Water pipe (Shisha,Hookah,Arghile) Smoking and Secondhand Tobacco Smoke Effects on CYP1A2 and CYP2A6 Phenotypes as Measured by Caffeine Urine Test

Abstract

Public policies to stop or reduce cigarette smoking and exposure to secondhand smoke and associated diseases have yielded successful results over the past decade. Yet, the growing worldwide popularity of another form of tobacco consumption, water pipe smoking, has received relatively less attention. To the best of our knowledge, no study to date has evaluated the effects of water pipe smoking on cytochrome P450 (CYP450) activities and drug interaction potential in humans, whereas only limited information is available on the impact of secondhand smoke on drug metabolism. In a sample of 99 healthy volunteers (28 water pipe smokers, 30 secondhand tobacco smoke exposed persons, and 41 controls), we systematically compared CYP1A2 and CYP2A6 enzyme activities in vivo using caffeine urine test. The median self-reported duration of water pipe smoking was 7.5 h/week and 3 years of exposure in total. The secondhand smoke group had a median of 14 h of self-reported weekly exposure to tobacco smoke indoor where a minimum of five cigarettes were smoked/hour for a total of 3.5 years (median). Analysis of variance did not find a significant difference in CYP1A2 and CYP2A6 activities among the three study groups (p > 0.05). Nor was there a significant association between the extent of water pipe or secondhand smoke exposure and the CYP1A2 and CYP2A6 activities (p > 0.05). Further analysis in a subsample with smoke exposure more than the median values also did not reveal a significant difference from the controls. Although we do not rule out an appreciable possible impact of water pipe smoke and secondhand smoke on in vivo activities of these two drug metabolism pathways, variability in smoke constituents from different tobacco consumption methods (e.g., water pipe) might affect drug metabolism in ways that might differ from that of cigarette smoke. Further studies in larger prospective samples are recommended to evaluate water pipe and secondhand tobacco smoke effects on CYP450 function, particularly at higher smoke exposure conditions.

Introduction

Numerous public policy and educational measures have been put into action to reduce cigarette smoking around the globe (Stone et al., 2016). Yet, new forms of smoking patterns and secondhand tobacco smoke exposure continue to pose a public health threat (Jones et al., 2014, 2016; Rajagopalan et al., 2016; Raoof et al., 2015). Chief among these is tobacco smoking using water pipe, also known as arghile, hookah, and shisha. Water pipe smoking has become a global epidemic, particularly among youth and college students, with serious repercussions for global health (Maziak et al., 2015).

The practices of water pipe smoking vary widely across the globe. It is important to note, however, water pipe smoking is not limited to Middle East but has gained worldwide popularity. This is, in part, because of several misperceptions that tobacco consumption by water pipe smoking is relatively safer and socially desirable (Grant and O'Mahoney, 2016). Filtering of the smoke through water before inhalation, use of aromatic flavored tobacco products (Maassel), relative paucity of regulations over water pipe smoking, together with the sought after social context in which smoking occurs contribute to popularity of water pipe smoking worldwide.

A traditional water pipe setup for tobacco smoking comprises the water pipe device that heats the tobacco with charcoal, usually through perforated aluminum wrapping. The generated smoke is filtered through a water container, which is then inhaled through a rubber pipe. Water pipe smoking, unlike cigarette smoking sessions that can last in the order of several minutes, is an intermittent long-term session lasting about 1 to 2 h per package of tobacco product.

There is a growing body of epidemiology data that support the idea that water pipe smoking has numerous adverse impacts on human health (Akl et al., 2010; Kim et al., 2016; Maziak et al., 2015; Salloum et al., 2016). Water pipe tobacco smoking was significantly associated, for example, with lung and other types of cancer, respiratory illness, low birth weight, and periodontal disease, to name but a few (Akl et al., 2010; Maziak, 2013). Apart from deleterious impacts on various health outcomes, tobacco smoke is one of the most notable modulators of clinical pharmacokinetics and drug action. In addition to active exposure to tobacco smoke, secondhand smoke is noteworthy (Frazer et al., 2016), for it has been associated, among others, with otitis media and asthma in children, and acute myocardial infarction, premature death, and lung cancer in adults (Institute of Medicine, 2009; U.S. Department of Health and Human Services, 2006).

