Combination Pharmacological Interventions for Smoking Cessation and Post-Cessation Weight Gain: A Meta-Analysis of Randomized Controlled Trials

Ⅰ. Introduction

Over 8 million deaths occur annually worldwide due to tobacco use, with approximately 7 million of those deaths resulting from direct use and the remainder resulting from exposure to second-hand smoke (World Health Organization [WHO], 2019). Increased health risks linked to cigarette smoking, such as the development of coronary heart disease, stroke, lung cancer, and diabetes, have been well documented (U.S. Department of Health and Human Services [HHS], 2014). Recent findings on the increased health risks among cigarette smokers have noted a two- to four-fold increased risk for developing coronary heart disease and stroke, a 25-fold increased risk for developing lung cancer, and a 30–40% increased risk for developing diabetes (HHS, 2014). In Korea, 36.7% of males and 7.5% of females were reported as active smokers in 2018 (Statistics Korea, 2020). The crude rate from lung cancer was 72.9 per 100,000 among males in 2017 (Korean Statistical Information Service, 2020). A higher smoking rate was identified in people with disabilities (35.9%) as compared to those without disabilities (19.0%) (Choi & Choi, 2016).

Weight gain after smoking cessation is a common concern among smokers. In one trial, 50% of females and 26% of male smokers were weight concerned (Clark et al., 2006). Despite the availability of effective smoking cessation treatments, evidence-based findings have indicated approximately 80% of smokers gain 6 to 8 pounds after quitting and some gain over 20 pounds, with heavy smokers at risk for the largest weight gain (Froom, Melamed, & Benbassat, 1998; Healton & Fiore, 2008; Lycett, Munafò, Johnstone, Murphy, & Aveyard, 2011; O’Hara et al., 1998; Williamson et al., 1991; Zhu et al., 2002). Moreover, post-cessation weight gain is associated with an increase in the incidence of type 2 diabetes and a 30% increased risk of hypertension (Gerace, Hollis, Ockene, Svendsen, & Group, 1991; Yeh, Duncan, Schmidt, Wang, & Brancati, 2010). Knowledge of potential adverse health effects associated with post-cessation weight gain coupled with a fear of weight gain among quitters can compromise adherence to smoking cessation interventions and play a role in early relapse (Alberg, Carter, & Carpenter, 2007).

Occupational therapy promotes smoking cessation by analyzing smoking triggers and introducing individualized strategies to develop new meaningful activities (Jordan, 2019). Occupational therapists provide client-centered lifestyle interventions, such as meal replacement, social support, education, facilitated exercise, and goal setting, to promote the desired outcomes, such as weight loss and smoking cessation (Conn, Bourke, James, & Haracz, 2019; Jordan, 2019). To be specific, the American Occupational Therapy Association (AOTA) describes smoking cessation as one of the mental health recovery practices in occupational therapy. AOTA suggests monitoring and reacting to changes in mental health conditions corresponding to individualized strategies to smoking cessation (Champagne et al., 2016). Of note, Wei and his colleagues proposed that the combination of pharmacological interventions and occupational therapy could encourage treatment results (Wei, Feldmann Jr, Brascher, & Benrath, 2014). Accordingly, acknowledging the expected pharmacological interventions and outcomes could be required to design individualized occupational therapy lifestyle strategies for enhancing and maintaining smoking cessation.

However, research addressing combined effects of both increased smoking abstinence and decreased post-cessation weight gain via combined pharmacotherapies is limited. According to the 2008 Public Health Service guideline update (Fiore et al., 2008), recommended combination pharmacotherapies known to delay weight gain for patients who are motivated to quit smoking include bupropion SR and 4-mg nicotine gum or 4-mg nicotine lozenge nicotine replacement therapies. However, the 2008 Public Health Service guideline update research did not address combination therapy of varenicline, an FDA first-line smoking cessation medication, with Nicotine Replacement Therapies (NRT) or bupropion SR (Fiore et al., 2008). In an RCT examining varenicline’s efficacy for smoking cessation versus bupropion SR and placebo, varenicline was more efficacious for smoking cessation, as measured by 9-12-week continuous abstinence (44% vs. 2.97%, p < .0001) (Nides et al., 2008).

