Abstract
Background: The prognostic impact of CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scores on clinical outcomes after drug-eluting stent (DES) placement has not been fully elucidated.
Methods and Results: The present study was a retrospective, non-randomized, single-center, and lesion-based study. Target lesion failure (TLF), comprising cardiac death, non-fatal myocardial infarction, and target vessel revascularization, occurred in 7.1% of 872 consecutive de novo coronary lesions in 586 patients. These patients were electively and exclusively treated by DESs from January 2016 to January 2022 until July 2022 with a mean (±SD) observational interval of 411±438 days. Multivariate Cox proportional hazard analysis revealed that CHA2DS2-VASc-HS scores ≥7 (hazard ratio [HR] 1.800; 95% CI 1.06–3.05; P=0.029) was a significant predictor of cumulative TLF among 24 variables evaluated. CHADS2 scores ≥2 (HR 3.213; 95% CI 1.32–7.80; P=0.010) and CHA2DS2-VASc scores ≥5 (HR 1.980; 95% CI 1.10–3.55; P=0.022) were also significant in the multivariate analysis. Pairwise comparisons of receiver operating characteristic curves for CHADS2 score ≥2, CHA2DS2-VASc score ≥5, and CHA2DS2-VASc-HS score ≥7 showed they were equivalent in terms of predicting the incidence of TLF, with areas under the curve of 0.568, 0.575, and 0.573, respectively.
Conclusions: All 3 cardiocerebrovascular thromboembolism risk scores were strong predictors of the incidence of cumulative mid-term TLF after elective DES placement, with cut-off values of 2, 5, and 7, respectively, and equivalent prognostic impacts.
Risk stratification for the incidence of major adverse cardiac events after drug-eluting stent (DES) placement in patients with coronary artery disease (CAD) is essential, despite the fact that many problems in the field of coronary revascularization have been resolved by advances in DESs. The Japanese Circulation Society (JCS) 2020 guidelines for antithrombotic therapy for CAD state that risk stratification is becoming an increasingly important part of the assessment of patients with CAD treated using DESs.1 There are many predictive risk scores available for use after DESs placement that take into account cardiorenal anemia syndrome (ACE score2), coronary severity (SYNTAX score3), stent thrombosis (CREDO-Kyoto risk score4), major bleeding (CREDO-Kyoto risk score,4 Academic Research Consortium for High Bleeding Risk [ARC-HBR],5,6 J-HBR1), and cardiocerebrovascular thromboembolism (CHADS2
and CHA2DS2-VASc scores7,8).
The CHADS2
and CHA2DS2-VASc scores have been widely used for the diagnosis, prognosis, and prediction of various cardiovascular diseases.7–10 The JCS/Japanese Heart Rhythm Society (JHRS) 2020 guidelines on pharmacotherapy for cardiac arrhythmias graded the CHADS2 score as a Class I recommendation for the risk stratification of cardiocerebrovascular thromboembolism in Japanese patients with atrial fibrillation instead of the CHA2DS2-VASc score (Class IIa).11 The prognostic value of the CHADS2 score for adverse cardiovascular events in CAD without atrial fibrillation has been reported previously.7 However, among the cardiocerebrovascular thromboembolism risk scores, the CHA2DS2-VASc-HS score (Table 1), comprising 10 variables by the addition of 2 risk factors (hyperlipidemia [H] and smoking [S]) to the CHA2DS2-VASc score, has been shown to be more accurate for the prediction of the severity of CAD than the CHADS2
and CHA2DS2-VASc scores.12 However, the prognostic impact of the CHA2DS2-VASc-HS score on clinical outcomes after DES placement compared with the CHADS2
and CHA2DS2-VASc scores is not fully understood. Thus, the prognostic impact of the CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scores on clinical outcomes after DES placement should be further examined by including the consistent predictors of target lesion failure (TLF) after DES placement, because their discriminatory powers seem to be unsatisfactory when these were excluded.
