Natural Course of Nicotine Dependence Among Adolescent Waterpipe And Cigarette Smokers

Abstract

Purpose:

Waterpipe (WP) smoking patterns and setting can result in a unique trajectory of nicotine dependence (ND) compared to cigarette smoking. This longitudinal study compared the development of ND symptoms among adolescent WP and cigarette smokers.

Methods:

A cohort of 647 8^th-and-9^th-graders in Lebanon were followed over 5 years. This study is based on 283 current-exclusive-WP and 146 current-exclusive-cigarette smokers. Kaplan-Meier survival analyses were conducted to evaluate 50% cumulative probability for the development of initial Hooked-on-Nicotine-Checklist (HONC) symptoms and the International-Classification-of-Diseases, 10th revision (ICD-10) ND.

Results:

An initial HONC symptom was endorsed by 59% of WP and 50% of cigarette smokers after smoking onset. Among those, 50% of both WP and cigarette smokers did so within 9.7 and 18.5 months, respectively. Approximately 28% of WP smokers and 22% of cigarette smokers developed ICD10 ND. Among those, 50% of both WP and cigarette smokers did so within 15 and 22 months, respectively. The most common 1^st-4^th-ICD10 criteria reported by WP smokers were “a strong desire to use tobacco”, “difficulties in controlling tobacco-taking behavior”, “neglect of alternative pleasure”, and “use despite harm”. The most common 1^st-4^th-ICD10 criteria reported by cigarette smokers were “a strong desire to use tobacco”, “difficulties in controlling tobacco-taking behavior”, “withdrawal” and “tolerance”.

Conclusions:

Compared to adolescent cigarette smokers, initial ND symptoms and ICD10 ND can develop sooner after starting to smoke and progress more rapidly among adolescent WP smokers. Developing, implementing, and evaluating intervention programs with adolescent WP smokers should be guided by the WP-specific trajectory of ND.

Introduction

Nicotine dependence (ND) is the fundamental reason that people persist in using tobacco products (1), which kill more than 8 million people every year globally(2). About 80% of the world’s 1.1 billion tobacco users live in low- and middle-income countries, where the burden of tobacco-related morbidity and mortality is heaviest(3). Contributing to this burden is waterpipe (WP; a.k.a., shisha, hookah, and narghile), a centuries-old method of smoking tobacco that has increased dramatically around the globe in the last twenty years especially among youth and young adults(4). WP smoking causes ND(5) and more than doubles the odds of initiating cigarette smoking(6).

According to the Tenth Revision of the International Classification of Diseases and Health Problems (ICD-10), dependence refers to a group of physiological, behavioral, and cognitive phenomena in which the use of a substance (e.g., tobacco products) takes on a much higher importance for a user than other behaviors that once had greater value(7). For example, after using tobacco products (e.g., WP and cigarette), nicotine increases the release of neurotransmitters (e.g., dopamine) that creates the pleasurable sensations which make smoking highly addictive, reinforces continued tobacco use, and makes person dependent on nicotine(1,7).

WP smokers have a notable increase in plasma nicotine concentration after smoking WP and exhibit behavioral symptoms of ND (e.g. craving, withdrawal) during abstinence that are relieved by smoking (5, 8). Although many WP-associated ND symptoms are identical to those seen with cigarette smoking, the development of ND symptoms in WP smokers is likely influenced by its unique features and use patterns (8). For example, long duration of smoking sessions (averaging an hour), limited access and portability, and unique sensory cues can shape a distinctive pattern of ND development (4, 8,9). Identifying this pattern and how it differs from cigarettes is essential to develop and adapt existing cessation and prevention interventions to curb WP smoking among youth. Moreover, WP is characterized by intermittent, nondaily smoking that is different from the typical daily smoking for cigarettes which can be associated with unique smoking patterns and ND trajectories for WP smokers compared with cigarette smokers(5,8,9). Consequently, the manifestations and sequence of ND symptoms may vary among adolescents WP and cigarette smokers, which is key for identifying and characterizing ND among this population at an early stage. Therefore, monitoring the natural process of ND development among youth WP and cigarette smokers will provide important insights for future interventions aimed at curbing WP use among adolescent smokers.

