Intervenciones para el abandono del hábito de fumar en pipa de agua
Resumen
Antecedentes
Mientras que el consumo de cigarrillos ha disminuido en todo el mundo, el de pipas de agua está aumentando, especialmente entre los jóvenes. El impacto de este aumento se ve amplificado por la creciente evidencia de su naturaleza adictiva y perjudicial. El consumo de pipa de agua está influido por múltiples factores, como los sabores atractivos, el marketing, el uso en contextos sociales y la percepción errónea de que la pipa de agua es menos perjudicial o adictiva que los cigarrillos. Las personas que utilizan pipas de agua están interesadas en dejar de fumar, pero a menudo no consiguen hacerlo por sí solas. Por lo tanto, el desarrollo y estudio de intervenciones para ayudar a las personas a dejar de fumar con pipas de agua se identificó como una prioridad para los esfuerzos mundiales de control del tabaquismo.
Objetivos
Evaluar la eficacia de las intervenciones para el abandono del tabaco en personas que fuman pipas de agua.
Métodos de búsqueda
Se realizaron búsquedas en el Registro especializado del Grupo Cochrane de Tabaquismo (Cochrane Tobacco Addiction Group) desde la creación de la base de datos hasta el 29 de julio de 2022, con la utilización de términos y ortografías variables ('waterpipe' o 'narguile' o 'arguile' o 'shisha' o 'goza' o 'narkeela' o 'hookah' o 'hubble bubble'). Se buscaron ensayos, publicados o no, en cualquier idioma.
Criterios de selección
Se buscaron ensayos controlados aleatorizados (ECA), ensayos controlados cuasialeatorizados o ECA por conglomerados de cualquier intervención para el abandono del hábito de fumar en personas que usan pipas de agua, de cualquier edad o sexo. Para ser incluidos, los estudios debían medir la abstinencia de la pipa de agua en un seguimiento de tres meses o más.
Obtención y análisis de los datos
Se utilizaron los métodos estándar de Cochrane. El desenlace principal de esta revisión fue la abstinencia del uso de la pipa de agua al menos tres meses después del inicio del estudio. También se recopilaron datos sobre los eventos adversos. Los efectos de los estudios individuales y los efectos agrupados se resumieron como razones de riesgos (RR) e intervalos de confianza del 95% (IC del 95%), y se utilizaron modelos de efectos aleatorios de Mantel‐Haenszel para combinar los estudios, cuando fue apropiado. La heterogeneidad estadística se evaluó mediante la estadística I 2 . Los desenlaces secundarios se resumieron de forma narrativa. Se utilizaron las cinco consideraciones del método GRADE (riesgo de sesgo, inconsistencia del efecto, imprecisión, medidas indirectas y sesgo de publicación) para evaluar la certeza del conjunto de evidencia para el desenlace principal en cuatro categorías: alta, moderada, baja o muy baja.
Resultados principales
Esta revisión incluyó nueve estudios con 2841 participantes. Todos los estudios se realizaron en adultos y se llevaron a cabo en Irán, Vietnam, Siria, Líbano, Egipto, Pakistán y EE. UU. Los estudios se realizaron en varios contextos, incluidos colegios/universidades, centros de salud comunitarios, hospitales de tuberculosis y centros de tratamiento del cáncer, mientras que dos estudios analizaron intervenciones de salud digital (intervención formativa en línea por Internet, intervención de mensajes de texto) (e‐health). En general, se consideró que tres estudios tenían bajo riesgo de sesgo y seis, alto.
Se agruparon los datos de cinco estudios (1030 participantes) que analizaron intervenciones conductuales intensivas presenciales en comparación con una intervención conductual breve (p. ej., una sesión de asesoramiento conductual), la atención habitual (p. ej., materiales de autoayuda) o ninguna intervención. En el metanálisis se incluyeron personas que utilizaban exclusivamente pipa de agua, o con otra forma de tabaco. En general, se encontró evidencia de certeza baja de un efecto beneficioso del apoyo conductual para la abstinencia de la pipa de agua (RR 3,19; IC del 95%: 2,17 a 4,69; I 2 = 41%; cinco estudios, n = 1030). La calidad de la evidencia se redujo debido a la imprecisión y al riesgo de sesgo.
Se agruparon los datos de dos estudios (n = 662 participantes) que probaron vareniclina combinada con intervención conductual en comparación con placebo combinado con intervención conductual. Aunque la estimación puntual favoreció a la vareniclina, los IC del 95% fueron poco precisos e incluyeron la posibilidad de ninguna diferencia y tasas de abandono inferiores en los grupos de vareniclina, así como un efecto beneficioso tan grande como el encontrado en el abandono del hábito de fumar cigarrillos (RR 1,24; IC del 95%: 0,69 a 2,24; I 2 = 0%; dos estudios, n = 662; evidencia de certeza baja). La calidad de la evidencia disminuyó debido a su imprecisión. No se encontró evidencia clara de una diferencia en el número de participantes que experimentaron eventos adversos (RR 0,98; IC del 95%: 0,67 a 1,44; I 2 = 31%; dos estudios, n = 662). Los estudios no informaron eventos adversos graves.
Un estudio analizó la eficacia de siete semanas de tratamiento con bupropión combinado con una intervención conductual. No hubo evidencia clara de un efecto beneficioso para el abandono del uso de la pipa de agua en comparación con el apoyo conductual solo (RR 0,77; IC del 95%: 0,42 a 1,41; un estudio, n = 121; evidencia de certeza muy baja) o con la autoayuda (RR 1,94; IC del 95%: 0,94 a 4,00; un estudio, n = 86; evidencia de certeza muy baja).
Dos estudios examinaron intervenciones de salud digital. Un estudio informó tasas más altas de abandono de la pipa de agua entre los participantes asignados al azar a una intervención de teléfono móvil personalizada o no personalizada en comparación con los asignados al azar a ninguna intervención (RR 1,48; IC del 95%: 1,07 a 2,05; dos estudios, n = 319; evidencia de certeza muy baja). Otro estudio informó mayores tasas de abstinencia de la pipa de agua tras una intervención formativa intensiva en línea en comparación con una intervención formativa breve en línea (RR 1,86; IC del 95%: 1,08 a 3,21; un estudio, n = 70; evidencia de certeza muy baja).
Conclusiones de los autores
Se encontró evidencia de certeza baja de que las intervenciones conductuales para el abandono del uso de la pipa de agua pueden aumentar las tasas de abandono entre los que las fuman. No se encontró evidencia suficiente para evaluar si la vareniclina o el bupropión aumentaron la abstinencia de la pipa de agua; la evidencia disponible es compatible con tamaños del efecto similares a los observados para el abandono del hábito de fumar cigarrillos.
Dado el alcance y la eficacia potenciales de las intervenciones de salud digital para el abandono de fumar con pipa de agua, se necesitan ensayos con muestras amplias y periodos de seguimiento prolongados. Los estudios futuros deberían utilizar la validación bioquímica de la abstinencia para evitar el riesgo de sesgo de detección. Por último, se ha prestado poca atención a los grupos de alto riesgo de fumar en pipa de agua, como los jóvenes, los adultos jóvenes, las mujeres embarazadas y los consumidores de dos o múltiples tipos de productos de tabaco. Estos grupos se beneficiarían de estudios específicos.
PICO
Resumen en términos sencillos
¿Es posible ayudar a los usuarios de pipas de agua a dejar el hábito mediante intervenciones para el abandono del hábito de fumar?
Mensajes clave
• La evidencia limitada mostró que el apoyo conductual podría ser útil para dejar de usar pipas de agua.
• No hubo evidencia suficiente para evaluar el efecto de la vareniclina, el bupropión ni las intervenciones de salud digital (e‐health) para el abandono del uso de pipas de agua.
• En todos los casos, estudios posteriores podrían modificar estas conclusiones. Se necesitan más ensayos grandes y bien diseñados sobre intervenciones conductuales y farmacológicas para dejar de fumar con pipa de agua.
• Dado el alcance y la eficacia potenciales de las intervenciones de salud digital para ayudar a dejar de fumar en pipa de agua, se necesitan ensayos con muestras amplias y periodos de seguimiento prolongados.
