Smoking cessation in primary care

Abstract

Evidence does not support routine use of combination therapy with nortriptyline

In the accompanying paper, Aveyard and colleagues report a pragmatic randomised controlled trial (RCT) of nortriptyline for smoking cessation.¹ The study randomised 901 smokers to a standard regimen of nortriptyline or placebo, and all participants were given the option of using nicotine replacement according to their preference (“pragmatic therapy”). The National Health Service stop smoking service provided group support in seven weekly sessions. The primary end point of the study was prolonged abstinence at six months. Nortriptyline plus nicotine replacement showed a modest but non-significant effect compared with placebo plus nicotine replacement at six months (relative risk 1.4; 95% confidence interval 1.00 to 1.98). This effect size is similar to that reported in a Cochrane systematic review and meta-analysis of two RCTs of nortriptyline for smoking cessation (n=318; odds ratio 1.48; 0.87 to 2.54).^{2 3 4} The results of the meta-analysis were also not significant, and significant heterogeneity occurred between studies.

Some of the major strengths of this research are its apparent methodological quality and large sample size, and it provides a substantial amount of new material for future meta-analyses of nortriptyline for smoking cessation. In particular, the sample size for the meta-analysis of combination nortriptyline will have increased at least fourfold compared with what has been published to date. These extra data should reduce statistical heterogeneity and provide greater statistical power to detect a modest effect of treatment if it exists, given the similarity between the effect size estimate of the present study and the existing pooled estimate from the most recent Cochrane review. This similarity supports the quality of the present study and suggests that the effect size estimate is close to the true value.

The study also highlights how little research has been published on combination treatments for smoking cessation. As the authors note, Cochrane systematic reviews and meta-analyses^{4 5} have identified only two RCTs of combination nortriptyline,^{2 3} and data indicating greater efficacy of nicotine replacement plus bupropion compared with nicotine replacement alone are derived from just one RCT.⁶ Furthermore, few published studies have compared the benefit of combining two or more types of nicotine replacement therapy with individual treatments,⁷ and we are not aware of any published trials of varenicline plus other pharmacotherapies for smoking cessation.⁸ In fact, evidence based guidelines from the World Health Organization,⁹ the NHS,¹⁰ and the US Public Health Service (USPHS)⁷ do not recommend combination bupropion therapy, and only the USPHS guideline explicitly recommends combination nicotine replacement therapy.⁷

Aveyard and colleagues are correct in concluding that on the basis of these results nortriptyline should not be routinely added to nicotine replacement therapy for smoking cessation. Moreover, although the contraindications for nicotine replacement are relatively few, those for nortriptyline are greater, so that combination therapy may not be appropriate for all. However, to date, more evidence supports a modest effect of combination nortriptyline than any other combined treatment for smoking cessation.

More research is needed to examine the relative efficacy of combined treatments for smoking cessation because even modest increases in the efficacy of drugs currently available to primary care doctors are likely to have a large effect on public health. For example, after the introduction of nicotine replacement therapy, the US Centers for Disease Control and Prevention estimated that attempts to quit smoking in the United States increased from around one million to seven million each year,¹¹ and the widespread use of this treatment in the US was projected to increase the number of ex-smokers by hundreds of thousands each year.¹²

Another important contribution of Aveyard and colleagues’ study is the use of a pragmatic methodology. Compared with conventional RCTs—which are often funded by the drug industry and based in tertiary centres—the methodology provides a more realistic assessment of drug use and reflects the shared decision making practices of primary care practitioners in the community. Patients’ experiences with drugs in “real world” clinical practices are particularly relevant. Ultimately, primary care doctors rely on clinical judgment and experience to weigh the risks and benefits of a drug such as nortriptyline and to identify subgroups for whom it is contraindicated, such as people at risk for suicide, or for whom it may confer advantages, such as people with mild depression. The pragmatic approach may make doctors more confident that they can generalise results to real world practice, with perhaps greater potential to affect standards of care. Thus, pragmatic designs should be considered in future studies, particularly once efficacy has been established in more traditional RCTs.

If we had an established standard of care for use of combined treatments, such as nortriptyline plus nicotine replacement, given the dramatic benefits of smoking cessation on reducing mortality, even a modest increase in efficacy relative to nicotine replacement alone (consistent with results from the present study) would be expected to prevent millions of premature deaths worldwide over time.