Given the 2008 Public Health Service guideline update calls for future research addressing the efficacy of medications to prevent weight gain during quit attempts, this meta-analysis aims to evaluate the effect of various combined pharmacotherapies on continuous smoking abstinence and post-cessation weight change. Findings from this meta-analysis seek to inform occupational therapists in their development of individualized lifestyle interventions for smoking abstinence and post-cessation weight change tailored to different pharmacotherapies.

This study will address the following research questions:

∐. Materials and methods

1. Data sources and study selection

In collaboration with an experienced librarian, the following databases were searched: PubMed, International Pharmaceutical Abstracts (IPA), Web of Science, and Cochrane Library. The core search criteria included: “Tobacco use cessation”, “Adult”, “Body weight changes”, and “Tobacco use cessation products”. Sample search strings used for PubMed are provided in Appendix B. In order to assess current literature, we limited the findings to articles (1) published between 1990 to July, 2016, (2) written in English, (3) that reported combined pharmacotherapy (at least one type of FDA approved smoking cessation medication), (4) with a study aim focused on weight change between baseline and follow-up, (5) reporting minimum 2-week follow-up, (6) involving randomized control trials (RCT), (7) and with subjects who were adult smokers with no current psychiatric or other serious illness, who had smoked 10 or more cigarettes per day for more than 6 months, and weighed at least 45 kilograms. Furthermore, included studies required that enrolled patients reported no recent experience with other cessation medication or successful smoking abstinence. Articles that were excluded were: (1) those that included a behavioral intervention, (2) had a sample size less than 30, (3) switched medications during the study, and (4) were not written in English. Grey literature and unpublished conference abstracts or manuscripts were not included in this meta-analytic review.

Appendix B.

PubMed Search Strategy

3. Data extraction

Two authors conducted the initial search and reviewed the results independently. After deletion of articles published in English before 1990 and that did not report RCTs, the initial search results for each database were: 505 for PubMed, 109 for IPA, 268 for Web of Science, and 313 for Cochrane Library. Two authors reviewed 1,195 articles independently. A search for article duplicates, yielded 138 article exclusions for duplication. Preliminary title-based exclusion resulted in 911 article exclusions. Among the abstracts reviewed, 114 were excluded for not meeting the inclusion criteria. Full text was examined for the remaining 32 articles, 19 of which were excluded for use of monotherapy, as opposed to combined pharmacotherapy, 4 of which were excluded for analysis of a behavioral component, and 1 of which was excluded for the sole use of nicotine patch non-responders. Consequently, eight peer-reviewed articles published between January, 1990 and July, 2016 were included in this meta-analysis (see Figure 1) (Biberman, Neumann, Katzir, & Gerber, 2003; Bohadana, Nilsson, Rasmussen, & Martinet, 2000; Ebbert et al., 2014; Hajek, Smith, Dhanji, & McRobbie, 2013; Jorenby et al., 1999; Koegelenberg et al., 2014; O’Malley et al., 2006; Piper et al., 2007). Short-term smoking abstinence outcomes were defined as continuous abstinence assessed before or between 6 and 12 weeks beyond treatment completion. Mid-term outcomes were defined as continuous abstinence assessed at 6 months and long-term outcomes were defined as 12 months after treatment completion. Short-term weight change outcomes were defined as mean weight change assessed before or between 4 and 8 weeks after combined pharmacotherapy and smoking cessation treatment completion. Mid-term outcomes were defined as mean weight change assessed at 6 months, and long-term outcomes were defined as mean weight change assessed at 12 months after treatment completion.

