Tobacco smoking is the leading behavioural risk factor for cardiovascular disease and may double the risk of long-term mortality after coronary artery bypass grafting (CABG) [1, 2]. Smoking cessation interventions, which combine pharmacological treatment and behavioural support, are effective at supporting abstinence at the time of surgery and at 12 months postoperatively [3]. However, smoking remains more prevalent among those patients undergoing surgery (25%) than in the general UK population (13%) [4, 5]. There is limited contemporary evidence about the burden of smoking in cardiac surgery. This study investigated the prevalence of smoking in people undergoing CABG surgery in the UK and the impact of smoking on postoperative outcomes. This will help with risk assessment of such patients, as well as resource allocation and strategic planning for addressing smoking-related issues in surgical contexts.

Permission was obtained for the National Institute for Cardiovascular Outcomes Research (NICOR) to release depersonalised patient data from the National Adult Cardiac Surgery Audit (NACSA) under an agreement between NHS England/GIG Cymru and King's College Hospital NHS Foundation Trust. The full process of data submission and processing by NICOR is described elsewhere [6]. All adults undergoing elective CABG between January 2012 and December 2022 were included. Other elective procedures were included only if performed in addition to CABG. Patients were not included if they underwent non-elective surgery, had no documented smoking status or had chosen not to have their data used for research. This study involved analysis of existing non-identifiable patient data and was, therefore, exempt from NHS ethics committee approval.

The primary outcome was prevalence of current smoking (one or more cigarettes per day); former smoking (not smoked within the last month); and non-smoking (never smoked) at the time of surgery. The secondary outcomes were trends in smoking prevalence over time; in-hospital mortality; postoperative duration of hospital stay; and complications. Smoking status was reported by year of operation and trends in smoking prevalence over time were assessed using multivariable logistic regression, adjusted for age, sex and procedure type. The incidence of postoperative complications and duration of hospital stay were compared between smoking status groups using multivariable logistic regression and Cox proportional hazards regression, respectively, adjusted for surgical risk using the European System for Cardiac Operative Risk Evaluation 2 (EuroSCORE 2) [7]. Statistical analysis was performed using R version 4.2.1 (R Studio, Vienna, Austria) with a two-sided significance level set at p < 0.05. Multiple imputation was performed using the multivariate imputation by chained equations package (version 4.2.3) to account for missing data with < 50% of missing values. Sensitivity analysis was done using complete case analysis.

A total of 96,071 patients were included in the analysis. Of these, 8237 (8.6%) were current smokers, 52,074 (54.2%) former smokers and 35,760 (37.2%) non-smokers. Patient characteristics and procedural details by smoking status are detailed in online Supporting Information Table S1. The prevalence of smoking remained stable at 8.6% through the study period (Fig. 1). However, the adjusted odds of smoking were marginally lower with each advancing year (OR 0.99, 95%CI 0.98–1.00, p = 0.03). The proportion of ex-smokers fell over this period from 56.3% to 49.0% (OR 0.97, 95%CI 0.97–0.98, p < 0.01), while the proportion of non-smokers increased from 35.0% to 42.4% (OR 1.03, 95%CI 1.03–1.04, p < 0.01). Smokers had higher odds of developing deep sternal wound infections and were more likely to need surgical debridement than non-smokers (Table 1). No significant differences were observed in the odds of in-hospital mortality, returning to the operating theatre, developing a new neurological deficit or renal replacement therapy postoperatively between smokers and non-smokers. The mean duration of hospital stay was slightly shorter for smokers than for non-smokers (hazard ratio 0.94, 95%CI 0.92–0.97), but no significant differences were observed in outcomes. Complete case analysis result estimates (online Supporting Information Table S2) were almost identical to those of the analysis using imputed data.

The main limitations of this study were self-reported smoking status and lack of granularity in the recorded smoking histories, with no stratification for number of cigarettes smoked or timing of cessation. Several important lifestyle risk factors and outcomes of interest, such as long-term mortality and pulmonary complications, were not available from the NACSA. Finally, some adults undergoing elective CABG in the UK may not be included in the NACSA, but this capture rate is not known.

In this nationwide cohort study, smoking was less prevalent among people undergoing CABG surgery than in the UK general population and other surgical cohorts within the study period. Future work should explore the patterns in smoking behaviour among people referred for cardiac surgery and the factors which support successful quitting in this cohort. Sharing learning with other surgical specialities may improve smoking cessation support within other peri-operative pathways and allow replication of this rare success story.

