Prognostic Value of the CHADS₂ Score for Adverse Cardiovascular Events in Coronary Artery Disease Patients Without Atrial Fibrillation—A Multi‐Center Observational Cohort Study

Introduction

Atrial fibrillation (AF) is a common cardiac arrhythmia associated with substantial morbidity and mortality from thromboembolisms, which can induce stroke. The occurrence of stroke is proportional to the presence of certain risk factors.1 The CHADS₂ score (scored as 1 point each for congestive heart failure, hypertension, diabetes mellitus, or age ≥75 years, and 2 points each for past stroke or transient ischemic attack) has been used to evaluate the risk for ischemic stroke in AF patients in terms of intracardiac thrombogenesis, allowing tailored initiation of antithrombotic treatments for risk reduction, balancing against the risk of bleeding from long‐term anticoagulation.1, 2, 3 This score is commonly used in clinical practice to guide decisions regarding anticoagulant as well as antiplatelet therapy.4 Recently, the CHA₂DS₂‐VASc (scored as 1 point each for congestive heart failure, hypertension, diabetes mellitus, vascular disease, age 65 to 75 years, or female sex, and 2 points each for age ≥75 years or past stroke/transient ischemic attack) and R₂CHADS₂ (adding 2 points for renal failure to the CHADS₂ score) scores were developed to improve the precision of risk stratification by including other important predictive factors. These newer scores have been reported to improve the accuracy of the CHADS₂ score in estimating the risk of stroke in AF patients.5, 6

Each of the components of the CHADS₂ score has long been associated with ischemic stroke in patients with coronary artery disease (CAD) in large, cohort studies.7, 8, 9 This suggests the possible utility of the score in predicting a wide range of cerebrovascular and cardiovascular diseases in the CAD population. Recently, several studies have reported on the use of CHADS₂'s predictive value for cardiovascular events in AF patients10 and cerebrovascular events in non‐AF patients.11 It has been suggested that the CHADS₂ score might predict cardiovascular events in non‐AF patients, though this has been investigated in relatively small populations or in a single‐center study.12

The CHADS2 score is the original score to predict cerebrovascular accidents in AF patients, it is more prevalent than the CHA₂DS₂‐VASc and R₂CHADS₂ scores, and, at present, the score is still described in the guideline and is in use.13 On the other hand, the predictive value of this original score in CAD patients without AF is not fully demonstrated in a large and multicenter study. Thus, we wished to investigate the value of the original CHADS₂ score in predicting cardiovascular/cerebrovascular events in CAD patients without AF, in a multicenter and in a large population.

Methods

We conducted a multicenter, observational cohort study of 7082 consecutive CAD patients (mean age, 69.7 years; male, 71.9%) without clinical evidence of AF requiring percutaneous coronary intervention (PCI). They were enrolled between June 2008 and March 2011 through the KICS (Kumamoto Intervention Conference Study) registry, a physician‐initiated, non‐company‐sponsored, multicenter registry involving 15 centers in Japan. The study complied with the Declaration of Helsinki regarding investigation in humans and was approved by each institutional ethics committee. Informed consent was obtained from all patients.

Baseline demographic data, cardiovascular risk factors, and medications on discharge were documented. We defined diabetes mellitus as symptoms of diabetes mellitus and a casual plasma glucose concentration ≥200 mg/dL, fasting plasma glucose concentration ≥126 mg/dL, 2‐hour plasma glucose concentration ≥200 mg/dL from a 75‐g oral glucose tolerance test, or taking medication for diabetes mellitus. Hypertension was defined as >140/90 mm Hg or taking antihypertensive medication, and dyslipidemia was defined as low‐density lipoprotein ≥140 mg/dL (≥3.63 mmol/L), high‐density lipoprotein <40 mg/dL (1.04 mmol/L), or triglycerides ≥150 mg/dL (≥1.7 mmol/L). We also evaluated the incremental effect of chronic kidney disease (CKD) on clinical outcome, following the recent suggestion that adding renal function to the score (R₂CHADS₂) improves stroke risk stratification in AF patients over that of CHADS₂.6 CKD was defined as an estimated glomerular filtration rate <60 mL/min per 1.73 m². Smoking status was determined by interview. Acute coronary syndrome was defined as either an acute myocardial infarction (MI; ST‐elevation MI or non‐ST‐elevation MI) or unstable angina pectoris. Patients with past or current intermittent claudication associated with an ankle‐brachial index value of <0.9 in either leg were categorized as having peripheral arterial disease. Patients with previous ischemic stroke or transient ischemic attack were defined as having cerebrovascular disease.