To the best of our knowledge, no study to date has evaluated the effects of water pipe smoking on cytochrome P450 (CYP450) family of drug metabolizing enzymes and, by extension, drug interaction susceptibility in humans. Insofar as secondhand smoke is concerned, be it from water pipe or cigarette consumption, there is limited information available on the extent to which drug metabolism is influenced after exposure to secondhand smoke. These are important gaps in medical and clinical pharmacology literature because likely different molecular composition of smoke produced by varied methods of tobacco consumption (e.g., cigarette versus water pipe smoking) may have potentially different effects on drug metabolism pathways. Moreover, because water pipe smoking and secondhand smoke exposure are both modifiable risk factors for public health and drug interactions, the knowledge of the extent of their impacts on multiple organ systems and biological pathways is important for rational therapeutics and clinical practice.

We report here, for the first time in the literature to the best of our knowledge, the effects of water pipe smoking and secondhand tobacco smoke exposure on CYP1A2 and CYP2A6 enzyme activities in a sample of 99 healthy volunteers using the caffeine urine test.

Materials and Methods

Subjects and study design

The study was approved by the Institutional Research Ethics Committee of Gaziantep University. A written informed consent was obtained from all subjects. Study subjects were healthy volunteers, men and women within the 18 to 65 years of age window, did not use chronic medication, herbal supplements, and illicit substances, and did not have chronic medical morbidity, acute infections, alcohol abuse, or dependence as determined by routine urine biochemistry and in-depth screening interviews and medical history. The subjects did not use grapefruit juice or other foodstuff that may interfere with drug metabolism.

The study design included three study groups comprising (1) controls who were nonsmokers of tobacco products and were not exposed to secondhand smoke, (2) water pipe tobacco users who smoked on a regular basis in the past 2 months at least 1 h per week or more, and did not use other methods of tobacco smoking, and (3) secondhand smoke group who were nonsmokers exposed to tobacco smoke on a regular basis in the past 2 months at least 1 h per week indoors where a minimum of five cigarettes were smoked per hour. These were the minimum requirements for study entry. Most subjects in our geographical area in southeast Turkey who have participated in the study had, however, higher levels of water pipe or secondhand smoke exposure as noted in the results section. Exposure data were collected by self-reports and interviews with the subjects.

Caffeine urine test procedure for CYP1A2 and CYP2A6 phenotypes

After an initial void of urine in the early evening, patients ingested 100 mg oral caffeine in the form of two cups of standardized instant coffee and collected their urine until the next morning, including the first morning urine, for a minimum of 8 h. Urine aliquots were stored at −20°C until the analysis for caffeine metabolites. On days when caffeine test was performed, caffeinated-beverage consumption was limited to a maximum of two cups of coffee, and the subjects abstained from consumption of alcohol on the day and the day before the urine test. The urine samples were analyzed for 5-acetylamino-6-formylamino-3-methyluracil (AFMU), 1-methylxanthine (1X), 1-methyluric acid (1U), and 1,7-dimethyluric acid (17U). The caffeine metabolic ratio (CMR), an in vivo index of CYP1A2 activity, was calculated with the ratio (AFMU +1X + 1U)/(17U) as described and applied previously in clinical biomarker studies by our group (Ozdemir et al., 1998a, 1998b). The denominator of the CMR (17U) controls for variability in habitual caffeine consumption (Ozdemir et al., 1998a, 1998b). CYP2A6 metabolic ratio was calculated by (17X/17U).

Of note, an increase in CMR is indicative of an increase in CYP1A2 enzyme activity, whereas an increase in the CYP2A6 metabolic ratio already noted (17X/17U) is indicative of a decrease in CYP2A6 enzyme activity in vivo.

Measurement of caffeine metabolites

Caffeine metabolites were measured by HPLC-MS/MS using a slight modification of the procedure described by Rybak et al. (2014). Chemicals 1X, 17X, 1U, and 17U, and the internal standard (IS) levallorphan were obtained from Sigma-Aldrich (St. Louis, MO, USA). AFMU was purchased from Santa Cruz Biotechnology (Dallas, TX, USA). LC-MS grade solvents (methanol, acetonitrile, and formic acid) were purchased from Fisher Scientific (Geel, Belgium). Deionized water was purified through a water purification system from Millipore (Bedford, MA, USA). The stock solutions of all analytes and IS were prepared at 10 mg/mL in milliQ purified water (0.1% formic acid) and stored in absence of light at −20°C. Working standard solutions were prepared at 10 μg/ml by appropriate dilution and stored at −20°C. Standard calibration samples were prepared by spiking the blank human urine with the working solution of each analyte. Quality control samples were similarly prepared from the stock solutions at three different concentrations within the designated low-, medium-, and high-concentration ranges.

