Article Text
PDF
Abstract
Introduction American Indians and Alaska Natives face disproportionately high rates of smoking and secondhand smoke (SHS) exposure. The Cheyenne River Sioux Tribe (CRST) is among the few Tribal Nations controlling commercial tobacco exposures in public and work places. We had an opportunity to explore effects of the new commercial tobacco-free policy (implemented in 2015) in an environmental health study (2014–2016) that collected information about commercial tobacco use and SHS prevalence and examined predictor variables of serum cotinine concentrations.
Methods Self-reported survey data were used in quantile regression statistical modelling to explore changes in cotinine levels, based on smoking status, smokeless tobacco consumption and SHS exposure.
Results From enrolled 225 adults, 51% (N=114) were current smokers. Among 88 non-tobacco users, 35 (40%) reported current SHS exposure. Significant differences in cotinine median concentrations were found among participants with and without current SHS exposure. Extremely high cotinine concentrations (~100 times larger than the median) were detected in some non-tobacco users. After implementing the new smoke-free air Tribal policy, cotinine decreased in participants with intermediate (3–15 ng/mL, non-tobacco users with SHS exposure) and high (>15 ng/mL, mainly tobacco users) cotinine levels showing association with an abatement of opportunities for SHS exposure. Significant predictors of cotinine levels were sampling year, current smoking and tobacco chewing. No gender differences were observed in cotinine.
Conclusions Our results show decrease in cotinine concentrations in CRST participants since implementation of their ‘Smoke-Free Clean Air Act’ in 2015.
- public policy
- cotinine
- disparities
- secondhand smoke
Statistics from Altmetric.com
Introduction
A declining trend in adult commercial cigarette smoking has been demonstrated in the USA across diverse ethnic groups.1–3 However, similar trends are not observed among American Indians and Alaska Natives (AI/ANs). Differences in commercial cigarette smoking prevalence4 and early initiation of smoking5 were found even across AI/AN Tribes. A prior study confirmed that Northern Plain Indians had higher smoking prevalence than Southwestern Tribes.6 Tribal communities also consume traditional tobacco as part of their ceremonies. Most native plants have increasingly been replaced in ceremonies and cultural settings by commercial tobacco products,4 5 therefore ceremonial practices can also influence commercial tobacco exposure.4 Furthermore, over the past decades, smoking rates have been rising in some Tribal communities that had historically low rates.4 7 This trend parallels increases in smoking-related diseases, such as lung cancer, cardiovascular and respiratory problems.8 9 Moreover, among Northern Plains American Indians lung cancer rates are the highest in the USA.10 In addition, cigarette smoking is shown to have many other adverse health effects,11 and mainstream cigarette smoke contributes exposure to mercury, cadmium and arsenic.12
In an effort to address the dangers of tobacco smoke exposure, many AI/AN Tribes are passing smoke-free policies on Tribal lands.13–19 This has been the case with the Cheyenne River Sioux Tribe (CRST) of South Dakota, in the Northern Plains. The CRST Canli Coalition—an anti-tobacco coalition—organised a public health campaign against cigarette smoking through dedicated efforts for over 6 years. This led to successful pass of the Smoke-Free Clean Air Ordinance #77, implemented by CRST in May 2015, that comprehensively banned smoking commercial tobacco in public spaces to reduce the high levels of tobacco smoke exposure.20 21 Literature has identified significant health benefits following implementation of tobacco policies such as reduction of risk for myocardial infarctions.22 However, the association between policy enactment and changes in cotinine concentrations has not been evaluated in Tribal communities. The aim of this analysis was to examine the impact of the new commercial tobacco-free policy on cotinine concentrations in the community.
Cotinine measurement quantifies commercial tobacco smoke exposure.23 The half-life of cotinine is ~16 hours, making it a reliable biomarker of tobacco smoke exposure in both smokers and non-smokers.24 Cotinine measurements have been strongly associated with self-reported smoking and with lung cancer risk in active smokers.25 26 Established cotinine thresholds discriminate active from passive tobacco smoke exposure. Cotinine levels >15 ng/mL were used to infer active smoking.27 Others suggested a more conservative cut-point of 3 ng/mL.28 Both thresholds were included in this analysis as appropriate in AI/AN communities, given the absence of smoke-free policies on most Tribal lands.
