Work

Are Some E-Cigarette Users “Blowing Smoke”?: Assessing the Accuracy of Self-Reported Smoking Abstinence in Exclusive E-Cigarette Users

Year: 2018

Type: article

Abstract: The impact of electronic cigarette (e-cigarette) use on health is frequently cited as an urgent research need.1 Population-based, longitudinal cohort studies can provide necessary evidence to inform researchers and the general public about potential health risks that e-cigarettes may pose. However, such studies frequently rely on self-reported data to infer e-cigarette user status, and such reports may not be completely accurate.2 Furthermore, there are populations that frequently misreport tobacco use, many of which constitute vulnerable groups, such as pregnant women and youth.3,4 Currently, data are lacking with respect to whether e-cigarette users may misreport occasional smoking. Recently, a study by Rubinstein et al.5 raised serious concerns regarding adolescents who vape e-cigarettes that also reported no tobacco cigarette use. In their study, self-reported e-cigarette users were found to have detectable levels of tobacco-specific biomarkers, suggesting that there may be potential issues in relying exclusively on self-reported data to establish tobacco use. We investigated this problem among adults by estimating the proportion of e-cigarette users that may be misreporting their smoking status. We analyzed data from Wave 1 of the Population Assessment of Tobacco and Health (PATH) Restricted Use Files via the Inter-University Consortium for Political and Social Research (ICPSR) secure Virtual Data Enclave (VDE).6 Details on the design, sampling, data collection, biospecimen collection, and weighting procedures are described elsewhere.6,7 For this analysis, data from the adult questionnaire were merged with biomarker data for those individuals who provided a urine sample for analysis and were included in the PATH biospecimen subsample at Wave 1. Our analysis focused on four groups: (1) exclusive e-cigarette users, who reported currently using e-cigarettes every day or some days and reported no current use of other tobacco products; (2) former cigarette smokers, who had ever smoked 100 cigarettes in their lifetime and reported no current cigarette or any other tobacco use; (3) current exclusive cigarette users who reported smoking at least 100 cigarettes in their lifetime and reported current every day or some days use of cigarettes, and no other current tobacco product use; and (4) never users, defined as those with no lifetime history of any tobacco use. To be included, all participants had to report no past 3-day use of nicotine replacement therapies, and, among users, reported smoking cessation ≥6 months prior to their interview. Biomarkers were measured using analytical methods described elsewhere.6 This analysis focused on urinary NNAL, a biomarker of exposure to tobacco-specific nitrosamine NNK, which has been demonstrated to distinguish between active and passive smoke exposure.8 Although trace levels of NNK have been detected in some e-cigarette products, levels are significantly lower than tobacco cigarettes, supporting the selection of NNAL as a specific biomarker to distinguish between smokers and exclusive e-cigarette users.9 Levels of urinary NNAL were corrected for creatinine, and measurements below the analytic limit of detection (LOD) were imputed using LOD/√2.10 Participants with urinary creatinine levels less than 10 mg/dL or more than 370 mg/dL were excluded from the analysis to adjust for confounding resulting from overly dilute or overly concentrated urine specimens, along with data points subject to interference in chromatographic peaks in lab assays. In the first step of these analyses, we performed a nonparametric Receiver Operating Curve (ROC) analysis on never tobacco users and current exclusive cigarette smokers using unweighted data to estimate the optimal cutoff value for urinary concentrations of NNAL (n = 4062).8 This cutoff would allow for a highly sensitive and specific classification of study participants into groups of smokers and nonsmokers. The participants with NNAL levels more than this determined cutoff value would be classified as smokers, whereas those with NNAL levels below the cutoff would be classified as nonsmokers. Following this, the determined cutoff value was applied to estimate the proportion of self-reported exclusive e-cigarette users and former cigarette smokers who had urinary NNAL levels that exceeded the cutoff. Those subjects were classified as having false negative results, meaning that although they reported being nonsmokers, their urinary NNAL level suggested that they recently smoked tobacco. In order to determine whether e-cigarette users are more likely to misreport smoking, we compared proportion of exclusive e-cigarette users above the cutoff value with the proportion of ex-smokers above the cutoff value using a χ2 test. All proportions were weighted using urine weights for the biospecimen subsample.6 Results indicated that the optimum cutoff value for NNAL in urine that distinguished between current smokers and never tobacco users was 14.5 pg/mg creatinine. This concentration allowed for 93.3% sensitivity and 95.9% specificity to differentiate between current exclusive cigarette smokers and never tobacco users, respectively. Put differently, 93.3% of those who reported smoking had urinary NNAL levels above 14.5 pg/mg (true positive) and 4.1% of those who reported never tobacco use had urinary levels above 14.5 pg/mg (100%–95.9%; false positive). Overall, applying this cutoff value allowed for correct classification of 94.3% cases (ROC Area: 0.9797, SE = 0.0021, asymptomatic normal CI = 0.975 to 0.984). In applying this cutoff to respondents who reported exclusive use of e-cigarette (no smoking or use of other tobacco products within the past 6 months) (n = 144), we found that 15% of this population had urinary NNAL levels above 14.5 pg/mg creatinine. The proportion of exclusive e-cigarette users above the established cutoff value was similar to the proportion of former smokers reporting no past 6-month tobacco product use (n = 156) who also had urinary NNAL levels that exceeded this threshold (15%; p = .85). Our analyses suggest that approximately one in six respondents who reported only using e-cigarettes may have smoked tobacco cigarettes. Our analyses also suggest that caution is needed when relying on self-reported tobacco use status among ex-smokers, as one in six respondents who reported quitting within the past 6 months may still smoke, at least occasionally. Studies that utilize the PATH and other survey data need to account for potential reporting biases due to misreported smoking status. Given the robust design of the PATH project in conjunct with its available biomarker data, we recommend using urinary NNAL as confirmation to objectively verify nonsmoking among e-cigarette users and former smokers. This work was supported by NCI P30 CA16056. Approval for this study was granted by the Institutional Review Board at Roswell Park Comprehensive Cancer Center. Maciej L. Goniewicz reports receiving fees for serving on an advisory board from Johnson & Johnson, and grant support from Pfizer. Danielle M. Smith has no conflicts of interest to declare.

Source: Nicotine & Tobacco Research

DOI: 10.1093/ntr/nty085

Authors: Maciej L Goniewicz, Danielle M Smith

Institution: Roswell Park Comprehensive Cancer Center

Language: English

Topic: Smoking Behavior and Cessation

Subfield: Physiology

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: bronze

Funders: National Institutes of Health, National Cancer Institute