The CHADS₂ score was calculated for each PCI patient at discharge. The previous studies to investigate the predictive value of the CHADS₂ score for cardiovascular events in AF patients10 and cerebrovascular events in non‐AF patients11 adopted categories of low (0–1), intermediate (2–3), and high (4–6) scores. This categorization had been reported to have a higher value of C‐statistic than that of categorization into low and high scores in the original literature of the CHADS₂ score.14 In the present study, we also divided subjects according to score values into 3 subgroups similarly to describe baseline poststenting characteristics and then to evaluate the effect of higher CHADS₂ score on clinical outcome. We classified the severity of CAD into single‐, double‐, and triple‐vessel or left main trunk (LMT) disease requiring PCI.

After coronary stent implantation, patients were followed prospectively at outpatient clinics in each institution. Cardiovascular and cerebrovascular events were ascertained from a review of medical records and confirmed by direct contact with the patients, their families, and physicians. The end point was a composite of cardiovascular or cerebrovascular death, nonfatal MI, and ischemic stroke at 1‐year follow‐up after PCI; it has been reported that patients with atherosclerotic arterial disease or at risk of atherothrombosis experience high incidence of cardiovascular events within 1 year.15 “Cardiovascular death” was defined as death attributed to MI, congestive heart failure, or documented sudden cardiac death. We used the universal definition of MI in this study.16 The diagnosis of ischemic stroke was based on clinical and radiological evidence of stroke. For subjects who had ≥2 cardiovascular events, only the first event was considered in the analysis.

The Shapiro–Wilk test was used to assess the normal distribution of continuous data. Continuous variables with a normal distribution are expressed as the mean (SD). The results of the CHADS₂ score were expressed by medians and interquartile ranges. Categorical data are presented as numbers or percentages. Differences between 2 groups were tested using Fisher's exact test or the chi‐squared test for categorical variables, as appropriate. Differences in continuous variables were analyzed by the ANOVA or the Kruskal–Wallis test, as appropriate. We used the Kaplan–Meier method to estimate the cardiovascular event probabilities at 365 days and also the log‐rank test to compare the distributions of survival times among groups. Cox proportional hazard models were used to calculate hazard ratios. Predictors of clinical outcome identified through univariate analysis were tested in a multivariate analysis. We selected variables of statistically significant in the univariate analyses (P<0.05) and to exclude variables that will cause internal correlations. The factors of age, hypertension, diabetes mellitus, cerebrovascular disease, and heart failure were components of the CHADS₂ score, and we thought these variables cause internal correlations with the CHADS₂ score variable. Estimates of the C‐statistic for the risk factors were calculated after the addition of the CHADS₂ score and CKD factor to the risk factors of independent predictive values identified in the multivariate Cox proportional hazards regression analyses; it is generally considered C‐statistic above 0.7 acceptable discriminatory power, that above 0.8 excellent discriminatory power, and that above 0.9 outstanding discriminatory power. The incremental effect of adding the CHADS₂ score and CKD factor or the to other risk factors in predicting future cardiovascular events was evaluated using the net reclassification improvement (NRI) as previously described.17 Considering potential center effects, we adjusted predictive models by including center variables as dummy variables. P<0.05 was considered to denote statistical significance. Statistical analyses were performed using commercial software (SPSS version 22; IBM Inc, Armonk, NY).

Results

A total of 7082 CAD patients requiring PCI were recruited into this study. The CHADS₂ scores of the patients are shown in Figure 1. The baseline laboratory and clinical findings for the study patients according to low, intermediate, and high scores are listed in Table 1. Patients with high CHADS₂ scores were older; had higher rates of dyslipidemia, CKD (including hemodialysis status), past coronary artery bypass graft, peripheral arterial disease, and use of angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers and beta‐blockers; and lower rates of current smoking.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Distribution of CHADS₂ scores over the total population.