An Agilent 1200 Series HPLC system (Agilent, Santa Clara, CA, USA) equipped with a G1311A binary pump, a G1329A autosampler, a G1322A degasser, and a G1316A column oven was utilized to analyze the urine samples. The chromatographic column was a polar-embedded C18 Synergi Fusion-RP column (50 × 2 mm internal diameter, 4 μm), which was kept at a constant temperature of 50°C, and preceded by a 4 × 2 mm Synergi Fusion guard column, both from Phenomenex (Torrance, CA, USA). Mass-spectrometric detection of the analytes was performed using an API2000 triple quadrupole mass spectrometer from AB Sciex (Framingham, MA, USA) equipped with an atmospheric pressure electrospray ionization (ESI) interface operated with Analyst software (version 1.5.1). Samples were centrifuged in a Spectrafuge 24D microcentrifuge from Labnet International (Woodbridge, NJ, USA).

For the extraction of analytes, urine samples were first thawed and then 3 mL of human urine was mixed with 12 mL of 0.1% formic acid aqueous solution and 10 μL of the IS (10 μg/mL). The resultant mixture was vortexed for 15 sec, then centrifuged for 5 min at 3500 g, and 1 mL of the supernatant was selected and filtered through a 0.45 μm nylon filter. Aliquots of 5 μL of the extract were injected into the chromatographic system.

The compounds were separated using 0.1% formic acid in water and 0.1% formic acid in acetonitrile as mobile phase. The flow rate was 0.2 mL min⁻¹. Elution was in a linear gradient, with acetonitrile content changing from 5% to 85% between 2 and 3 min. Acetonitrile content was maintained at 85% from 3 to 4 min, and then decreased to 5% over 1 min, followed by column equilibration for 7.5 min. The total run time was thus 12.5 min per sample. Data acquisition was performed in selected reaction monitoring (SRM) mode with negative ESI for all analytes except 17X and the IS whose data were acquired with positive ESI. The ESI inlet conditions were as follows: curtain gas, nitrogen (20 psi); collision gas, nitrogen (6 psi); ion spray voltage, 4500 V in positive mode and −4000 V in negative mode; ion source temperature, 500°C; ion source gas 1, nitrogen (50 psi); and ion source gas 2, nitrogen (90 psi). SRM was optimized for each analyte/standard by direct infusion using the ESI source. Unit mass resolution was used throughout.

Analytes were quantified by interpolation of peak area ratios for MS/MS transitions against 10-point calibration curves (quadratic regression, 1/x weighting). The response was linear from 2.5 ng/mL to 4 μg/mL for all analytes (r > 0.99 in all cases). A summary of mass spectrometric conditions used is presented in Table 1. The within-day and between-day coefficients of variation (CVs) were less than 12%, with a lower limit of quantitation assigned as the low standard (2.5 ng/mL). Samples were analyzed in triplicate and the mean values were used in the statistical analysis.

Table 1.

Analysis Parameters and Conditions for the Measurement of Caffeine Metabolites

		MS parameters (V)
Compound	SRM transition (m/z)	DP	CE	CXP
Negative ESI mode
1X	164.9 > 108.1	−25	−24	−4
1U	180.9 > 137.8	−21	−20	−8
17U	194.9 > 179.9	−21	−22	−6
AFMU	224.9 > 196.9	−26	−14	−12
Positive ESI mode
17X	181.1 > 124.1	31	20	4
LL (IS)	284.2 > 157.1	30	51	12

1U, 1-methyluric acid; 17U, 1,7-dimethyluric acid; 1X, 1-methylxanthine; 17X, 1,7-dimethylxanthine; AFMU, 5-acetylamino-6-formylamino-3-methyluracil; CE, collision energy (eV); CXP, collision cell exit potential (V); DP, declustering potential (V); ESI, electrospray ionization; IS, internal standard; MS, mass spectrometry; SRM, selected reaction monitoring.

Statistical analysis

Data were assessed for normality and log transformations were used when necessary. Putative differences among the study groups for CYP1A2 and CYP2A6 activities were evaluated by analysis of variance (ANOVA). Subjects' age, weight, height, sex, and self-reported ethnicity data were entered as covariates in ANOVA. For the secondhand smoke group, the square meter dimensions of the room or the closed indoor space where the subjects were exposed to secondhand smoke was entered as a co-variate to explain the between-subject variation in enzyme activities. Nonparametric Spearman's rank order correlations were used to evaluate associations among the study variables. Percentage coefficient of variation (CV%: standard deviation/mean × 100) was utilized to express between-subject variability in CYP1A2 and CYP2A6 activities. A p value of less than 0.05 was accepted as statistical significance threshold.

