Exercise interventions for smoking cessation
Abstract
Background
Taking regular exercise may help people give up smoking by moderating nicotine withdrawal and cravings, and by helping to manage weight gain.
Objectives
To determine whether exercise‐based interventions alone, or combined with a smoking cessation programme, are more effective than a smoking cessation intervention alone.
Search methods
We searched the Cochrane Tobacco Addiction Group Specialized Register in April 2014, and searched MEDLINE, EMBASE, PsycINFO, and CINAHL Plus in May 2014.
Selection criteria
We included randomized trials which compared an exercise programme alone, or an exercise programme as an adjunct to a cessation programme, with a cessation programme (which we considered the control in this review). Studies were required to recruit smokers or recent quitters and have a follow‐up of six months or more. Studies that did not meet the full inclusion criteria because they only assessed the acute effects of exercise on smoking behaviour, or because the outcome was smoking reduction, are summarised but not formally included.
Data collection and analysis
We extracted data on study characteristics and smoking outcomes. Because of differences between studies in the characteristics of the interventions used we summarized the results narratively, making no attempt at meta‐analysis. We assessed risk of selection and attrition bias using standard methodological procedures expected by The Cochrane Collaboration.
Main results
We identified 20 trials with a total of 5,870 participants. The largest study was an internet trial with 2,318 participants, and eight trials had fewer than 30 people in each treatment arm. Studies varied in the timing and intensity of the smoking cessation and exercise programmes offered. Only one included study was judged to be at low risk of bias across all domains assessed. Four studies showed significantly higher abstinence rates in a physically active group versus a control group at end of treatment. One of these studies also showed a significant benefit for exercise versus control on abstinence at the three‐month follow‐up and a benefit for exercise of borderline significance (p = 0.05) at the 12‐month follow‐up. Another study reported significantly higher abstinence rates at six month follow‐up for a combined exercise and smoking cessation programme compared with brief smoking cessation advice. One study showed significantly higher abstinence rates for the exercise group versus a control group at the three‐month follow‐up but not at the end of treatment or 12‐month follow‐up. The other studies showed no significant effect for exercise on abstinence.
Authors' conclusions
Only two of the 20 trials offered evidence for exercise aiding smoking cessation in the long term. All the other trials were too small to reliably exclude an effect of intervention, or included an exercise intervention which may not have been sufficiently intense to achieve the desired level of exercise. Trials are needed with larger sample sizes, sufficiently intense interventions in terms of both exercise intensity and intensity of support being provided, equal contact control conditions, and measures of exercise adherence and change in physical activity in both exercise and comparison groups.
PICOs
Plain language summary
Do exercise interventions help people quit smoking
Review question
We reviewed the evidence about the effect of exercise programmes in people who want to quit smoking. We looked at whether exercise programmes, either alone or combined with stop smoking programmes, helped more people to quit at six months or longer than stop smoking programmes alone or stop smoking programmes combined with health education.
Background
Specialist clinics and self‐help materials regularly recommend exercise to people who want to quit smoking. Taking regular exercise may help people give up smoking by helping with withdrawal and cravings, and by helping to manage weight gain.
Study characteristics
The evidence is current to April 2014. We found 20 trials with a total of 5,870 participants. Nine studies were in women only and one study was in men only. Studies varied in the timing and intensity of programmes offered. We only included studies that measured smoking at six months or longer. In most of the trials, the exercise programmes included group and home‐based exercise.
Key results
Since these studies used different types and intensities of exercise programmes, the results were not combined.
In four studies, people who received the exercise programme were significantly more likely to quit smoking at end of treatment than people who only received a stop smoking programme. Only two of the 20 trials offered evidence for exercise helping people to quit smoking in the long term. In one of these studies, the people in the exercise group had significantly higher quit rates at three‐month follow‐up and at 12 months, the results from this study were borderline significant. In this study, people who received the exercise programme were more than twice as likely to still be quit at 12 months. Another study reported significantly higher quit rates at six month follow‐up for a combined exercise and smoking cessation programme compared with brief smoking cessation advice. The other studies did not find an effect of exercise programmes on quit rates but that could have been because they were small studies or because the exercise programmes were not intense enough.
Quality of evidence
The level of evidence for whether exercise programmes help people quit smoking is very low and more research is needed. There are issues with study design, possible risk of bias, and differences between the studies.
Authors' conclusions
Summary of findings
| Exercise interventions for smoking cessation | |||
| Population: People who smoke or people who have recently quit smoking Intervention: Exercise programmes alone or as adjuncts to smoking cessation programmes Comparison: Smoking cessation programmes without exercise components | |||
| Outcomes | Effects of exercise interventions for smoking cessation | No of Participants | Quality of the evidence |
| Smoking cessation at longest follow‐up (6+ months) | At longest follow‐up, one study detected a difference of borderline significance in favour of the intervention group. Another study reported significantly higher abstinence rates at six month follow‐up for a combined exercise and smoking cessation programme compared with brief smoking cessation advice, but not when compared to the full smoking cessation programme. No other studies detected a significant difference between intervention and comparison groups at longest follow‐up. | 5870 (20 studies) | ⊕⊝⊝⊝ very low1 |
| GRADE Working Group grades of evidence | |||
| 1 Risk of bias unclear or high for all but one included study. High level of clinical heterogeneity precluded meta‐analysis. Issues with inadequate sample size in majority of included studies. | |||
Background
Cigarette smoking is an important risk factor for cardiovascular disease, cancer and hypertension, and is one of the major causes of premature mortality in industrialized nations (Doll 2004; Peto 1996). Stopping smoking prolongs life and reduces morbidity (USDHHS 1990; Taylor 2002). Many attempts to stop smoking are made unaided (West 1997; Hughes 2004), with a success rate (6 to 12 months prolonged abstinence) of around 3 to 5% (Hughes 2004). Aided quit attempts, particularly through a combination of behavioural counselling and nicotine replacement therapy (NRT), bupropion or varenicline can improve success rates, but these remain low (Cahill 2012; Hughes 2014; Stead 2012). More effective smoking cessation interventions are needed.
Effect of exercise on tobacco withdrawal and cravings
Exercise has been proposed as an aid for smoking cessation (Hill 1981). In this review the terms exercise and physical activity (PA) are used interchangeably and refer to both 'lifestyle' physical activities, such as walking, as well as more formal structured activities, such as using a stationary cycle. The severity of 'desire to smoke' reliably predicts relapse in smokers who are trying to stop (Doherty 1995; West 1989) and interventions are required which reduce the desire to smoke. In experimental studies, cardiovascular‐type exercise has been shown to have an acute effect on reducing both psychological withdrawal symptoms and desire to smoke in abstinent smokers. This has been shown to be the case for both brief (5 to 10 minute) bouts of moderate intensity exercise among smokers who have been abstinent overnight and for 30 to 40 minute bouts of vigorous intensity among smokers who are trying to quit smoking (Haasova 2013; Haasova 2014; Taylor 2007b; Roberts 2012 ‐ also see the table of acute studies in Appendix 1). The mechanism underlying the observed beneficial effect of exercise on withdrawal and cravings is not clear. Exercise has been shown to have some similarities to smoking in its effects on stimulating the central nervous system (Russell 1983) and on neurobiological processes in the brain (Dishman 2009), including increasing beta‐endorphin levels in smokers (Leelarungrayub 2010), and consequently it has been argued that exercise may provide an alternative reinforcer to smoking (Marlatt 1985). This argument is consistent with behavioural theories of choice (Correia 1998) and animal studies have demonstrated that exercise is an effective alternative reinforcer to illicit substances for rats (e.g. Cosgrove 2002), but no studies could be identified which have investigated the role of exercise as an alternative reinforcer to smoking. It seems plausible that the attention to somatic cues during exercise presents a unique strategy for distracting smokers from the cravings and negative cognitions experienced during smoking abstinence, although the findings from one study suggest that distraction is unlikely to play a major role (Daniel 2006). Another possible mechanism is that exercise influences cognitive functioning in smokers; for example, exercise appears to reduce attentional bias to smoking images (Janse van Rensburg 2009a; Oh 2014).
