, our study extends the literature-in which, serum COT was generally used for biomonitoring [27,28], followed by TNEs [24]-by presenting users identifying as Hispanic or Mexican American to have lower concentrations of urinary biomarkers than non-Hispanic Whites and non-Hispanic Blacks. We also find that non-Hispanic Whites had higher urinary nicotine biomarkers compared with other race/Hispanic origin groups, perhaps because the tobacco use group included people who use smokeless tobacco. Smokeless tobacco use is associated with higher nicotine exposure compared with other tobacco products [24,[29][30][31], and smokeless tobacco users are disproportionately non-Hispanic White [29,30]. Higher education attainment among individuals who use tobacco was not associated with lower nicotine exposure in the weighted multiple linear regression models; however, we note an overall pattern in decreasing exposure levels among the higher education attainment groups, which generally followed the pattern noted in previous studies using serum COT [27] and urinary TNEs [24].
Several recent studies have characterized non-Hispanic Black cigarette smokers as having higher exposure levels than non-Hispanic Whites [27,28,32,33] when comparing their serum COT or urinary COT (unconjugated) concentrations. Other studies that accounted for total urinary NIC and its two major metabolites generally found higher exposure levels among non-Hispanic White tobacco users [24] and exclusive cigarette users [25] compared with non-Hispanic Black users, which is consistent with the results presented in this report. The above observations may potentially reflect a difference in the type of biomarker used to characterize the population exposure levels (i.e., unconjugated vs. total measurements), rather than any inherent differences between the study type. In addition, non-Hispanic Blacks are reported to have lower COT (and NIC) glucuronidation rates compared to non-Hispanic Whites [34,35], which may explain the higher serum COT concentrationsand further explain the lower urinary COT concentrations-reported for non-Hispanic Blacks compared to non-Hispanic Whites. Because substantial racial/ethnic differences are observed in nicotine metabolism [36][37][38]-which may be influenced by both genetic and environmental factors [39,40]-the use of a nicotine metabolite that is susceptible to metabolic-related differences may not be comprehensive in characterizing exposure levels within large, population-wide studies.
We analyzed nicotine metabolites in both urine and serum collected from the same NHANES study participants. COT in these two matrices was highly correlated, which is consistent with urine/blood measure correlations in other studies (r = 0.69 to 0.91, p < 0.05) [41][42][43]. Importantly, COT concentrations are typically 4-to 5-fold higher in urine compared with blood plasma or serum [42]; the difference between urine and blood matrices is mostly attributable to renal clearance processes [9]. These findings underscore the value of urinary nicotine biomarkers in extending the range of trace-level nicotine exposures that can be characterized. In our current dataset, for example, COT was detected in 94% of non-user urine samples collected from study participants with serum COT that was ≤0.015 µg/L. The high COT detection rate obtained for the urine samples analyzed would suggest persistent exposure to nicotine in the U.S. population; however, such an observation would not be as apparent when comparing the serum detection rate from the same population sub-group.
Among urinary biomarkers of nicotine exposure, urinary TNEs are better suited for characterizing nicotine exposure than any single metabolite, such as COT. Our analysis showed strong TNE2-COT and TNE2-HCT correlations (r ≥ 0.92), among both the nonuser and user populations, and very strong correlations between TNE2, TNE3, and TNE7 (r ≥ 0.98) when surveying the user population. Among non-users, TNE2 could estimate nicotine exposure reasonably well, as the molar sum of COT and HCT typically accounts for~70% of the total nicotine dose [9]. TNE2 may also be sufficient for estimating active usage, as it is strongly correlated to both TNE3 and TNE7. TNE3 or TNE7 would be good biomarkers to estimate exposure among individuals who use tobacco; however, TNE7 would provide the best estimate of recent nicotine exposure as it accounts for~85% to 90% of the total nicotine dose [11]. Overall, the use of TNEs may provide a more suitable assessment of nicotine exposure because these measurements are not significantly affected by individual differences in metabolism.
Some of the constraints in the current analysis included the limited sample size obtained for individuals who use tobacco after applying all necessary criteria for data analyses, and limited sample size obtained for users of tobacco products other than cigarettes. In addition, by using the serum COT cutoff to categorize the non-user and user populations, we introduced the potential for a portion of the self-reported non-daily users to be classified as non-users. Furthermore, misclassification resulting from misreporting of information in self-reported questionnaire responses is likely. Lastly, we measured nicotine exposure biomarkers with relatively short half-lives (15-20 h for the major metabolites) in a spot urine sample, and thus introduce a degree of imprecision for non-daily users due to potential variations in the time elapsed between last tobacco use and urine collection.

Conclusions
We characterized nicotine exposure among individuals who use tobacco and non-users with passive exposure to nicotine in a representative sample of the U.S. population, based on data collected from the NHANES 2015-2016 study cycle. This paper provides pertinent biomonitoring data to assess public health risk and identify population sub-groups that are at a higher risk of being exposed to tobacco. Our current analysis documents important differences in nicotine exposure and shows that, along with certain demographic factors such as age, race/Hispanic origin, and education attainment, tobacco use frequency and passive exposure to nicotine are major contributors to increased nicotine exposure. These data provide a crucial baseline against which future analyses of urinary nicotine biomarkers can be compared to document variations caused by changes in tobacco products, use behaviors, and/or policies/regulations.