Figure 1

Flow Diagram of Meta-Analysis

5. Publication bias

Potential for publication bias was assessed via both funnel plots and the Egger’s regression test (Sterne, Gavaghan, & Egger, 2000). Asymmetry in a funnel plot was considered a risk of publication bias and the Egger’s regression test was used for statistical verification of publication bias. A symmetrical funnel plot of the aggregate fixed effect model estimate across all continuous abstinence time points suggested low risk for publication. A non-significant Egger’s regression test (p = .55) confirmed low potential for publication bias on continuous abstinence outcomes. Funnel plot symmetry for the aggregate fixed effect model estimate across all weight change time points was confirmed with a non-significant Egger’s regression test (p = .19) (Supplementary Figure 2). The test results were generated by R 0.99.902.

Supplementary Figure 1.

Continuous Abstinence

Supplementary Figure 2.

Weight Change

Ⅲ. Results

1. Characteristics of included studies

Eight randomized controlled combination pharmacological trials with 2,513 participants were examined in this meta-analysis. Included studies consisted of various combinations of pharmacological therapies compared against monotherapy counterparts, including: oral selegiline plus nicotine patch vs nicotine patch alone, nicotine inhaler plus nicotine patch vs nicotine inhaler alone, varenicline plus bupropion SR vs varenicline alone, naltrexone hydrochloride plus nicotine patch vs nicotine patch alone, varenicline plus nicotine patch vs varenicline alone, bupropion plus nicotine patch vs bupropion alone, and bupropion plus nicotine gum vs bupropion alone (see Table 1). Table 1 summarizes the baseline demographics, number of participants, weight, and Fagerström Score for included studies. All eight trials recruited smokers who were aged 18 and over, smoked 10 or more cigarettes per day for more than 6 months and weighed at least 45 kilograms, were not breastfeeding or pregnant, had no current psychiatric or other serious illness, and were motivated to quit smoking.

Table 1

Baseline Characteristics of Included Studies

2. Main effects

The fixed effect model (Borenstein, Hedges, Higgins, & Rothstein, 2009) was used to analyze pooled continuous smoking abstinence and post-cessation weight change estimates across a total of eight studies with 2,513 participants. Study heterogeneity was not statistically significant (Q = 6.87; p = .99) across continuous smoking abstinence outcomes. An aggregate fixed effect model estimate across all continuous abstinence time points revealed combined pharmacotherapy yielded a significant increase in continuous smoking abstinence (OR = 1.81, 95% CI 1.38 to 2.24, p < .001) (see Figure 2). Statistically significant heterogeneity (Q = 46.8; p = .001) was found across weight change time points. An aggregate fixed effect model estimate across all weight change time points suggested combined pharmacotherapy yielded a significant mean decrease (-.15 kg) in post-cessation weight change (95% CI -.22 to –.08, p = .001) (see Figure 3).

Figure 2

Aggregate Continuous Abstinence Odds Ratios

Figure 3

Aggregate Weight Change Kilograms

3. Subgroup analyses

1) Continuous abstinence

(1) Short-, mid-, and long-term continuous abstinence

Fixed effect model subgroup analyses of pooled continuous smoking abstinence at 6-12 weeks (OR = 1.7, 95% CI .51 to 2.88, p = .005), 6 months (OR = 1.89, 95% CI .72 to 3.05, p = .001), and 12 months (OR = 2.03, 95% CI .52 to 3.54, p = .008) post-cessation all indicated significant increases in continuous smoking abstinence.

(2) Varenicline plus nicotine patch vs bupropion plus NRT

Fixed effect model varenicline subgroup analyses of continuous abstinence indicated significant increases in continuous smoking abstinence among both combined varenicline plus nicotine patch (OR = 1.74, 95% CI .28 to 3.2, p = .02) and bupropion plus NRT pharmacotherapy (OR = 1.46, 95% CI .04 to 2.87, p = .04).