Table 1. Factors Constituting the CHA
2DS
2-VASc-HS Scores
| Factor |
Abbreviation |
Definition |
Points |
| Congestive heart failure |
C |
Heart failure (signs/symptoms of heart failure confirmed with
objective evidence of cardiac dysfunction |
1 |
| Hypertension |
H |
Measurements of systolic and diastolic blood pressure ≥140/90 mm
Hg or taking antihypertensive medications |
1 |
| Age |
A2 |
Age ≥75 years |
2 |
| Diabetes |
D |
Fasting blood glucose >126 mg/dL or blood glucose ≥200 mg/dL or
using antidiabetic drugs |
1 |
| Previous stroke or TIA |
S2 |
Previous ischemic stroke or TIA |
2 |
| Vascular disease |
V |
MI and peripheral artery disease including prior revascularization,
amputation, or angiographic evidence or aortic plaque |
1 |
| Age |
A |
65–74 years |
1 |
| Sex category |
Sc |
Male sex |
1 |
| Hyperlipidemia |
H |
Increased LDL-C level according to the NCEP-3 recommendations
and history of using lipid-lowering medications |
1 |
| Smoking |
S |
Smoking >10 cigarettes a day for at least 1 year without an attempt
to quit |
1 |
The maximum CHA2DS2-VASc-HS score is 11. LDL-C, low-density lipoprotein cholesterol; MI, myocardial infarction; NCEP-3, National Cholesterol Education Program 3; TIA, transient ischemic attack.
Thus, the aim of the present study was to compare the prognostic impacts of the 3 cardiocerebrovascular thromboembolism risk scores (CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS) for the incidence of cumulative mid-term TLF after successful elective DES placement in our pooled, single-center cohort.
Methods
Study Design, Population, and Percutaneous Coronary Intervention (PCI) Procedure
This was a retrospective observational cohort study that enrolled consecutive patients who underwent elective procedures for de novo coronary lesions between January 4, 2016 and January 29, 2022 at Dokkyo Medical University Saitama Medical Center. All patients provided informed consent for the collection and publication of clinical data in accordance with the Declaration of Helsinki. The study was approved by the Ethics Committee of Dokkyo Medical University Saitama Medical Center in March 2020 (Accepted no. 19105).
Of the 2,422 lesions in 1,297 patients treated with PCI over 6 years, DESs were used for 1,389 lesions. During this period, paclitaxel-coated balloon angioplasty was performed in approximately 40% of patients.13,14 Lesions were excluded in the following cases: emergency cases (n=423); non-de novo lesions (n=73); failed treatment due to the occurrence of the final slow/no-reflow phenomenon (n=20); and lesions treated with bypass grafts (n=1). Thus, the present study included 872 consecutive, successfully DES-treated lesions in 586 patients who underwent elective procedures.
All procedures were conducted under intravascular ultrasound (IVUS). The use of IVUS was defined as intravascular assessments conducted once during the PCI procedure. The following procedures were not prospectively randomized and were performed at the discretion of the physician: (1) lesion preparation and device selection for predilation using a semicompliant balloon or scoring balloon before DES placement; (2) balloon inflation time, DES dilation time per lesion (usually 2 or 3 times), and dilation pressure (mostly rated pressure; Table 2); (3) duration of dual antiplatelet therapy (DAPT), type of single antiplatelet therapy (SAPT) following DAPT, and antiplatelet therapy with oral anticoagulant agents; (4) assignment to follow-up coronary angiographies (usually planned at 9–14 months after DES placement); and (5) drugs administered for secondary prevention.