• Informar con más detalle sobre las estrategias conductuales utilizadas en las intervenciones de estudio ayudaría a identificar componentes vitales en las intervenciones con pipas de agua.
¿Qué es una pipa de agua?
Las pipas de agua son dispositivos utilizados para fumar tabaco. Tal y como se utiliza hoy en día, la pipa de agua (también conocida como hookah, shisha o narghile) consta de una cabeza (donde se coloca el tabaco), un cuerpo, una base de agua y un tubo flexible que termina en una boquilla. Los trozos de carbón encendido suelen colocarse encima de una lámina de aluminio perforada que cubre la cabeza llena de tabaco, lo que permite que el aire calentado por el carbón pase a través del tabaco, mientras que los orificios de la parte inferior de la cabeza permiten que el humo baje por el tallo de la pipa de agua. La parte inferior del tallo está bajo el agua, lo que hace que el humo burbujee a través de él, en su camino a través del tubo flexible y la boquilla hacia el fumador. La punta de la boca suele cubrirse con una boquilla desechable para usuarios individuales.
¿Cómo se trata la adicción a la pipa de agua?
Las formas de apoyo para ayudar a las personas a dejar de fumar en pipa de agua son similares a las que se dirigen a otras formas de consumo de tabaco. Entre ellos se incluyen el apoyo conductual, los medicamentos para dejar de fumar y combinaciones de ambos.
¿Por qué se ha realizado esta revisión Cochrane?
El consumo de pipas de agua ha aumentado drásticamente, convirtiéndose en un problema de salud pública mundial, especialmente entre los jóvenes. La evidencia actual sugiere que fumar en pipa de agua es tan adictivo como los cigarrillos y que los usuarios experimentarán riesgos para su salud similares a los de los fumadores de cigarrillos. Por lo tanto, es importante desarrollar y analizar intervenciones específicas para dejar de fumar en pipa de agua.
¿Qué se hizo?
Se incluyeron ensayos de cualquier intervención para ayudar a los fumadores de pipa de agua a dejar de fumar. Podían ser de cualquier edad o sexo. Se midió si los participantes habían dejado de utilizar una pipa de agua a los tres meses de la intervención o más tarde. Se incluyeron intervenciones dirigidas a individuos o grupos de personas.
¿Qué se encontró?
Se encontraron nueve estudios que analizaron intervenciones para ayudar a los fumadores de pipa de agua a dejar de fumar. Entre ellos, cinco estudios examinaron el apoyo conductual; dos estudios examinaron un medicamento para dejar de fumar, llamado vareniclina; un estudio examinó un medicamento para dejar de fumar llamado bupropión; y dos estudios examinaron el apoyo con salud digital proporcionado a través de Internet o teléfono móvil.
¿Cuáles son los resultados principales de esta revisión?
Se tiene evidencia de calidad baja de que el apoyo conductual presencial ayuda a más personas a dejar de fumar en pipa de agua que ningún apoyo. No está claro el efecto del bupropión, la vareniclina ni las intervenciones de salud digital debido a la escasez de evidencia.
¿Cuáles son las limitaciones de la evidencia?
Los resultados se basan en los datos de unos pocos estudios. Seis de los nueve estudios presentaron problemas de diseño que podrían afectar la fiabilidad de los resultados. Además, las intervenciones conductuales incluidas difirieron entre sí y no siempre estuvieron bien descritas. Esto significa que los resultados podrían variar cuando se disponga de más estudios.
¿Cuál es el grado de actualización de esta evidencia?
Esta búsqueda de evidencia se realizó en julio de 2022.
Authors' conclusions
Summary of findings
| Patients or population: people who use waterpipe Settings: clinics/communities/college campuses Intervention: behavioral support and pharmacotherapy Comparison: usual care/brief intervention/placebo | |||||
| Outcome | Anticipated absolute effects (95% CI) | Relative effect (Random effect) | Number of participants (number of studies) | Certainty of the evidence (GRADE) | |
|---|---|---|---|---|---|
| Risk without intervention, usual care | Risk with Interventions | ||||
| Face‐to‐face behavioral intervention vs minimal or no support (Measured 3 to 6 months with CO‐validated prolonged and continuous abstinence, and self‐reported abstinence) | 112 per 1000 | 297 per 1000 (187 to 472) | RR 2.64 (1.66 to 4.20) | 851 (5 cluster‐RCTs) | ⨁⨁⊝⊝ |
| Varenicline plus behavioral intervention vs placebo plus behavioral intervention ‐ (repeated point prevalence abstinence (not even a puff/chew/session in the last 7 days) at each of weeks 5, 12, and 25 (combined), verified by CO cutoff < 10 ppm) | 56 per 1000
| 70 per 1000 (39 to 126)
| RR 1.24 (0.69 to 2.24)
| 662 (2 RCT)
| ⨁⨁⊝⊝
|
| Bupropion therapy plus behavioral intervention vs behavioral intervention alone (continuous waterpipe smoking abstinence, defined as an expired CO of 9 ppm or less at the 6‐month postintervention follow‐up visits) | 449 per 1000 | 364 per 1000 (189 to 633) | RR 0.77 (0.42 to 1.41) | 121 (1 cluster‐RCT) | ⨁⊝⊝⊝
|
| Bupropion therapy plus behavioral support versus self‐help (continuous waterpipe smoking abstinence, defined as an expired CO of 9 ppm or less at the 6‐month postintervention follow‐up visits) | 194 per 1000 | 375 per 1000 (182 to 774) | RR 1.94 (0.94 to 4.00)
| 86 (1 cluster‐RCT) | ⨁⊝⊝⊝ |
|
| Impact | Number of participants (number of studies) | Certainty of the evidence (GRADE) | ||
| E‐health interventions | Due to the variation in the studies, we did not pool the results. Quit rates were higher among participants randomized to receive either the tailored or untailored intervention groups compared with those in the no intervention control group (RR 1.48, 95% CI 1.07 to 2.05; 1 study, N = 319). Smoking abstinence rates were higher among those randomized to receive a 20‐slide presentation compared with those randomized to receive an 8‐slide presentation (RR 1.86, 95% CI 1.08 to 3.21; 1 study, N = 70). | 389 (2 RCTs)
| ⨁⊝⊝⊝
| ||
| GRADE Working Group grades of evidence CI: confidence interval; CO: carbon monoxide; ppm: parts per million; RCT: randomized controlled trial; RR: risk ratio | |||||
| a Downgraded one level for risk of bias: no studies were at low risk of bias | |||||
Background
Description of the condition
Waterpipe smoking (aka hookah, shisha, narghile) has become a global public health problem (Maziak 2015c). Despite the substantial success in reducing cigarette smoking in some parts of the world, multiple surveillance data from around the globe have shown that waterpipe smoking has increased dramatically, and become one of the leading tobacco use methods (Akl 2011; Akl 2015; Jawad 2018; Maziak 2015a; Maziak 2015b; Maziak 2015c). For example cigarette smoking rates have decreased by approximately 33% in the USA, while waterpipe smoking has taken the opposite course, with a 123% increase among young adults in the USA (Arrazola 2015; Johnston 2016). A 2018 systematic review, which included 68 countries in the World Health Organization’s (WHO) six regions, indicated that waterpipe smoking among adults was highest in the Eastern Mediterranean (EMR), followed by Europe, and is predominantly popular among young people (Jawad 2018). For example, waterpipe use was the highest among young Lebanese adults (ages 18 to 24; 65.3%); current waterpipe use (past 30‐day use) was the highest among Lebanese youth (ages 12 to 17; 37.2%); and daily use was the highest among Egyptian youth (10.4%). A similar surge was documented in western countries. For example, based on data from the third wave (2016) of the Population Assessment of Tobacco and Health (PATH) study in the USA, young adults (ages 18 to 24) had a higher prevalence of ever used (47.5%), past 12‐month use (25.8%), and current (past 30‐day) waterpipe use (9.2%) than youth, aged 12 to 17 (ever used [5.2%], past 12‐month use [2.9%], and current use [0.7%]); or adults 25+ (15.5% ever used, 3.5% used within past 12 months, 1.2% current use [Sharma 2020]). A similar profile was documented in Germany. Using data from the German Health Interview and Examination Surveyfor Children and Adolescents during 2014 to 2017 (N = 6599), the prevalence of ever used was 25.8%, use in the last 12 months was 19.7%, and current waterpipe use was 8.5% (Klosterhalfen 2020).