Our 9-year community-based participatory collaboration with the CRST examined exposures to various mine waste metals and associated health effects. Over 900 gold and silver mines had been operated for several decades within the CRST watershed. The amalgamation production process released inorganic mercury into the Cheyenne River, the CRST’s only surface water source. A dozen years after the mercury contamination warnings were placed by United States Environmental Protection Agency, the Tribal environmental protection agency identified another contaminant, arsenic, in the river sediment as well.29 The focus of the study was on fishing and fish consumption as culturally significant activities exposing community members to mercury and arsenic toxicants. Since cigarette smoke can also be a source of metal exposures, self-reported smoking and secondhand smoke (SHS) exposures were assessed using a short smoking survey. To establish validity of this questionnaire, we used cotinine measurements. This relationship was confirmed both before and after the implementation of ‘Ordinance #77—The CRST Smoke-Free Clean Air Act of 2015’ on the CRST Reservation.
Methods
Population and sample collection
The study enrolled 225 adults, who provided written informed consent, in the summers of 2014 (prior to the ordinance) and 2015 and 2016 (post ordinance). Participants were living on the CRST Reservation in the Northcentral part of South Dakota. Wildlife is an important dietary source for this community. Approximately 7000 adults live on the Reservation, but more than 5000 hunting and fishing licenses are sold annually.30 We recruited anglers, fishermen, and others who participated in outdoor activities that brought them into contact with the Cheyenne River and metal toxicants. We used community centres, radio announcements, flyers and public events in the rural areas to publicise enrolment.
Each participant was interviewed using fishing, land-use, and smoking surveys, and provided a blood sample for laboratory analysis. Participants were reimbursed for their time and for donating biological samples. The smoking survey is available (see online supplementary material). Participants’ age, gender and community location were recorded. The Reservation’s geographical area was subdivided into exposure zones based on the predominant environmental toxicants.
Supplemental material
Laboratory analysis
Serum samples were obtained at enrolment by venipuncture and stored in a −80°C freezer until laboratory use. Competitive ELISA assay was carried out to measure serum cotinine concentrations (ng/mL) following the manufacturer’s instructions (Calbiotech Inc., El Cajon, California, USA). As the primary and stable metabolite of nicotine, cotinine is used to measure nicotine absorption (within ~16 hours) and metabolism by the body.23 24 Cotinine is an exposure biomarker of active smoking (>15 ng/mL) and SHS (3–15 ng/mL) exposure in population-based studies,6 28 although the cut-off values vary.31
Data collection and statistical analysis
Participants’ current smoking and tobacco chewing were assessed based on answers to the survey questions (see online supplementary material). The survey also assessed the number of cigarettes smoked daily. The effects of cigarette dose on cotinine were examined. Dual tobacco use was defined as current smoking and chewing smokeless tobacco products. Current SHS exposures at home, in the workplace, and during leisure time activities were combined to create a composite SHS variable (CoSHS). In addition, childhood home SHS was recorded as a historical measure of exposure. Former smoking and tobacco chewing measures were not included.
For the main analysis, survey information was coded to create binary tobacco exposure variables (yes/no) for current smokers (figure 1F) and chewers (figure 1D). The amount of smoking both at the personal (online supplementary figure S2) and in-home SHS levels was considered in additional analysis (see online supplementary material).
Quantile regression modelling: selected estimated parameters by quantile for serum cotinine natural log (with 95% CI).
In order to examine the influence of age on cotinine concentrations, participant’s age was used as a continuous variable. In addition, participants were stratified into two groups: <42 and ≥42 years old. The threshold of 42 years was based on the mean age ±SD (41.8±13.4 years) of study sample. This was applied to capture susceptible age groups to smoking. χ2 and Fisher’s exact tests were used to examine proportional differences between groups in the sample. The generated p values of these tests are shown in the text. To assess whether cotinine levels of one group stochastically dominate levels of the other group(s), the Wilcoxon rank-sum or the Kruskal-Wallis test was employed.
Quantile regression provides a combined set of results for multiple linear regression analyses that are calculated at each percentile (or quantile) of the outcome. The quantile regression modelling (QRM) allows visualising the effects of each predictor at multiple points of the outcome.32–35 These effects can vary and may even cancel each other throughout the distribution of the outcome variable; this method helps to see the entirety of effects. A summary of quantile regression methodology is available elsewhere.33
In this paper, the natural log of serum cotinine was the outcome variable in modelling. It had a skewed distribution, with some extreme high concentrations detected. QRM was used, because the method is little affected by outliers. Interaction between active smoking and tobacco chewing was also included as a predictor variable of dual tobacco use. Selected QRM plots are provided for visual examination. The following three areas are marked on the plots using literature information on cotinine concentration values as thresholds27 28: low cotinine (<3 ng/mL, non-tobacco users without SHS), intermediate cotinine (3–15 ng/mL, non-tobacco users with SHS) and high cotinine levels (>15 ng/mL, mainly tobacco users).