In Lebanon, WP smoking among youth has reached epidemic proportions, with the prevalence of current WP smoking being more than three times (37.2%) that of cigarette smoking (11.2%) (10, 11). Therefore, this country is in urgent need for WP smoking interventions, and at the same time provides an opportunity to study the trajectory of ND development among adolescent WP and cigarette smokers. In a previous report using baseline data from the Waterpipe Dependence in Lebanese Youth (WDLY) study, Bahelah et al(12) reported that ND symptoms among adolescent WP smokers develop at low levels of consumption and frequency of use. However, the longitudinal trajectory of ND development among young WP smokers and how it differs from cigarettes has not been studied yet. The WDLY prospective study was designed to map the natural course of ND development among young WP and cigarette smokers and will be used to characterize the early and most common symptoms of ND and their sequence in adolescent WP and cigarette smokers.

Methods

1. Sample

For this study, data were obtained from the WDLY, a 5-year prospective study of 647 adolescent WP and cigarette smokers and never-smokers. WDLY data were collected from 38 public and private schools using a list from the Lebanese Ministry of Education beginning when the students were in 8^th or 9^th grade. To determine eligibility, a brief in-class, self-administered recruitment survey about students’ smoking status was administrated. In order to compare cigarette and WP smokers in terms of ND pattern, students were eligible to participate if they currently (past 30 days) smoked either cigarettes or WP, but not both. We also included non-smokers who were susceptible to smoking initiation in the future. Data were collected in six survey waves between May 2015 and December 2017 at 6-month intervals (Figure 1). A 7^th wave was collected in March 2019. A full description of the methodology of WDLY is available elsewhere(12). The Institutional Review Boards of Florida International University and the American University of Beirut approved the study protocol.

Figure 1.

Data collection process for the Waterpipe Dependence in Lebanese Youth (WDLY) study, 2015–2019. ^† the numbers and percentages represent participants who took part in the interview for the same wave. *Note: 149 participants were added to the study during waves 4, 5, and 6; including WP smokers (n=9), cigarette smokers (n=104) as well as susceptible nonsmokers (n=36) with the retention rate of 78% (n=116) at wave 7. Note: This study characterizes the natural course of ND symptoms among exclusive WP and exclusive cigarette users, therefore, nonsmokers and subjects who smoked both WP and cigarettes or other tobacco products at various times were excluded from the analysis.

Figure 1 displays a detailed summary of data collection and a timeline for waves 1 to 7. At baseline (wave 1), a total of (498) 8^th-and-9^th-graders were enrolled. Of these, 81.5% (n=406) were retained through 7 waves of data collection. Due to the small number of cigarette smokers at baseline (n=32), 104 current cigarette smokers were added to the study during waves 4–6 to provide statistical power to detect differences between WP and cigarette smokers in regard to the timing of ND symptoms. Overall, 149 participants were added to the study during waves 4–6, including cigarette smokers (n=104), WP smokers (n=9), and nonsmokers (n=36), with a retention rate of 78% (n=116) at wave 7.

Between waves 1 and 7, 647 (498 at baseline + 149 added at follow-ups) participants completed at least one wave of the WDLY study. Of the 647 participants, 429 (66.3%) who reported WP-only (n=283) or cigarette-only (n=146) in the past 30-days preceding any interview time were included in the current analyses. Non-smokers (n=87) and participants who switched from WP to cigarettes or vice versa or were both WP and cigarette users (n=126) and other tobacco product users (n=5; e-cigarettes/e-hookah) were excluded from the analysis.

2. Measures

2.1. The Hooked-on-Nicotine-Checklist (HONC)

The HONC (12,13) was employed to measure loss of autonomy over tobacco use as one of the main outcomes of the present study. The HONC is a validated and widely used instrument that assesses whether any of 10 symptoms of dependence have been experienced (indicated by yes or no responses), including an unsuccessful quit attempt, finding it really hard to quit, and strong cravings to smoke (13). The HONC’s focus on symptoms allows for its use with any nicotine delivery method and is sensitive to detecting initial symptoms of ND among youth (12,13). The Arabic version of HONC used in the WDLY study showed acceptable internal reliability among WP smokers (α=0.74) (12).

