What is the clinical effectiveness and cost-effectiveness of cytisine compared with varenicline for smoking cessation? A systematic review and economic evaluation

Background: Tobacco smoking is one of the leading causes of deaths worldwide. Nearly one-fifth of adults in the UK regularly smoke cigarettes. The ill-health associated with smoking costs the NHS over £3B every year. A number of pharmacological interventions are available that can help people to quit smoking. These include nicotinic receptor partial agonists such as varenicline or cytisine. Varenicline is a synthetic product licensed for use in the UK, while cytisine is derived naturally from the seeds of the plant Cytisus laborinum L. (golden rain acacia). Objectives: To review the evidence on the clinical effectiveness and safety of cytisine from smoking cessation compared with varenicline; to develop an economic model to estimate the cost-effectiveness of cytisine and varenicline; and to provide recommendations based on value of information analyses as to whether or not a head-to-head trial of cytisine and varenicline would represent effective use of resources. Data sources: Efficacy and adverse events data were sourced from a recent Cochrane review. These data were supplemented with an updated search of twelve electronic databases, including MEDLINE, EMBASE, Cumulative Index to Nursing and Allied Health Literature and The Cochrane Library, for the period from December 2011 to January 2013. The review included randomised controlled trials (RCTs) of adult smokers attempting to quit using varenicline or cytisine. Further interventions were considered (placebo, nicotine replacement therapy, bupropion) to allow an indirect comparison between varenicline and cytisine. The primary outcome was abstinence at a minimum of 6 months’ follow-up. Secondary outcomes were common adverse events such as abnormal dreams, headache, nausea, insomnia and serious adverse events. Review methods: A systematic review and network meta-analysis of the clinical evidence was undertaken. A random-effects model was used to allow for heterogeneity between studies. The economic model structure was based on a published model. Probabilistic sensitivity analyses were undertaken to estimate the treatment expected to be most cost-effective given current information. Formal expected value of perfect information, perfect partial information and of sample information were performed. Results: Twenty-three (RCTs) were included in the systematic review, comprising a total of 10,610 participants. Twenty-one trials of varenicline of differing dosing schedules and two trials of cytisine at standard dose met the inclusion criteria. No head-to-head trials comparing varenicline with cytisine were identified. The methodological quality of the studies was judged to be moderate to good. Cytisine was more efficacious than placebo [hazard ratio (HR) 4.27, 95% credible interval (CrI) 2.05 to 10.05], as was standard-dose varenicline (HR 2.58, 95% Crl 2.16 to 3.15). Standard-dose varenicline treatment was associated with significantly higher rates of headache, insomnia and nausea than placebo; there was no significant difference in the rates of abnormal dreams. There were no significant differences in the rates of headache or nausea between cytisine and placebo; data were identified for neither abnormal dreams nor insomnia. Using expected values, cytisine is anticipated to dominate varenicline, in that it produces more quality-adjusted life-years at a lower associated cost. This occurred in approximately 90% of the scenarios performed. However, owing to the large number of people who wish to quit smoking (estimated to be 3 million over a 10-year period), the implications of making an incorrect decision is large. The expected value of sample information indicated that conducting a head-to-head trial of cytisine and varenicline was worthwhile, and that 1000 smokers per arm was an appropriate number to recruit. Conclusions: On the basis of the evidence included in this review, varenicline and cytisine are both effective interventions to aid smoking cessation when compared with placebo. Cytisine is estimated to be both more clinically effective and cost-effective than varenicline. However, there is uncertainty in the decision, and a head-to-head trial of cytisine and varenicline would appear to be an effective use of resources. Study registration: The study was registered as PROSPERO CRD42012003455. Funding details: The National Institute for Health Research Health Technology Assessment programme