The impact of the dramatic rise of waterpipe smoking globally is amplified by the mounting evidence of its dependence‐inducing and harmful nature (Gathuru 2015; Sutfin 2018). Evidence suggests that waterpipe smoking is addictive (Aboaziza 2015). A typical waterpipe smoking session delivers 1.7 times the nicotine dose of a single cigarette (Eissenberg 2009), and even intermittent waterpipe use can produce symptoms of nicotine dependence (Bahelah 2017; Ebrahimi 2021). Several laboratory studies have indicated that regular and long‐term waterpipe smokers experience withdrawal symptoms while abstaining (Maziak 2009; Rastam 2011). In terms of the health effects of waterpipe smoking, evidence suggests that compared with non‐smokers, waterpipe smokers have an increased risk of respiratory disease (odds ratio (OR) 3.18; 95% confidence interval (CI) 1.25 to 8.08), oral cancer (OR 2.12, 95% CI 1.32 to 3.42), lung cancer (OR 6.0, 95% CI 1.78 to 20.26), and cardiovascular disease (OR 2.4, 95% CI 1.20 to 2.80 [Asfar 2020; Bhatnagar 2019; El‐Zaatari 2015; Jawad 2013; Waziry 2017]). Using charcoal to heat the tobacco in the waterpipe results in additional exposure to toxins and cancer‐causing chemicals (Al‐Faham 2018; Monzer 2008; Nguyen 2013). Sharing the same waterpipe can also expose users to communicable disease risks (e.g. tuberculosis, herpes [Al‐Faham 2018; Martin 2009]). Furthermore, there is evidence that waterpipe smoking can roll back successes achieved in tobacco control by providing a gateway to cigarette smoking among youth (Jaber 2015; Soneji 2015). For example in Soneji 2015, waterpipe smoking among adolescents was associated with cigarette smoking initiation (adjusted odds ratio (AOR) 2.56, 95% CI 1.46 to 4.47), current cigarette smoking (AOR 2.48, 95% CI 1.01 to 6.06), and higher intensity of cigarette smoking (AOR 2.55, 95% CI 1.48 to 4.38). Hence, waterpipe smoking cessation interventions are a high priority.
Description of the intervention
Interventions to help people quit smoking waterpipes may take similar forms to interventions targeting other forms of tobacco use (Livingstone‐Banks 2022; Rigotti 2022). These include treatments available for people wanting to quit, such as behavioral support (Hartmann‐Boyce 2021), pharmacotherapy (Cahill 2013), and combinations of the two (Hartmann‐Boyce 2019; Stead 2016); and interventions targeting users to induce quit attempts, such as brief advice from a health professional (Stead 2013). Behavioral support can include counseling (Lancaster 2017; Stead 2017), providing self‐help materials (Livingstone‐Banks 2019), and incentives (Notley 2019). Pharmacotherapies include nicotine replacement therapy (Hartmann‐Boyce 2018; Lindson 2019), bupropion (Hajizadeh 2023), and varenicline (Livingstone‐Banks 2023). Interventions can be delivered via different media (Matkin 2019; Taylor 2017; Tzelepis 2019; Whittaker 2019), and by various healthcare professionals, including nurses (Rice 2017), dentists (Holliday 2021), and pharmacists (Carson‐Chahhoud 2019).
How the intervention might work
Different smoking cessation treatments incorporate different mechanisms, but the underpinning principles are to mitigate the craving and withdrawal symptoms often associated with a quit attempt; reduce the reward derived from smoking, by indirectly disrupting dopamine release or by desensitizing receptors; and delivering positive reinforcement through means other than smoking (Adams, 2020; Chang, 2020).
Why it is important to do this review
In response to the increase in waterpipe use, the WHO released two advisory notes in 2005 and 2015 about the risks of waterpipe smoking and ways to address them (WHO 2015a; WHO 2015). One of the main WHO recommendations was based on Article 14 of the Framework Convention on Tobacco Control (FCTC), highlighting the need to develop waterpipe‐specific smoking cessation programs (WHO 2015), which take into account the emerging evidence about the unique aspects of waterpipe smoking that contribute to dependence among smokers (Ward 2015). For example the social aspect of waterpipe smoking and the prominent role of waterpipe venues (e.g. hookah cafés, restaurants) in shaping waterpipe experience and dependence need to be addressed (Ebrahimi 2020). Yet, unlike cigarettes, guidelines for waterpipe‐specific treatments have not been established, and waterpipe cessation services are still not well‐developed. One of the main barriers to developing such cessation guidelines and evidence‐based cessation programs lies in the scarcity of studies that have tested waterpipe‐specific cessation interventions. When such studies exist, their quality and suitability to guide waterpipe cessation services are not evaluated. Therefore, this Cochrane Review aims to assess the emerging literature in this area (Maziak 2007; Maziak 2015a).
The previous version of this Cochrane Review, published in 2015, found only three published randomized controlled trials (RCTs) for waterpipe cessation interventions (Maziak 2015a). For this review version, we expanded our inclusion criteria to increase the value of the evaluation. The current review includes studies that (1) measured abstinence at three‐month follow‐up or more, rather than six months or more; (2) used any definition of waterpipe smoking rather than past‐month use only; and (3) reported cessation in dual users (waterpipe and any other tobacco product), as well as in people who exclusively smoked waterpipes.
Objectives
To evaluate the effectiveness and safety of tobacco cessation interventions for people who use waterpipes.
Methods
Criteria for considering studies for this review
Types of studies
-
Randomized controlled trials (RCTs)
-
Cluster‐randomized controlled trials (cluster‐RCTs)
-
Quasi‐randomized controlled trials (quasi‐RCTs)
Types of participants
People of any age and gender who smoke waterpipes, including people who use waterpipes (daily or occasionally) alongside e‐cigarettes or other tobacco products (e.g. people who use a combination of waterpipes and cigarettes or smokeless tobacco).
Types of interventions
We included any interventions aimed at helping people quit waterpipe smoking (or the dual use of tobacco products, including waterpipes). Therefore, eligible interventions could consist of brief advice to quit, provision of self‐help materials, brief or intensive behavioral counseling interventions, pharmacotherapies (e.g. nicotine replacement therapy, varenicline, bupropion), referral to other sources of support, and any combination of the above.
Comparators
We set out to include trials that compared waterpipe cessation interventions with any of the following comparators:
-
No intervention;
-
Wait‐list controls;
-
Usual care, including brief advice interventions;
-
Other active interventions (as defined above).
Types of outcome measures
Primary outcomes
The primary outcome was abstinence from waterpipe use. To be eligible, studies had to report abstinence rates from waterpipe use at least three months from baseline. We excluded trials that did not investigate waterpipe‐use outcomes, or did not have a sufficiently long follow‐up. In trials with more than one measure of abstinence, we selected the measure with the longest follow‐up and the strictest definition, in line with the Russell Standard (West 2005), and the SRNT Treatment Research Network guidance (Piper 2006). Therefore, we preferred biochemically validated over self‐reported abstinence, and prolonged or continuous abstinence over point prevalence abstinence. Abstinence rates were based on intention‐to‐treat analyzes, with dropouts and losses to follow‐up assumed to be continuing or relapsed waterpipe users.
Secondary outcomes
Secondary outcomes were the number of people who experienced adverse events, and the number of people who experienced serious adverse events, as reported by the authors.
Search methods for identification of studies
Electronic searches
We updated searches of the Cochrane Tobacco Addiction Group Specialised Register via the Cochrane Register of Studies (CRS‐Web) from database inception to 29 July 2022, using the terms 'waterpipe' or 'narghile' or 'arghile' or 'shisha' or 'goza' or 'narkeela' or 'hookah' or 'hubble bubble'. We searched for these terms in the title, the abstract, or as keywords. At the time of the search, the Register included the results of searches of the following databases from inception: the Cochrane Central Register of Controlled trials (CENTRAL, which includes records from both the US National Library of Medicine's clinicaltrials.gov trial registry, and the WHO International Clinical Trials Registry Platform [ICTRP; www.who.int/ictrp/], Issue 6, 2022; MEDLINE OVID to update 15 July 2022; Embase OVID to 29 July 2022; PsycINFO OVID to 11 July 2022. For further details of the searches used to populate the Cochrane Tobacco Addiction Group Specialised Register, see tobacco.cochrane.org/resources/cochrane-tag-specialised-register. Our search strategy can be found in Appendix 1.
