Study design

The CCC-ACS project, launched in 2014, is a large nationwide registry and quality-improvement study with an ongoing database focusing on quality of ACS care. It is a collaborative initiative of the American Heart Association and the Chinese Society of Cardiology. Details of the design and methodology of the CCC project have been published [16]. In brief, the study included 158 tertiary hospitals and 82 secondary hospitals from 30 provinces, autonomous regions, or municipalities of different geographic-economic levels in China. Trained data abstractors at participating hospitals reported the required data from medical records via a web-based data-collection platform (Oracle Clinical Remote Data Capture; Oracle Corp, Redwood City, CA, USA). Third-party clinical research associates performed quality audits to ensure that cases were reported consecutively rather than selectively. Around 5% of reported cases were randomly selected, and their data were compared with the original medical records to ensure accuracy and completeness (Supplementary Methods).

Study population

A total of 92,509 ACS inpatients, identified using their principal discharge diagnosis, were enrolled across China from November 2014 to December 2018. ACS was defined according to guidelines issued by the Chinese Society of Cardiology for the diagnosis and management of patients with ST-elevation myocardial infarction (STEMI) and non-ST-elevation acute coronary syndrome (NSTE-ACS) (including non-STEMI and unstable angina) [7, 8].

Study variables

Statistical analysis

We first assessed smoking prevalence among male and female ACS patients and further in different subgroups. Categorical variables were presented as percentages and numbers of patients by group and compared using chi-square test or Cochran-Armitage trend test. Smoking prevalence of ACS patients was standardized by age according to the 2010 national census when compared with that in the Chinese general population [18]. We used logistic regression analyses to evaluate the association between smoking status and critical cardiac symptoms at admission and in-hospital outcomes among all ACS patients and in STEMI and NSTE-ACS patients, respectively. All model covariates are presented in Supplementary Table 1, and their definitions are presented in Supplementary Methods. We then compared the proportions of smoking patients receiving cessation interventions and other medications prescribed at discharge.

For variables with a missing rate <15%, multiple imputation with chained equations was used to impute missing values by IVEware software version 0.2 (Survey Research Center, University of Michigan, Ann Arbor, MI, USA). Missing rates of variables and strategies for management of missing data are presented in Supplementary Table 2.

Statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA). Two-tailed P values of <0.05 were considered statistically significant.

Smoking prevalence

Of our sample of 92,509 ACS patients, 68,371 (73.9%) were men and 24,138 (26.1%) were women. Smoking prevalence was 52.4% in male patients and 8.0% in female patients. Smoking was most common among patients younger than 45 years (men: 68.0%; women: 14.9%) (Figure 1). Among patients with recurrent ACS, 40.2% of male patients and 5.9% of female patients were smokers (Supplementary Table 3). Smoking prevalence was over 50% in 21 of the 30 Chinese provinces among male patients and varied widely between provinces among female patients, from 1.2% in Zhejiang to 26.4% in Tianjin.

The age-standardized smoking prevalence was five times higher in female ACS patients than in Chinese women in general (12.6% versus 2.7%) and 1.3 times higher in male ACS patients than in Chinese men in general (63.0% versus 47.2%) (Supplementary Figure 1).

There was a decline in smoking prevalence among ACS patients from 2014 to 2018, decreasing from 54.6% to 47.3% in male patients and from 9.1% to 6.4% in female patients (Supplementary Figure 2). However, smoking was still the most prevalent risk factor among younger male patients with initial ACS, and smoking prevalence among female patients younger than 45 years increased from 11.8% to 16.0% during the study period.

Association between smoking and patients’ clinical conditions

Association between smoking and severe clinical manifestations

Compared with non-smokers, a higher proportion of patients who smoked had STEMI (67.8% versus 54.8%, p < 0.001) and substantially elevated myocardial injury markers (86.1% versus 83.0%, p < 0.001) (Table 1). Smoking was associated with increased odds of critical cardiac symptoms at admission after multivariable adjustment (odds ratio (OR) = 1.14, 95% confidence interval (CI): 1.08–1.20; p < 0.001), especially among NSTE-ACS patients (OR = 1.35, 95% CI: 1.22–1.49; p < 0.001) (Figure 2). The similar findings were identified in both male and female patients. Compared with non-smokers, both male and female patients who smoked had a higher proportion of STEMI (68.0% versus 58.0%, p < 0.001 in men; 63.6% versus 50.0%, p < 0.001 in women) and a higher proportion of the patients with substantially elevated myocardial injury markers (86.1% versus 83.6%, p < 0.001 in men; 87.3% versus 82.2%, p < 0.001 in women). In separated analyses for male and female patients, the association of smoking with higher risk of critical cardiac symptoms at admission remained significant in both male (OR = 1.15, 95% CI: 1.08–1.21; p = 0.004) and female (OR = 1.16, 95% CI: 1.00–1.35; p = 0.048) patients.