Table 2. Baseline Characteristics, Including Cardiocerebrovascular Thromboembolism Risks Overall and in TLF and Non-TLF Groups Separately
| |
Total
(n=872) |
TLF
(n=62) |
Non-TLF
(n=810) |
P value |
| Cardiocerebrovascular thromboembolism risk |
| CHADS2 score |
2.25±1.00 |
2.52±0.90 |
2.23±1.01 |
0.015 |
| CHADS2 score ≥2 (%) |
77.6 |
90.3 |
76.7 |
0.013 |
| CHADS2 score ≥3 (%) |
37.7 |
50.0 |
36.8 |
0.039 |
| CHADS2 score ≥4 (%) |
9.6 |
11.3 |
9.5 |
0.646 |
| CHA2DS2-VASc score |
4.87±1.66 |
5.11±1.52 |
4.85±1.67 |
0.188 |
| CHA2DS2-VASc score ≥4 (%) |
78.9 |
82.3 |
78.6 |
0.501 |
| CHA2DS2-VASc score ≥5 (%) |
57.1 |
71.0 |
56.0 |
0.022 |
| CHA2DS2-VASc score ≥6 (%) |
34.9 |
37.1 |
34.7 |
0.702 |
| CHA2DS2-VASc-HS score |
6.29±1.71 |
6.69±1.57 |
6.26±1.71 |
0.039 |
| CHA2DS2-VASc-HS score ≥6 (%) |
66.7 |
75.8 |
66.0 |
0.116 |
| CHA2DS2-VASc-HS score ≥7 (%) |
44.5 |
58.1 |
43.5 |
0.026 |
| CHA2DS2-VASc-HS score ≥8 (%) |
24.8 |
24.2 |
24.8 |
0.913 |
| Score componentsA |
| C (%) |
21.4 |
43.5 |
19.8 |
<0.001 |
| H (%) |
91.1 |
91.9 |
91.0 |
0.801 |
| Age (year) |
71.2±10.0 |
70.6±9.8 |
71.2±10.0 |
0.643 |
| A2 (%) |
44.6 |
43.5 |
44.7 |
0.861 |
| D (%) |
58.7 |
69.4 |
53.3 |
0.015 |
| S2 (%) |
4.4 |
1.6 |
4.6 |
0.272 |
| V (%) |
63.5 |
67.7 |
63.2 |
0.475 |
| A (%) |
31.0 |
32.3 |
30.9 |
0.819 |
| Sc (%) |
78.4 |
72.6 |
78.9 |
0.244 |
| H (%) |
98.5 |
98.4 |
98.5 |
0.934 |
| S (%) |
49.7 |
59.7 |
43.0 |
0.011 |
| Other patient characteristics |
| eGFR (mL/min/1.73 m2) |
54.7±23.1 |
46.1±26.4 |
55.3±22.7 |
0.002 |
| CKD (%) |
| Stage 5D (hemodialysis) |
8.7 |
21.0 |
7.8 |
<0.001 |
| Stage 3b to 5 exp 5D |
19.5 |
21.0 |
19.5 |
0.780 |
| Serum hemoglobin (g/dL) |
12.7±2.0 |
11.9±1.91 |
12.8±1.99 |
<0.001 |
| Anemia (%) |
47.0 |
69.4 |
45.3 |
<0.001 |
| BMI (kg/m2) |
24.0±3.9 |
22.8±4.18 |
24.1±3.85 |
0.015 |
| Low (≤18.4 kg/m2) BMI (%) |
5.3 |
14.5 |
4.6 |
<0.001 |
| Lesion characteristics (%) |
| Left anterior descending artery |
42.8 |
37.1 |
43.2 |
0.348 |
| Left circumflex artery |
19.8 |
17.7 |
20.0 |
0.667 |
| Right coronary artery |
33.7 |
38.7 |
33.3 |
0.388 |
| Left main coronary artery |
3.4 |
6.5 |
3.2 |
0.177 |
| Diffuse |
46.3 |
61.3 |
45.1 |
0.014 |
| Proximal tortuosity |
6.4 |
8.1 |
6.3 |
0.584 |
| Bending |
18.2 |
29.0 |
17.4 |
0.022 |
| Calcification |
28.3 |
50.0 |
26.7 |
<0.001 |
| Chronic total occlusion |
5.8 |
11.3 |
5.4 |
0.058 |
| Ostium |
15.6 |
19.4 |
15.3 |
0.397 |
| Bifurcation |
28.7 |
27.4 |
28.8 |
0.821 |
| Procedural characteristics |
| No. DES |
1.33±0.58 |
1.47±0.67 |
1.32±0.57 |
0.082 |
| Balloon size (mm) |
3.58±0.63 |
3.44±0.59 |
3.59±0.63 |
0.049 |
| Total DES length (mm) |
37.8±23.7 |
44.0±25.7 |
37.3±23.4 |
0.047 |
| Total DES length ≥60 mm (%) |
18.6 |
27.4 |
17.9 |
0.063 |
| Pressure (atm) |
16.1±3.3 |
15.5±3.1 |
16.1±3.3 |
0.154 |
| Debulking device use (%) |
10.2 |
22.6 |
9.3 |
<0.001 |
| Quantitative coronary angiography |
| Preprocedural MLD (mm) |
1.05±0.557 |
0.876±0.475 |
1.06±0.561 |
0.004 |
| Preprocedural %DS |
56.5±21.0 |
59.7±20.9 |
56.2±21.0 |
0.211 |
| Post-procedural MLD (mm) |
2.67±0.54 |
2.49±0.47 |
2.68±0.55 |
0.003 |
| Post-procedural %DS |
11.9±6.9 |
15.4±8.9 |
11.7±6.6 |
0.001 |
| Post-procedural %DS ≥20 (%) |
11.1 |
22.6 |
10.2 |
0.003 |
| Post-procedural RD (mm) |
3.04±0.59 |
2.96±0.54 |
3.04±0.60 |
0.272 |
| Clinical outcomes |
| Clinical observational interval (days) |
411±438 |
382±299 |
414±447 |
0.436 |
| All-cause death (%) |
4.7 |
32.3 |
2.6 |
<0.001 |
| TLF (%) |