Searching other resources
We also used our existing bibliography, compiled from earlier exhaustive reviews of the literature on waterpipe smoking (Akl 2015; Asfar 2016; Jawad 2016; Jawad 2018; Maziak 2015a; Ward 2015).
Data collection and analysis
Selection of studies
Two review authors (TG and OO) independently prescreened titles and abstracts of articles identified in the search, using a screening checklist based on our eligibility criteria. We resolved disagreements through discussion, or referral to a third review author (TA). We conducted screening using Covidence software (Covidence).
Two review authors (TG and OO) independently screened the full text of articles that passed prescreening. We consulted a third review author (TA) to resolve any disagreements that were not resolved through discussion.
Data extraction and management
Two review authors (TG and OO) independently extracted data from each included study, using a standardized electronic data collection form. Review authors then cross‐checked this information between themselves, and resolved disagreements through discussion. If the current review team authors were also the authors of an included study, we ensured that other review authors or other researchers did the data extraction and risk of bias assessment. We extracted the following information about each study, where possible, which is presented in the Characteristics of included studies tables.
-
Methods: study design; study location (i.e. country); study setting (e.g. hospital, university, community); and study recruitment procedure
-
Participants: number of participants (N); if this was a specialist population (e.g. people with tuberculosis); if participants were selected based on motivation to quit; and participant characteristics (including gender, age, baseline average cigarettes/day, nicotine dependence, baseline motivation to quit, and baseline self‐efficacy/confidence in quitting)
-
Interventions: comparator (control) and intervention details, including modality of support; details of provider training; overall contact time; the number of sessions; and use of pharmacotherapy
-
Outcomes: definition of waterpipe abstinence; longest follow‐up time; use of biochemical validation; and adverse events
-
Study funding sources
-
Author conflicts of interest
Assessment of risk of bias in included studies
We assessed and reported the risk of bias in the Characteristics of included studies tables for each eligible study, following the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017), and the risk of bias guidance developed by the Cochrane Tobacco Addiction Group (Higgins 2011). The latter states that performance bias (relating to blinding participants and providers) should not be assessed for behavioral interventions, as it is impossible to blind people to these types of interventions. Therefore, we reported on the following individual domains:
-
Random sequence generation;
-
Allocation concealment;
-
Blinding of participants and personnel (performance bias) for studies testing pharmacotherapy interventions only;
-
Blinding of outcome assessment (detection bias);
-
Incomplete outcome data (attrition bias);
-
Selective reporting (reporting bias);
-
Other bias.
Two review authors (TG and OO) independently assessed the risk of bias for each included study, with any disagreements resolved by discussion and inclusion of a third review author (TA). A summary risk of bias judgment was derived for each study by applying an algorithm suggested in Section 8.7 of the Cochrane Handbook (Higgins 2017). Specifically, if the judgment for at least one of the domains was judged to be high risk of bias, we determined the summary risk of bias to be high. If there were no judgments of high risk, but the judgment in at least one domain was at unclear risk of bias, then we determined the summary risk of bias to be unclear. The summary risk of bias was deemed low if our judgments in all domains were low risk of bias.
Measures of treatment effect
For waterpipe use abstinence, we calculated a risk ratio (RR) and associated 95% confidence interval (CI) for each trial included in the meta‐analyzes. We calculated RRs as follows: (number of participants abstinent from waterpipe in the intervention group/number of participants in the intervention group)/(number of participants abstinent from waterpipe in the control group/number of participants in the control group). We would have used the same methods to calculate the RRs, and 95% CI for the numbers of participants experiencing adverse events and serious adverse events for each study were the relevant data presented.
Unit of analysis issues
The unit of analysis was the individual. Where we deemed it possible and appropriate to the structure of the analysis, we combined all relevant experimental intervention groups of a given multiple‐arm study into a single intervention group, and all relevant controls of that study into a single control group. When extracting data from cluster‐RCTs, we considered whether study authors had made allowances for clustering in the data analysis reported, and when available, used data adjusted for clustering effects. When studies reported analyzes that accounted for the clustered study design, we estimated the effect on this basis. When this was not possible, and the information was unavailable from the authors, we carried out an approximately correct analysis (Higgins 2022). We imputed estimates of the intra‐cluster correlation coefficient (ICC), as reported in the study, using estimates derived from similar studies, or using general recommendations from empirical research.
Dealing with missing data
We calculated quit rates on an intention‐to‐treat basis, and assumed that participants lost to follow‐up were still smoking. However, we excluded deaths from the denominators of studies. Where abstinence data were missing, we contacted study authors for further information or clarifications.
Assessment of heterogeneity
We assessed the characteristics of included studies to identify any clinical or methodological heterogeneity before pooling studies and conducting meta‐analyzes. Where we deemed studies homogeneous enough to be meaningfully combined, we conducted a meta‐analysis and assessed statistical heterogeneity, using the I2 statistic. We conducted subgroup and sensitivity analyzes to investigate potential causes of the observed heterogeneity.
Assessment of reporting biases
If we had been able to meta‐analyze abstinence rates from at least ten studies, we planned to assess reporting bias using funnel plots. Funnel plots illustrate the relationship between the effect estimates from individual studies against their size or precision. The greater the degree of asymmetry, the greater the risk of reporting bias.
Data synthesis
We conducted our analyzes in Review Manager 5.4 (Review Manager 2020). We pooled studies for our waterpipe abstinence outcome using Mantel‐Haenszel random‐effects methods to generate pooled RRs with 95% CIs. We used a random‐effects model to accommodate anticipated variations across included studies in the interventions and comparators administered to participants. An RR greater than one indicated that people were more likely to quit in the intervention group than in the control group.
Subgroup analysis and investigation of heterogeneity
We conducted subgroup analyzes to account for clinical heterogeneity among studies.
We grouped studies by:
-
Type of tobacco use (exclusive waterpipe users; dual waterpipe users)
-
Level of participant motivation (recruited based on being motivated to quit; recruited regardless of motivation)
Sensitivity analysis
We conducted a sensitivity analysis removing studies deemed to be at an overall high risk of bias.
Summary of findings and assessment of the certainty of the evidence
Following standard Cochrane methodology, we used the five GRADE considerations (risk of bias, inconsistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence for our primary outcome (Schünemann 2022; Schünemann 2022a). Depending on our assessment of these considerations in each intervention comparison, we judged the certainty of the evidence for this outcome to be high, moderate, low, or very low. To present these judgments, we used GRADEpro GDT to create a summary of findings table with the following intervention comparisons: behavioral interventions, pharmacotherapy interventions, and e‐health interventions (GRADEpro GDT).
Results
Description of studies
See Characteristics of included studies; Characteristics of ongoing studies; and Characteristics of excluded studies for specific details.
Results of the search
The new searches for this update of the review retrieved 153 unique records. We also included two out of three studies from the previous review, and one previously excluded study, which was now eligible due to changes in our eligibility criteria, documented in the Differences between protocol and review (reported abstinence at three‐month follow‐up). After title and abstract screening, we classified 23 studies as potentially eligible for inclusion. After full‐text screening, we identified nine studies that met inclusion criteria. The flow of studies through the systematic review process for this update is shown in Figure 1.
Study selection PRISMA flow chart
Included studies
This review included nine studies, involving 2841 participants. We identified six new studies in the current update (Chami 2022; Dadipoor 2022; Dogar 2018; Jiang 2019; Joveini 2020; Mays 2021), in addition to two studies included in the previous review (Dogar 2014; Lipkus 2011). We excluded one previously included study because it also recruited participants who did not smoke a waterpipe (Mohlman 2013), and included one study that was previously excluded but is now eligible as it reported abstinence at three months follow‐up (Asfar 2014).