Besides the potential benefits of exercise for moderating psychological withdrawal symptoms and cravings, there is evidence for exercise reducing post‐smoking cessation weight gain in the long‐term (Farley 2012), and for reducing cravings for sweet foods during the first week of abstinence (Teo 2014). The weight control benefits of exercise may be of particular importance to female smokers (see Linke 2013 for discussion of exercise interventions for female smokers) who report smoking to control weight (USDHHS 2001; Weekley 1992), and report fear of post‐cessation weight gain as a motivation for continued smoking (Clark 2004; Sorenson 1992; USDHHS 2001) and for smoking relapse (Gritz 1989; Klesges 1992). Exercise has also been shown to have a positive effect on other factors that may protect against smoking relapse, including perceived coping ability (Steptoe 1989) and self esteem (Spence 2005). In addition, being physically active has many general health benefits (Garber 2011), which have been observed for smokers who have quit (Albrecht 1998; Niaura 1998; Shinton 1997) and for continuing smokers (Colbert 2001; Hedblad 1997; Senti 2001). Moreover, a review suggests that participation in regular physical activity satisfies eight of the principles characterising a tobacco harm reduction strategy (deRuiter 2006). For example, one study observed that physical activity levels were inversely associated with lung carcinoma among current and former smokers (Leitzmann 2009).
Associations between exercise and smoking behaviours
Evidence from a number of large cross‐sectional surveys indicates that levels of PA are inversely related to smoking rates (e.g. Boutelle 2000; Boyle 2000; Hu 2002; Picavet 2010; Schuman 2001; Takemura 2000). Among smokers with a depressive disorder attempting to quit, lower physical activity levels at baseline predicted an increased likelihood of relapsing (Bernard 2012). One survey has shown an association between higher levels of cigarette dependence and lower levels of physical activity (Loprinzi 2014). Other evidence from cross‐sectional studies suggests that this relationship may be influenced by both gender and mode of PA. For example, when only examining leisure‐time PA, heavy smoking has been shown to be inversely related to PA in men but not in women (Schroder 2003). Elsewhere, participation in sport has been negatively associated with smoking in men but not in women (Helmert 1994). Additionally, some earlier studies have shown a weak relationship or no relationship between PA and smoking (Blair 1985; King 1992).
We only found one study (Sasco 2002) which examined the relationship between smoking and exercise in pre‐adolescents, and this cross‐sectional study reported a positive association between engaging in PA and 'ever smoking'. Among adolescents, cross‐sectional studies have consistently shown that smoking is negatively associated with participation in sport (Escobedo 1993; Peretti‐Watel 2003; Rodriguez 2004; Rodriguez 2008) and with overall levels of PA (Coulson 1997; Pate 1996; Verkooijen 2008; Ward 2003). There is some evidence to suggest that this pattern may be different for boys versus girls and some of the evidence is contradictory. For example, a cross‐sectional study of adolescents found a negative association between sporting activity and smoking for boys and heavy smoking, but not for girls or for lighter smokers (Peretti‐Watel 2002). Another study observed no association between sports participation and smoking levels in males (Davis 1997), while a prospective study found that leisure‐time PA was positively associated with initiating smoking for girls but not for boys (Aaron 1995). Two prospective studies found that higher levels of PA reduced the odds of starting smoking for boys and girls both during childhood (Audrain‐McGovern 2003) and adulthood (Kujala 2007). One study showed that the negative association between physical activity and smoking is mediated by having a physically active identity (Verkooijen 2008). A detailed review of studies examining associations between smoking and physical activity has been published by Kaczynski 2008.
Smokers trying to quit are likely to be more receptive to an active lifestyle than smokers in general (Doherty 1998; King 1996). An exercise‐based smoking cessation intervention has been found to significantly improve cardiovascular disease biomarkers (e.g. inflammatory markers) within three months (Korhonen 2011) and smokers report that they value exercise as a strategy for reducing the risk of developing tobacco‐related disease (Haddock 2004). In addition, higher levels of exercise are associated with less depression in smokers (Vickers 2003; Williams 2008). Being physically active has been positively associated with initiating a quit attempt (deRuiter 2008; Gaulthier 2012; Haddock 2000), with confidence to maintain smoking abstinence (King 1996) and with success at stopping smoking (Abrantes 2009; Derby 1994; Paavola 2001; Sedgwick 1988), although one large survey found no association between exercise levels and intention to quit smoking (Nguyen 1998). Other work shows a positive trend between avoiding relapse to smoking and physical health and fitness (Metheny 1998) and a significantly reduced risk of smoking relapse among those who are more physically active (McDermot 2009), including among those with depression (Bernard 2012).
Overall, from the above evidence one might hypothesize that pursuing regular exercise during an attempt to stop smoking could act both to reduce nicotine withdrawal symptoms and cravings and to increase rates of smoking cessation. In practice, exercise has for many years been routinely recommended as an aid to smoking cessation by specialist smoking clinics (e.g. Everson 2010; Hurt 1992), by pharmaceutical companies (e.g. Boots 1998), in self‐help guides (Ashelman 2000; Marcus 2004 ), by physical therapists (Pignataro 2012), and in national guidelines (e.g. Quit 1994; Woodhouse 1990; USDHHS 2008), and many smokers are likely to view physical activity as an aid to quitting (Everson‐Hock 2010a). In the short term, most smokers are unlikely to spontaneously increase their levels of PA after quitting (Allen 2004; Hall 1989; Vander Weg 2001), and the present review examines studies which have evaluated exercise interventions as an aid to smoking cessation.
The objective of the main review is to evaluate studies on the effect of exercise on smoking cessation after at least six months. Two secondary questions are also addressed: the acute effects of exercise on smoking related outcomes, and, for this update of the review, the effects of physical activity on levels of cigarette consumption. Evidence for the acute effects of exercise on cigarette cravings and withdrawal provides the main rationale for promoting exercise for smoking cessation; therefore we consider it important to present these studies. The effect of a physical activity intervention on levels of cigarette consumption is important as there is increasing interest in the effectiveness of interventions (pharmacological and behavioural) targeting smoking harm reduction (NICE 2013). While smoking cessation is the ultimate goal of interventions, there is evidence that cessation can be induced through smoking reduction, among both smokers who wish to quit and those who do not. Exercise interventions that seek to reduce cigarette consumption, whether it be to implicitly or explicitly lead to a quit attempt, are therefore reviewed.
Objectives
The primary objective was to establish whether exercise‐based interventions alone, or combined with a smoking cessation programme, are more effective than a smoking cessation intervention alone.
Methods
Criteria for considering studies for this review
Types of studies
Randomized controlled trials.
Types of participants
Smokers wishing to quit or recent quitters.
Types of interventions
Programmes aimed at increasing physical activity, either alone or as an adjunct to a smoking cessation intervention, compared with a smoking cessation programme alone. Interventions which included exercise in a multiple component smoking cessation programme were excluded since the specific effects of exercise on smoking abstinence could not be addressed. Multiple risk factor interventions where smoking cessation was one of a number of health‐related outcomes were excluded for the same reason.
Types of outcome measures
Smoking cessation at the longest follow‐up reported. Trials with less than six months' follow‐up were not included (i.e. a study was included if follow‐up was at least six months post‐baseline, six months post‐quit or six months post‐treatment).
Search methods for identification of studies
We searched the Specialized Register of the Cochrane Tobacco Addiction Group for studies including the terms 'exercise' or 'physical activity' in the title, abstract or keyword fields. At the time of the search in April 2014 the Register included the results of searches of the Cochrane Central Register of Controlled trials (CENTRAL), issue 3, 2014; MEDLINE (via OVID) to update 20140321; EMBASE (via OVID) to week 201413; and PsycINFO (via OVID) to update 20140317. See the Tobacco Addiction Group Module in the Cochrane Library for full search strategies and list of other resources searched. We also searched MEDLINE, PubMed, EMBASE, PsycINFO, and CINAHL Plus, using the terms 'smoking', 'smoking cessation', 'exercise', 'physical activity', and 'intervention' (searches completed May 2014). We also carried out a hand search of reference lists and conference abstracts, conducted additional searches on key authors, and contacted key authors.