7.1 |
100 |
0 |
1.000 |
| Cardiac death (%) |
2.2 |
30.6 |
0 |
<0.001 |
| Non-fatal MI (%) |
0.2 |
3.2 |
0 |
<0.001 |
| Angiographic TVR (%) |
4.4 |
66.1 |
0 |
<0.001 |
| Clinical TLR (%) |
1.8 |
25.8 |
0 |
<0.001 |
| Angiographic TLR (%) |
3.9 |
54.8 |
0 |
<0.001 |
Unless specified otherwise, data are given as the mean±SD or as percentages. AFor definitions of individual score components, see Table 1. BMI, body mass index; CKD, chronic kidney disease; DES, drug-eluting stent; %DS, percentage diameter stenosis; eGFR, estimated glomerular filtration rate; MI, myocardial infarction; MLD, minimum lumen diameter; RD, reference vessel diameter; TLF, target lesion failure; TLR, target lesion revascularization; TVR, target vessel revascularization.
The incidence of TLF was the main endpoint. TLF was defined as follows: (1) cardiac death, including sudden death of unknown origin; (2) non-fatal myocardial infarction (MI); and (3) the occurrences of target vessel revascularization (TVR) and target lesion revascularization (TLR), assessed during follow-up angiograms, including clinical TLR and definite stent thrombosis after DES placement. Angiographic TVR and TLR observed on follow-up angiograms were defined as elective, emergent, or repeated PCI or coronary artery bypass grafting performed for in-stent restenosis at both the 5-mm proximal and distal stented and balloon-inflated margins, and definite stent thrombosis. Clinically indicated TLR (clinical TLR) conducted prior to the planned follow-up CAG was defined as follows: (1) percentage diameter stenosis (%DS) >50 by quantitative coronary angiography (QCA; defined below) together with angina presumably due to the target vessel or objective signs of ischemia; and/or (2) TLR with %DS >70. The need for TLR was determined based on clinical symptoms, visual and quantitative angiographic outcomes, physiological examinations, and IVUS assessment. The all-cause mortality rate was calculated as an additional clinical outcome.
Quantitative Coronary Artery Angiography
QCA parameters were measured using the Goodnet Cardiovascular Network system (QAngio XA 7.3; Medis Medical Imaging Systems, Leiden, Netherlands), as described previously.13,14 Values were obtained at 2 time points: before PCI and immediately after successful PCI. Minimum lumen diameter (MLD), %DS, and reference vessel diameter (RD) were measured. For occluded lesions, %DS was defined as 100% and MLD was defined as 0 mm. All QCA parameters were measured using the T.I.
Estimated Variables
The letters in the “CHADS2” score represent congestive heart failure (C), hypertension (H), age (A), diabetes (D), and stroke (S).7,11 The CHADS2 score is calculated by assigning 1 point each for the presence of chronic heart failure, hypertension, age ≥75 years, and diabetes, and by assigning 2 points for a history of stroke or transient ischemic attack (TIA). The CHA2DS2-VASc score was established by including additional stroke risk factors to the CHADS2 score, namely vascular disease (V), age 65–74 years (A), and female sex (as a sex category [Sc]).11 In the CHA2DS2-VASc score, age ≥75 years (A2) is assigned 2 points. In the CHA2DS2-VASc-HS score, hyperlipidemia [H] and smoking [S] have been added to the CHA2DS2-VASc score, and male sex rather than female sex is used. The variables and abbreviations in the CHA2DS2-VASc-HS score are listed in Table 1. The maximum CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scores are 6, 9, and 11, respectively.