Types of studies
Five studies were randomized controlled trials (RCT), in which the individual participant was the unit of randomization (Asfar 2014; Chami 2022; Dogar 2018; Lipkus 2011; Mays 2021). Among these, four studies were two‐arm RCTs (Asfar 2014; Chami 2022; Dogar 2018; Lipkus 2011), and one was a two‐site, three‐arm, parallel‐group RCT (Mays 2021). Three studies were quasi‐experimental studies, using community health centers or colleges as the unit of randomization (Dadipoor 2022; Jiang 2019; Joveini 2020). Another study was a cluster‐RCT, using tuberculosis health centers as the unit of randomization (Dogar 2014). Six studies were funded by government agencies (the American National Cancer Institute, National Institute on Drug Abuse, and Fogarty International Center, and the Canadian International Development Research Centre [Asfar 2014; Dadipoor 2022; Dogar 2014; Jiang 2019; Lipkus 2011; Mays 2021]), two studies were funded by a pharmaceutical company (Pfizer [Chami 2022; Dogar 2018]), and one study was funded by a university (Tehran University of Medical Sciences, Iran [Joveini 2020]).
Type of settings and participants
Studies were carried out in Iran (Dadipoor 2022; Joveini 2020), Vietnam (Jiang 2019), Syria (Asfar 2014), Pakistan (Dogar 2014; Dogar 2018), Lebanon (Chami 2022), and the USA (Lipkus 2011; Mays 2021). Studies were conducted in a range of settings, including colleges/universities (Chami 2022; Joveini 2020), community healthcare centers (Dadipoor 2022), tuberculosis healthcare centers (Dogar 2014), cancer treatment centers (Asfar 2014), rural health centers (Jiang 2019), and hospitals (Dogar 2018). One study tested an online web‐based intervention (Lipkus 2011), and another tested a text messages intervention (Mays 2021). Eight studies were conducted on adults (Asfar 2014; Chami 2022; Dogar 2014; Dogar 2018; Jiang 2019; Joveini 2020; Lipkus 2011; Mays 2021), and one study was conducted among women aged 15 years and older (Dadipoor 2022). Four studies recruited participants from specific population groups, including people with suspected tuberculosis (Dogar 2014), and college students (Joveini 2020; Lipkus 2011; Mays 2021).
Studies varied regarding participants’ smoking status and motivation to quit using the waterpipe. Seven studies targeted people who exclusively smoked a waterpipe (Asfar 2014; Chami 2022; Dadipoor 2022; Dogar 2018; Joveini 2020; Lipkus 2011; Mays 2021), and two studies targeted people smoking combustible cigarettes, waterpipe, or both (Dogar 2014; Jiang 2019). Abstinence outcomes for the three types of participants were reported separately in Jiang 2019 and Dogar 2014. We use waterpipe cessation specifically for our analyses. The waterpipe smoking status definition used to determine eligibility differed across studies. Participants in Mays 2021 and Joveini 2020 smoked a waterpipe at least once a month, participants in Jiang 2019 smoked cigarettes and a waterpipe some days or every day, participants in Dogar 2018 smoked a waterpipe daily, participants in Dogar 2014 smoked a local form of unflavored waterpipe tobacco ten times a day on average, participants in Asfar 2014 smoked a waterpipe at least three times a week in the previous year, participants in Chami 2022 smoked a waterpipe five to seven times a week, participants in Dadipoor 2022 smoked waterpipe four times a week for at least six months, and participants in Lipkus 2011 smoked a waterpipe at least once a month. Only four studies only selected participants based on their motivation (or readiness) to quit waterpipe (Asfar 2014; Chami 2022), or all forms of tobacco (Dogar 2018; Jiang 2019). The other studies did not specify this criterion in their eligibility criteria (e.g. the participant is willing to participate in the study, but is not required to be ready to quit waterpipe). No studies focussed on inducing quit attempts among people not seeking support.
Types of interventions
Eight studies tested individual‐level interventions (Asfar 2014; Chami 2022; Dogar 2014; Dogar 2018; Jiang 2019; Joveini 2020; Lipkus 2011; Mays 2021), and one tested group‐based interventions at the community level (Dadipoor 2022). Five studies tested behavioral interventions at different intensities (Asfar 2014; Dadipoor 2022; Dogar 2014; Jiang 2019; Joveini 2020). Three of those interventions were based on the 5 As model theoretical framework for cigarette smoking (Asfar 2014; Dogar 2014; Jiang 2019). Three studies tested pharmacological interventions (Chami 2022; Dogar 2014; Dogar 2018). Two tested varenicline compared to placebo (Chami 2022; Dogar 2018), and one tested bupropion combined with behavioral support compared to behavioral support alone, or self‐help materials alone (Dogar 2014). Two studies tested e‐health interventions (Lipkus 2011; Mays 2021). One study tested a 6‐week tailored mobile messaging intervention conveying the risks of waterpipe tobacco and strategies to enhance motivation and support quitting (Mays 2021). The frequency and duration of the messages were tailored, based on participants’ waterpipe tobacco smoking patterns. The second study tested brief versus intensive online educational interventions targeting college students in the USA (Lipkus 2011).
Types of outcome measures
A wide range of waterpipe abstinence definitions were used. Four studies used seven‐day point prevalence abstinence (Chami 2022; Dogar 2018; Jiang 2019; Mays 2021), three studies used self‐reported waterpipe abstinence (Dadipoor 2022; Joveini 2020; Lipkus 2011), one study used continuous abstinence (Dogar 2014), and one study used prolonged abstinence (Asfar 2014). Five studies out of the nine included studies reported biochemically validated waterpipe abstinence, using exhaled carbon monoxide (CO) levels of < 10 ppm (Asfar 2014; Chami 2022; Dogar 2014; Dogar 2018; Jiang 2019). In six studies, participants were followed for a maximum of six months (Dogar 2014; Dogar 2018; Jiang 2019; Lipkus 2011; Mays 2021, Dadipoor 2022), and in one for twelve months (Joveini 2020). In two studies, the final follow‐up was three months after the intervention (Asfar 2014; Chami 2022).
Two of the three pharmacologic studies reported adverse events (Chami 2022; Dogar 2018).
Excluded studies
We listed nine studies from our searches as potentially relevant studies that did not meet all inclusion criteria. We listed these studies with reasons for exclusion in the Characteristics of excluded studies table. Six studies tested prevention interventions for waterpipe smoking and did not consider waterpipe cessation as an outcome (Leavens 2018; Mohammadi 2019; Mojahed 2018; Nakkash 2018; Shelley 2021; Thomas 2015), one recruited non‐waterpipe smokers (Mohlman 2013), and two followed participants for less than three months (Alzyoud 2018; Shishani 2018).
Risk of bias in included studies
A summary of the risk of bias assessments for the included studies can be found in Figure 2. Overall, we considered three studies at low risk of bias (Asfar 2014; Chami 2022; Dogar 2018), and six at high risk of bias (Dadipoor 2022; Dogar 2014; Jiang 2019; Joveini 2020; Lipkus 2011; Mays 2021).
Risk of bias summary: review authors’ judgments about each risk of bias item for each included study
+ Low risk, ‐ High risk and? Unclear risk
Allocation
We judged six studies at low risk for random sequence generation bias, as they adequately described the randomization process (Asfar 2014; Chami 2022; Dogar 2014; Dogar 2018; Lipkus 2011; Mays 2021), while three studies were at high risk (Dadipoor 2022; Jiang 2019; Joveini 2020).
For allocation concealment, we considered three studies at low risk of bias (Chami 2022; Dogar 2014; Dogar 2018), and three at high risk of bias, because the participants were recruited after the clusters were randomized, which could have influenced recruitment (Dadipoor 2022; Jiang 2019), or because the study was only quazi‐randomized (Joveini 2020). We considered the remaining three studies as unclear risk of allocation concealment due to a lack of information (Asfar 2014; Lipkus 2011; Mays 2021).
Blinding
We assessed performance bias in only three studies (Chami 2022; Dogar 2014; Dogar 2018) because the remaining six studies tested behavioral interventions alone (as specified in the methods). We judged Dogar 2018 and Chami 2022 at low risk of bias as they used a placebo to facilitate blinding, whereas Dogar 2014 was considered at high risk due to lack of a placebo.