Details on other QRMs that considered age as a continuous variable, or dose of daily cigarette consumption are also submitted in online supplementary material.
Data were stored in REDCap.36 Statistical analyses were performed using SAS V.9.2 and R, V.3.2.2 software (The R project for statistical computing, https://www.r-project.org). All tests were two-sided and the p value of <0.05 was considered for statistical significance. No multiplicity adjustment was carried out, as predictor variables used in the models have documented influence on cotinine values and therefore were not randomly chosen for modelling.
Results
Characteristics of the study subjects are presented in table 1. A large proportion of the participants lived in the vicinity of Eagle Butte, SD, the largest town in CRST Reservation. Those who reported fishing had higher mean and median for cotinine concentrations. These differences, however, were not statistically significant (table 1).
Cheyenne River Sioux Tribe participants’ serum cotinine concentrations by tobacco-related and demographic variables
Current smokers and tobacco chewers had significantly increased cotinine concentrations compared with non-smokers and non-chewers (p<0.0001 and p=0.0457, respectively). Furthermore, participants with CoSHS had higher cotinine concentrations than others without CoSHS (p=0.0357).
There was a significant decrease over the years in overall cotinine levels among all CRST participants (2014–2016, p=0.0062) and also among tobacco users (p<0.0001). Among non-tobacco users, cotinine concentrations decreased in 2015 (p<0.0001) but were higher in 2016 (p<0.0001), with an overall increase from 2014 to 2016 (p=0.0041) (not shown in table 1). The proportion of non-tobacco users with cotinine levels in the active smoking range (>15 ng/mL) decreased over the years (25.9% in 2014, 6.7% in 2015 and 3.2% in 2016, p=0.0268) (not in table 1). Similar trend was observed in the proportion of tobacco users with cotinine >15 ng/mL: 85.4% in 2014, 77.8% in 2015 and 40.9% in 2016, p<0.0001 (not in table 1).
The ‘Tobacco use status’ section of table 1 demonstrated that for non-users, maximum cotinine concentration was about 100 times larger than the median, with or without CoSHS. Cotinine levels of non-tobacco users with CoSHS were not significantly different from levels of those without CoSHS (p=0.4783) (online supplementary table S1). CoSHS exposure contributed to highly right-skewed cotinine concentration distribution among non-tobacco users because 40% of non-smokers had CoSHS exposure (online supplementary table S2). A decreasing, although not statistically significant, trend was detected in the frequency of reported CoSHS over the years: from 66% to 52% (online supplementary table S2). While there were no direct questions in the survey addressing no-indoor smoking rules, 10 non-smokers and 13 smokers had rules at home and four had them at work (not in tables) as were revealed during the interview process.
There was no difference in the proportion of participants with extreme cotinine concentrations based on tobacco use status (4.4% among tobacco users and 4.5% in non-users). There were four non-tobacco users (three in 2014, one in 2015 and 0 in 2016) with extreme cotinine levels (above 44.64 ng/mL, the 95th percentile) and six among tobacco users (three in 2014, three in 2015 and 0 in 2016) were above 544.75 ng/mL, the 95th percentile cotinine concentrations (data not in table 1).
Active smoking stayed high among participants overall (51%, table 1), as well as in each sampling year (57%, 48% and 47%) (online supplementary table S2). A substantial portion (43.9%) of tobacco chewers also smoked cigarettes. As was true for active smokers, the proportions of tobacco chewers and tobacco users were similar across study years. There were more tobacco chewers in the younger (25%) than in the older age group (10%). Also, 8.2% (N=5) of younger smokers (N=61) reported consumption of more than 20 commercial cigarettes per day. However, there were no older smokers with such as high daily cigarette use. While we enrolled a new set of participants across the CRST Reservation in each sampling year, their age group, gender, fishing and smoking status were not significantly different across the years (online supplementary table S2).
Distribution of cotinine levels among CRST participants is presented in figure 2 as a bean plot with a log scaled y axis. Cotinine percentiles and the 3 and 15 ng/mL literature-based thresholds were marked for comparison. As it was shown in the bean plot, almost a half of CRST participants had high cotinine (>15 ng/mL) and less than a quarter had low cotinine (<3 ng/mL) concentrations.