In addition to CHADS2 score ≥2, CHA2DS2-VASc score ≥5, and CHA2DS2-VASc-HS score ≥7 (cut-off values determined according to the single Cox proportional hazard model [see below]; Table 3), the following patient characteristics were measured at the index PCI: mean estimated glomerular filtration rate (eGFR), chronic kidney disease (CKD) Stage 5D (CKD5D: hemodialysis),15,16 CKD from Stage 3b to 5, excluding 5D (CKD Stage 3b to 5 exp 5D),2,17 serum hemoglobin (Hb) levels, anemia (Hb <13 g/dL in males, <12 g/dL in females),2,17 body mass index (BMI), and low BMI (BMI ≤18.4 kg/m2).18
Table 3. HRs of Baselines Characteristics for Predicting TLF Using a Single Cox Proportional Hazard Model
| |
HR |
95% CI |
P value |
| Patient characteristics |
| CHADS2 score |
1.489 |
1.18–1.88 |
0.001 |
| CHADS2 score ≥1 |
1.002 |
0.14–7.23 |
0.999 |
| CHADS2 score ≥2 |
3.783 |
1.62–8.81 |
0.002 |
| CHADS2 score ≥3 |
2.180 |
1.32–3.59 |
0.002 |
| CHADS2 score ≥4 |
1.513 |
0.69–3.32 |
0.303 |
| CHA2DS2-VASc score |
1.208 |
1.04–1.41 |
0.015 |
| CHA2DS2-VASc score ≥4 |
1.493 |
0.78–2.87 |
0.229 |
| CHA2DS2-VASc score ≥5 |
2.464 |
1.42–4.29 |
0.001 |
| CHA2DS2-VASc score ≥6 |
1.596 |
0.95–2.68 |
0.076 |
| CHA2DS2-VASc-HS score |
1.254 |
1.08–1.45 |
0.002 |
| CHA2DS2-VASc-HS score ≥6 |
1.946 |
1.09–3.48 |
0.025 |
| CHA2DS2-VASc-HS score ≥7 |
2.409 |
1.45–4.00 |
0.001 |
| CHA2DS2-VASc-HS score ≥8 |
1.510 |
0.84–2.71 |
0.167 |
| Score componentsA |
| C |
3.393 |
2.05–5.61 |
<0.001 |
| H |
1.489 |
0.60–3.72 |
0.394 |
| Age |
1.003 |
0.98–1.03 |
0.815 |
| A2 |
1.181 |
0.71–1.96 |
0.516 |
| D |
1.787 |
1.04–3.07 |
0.035 |
| S2 |
0.392 |
0.05–2.83 |
0.352 |
| V |
1.293 |
0.76–2.20 |
0.344 |
| A |
1.143 |
0.64–2.05 |
0.655 |
| Sc |
0.801 |
0.46–1.40 |
0.437 |
| H |
0.809 |
0.11–5.85 |
0.834 |
| S |
1.807 |
1.09–3.01 |
0.023 |
| eGFR |
0.982 |
0.97–0.99 |
0.001 |
| CKD |
| Stage 5D (hemodialysis) |
3.795 |
2.05–7.01 |
<0.001 |
| Stage 3b to 5 exp 5D |
1.887 |
1.14–3.13 |
0.014 |
| Serum hemoglobin |
0.800 |
0.71–0.90 |
<0.001 |
| Anemia |
3.033 |
1.77–5.21 |
<0.001 |
| BMI |
0.902 |
0.84–0.97 |
0.006 |
| Low (≤18.4 kg/m2) BMI |
3.476 |
1.71–7.06 |
0.001 |
| Lesion characteristics |
| Left anterior descending artery |
0.767 |
0.46–1.29 |
0.314 |
| Left circumflex artery |
0.898 |
0.47–1.73 |
0.747 |
| Right coronary artery |
1.259 |
0.75–2.10 |
0.378 |
| Left main coronary artery |
1.939 |
0.70–5.35 |
0.201 |
| Diffuse |
1.971 |
1.18–3.29 |
0.009 |
| Proximal tortuosity |
1.365 |
0.55–3.41 |
0.506 |
| Bending |
1.619 |
0.94–2.80 |
0.085 |
| Calcification |
2.553 |
1.55–4.20 |
<0.001 |
| Chronic total occlusion |
1.593 |
0.72–3.51 |
0.247 |
| Ostium |
1.719 |
0.91–3.23 |
0.093 |
| Bifurcation |
1.309 |
0.76–2.33 |
0.320 |
| Procedural characteristics |
| No. DES |
1.544 |
1.07–2.24 |
0.022 |
| Balloon size |
0.910 |
0.61–1.36 |
0.649 |
| Total DES length |
1.014 |
1.00–1.02 |
0.004 |
| Total DES length ≥60 mm |
1.930 |
1.10–3.38 |
0.021 |
| Pressure |
15.5±3.1 |
16.1±3.3 |
0.154 |
| Debulking device use |
3.392 |
1.86–6.17 |
<0.001 |
| Quantitative coronary angiography |
| Preprocedural MLD |
0.765 |
0.48–1.22 |
0.263 |
| Preprocedural %DS |
0.999 |
0.99–1.01 |
0.866 |
| Post-procedural MLD |
0.610 |
0.37–1.0005 |
0.050 |
| Post-procedural %DS |
1.056 |
1.03–1.09 |
<0.001 |
| Post-procedural %DS ≥20 |
2.206 |
1.22–4.00 |
0.009 |
| Post-procedural RD |
0.902 |
0.59–1.39 |
0.638 |
AFor definitions of individual score components, see Table 1. CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 2.
