Combination Therapies for Smoking Cessation: A Hierarchical Bayesian Meta-Analysis

Abstract

Context: Treatment guidelines recommend the use of combination therapies for smoking cessation, particularly behavioral therapy (BT) as an adjunct to pharmacotherapy. However, these guidelines rely on previous reviews with important limitations. This study's objective was to evaluate the efficacy of combination therapies compared with monotherapies, using the most rigorous data available.

Evidence acquisition: A systematic review and meta-analysis of RCTs of pharmacotherapies, BTs, or both were conducted. The Cochrane Library, Embase, PsycINFO, and PubMed databases were systematically searched from inception to July 2015. Inclusion was restricted to RCTs reporting biochemically validated abstinence at 12 months. Direct and indirect comparisons were made in 2015 between therapies using hierarchical Bayesian models.

Evidence synthesis: The search identified 123 RCTs meeting inclusion criteria (60,774 participants), and data from 115 (57,851 participants) were meta-analyzed. Varenicline with BT increased abstinence more than other combinations of a pharmacotherapy with BT (varenicline versus bupropion: OR=1.56, 95% credible interval [CrI]=1.07, 2.34; varenicline versus nicotine patch: OR=1.65, 95% CrI=1.10, 2.51; varenicline versus short-acting nicotine-replacement therapies: OR=1.68, 95% CrI=1.15, 2.53). Adding BT to any pharmacotherapy compared with pharmacotherapy alone was inconclusive, owing to wide CrIs (OR=1.17, CrI=0.60, 2.12). Nicotine patch with short-acting nicotine-replacement therapy appears safe and increases abstinence versus nicotine-replacement monotherapy (OR=1.63, CrI=1.06, 3.03). Data are limited concerning other pharmacotherapy combinations and their safety and tolerability.

Conclusions: Evidence suggests that combination therapy benefits may be less than previously thought. Combined with BT, varenicline increases abstinence more than other pharmacotherapy with BT combinations.