Effects of varenicline on adverse outcomes: an observational cohort study using electronic medical records. (November 2016)

• Record Type: Journal Article
• Authors: Davies, Neil M; Taylor, Gemma M J; Taylor, Amy E; Jones, Tim; Martin, Richard M; Munafò, Marcus R; Windmeijer, Frank; Thomas, Kyla H
• Is Part Of: Lancet. Volume 388(2016)Supplement 2
• Page Start: S42
• Publication Date: 2016-11
• Subjects: Medicine -- Periodicals; Medicine -- Periodicals; Medicine; Medicine; Electronic journals; Periodicals; 610.5
• DOI: 10.1016/S0140-6736(16)32278-4
• Languages: English
• ISSNs: 0140-6736

Abstract

Abstract: Background: Regulators have issued warnings about the potential adverse effects of the smoking cessation medication varenicline. However, there is little evidence of the long-term effects of varenicline when prescribed in everyday clinical practice. We aimed to investigate the association of varenicline and nicotine replacement therapy (NRT) with adverse outcomes when prescribed in primary care. Methods: Using the Clinical Practice Research Datalink (CPRD), we conducted an observationalcohort study of electronic medical records and linked mortality and hospital admission data of 126 718 primary care patients in the UK prescribed varenicline (41 742) and NRT (84 976). We used three approaches to overcome potential residual confounding: multivariable adjusted regression, propensity score matching, and instrumental variable regression. The exposure was first prescription of varenicline or nicotine replacement therapy. The primary outcome was all-cause mortality 2 years after first prescription. Secondary outcomes were mortality from cardiovascular and respiratory disease, all-cause hospital admission and for cardiovascular and respiratory disease, incident primary care diagnosis of myocardial infarction or chronic obstructive pulmonary disease, and frequency of primary care attendance. Findings: In multivariable adjusted analysis, mortality was lower among patients prescribed varenicline than in those prescribed NRT (OR 0·78, 95% CI 0·69 to 0·87; p<0·0001), as was hospital admission (0·85, 0·83 to 0·88; p<0·0001). There were similar rates of incident myocardial infarction (0·92, 0·75 to 1·14) and chronic obstructive pulmonary disease (1·05, 0·96 to 1·16). Patients prescribed varenicline were 13·9 percentage points less likely to subsequently attend primary care (95% CI 11·8 to 16·0, p<0·0001). The instrumental variable analysis found little evidence of a substantial increase in risk of adverse outcomes of mortality (mean difference per 100 patients treated 0·50, 95% CI −0·50 to 1·50), hospital admission (−0·91, −3·64 to 1·81), myocardial infarction (0·42, −0·05 to 0·88), or chronic obstructive pulmonary disease (−0·21, −1·39 to 0·96). Interpretation: There was little evidence of a substantial increase in risk of adverse outcomes in patients prescribed varenicline as part of their standard clinical care. The regulatory warnings associated with varenicline do not accurately reflect the scientific evidence and should be reconsidered by policy makers and regulators. Funding: The research described in this paper was funded by the Medical Research Council (MR/N01006X/1), the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) programme (project number 14/49/94). KHT is funded by a clinical lectureship award from the NIHR. RMM is supported by Cancer Research UK (programme grant C18281/A19169) (the Integrative Cancer Epidemiology Programme). TJ is supported by the NIHR Collaboration for Leadership in Applied Health Research and Care West, University Hospitals Bristol NHS Foundation Trust. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the abstract for submission.