Results

The total study sample size comprised 99 healthy volunteers: 41 controls (n = 18 women and 23 men), 30 secondhand tobacco smoke exposed persons (n = 8 women and 22 men), and 28 water pipe smokers (n = 3 women and 25 men). The greater number of men among the sample of water pipe smokers reflects the larger prevalence of water pipe use by men in the study region in southeast Turkey. The median age was 30, 26, and 27 years, respectively, in the control, secondhand smoke, and water pipe groups. Most subjects (85 out of 99) had a self-reported ethnicity of Turkish, whereas 14 were of Syrian-Arab ethnicity. All females in the study were premenopausal and did not use oral contraceptives.

In terms of the tobacco consumption and exposure patterns, all water pipe smokers consumed a flavored tobacco product. The amount of tobacco consumed by water pipe was 15 g per package, and each smoking session lasted 90 min. The subjects did not share their water pipe with other smokers and consumed each tobacco package on their own. The median self-reported water pipe smoking was 7.5 h/week, with a median duration of overall exposure dating back to previous 3 years, in total. Most subjects, 75% of them (21 out of the 28 study participants), smoked water pipe indoors.

The secondhand smoke group had a median of 14 h of self-reported weekly exposure to tobacco smoke indoors where a minimum of five cigarettes were smoked/hour, for a total of 3.5 years (median). All secondhand smoke exposed subjects had passive exposure to cigarette smoke only (n = 29), except one subject who was exposed to secondhand water pipe smoke.

ANOVA did not find a significant difference in CYP1A2 and CYP2A6 activities among the three study groups (p > 0.05) (Figs. 1 and 2). Nor was there a significant association between the duration of water pipe smoking sessions or the secondhand smoke exposure duration versus the CYP1A2 and CYP2A6 enzyme activities (p > 0.05). A further analysis in a subsample with smoke exposure levels more than the median values in the water pipe and secondhand smoke groups also did not reveal a significant difference from the controls (p > 0.05). Between-subject variability in both CYP1A2 and CYP2A6 activities appeared to be larger in the water pipe group as suggested by a greater %CV (Table 2)

FIG. 1.

Distribution of CYP1A2 enzyme activity among healthy controls (squares), secondhand (passive) tobacco smoke exposed group (circles), and water pipe smokers (triangles).

FIG. 2.

Distribution of CYP2A6 metabolic ratio among healthy controls (squares), secondhand (passive) tobacco smoke exposed group (circles), and water pipe smokers (triangles).

Table 2.

CYP1A2 and CYP2A6 In Vivo Activities as Measured by Caffeine Urine Test, and Between-Subject Variability in Three Groups of Healthy Volunteers

	CYP1A2 metabolic ratio ^a	CYP2A6 metabolic ratio ^a
Controls
Median	3.99	1.21
Mean	4.02	1.39
Standard deviation	1.42	0.78
CV (%)	35	56
Secondhand tobacco smoke exposed
Median	3.89	1.13
Mean	4.07	1.22
Standard deviation	1.44	0.62
CV (%)	35	51
Water pipe smokers
Median	4.20	1.20
Mean	4.62	1.69
Standard deviation	2.26	1.32
CV (%)	49^b	78^b

An increase in CYP1A2 metabolic ratio ([AFMU +1X + 1U]/[17U]) is indicative of an increase in enzyme activity, whereas an increase in CYP2A6 metabolic ratio (17X/17U) is indicative of a decrease in enzyme activity in vivo.

Note that the CV and between-subject variability of CYP1A2 and CYP2A6 activities appear to be markedly larger in water pipe smokers than in the control and the secondhand tobacco smoke exposed group.

CV, coefficient of variation.

Discussion

This is the first clinical phenotyping report, to the best of our knowledge, that examined the effects water pipe smoking and secondhand tobacco smoke exposure on two key CYP450 enzymes, CYP1A2 and CYP2A6. Both pathways make critical contributions to drug and xenobiotic metabolism, display population variability, and influence a host of disease susceptibilities and health outcomes (Aklillu et al., 2014; de Andrés et al., 2016). Unlike cigarette smoking that lasts several minutes with relatively continuous and intense exposure to smoke, water pipe smoking sessions last longer, typically in the order of 1 or 2 h, and involve intermittent exposure to smoke inhalation separated by socialization and related rituals. The depth and frequency of smoke inhalation can additionally determine the outcome of CYP450 enzyme phenotype changes after exposure to tobacco smoke during a water pipe smoking session or secondhand exposure to smoke. Self-reports can be limited in capturing smoke exposure data as well.

