Differences in puff topography, toxicant exposure, and subjective response between waterpipe tobacco smoking men and women

Abstract

Waterpipe tobacco smoking (WTS) exposes users to toxicants in much greater amounts than a cigarette. Little is known about how gender affects WTS toxicant exposure and subjective response. Data from three clinical laboratory studies examining WTS conducted between 2008 and 2013 were combined for analysis. Participants (n=99; 38 women) completed a 45-minute WTS session where they smoked tobacco from a waterpipe ad libitum. Puff topography was measured throughout sessions, and plasma nicotine concentration, expired CO, and subjective responses were measured pre- and post-WTS. There was a gender effect for total puff volume with men inhaling a greater volume of smoke, on average (mean=59.9 L, SD=40.7), compared to women (mean=38.8 L, SD=27.8; p<0.01). Men had greater post-WTS mean plasma nicotine concentrations (mean=10.0 ng/ml, SD=7.1) compared to women (mean=6.9 ng/ml, SD=5.2; p<0.05). Post-WTS expired air CO was not associated with gender (men mean=27.6 ppm, SD=18.9; women mean=22.7 ppm, SD=17.0, n.s.). Relative to men, women had higher post-WTS scores for subjective measures of “Nauseous”, “Dizzy”, “Nervous”, “Headache”, and “Heart pounding”. Men and women are exposed to toxicants during WTS, and men may achieve higher post-WTS plasma nicotine concentrations than women, likely as a result of gender differences in the amount of smoke inhaled. However, similar post-WTS expired air CO concentration between men and women and higher ratings of negative subjective responses among women may indicate that other factors beyond smoke exposure (e.g., sensitivity to certain toxicants) may impact toxicant exposure and subjective response to WTS.

Waterpipe tobacco smoking (WTS), also known as hookah, shisha, or narghila, is a centuries old form of tobacco use (World Health Organization, 2005). While typically associated with the Eastern Mediterranean Region (Akl et al., 2011), WTS has increased in frequency in the United States (US), particularly among young adults (Akl et al., 2015; Soule, Lipato, & Eissenberg, 2015). Cross-sectional surveys performed from 2008 to 2012 among young adults and college students in the US between the ages of 18 and 24 indicate past 30-day WTS use ranges from 9.9% to 17.0% (Barnett et al., 2013; Goodwin et al., 2014; Rahman, Chang, Hadgu, Salinas-Miranda, & Corvin, 2014; Sutfin et al., 2011). Importantly, while this trend has slowed (Phillips et al., 2017), more recent national data suggest that WTS is still a major public health issue: 10.7% of young adults reported WTS in the past 30 days and 18.2% reported WTS at least once per month or more in 2013-2014 (Kasza et al., 2017).

The sustained popularity of WTS among young adults in the US represents a major public health concern because of negative health effects associated with WTS (Waziry, Jawad, Ballout, Al Akel, & Akl, 2017). Many of the same negative health effects linked to cigarette smoking are also associated with WTS including likelihood for nicotine/tobacco dependence (Aboaziza & Eissenberg, 2015), lung cancer (Aoun, Saleh, Waked, Salamé, & Salameh, 2013; Koul et al., 2011), and chronic obstructive pulmonary disease (COPD; Mohammad et al., 2013; Tageldin et al., 2012; Waked, Khayat, & Salameh, 2011). Additionally, WTS poses unique acute health effects including carbon monoxide (CO) poisoning (Eichhorn, Michaelis, Kemmerer, Jüttner, & Tetzlaff, 2017). Despite the documented health risks of WTS, many young adults report perceptions of decreased harm associated with WTS (Primack et al., 2008; Sutfin et al., 2011).