We rated five studies at low risk of detection bias (Asfar 2014; Chami 2022; Dogar 2014; Dogar 2018; Jiang 2019), while the remaining four studies were judged at high risk of bias because they did not biochemically verify abstinence, and the intervention and control arms did not receive similar amounts of contact.
Incomplete outcome data
We rated eight studies at low risk of attrition bias (Asfar 2014; Chami 2022; Dadipoor 2022; Dogar 2014; Dogar 2018; Jiang 2019; Lipkus 2011; Mays 2021). Missing data ranged from 7% in Dogar 2014 to 41% in Chami 2022. We judged one study at unclear risk for incomplete outcome data because attrition rates were not reported.
Selective reporting
We judged five studies at low risk for selective reporting (Asfar 2014; Chami 2022; Dadipoor 2022; Dogar 2014; Dogar 2018) because reported outcomes were consistent with the methods sections of protocols or trial registry entries. We could not find prespecified methods in the form of protocols or trial registry entries for the remaining studies, so we judged them at unclear risk.
Other potential sources of bias
Baseline characteristics were unbalanced between treatment groups in one cluster‐randomized study, and therefore, we deemed this study at high risk (Dogar 2014).
Effects of interventions
See: Summary of findings 1 Interventions for waterpipe smoking cessation
Face‐to‐face behavioral support
After adjusting for cluster randomization in Dadipoor 2022; Dogar 2014, Jiang 2019, and Joveini 2020, we pooled the five studies on this intervention (N = 851 effective participants sample size) in our meta‐analysis (Asfar 2014; Dadipoor 2022; Dogar 2014; Jiang 2019; Joveini 2020). See Figure 3. These studies tested more than one session of face‐to‐face behavioral interventions compared with self‐help, brief advice, or no intervention. Three studies included only exclusive waterpipe users (Asfar 2014; Dadipoor 2022; Joveini 2020); the other two studies included exclusive waterpipe users, exclusive cigarette users, and dual waterpipe and cigarette users. We included only exclusive waterpipe users and dual users from these studies in our meta‐analysis. Dogar 2014 was a three‐arm cluster‐RCT testing a self‐help intervention versus a behavioral support combined with bupropion medication. We included only the self‐help and behavioral intervention alone study arms in this meta‐analysis. We used an assumed intracluster correlation coefficient (ICC) of 0.036 to adjust for the cluster‐randomized designs, ending with these estimated design effects for Dadipoor 2022 (2.872), Dogar 2014 (1.792), Jiang 2019 (1.832), and Joveini 2020 (3.664). Therefore, the effective samples for the intervention and comparison arms were 37 versus 37 (Dadipoor 2022), 150 versus 132 (Dogar 2014), 278 versus 125 (Jiang 2019), and 21 versus 21 (Joveini 2020), using the Cochrane Handbook (Higgins 2022). We used these adjusted numbers for the meta‐analysis.
Overall, we found evidence of benefit from behavioral support (risk ratio (RR) 2.64, 95% confidence interval (CI) 1.66 to 4.20; I2 = 44%; 5 studies, N = 851; Analysis 1.1; Figure 3; low‐certainty evidence).
We conducted two subgroup analyses. First, we divided studies by whether they included exclusive waterpipe users or dual users. We used data from five studies in the exclusive waterpipe users subgroup (Asfar 2014; Dadipoor 2022; Dogar 2014; Jiang 2019; Joveini 2020), and two studies in the dual users subgroup (Dogar 2014; Jiang 2019). We found no evidence of subgroup difference (P = 0.36; I2 = 0%; Analysis 1.2; Figure 4). Second, we divided studies by participants' selection based on their motivation to quit, as indicated by the method of recruitment. Two studies recruited participants who were motivated to quit (Asfar 2014; Jiang 2019), and three studies recruited participants regardless of their motivation status (Dadipoor 2022; Dogar 2014; Joveini 2020). We did not find clear evidence of a difference in how effects should be interpreted across subgroups (P = 0.17; I2 = 47.9%; Analysis 1.3; Figure 5). In both cases, the subgroup effect estimates favored the intervention. We could not conduct our planned sensitivity analysis because only one study was not at high risk of bias (Asfar 2014). The point estimate for Asfar 2014 demonstrated a benefit of more intensive behavioral intervention; however, the CI incorporated the possibility of no difference and a benefit of brief intervention, as well as a potential benefit of more intensive intervention (RR 1.46, 95% CI 0.69 to 3.09; N = 50; Analysis 1.3).
Varenicline
We pooled two studies (N = 662 participants) in our meta‐analysis (Chami 2022; Dogar 2018). The two studies tested varenicline plus brief behavioral intervention versus placebo plus brief behavioral intervention among exclusive waterpipe smokers. The point estimate showed more people randomized to receive varenicline quit waterpipe smoking (RR 1.24, 95% CI 0.69 to 2.24; 2 studies, N = 662; Analysis 2.1, Figure 6; low‐certainty evidence). However, the results were limited by imprecision, with CIs incorporating the potential for harm, as well as benefit.
Dogar 2018 reported 34 adverse events in 27 participants (11 in the varenicline arm and 16 in the placebo arm), while Chami 2022 reported 410 adverse events in 103 participants (50 in the varenicline arm and 53 in the placebo arm). When pooled, we did not find clear evidence of a difference between arms (RR 0.98, 95% CI 0.67 to 1.44; I2 = 31%; 2 studies, 662 participants; Analysis 2.2), although there was substantial imprecision. None of the events were deemed serious in either study (Analysis 2.3). Reported adverse events included respiratory disorders, gastrointestinal disorders, general disorders, immune system disorders, renal and urinary disorders, nervous system disorders, psychiatric disorders, headache, sleep disturbance, fatigue, skin reaction, change in appetite, and ear, nose, and throat symptoms.
Bupropion
Dogar 2014 was a three‐arm cluster‐RCT that tested self‐help materials versus brief behavioral intervention, and versus brief behavioral intervention plus bupropion among cigarette or waterpipe smokers. We considered the data provided by waterpipe smokers only. Compared to the behavioral support alone group, behavioral support plus bupropion did not provide clear evidence of increased quitting at the 25‐week follow‐up (RR 0.77, 95% CI 0.42 to 1.41; 1 study, N = 121; Analysis 3.1; very low‐certainty evidence). Dogar 2014 also tested a combined pharmacotherapy and behavioural support intervention. Compared to the self‐help materials group, behavioral support plus bupropion resulted in higher quit rates at 25 weeks of follow‐up. However, this was inconclusive due to imprecision and risk of bias (RR 1.94, 95% CI 0.94 to 4.00; 1 study, N = 86; Analysis 4.1; very low‐certainty evidence).
Dogar 2014 did not report on adverse events.
E‐health interventions
Two studies tested e‐health interventions for waterpipe use (Lipkus 2011; Mays 2021). Due to the variation in the studies, we did not pool the results of these studies using meta‐analysis. Mays 2021 tested a tailored mobile phone messaging waterpipe intervention compared with an untailored intervention or no intervention in young adults in the USA. Quit rates were higher among participants randomized to either the tailored or untailored intervention groups compared with those in the no intervention control group (RR 1.41, 95% CI 1.01 to 1.97; 1 study, N = 349; Analysis 5.1), but we did not detect a clear difference in quit rates between those receiving tailored messages and those receiving untailored messages (RR 1.24, 95% CI 0.89 to 1.71; 1 study, N = 130; Analysis 5.1). Lipkus 2011 tested an online educational intervention targeting college students in the USA, comparing the effect of an intensive twenty‐slide presentation with a brief eight‐slide presentation. Smoking abstinence rates were higher in the more intensive intervention group (RR 1.91, 95% CI 1.06 to 3.43; 1 study, N = 88; Analysis 5.1). Overall, we deemed this body of evidence as very low certainty.
Discussion
Summary of main results
We conducted two meta‐analyses for the primary outcome of waterpipe abstinence at three months or longer. The first analysis included five studies (N = 1030 participants) that tested face‐to‐face behavioral interventions with different intensities (2 to 7 counseling sessions) in comparison to lower intensity interventions (brief advice, self‐help materials, no support). The interventions were based on best practice cessation interventions for people who smoke combustible cigarettes in three studies, and adapted to people who used waterpipes in one study. Overall, we found low‐certainty evidence that behavioral interventions increase the chances of achieving waterpipe smoking abstinence. Our subgroup analyses tested the differences between exclusive waterpipe users and dual users, and between those selected based on their motivation to quit and those not selected based on their motivation, and found no clear evidence of subgroup differences.