2.2. WHO’s International Classification of Diseases, 10th version (ICD-10)

The ICD-10 uses a cluster of behavioral, cognitive and physiological phenomena that develop after repeated use of tobacco products to diagnose ND syndrome(14, 15). The ICD-10 includes 19 dichotomous (yes, no) items with 6 criteria (Table 2). The ICD-10 diagnosis of ND requires the clustering of 3 or more criterion symptoms of dependence during a 1-year period(15). The ICD-10 has been validated among adolescent WP smokers in this cohort using an Arabic version of the 19-item instrument that had acceptable internal reliability (α=0.76)(12). We selected the HONC and ICD-10 to assess ND in our study because these two measures are not tobacco method-specific, making them appropriate for examining ND trajectories irrespective of the type of tobacco product (16,18). See supplementary materials-Appendix I for ICD10 and HONC items.

Table 2.

Time (months) to 50% (median) cumulative probability of endorsement of HONC symptoms since first puff on WP/cigarette

	Proportion reporting symptoms %†			Time to 50% cumulative probability, Estimate (95% CI), months
HONC Symptomsa	Waterpipe smokers (n=283)	Cigarette smokers (n=146)	Chi-Square p-value	Waterpipe smokers	Cigarette smokers	Log-Rank p-value
				First puff on WP	First puff on cigarette
Diminished autonomyb	58.7	50.0	0.054	9.7 (7.5–11.8)	18.5 (14.7–26.3)	<0.001
Craving	38.5	33.6	0.313	14.9 (11.9–20.2)	27.4 (18.4–30.5)	0.010
Felt addicted	36.7	34.9	0.710	14.1 (11.5–17.9)	24.4 (16.7–35.5)	<0.001
Failed quit attempt	29.0	26.7	0.622	17.1 (13.8–21.1)	20.6 (12.1–31.6)	0.179
Strong urge to smoke	28.6	19.9	0.049	19.2 (13.9–23.0)	27.6 (16.7–36.5)	0.029
Restless, anxious, or nervous	21.6	18.5	0.457	17.8 (14.2–23.0)	25.7 (17.9–33.7)	0.261
Needed a WP/cigarette	17.7	21.2	0.423	21.0 (11.9–27.5)	24.6 (16.7–32.1)	0.643
Hard to quit	17.3	17.1	0.960	21.6 (14.5–29.4)	28.8 (18.6–40.7)	0.749
Feeling irritable	16.6	16.4	0.964	17.0 (12.4–26.6)	24.1 (16.7–36.5)	0.213
Hard to refrain	11.7	21.9 *	0.005	14.5 (11.1–22.4)	33.3 (21.2–39.1)	0.010
Impaired concentration	9.9	15.8	0.076	28.4 (18.1–37.5)	32.7 (18.6–40.7)	0.298
ICD10 criteria 1: A strong desire or sense of compulsion to use tobacco	30.7	25.3	0.242	12.1 (9.6–15.1)	18.5 (15.6–32.8)	0.025
ICD10 criteria 2: Difficulties in controlling tobacco-taking behavior in terms of its onset, termination, and levels of use	29.3	24.0	0.239	16.1 (13.1–18.6)	28.6 (18.4–31.6)	0.068
ICD10 criteria 3: A physiological withdrawal state	18.0	16.4	0.683	18.6 (13.2–28.7)	24.0 (14.1–31.9)	0.479
ICD10 criteria 4: Evidence of tolerance	15.5	11.6	0.273	15.8(11.1–23.0)	19.4 (12.3–33.3)	0.104
ICD10 criteria 5: Neglect of alternative pleasure	25.1	17.1	0.061	14.8 (11.06–19.5)	21.7 (12.3–35.4)	0.083
ICD10 criteria 6: Use despite harm	16.6	8.2 *	0.017	17.5 (12.1–22.1)	27.8 (3.0–46.7)	0.273
Full ICD10 ND syndrome	27.9	21.2	0.133	15.1 (13.1–18.2)	22.3 (15.6–31.7)	0.045

^†Proportion of the cohort reporting each symptom.

^*Chi-square tests indicate the significant differences between proportion of each HONC item/or each ICD10 Criteria for WP and cigarettes.