There is a need to understand the factors that influence the risk posed by WTS. One factor is puff topography, or smoking behaviors such as puff count, puff volume, or interpuff interval (IPI). During a typical WTS session users inhale many of the same toxicants found in cigarette smoke – such as CO (Eissenberg & Shihadeh, 2009; Monzer, Sepetdjian, Saliba, & Shihadeh, 2008; Shihadeh & Saleh, 2005) and polycyclic aromatic hydrocarbons (PAHs; Monzer et al., 2008; Nguyen et al., 2013; Sepetdjian, Saliba, & Shihadeh, 2010) – but because waterpipe user puffs are so large and numerous, the amount of each toxicant is much greater than is inhaled when smoking a single cigarette. Indeed, the fact that waterpipe tobacco smokers may inhale between 39-172 times the volume of smoke during a single WTS session compared to a cigarette, puts even occasional users at risk for negative health outcomes (Blank et al., 2011; Cobb, Sahmarani, Eissenberg, & Shihadeh, 2012; Eissenberg & Shihadeh, 2009; Maziak et al., 2009). There may be other factors that are associated with toxicant exposure, including gender of the waterpipe smoker. With regard to tobacco cigarettes, men and women smoke differently, resulting in differential exposure to cigarette smoke and associated toxicants (Melikian et al., 2007). In one study, women took more frequent cigarette puffs of smaller puff volume and shorter puff duration compared to men, and when smoke emissions were compared (using analytical chemistry methods), the puff topography of women smokers yielded lower levels of nicotine and select tobacco-related toxicants (e.g., NNK) compared to men (Melikian et al., 2007).

These differences may occur in WTS as well, but to date, only one published report has compared physiological effects (vital signs, spirometry, complete cell count, expired nitric oxide, oxidative stress indices), carboxyhemoglobin (COHb; blood measure of CO exposure), and subjective effects among male and female waterpipe smokers (Hakim et al., 2011). In this study, only CO exposure differed significantly by gender with men having a larger post-smoking COHb level (10.4% vs. 6.8%). The authors attributed this difference is CO exposure to potential differences in lung physiology and/or smoking patterns (Hakim et al., 2011). This foundation of evidence suggests the potential for gender differences in WTS puff topography and toxicant exposure outcomes and supports the investigation of these outcomes as well as others that may differ by gender including the subjective effects associated with WTS.

Understanding gender differences in WTS behaviors may be useful for understanding which subgroups of the WTS population are at greatest risk for toxicant exposure. A greater understanding of WTS behaviors also could aid in developing appropriate prevention approaches and regulatory policies that would limit toxicant exposure and health risks associated with WTS. Clinical laboratory methods are ideal for examining WTS behaviors in a controlled setting, however, the smaller sample sizes used in these studies often limit the ability to conduct between group analyses that are powered sufficiently. Combining data from clinical laboratory studies on WTS that employ similar inclusion criteria and study protocols allows investigation of these important between group effects. Therefore, the purpose of this study was to examine gender differences in puff topography, toxicant exposure (nicotine and CO), and subjective response to WTS by conducting a secondary data analysis of three WTS clinical laboratory studies (Ramôa, Shihadeh, Salman, & Eissenberg, 2016; Cobb et al., 2012; Cobb, Shihadeh, Weaver, & Eissenberg, 2011).

Methods

Participants

Data from 108 participants from three IRB-approved clinical laboratory studies of waterpipe tobacco smokers performed at the same study site between 2008 and 2013 were analyzed for the current study (Study 1: Ramôa et al., 2016; Study 2: Cobb et al., 2012; Study 3: Cobb et al., 2011). Across studies, participants were required to be between 18-50 years of age and report a history of WTS of at least two times (Study 3) or four times per month (Study 1 and 2) for the past six months. Exclusion criteria included self-reported history of chronic health conditions, regular use of prescription medications (other than vitamins or birth control), marijuana use on more than 5 days or alcohol use on more than 25 days (20 days for Study 2) of the past 30 days, and current pregnancy or breastfeeding. Study 1 required that participants had a friend who also reported WTS. Study 2 additionally excluded individuals who reported smoking 5 or more cigarettes per day for the past year and individuals who consumed less than 100 mg caffeine per day over the past year. Study 3 participants had to report smoking at least 5 cigarettes per week, but could not report past month use of other tobacco products besides WTS or intentions to quit smoking. Examination of sample characteristics by study (see Table 1; total n=99 with complete data for all study outcomes) indicated only average WTS sessions/month differed significantly between studies with fewer WTS sessions reported for Study 3 participants (p=0.001). This difference was likely due to the inclusion criterion difference described above. Importantly the distribution of gender did not significantly differ between studies.

Table 1.

Sample characteristics by study and gender (n=99).