Data collection and analysis
We extracted the following data from each study report: study design; recruitment and randomisation method; subject characteristics including age, gender, smoking behaviour, and exercise levels at entry; sample size; description of exercise and smoking cessation programmes (including number of sessions and duration); rates of exercise adherence; control conditions; length of follow‐up; definition of cessation; and method of validation. The primary outcome was quitting at longest follow‐up using the strictest definition of abstinence reported in the study. Data on change in cigarette consumption was also extracted.
Due to the differences in study design and intervention, we did not conduct a meta‐analysis. For each study the risk ratio for quitting at longest follow‐up ((number of events in intervention condition/intervention denominator)/ (number of events in control condition/control denominator)) and the 95% confidence interval were displayed graphically. Where the event is defined as smoking cessation, an RR greater than one indicates that more people successfully quit in the treatment group than in the control group.
Unless noted otherwise, quit rates are calculated based on numbers randomised to an intervention or control, and exclude any deaths or untraceable moves. We regard participants who dropped out or were lost to follow‐up as continuing to smoke. We have where possible conducted intention‐to‐treat analyses (i.e., all participants initially assigned to intervention or control are included in their original groups).
Included studies were assessed for risk of selection bias (random sequence generation and allocation concealment) and attrition bias (incomplete follow‐up) in accordance with Cochrane guidelines.
Results
Description of studies
The searches identified 20 studies which met the inclusion criteria, with a total of 5,870 participants, the largest study being an internet trial with 2,318 participants (McKay 2008). Eight trials had fewer than 30 individuals in each treatment arm (Bock 2012; Ciccolo 2011; Hill 1985; Hill 1993; Marcus 1991; Marcus 1995; Russell 1988; Taylor 1988). Five studies have been added since the last version of this review (Abrantes 2014; Bock 2012; Horn 2011; Maddison 2014; Whiteley 2012). Full details for each study are given in the Characteristics of included studies table. Ten studies had more than one associated publication or abstract (Bize 2010; Horn 2011; Kinnunen 2008; Maddison 2014; Marcus 1999; Marcus 2005; Martin 1997; Prapavessis 2007; Ussher 2003; Whiteley 2012) and these are listed under the study identifier in the reference section. Nine trials were limited to women (Bock 2012; Kinnunen 2008; Marcus 1991; Marcus 1995; Marcus 1999; Marcus 2005; Prapavessis 2007; Russell 1988; Whiteley 2012) and one was limited to men (Taylor 1988).
In all but two of the studies (McKay 2008; Taylor 1988), a multi‐session cognitive behavioural smoking cessation programme was provided for intervention and control conditions. In ten studies this was described as beginning prior to quit day (Abrantes 2014; Bock 2012; Hill 1993; Kinnunen 2008; Maddison 2014; Marcus 1999; Marcus 2005; Prapavessis 2007; Ussher 2003; Whiteley 2012). One study provided only a single session cessation programme and participants were post‐acute myocardial infarction (AMI) patients, with the intervention being for relapse prevention (Taylor 1988). One study delivered a smoking cessation programme via the Internet and this was only available for the non‐exercise condition (McKay 2008). Six studies included nicotine patches as part of the smoking cessation programme (Abrantes 2014; Ciccolo 2011; Marcus 2005; Prapavessis 2007; Ussher 2003; Whiteley 2012), one study used nicotine gum (Kinnunen 2008), three promoted nicotine replacement therapy in general (Bize 2010; Maddison 2014; McKay 2008), and one advocated smoking cessation medicines in general (Bock 2012).
Sixteen of the studies recruiting current smokers set a quit date, and one set a quit date for the non‐exercise condition but did not specify whether the exercise group set a quit date (McKay 2008). The exercise programme began before the quit date in 12 studies (Abrantes 2014; Bize 2010; Bock 2012; Hill 1993; Kinnunen 2008; Marcus 1991; Marcus 1995; Marcus 1999; Marcus 2005; Prapavessis 2007; Ussher 2003; Whiteley 2012), on the quit date in three (Ciccolo 2011; Hill 1985; Martin 1997), and after the quit date in three (Maddison 2014; Russell 1988; Taylor 1988). Two studies did not state the timing of the exercise programme relative to quit date (Horn 2011; McKay 2008). Two studies involved exercise programmes lasting for less than six weeks (Hill 1985; Martin 1997) and the length of one programme was not given (McKay 2008). Most of the trials employed supervised, group‐based cardiovascular‐type exercise supplemented by a home‐based programme. Five studies did not provide a home programme (Bock 2012; Ciccolo 2011; Marcus 1991; Marcus 1995; Marcus 1999), one study used only brief one‐to‐one counselling towards pursuing home‐based exercise (Ussher 2003), one focused on telephone‐based physical activity counselling (Maddison 2014), and one provided a web‐based program designed to encourage engagement in a personalized fitness program, although specific detail was not provided regarding the type of exercise promoted (McKay 2008). Ciccolo 2011 focused exclusively on an individual programme of resistance exercise (i.e. weight training) and Bock 2012 delivered a yoga intervention.
Excluded studies
The literature search revealed a number of trials which did not satisfy the inclusion criteria (see Characteristics of excluded studies table), but had exercise as an independent variable and smoking cessation behaviour as a dependent variable. These studies mainly fell into four categories:
(a) Multiple independent and dependent variables: a number of studies were identified in which exercise was one element in a multiple risk factor intervention, with smoking cessation behaviour as one of a number of health‐related outcomes. The specific effects of exercise on smoking cessation could not be determined due to possible interaction and confounding between the independent variables. For example, it is not possible to separate the effects on smoking cessation due to a change in diet versus a change in exercise.
(b) Multiple independent variables and a single dependent variable: these studies included multiple smoking cessation elements one of which was exercise. In these studies the specific effects of exercise on smoking abstinence were not addressed.
(c) Single independent variable and multiple dependent variables: in these studies exercise was encouraged without a smoking cessation programme, and changes in various health and behavioural indices including smoking cessation were examined. None of these studies found a significant effect on smoking abstinence for the active condition. However, as these studies did not record the number of smokers who were trying to stop, it is difficult to evaluate their success.
(d) Acute studies: These experimental studies assessed the acute impact of an exercise intervention on withdrawal symptoms and desire to smoke, mostly following temporary abstinence. Details and findings of these studies are shown in Appendix 1 and summarised in the Discussion.
(e) Did not meet other inclusion criteria: These studies either had a follow‐up of less than six months, did not include all smokers who wish to quit at the outset, did not include a non‐exercise control group or did not have smoking abstinence as an outcome. Studies which reported change in cigarette consumption are included in Appendix 2.
Risk of bias in included studies
Only one of the 20 studies was at low risk of bias across all domains (Figure 1). All other studies were rated at unclear risk in at least one domain due to insufficient detail in the study report. Eleven studies described the randomization method in detail (Abrantes 2014; Bize 2010; Bock 2012; Ciccolo 2011; Marcus 1999; Maddison 2014; Marcus 2005; McKay 2008; Prapavessis 2007; Ussher 2003; Whiteley 2012). The strictest measure of abstinence was continuous in eight studies, prolonged abstinence in two, point prevalence in eight, and was not specified in two.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Thirteen studies stated that those lost to follow‐up were counted as having relapsed to smoking (Abrantes 2014; Bize 2010; Bock 2012; Ciccolo 2011; Horn 2011; Hill 1985; Maddison 2014; Marcus 1991; Marcus 1999; Marcus 2005; McKay 2008; Ussher 2003; Whiteley 2012). Post‐randomization dropouts were excluded from the denominator in six studies (Bize 2010; Ciccolo 2011; Hill 1993; Kinnunen 2008; Prapavessis 2007; Taylor 1988). Two studies were judged to be at high risk of attrition bias as less than half of the participants were followed up (Marcus 2005; McKay 2008).
