A systematic review and network meta-analysis of the safety and effectiveness of smoking cessation medicines and electronic cigarettes

Aim: To assess the comparative safety and effectiveness of smoking cessation medicines.

Background: Cigarette smoking is one of the leading causes of early death. Of UK-licensed stop smoking medicines (Varenicline [Champix], Bupropion [Zyban], Nicotine Replacement Therapy [NRT]), Varenicline has been shown to be the most effective monotherapy for long-term smoking abstinence. However, there is ongoing debate regarding the neuropsychiatric safety of these medicines and of the newly licensed electronic cigarettes (E-Cigarettes).

Methodology: We conducted systematic reviews and network meta-analyses of randomised controlled trials (RCTs) of medicines delivered as monotherapy and in combination treatment (e.g. Varenicline+NRT) with control groups. Studies published before March 2017 were identified from 8 databases. The primary composite safety outcome was serious adverse events (SAEs) and secondary composite outcomes were major adverse neuropsychiatric events (MANEs) and major adverse cardiovascular events (MACEs). The primary effectiveness outcome was sustained smoking cessation abstinence (= 6?months, biochemically validated). We also examined ranking of treatments (monotherapies or combinations of therapies) by estimating mean rank, and the probability that each intervention is best, second best, and so on, for safety and effectiveness outcomes. Finally, we performed sensitivity analyses excluding different subsets of studies, including studies assessed at high risk of bias on any domain, studies including samples with all (or most) participants having one or more current psychiatric conditions or comorbidities, studies focused on smokeless tobacco consumers, and studies where patients were heavy smokers (> 20 cigarettes per day).

Results: We screened 10,779 records and analysed 334 and 346 RCTs for safety and effectiveness outcomes, respectively. We found evidence that Bupropion at standard dose increased the odds of SAEs (OR=1.31, 95% CI 1.06-1.60) but reduced the odds of MANEs (0.62, 0.42-0.87) compared to placebo, whereas E-Cigarette users had higher odds of SAEs compared to those who received placebo (2.34, 1.09-5.16). No treatment increased the odds of MACEs compared to placebo. Ranking treatments according to serious adverse events, where lower number rankings are preferred (1=best, etc), placebo had the best average ranking (1.97) whereas the worst average ranks were for E-Cigarette (6.20) and Varenicline & Bupropion Std (5.48), suggesting patients randomised to these interventions were most likely to experience SAEs. Regarding effectiveness, the odds of achieving sustained abstinence were higher for smokers randomised to standard doses of Bupropion (1.65, 1.39-1.95), Varenicline (OR=2.75, 95% CI 2.29 to 3.29), Varenicline+NRT (5.58, 2.25-14.0) or Varenicline+Bupropion (2.80, 1.15-6.96), alongside those using any E-Cigarette (2.75, 1.26-6.11) compared to smokers treated with placebo. Ranking treatments according to sustained abstinence, Varenicline Std + NRT Std had the best average ranking (mean rank 1.35) while placebo had the lowest rank with a mean rank of 6.98. Sensitivity analyses did not yield results that differed from those of the main analyses.

Conclusions: Combination treatment with varenicline and NRT (currently unlicensed in the UK) is most likely to be effective at achieving sustained abstinence from smoking. E-cigarettes for smoking cessation showed promise, but also increased odds for a SAE based on few studies.