Notwithstanding these caveats and potential limitations, our sample of water pipe smokers had an appreciable median weekly smoking session of 7.5 h, and mostly indoors without use of other tobacco products, a practice that they reported dating back a median of 3 years before the study. The key observation made in this study is that there did not appear to be a significant difference in CYP1A2 and CYP2A6 activities in the water pipe group compared with those in the controls. It is conceivable that at even higher intensity and frequency of water pipe smoke exposure, a robust change in CYP450 activities may occur, particularly in the CYP1A2 pathway. Alternatively, variability in smoke constituents from different tobacco consumption methods such as water pipe might affect drug metabolism in ways that differ from that of cigarette smoke. The latter hypothesis and explanation are, in part, supported by the observation that we did not find an association between the duration of water pipe smoking and CYP1A2 and CYP2A6 activities.

In the water pipe group, two subjects were outliers with a CMR value of 10.01 and 11.64, indicating high CYP1A2 activity that was markedly more than the sample median CMR value of 4.20 (Fig. 1). Their weekly water pipe smoking session duration data were 1 and 10 h, respectively (median value 7.5 h/week; nine subjects had in fact a self-reported weekly use history of greater than 10 h but with lower CMR values). In CYP2A6 metabolic ratio distributions, we identified two outlier subjects who appeared to display distinct enzyme activity (Fig. 2). However, similar to the case with CYP1A2 outliers, CYP2A6 outliers were not explained nor accompanied by outlier water pipe smoke exposure data.

The effect of secondhand cigarette smoke exposure on CYP1A2-mediated phenacetin metabolism was previously studied in a Chinese healthy volunteer sample by Dong et al. (1998). The study in nonsmoker subjects who lived in families where at least one family member smoked 10 cigarettes/day for at least 8 h/day showed a 24% increase in phenacetin apparent oral clearance, whereas the effect of active cigarette smoking was markedly greater as evidenced by a 250% increase in phenacetin apparent oral clearance (Dong et al., 1998). In this study, our sample of healthy subjects had a median of 14 h of self-reported weekly exposure to tobacco smoke indoors. Importantly, in each hour of secondhand tobacco smoke exposure, a minimum of five cigarettes were smoked/hour indoors, which is considerable. In our sample, seven volunteers had 28 or more hours of weekly heavy exposure to secondhand smoke. Despite these secondhand smoke exposure rates, we did not observe a significant effect on CYP1A2 and CYP2A6 activities. Secondhand smoke continues to be an important public health threat worldwide (Dai and Hao, 2016). Studies in subjects with very high levels of secondhand smoke exposure might be required to dissect the putative effects on drug metabolism pathways.

In all, although these observations should be considered preliminary, they raise new hypothesis and avenues for future research on the effects of water pipe smoking and secondhand smoke on CYP450 drug metabolism pathways. The anticipated water pipe smoke effects may be uniquely different on each CYP450, in part, because of the likely different composition of the tobacco smoke after preparation by different methods of tobacco combustion and consumption. Larger exposure to water pipe and secondhand smoke might be necessary to observe an appreciable effect on CYP450 function. Prospective studies with larger samples that measure the water pipe use and smoke exposure data are warranted to answer these new hypotheses. The answers identified will ultimately play important roles in shaping future public and preventive medicine policies targeting the current worldwide epidemic of water pipe smoking as well as secondhand smoke exposure. New public policies are particularly important considering the numerous adverse impacts of water pipe and secondhand smoke (Wright et al., 2016). They are modifiable and preventable risk factors once their effects on a broader range of biological systems are mapped out as suggested in this report.

Footnotes

Acknowledgment

Prof. Vural Özdemir is recipient of a career award from the Scientific and Technological Research Council of Turkey (2232 Program). The views expressed represent the personal opinions of the authors and do not necessarily reflect the positions of their affiliated institutions.

Author Disclosure Statement

The authors declare that no conflicting financial interests exist.

Abbreviations Used

References

Akl

, Gaddam

, Gunukula

, Honeine

, Jaoude

, and Irani

. (2010). The effects of waterpipe tobacco smoking on health outcomes: A systematic review. Int J Epidemiol, 39, 834–857.