Effects of interventions
See: Summary of findings for the main comparison
We defined the efficacy of the intervention in terms of the risk ratio (RR) for quitting in the treatment group versus the controls. The RRs with 95% confidence intervals (CI) for cessation at longest follow‐up for each study are shown in Analysis 1.1. Four studies showed significantly higher abstinence rates in a physically active group versus a control group at end of treatment (Bock 2012; Marcus 1991; Marcus 1999; Martin 1997). One of these studies also showed a benefit for exercise versus control on abstinence at the three‐month follow‐up and a benefit for exercise of borderline significance at the 12‐month follow‐up point (Marcus 1999). The latter study showed a difference in abstinence rates for the exercise condition compared with the control of 11.9% versus 5.4% (p = 0.05, RR 2.19, 95% CI 0.97 to 4.96) at the 12‐month follow‐up. One study observed significantly higher cessation rates for a condition combining exercise with a cessation programme compared with a brief advice only control group at three and six month follow‐ups (at six months: p = 0.013, RR 2.81, 95% CI 1.44 to 5.49), but the differences were not significant when comparing the exercise condition with a cessation programme plus brief advice (Horn 2011). In this study, abstinence was not biochemically validated at six months. One study showed significantly higher abstinence rates for the exercise group versus a control group at the three‐month follow‐up but not at the end of treatment or 12‐month follow‐up (Marcus 2005). The latter study also found that those with higher levels of exercise adherence were significantly more likely to achieve smoking abstinence at the end of treatment. The other studies showed no significant effect for exercise on abstinence. Several of the studies showed a trend for higher rates of abstinence in the exercise condition compared with the controls (Abrantes 2014; Ciccolo 2011; Hill 1985; Kinnunen 2008; Marcus 1995; Prapavessis 2007; Whiteley 2012). Only seven studies had a sufficiently large sample size to have a good prospect of detecting a significant difference between the treatment and control conditions (Bize 2010; Maddison 2014; Marcus 1999; Marcus 2005; Martin 1997; McKay 2008; Ussher 2003). One of the studies did not provide separate abstinence data for the experimental and control groups, although it was reported that no significant difference was found between the groups (Russell 1988, not included in Analysis 1.1).
In addition to comparing the exercise condition with a control group, four of the studies examined the effectiveness of exercise versus nicotine replacement therapy (NRT) (Hill 1993; Kinnunen 2008; Martin 1997; Prapavessis 2007). In one study at end of treatment and at 12‐month follow‐up abstinence rates were significantly higher in the exercise‐plus‐patch group than in the exercise‐only group (Prapavessis 2007). The other studies observed no significant differences.
Discussion
The evidence on the effectiveness of exercise interventions for smoking cessation is very limited (summary of findings Table for the main comparison). Four of the 20 studies found an effect in favour of the intervention at end of treatment, one study detected a borderline significant difference in favour of the intervention at 12 months, and one study detected a significant difference at six months when evaluating an exercise programme as an adjunct to brief advice, but not when comparing the exercise programme to a more intensive smoking cessation programme. Key limitations to the body of evidence include clinical heterogeneity, unclear risk of bias, issues with generalisabilty to other population groups, and issues with study design and outcome measures. These are explored in more detail below.
Cessation programmes
In one study the effect of the treatment may have been compromised by the smoking cessation programme being limited to a single counselling session (Taylor 1988). In this study the interventions were not intended to initiate smoking abstinence but rather to maintain abstinence in smokers following acute myocardial infarction (AMI). Thus the results, which did not show any benefit for exercise, cannot easily be generalized beyond abstaining post‐AMI smokers. This trial also compared the combined effect on smoking abstinence of four different exercise interventions with the combined effect of two different control interventions; therefore it was not possible to relate outcomes for smoking cessation to specific interventions. This study is further limited by providing smoking cessation counselling for only one of the two control conditions.
The results of one of the studies, showing a positive effect for exercise on smoking abstinence at end of treatment, may have been confounded by the exercise group receiving a different cessation programme than the control group (Martin 1997). In two studies showing a significant benefit of exercise versus control at end of treatment the exercise condition received more staff contact time than the control (Marcus 1991; Martin 1997), leading to the question of whether the outcomes for abstinence were due to exercise alone or due to additional social support.
It has been recommended that a smoking cessation programme should start before the quit date and continue into the period of abstinence (Raw 1998). Yet almost half of the trials did not do this (Ciccolo 2011; Hill 1985; Maddison 2014; Marcus 1991; Marcus 1995; Martin 1997; McKay 2008; Russell 1988; Taylor 1988). With the provision of more extensive cessation programmes the impact of the interventions may have been more pronounced. Furthermore, only one of the studies (Ussher 2003) described an intervention in which the smoking cessation and exercise components were integrated in such a way as to reinforce exercise as a coping strategy for smoking cessation (Marlatt 1985; Taylor 2010). For example, the potential for exercise to be used to reduce cigarette cravings and withdrawal symptoms (Taylor 2007b) was not made explicit in the majority of studies.
Target populations
Demographic factors, such as age, gender, weight, fitness level, socioeconomic status and occupation could influence outcomes for both smoking cessation (Jarvis 1997; Vangeli 2011) and exercise behaviour (Caspersen 1994; Pate 1995). Of the nine trials which recruited men and women, four compared outcomes by gender. Of these studies, Abrantes 2014, Hill 1993, and Ussher 2003 reported no gender differences, and Horn 2011 observed significantly higher abstinence rates for males, but not for females, at three and six month follow‐ups, in a teenage population. None of the studies considered outcomes relative to occupation, socioeconomic status or age. It is possible that the relationship between demographic variables and outcomes was not explored in some of the studies because of small sample sizes. All but four of the studies were North American. Thirteen studies recorded ethnic status, and all reported a predominantly white sample. Researchers must consider whether these results can be generalized to other national and ethnic populations (Caspersen 1994; King 1997; Mackay 1996). One trial recruited post‐acute myocardial infarction (AMI) patients, another targeted teenage smokers (aged 14 to 19 years), and the remaining trials recruited from the general population of smokers. Trials are needed among other populations of smokers who might especially benefit from an exercise intervention. Given the high prevalence of smoking among those with mental illness, and the established benefits of regular physical activity for mental health (Stathopoulou 2006), research is needed to examine the role that physical activity may play as an aid to quitting. Those with serious mental illness are likely to be receptive to exercise as an aid to cessation (Arbour‐Nicitopoulos 2011; Arbour‐Nicitopoulos 2011b; Faulkner 2007) and an exercise intervention has been successfully piloted among women smokers with depression (Bernard 2013; Vickers 2009). Horn 2011 showed that teenage smokers are likely to benefit from an exercise intervention. An excluded study explored a sport‐based intervention for smoking prevention in pre‐adolescents (Trigwell 2014). Obese quitters may have a particular need for weight control interventions, such as exercise (Lycett 2011), and we have yet to see a trial of exercise focusing on this population. Additionally, a non‐pharmaceutical intervention such as exercise is likely to appeal to pregnant smokers (Ussher 2004; Ussher 2007) and a recently completed trial is assessing the effects of an exercise intervention in this population (Ussher 2008; Ussher 2012).
Seven studies did not present the participants' level of exercise at baseline (Abrantes 2014; Bock 2012; Ciccolo 2011; Hill 1985; McKay 2008; Russell 1988; Taylor 1988). All the remaining studies, except Horn 2011, reported that they had recruited fairly sedentary smokers. A substantial proportion of smokers may be physically active (deRuiter 2008; Emmons 1994; Prochaska 1992; Ward 2003; Scioli 2009) and there is some evidence that regular exercisers may be more successful at quitting (Abrantes 2009; Derby 1994; Paavola 2001; Sedgwick 1988), yet it is not clear whether exercise interventions are effective as an aid to smoking cessation for a more active population.
