Our meta-analysis was designed to estimate the efficacy of various behavioural therapies and pharmacotherapies for smoking cessation in cardiac patients. We found that both behavioural and pharmacotherapy treatments are efficacious in cardiac patients. However, the magnitude of the effect was small for such a high-risk group.
Although behavioural therapies as a group are superior to usual care, there are insufficient data to draw conclusions regarding the optimal length, duration and type of behavioural therapy to administer. The intensity of the behavioural intervention applied in each of the studies varied widely. Furthermore, RCTs also tested a broad spectrum of behavioural therapies – some studies tested smoking cessation advice and others tested multiple individual or group counselling sessions. Consequently, further studies are required before we can develop guidelines for smoking cessation in cardiac patients. In particular, large, multicentre, head-to-head RCTs are required to identify which types of behavioural therapies are most efficacious in cardiac patients. Additional RCTs are also required to identify the optimal length and duration of each type of behavioural intervention.
Additional RCTs are also needed to examine the effect of combination therapy involving both behavioural therapy and pharmacotherapy. Combination therapy has been shown to improve abstinence rates in the general population of smokers (33
) and published guidelines now recommend combining multiple individual or group counselling sessions with NRT (35
). Combination therapy may prove to be more efficacious than either behavioural therapy or pharmacotherapy alone in cardiac patients. However, we identified only one pharmacotherapy RCT in cardiac patients that included a combined behavioural component. However, both the active and placebo arms received the same behavioural intervention and thus, the effect of pharmacotherapy alone in comparison with combination therapy could not be examined (36
We identified only four RCTs examining the efficacy of pharmacotherapy in cardiac patients, and we could not identify any previous meta-analyses that examined the efficacy of smoking cessation pharmacotherapies in cardiac patients. The paucity of research in this area may relate to several factors. First, physicians may be reluctant to enroll cardiac patients in pharmacotherapy RCTs due to safety concerns. However, the two safety trials conducted to date suggest that smoking cessation pharmacotherapy use has a similar safety profile in cardiac patients as that observed in the general population (30
). Second, researchers may believe that there is no need to replicate studies performed among general populations in cardiac populations. However, cardiac patients may have different safety profiles than the general population and are at a high risk of cardiac events if they continue to smoke. Furthermore, they often have different motivations to quit smoking than the general population. Receiving a cardiac diagnosis is thought to be a ‘teachable moment’, a naturally occurring health event thought to motivate individuals to spontaneously adopt risk-reducing health behaviours (37
). Evidence of this phenomenon has been found in a number of studies (10
). For example, in the Framingham Heart Study (38
), men were 1.9 times more likely to quit smoking than the general smoking population following the development of CAD. In another study, Wilkes and Evans (10
) found that patients with chronic disease, including heart disease, expressed a greater desire (45% versus 30%) and need for assistance (38% versus 23%) to quit smoking than age-matched controls in the general population. Finally, the magnitude of the effect sought in this population must be greater given that they are at such high immediate risk. Thus, there remains an important need to examine the safety and efficacy of smoking cessation therapies in this patient population.
Several potential limitations of our meta-analysis should be noted. First, there were substantial methodological variations in the RCTs included in the present meta-analysis. RCTs varied in their definitions of behavioural therapy, usual care, patient characteristics, and the intensity and duration of therapy. Nevertheless, we considered them to be similar enough to be pooled. Furthermore, we used random effects models rather than fixed effects ones. Thus, our models incorporate both between-study and within-study variability, and account for heterogeneity. Second, we identified only four pharmacotherapy studies in cardiac patients. Thus, the estimates produced in our meta-analysis of pharmacotherapies have wide CIs. Third, the four pharmacotherapy RCTs examined three different medications. Consequently, heterogeneity was present and meta-analysis, even via random effects, may not have been fully appropriate. Nevertheless, it represents the totality of available evidence for pharmacotherapies to date in this specific patient population. Fourth, there were insufficient data available to fully examine the safety of these therapies in cardiac patients. Finally, as is true for all systematic reviews and meta-analyses, the results of the present study are also limited by the possibility of publication bias, particularly among the behavioural therapy RCTs. The two largest behavioural therapy RCTs were not statistically significant, while the results of the smallest RCTs were significant. These findings support the theory that studies with null results are less likely to be published than those with significant results (ie, publication bias) (39
). This is particularly true for smaller studies. However, our meta-analysis produced relatively strong ORs for efficacy and publication bias would have to be quite strong to overturn these results. Thus, it is highly unlikely that our results are due to the effects of publication bias.