^aParticipants reported HONC symptoms by responding yes/no to the questions: Craving; “Have you ever had strong cravings to smoke WP/cigarette?” Felt addicted; “Have you ever felt like you are addicted to smoking WP/cigarette?” Failed quit attempt; “Have you ever tried to quit smoking but could not do it?” Strong urge to smoke; “Did you feel a strong need or urge to smoke WP/cigarette?” Restless, anxious, nervous; “Did you feel restless, anxious, or nervous because you could not smoke?” Needed a WP/cigarette; “Have you ever felt like you really needed a WP/cigarette?” Hard to quit; “Do you smoke now because it is really hard to quit?” Feeling irritable; “Did you feel more irritable because you could not smoke WP/cigarette?” Hard to refrain; “Is it hard to keep from smoking in places where you are not supposed to, like school?” Impaired concentration; “Did you find it hard to concentrate because you could not smoke WP/cigarette?”

^bEndorsing ≥ one HONC symptoms.

The dates of first puff on WP/cigarette and onset of each HONC symptom and ICD10 criteria were recorded. The first puff on WP/cigarette was collected with the question “Have you ever smoked a WP/cigarette even just one puff or two?” Possible responses were no/yes; if yes, the date of the first puff was recorded as day/month/year. We also recorded the frequency of use (days smoked per month) and quantity (the number of WP/cigarettes smoked per month) at the time of the appearance of each individual HONC items and ICD10 criteria.

2.3. Frequency and quantity of use

To maintain consistency with previously published work (16) and simplify the presentation of results, we categorized the quantity of cigarette use per month as 1, 2, 3–4, 5–9, 10–19, 20–99, 100 or more. The quantity of WP use was categorized as smoking ≤1 bowl/head of WP per month, 2, 3–4, 5–9, 10–19, 20–29, or ≥30. The frequency of WP or cigarette smoking per month was adopted from the Global Youth Tobacco Survey (17), which was categorized as smoking a cigarette or WP on 1 or 2 days per month, 3–5, 6–9, 10–19, 20–29, or all 30 days.

To calculate the time intervals (i.e. number of months) between first puff on WP/cigarette and the appearance of each HONC item, ICD10 criteria, and full ICD10 ND syndrome, the date of the first puff was subtracted from the date when the appearance of 1^st symptom/criteria or ICD10 ND syndrome took place (12).

Statistical analysis

Baseline characteristics were compared between WP and cigarette smokers using frequencies and percentages for categorical variables and mean ± standard deviation (SD) for continuous variables. Categorical variables (proportions) were compared using chi-squared tests, while t-test or Mann-Whitney U/Kruskal-Wallis tests, where applicable, were used to test group differences in continuous variables.

The time-to-event intervals (months) were computed using Kaplan-Meier (K-M) survival analyses stratified by smoking group (WP and cigarette smokers). Because of the positive skew frequently seen with follow-up times, medians are often a better indicator of “average” time-to-event intervals (18). Therefore, K-M product-limit estimates focused on the time to 50% (median) and corresponding 95% CI of cumulative probability. The log-rank test was used to evaluate differences between WP and cigarette smokers in time (months) until the appearance of initial symptom of HONC, each individual HONC symptom, each ICD10 criteria, and full ND syndrome since first puff.

For this cohort, each participant could endorse 10 HONC items and 6 ICD10 criteria at different time points (i.e., dates) and therefore, produce a sequence of the 1^st to 10^th HONC symptom and 1^st to 6^th ICD10 criteria experienced. However, to simplify the presentation of results, we reported the percentage of individuals who endorsed each HONC item and ICD10 criteria only for 1^st to 4^th symptoms that appeared after the first puff on WP or cigarette. We obtained the quantity and frequency of WP and cigarette use at the time of appearance of initial HONC symptom and full ND syndrome. To control the effect of age, gender, and school type on timing differences between WP and cigarette smokers, we performed a sensitivity analysis using the Cox proportional-hazards model through the PHREG Procedure in SAS/STATv14.2 (19, 20). This sensitivity analysis indicated that age differences at onset of smoking and at the time of entry to study, as well as gender and school type, did not affect the observed timing differences between WP and cigarette smokers in experiencing initial HONC symptom or developing the full ICD10 ND during follow up (see supplementary materials-Appendix II). A two-tailed α=0.05 was set a priori for all analyses, which were computed using SAS/STATv14.2 (SAS Institute Inc., NC; USA) and SPSS v.22 (IBM Corp., Armonk, NY, USA).

Results

1.1. Demographic characteristics

Table 1 shows selected characteristics of current (past 30-day) WP and cigarette smokers at the time of study entry. Overall, 174 (61.5%) of WP smokers and 26 (17.8%) of cigarette smokers were females, with a mean age of 13.3 years (SD, 1.9) and 13.8 years (SD, 2.0, p<.004), respectively.