Distribution of serum cotinine levels among CRST participants. CRST, Cheyenne River Sioux Tribe.
Results of the QRM are presented in table 2 by predictor variables. The 0.10th quantile (or the 10th percentile of cotinine natural log distribution) was associated with low cotinine concentration (<3 ng/mL, non-tobacco users without SHS), the 0.25th and 0.50th quantiles were linked with intermediate cotinine levels (3–15 ng/mL, non-tobacco users with SHS), and the 0.75th and 0.90th quantiles were associated with high cotinine concentrations (>15 ng/mL, tobacco users). Results showed that sampling years were critical contributors of increasing likelihood of low cotinine values in the study. In 2015, approximately 2 months after implementing the new smoke-free air Tribal policy, cotinine levels among non-tobacco users, both with and without SHS, were lower than in the previous sampling year. Yet, there was no statistically significant change in cotinine concentrations in tobacco users. Compared to 2015, 2016 cotinine levels increased among non-tobacco users without SHS, remained unchanged in non-users with SHS, and decreased in tobacco users. Overall, from 2014 to 2016, there was an increase in the lowest part of cotinine distribution, among non-users without SHS (1.69 in 0.10th quantile), and cotinine levels decreased in non-users with SHS exposure (−1 and −1.3 in 0.25th and 0.50th quantiles) and in tobacco users (−1.47 and −2.33 in 0.75th and 0.90th quantiles) (table 2).
Serum cotinine natural log: quantile regression estimates (95% CI) for selected quantiles
For almost all quantiles of cotinine concentrations, self-reported smoking and chewing were predictors of increased cotinine within a quantile. The interaction between active smoking and tobacco chewing was taken as an indicator of dual tobacco use. Dual users had lower cotinine levels, compared with their respective counterparts, in most of the quantiles (see the ‘Interaction between smoking and chewing’ row in table 2). This suggested that dual tobacco use had a lesser impact on cotinine than smoking without chewing.
Childhood SHS significantly elevated cotinine in non-tobacco users who were currently exposed to SHS (0.76 at 0.50th quantile). Older users had significantly higher cotinine levels compared with younger users (0.56 at 0.90th quantile). The comparison of age effect as continuous vs binary variable was shown in online supplementary figure S1. There was no significant effect of gender, fishing status, CoSHS, or environmental exposure zone on the detected cotinine concentrations.
The QRM plots were presented in figure 1. Low cotinine threshold of 3.08 ng/mL was the 0.23th cotinine quantile of the study sample while the higher cotinine threshold of SHS exposure value, 15.02 ng/mL, was the 0.54th sample quantile of cotinine. The 0.23th and 0.54th quantiles were shown in figure 1 as the two solid vertical lines in each plot. These lines visually divided each QRM plot into the following three areas (from left to right): low (<3 ng/mL), intermediate (3–15 ng/mL) and high (>15 ng/mL) cotinine concentrations.
QRM results (figure 1A–C) of sampling year as a predictor of cotinine were similar to the results based on descriptive analysis (table 1). Cotinine ranges narrowed down over the data collection years. This was especially demonstrated by the decrease of high cotinine levels in the graph (figure 1B,C).
Smokers (figure 1D) and tobacco chewers (figure 1F) had elevated cotinine levels in majority of cotinine quantiles, compared with non-smokers and non-chewers, respectively. Dual tobacco users were predicted to have lower cotinine concentrations in all three cotinine areas, when compared with other participants (figure 1E), although in some quantiles this was not significant.
The effect of daily dose of commercial cigarette consumption was also investigated (see online supplementary figure S2); as was expected, higher cigarette consumption had a greater effect on cotinine levels. However, the number of cigarettes smoked by others in their homes was not a significant predictor.
Conclusions
This is the first study to our knowledge examining the impact of a smoke-free policy on cotinine levels in an AI/AN community. The results demonstrate the usefulness of implementing such a policy against smoking. Besides the immediate benefits of decreased indoor nicotine exposure, long-term reduction in cardiovascular risk reported in other populations22 might also be expected. Evidence supports that smoke-free policies provide benefits to society11 and can diminish early smoking initiation and eventually curb active tobacco use.
