Skip to content

Abstract 17059: Smoking and Cessation, DNA Methylation, and Mortality

Nov 2026 · Circulation · 0 citations

Abstract

Background: Self-reported smoking history is susceptible to measurement errors. Although DNA methylation (DNAm) may serve as a biologically relevant long-term biomarker of smoking exposure, there remains a lack of a signature that accurately quantifies DNAm profiles after cessation. Moreover, it remains unclear whether such signatures can inform subsequent mortality risk, especially for residual risk after cessation. Hypothesis: DNAm can objectively quantify smoking history and related mortality risk. Methods: This prospective population-based cohort included 10,127 US male physicians (mean age, 67.2±8.47 years) from the Physicians’ Health Study (PHS) who provided blood samples in 1997 and were followed up through August 2022 (median follow-up of 16.5 years). 1,327 participants from COSMOS/VITAL were also included as validation datasets. Smoking phenotypes (smoking status, pack-years, and years since quitting [YSQ]) were collected through self-reported questionnaires, and DNA methylation profiles were quantified using the Illumina EPIC 850K array. We developed two DNAm signatures on pack-years and YSQ and then evaluated their associations with all-cause, cardiovascular, and cancer mortality. Results: We identified 7,146 smoking-related CpGs, including 897 newly identified (Holm-adjusted p < 0.05). 1,463 of these methylation alterations remained significant among former smokers who quit smoking over 30 years, and 14 did not completely reverse after 40 years. We identified two different DNA methylation signatures associated with pack-years (1248 probes; r=0.70-0.77) and YSQ (437 probes; r=0.66-0.69), both robustly correlated with self-reported smoking exposure and accurately distinguishing smoking status. In PHS, DNAm pack-years and DNAm YSQ were both significantly associated with all-cause mortality, independent of their self-reported phenotypes (HR per SD increment: 1.21 [1.16–1.26] for DNAm pack-years; 0.76 [0.72–0.80] for DNAm YSQ). These two signatures were also jointly and independently associated with mortality risk (p<0.0001 for interaction). Based on DNAm predicted YSQ, former smokers had significantly higher risks until 20-25 years after cessation for all-cause, CVD, and cancer mortality, and 25-30 years for lung cancer and pulmonary mortality. Conclusion: These DNAm signatures robustly reflect smoking and cessation patterns and related epigenetic responses, quantify smoking-related mortality risk without relying on imprecise self-reported information on smoking history, and reveal a persistent residual mortality risk after decades of cessation.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.