2) Weight change

(1) Short-, mid-, and long-term continuous abstinence

Fixed effect model subgroup analyses of pooled mean weight change from 6 to 8 weeks (-.14 kg, 95% CI -.26 to -.02, p = .02) post-cessation suggested a significant decrease in mean weight change. Pooled mean estimates for 6 months post-cessation weight change indicated a non-significant increase in weight (.08 kg, 95% CI -.06 to .23, p = .23). However, 12-month pooled mean weight change suggested a non-significant decrease in weight change (-.15 kg, 95% CI -.32 to .03, p = .09).

(2) Combined varenicline plus nicotine patch vs bupropion plus NRT

Fixed effect model varenicline subgroup analyses of mean weight change indicated significant increases in weight among the combined varenicline pharmacotherapies (.21 kg, 95% CI .03 to .39, p = .01) (Supplementary Figure 3). On the other hand, bupropion plus NRT combined pharmacotherapies under study were found to yield a significant decrease in post-cessation mean weight change (-.15 kg, 95% CI -.28 to -.03, p = .01) (see Supplementary Figure 4).

Supplementary Figure 3.

Weight Change Varenicline Plus Nicotine Patch

Supplementary Figure 4.

Weight Change Bupropion Plus NRT

Ⅳ. Discussion

Aggregate fixed effect model estimates across continuous abstinence and mean post-cessation weight change time points revealed combined pharmacotherapy yielded a significant increase in continuous smoking abstinence and a significant mean decrease in post-cessation weight change among abstainers. Subgroup analyses of continuous smoking abstinence based on 4-8 week, 6 month, and 12 month follow-up outcomes also indicated significant increases in smoking abstinence for combination pharmacotherapy as compared to NRT, varenicline, or bupropion monotherapy counterparts. On the other hand, a post-cessation decrease in mean weight change only remained significant at 6-8 weeks follow-up among abstainers. These findings are consistent with a previous systematic review indicating a short-term decrease in mean weight change among abstainers in combined smoking cessation pharmacological interventions as compared to monotherapy counterparts (Yang, Bhowmik, Wang, & Abughosh, 2013). In addition to providing an updated review of combined pharmacotherapies, this meta-analysis extends prior monotherapy findings with evidence for significant increased post-cessation weight gain when comparing combined varenicline plus nicotine patch vs varenicline alone to combined bupropion plus NRT vs bupropion alone (Gonzales et al., 2006; Hays & Ebbert, 2003; Jorenby et al., 2006).

Subgroup analyses comparing combined varenicline plus nicotine patch vs bupropion plus nicotine patch/gum examined the 2008 Public Health Service guideline update’s apparent gap in addressing the role of combination therapy of varenicline plus NRT or bupropion SR plus NRT for weight-concerned smokers. Subgroup analyses comparing combined varenicline plus nicotine patch vs bupropion plus nicotine patch/gum suggested significant increases in continuous smoking abstinence among both combined varenicline plus nicotine patch and bupropion plus NRT pharmacotherapy. However, subgroup analyses comparing combined varenicline plus nicotine patch vs bupropion plus nicotine patch/gum suggested significant increases in mean weight gain among varenicline plus nicotine patch pharmacotherapy abstainers, whereas bupropion plus NRT combined pharmacotherapies demonstrated a small yet statistically significant decrease in post-cessation mean weight gain among abstainers. These findings corroborate prior research suggesting bupropion SR monotherapy rendered significant decreased mean post-cessation weight gain among abstainers when compared to varenicline monotherapy (Gonzales et al., 2006; Hays & Ebbert, 2003; Jorenby et al., 2006).