Weight gain
Marcus 1999 reported a significantly smaller weight gain for those in the exercise condition compared with the controls at the end of treatment; however, those in the exercise condition weighed more than the controls at baseline, and this difference was not controlled for, which makes interpretation of the finding problematic. Marcus 1999 did not find any significant differences in weight change between the treatment conditions at the three‐month or 12‐month follow‐ups. Prapavessis 2007 observed no difference in weight gain at end of treatment when comparing cognitive‐behavioural support plus nicotine patches with exercise plus nicotine patches; however, at end‐of‐treatment those in the exercise only condition gained significantly less weight than those receiving only cognitive‐behavioural support. Other studies found no difference in weight gain for the exercise versus controls at end of treatment (Horn 2011 (BMI); Marcus 1991; Marcus 1995; Marcus 2005; Ussher 2003; Whiteley 2012), at three‐ and six‐month follow‐ups (Ciccolo 2011), or at 12 months post‐cessation (Bize 2010; Ussher 2003). The studies by Ciccolo 2011, Marcus 1991, Marcus 1995, and Whiteley 2012 were too small to have a realistic chance of detecting significant differences. The studies by Bize 2010, Marcus 2005, Ussher 2003, and Whiteley 2012 included nicotine replacement therapy (NRT) and post‐cessation weight gain is likely to be less pronounced when using NRT (Jorenby 1996). Therefore, the potential for exercise to moderate weight gain was reduced. It is possible that exercise provides a role in weight management once an individual has stopped using NRT, but this has yet to be determined.
When pooling the studies, Farley 2012 found no evidence for exercise moderating weight gain at end of treatment, but reported a benefit at 12 months follow‐up when combining three studies (Bize 2010; Marcus 1999; Ussher 2003). The authors concluded that 'More studies are needed to clarify whether this is an effect of treatment or a chance finding'. An earlier publication conducted a meta‐analysis with 10 studies of weight management interventions during smoking cessation, including five of the studies included in the current review (Marcus 1991; Marcus 1995; Marcus 1999; Marcus 2005; Ussher 2003), and observed a significant benefit for the intervention in the short‐term (< three months), but not in the long‐term (> six months) (Spring 2009).
Nicotine replacement therapy
Prapavessis 2007 provides some indication that combining nicotine patches and exercise enhances abstinence compared with exercise alone, as would be expected given the established efficacy of NRT (Stead 2012). Future studies need to establish whether exercise offers additional benefits to those provided by NRT, and other smoking cessation medications, alone. It is feasible that exercise could address psychosocial and physical needs that are not currently met by NRT‐based programmes.
Exercise programmes
For those beginning exercise either on or after the quit date (Ciccolo 2011; Hill 1985; Maddison 2014; Martin 1997; Russell 1988) success rates may have been hampered by the demand to cope simultaneously with two major changes in health behaviour (Emmons 1994; King 1996; Patten 2001). In studies where the exercise programme started after a period of smoking abstinence the potential for exercise to moderate withdrawal symptoms during this period was lost (Taylor 2007b). In practice, when the exercise programme begins may depend on individual capabilities and preferences (Everson‐Hock 2010b).
In the two studies with exercise programmes lasting for less than six weeks (Hill 1985; Martin 1997), the intervention may have been of insufficient length to encourage long‐term exercise adherence. Most of the trials employed supervised, group‐based exercise supplemented by a home‐based programme. Where home programmes were not provided (Bock 2012; Ciccolo 2011; Marcus 1991; Marcus 1995; Marcus 1999) it is possible that the participants' high level of dependence on supervised exercise reduced their level of post‐intervention activity. An excluded study compared facility and home‐based exercise and found no group differences in smoking abstinence at 12 months (Kinnunen 2013).
Those adequately powered trials not showing a consistent effect of exercise on smoking abstinence (Bize 2010; Maddison 2014; Marcus 2005; McKay 2008; Ussher 2003) had interventions of a low intensity, in that they promoted moderate intensity rather than vigorous intensity exercise. In one case they relied solely on fairly brief exercise counselling (Ussher 2003), another focused on telephone‐based counselling (Maddison 2014), in two other studies supervised exercise was only provided once per week (Bize 2010; Marcus 2005), and the remaining study relied on a web‐based programme (McKay 2008). In these studies the exercise intervention may have been insufficiently intense to benefit smoking abstinence. Further studies are required to establish the optimum intensity of exercise intervention required as an aid to smoking cessation. Intensity here refers to both the exercise intensity per se (i.e. light, moderate or vigorous) and the extensiveness of the support being provided (e.g. number of supervised exercise sessions). The findings from Marcus 2005 suggest that abstaining smokers may need to accumulate at least 110 minutes of activity per week to maintain abstinence (at least during the intervention period), and supervised exercise on two or three days a week may be necessary to achieve this. A recent pilot study showed promising findings for an intervention involving moderate intensity exercise supervised on three days a week over eight weeks (Williams 2010) and this needs to be tested in a larger trial.
Only three of the studies provided any post‐intervention exercise programming (Hill 1993; Maddison 2014; Ussher 2003), and this may have reduced post‐intervention exercise adherence (King 1989). However, it is not possible to draw any conclusions about whether various aspects of the intervention affected levels of exercise adherence after the formal supervised programme ended because none of the studies reported rates of adherence for this period.
Ciccolo 2011 promoted resistance exercise, Whiteley 2012 covered both aerobic and resistance exercise, and yoga classes were provided by Bock 2012. The remaining studies focused on cardiovascular‐type exercise. More studies are required with non‐cardiovascular exercise. For example, isometric exercise has been shown to reduce tobacco cravings and urges to smoke (Ussher 2006; Ussher 2009), and has been successfully piloted (Al‐Chalabi 2008).
Exercise adherence issues
During the treatment period a range of cognitive‐behavioural methods were employed to improve adherence to the exercise programme. All but seven of the studies used group‐based exercise (Ciccolo 2011; Horn 2011; Kinnunen 2008; Maddison 2014; McKay 2008; Ussher 2003; Whiteley 2012). Only five studies did not provide full supervision of facility‐based exercise (Horn 2011; Kinnunen 2008; Maddison 2014; McKay 2008; Ussher 2003). All the studies included goal setting, nine used self‐monitoring (Abrantes 2014; Horn 2011; Hill 1985; Kinnunen 2008; Maddison 2014; Martin 1997; Russell 1988; Taylor 1988; Whiteley 2012), one used reinforcement (Martin 1997), Hill 1993 used telephone follow‐up in the case of non‐attendance, Taylor 1988 used remote monitoring of heart rate, and two studies provided pedometers (Horn 2011; Maddison 2014). Two trials employed exercise counselling, including a broad range of cognitive‐behavioural techniques (Maddison 2014; Ussher 2003).
Four studies did not report overall activity levels for the treatment group during the treatment period (Bock 2012; Ciccolo 2011; Hill 1993; McKay 2008). Where supervised exercise was offered, attendance at these sessions was high. Where the emphasis was on home‐based exercise (Bize 2010; Horn 2011; Maddison 2014; Marcus 2005; McKay 2008; Ussher 2003) only a minority of the participants achieved the criterion level of exercise. For example, in one study combining home‐based exercise with one supervised session of exercise per week, 50% of those in the exercise group were still classed as sedentary at the end of treatment (Bize 2010). One study reported greater attrition for the exercise group compared with the controls (Marcus 1999 ‐ see Borrelli 2002). Two studies reported lower attendance for the exercise intervention compared with the health education programme (Kinnunen 2008; Whiteley 2012). The one internet‐based trial observed very similar levels of physical activity for the two groups at the six‐month follow‐up (McKay 2008). Future studies need to consider other methods for increasing 'home‐based' physical activity. For example, pedometers have been used to increase participation in a walking‐based intervention during smoking cessation (Prochaska 2008). However, when using a pedometer and counselling‐based intervention Horn 2011 observed no effect on physical activity levels in teens, while Maddison 2014 used a similar intervention and found a significant increase in leisure time physical activity for the exercise group versus control group, but no effect on overall physical activity.