Aklillu

, Djordjevic

, Carrillo

, Makonnen

, Bertilsson

, and Ingelman-Sundberg

. (2014). High CYP2A6 enzyme activity as measured by a caffeine test and unique distribution of CYP2A6 variant alleles in Ethiopian population. OMICS, 18, 446–453.

Dai

, and Hao

. (2016). The Prevalence of exposure to workplace secondhand smoke in the United States: 2010 to 2015. Nicotine Tob Res pii: ntw306. [Epub ahead of print].

de Andrés

, Terán

, Bovera

, Fariñas

, Terán

, and LLerena

. (2016). Multiplex phenotyping for systems medicine: A one-point optimized practical sampling strategy for simultaneous estimation of CYP1A2, CYP2C9, CYP2C19, and CYP2D6 activities using a cocktail approach. OMICS, 20, 88–96.

Dong

, Ping

, Xiao

, et al. (1998). Effect of active and passive cigarette smoking on CYP1A2-mediated phenacetin disposition in Chinese subjects. Ther Drug Monit, 20, 371–375.

Frazer

, Callinan

, McHugh

, et al. (2016). Legislative smoking bans for reducing harms from secondhand smoke exposure, smoking prevalence and tobacco consumption. Cochrane Database Syst Rev, 2, CD005992.

Grant

, and O'Mahoney

. (2016). Portrayal of waterpipe (shisha, hookah, nargile) smoking on Twitter: A qualitative exploration. Public Health, 140, 128–135.

Institute of Medicine. (2009). Secondhand Smoke Exposure and Cardiovascular Effects: Making Sense of the Evidence. Institute of Medicine; Committee on Secondhand Smoke Exposure and Acute Coronary Events. Washington, DC: National Academies Press (USA).

Jones

, St Helen

, Meyers

, et al. (2014). Biomarkers of secondhand smoke exposure in automobiles. Tob Control, 23, 51–57.

10.

Jones

, Magid

, Al-Rifai

, et al. (2016). Secondhand smoke exposure and subclinical cardiovascular disease: The multi-ethnic study of atherosclerosis. J Am Heart Assoc, 5, pii: e002965.

11.

Kim

, Kabir

, and Jahan

. (2016). Waterpipe tobacco smoking and its human health impacts. J Hazard Mater, 317, 229–236.

12.

Maziak

. (2013). The waterpipe: An emerging global risk for cancer. Cancer Epidemiol, 37, 1–4.

13.

Maziak

, Taleb

, Bahelah

, et al. (2015). The global epidemiology of waterpipe smoking. Tob Control, 24 Suppl 1, i3–i12.

14.

Ozdemir

, Naranjo

, Herrmann

, et al. (1998a). The extent and determinants of changes in CYP2D6 and CYP1A2 activities with therapeutic doses of sertraline. J Clin Psychopharmacol, 18, 55–61.

15.

Ozdemir

, Naranjo

, Shulman

, et al. (1998b). Determinants of interindividual variability and extent of CYP2D6 and CYP1A2 inhibition by paroxetine and fluvoxamine in vivo. J Clin Psychopharmacol, 18, 198–207.

16.

Rajagopalan

, Nanjappa

, Raja

, et al. (2016). How does chronic cigarette smoke exposure affect human skin? A global proteomics study in primary human keratinocytes. OMICS, 20, 615–626.

17.

Raoof

, Agaku

, and Vardavas

. (2015). A systematic review of secondhand smoke exposure in a car: Attributable changes in atmospheric and biological markers. Chron Respir Dis, 12, 120–131.

18.

Rybak

, Pao

C-I

, and Pfeiffer

. (2014). Determination of urine caffeine and its metabolites by use of high-performance liquid chromatography-tandem mass spectrometry: Estimating dietary caffeine exposure and metabolic phenotyping in population studies. Anal Bioanal Chem, 406, 771–784.

19.

Salloum

, Asfar

, and Maziak

. (2016). Toward a regulatory framework for the waterpipe. Am J Public Health, 106, 1773–1777.

20.

Stone

, Zhou C; International Association for the Study of Lung Cancer Tobacco Control

Committee

. (2016). Slowing the titanic: China's epic struggle with tobacco. J Thorac Oncol, 11, 2053–2065.

21.

U.S. Department of Health and Human Services. (2006). The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Centers for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Atlanta, GA: Centers for Disease Control and Prevention (USA).

22.

Wright

, Burrow

, and Hurst

. (2016). Should dental teams be doing more to make adolescents aware of the health risks of water pipe tobacco smoking (shisha)?. Br Dent J, 221, 697–699.