The following angiographic and lesion characteristics were evaluated at the index PCI: lesions located in the 4 main coronary arteries (left anterior descending artery, left circumflex artery, right coronary artery, and left main coronary artery); American College of Cardiology (ACC)/American Heart Association (AHA) lesion characteristics, diffuse lesion length (>20 mm; based on visual estimation using IVUS), and balloon length (diffuse);19 proximal tortuosity; moderate or severe calcification (calcification);20 chronic total occlusion; ostial lesions in the 4 main coronary arteries (ostium); and bifurcative lesions (bifurcation).
The following information was also collected at the time of the index PCI: total number of DES used per lesion, maximum dilated balloon diameter (balloon size), total stent length regardless of overlap, total DES length ≥60 mm,21 maximum pressure used to inflate the balloon to its maximum size, and the debulking device use.22
The QCA variables described above were recorded, and post-procedural %DS ≥2023 was added as a categorical variable.
Statistical Analyses
Baseline characteristic variables are presented as mean±SD and percentages. The parameters and clinical outcomes of the TLF group were compared with those of the non-TLF group using unpaired t-tests and χ2
or Fisher’s tests for continuous and categorical values, respectively (Table 2).
To define cut-off values for the CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scored, the effects of CHADS2 score ≥1–4, CHA2DS2-VASc score ≥4–6, and CHA2DS2-VASc-HS score ≥6–8 on the cumulative incidence of TLF were examined using the single Cox proportional hazard model. Owing to the highest significance (CHA2DS2-VASc score ≥5 and CHA2DS2-VASc-HS score ≥7) and the highest significance with a higher hazard ratio (CHADS2 score ≥2), the 3 scores were divided into 2 groups as categorical variables: CHADS2 score ≥2, CHA2DS2-VASc score ≥5, and CHA2DS2-VASc-HS score ≥7 (Tables 2–5; Figure).
Table 4. Predictors of TLF Using a Multivariate Cox Proportional Hazard Model With the 3 Scores
| |
HR |
95% CI |
P value |
| With CHADS2 score |
| CHADS2 score ≥2 |
3.213 |
1.32–7.80 |
0.010 |
| V |
1.153 |
0.67–2.00 |
0.612 |
| S |
1.945 |
1.14–3.31 |
0.014 |
| CKD Stage 5D (hemodialysis) |
2.587 |
1.17–5.70 |
0.018 |
| CKD Stage 3b to 5 exp 5D |
0.732 |
0.37–1.44 |
0.364 |
| Anemia |
2.414 |
1.32–4.42 |
0.004 |
| Low (≤18.4 kg/m2) BMI |
2.070 |
0.97–4.43 |
0.061 |
| Diffuse |
1.275 |
0.70–2.33 |
0.429 |
| Calcification |
1.418 |
0.77–2.61 |
0.263 |
| Total DES length ≥60 mm |
1.643 |
0.86–3.13 |
0.132 |
| Debulking device use |
1.559 |
0.76–3.20 |
0.226 |
| Post-procedural %DS ≥20 |
1.985 |
1.06–3.72 |
0.032 |
| With CHA2DS2-VASc score |
| CHA2DS2-VASc score ≥5 |
1.980 |
1.10–3.55 |
0.022 |
| S |
1.949 |
1.16–3.28 |
0.012 |
| CKD Stage 5D (hemodialysis) |
2.839 |
1.28–6.30 |
0.010 |
| CKD Stage 3b to 5 exp 5D |
0.754 |
0.38–1.48 |
0.414 |
| Anemia |
2.120 |
1.16–3.88 |
0.015 |
| Low (≤18.4 kg/m2) BMI |
2.304 |
1.07–4.94 |
0.032 |
| Diffuse |
1.296 |
0.71–2.36 |
0.396 |
| Calcification |