Table 1.

Selected characteristics of adolescent WP and cigarette smokers at entry into the WDLY study^*

	Total (n=429)	Waterpipe (n=283)	Cigarette (n=146)	p-value
Age, years, mean (SD)	14.7 (1.6)	14.1 (1.3)	15.8 (1.5)	<0.001
Sex, n (%)				<0.001
Male	229 (53.4)	109 (38.5)	120 (82.2)
Female	200 (46.6)	174 (61.5)	26 (17.8)
School type, n (%)				0.003
Public	224 (52.2)	134 (47.3)	90 (61.6)
Private	205(47.8)	149 (52.7)	56 (38.4)
School grades, n (%)				<0.001
8th	164 (38.4)	141 (49.8)	29(17.4)
9th	154 (36.1)	131 (46.3)	29(17.4)
10th	24 (5.6)	2(0.7)	28(16.8)
11th	59 (13.8)	5 (1.8)	59(35.3)
12th	26 (6.1)	4(1.4)	22(13.2)
BMI, mean (SD)	21.7 (4.2)	21.3 (4.0)	23.1 (4.4)	0.015
Age first smoked tobacco, mean (SD)	–	13.3 (1.9)	13.8(2.0)	0.004
Age first HONC item was experienced mean (SD)a	–	14.0(1.6)	15.6 (1.6)	<0.001
Age of attaining ICD-10 ND	–	14.5 (1.4)	15.6 (1.7)	0.002

HONC, Hooked on Nicotine Checklist; ICD-10 ND, International Classification of Diseases, Nicotine Dependence (≥3 criteria),

^*274 WP smokers and 42 Cigarette smokers entered the study in Wave 1, 9 WP smokers were added to the study during waves four (n=3), five (n=1), and six (n=5). 104 cigarette smokers were added to the study during waves four (n=70), five (n=12), and six (n=22).

^aAge first HONC item was experienced and age of attaining ICD-10 ND were among 166 WP and 73 cigarette smokers who experienced initial HONC symptom and attained ICD10 ND.

1.2. Course of symptom development

The proportion of participants reporting each HONC symptom and ICD10 criteria since first puff on WP or cigarette and the number of months after smoking initiation when the cumulative probability of developing each symptom was 50% are shown in Table 2.

1.2.1). Course of symptom development (HONC)

There was a marginally significant difference in the proportion of WP and cigarette smokers who experienced an initial HONC symptom (WP=58.7% (n=166) vs cigarette=50.0% (n=73), p=0.054).

At least one HONC symptom was reported by 50% of symptomatic WP and cigarette smokers within 9.7 and 18.5 months since smoking initiation, respectively (p<0.001). (Table 2). Similarly, the time since smoking initiation to the point at which 50% of participants developed individual HONC symptoms such as “having a strong craving to smoke”: (WP: 14.9 vs cigarette: 27.4 months, p=0.010); “feeling addicted”: (WP:14.2 vs cigarette: 24.4 months, p<0.001); “having a strong urge to smoke” (WP: 19.2 vs cigarette: 27.6 months, p=0.029), and “hard to refrain from smoking where it is not allowed” (WP: 14.5 vs cigarette: 33.3 months, p<0.010) were significantly shorter among WP smokers compared to cigarette smokers (Table 2).

1.2.2). Course of symptom development (ICD10)

ICD10 dependence criteria were met by 27.9 % (n=79) of WP smokers and 21.2% (n=31) of cigarette smokers (P=0.082). Among those with ICD10 ND, this milestone was reported by 50% of WP and cigarette smokers within 15.1 and 22.3 months since smoking initiation, respectively (p=0.045) (Table 2). The time since smoking initiation to the point at which 50% of participants attained individual criteria of ICD 10 was significantly different between WP and cigarette smokers only for the criteria #1: “A strong desire or sense of compulsion to use tobacco” (WP: 12.1 vs cigarette: 18.5 months, p=0.025) (Table 2).

