Genome-wide association study of varenicline-aided smoking cessation

Abstract Introduction Varenicline is an α4β2 nicotinic acetylcholine receptor partial agonist with the highest therapeutic efficacy of any pharmacological smoking cessation aid and a 12-month cessation rate of 26%. Genetic variation may be associated with varenicline response, but to date no genome-wide association studies of varenicline response have been published. Methods In this study, we investigated the genetic contribution to varenicline effectiveness using two electronic health record-derived phenotypes. We defined short-term varenicline effectiveness (SVE) and long-term varenicline effectiveness (LVE) by assessing smoking status at 3 and 12 months, respectively, after initiating varenicline treatment. In Stage 1, comprising five European cohort studies, we tested genome-wide associations with SVE (1,405 cases, 2,074 controls) and LVE (1,576 cases, 2,555 controls), defining sentinel variants (the most strongly associated variant within 1 megabase) with p-value <5×10–6 to follow up in Stage 2. In Stage 2, we tested association between sentinel variants and comparable smoking cessation endpoints in varenicline randomised controlled trials. We subsequently meta-analysed Stages 1 and 2. Results No variants reached genome-wide significance in the meta-analysis. In Stage 1, 10 sentinel variants were associated with SVE and five with LVE at a suggestive significance threshold (p-value <5×10–6); none of these sentinels were previously implicated in varenicline-aided smoking cessation or in genetic studies of smoking behaviour. Conclusions We provide initial insights into the biological underpinnings of varenicline-aided smoking cessation, through implicating genes involved in various processes, including gene expression, cilium assembly and early-stage development. Implications Leveraging electronic health records, we undertook the largest genetic study of varenicline-aided smoking cessation to date, and the only such study to test genome-wide associations. We showed distinct genetic variants associated (p-value <5×10–6) with varenicline-aided smoking cessation which implicate diverse cellular functions, including transcriptional regulation, RNA modification and cilium assembly. These provide insights which, if independently corroborated, will improve understanding of varenicline response. The growing availability of biobank resources with genetic and varenicline response data will provide future opportunities for larger studies using the approach we developed.