Characteristic	Study 1	Study 2	Study 3	p	Women	Men	p
	(n=22)	(n=29)	(n=48)		(n=38)	(n=61)
Gender, N (%)				n.s.			-
Women	14 (63.6)	13 (44.8)	34 (70.8)		38 (100)	-
Men	8 (36.4)	16 (55.2)	14 (29.2)		-	61 (100)
Age in years, M (SD)	21.0 (2.2)	21.6 (2.8)	21.2 (2.4)	n.s.	21.0 (2.2)	21.4 (2.6)
Race, N (%)				NA			0.016
Asian	8 (36.4)	8 (27.6)	6 (12.5)		7 (18.4)	15 (25)
Black/African American	4 (18.2)	7 (24.1)	4 (8.3)		11 (28.9)	4 (6.6)
White	4 (18.2)	10 (34.5)	32 (66.7)		13 (34.2)	33 (54.1)
Other/More than one race	6 (27.3)	4 (13.8)	6 (12.5)		7 (18.4)	9 (14.8)
Hispanic, N (%)	2 (9.1)	4 (13.8)	6 (12.5)	n.s.	6 (15.8)	6 (9.8)	n.s.
Body weight in lbs, M (SD)	152.3 (31.8)	155.7 (33.5)	160.6 (30.6)	n.s.	138.6 (25.2)	169.0 (29.6)	<0.001
WTS sessions/month, M (SD)	13.7 (5.7) *	11.2 (11.4)#	5.8 (4.4) *#	0.001	9.9 (7.4)	8.7 (8.4)	n.s.
Length of WTS in months, M (SD)	18.0 (13.0)	27.7 (19.4)	21.9 (13.3)	n.s.	22.8 (15.5)	22.6 (15.6)	n.s.

Note. ANOVA, t-test, or chi-square used for bivariate comparisons; NA=not applicable due to cell frequency; Symbols (*^#) indicate two means that differed (p<0.05).

Collapsed across studies, there were no significant differences in mean age, Hispanic status, WTS sessions per month, or length of WTS between women and men (see Table 1). Women and men differed significantly with regard to the distribution of race (p=0.016) with more women identifying as African American (29%) compared to men (7%) and fewer women identifying as White (34%) compared to men (54%). Similar percentages of women and men identified as Asian (women: 18%; men: 25%), and other or more than one race (women: 18%; men: 15%). On average, women also weighed significantly less than men (p<0.001).

Materials

For all WTS sessions, a one-hose waterpipe with a chrome body (height=43 cm) screwed into an acrylic base (height=24 cm, volume=1230 ml; www.myasaray.com) was used. The base was filled with 870 ml of water, submerging about 2.5 cm of the body’s conduit. The 6 cm diameter fired ceramic head included 5 holes in the base and was packed with tobacco (10 grams in Study 2 and 15 grams in Studies 1 and 3) and covered with a sheet of aluminum foil perforated using a “screen pincher” (www.smoking-hookah.com) and one piece of lit, “quick lighting” charcoal (Three Kings, Holland; 33 mm diameter) was placed on top of the aluminum foil. The ma’assal tobacco used for each study was the participants’ preferred brand and flavor for Studies 1 and 3 and Tangiers Melon Blend flavor (http://www.tangiers.us./) for Study 2 for experimental control purposes (melon was reported as a preferred flavor by many participants in Studies 1 and 3). The single leather hose was fitted with a topography measurement device and a sterile plastic mouthpiece tip for each session. (www.hookahcompany.com).

Procedures

During each study, participants reported to the laboratory at a pre-determined time after a period of 12-hour tobacco smoking abstinence that was verified by expired air CO of ≤10 ppm. For Studies 2 and 3, a catheter was inserted into a forearm vein which was followed by a 30-minute rest period. After the rest period, a pre-WTS expired air CO measurement was taken, blood was sampled for plasma nicotine concentration, and subjective measures were administered. Approximately 5 minutes later, the WTS session was started (Time 0/Baseline) and a 45-minute ad libitum WTS session commenced. Blood was sampled again at 45 minutes after the onset of the WTS session (results from additional samples taken at 5, 15, and 30 minutes were excluded from the current study) and subjective measures were administered immediately following this blood draw. Expired air CO was recorded 5 minutes after the WTS session was completed. For Study 1, participants also had a 30-minute rest period prior to the session, however, a needle was inserted into a forearm vein and blood was sampled at Time 0/Baseline. Study 1 participants then proceeded with the 45-minute ad libitum WTS session and at the completion of the session a needle was inserted in a forearm vein again and blood was drawn and expired air CO was measured 5 minutes post-session. In summary, across all studies, pre-WTS expired air CO readings were taken approximately 5 minutes before the beginning of the WTS session and then 5 minutes after completion of the 45-minute WTS session. Observations of waterpipe smokers in waterpipe smoking cafés indicate that while there are variations between WTS sessions, 45 minutes represents a typical session (Shihadeh, Azar, Antonios, & Haddad, 2004). Blood draws for plasma nicotine data and subjective ratings were collected approximately 5 minutes before the WTS session started and again at the conclusion of the 45-minute WTS session. Participants were instructed that they could smoke as much or as little as they wanted during the WTS session and were allowed to watch movies throughout.