1.3. Sequence of ND symptoms

Figure 2 (panel A&B) shows variability in the order of appearance of HONC symptoms based on the percentages of participants who endorsed each HONC item as their 1^st to 4^th symptoms after smoking initiation. Among 166 symptomatic WP smokers, “having a strong craving to smoke WP” was the first most commonly reported symptom (by 33.7% of WP smokers), followed by “feeling addicted” (17.5%), “failed quit attempt” (10.2%), and “strong urges to smoke WP” (7.2%). Among 73 symptomatic cigarette smokers, “feeling addicted” was the first most common presenting symptom (39.7%), followed by “having a strong craving to smoke a cigarette” (20.5%), “failed quit attempt” (12.3%), and “feeling irritable” (11%).

Figure 2.

Order of appearance of 10 HONC symptoms and 6 ICD10 criteria. The bars display percentages of 166 WP (panel A) and 73 cigarettes (panel B) smokers who endorsed HONC symptoms as their first, second, third, and fourth symptoms based on temporality sequence since smoking initiation. The bars also display percentages of 79 WP (panel C) and 31 cigarettes (panel D) smokers who attained each ICD10 criteria as their first, second, third, and fourth ND symptoms based on temporality sequence since smoking initiation. Six criteria of ICD-10 includes criteria 1: “a strong desire or sense of compulsion to use tobacco”; criteria 2: “difficulties in controlling tobacco-taking behavior in terms of its onset, termination, and levels of use”; criteria 3: “a physiological withdrawal state”; criteria 4: “evidence of tolerance”; criteria 5: “neglect of alternative pleasure”; and criteria 6: “use despite harm”.

Figure 2 (panel C&D) shows variability in the order of developing ICD10 criteria based on the percentages of participants who attained each ICD10 criteria as their 1^st to 4^th symptoms after smoking initiation. Among 79 WP smokers who met the criteria for ICD10 ND (panel C), “a strong desire or sense of compulsion to use tobacco” was the first most commonly reported symptom (by 51.9% of WP smokers), followed by “difficulties in controlling tobacco-taking behavior in terms of its onset, termination, and levels of use” (41.8%), “neglect of alternative pleasure” (19.0%), and “use despite harm” (11.4%). Among 31 cigarette smokers who developed ICD10 ND (panel D), “a strong desire or sense of compulsion to use tobacco” was the first most common presenting symptom (51.6%), followed by “neglect of alternative pleasure” (35.5%), “a physiological withdrawal state” (25.8%), and “evidence of tolerance” (9.7%).

1.4. Frequency and quantity of use

Tables 3 & 4 display the cumulative percentages of quantity and frequency of use at the time of experiencing the initial HONC symptom and attaining ICD10 ND. As shown in Table 3, among the 166 symptomatic WP smokers who experienced one or more HONC symptoms, 14.5% presented an initial HONC symptom after smoking ≤1 WP (head/bowl)/month. Only 5.1% of 79 WP smokers who attained ICD10 ND did so after smoking ≤1 WP (head/bowl)/month. Among the 73 symptomatic cigarette smokers who experienced one or more HONC symptoms, 4.1% experienced the initial HONC symptom after smoking only 1–2 cigarettes/month. Only 0.7% of 31 symptomatic cigarette smokers who met ICD10 ND criteria did so after smoking only 1–2 cigarettes/month.

Table 3.

Cumulative percentages of quantity of use at the time of experiencing initial HONC symptom and attaining ICD10 ND among adolescent waterpipe and cigarette smokers

	WP smokers			Cigarette smokers
Quantitya	HONC (n=166)b	ICD10 (79)c	Quantitya	HONC (n=73)d	ICD10 (n=31)e
≤1	14.5	5.1	1	1.4	0.7
2	28.3	12.7	2	4.1	0.7
3 to 4	51.8	25.3	3–4	9.6	0.7
5 to 9	67.5	45.6	5–9	13.7	19.9
10 to 19	83.7	63.3	10–19	20.5	19.9
20 to 29	89.8	69.6	20–99	35.6	41.1
≥30	10.2	29.4	≥100	64.4	58.9

^aQuantity of use (the number of WP/cigarettes smoked in the past month) at the time of the appearance of initial HONC symptom and attaining ICD10 ND.

^b166 symptomatic WP smokers who endorsed at least one HONC symptom.

^c79 symptomatic WP smokers who attained ICD10 ND,

^d73 symptomatic cigarette smokers who endorsed at least one HONC symptom.

^e31 symptomatic WP smokers who attained ICD10 ND.

Note: Percentages are based on cumulative order.

Table 4.