Figure 2 

Multivariable analysis of association between smoking and critical cardiac symptoms at admission and in-hospital outcomes among ACS patients. This forest plot shows patients’ critical cardiac symptoms at admission and in-hospital outcomes according to smoking status among all ACS patients, and among patients by subtypes of ACS, using data from the CCC-ACS project. Critical cardiac symptoms at admission included acute heart failure, cardiogenic shock, and cardiac arrest. ACS: acute coronary syndrome; CI: confidence interval; NSTE-ACS: non-ST-elevation acute coronary syndrome; OR: odds ratio; STEMI: ST-elevation myocardial infarction.

Association between smoking and in-hospital outcomes

Smokers had lower frequency of in-hospital outcomes than non-smokers (2.4% versus 3.9%, p < 0.001), particularly among STEMI patients (2.7% versus 5.1%, p < 0.001) (Supplementary Figure 3). And smokers had lower proportion of in-hospital outcomes than non-smokers in both male (2.3% versus 3.5%, p < 0.001) and female (4.1% versus 4.4%, p = 0.518) patients. However, after adjusting for patients’ demographic characteristics, medical history, clinical manifestations, in-hospital treatment, and other confounding factors, the odds of experiencing the in-hospital outcomes investigated was close to 1 for ACS patients in general (OR = 0.93, 95% CI: 0.84–1.02; p = 0.107), as well as for those with STEMI (OR = 0.91, 95% CI: 0.82–1.02; p = 0.103) or NSTE-ACS (OR = 0.99, 95% CI: 0.82–1.20; p = 0.934) (Figure 2). The association between smoking and in-hospital outcomes remained unchanged in both male (OR = 0.95, 95% CI: 0.85–1.05; p = 0.286) and female (OR = 0.93, 95% CI: 0.72–1.20; p = 0.585) patients.

Smoking cessation interventions before discharge

Provision of smoking cessation interventions was evaluated in 37,336 smokers with ACS who survived discharge during the study period. Of these patients, 35.3% received smoking cessation interventions before discharge. The proportions of patients receiving smoking cessation interventions varied greatly between hospitals (from 0% to 100%). The rate of provision was <10% in around 40% of hospitals, and overall was inadequate for both male (35.6%) and female (28.7%) smokers. Only 36.0% of smokers younger than 45 years received smoking cessation interventions. Even among smokers with recurrent ACS, only 31.8% received interventions (Supplementary Table 4).

Providing smoking cessation brochures to smokers was the most common intervention (30.8%), followed by making personalized cessation plans (9.1%) and prescription of pharmacotherapy (0.6%) (Figure 3).

We found a promising improvement in provision of smoking cessation interventions, with the yearly rate increasing from 29.4% to 41.9% during the study period (Supplementary Figure 4). However, these proportions were still significantly lower than other quality-of-care indexes, including discharge on aspirin (92.6%), statins (92.8%), beta-blockers (67.4%), and ACEI or angiotensin-receptor blocker (56.4%) (Figure 4).

High smoking prevalence among ACS patients in China

This study demonstrated that smoking was highly prevalent in ACS patients across China. More than half of male patients and around 70% of younger male patients were smokers. Age-standardized smoking prevalence in female patients was five times higher than in Chinese women in general.

Smoking prevalence among ACS patients in China was notably higher than that found in many other countries. The recent European Action on Secondary and Primary Prevention by Intervention to Reduce Events (EUROASPIRE) V survey reported that the mean smoking prevalence among CHD patients was 19.0% in 27 countries [19]. Even as early as 1995, the EUROASPIRE I survey reported a mean smoking prevalence of 19.4% among CHD patients in nine countries [20]. In the United States, the National Cardiovascular Data Registry reported that smoking prevalence among AMI patients was 32.4% in 2017 [21].

Various factors could account for the high smoking prevalence among ACS patients in China. First, smoking among the general Chinese population represents a major public health issue, particularly among men, with overall smoking prevalence significantly exceeding the global average (47.2% versus 25.0%) [22, 23]. Second, smoking is a major risk factor for ACS. Previous studies have reported that smokers have a 75% higher risk of ACS than non-smokers [24]. The highly epidemic smoking problem among ACS patients suggests an urgent need to strengthen smoking-related public health interventions in China. Recent population-level tobacco interventions include adopting tobacco clauses (e.g., Advertising Law of 2015) and local smoke-free regulations [22]. While these have been met with declines in smoking prevalence, further efforts are needed to achieve the goal of reducing the nationwide smoking prevalence to 20% by 2030 [25].