The left side of Figure 2A shows the distribution of triple‐vessel or LMT CAD among the low, intermediate, and high CHADS₂ scores. The rate of triple‐vessel or LMT disease positively correlated with a higher CHADS₂ score. The CHADS₂ scores among single‐, double‐, and triple‐vessel or LMT CADs were 2 (1–2), 2 (1–3), and 2 (2–3), respectively (P<0.001; the right side).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

A, CHADS₂ scores and severity of CAD. The left side; rates of triple‐vessel or left main trunk disease among low, intermediate, and high CHADS₂ scores. The right side, CHADS₂ scores among patients with single, double, and triple‐vessel or left main trunk disease. B, CHADS₂ scores and rates of adverse cardiovascular events. The left side; rates of primary outcome among low, intermediate, and high CHADS₂ scores. The right side; CHADS₂ scores compared between patients with and without cardiovascular events.

As detailed in Table 2, a total of 194 patients (2.8%) suffered a cardiovascular or cerebrovascular event. The rates of cardiovascular events among low, intermediate, and high CHADS₂ score groups were 1.5%, 3.3%, and 4.9%, respectively (Figure 2B, left side). Patients with cardiovascular/cerebrovascular events had significantly higher CHADS₂ scores than those without events, at 2 (2–3) versus 2 (1–3; P<0.001; Figure 2B, right side). We performed a Kaplan–Meier analysis and observed that a significantly higher incidence of cardiovascular/cerebrovascular events was in proportion to a higher CHADS₂ score (log‐rank test, P<0.001; Figure 3A). Rates of cardiovascular or cerebrovascular death and ischemic stroke were higher in proportion to CHADS₂ score (P<0.001; Figure 3B and 3D), respectively, but there was no significant difference in the rates of non‐fatal MI among the low, intermediate, and high CHADS₂ score groups (P=0.331; Figure 3C). We also evaluated the occurrence of cardiovascular events during 2 periods, from discharge to 30 days and from 30 days to 1 year (Figure 4). The results of Cox proportional hazards analyses are shown in Table 3. We included the CHADS₂ score, body mass index, acute coronary syndrome, dyslipidemia, CKD, and peripheral arterial disease, and excluded hemodialysis, 1 vessel disease, LMT/3 vessel disease, and statin attributed to the internal correlation. Multivariate analyses identified CHADS₂ score (per 1 point) as an independent and significant predictor of the primary outcome (hazard ratio, 1.31; 95% CI, 1.17–1.47; P<0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Kaplan–Meier analyses at 1‐year follow‐up. Kaplan–Meier analyses of primary outcome (A), cardiovascular or cerebrovascular death (B), nonfatal myocardial infarction (C), and ischemic stroke (D). MI indicates myocardial infarction.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Kaplan–Meier analyses during 2 periods, from 30 days and from 30 days to 1 year.

From the results of the Cox proportional hazards regression analyses, we calculated the C‐statistic for the predictive value of future cardiovascular events. The C‐statistic of the variables, including body mass index, acute coronary syndrome, dyslipidemia, and peripheral arterial disease peripheral arterial disease, was 0.69 versus 0.72 when CHADS₂ score was included; the continuous NRI was 16.5% (8.7–23.4%; P<0.001). After including the CKD with these factors, we found an increase in C‐statistic from 0.72 to 0.74, and the continuous NRI was 20.8% (13.0–27.6%; P<0.001).

Discussion

The main findings of this study were as follows: (1) The CHADS₂ score was higher in relation to the severity of CAD; (2) CAD patients with higher CHADS₂ score points had a significantly higher probability of adverse cardiovascular/cerebrovascular events by the log‐rank test; (3) multivariate Cox proportional hazards analysis showed that the CHADS₂ score and presence of CKD were independent and significant predictors of clinical outcome in CAD patients; and (4) predictive model improvements by significant NRIs after adding the CHADS₂ score and CKD status to the model of other independent predictive values.