Cumulative percentages of frequency of use at the time of experiencing initial HONC symptom and attaining ICD10 ND among adolescent waterpipe and cigarette smokers

	HONC comparing WP to cigarettes		ICD10 ND comparing WP to cigarettes
Frequencya	WP (n=166)b	Cigarettes (n=73)c	WP (79)d	Cigarettes (n=31)e
1–2 days	25.9	5.5	15.2	3.2
3–5	55.4	20.5	30.4	3.2
6–9	66.9	21.9	44.3	6.5
10–19	80.7	34.2	64.6	12.9
20–29	87.3	42.5	70.9	25.8
30 days*	12.7	57.5	29.1	74.2

^aFrequency of use (days smoked per month) at the time of the appearance of initial HONC symptom and attaining ICD10 ND.

^b166 symptomatic WP smokers who endorsed at least one HONC symptom.

^c73 symptomatic cigarette smokers who endorsed at least one HONC symptom.

^d79 symptomatic WP smokers who attained ICD10 ND.

^e31 symptomatic WP smokers who attained ICD10 ND.

Note: Percentages are based on cumulative order.

As shown in Table 4, approximately 88% of 166 symptomatic WP smokers experienced an initial HONC symptom before daily smoking, and about 71% of 79 WP smokers who met ICD10 ND criteria did so before daily smoking. While approximately 43% of 73 symptomatic cigarette smokers experienced an initial HONC symptom before daily smoking, about 25.8% of 31 symptomatic cigarette smokers who met ICD10 ND criteria did so before daily smoking.

Discussion

This is the first longitudinal study to compare the trajectories of ND symptoms between adolescent WP and cigarette smokers. Our findings show that adolescent WP smokers experience initial ND symptoms and meet ICD10 dependence criteria earlier and with less frequent use than cigarette smokers. Half of adolescent WP smokers who experienced one or more HONC symptoms did so within 10 months of the first puff on WP, compared to 19 months for cigarette smokers. Similarly, half of adolescent WP smokers who met ICD10 criteria did so within 15.1 months after smoking initiation, compared to 22.3 months for cigarette smokers. In terms of specific symptom appearance, we found that 3 HONC symptoms (i.e., craving, feeling addicted, and urge to smoke) and one ICD10 criterion (i.e., a strong desire or sense of compulsion to use tobacco) develop faster in WP smokers compared to cigarette smokers. We also found that as initial HONC symptoms develop and ICD10 ND criteria are met, adolescent cigarette smokers show more accelerated patterns of increasing use frequency compared to WP smokers. Interestingly, at a comparable stage of use, more cigarette smokers expressed difficulty in refraining from smoking in places where it is not allowed compared to WP smokers. These findings highlight the potential role of WP-unique features, use patterns, and social setting in ND development, and the need to address early ND symptoms appearance, composition, and contextual factors in WP smoking cessation interventions.

The finding that ND develops more quickly and with a low frequency of use among WP smokers compared to cigarette smokers may be explained by several unique features of WP smoking, including deep inhalation that is required to operate the WP and overcome the device and tubing space (21–23). For example, a smoker can inhale approximately 50–80 L of tobacco smoke during a one-hour WP smoking session compared with 0.5–0.8 L from ~5 min smoking single cigarette (24); greater intake of smoke is associated with greater intake of nicotine (23). Other potential contributors include the pleasant and social nature of WP smoking, and its sensory cues (e.g. aromatic smell, bubbling sound, the attractive silhouette of the device), which may provide conditioned cues that enhance the development of ND in adolescent WP smokers (5,8,25). These factors should be interpreted in light of the differences in the dose of nicotine delivered and other differences between these two smoking methods.

Although there are differences in the timing and order of appearance of ND symptoms in WP vs. cigarette smokers, the earliest appearing symptoms are generally the same. Three of the four most common ones are identical for WP and cigarette smokers (i.e., strong craving, feeling addicted, and failed quit attempts) suggesting largely similar processes but distinct timing, with ND onset being much more rapid in WP smokers. Additionally, where differences in appearance of ICD10 criteria exist (difficulty controlling use and use despite harm, appear as 1^st-4^th for WP, whereas physiological withdrawal and tolerance appear for cigarette smokers), it suggests that conditioned responses to the social, cognitive, and sensory cues may be more important determinants of early manifestation of ND symptoms in WP whereas physiological changes are more important early on for cigarette smokers. These WP-specific patterns highlight the importance of targeting social cues (e.g., laws prohibiting WP in restaurants) and sensory cues (e.g., banning WP tobacco flavors, just as flavored cigarettes are banned in the US) for WP cessation interventions.