Fitness measures
Although many of the studies reported fitness measures for the control group during the treatment period (Abrantes 2014; Ciccolo 2011; Hill 1985; Kinnunen 2008; Marcus 1991; Marcus 1995; Marcus 1999; Prapavessis 2007; Russell 1988; Taylor 1988; Whiteley 2012), less than half of them reported physical activity (PA) levels for the control group at this time (Abrantes 2014; Bize 2010; Hill 1985; Horn 2011; Kinnunen 2008; Maddison 2014; Ussher 2003; Whiteley 2012). Therefore, in the majority of the studies the relative increase in PA in the treatment group versus any spontaneous increase in activity in the control group could not be accurately monitored. During the follow‐up period none of the studies described using cognitive‐behavioural techniques to encourage regular exercise, although Maddison 2014 offered physical activity consultations up until the end of a six month intervention. Only three of the studies assessed fitness during the follow‐up period (Ciccolo 2011; Prapavessis 2007; Russell 1988) and only two studies reported levels of activity at 12‐month follow‐up (Bize 2010; Ussher 2003). Therefore, for the vast majority of studies it was not possible to relate long‐term smoking abstinence to exercise behaviour.
Fitness measures are useful as a confirmation of exercise adherence. However, the significance of changes in fitness in the context of smoking cessation is debatable. Since exercise has been shown to benefit psychological and general health without increases in fitness (Pate 1995; Taylor 2008) it is possible that exercise could aid smoking cessation independently of changes in physical capacity. A number of the trials reported a significant increase in fitness levels at the end of the treatment period within the active exercise condition (Marcus 1991; Marcus 1995; Marcus 1999 (see also Albrecht 1998); Marcus 2005; Prapavessis 2007). Four studies showed an increase in fitness for the intervention conditions compared with the controls at end of treatment (Marcus 1999; Prapavessis 2007; Taylor 1988; Whiteley 2012); others showed no differences at end of treatment (Abrantes 2014; Kinnunen 2008), at a four‐month follow‐up (Russell 1988), or a 12‐month follow‐up (Prapavessis 2007).
Psychological measures
The majority of the studies used psychological measures at baseline and eleven trials reported changes in these measures (Abrantes 2014; Bock 2012; Kinnunen 2008; Maddison 2014; Marcus 1999; Marcus 2005; Martin 1997; Prapavessis 2007; Russell 1988; Ussher 2003; Whiteley 2012). Russell 1988 found a significant increase in Profile of Mood States (POMs) tension‐anxiety scores for the active group compared with the controls at four months follow‐up. These findings are not consistent with the general consensus that exercise reduces mood disturbance, stress, and anxiety (Stathopoulou 2006; Taylor 2000; Taylor 2008). The reported effect on psychological outcomes may have been caused by extraneous variables which could not be controlled for with a small sample size. Martin 1997 found no significant treatment differences on mood (POMs) or depression (Beck Depression Inventory) when comparing measures taken at baseline and seven days post‐quit; these findings may have been influenced by the sample including a large number of individuals with a history of major depression. Prapavessis 2007 showed that reports of self efficacy for stopping smoking were higher in a cognitive‐behavioural support condition compared with an exercise‐only condition. Marcus 1999 did not find a significant change in reports of tobacco withdrawal symptoms and cigarette cravings for exercise versus controls across the treatment period. Kinnunen 2008 did not find any difference in reports of withdrawal symptoms for the exercise group versus the controls at one week post‐cessation. Bize 2010 found no significant differences in reports of withdrawal symptoms, depression, urges to smoke, or perceived stress for the exercise group versus the control group. Marcus 2005 observed that, among 40 women who were abstinent at the end of treatment, those who increased their fitness were more likely to report decreases in depressive symptoms (see Williams 2008). Ussher 2003 observed a reduction in some withdrawal symptoms for exercise versus controls up to three weeks post‐cessation. Bock 2012 observed no effects of a yoga intervention on anxiety, depression, or temptations to smoke. Maddison 2014 found no group differences in tobacco withdrawal symptoms. Abrantes 2014 reported significantly lower somatic withdrawal symptoms and sleep disturbance for the exercise versus control group; there were no group differences for craving, mood disturbance, or positive affect. Abrantes 2014 was the only study which examined the effect of exercise on sleep disturbance, and this may be a worthwhile objective. For example, Grove 2006 observed that, compared with controls, regular participation in exercise, during the period of tobacco withdrawal, did not affect the ability to stay asleep but exercisers reported significantly less difficulty falling asleep. It would also be valuable if affective changes after exercise were assessed among different subgroups of smokers. For example, one study observed that, among women smokers with increased concern about weight gain, engagement in exercise was associated with less of an increase in negative affect following smoking cessation (Schneider 2007).
Acute effect of exercise on tobacco withdrawal and cravings
Appendix 1 presents a summary of 41 studies we identified which have assessed the acute effects of exercise on smoking outcomes. Since the previous version of this review a further 14 studies have been added. Three studies focused on outcomes related to smoking intake (Kurti 2014; Mikhail 1983; Reeser 1983). The remaining 38 studies included outcomes related to tobacco withdrawal/mood and/or tobacco cravings. Six studies assessed outcomes during an attempt to quit smoking (Abrantes 2014; Arbour‐Nicitopoulos 2011; Bock 1999; Harper 2012; Harper 2013; Williams 2011). Of these studies, compared with a passive control group, two reported a significant reduction in mood related withdrawal symptoms and cigarette cravings (Abrantes 2014; Bock 1999), and one showed that exercise increased energy and reduced tiredness but had no effect on cravings (Williams 2011). In addition, two single‐group studies observed a significant post‐exercise reduction in withdrawal and cravings (Harper 2012; Harper 2013). Just one of the studies involving a quit attempt found no effect of exercise on cravings or withdrawal symptoms and this was among those with serious mental illness (Arbour‐Nicitopoulos 2011). We found 32 studies that examined the acute effects of exercise on withdrawal symptoms and/or cravings among temporarily abstinent smokers and all but six of these studies (Daley 2004; Daniel 2007; Everson 2006; Faulkner 2010; Oh 2014; Pomerleau 1987) observed a significant reduction in cravings and/or withdrawal symptoms compared with a passive control. One of these studies, for the first time, demonstrated a reduction in cravings, for an exercise versus passive group, among pregnant smokers (Prapavessis 2014). Previous reviews of 14 or 15 studies (Haasova 2013; Roberts 2012; Taylor 2007b) provide a more detailed discussion but this section highlights some findings from more recent studies which have shown an acute benefit of exercise during temporary smoking abstinence. These studies showed that, compared with a passive condition, after periods of up to 24 hours without smoking, smokers have lower cravings, withdrawal symptoms and negative affect during and for up to 30 minutes post‐exercise. The effects are evident for moderate and vigorous intensity exercise, for Hatha Yoga, and for durations from five minutes of seated isometric exercise to 20 to 30 minutes of cardiovascular activity. Encouragingly, findings suggest relatively convenient forms of physical activity (e.g. 10 to 15 minutes of brisk walking) can be effective.
Haasova 2013 and Roberts 2012 have quantified the effects of a single bout of exercise on reducing strength of desire to smoke using 15 studies; the pooled estimates for treatment effect (non‐standardised mean difference) were ‐1.91 and ‐2.41, respectively, with a high degree of between‐study heterogeneity. There has been a tendency for studies with shorter bouts of exercise to show a less sustained effect on reducing cravings and withdrawal and further research is needed to understand how the dose of exercise impacts on the duration of acute effects. However, even brief bouts of exercise, with a brief effect, may be useful to cope with a temporary spike in cravings.