Unfortunately, our literature searches identified no available studies of specific combined pharmacological smoking cessation interventions that examined the efficacy of bupropion SR plus varenicline vs bupropion alone for comparison. One included study examined the efficacy of combined varenicline plus bupropion SR vs varenicline alone (Ebbert et al., 2014). Findings from this study indicated significant increased continuous smoking abstinence at 12 months and significant short-term decreased weight gain at 12 weeks with non-significant decreases in weight gain among abstainers at 6 and 12 months of follow-up. Given comparable rates of significantly increased continuous smoking abstinence and decreased post-cessation weight gain evidenced among combined varenicline plus bupropion SR vs varenicline alone and bupropion SR plus nicotine patch vs bupropion alone (Ebbert et al., 2014), future research is warranted to assess the efficacy of varenicline plus bupropion SR smoking cessation interventions to attenuate post-cessation weight gain among abstainers.

Limitations to this meta-analysis consist of the restricted availability of multiple weight change follow-up time points for each included study. Given that post-cessation weight change was largely measured as a secondary outcome, weight change follow-up time points were not fully reported analogously to continuous smoking abstinence follow-up time points. Consequently, except for Ebbert et al. (2014), each study included only one weight change time point. Furthermore, the limited number of combined varenicline and bupropion pharmacotherapies affected our subgroup analyses addressing specific pharmacologic combinations. For instance, only two included studies (Ebbert et al., 2014; Koegelenberg et al., 2014) reported mean post-cessation weight change at 6 months of follow-up. However, since one of these studies (Ebbert et al., 2014) was a combination of varenicline plus bupropion SR (with a non-significant decrease in mean weight change at 6 months post-cessation) while the other 6-month time point provided (Koegelenberg et al., 2014) was a combination of varenicline plus nicotine patch (with a nonsignificant increase in weight gain), the analysis was swayed to a non-significant weight gain at 6-months. Also, some of the included studies (Jorenby et al., 1999; O’Malley et al., 2006; Piper et al., 2007) included more treatment conditions than were analyzed in this study. For instance, Jorenby et al. (1999) included nicotine-patch only, bupropion only, bupropion plus nicotine patch, and placebo only conditions in their study. Given that the purpose of this study was to examine combined pharmacotherapy, the combined therapies were compared against monotherapy counterparts. In addition, due to our inclusion criteria, this study may only be generalized to adult smokers with no current psychiatric or other serious illnesses.

While the primary purpose of this study was to provide an updated review of the effectiveness of combined pharmacotherapies for increased smoking abstinence and post-cessation weight gain prevention, this meta-analytic review also accounted for different pharmacological combinations (i.e., mechanisms of action) via subgroup analyses. Subgroup analyses further strengthened our interpretation of the data by separating bupropion combinations and varenicline plus nicotine patch combinations. In addition, missing standard deviations were requested for weight change analyses from study authors and all (but one 52-week follow-up data point) were retrieved (Hajek et al., 2013; Jorenby et al., 1999; Koegelenberg et al., 2014; Piper et al., 2007). Lastly, revised means and standard deviations for post-cessation weight change were also requested in order to calculate an unbiased effect size excluding one study’s reported outlier (Hajek et al., 2013).

In comparison to monotherapy, combination pharmacotherapy generates significantly increased smoking abstinence and decreased post-cessation weight change among abstainers, particularly among bupropion plus NRT combined pharmacotherapies when compared against varenicline plus nicotine patch. However, significant decreases in post-cessation weight change have been observed only short-term, in 6- to 8-week follow-up. Bupropion plus NRT combined pharmacotherapy was also associated with decreased post-cessation weight change among abstainers at 52-week follow-up, but this difference was not statistically significant. Significantly increased continuous smoking abstinence and significantly decreased weight gain evidenced with varenicline plus bupropion SR vs varenicline alone warrants future research to assess the efficacy of varenicline plus bupropion SR pharmacological smoking cessation interventions to reduce post-cessation weight gain among abstainers. Combination pharmacotherapy seems to provide weight-concerned smokers with a viable clinical option for delaying short-term weight gain which might otherwise compromise adherence to smoking cessation.