To our knowledge, this is the largest study to examine the association of CHADS₂ score with severity of CAD and with future adverse cardiovascular/cerebrovascular events in CAD patients without clinical evidence of AF. The CHADS₂ score was originally developed for risk prediction and stratification of ischemic stroke in patients with nonvalvular AF and for guiding anticoagulant therapy.1, 2, 5 The CHADS₂ score is also reportedly useful in the prediction of cardiovascular events in AF patients10 and of ischemic stroke in non‐AF patients.11, 18 Moreover, although in a relatively small population in a single‐center study, 1 report has indicated that the CHADS₂ score might predict adverse cardiovascular events in non‐AF patients with vascular dysfunction.19 These reports suggest that the CHADS₂ score has the ability to predict severe atherosclerosis and cerebrovascular and cardiovascular events in the presence or absence of AF. It would be reasonable that in this specific population in the present study, the distribution would shift to the higher CHADS₂ score category. Actually, however, we found that the distribution in the present study was similar to the 1 reported on in the original literature.14 Not surprisingly, we observed high frequencies of severe CAD (triple‐vessel or LMT) in our subjects, and it is well known that adverse cardiovascular events are more frequent in CAD patients with multivessel disease.20 Indeed, in the present study, we found that CAD patients with multivessel disease had a significantly higher CHADS₂ score than those with single‐vessel disease, and that the CHADS₂ score was a predictor of future cardiovascular/cerebrovascular events in CAD patients by 2 metrics: the C statistic derived from multivariate Cox proportional hazards models, and the NRI. Thus, our results support and demonstrate previous reports in a large population in a multicenter study. We also evaluated the occurrence of cardiovascular events during 2 periods, from discharge to 30 days and from 30 days to 1 year, and found that the impact of the CHADS₂ score on clinical outcome are clearer after 30 days within 1‐year follow‐up. Within 30 days, many other factors such as medications and intervention results might affect the outcome, and the impact of the CHADS₂ score is more significant in the chronic phase.

A recent study in high‐risk patients (CAD, ischemic stroke, and diabetes mellitus) without AF reported that the CHADS₂ score might have clinical applications for prediction of cardiovascular/cerebrovascular events, and that the CHADS₂ score was associated with other biological markers of vascular injury, such as brachial flow‐mediated dilation, carotid intimal thickness, and pulse wave velocity.19 It has been reported that impaired vascular endothelial function assessed by flow‐mediated dilation is related to the severity of CAD,21 and that endothelial dysfunction may predict cardiovascular events in patients with CAD.22 Impaired vascular endothelial function generally triggers the platelet adhesion and aggregation and fibrin formation that play a critical role in systemic hypercoagulability.23 Vascular endothelial dysfunction is associated with cardiovascular risk factors24, 25 and is 1 of the key agents of not only coronary atherosclerosis/plaque vulnerability, but also other cardiovascular complications such as vascular remodeling.26 Chan et al previously reported the significant association of the CHADS₂ score with vascular endothelial function assessed by flow‐mediated dilation in non‐AF patients.19 Even in the absence of AF, patients with heart failure, hypertension, older age, and diabetes mellitus have elevated markers of endothelial dysfunction and hypercoagulability,24, 27, 28, 29 indicating that platelet activation might be attributed to underlying risk factors other than AF. Therefore, the combined factors of the CHADS₂ score can predict adverse events in the absence of AF. In this study, we further evaluated subjects' peripheral endothelial function in 698 CAD patients without AF using a reactive hyperemia‐peripheral arterial tonometry system and found that CAD patients with higher CHADS₂ score had significantly impaired peripheral endothelial function (P<0.001, data not shown). This suggests that more‐careful observation and intensive risk reduction treatment might be needed to treat CAD patients with a high CHADS₂ score.

A recent study has reported that the R₂CHADS₂ score, in which the presence of CKD is factored into the original CHADS₂ score, improves risk stratification for stroke occurrence in AF patients.6 In our study, adding CKD status to the other predictors of clinical outcome resulted in an improvement in prognostic ability. It is well known that renal dysfunction is associated with cardiovascular events, and we previously reported that peripheral endothelial function was impaired in CKD patients and was associated with cardiovascular events.28 Renal dysfunction–induced hypertension and vascular calcification, leading to increased cardiac afterload, were reported to be associated with cardiovascular events in CKD patients partially through associated lipid disorders, oxidative stress, and abnormal levels of homocysteine and fibrinogen,30 suggesting that the R₂CHADS₂ score is more accurate than the CHADS₂ score in prediction of subsequent cardiovascular/cerebrovascular events in CAD patients. Further investigation on the newer scores, such as CHA₂DS₂‐VASc and R₂CHADS₂ scores, will be required in CAD patients without AF.