Several mechanisms have been tested among these studies for how exercise reduces cravings. Distraction (Daniel 2006) does not appear to explain the effects. Exercise expectancy was modestly associated with psychological symptoms, but not with cigarette cravings, in one study (Harper 2013) and did not explain any of the effects in another study (Daniel 2007). In two studies changes in cortisol concentration were unrelated to changes in cravings (Janse Van Rensburg 2013; Scerbo 2010). This suggests that cortisol changes do not mediate the effects of exercise on cravings. Taylor 2006a reported that reductions in urges to smoke in response to exercise were mediated by reductions in tension. Three studies involving functional Magnetic Resonance Imagery (fMRI) scanning suggested that parts of the brain that are typically activated by smoking cues (images) were less activated following moderate intensity exercise (Janse van Rensburg 2009b; Janse van Rensburg 2010; Janse Van Rensburg 2012). Finally, two studies (Janse van Rensburg 2009a; Oh 2014) reported that after exercise, compared with rest, abstinent smokers had less attentional bias (gaze or dwell time, measured using eye‐tracker technology) towards smoking images, compared with neutral images presented simultaneously. Shifts in attentional bias away from smoking‐related cues, after exercise, are in line with other studies in which participants report improvements in concentration (as a withdrawal symptom) after exercise (e.g. Daniel 2006; Ussher 2001; Ussher 2006). Further work is needed to understand how different types of exercise (e.g. isometric, resistance, cardiovascular) influence symptoms known to cause relapse among actual quitters, and among those using pharmaceutical aids to cessation, in which case symptoms may be lower at the outset. In addition to studies focusing on self‐reported cravings, seven studies (Faulkner 2010; Katomeri 2007; Kurti 2014; Mikhail 1983; Reeser 1983; Taylor 2007a; Thayer 1993;) reported that a bout of exercise delayed ad libitum smoking, or favourably influenced smoking topography, although three other studies observed no significant effect on ad libitum smoking or smoking topography (De Jesus 2014; Fong 2014; Schneider 2014). Overall, given this experimental evidence further research is needed to understand how best to promote the use of acute bouts of physical activity, in contrast to longer scheduled bouts of exercise, as a momentary aid to smoking cessation.
Effects of exercise interventions on smoking reduction
In this update of the review, for the first time, we include a review of studies that assessed the effect of a PA intervention on levels of cigarette consumption. This is the first review we are aware of on this topic. This issue is important as many smokers say that they wish to reduce cigarette consumption before quitting, reduction for those who want to quit has been shown to be as effective for smoking cessation as abrupt approaches (Lindson 2010), and reduction approaches are recommended in national guidelines (NICE 2013). The NICE 2013 review that informed the guidelines did not explicitly search for ‘exercise’ studies that may have aided reduction, and none were reported. This section of the review includes any studies, irrespective of study design, which report the effects of a PA intervention on rates of cigarette consumption. The ultimate aim of the majority of these studies was quitting; therefore, we also report outcomes for attempts to quit and for smoking abstinence and state whether these were primary or secondary outcomes. A table of studies assessing the effect of physical activity on levels of cigarette consumption is presented in Appendix 2. Our searches identified 14 studies and five of these are also included in the main review of smoking cessation studies (Hill 1985; Taylor 1988; Horn 2011 (smoking reduction reported in Horn 2013); Prapavessis 2007; Maddison 2014). Two studies were published in the 1980s and the remainder were reported since 2007.
In the majority of studies smoking cessation was defined as the primary outcome and smoking reduction as a secondary outcome (Hill 1985; Horn 2013; Maddison 2014; Prapavessis 2007; Taylor 1988; Taylor 2014; Ussher 2012; Whiteley 2007; Ybarra 2013). Two studies defined smoking reduction as the sole primary outcome (Bernard 2013b; Leelarungrayub 2010). Changes in PA levels were primary and smoking levels were secondary in one study (Kovelis 2012). McClure 2011 defined mental and physical well‐being as the main outcome and cigarette consumption as secondary. Finally, one study identified both smoking abstinence and cigarette consumption as primary outcomes (Gorini 2012).
All ten studies with smoking cessation as an outcome provided a smoking cessation intervention; in addition, McClure 2011 provided a smoking cessation intervention but did not assess rates of smoking abstinence. In most of these cessation studies the participants wished to quit smoking immediately and smoking reduction was measured ‘incidentally’ without a specific goal to reduce. In one study the participants wished to reduce but not quit immediately; PA was used to assist reduction and inducing quit attempts was a research goal but this was not an explicit goal for the participants (Taylor 2014). In one case participants were seriously thinking about quitting in the next 30 days and exercise was used to prepare for and aid a quit attempt (Ybarra 2013). McClure 2011 reported that 71% of participants were in the preparation or contemplation stage of change for quitting smoking, and the level of intention to quit was not stated in one study (Taylor 1988). Among the three non‐smoking cessation studies, in one study participants’ intention to quit was not stated and the aim was solely to reduce smoking levels as a harm reduction strategy among individuals with schizophrenia (Bernard 2013b) and in the studies by Leelarungrayub 2010 and Kovelis 2012 intention to quit was not reported, nor were participants advised to reduce their smoking (i.e. smoking reduction was assessed incidentally following a PA programme). Only one study encouraged specific strategies for smoking reduction, and in this case four strategies (hierarchical reduction, smoke free periods, scheduled reduction, planned reduction) were offered to smokers (Taylor 2014).
Twelve of the 14 studies were RCTs and the remaining two had a single group within‐subjects design (Bernard 2013b; Whiteley 2007). Seven had a sample of less than 60 participants (Bernard 2013b; Hill 1985; Horn 2013; Kovelis 2012; McClure 2011; Taylor 1988; Whiteley 2007). No studies reported power calculations related to examining the effect on smoking reduction, and among the smoking cessation studies only three had large enough samples to have a realistic chance of detecting an effect on smoking cessation (Gorini 2012; Maddison 2014; Ussher 2012). Five studies were identified as pilot or feasibility studies (Bernard 2013b; McClure 2011; Taylor 2014; Whiteley 2007; Ybarra 2013).
The large majority of studies were based in North America. Four studies targeted more sedentary smokers (Maddison 2014; McClure 2011; Prapavessis 2007; Whiteley 2007), and all but two studies (Taylor 1988; Horn 2013) reported that participants were smoking a mean or median of at least 10 cigarettes a day at baseline. Most studies targeted smokers in general, while five targeted the following specific populations: post‐acute myocardial infarction (Taylor 1988); diagnosis of depression (McClure 2011); diagnosis of schizophrenia/schizoaffective disorder (Bernard 2013b); teenagers (Horn 2013); and pregnant smokers (Ussher 2012). Males and females were recruited in most studies, one trial only included men (Taylor 1988), and four were women only trials (Gorini 2012; Prapavessis 2007; Ussher 2012; Whiteley 2007). Half the studies provided supervised exercise (i.e. under the guidance of an instructor) (Bernard 2013b; Hill 1985; Leelarungrayub 2010;Prapavessis 2007; Taylor 1988; Whiteley 2007; Ussher 2012), four provided only PA counselling (Gorini 2012; Horn 2013; Maddison 2014; Taylor 2014), two provided just a pedometer‐based programme (Kovelis 2012; McClure 2011), and one was based solely on a text message‐based PA programme (Ybarra 2013). All the studies focused on cardiovascular type exercise (e.g. brisk walking). Four of the studies did not report levels of adherence to the PA intervention (Bernard 2013b; Hill 1985; Leelarungrayub 2010; Ybarra 2013). In the studies reporting PA adherence, three observed no significant difference in PA levels for the PA versus control group (Gorini 2012; Horn 2013; McClure 2011). Some studies included a multi‐component intervention targeting behaviours other than smoking and PA (McClure 2011; Ybarra 2013). Only three studies reinforced PA as a method for reducing cigarette consumption or increasing cessation (e.g. through promoting exercise as a means for coping with cigarette cravings, withdrawal symptoms, and weight gain) (Bernard 2013b; Taylor 2014; Ussher 2012).