Serious clinical conditions associated with smoking in ACS patients

This study showed that smoking was associated with serious clinical conditions in ACS patients. First, among male patients with initial ACS, smokers experienced ACS approximately six years earlier than non-smokers. Patients in whom ACS occurs in young or middle age have a higher risk of premature mortality and sudden cardiac death outside of hospital. Previous study showed that among younger patients who died of AMI, 90% died outside of hospital [26]. Those patients who survived ACS are more likely to experience a lower quality of life and a limited ability to work [27].

Second, although smokers were younger and had fewer other cardiovascular disease risk factors than non-smoking patients, a higher proportion had a diagnosis of STEMI (67.8% versus 54.8%) and substantially elevated cardiac injury markers (86.1% versus 83.0%) indicating more severe myocardial damage to the heart. Smoking was also associated with an increased risk of critical cardiac symptoms, which are important predictors of poor outcomes at admission [3, 5].

No direct association between smoking and in-hospital outcomes in ACS patients

We did not find that smoking has a direct association with in-hospital outcomes among ACS patients. Previous studies have reported that smokers had lower crude rates of in-hospital outcomes than non-smokers [12, 13]. Similar to previous studies, we found that smokers in our study population were younger than non-smokers and less likely to have other cardiovascular risk profiles or experience the in-hospital outcomes investigated. The fact that our multivariable analysis found no difference in in-hospital outcomes between smokers and non-smokers suggests that observed differences of in-hospital outcomes between smokers and non-smokers could be attributable to confounding covariates.

Moreover, after reviewing the studies which reported smokers had lower risk of in-hospital outcomes after ACS than non-smokers [12, 13], we found some limitations in their analytical approaches; they did not comprehensively adjust for relevant confounders for the associations between smoking and in-hospital outcomes (e.g., clinical manifestations and in-hospital treatment).

In addition, some studies with longer follow-up periods have found strong associations between smoking and adverse long-term clinical outcomes among ACS patients [28, 29, 30]. These findings highlight the value of providing smoking cessation interventions for ACS patients.

Inadequate provision of smoking cessation interventions

We found a low rate of provision of smoking cessation interventions among ACS patients who smoked in this study, with two-thirds of patients not receiving formal intervention before discharge.

Although quitting smoking has been recognized as an effective and cost-effective secondary prevention strategy for CHD patients, the rate of provision of smoking cessation interventions was still much lower than that of other secondary prevention medications. This suggests that smoking cessation interventions remain an underutilized option for secondary ACS prevention in China. Gaps in care might result from inadequate awareness of providers, structural inadequacies of care systems, or physicians’ limited experience in successful provision of smoking cessation interventions. One systematic review found that intensive smoking cessation interventions initiated during hospitalization could help smokers quit smoking, and that the period of hospitalization was a ‘teachable moment’ for patients [9]. Physicians should therefore capitalize on opportunities to provide smoking cessation interventions before discharge for patients who smoke.

High rates of provision of smoking cessation interventions have been achieved in other countries. The EUROASPIRE IV survey reported that around 90% of CHD patients across 24 countries received smoking cessation interventions [31]. Studies from the United States from 2007 to 2017 have reported consistently high rates of provision of smoking cessation interventions for ACS patients (of nearly 100%) [21, 32], whereas historically only 40% to 58% of AMI patients were offered smoking cessation interventions [33, 34]. This notable increase can be attributed to implementation of quality-improvement programs, which can bridge knowledge gaps or change physicians’ attitudes toward smoking cessation interventions [32]. National tobacco control efforts, including mass media campaigns, tobacco taxes, smoking bans, and plain packaging with graphic warnings, are essential for the achievement of society-wide reductions in smoking prevalence. Although there may be some differences in medical record documentation among different countries, the lower rate of provision of smoking cessation interventions for ACS patients at discharge in China compared with other countries suggested that it remains a significant area for improvement.

Limitations

This study had some limitations. First, the data about smoking durations and daily numbers of cigarettes of smokers were unavailable and we were unable to study the dose-response association of total smoking exposure with the onset age and in-hospital outcomes among ACS patients. Second, as all patients did not smoke during hospitalization, we were unable to evaluate the effect of smoking cessation during hospitalization on in-hospital outcomes of patients. It is possible that quitting smoking in short term have a potential effect on in-hospital outcomes of patients. Third, this study did not collect information on patients who died before admission. It is possible that smokers with ACS had a greater risk of sudden cardiac death before admission than non-smokers [35], with those who were hospitalized already representing ‘survivors.’