Puff topography measures

Puff topography was measured using a waterpipe hose with an integrated pressure transducer and attached transducer (Shihadeh, Antonios, & Azar, 2005). This setup allows pressure changes produced by waterpipe user inhalation to be amplified, digitized, and recorded. Accompanying computer software converts the recorded signals into air flow (ml/s) that can be used to calculate average puff volume, duration, number, and IPI.

Toxicant exposure measures

Blood samples were centrifuged and stored at -70°C until they were sent to the Virginia Commonwealth University Bioanalytical Laboratory for analysis of nicotine concentration (as in Breland, Kleykamp, & Eissenberg, 2006a; Naidong, Shou, Chen, & Jiang, 2001). The limit of quantitation (LOQ) for this analysis was 2.0 ng/ml. Plasma nicotine concentration values below this value were replaced with the LOQ. Expired air CO was measured using a breath CO monitor (Vitalograph, Lenaxa, KS).

Subjective measures

Subjective measures were administered via computer (as in Breland, Kleykamp, & Eissenberg, 2006b; Cobb et al., 2015). The Direct Effects of Nicotine Scale (DENS) is a 15-item visual analogue scale developed to assess incidence of nicotine-related side effects (Evans, Blank, Sams, Weaver, & Eissenberg, 2006) and has been used to examine side effects in WTS (e.g., Blank et al., 2011). Data for eight of these side effects were collected across all three included studies (“Nausea”, “Dizzy”, “Lightheaded”, “Nervous”, “Sweaty”, “Headache”, “Excessive salivation”, and “Heart pounding”) and were included in the current analysis.

Data preparation and statistical analysis

Participant characteristics, toxicant exposure data, and puff topography were identified from each individual study and combined into a single dataset for analysis. Only individuals who were not missing outcomes from toxicant exposure (plasma nicotine and expired air CO), subjective measures, and puff topography data were included in the current study. Of the original 108 participants, 9 did not have complete data across all outcomes and were excluded bringing the final sample from Studies 1-3 included in the current study to 99 participants (38 women, 61 men).

Independent t-tests were performed to examine differences in puff topography outcomes by gender. For plasma nicotine, expired air CO, and subjective outcomes mixed ANOVA models were conducted with one repeated factor (time; pre- and post-WTS) and one between subjects factor (gender; women and men) using Type III Sum of Squares to help adjust for the unbalanced gender distribution of the sample. For the two toxicant exposure outcomes (plasma nicotine and expired air CO), the influence of an additional between-subjects covariate, body weight, was examined (due to the differences between men and women noted in Table 1). This covariate did not impact gender effects nor were there any main effects or interactions with gender or time in our mixed ANOVA models; thus, this covariate was dropped for parsimony and to maximize power. For all study outcomes, the influence of study number as a between-subjects covariate was examined in our mixed ANOVA models (due to minor differences between study population/design as described above). The effects of gender on study outcomes were similar with and without study number included; thus, this covariate also was dropped for parsimony and to maximize power. Paired samples t-tests were used to compare mean pre- and post-WTS differences in plasma nicotine, expired air CO, and subjective measures for each gender where significant main effects/interactions were observed. Independent t-test planned contrasts were used to make cross gender comparisons for plasma nicotine, expired air CO, subjective measures at each time point. Because these planned contrasts were orthogonal at each time point, no corrections were made to type I error rate (Keppel, 1991). All analyses were conducted with IBM SPSS Statistics (24) and an alpha level of 0.05 was used for all statistical analyses.