1.404 |
0.76–2.59 |
0.277 |
| Total DES length ≥60 mm |
1.589 |
0.84–3.01 |
0.156 |
| Debulking device use |
1.606 |
0.78–3.30 |
0.198 |
| Post-procedural %DS ≥20 |
2.024 |
1.08–3.78 |
0.027 |
| With CHA2DS2-VASc-HS score |
| CHA2DS2-VASc-HS score ≥7 |
1.800 |
1.06–3.05 |
0.029 |
| CKD Stage 5D (hemodialysis) |
2.403 |
1.27–4.56 |
0.007 |
| CKD Stage 3b to 5 exp 5D |
0.870 |
0.45–1.70 |
0.683 |
| Anemia |
2.111 |
1.17–3.82 |
0.014 |
| Low (≤18.4 kg/m2) BMI |
2.236 |
1.06–4.82 |
0.040 |
| Diffuse |
1.350 |
0.75–2.44 |
0.322 |
| Calcification |
1.459 |
0.79–2.69 |
0.225 |
| Total DES length ≥60 mm |
1.397 |
0.74–2.64 |
0.302 |
| Debulking device use |
1.630 |
0.79–3.34 |
0.182 |
| Post-procedural %DS ≥20 |
1.829 |
0.96–3.47 |
0.065 |
Abbreviations as in Tables 2,3.
Table 5. Pairwise Comparisons of AUCs for the 3 Scores
| Scores |
Difference in AUCs |
P value |
| CHADS2 ≥2 vs. CHA2DS2-VASc score ≥5 |
0.006 |
0.811 |
| CHADS2 ≥2 vs. CHA2DS2-VASc-HS score ≥7 |
0.005 |
0.878 |
| CHA2DS2-VASc score ≥5 vs. CHA2DS2-VASc-HS score ≥7 |
0.002 |
0.944 |
AUC, area under the curve.
To define the predictors of TLF, the effects of patient-related variables (cardiocerebrovascular thromboembolism risk and others), lesion-related variables, procedural characteristics, and QCA on cumulative incidence of TLF were examined using a single Cox proportional hazard model (Table 3). Single Cox proportional hazard models were also used to show the effects of 8 categorical variables (CHADS2 score, CHA2DS2-VASc score, CHA2DS2-VASc-HS score, eGFR, Hb, BMI, total DES length, and postprocedural %DS) when considered as continuous variables (Table 3). A multivariate Cox proportional hazard model was created by including the significant variables found in the univariate analysis (13 variables in the analysis of the CHADS2 score, 11 variables in the analysis of the CHA2DS2-VASc score, and 10 variables in the analysis of the CHA2DS2-VASc-HS score; Table 4). To simplify the constituent factors of the CHA2DS2-VASc-HS in Tables 3,4, Results, and Discussion, all the units were abbreviated, and only significant variables are described in the Results section, without describing the actual statistical results.
Receiver operating characteristic (ROC) curves for the CHADS2 score ≥2, CHA2DS2-VASc score ≥5, and CHA2DS2-VASc-HS score ≥7 are shown in the Figure. The accuracies of the 3 categorical scores on cumulative TLF were assessed by pairwise comparison of areas under the curve (AUC) using the “ROCCOMP” command in STATA (Table 5).
Statistical significance was set at P<0.05. All statistical analyses were conducted using the STATA for Windows version 16 software (StataCorp, College Station, TX, USA).
Results
Constituent Factors of the CHA2DS2-Vasc-HS Score
Table 1 lists the 10 factors making up the CHA2DS2-VASc-HS scores, with their abbreviations, definitions, and number of points allocated. The mean CHA2DS2-VASc-HS score and the percentage of patients with different CHA2DS2-VASc-HS scores and positive for different CHA2DS2-VASc-HS score components are presented in Table 2.