Of the 14 studies identified, just one reported a significant smoking reduction for the PA versus control group, at 16 weeks post‐baseline (Taylor 2014). Two further studies observed significantly lower absolute levels of smoking for a PA group versus control at 23 week post‐treatment (Taylor 1988) and 24 weeks after quit day (Maddison 2014), although they did not analyse the changes in smoking relative to baseline. One study reported a significant smoking reduction in the PA group but not in the control group (group differences were not analysed) (Leelarungrayub 2010), two studies with a single PA group reported a significant smoking reduction (Bernard 2013b; Whiteley 2007), one trial reported significantly lower smoking levels for a cessation programme versus a PA programme (Prapavessis 2007), and one study did not report the group effect but observed that those with higher rates of PA adherence were more likely to reduce their cigarette consumption (Horn 2013).
Among the remaining five studies, all not reporting any significant smoking reductions, there were notable methodological challenges which hampered their chances of detecting any effects: two had samples of less than 60 participants (Hill 1985; McClure 2011); two had PA interventions may have been insufficiently intense to significantly raise PA levels (Gorini 2012; Ybarra 2013) (brief PA counselling alone, text message support alone, respectively); and one targeted pregnant smokers at around 16 weeks gestation, the majority of which had already markedly reduced their smoking levels relative to before their pregnancy (Ussher 2012). All these five studies that did not report a significant effect did report a trend such that the smoking reduction tended to be higher for the PA group versus the control group.
In conclusion, we observed that only one of 14 studies observed a significant smoking reduction for the PA versus control group (Taylor 2014). Notably, this study also reported that, compared with the control group, a significantly greater proportion of those in the PA group made a quit attempt. The remaining studies either did not report reduction levels for a PA versus control group or had methodological challenges which limited their chances of detecting a significant difference. Taylor 2014 was by far the most rigorous study as it was an RCT, with a sufficiently large sample size to have a realistic chance of detecting a significant effect on short‐term reduction outcomes, recruiting men and women who wished to reduce but not quit in the next month. Reduction strategies were incorporated, exercise was promoted as a reduction strategy, the intervention was sufficient to result in a significant increase in PA levels for the PA versus control group, follow‐up support was offered and, as well as smoking reduction, quit attempts and smoking cessation were assessed. However, this was a feasibility study which only followed‐up participants to 16 weeks post‐baseline. Larger trials with methods comparable to those used by Taylor 2014 are needed with a follow‐up of at least six months.
Overall commentary
A comparison of the studies was complicated by differences in study design and intervention, and by the relative paucity of rigorous research in this field. There were marked variations between the studies in the length, type, and timing of the exercise intervention, in the design of the control condition and cessation programme, and in the demographic factors recorded. In addition, there was often a lack of data relating to the physical activity levels of the control groups, and of either group during the follow‐up period. Together, these factors restricted meaningful comparison of results between studies. The findings presented in this review have implications for future research in this field. One of the first requirements for future work must be to have trials with larger sample sizes.
It is possible that a greater integration between the smoking cessation and exercise programmes may have enhanced abstinence rates (Taylor 2010). In future research exercise could be presented more as a self‐control strategy as well as as a means of increasing fitness and general health and of managing body weight (Marlatt 1985). For example, in initiating abstinence, exercise could be presented as a strategy for managing withdrawal symptoms and overcoming physical dependency (Taylor 2007b). The evidence we have reviewed consistently demonstrates the benefits of an acute bout of exercise on alleviating cravings and withdrawal symptoms under optimum conditions for observing such an effect (i.e. with experimentally manipulated increased baseline cravings – through temporary abstinence, and in some cases in the presence of smoking related cues, prior to exercising). As regards relapse prevention, exercise could be presented as a strategy which increases self esteem and pride in one's health, and reinforces an identity as a non‐smoker and as a physically active person (Verkooijen 2008) in such a way that being a smoker is incompatible with these perceptions (Fox 1998). Critically, it is likely that exercise needs to be maintained for it to continue to aid smoking cessation. An ongoing trial is assessing the effectiveness of a home and community‐based lifestyle exercise maintenance intervention in assisting women to maintain exercise following the termination of an exercise aided smoking cessation program, and hence reduce smoking relapse (Fitzgeorge 2011; Jung 2010).
At what point should the smoker who is trying to quit begin an exercise programme? In the studies reviewed, there was wide variation in the timing of the exercise programme. Some recommendations for changes in exercise and smoking behaviour are for sequential rather than simultaneous changes but this is likely to be specific to the individual's needs (Emmons 1994; Everson 2008b; King 1996; McEwen 2006). Another study showed a tendency for higher quit rates among those increasing exercise simultaneously rather than sequentially (Hyman 2007). It has been argued that a physical activity programme should begin prior to quitting, thereby allowing individuals to adjust to the demands of being more active before significantly changing their smoking behaviour (Marcus 1995). Elsewhere, it has been shown that abstaining smokers are more confident about adopting exercise than those preparing to quit (King 1996), which would support beginning an exercise programme when already abstinent, although delaying the start of the programme reduces the potential for managing withdrawal symptoms (Taylor 2007b). A quasi‐experimental study has reported higher adherence rates for smokers who undergo an exercise regimen commencing eight weeks before the quit day compared with those starting exercise on the quit day (Patten 2001). Further empirical work is required in order to ascertain the relative benefits of initiating exercise at different points in the cessation schedule. All of the 20 studies in the main review only included smokers who wished to attempt to quit smoking and to do so 'abruptly'; a recent feasibility study has shown the potential for exercise being used to increase quit attempts, through first gradually reducing their smoking intake, among those who are not motivated to initiate such an attempt (Taylor 2014) and further studies are needed in this area (see above section 'Effect of exercise interventions on smoking reduction'). Certainly, the majority of the studies we have reviewed did not demonstrate a significant smoking reduction among individuals in a physical activity treatment arm.
Only one study with balanced contact time showed a long‐term effect of exercise on smoking cessation (Marcus 1999). This study combined a vigorous intensity, thrice weekly supervised exercise programme with cognitive‐behavioural support. It has yet to be determined whether a less intensive exercise intervention can aid smoking cessation. Additionally, among teenagers, Horn 2011 reported significantly higher abstinence rates at six‐month follow‐up for a combined exercise and smoking cessation programme compared with brief smoking cessation advice. Further studies are needed to investigate the role of exercise for smoking cessation amongst young smokers. Finally, there is no evidence of harm in promoting physical activity to smokers. That is, no studies report reduced smoking cessation rates in an exercise group compared with control conditions and exercise has many benefits as a harm reduction strategy for smokers (deRuiter 2006).
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Exercise component versus smoking cessation programme only, Outcome 1 Smoking cessation at longest follow‐up.
| Exercise interventions for smoking cessation | |||
| Population: People who smoke or people who have recently quit smoking Intervention: Exercise programmes alone or as adjuncts to smoking cessation programmes Comparison: Smoking cessation programmes without exercise components | |||
| Outcomes | Effects of exercise interventions for smoking cessation | No of Participants | Quality of the evidence |
| Smoking cessation at longest follow‐up (6+ months) | At longest follow‐up, one study detected a difference of borderline significance in favour of the intervention group. Another study reported significantly higher abstinence rates at six month follow‐up for a combined exercise and smoking cessation programme compared with brief smoking cessation advice, but not when compared to the full smoking cessation programme. No other studies detected a significant difference between intervention and comparison groups at longest follow‐up. | 5870 (20 studies) | ⊕⊝⊝⊝ very low1 |
| GRADE Working Group grades of evidence | |||
| 1 Risk of bias unclear or high for all but one included study. High level of clinical heterogeneity precluded meta‐analysis. Issues with inadequate sample size in majority of included studies. | |||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Smoking cessation at longest follow‐up Show forest plot | 19 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