Results

Puff topography

Table 2 presents summary statistics and results of statistical analysis for puff topography outcomes by gender. Across gender, participants took, on average, 85.8 puffs (SD=74.4) with an IPI of 39.9 seconds (SD=25.6), and puff volume of 0.7 L (SD=0.4) resulting in an average total puff volume of 51.8 L (SD=37.6). Descriptively women took more puffs than men with a shorter IPI, but these measures did not differ significantly by gender. Men had a significantly larger mean puff volume and total puff volume inhaled compared to women (ps<0.01).

Table 2.

WTS puff topography measures by gender (n=99).

Puff topography measure	Women	Men	p
	(n=38)	(n=61)
Number of puffs, M (SD)	96.9 (107.0)	78.9 (43.0)	n.s.
Puff volume in L, M (SD)	0.5 (0.3)	0.8 (0.5)	<0.001
Total puff volume in L, M (SD)	38.8 (27.8)	59.9 (40.8)	0.006
Interpuff interval in seconds, M (SD)	37.7 (20.1)	41.3 (28.5)	n.s.

Note: df = (1, 97).

Toxicant exposure effects and subjective effects

Statistical analysis results for the mixed ANOVA models performed for plasma nicotine, expired air CO, and subjective measures are displayed in Table 3. A significant interaction of gender X time was observed [F(1,97)=5.3, p<0.05] for plasma nicotine concentration. No significant differences in plasma nicotine concentration were observed pre-WTS between men (mean=2.1 ng/ml, SD=0.5) and women (mean=2.0 ng/ml, SD=0.3), but post-WTS men (mean=10.0 ng/ml, SD=7.1) had significantly greater mean plasma nicotine concentrations compared to women (mean=6.9 ng/ml, SD=5.2; see Figure 1, Panel A; p=0.022). Post-WTS plasma nicotine concentrations for both genders were significantly elevated relative to pre-WTS (ps<0.001).

Table 3.

Statistical analysis results of mixed ANOVA models for toxicant exposure and subjective measures (n=99).

	Gender		Time		Gender × Time
	F	p	F	p	F	p
Toxicant exposure
Plasma nicotine	5.4	0.022	92.3	<0.001	5.3	0.023
Expired air CO	2.8	n.s.	136.1	<0.001	0.8	n.s.
Subjective measures
Nauseous	5.0	0.027	0.6	n.s.	0.1	n.s.
Dizzy	9.1	0.003	15.8	0.001	4.6	0.035
Lightheaded	3.5	n.s.	33.2	<0.001	0.8	n.s.
Nervous	8.7	0.004	0.0	n.s.	0.8	n.s.
Sweaty	2.0	n.s.	0.1	n.s.	0.3	n.s.
Headache	5.7	0.019	0.9	n.s.	2.0	n.s.
Excessive salivation	1.5	n.s.	2.1	n.s.	0.0	n.s.
Heart pounding	6.2	0.015	3.6	n.s.	2.2	n.s.

^aGender (male, female) included as a between-subjects covariate with time as a within-subjects factor (2 levels; pre-WTS, post-WTS); df_gender = (1, 97); df_time = (1, 97); df_{gender x time} = 1, 97).

Figure 1.

Pre- and Post-WTS plasma nicotine concentration and expired air CO. Filled symbols indicate differences from baseline and * indicates differences between men and women.

A significant main effect of time for expired air CO was observed with significant increases in expired air CO from pre-WTS to post-WTS for both genders [F (1,97)=136.1, p<0.001]; see Figure 1, Panel B). Pre-WTS, men had slightly higher CO levels (mean=4.5 ppm, SD=2.8) than women (mean=2.9 ppm, SD=2.4; p=0.003), but no significant differences in CO levels were observed between men and women post-WTS (men mean=27.6 ppm, SD=18.9; women mean=22.7 ppm, SD=17.0).

There was a significant main effect of gender on “Nauseous”, “Dizzy”, “Nervous”, “Headache”, and “Heart pounding.” Women rated each of these subjective measures higher than men post-WTS (ps<0.05). There was a significant main effect of time for “Dizzy” and “Lightheaded”, [Fs(1,97)≥15.8, ps<0.05)] and one significant interaction of gender X time for “Dizzy”, [F(1,97)=4.6, p<0.05)]. For scores for “Dizzy”, women increased significantly from pre-WTS (mean=5.3, SD=14.2) to post-WTS (mean=14.3, SD=19.5; p=0.003), but there was no significant difference by time for men. For scores for “Lightheaded”, women increased from pre-WTS (mean=6.13, SD=12.4) to post-WTS (mean=22.2, SD=24.9; p<0.001). Men also increased significantly on “Lightheaded” from pre-WTS (mean=2.9, SD=7.7) to post-WTS (mean=14.7, SD=22.9; p<0.001).