吸烟是心血管疾病的主要行为危险因素，冠状动脉旁路移植术 （CABG） 后的长期死亡风险可能增加一倍 [1， 2]。戒烟干预结合了药物治疗和行为支持，可有效支持手术时和术后 12 个月的戒烟 [3]。然而，吸烟在接受手术的患者 （25%） 中仍然比在英国普通人群 （13%） 中更普遍 [4， 5]。关于心脏手术中吸烟负担的当代证据有限。本研究调查了英国接受 CABG 手术的患者吸烟率以及吸烟对术后结局的影响。这将有助于对此类患者进行风险评估，以及解决手术环境中吸烟相关问题的资源分配和战略规划。

根据 NHS England/GIG Cymru 与国王学院医院 NHS 基金会信托基金之间的协议，国家心血管结果研究所 （NICOR） 获得许可，可以发布国家成人心脏手术审计 （NACSA） 中的非个性化患者数据。NICOR 提交和处理数据的完整过程详见其他 [6]。纳入 2012 年 1 月至 2022 年 12 月期间接受择期 CABG 的所有成人。只有在 CABG 之外进行的其他择期手术才被纳入。如果患者接受了非择期手术、没有记录的吸烟状况或选择不将其数据用于研究，则不被纳入。这项研究涉及对现有不可识别的患者数据的分析，因此免于NHS伦理委员会的批准。

主要结局是当前吸烟率（每天一支或多支香烟）;以前吸烟（上个月未吸烟）;手术时不吸烟（从不吸烟）。次要结局是吸烟率随时间变化的趋势;院内死亡率;术后住院时间;和并发症。按手术年份报告吸烟状况，并使用多变量 logistic 回归评估吸烟流行率随时间的趋势，并根据年龄、性别和手术类型进行调整。使用多变量 logistic 回归和 Cox 比例风险回归分别比较吸烟状况组之间的术后并发症发生率和住院时间，并使用欧洲心脏手术风险评估系统 2 （EuroSCORE 2） 调整手术风险 [7]。使用 R 4.2.1 版（R Studio，维也纳，奥地利）进行统计分析，双侧显着性水平设置为 p < 0.05。使用通过链式方程的多变量插补包（版本 4.2.3）进行多重插补，以解释 < 缺失值为 50% 的缺失数据。使用完整的病例分析进行敏感性分析。

共有 96,071 名患者被纳入分析。其中，8237 人 （8.6%） 是当前吸烟者，52,074 人 （54.2%） 是前吸烟者，35,760 人 （37.2%） 是非吸烟者。按吸烟状况划分的患者特征和操作细节详见在线支持信息表 S1。在整个研究期间，吸烟率稳定在 8.6%（图 1）。然而，随着年龄的增长，调整后的吸烟几率略低 （OR 0.99,95%CI 0.98–1.00，p = 0.03）。在此期间，戒烟者的比例从 56.3% 下降到 49.0% （OR 0.97,95%CI 0.97-0.98，p < 0.01），而不吸烟者的比例从 35.0% 增加到 42.4% （OR 1.03,95%CI 1.03-1.04，p < 0.01）。吸烟者发生胸骨深部伤口感染的几率更高，并且比不吸烟者更可能需要手术清创（表 1）。吸烟者和非吸烟者在术后住院死亡率、重返手术室、出现新的神经功能缺损或肾脏替代治疗的几率未观察到显著差异。吸烟者的平均住院时间略短于非吸烟者 （风险比 0.94,95% CI 0.92-0.97），但未观察到结局的显著差异。完整的病例分析结果估计值（在线支持信息表 S2）与使用插补数据的分析几乎相同。

这项研究的主要局限性是自我报告的吸烟状况和记录的吸烟史缺乏粒度，没有对吸烟数量或戒烟时间进行分层。NACSA 无法获得几个重要的生活方式风险因素和感兴趣的结局，例如长期死亡率和肺部并发症。最后，一些在英国接受选择性 CABG 的成年人可能不包括在 NACSA 中，但这个捕获率尚不清楚。

在这项全国性的队列研究中，吸烟在接受 CABG 手术的人群中低于英国普通人群和研究期间的其他手术队列。未来的工作应探讨转诊接受心脏手术的人的吸烟行为模式以及支持该队列成功戒烟的因素。与其他外科专业分享学习可能会改善其他围手术期途径中的戒烟支持，并允许复制这一罕见的成功故事。