We also found low‐certainty evidence with a point estimate showing higher waterpipe quit rates in response to varenicline; however, confidence intervals included the potential for no difference or lower quit rates as well as benefit. Adverse events in the two studies were reported, but there was no clear evidence of a difference between study arms, and none of the events reported were severe. In addition, data on the effectiveness of bupropion on waterpipe smoking cessation were inconclusive. We found only one study that tested the effect of bupropion. This study showed evidence of benefit from a combined intervention of bupropion and behavioral support compared with self‐help materials, but not compared with behavioral support alone. These results do not match the well‐established effect of varenicline and bupropion in cigarette smoking cessation (Hajizadeh 2023; Livingstone‐Banks 2023; Stead 2016). They should be treated with caution due to the limitations of the evidence, including imprecision and risk of bias, so we cannot rule out that varenicline and buproprion work as well in waterpipe users as in cigarette smokers. On the other hand, the profound differences between cigarette and waterpipe smoking behaviors and addiction patterns could mean a difference in how well these interventions work among waterpipe users. Waterpipe smoking has a solid social dependence dimension (Ward 2015). While cigarette smoking can be a solitary behavior, waterpipes are more often shared and smoked in socially enjoyable settings (Akl 2015; Alam 2020). In addition, the local sociocultural norms favoring waterpipe smoking and the lack of waterpipe control policies and regulations (e.g. taxation, enforcement of waterpipe smoking ban in public places, and applying health warning labels on waterpipe devices and tobacco) may contribute to lower waterpipe smoking cessation rates. Mass media campaigns and community outreach interventions that consider waterpipe‐specific social cues have the potential to challenge current waterpipe social norms and encourage cessation (Maziak 2019; Salloum 2016).
Two studies tested e‐health interventions for waterpipe use. Due to the variation in the studies, we did not pool the results of the studies using a meta‐analysis. The first study tested a brief versus intensive online educational intervention among college students. Smoking abstinence rates were higher in the intensive intervention compared with the brief intervention. The second study tested a tailored mobile text‐messaging intervention compared with either an untailored or no intervention. Quit rates were higher among participants in the tailored and untailored intervention groups compared with those in the no intervention control group. While we judged this evidence to be of very low certainty, e‐health is a promising strategy for waterpipe smoking cessation interventions. E‐health interventions might have a high reach because most waterpipe smokers are young adults who are receptive to e‐health interventions (Villanti 2017). In addition, e‐health interventions are highly scalable, with the potential to be available at a low cost at the population level. Therefore, further research is needed to explore the effectiveness of this setting.
Overall completeness and applicability of evidence
Given the strong cultural component in waterpipe use, which can vary substantially by country, there are potential limitations for the generalizability of our results. Studies were carried out in Iran, Vietnam, Syria, Lebanon, Pakistan, Egypt, and the USA. These countries have significant differences in culture, healthcare systems, and socioeconomic status, which should be taken into account when interpreting these results. In addition, although most studies described how the interventions were delivered (e.g. text messaging, in‐person, number of sessions, etc.), details about the content and specific behavior change techniques were often lacking. Future trials should include comprehensive descriptions of the interventions delivered. Finally, the studies were not comparable in terms of participants’ smoking status at baseline. For example, several studies recruited exclusive regular waterpipe smokers but used different definitions, such as smoked waterpipe at least once monthly, smoked waterpipe daily, or smoked waterpipe at least three times a week. Other studies recruited dual waterpipe users, defined as current dual users (smoked cigarettes and waterpipe in the past seven days) or daily dual users (smoked cigarettes and waterpipe more than once a day). Adopting standard definitions and assessments of waterpipe use and dependence would facilitate comparing waterpipe cessation interventions in the future (Maziak 2015b).
Quality of the evidence
We judged three studies to be at low risk of bias and six studies to be at high risk of bias. The most common reason was high risk of detection bias, because waterpipe‐use status was self‐reported, and there was differential face‐to‐face contact with investigators across study arms. For the primary abstinence outcome, we assessed the certainty of the evidence for our analysis using the GRADE system (Schünemann 2022). In the face‐to‐face behavioral intervention comparison, the certainty was downgraded one level due to imprecision (i.e. fewer than 300 events in the analysis), and another level because none of the included studies were at low risk of bias. In addition, the included behavioral interventions were heterogeneous and not always well described, and the trials used inconsistent definitions of baseline waterpipe smoking status. In the varenicline comparison, the certainty was downgraded by two levels due to imprecision (i.e. fewer than 150 events in the analysis). We downgraded the comparisons testing bupropion and e‐health interventions one level each, because of high risk of bias, and by two levels due to imprecision.
Potential biases in the review process
Cochrane's methods are designed to minimize reviewer bias where possible. For example, at least two review authors independently conducted study selection, data extraction, and risk of bias assessments. A potential key limitation of the review is that we may have failed to identify all relevant research for inclusion in the review. However, given the nature of the Cochrane methods, we are confident that any failures in the identification of studies for inclusion will not be systematic, and therefore, should not have a significant impact on the validity of our results.
We did not find enough studies to conduct a formal assessment of the risk of publication bias. However, this risk cannot be ruled out, and the possibility of publication bias should be taken into account when interpreting the results of this review.
Agreements and disagreements with other studies or reviews
Three other reviews have explored the effects of waterpipe smoking cessation interventions (Babaie 2021; Jawad 2016; Kader 2019). None of these reviews included meta‐analyses. The main objective of Jawad 2016 was to systematically review the literature for interventions targeting the prevention and cessation of waterpipe tobacco smoking, including legislative interventions. This review considered a broader range of study designs (randomized controlled trials (RCT), non‐RCT, observational studies, qualitative studies). Four of our included studies were reported descriptively in this review. They concluded that few studies showed promising results, but that small sample sizes and high risk of bias meant that better designed and larger trials were required for firmer conclusions. Kader 2019 attempted to systematically review interventions aimed at reducing waterpipe use with a focus on the RE‐AIM framework (reach, efficacy, adoption, implementation, maintenance). Five of our included studies were reported descriptively in this review. In line with our findings, they concluded that despite limited evidence on interventions for reducing waterpipe use, counseling and educational support appeared the most promising. Babaie 2021 aimed to identify the types of interventions used for preventing and controlling waterpipe smoking. This review identified 27 interventions that were categorized into four main themes, including preventive interventions, control interventions (policies and regulations), and actual enactment, and implementation of legislation and policies.
Prior research identified several high‐risk groups for waterpipe smoking. As evident from several epidemiological studies worldwide, waterpipe use has increased the most among youth and young adults (Maziak 2015b). Our search identified several potential studies targeting these high‐risk groups; however, most of these studies were not eligible for our review for several reasons. Most of these studies were uncontrolled pre‐post studies (Alzyoud 2018), had a focus on prevention outcomes (e.g. knowledge, harm perception, initiation, intention to quit [Leavens 2018; Mays 2016]), or a focus on reducing craving (Pourmohammad 2016), or the studies only reported short‐term (less than three months) abstinence (Alzyoud 2018; Johnson 2019; Nakkash 2018; Shishani 2018).
Another high‐risk group for waterpipe smoking is pregnant women. Waterpipe smoking during pregnancy increases the risk of spontaneous miscarriage, premature birth, and low birth weight (Al‐Sheyab 2016; Khabour 2016; Nematollahi 2018; Tamim 2008). Waterpipe smoking prevalence is relatively high among women in the Eastern Mediterranean region, ranging from 6% to 18%, due to the social acceptability of waterpipe smoking among women in that region compared with cigarettes (Azab 2013; Baheiraei 2012; Mirahmadizadeh 2008). We found only one study that was conducted in 2017, among 140 suburban pregnant women, who were identified as waterpipe users during their visit to receive prenatal care in healthcare centers in the outskirts of Zahedan, Iran (Mojahed 2018). The study was a quasi‐experimental study that involved a control group and a pre‐post‐test design, and assessed a behavioral intervention aimed at reducing waterpipe craving and dependence. Participants were assigned to either the intervention or control group. The intervention group received five motivational group interview sessions about waterpipe smoking withdrawal. The control group was given the usual care (brief advice) to quit the waterpipe. A notable reduction in waterpipe craving and dependence was observed between the two groups at two‐month follow‐up. However, the study did not measure quitting.