Discussion

In the current study, during a 45-minute WTS session, on average men inhaled significantly more smoke than women (1.5 times more; 59.9 L vs. 38.8 L). Men also had significantly higher plasma nicotine concentrations post-WTS compared to women, but post-smoking expired air CO levels did not differ significantly by gender (men: 27.6 ppm; women: 22.7 ppm). Women had significantly higher mean ratings on the subjective effect item “Dizzy” compared to men. Importantly, though initial analyses indicated gender differences on nicotine exposure, this effect was no longer present after running a post-hoc mixed ANOVA model including total puff volume as an additional between-subjects covariate (men adjusted mean=9.3 ng/ml, women adjusted mean=8.1 ng/ml; n.s.) suggesting that gender differences in plasma nicotine were likely caused by the greater amount of smoke inhaled by men. Conversely, women had higher post-WTS negative subjective response ratings (i.e., Dizzy) compared to men that remained when controlling for total puff volume. This finding suggests that other factors may impact differences in subjective responses to WTS between men and women.

In comparison to the effects of gender and WTS-associated CO exposure previously reported (Hakim et al., 2011), while expired air CO was higher among men it was not significantly so in this study. This previous report also included individuals with a much higher WTS rate (62% smoked waterpipe 3 or more times per week) and had other methodological differences from the current design (group smoking allowed; 30-minute limit of smoking session). Other differences between the studies collapsed for this analysis may have influenced the results obtained. Study 2 required that participants use “melon” flavor whereas Studies 1 and 3 allowed participants to use a preferred flavor, most of whom chose some type of fruit flavor. Study 3 included individuals who used smoked WTS at significantly lower frequencies than Study 1 and were at least weekly cigarette smokers. Importantly, when study number was included as a covariate in our mixed models, effects for gender on study outcomes were largely unaffected indicating that gender effects remained after controlling for study number, but power and sample size may be limiting factors for examining these types of interactions. Still, there may be other unmeasured factors that may influence differences between men and women in terms of smoking behaviors, toxicant exposure, and subjective responses such as differences in lung capacity or total blood volume (Hakim et al., 2011). Future research that examines these variables in addition to puff topography may account for greater variance in toxicant exposure and subjective response outcomes associated with WTS.

The results of this study have important policy and clinical implications. Men and women inhale large volumes of toxicant laden smoke during a single WTS session while also reporting relatively low negative subjective effects, an occurrence that is concerning given the health risks that can occur after even a single WTS session (Eichhorn et al., 2017). Furthermore, there may be some groups, such as men in the current study, that are able to expose themselves to higher amounts of toxicants and still report relatively low adverse experiences while engaging in WTS. Indeed, greater time spent in a waterpipe café has been shown to be associated with increased expired CO post-WTS (Leavens et al., 2017) indicating many waterpipe smokers are able to continue to smoke despite increased toxicant exposure. Therefore, in addition to prevention efforts that encourage individuals to decrease their WTS frequency to limit smoker exposure to waterpipe smoke toxicants and associated health risks, efforts that limit the amount of time waterpipe smokers can spend in a waterpipe café may have potential for positive public health impacts.

Some of the appealing qualities of WTS include the positive physical effects associated with WTS including feelings of being relaxed, “buzzed”, dizzy, lightheaded, or “high” (Braun, Glassman, Wohlwend, Whewell, & Reindl, 2012; Sharma, Clark, & Sharp, 2014; Soule, Barnett, Curbow, Moorhouse, & Weiler, 2015). While some of these side effects of WTS may be attributable to nicotine exposure, a placebo-controlled clinical laboratory study found that WTS using typical waterpipe tobacco and nicotine-free “herbal” tobacco resulted in similar subjective ratings of dizziness (Blank et al., 2011). This finding suggests that subjective effects of dizziness may be attributable to smoke constituents other than nicotine such as CO. Higher ratings of adverse side effects of WTS including “Dizzy” were observed post-WTS in women relative to men. These side effects may have resulted in less inhalation of waterpipe smoke in response to these subjective feelings (possibly having achieved a desired level of dizziness but wanting to prevent increases in other feelings such as headache or nausea). Conversely, men in the study may have also been seeking desired physical effects and therefore inhaled greater volumes of smoke. Perhaps because these men were able to inhale greater volumes of smoke because they did not receive the same negative feedback of adverse side effects as the WTS women. Due to the measurement time course in this study, the directionality of subjective effects and puffing behaviors is not able to be assessed. This possible explanation for the observed differences in select physiological outcome measures, subjective measures, and puffing behaviors between men and women should be examined further.