Several studies documented that self-reported surveys can significantly underestimate tobacco consumption, suggesting the need for confirmation of actual exposures in various populations.25 37–39 Self-reported tobacco use among CRST participants was successfully confirmed by using cotinine measurement. Not only high rate of active smoking was found but non-tobacco users with extreme cotinine concentrations were also identified. This suggested that there were other potential exposure sources to SHS not evaluated in this study. Evidence suggests that 34% of Tribal members do not own reliable transportation, therefore they practice regular hitchhiking and rely on carpooling.30 That can increase smoke exposure given the prevalence of active smokers in the community. Passengers’ exposure to SHS in automobiles with smokers has been documented in other ethnic groups.40 The extremely high cotinine concentrations (highest concentration of 859.1 ng/mL) warrant further work to reduce health disparity in the CRST communities.
The authors’ previous research has shown that non-commercial ceremonial tobacco use had no influence on cotinine in either smokers or non-smokers.6 Ceremonial use of commercial tobacco products could have contributed to the detected high cotinine levels, but that effect was not evaluated.
Furthermore, we lacked information about electronic cigarette (e-cig) use. At the time of the study, there were no e-cig data available on CRST nor among AI/ANs nationwide.41 e-Cig and tobacco cigarette consumption had a similar effect on cotinine concentrations in both active smokers and non-smokers with SHS.42–44 Recent data among AI/AN smokers confirmed that active tobacco cigarette smoking and e-cig use both generated comparable salivary cotinine levels.45
We had no information about other possible sources of nicotine consumptions or nicotine replacement therapy even though the later is prescribed in various Tribal communities.46–48 Because cigarette smoking is the dominant form of combustible tobacco use on the CRST, our short survey did not distinguish between various forms of combustible consumption. Another limitation of this study was the lack of direct survey questions regarding no-indoor smoking rules established at work and in homes. However, only a small number of CRST participants reported having indoor smoking restrictions (see Results section).
Genetic factors have been shown to be molecular drivers of high nicotine metabolising capacity. They were not investigated under this study, but previously were shown in the same group of Northern Plains Tribal members.6 49 This is likewise supported by the similarity of frequency of extreme cotinine concentrations among tobacco users and non-users (see Results section).
Although historical childhood home SHS exposure and age had weak effects at certain quantiles of cotinine, no other known demographic variables (gender, fishing status, or geographical location) altered the predictive association of sampling years and tobacco-related covariates on cotinine concentrations. We hypothesised that the cotinine level increase in 2016 may come from less smoking in designated public places, more in homes and cars. Further research might investigate changes in cotinine levels via SHS and thirdhand smoke in this Tribal community.
Based on this analysis, self-reported CRST smoking information was accurate and confirmed previous findings of 51% active smoking rate in this Tribe.4 50 The detected high active smoking rates support the importance to continue the anti-smoking campaigns and efforts of the Canli Coalition protecting the enactment of Ordinance #77. This analysis suggests that other Tribal communities would rapidly benefit from implementation of similar, anti-smoking public health regulations.
What this paper adds
This is the first study to our knowledge that demonstrates the importance and immediate impact of a smoke-free Tribal policy on serum cotinine levels in a high-risk, high tobacco consuming American Indian community.
We confirm the effectiveness of using serum cotinine levels in the context of a Tribal tobacco control policy.
Decrease in secondhand smoke exposures shortly after enacting the Tribal tobacco control policy can be detected. This result supports the protection of a Sovereign Nations’ comprehensive smoke-free law.
Acknowledgments
The analyses presented here received support from the UNM Center for Native Environmental Health Equity Research to EOD, funded by NIH 1P50ES026102 and assistance agreement no. 83615701 from the Environmental Protection Agency (EPA). This work has not been formally reviewed by the EPA. The views expressed are solely those of the authors and do not necessarily reflect those of the Agency. The first author is indebted to Ji-Hyun Lee, DrPh for her time and mentorship on statistical methods applied in this paper. We also thank Ms Corrine Huber, MA for her dedicated work with the Canli Coalition and its partners and for providing comments.
References
Footnotes
Contributors As corresponding author I state that all members of the team who named as authors on this paper agreed to the submission, reviewed the text and tables and are in agreement with the manuscript’s conclusions.
Funding This work was supported by NARCHVII study 'Complex metal exposures and immune status on the Cheyenne River'—of which data were presented in this paper—U261IHS0076 IHS/NIH/NIAID; Black Hills Center for American Indian Health (Director: Dr Jeffrey A Henderson, MD, MPH) and subaward to UNM HSC PI: Dr Erdei.
Competing interests No, there are no competing interests.
Patient consent for publication Not required.
Ethics approval This community-academic collaboration, including community-based outreach, recruitment and enrolment was approved by UNM HSC HRRC (HRPO# 08–486) and CRST Tribal Executive Resolution (E-135–2014-CR).
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement No data are available.