A growing high‐risk group that should be the focus of waterpipe smoking cessation efforts is dual users. Data from the first three waves (2013 to 2016) of the Population Assessment of Tobacco and Health Study, a nationally representative, longitudinal cohort study of adults and youth in the USA, indicated that most waterpipe poly tobacco users used cigarettes (youth: 49.4%, young adults: 59.4%, adults 25+: 63.2%), and had lower rates of quitting all tobacco than exclusive hookah users or hookah poly tobacco users who did not use cigarettes (Sharma 2020). In a study in Jordan, the prevalence of dual cigarette and waterpipe users among adolescents (age 11 to 17 years) was 30%, which was greater than waterpipe alone (21.1%), or cigarettes alone (6.7% [Alomari 2018]). Compared with exclusive waterpipe users, dual users have higher nicotine dependence, are less interested in quitting, and are less likely to make a quit attempt (Abbadi 2020; Shelley 2017). This group seems to be receiving increasing attention among researchers, as three of the included studies in our review enrolled waterpipe dual users. Our subgroup analysis testing the difference in the intervention effects between exclusive and dual waterpipe users showed no clear difference. However, only two studies were included in the dual users group and the study designs were heterogeneous, so the results must be interpreted cautiously.
Study selection PRISMA flow chart
Risk of bias summary: review authors’ judgments about each risk of bias item for each included study
+ Low risk, ‐ High risk and? Unclear risk
Comparison 1: Face‐to‐face behavioral interventions versus control, Outcome 1: Waterpipe abstinence
Comparison 1: Face‐to‐face behavioral interventions versus control, Outcome 2: Waterpipe abstinence: subgrouped by exclusive waterpipe use vs dual tobacco use
Comparison 1: Face‐to‐face behavioral interventions versus control, Outcome 3: Waterpipe abstinence: subgrouped by baseline motivation to quit
Comparison 2: Varenicline versus placebo, Outcome 1: Waterpipe abstinence
Comparison 2: Varenicline versus placebo, Outcome 2: Adverse events
Comparison 2: Varenicline versus placebo, Outcome 3: Serious adverse events
Comparison 3: Bupropion + behavioral support versus behavioral support alone, Outcome 1: Waterpipe abstinence
Comparison 4: Bupropion + behavioral support versus self‐help, Outcome 1: Waterpipe abstinence
Comparison 5: E‐health interventions for waterpipe smoking cessation, Outcome 1: Waterpipe abstinence
| Patients or population: people who use waterpipe Settings: clinics/communities/college campuses Intervention: behavioral support and pharmacotherapy Comparison: usual care/brief intervention/placebo | |||||
| Outcome | Anticipated absolute effects (95% CI) | Relative effect (Random effect) | Number of participants (number of studies) | Certainty of the evidence (GRADE) | |
|---|---|---|---|---|---|
| Risk without intervention, usual care | Risk with Interventions | ||||
| Face‐to‐face behavioral intervention vs minimal or no support (Measured 3 to 6 months with CO‐validated prolonged and continuous abstinence, and self‐reported abstinence) | 112 per 1000 | 297 per 1000 (187 to 472) | RR 2.64 (1.66 to 4.20) | 851 (5 cluster‐RCTs) | ⨁⨁⊝⊝ |
| Varenicline plus behavioral intervention vs placebo plus behavioral intervention ‐ (repeated point prevalence abstinence (not even a puff/chew/session in the last 7 days) at each of weeks 5, 12, and 25 (combined), verified by CO cutoff < 10 ppm) | 56 per 1000
| 70 per 1000 (39 to 126)
| RR 1.24 (0.69 to 2.24)
| 662 (2 RCT)
| ⨁⨁⊝⊝
|
| Bupropion therapy plus behavioral intervention vs behavioral intervention alone (continuous waterpipe smoking abstinence, defined as an expired CO of 9 ppm or less at the 6‐month postintervention follow‐up visits) | 449 per 1000 | 364 per 1000 (189 to 633) | RR 0.77 (0.42 to 1.41) | 121 (1 cluster‐RCT) | ⨁⊝⊝⊝
|
| Bupropion therapy plus behavioral support versus self‐help (continuous waterpipe smoking abstinence, defined as an expired CO of 9 ppm or less at the 6‐month postintervention follow‐up visits) | 194 per 1000 | 375 per 1000 (182 to 774) | RR 1.94 (0.94 to 4.00)
| 86 (1 cluster‐RCT) | ⨁⊝⊝⊝ |
|
| Impact | Number of participants (number of studies) | Certainty of the evidence (GRADE) | ||
| E‐health interventions | Due to the variation in the studies, we did not pool the results. Quit rates were higher among participants randomized to receive either the tailored or untailored intervention groups compared with those in the no intervention control group (RR 1.48, 95% CI 1.07 to 2.05; 1 study, N = 319). Smoking abstinence rates were higher among those randomized to receive a 20‐slide presentation compared with those randomized to receive an 8‐slide presentation (RR 1.86, 95% CI 1.08 to 3.21; 1 study, N = 70). | 389 (2 RCTs)
| ⨁⊝⊝⊝
| ||
| GRADE Working Group grades of evidence CI: confidence interval; CO: carbon monoxide; ppm: parts per million; RCT: randomized controlled trial; RR: risk ratio | |||||
| a Downgraded one level for risk of bias: no studies were at low risk of bias | |||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Waterpipe abstinence Show forest plot | 5 | 851 | Risk Ratio (M‐H, Random, 95% CI) | 2.64 [1.66, 4.20] |
| 1.2 Waterpipe abstinence: subgrouped by exclusive waterpipe use vs dual tobacco use Show forest plot | 5 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.2.1 People who exclusively use waterpipe | 5 | 482 | Risk Ratio (M‐H, Random, 95% CI) | 2.36 [1.62, 3.43] |
| 1.2.2 People who use both waterpipe and combustible cigarettes (dual use) | 2 | 485 | Risk Ratio (M‐H, Random, 95% CI) | 3.62 [1.57, 8.36] |
| 1.3 Waterpipe abstinence: subgrouped by baseline motivation to quit Show forest plot | 5 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.3.1 Participants motivated to quit | 2 | 218 | Risk Ratio (M‐H, Random, 95% CI) | 1.72 [0.96, 3.08] |
| 1.3.2 Participants not selected based on motivation to quit | 3 | 264 | Risk Ratio (M‐H, Random, 95% CI) | 2.94 [1.81, 4.77] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 Waterpipe abstinence Show forest plot | 2 | 662 | Risk Ratio (M‐H, Random, 95% CI) | 1.24 [0.69, 2.24] |
| 2.2 Adverse events Show forest plot | 2 | 662 | Risk Ratio (M‐H, Random, 95% CI) | 0.98 [0.67, 1.44] |
| 2.3 Serious adverse events Show forest plot | 2 | 0 | Risk Ratio (M‐H, Random, 95% CI) | Not estimable |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3.1 Waterpipe abstinence Show forest plot | 1 | 121 | Risk Ratio (M‐H, Random, 95% CI) | 0.77 [0.42, 1.41] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 4.1 Waterpipe abstinence Show forest plot | 1 | 86 | Risk Ratio (M‐H, Random, 95% CI) | 1.94 [0.94, 4.00] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 5.1 Waterpipe abstinence Show forest plot | 2 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 5.1.1 Text messages vs no intervention | 1 | 349 | Risk Ratio (M‐H, Random, 95% CI) | 1.41 [1.01, 1.97] |
| 5.1.2 Tailored vs untailored text messages | 1 | 230 | Risk Ratio (M‐H, Random, 95% CI) | 1.24 [0.89, 1.71] |
| 5.1.3 20‐slide vs 8‐slide presentation | 1 | 88 | Risk Ratio (M‐H, Random, 95% CI) | 1.91 [1.06, 3.42] |