One interesting finding is that although the number of participants endorsing different HONC symptoms was similar between the two products, difficulty in refraining from tobacco was reported by a higher proportion of cigarette smokers than WP despite taking longer time to endorse this symptom. Our earlier results of predictors of progression of WP smoking showed difficulty refraining from smoking WP while in a restaurant strongly predicted progression in ND symptoms in adolescent WP smokers, underling the role of smoking cues in WP-serving venues and the important role of these venues in promoting ND (22). These findings reflect a new nuance in the difference in addiction patterns between cigarettes and WP smokers. They can also mean that although WP smokers develop ND symptoms faster, they have less difficulty refraining from smoking than cigarette smokers in tobacco-free settings. This notion has important implications for WP smoking cessation as it means that earlier intervention (because WP smokers develop ND symptoms faster) potentially may lead to higher rates of cessation given that WP smokers also have less difficulty refraining from smoking WP than cigarette smokers. Also, as discussed above, among ICD10 criteria, “use despite harm” was reported in higher proportion and earlier for WP smokers than cigarettes. This emphasizes that even for a tobacco use method that is considered by many as safer than cigarettes (7, 25–27), the more ND smokers continue their WP smoking despite full awareness of its harms.

This study has some limitations. First tobacco use data were collected by self-report and may be intentionally distorted by social desirability response bias. However, self-reported smoking behavior has been shown to be a reliable and valid method in adolescent survey studies (28, 29). Second, the longitudinal nature of the study makes possible errors in recalling the dates of events which happened remotely. We minimized the likelihood of this error by means of methods that improve recall of events (e.g., personal landmarks, bounded recall, decomposition, and a visual aid) (30–32). Finally, most of the cigarette smokers that we recruited during waves 4, 5, and 6 of the study were 11^th or 12^th graders that raised the age of cigarette smokers compared to WP smokers at the time of entry to study. Also, the age of smoking initiation was slightly higher for cigarettes compared to WP smokers. However, our sensitivity analyses revealed that after adjusting for age of smoking initiation and age at the time of entry to study, differences in time-to-event (initial HONC symptom or full ICD10 ND) between WP and cigarette smokers remained significant.

Conclusions

This longitudinal study, for the first time, provides a robust examination of the development of ND and sequence of symptom presentation among adolescents who smoke WP compared to cigarettes. It shows that early development of symptoms of dependence and the full syndrome of ND that are important in relapse and cessation appear earlier with low frequency and quantity of use in adolescent WP smokers compared to cigarette smokers. We also documented that early HONC symptoms and ICD10 criteria develop in almost similar sequences for WP and cigarette smokers, with faster appearance of ND symptoms among WP smokers compared with cigarette smokers. Accordingly, WP smokers who manifest early ND symptoms should be identified and offered cessation treatment as early as possible to avoid further development of ND syndrome and to increase the likelihood of successful quitting. Failed quit attempts manifests as an early symptom of ND among WP smokers which may warrant intervention efforts to boost quitting self-efficacy, which at least among cigarette smokers is an important determinant of success (37,38) and predicts adherence to WP cessation treatment (39). Our results also suggest that cessation efforts with adolescents early in their WP use may benefit from targeting the extinction of social and sensory cues that are associated with ND rather than physiological symptoms which tend to manifest somewhat later. It is also important to educate youths who smoke to recognize their early symptoms of ND which in turn will prompt earlier and more-effective cessation efforts. To achieve higher effectiveness, WP-tailored interventions need to consider the unique trajectories of appearance of ND symptoms.

IMPLICATIONS AND CONTRIBUTION.

This study shows that early development of symptoms of nicotine dependence that are important in relapse and cessation appear earlier with low frequency and quantity of use in adolescent waterpipe smokers compared to cigarette smokers. Waterpipe-tailored interventions need to consider the unique trajectories of appearance of nicotine dependence symptoms.

Abbreviations:

ND

Nicotine Dependence

WP

Waterpipe

HONC

Hooked on Nicotine Checklist

ICD10

International-Classification-of-Diseases, 10th revision

WDLY

Waterpipe Dependence in Lebanese Youth