There also may be differences in reasons for WTS between women and men that result in different puffing behaviors and in turn differences in physiological outcomes as well as differences in subjective responses to WTS. While not examined in the current study, social factors may influence WTS behaviors and outcomes. For instance, past research suggests the gender expression of masculinity is associated with the prevalence of smokeless tobacco and cigar use (Roberts et al., 2014). Masculinity may be associated with WTS behavior as well. Additionally, the social aspect of smoking may be less important for men compared to women. In an observational study of waterpipe café patrons, only men were seen engaging in WTS alone, and women were always a part of a group (Carroll et al., 2014). In this same study, men were also observed engaging in more leadership roles (e.g., “One guy seems to be in charge. He orders the next flavor for the group.”) or engaging in excessive use (“He seemed like he was in a hookah-induced stupor.”). Some observations indicated that social connections were important for women (“The girls, when they first arrive, take pictures of each other and the whole group on their phones.”; Carroll et al., 2014). WTS behaviors also may be influenced by group size with more frequent puffing behaviors being observed in smaller WTS groups (Blank, Brown, Goodman, & Eissenberg, 2014). However, women may be more sensitive to the aversive effects of nicotine compared to men (Perkins & Karelitz, 2015). More research is needed to confirm gender differences in reasons for and appeals of WTS, but these potential differences may provide insight into differences in WTS behaviors and outcomes. In particular, studies that consider the many factors that are associated with WTS in natural settings are likely to provide data that are more representative of typical WTS toxicant exposures and subjective responses.

The findings from this study should be taken with consideration of several limitations. Analyses were based on measurements taken pre- and post- a 45-minute WTS session. Though this time period is often used to describe a “typical” WTS session, WTS sessions can sometimes last much longer. A recent field study indicated WTS café patrons spend on average close to 2 hours in waterpipe cafés (Leavens et al., 2017). This same study reported higher toxicant exposure from WTS was associated with alcohol use, a behavior that co-occurs commonly with WTS in the US (Soule, Barnett, & Curbow, 2012; Soule, Barnett, et al., 2015). Similarly, WTS typically occurs in groups (Blank et al., 2014; Carroll et al., 2014; Maziak, 2014), though the current study examined WTS among singletons. Lab assessments that resemble real-world conditions more closely may provide more accurate representations of user behaviors and outcomes. Additionally, while this analysis combined results from three clinical laboratory studies to obtain a sample size needed to examine gender effects, additional participants would have increased statistical power and allowed for further subgroup analyses. More men than women were included in the current analysis, and although the results highlight within group differences when considering the effect of gender, future analyses would benefit from equal number of participants in each group. A final limitation is that gender was dichotomized; future studies of gender differences may need to consider non-cisgender categories that reflect the diversity of the underlying population.

Conclusions

Women and men engage in different smoking behaviors during WTS which may influence and be influenced by physiological and subjective responses to WTS. Thus, it may be useful to take into account gender differences in reasons for WTS and WTS behaviors in developing prevention and cessation efforts. Other opportunities to reduce risk may be to include limiting promotions such as “all-you-can-smoke” in waterpipe cafés to help limit toxicant exposure. While there may be challenges associated with addressing WTS via policy (Salloum, Asfar, & Maziak, 2016; Sutfin, Soule, McKelvey, & Jenson, 2017), the present environment that supports tobacco use prevention including the US Food and Drug Administration’s expanded authority to regulate WTS represents an opportunity to reduce WTS-related morbidity and mortality.

Public Health Significance.

In a single waterpipe tobacco smoking (WTS) session, women and men do not differ on total puffs taken. Men on average take larger puffs during a WTS session likely resulting in greater nicotine exposure. However, women rated negative subjective responses to WTS higher than men. These findings indicate that while other factors may influence WTS outcomes, smoking behaviors differ between women and men and may influence and be influenced by physiological and subjective responses to WTS smoking.