Baseline and Clinical Outcomes in the TLF and Non-TLF Groups
The baseline characteristics and clinical outcomes for the TLF and non-TLF groups are presented in Table 2. Of the 24 variables related to cardiocerebrovascular thromboembolism risk, 9 variables differed significantly between the 2 groups: mean CHADS2 score, percentage of patients with a CHADS2 score ≥2, percentage of patients with a CHADS2 score ≥3, mean CHA2DS2-VASc score, percentage of patients with a CHADS2-VASc score ≥5, mean CHA2DS2-VASc-HS score, percentage of patients with a CHA2DS2-VASc-HS score ≥7, and the percentage of patients with C, D, and S. Of the 7 variables not related to cardiocerebrovascular thromboembolism risk, the following 6 differed significantly between the TLF and non-TLF groups: mean eGFR, percentage of patients on hemodialysis, mean serum Hb, percentage of patients with anemia, mean BMI, and the percentage of patients with a low (≤18.4 kg/m2) BMI.
Among the 11 lesion-related characteristics, the percentage of patients with diffuse, bending, and calcified lesions differed significantly between the TLF and non-TLF groups. Of the 6 procedure-related variables, mean balloon size, total DES length, and the percentage of patients in whom a debulking device was used differed significantly between the 2 groups. Four of the 6 QCA parameters differed significantly between the TLF and non-TLF groups, namely mean preprocedural MLD, post-procedural MLD, post-procedural %DS, and the percentage of patients with a post-procedural %DS ≥20.
With regard to clinical outcomes, there was no significant difference in clinical observational interval between the TLF and non-TLF groups, but the incidence of all-cause death, cardiac death, non-fatal MI, angiographic TVR, clinical TLR, and angiographic TLR differed significantly between the 2 groups.
Predictors of TLF Determined by the Single Cox Hazard Model
The results of the single Cox proportional hazard model are presented in Table 3. Of the 24 variables related to the cardiocerebrovascular thromboembolism risk, 9 were significant: CHADS2 score, CHADS2 score ≥2, CHADS2 score ≥3, CHA2DS2-VASc score, CHA2DS2-VASc score ≥5, CHA2DS2-VASc-HS score, CHA2DS2-VASc-HS score ≥6, CHA2DS2-VASc-HS score ≥7, C, D, and S. Among the 7 patient characteristics, only cardiocerebrovascular thromboembolism risk was not significant; eGFR, hemodialysis, CKD Stage 3b to 5 exp 5D, serum Hb, anemia, BMI, and low (≤18.4 kg/m2) BMI were all significant. Two of the 11 lesion characteristic variables (diffuse and calcified lesions), 4 of the 6 procedure characteristic variables (number of DESs, total DES length, total DES length ≥60 mm, and debulking device use), and 2 of 7 QCA parameters (post-procedural %DS and post-procedural %DS ≥20) were significant in the single Cox proportional hazard model.
Predictors of TLF by Multivariate Cox Hazard Model for the 3 Scores
The results of the multivariate Cox proportional hazard model for the 3 scores are presented in Table 4. Among the 12 CHADS2 scores variable, 5 ( CHADS2 score ≥2, S, hemodialysis, anemia, BMI, and post-procedural %DS ≥20) were significant in the multivariate Cox hazard model; among the 11 CHA2DS2-VASc score variables, 6 (CHA2DS2-VASc score ≥5, S, hemodialysis, anemia, low (≤18.4 kg/m2) BMI, and post-procedural %DS ≥20) were significant; and among 10 CHA2DS2-VASc-HS score variables, 5 (CHA2DS2-VASc-HS score ≥7, S, hemodialysis, anemia, low (≤18.4 kg/m2) BMI, and post-procedural %DS ≥20) were significant.
Pairwise Comparison of the Accuracy of the 3 Scores
The Figure shows ROC curves for the 3 categorical scores with AUCs. Pairwise comparisons showed that the scores were equivalent, although the AUCs indicated mild-to-moderate accuracies of the 3 scores for the prediction of the cumulative incidence of TLF (Table 5).
Discussion
The present study was conducted to examine the prognostic impacts of 3 major cardiocerebrovascular thromboembolic risk scores for the mid-term cumulative incidence of TLF after elective DESs placement. This was based on the need for multifaceted risk stratification in patients with CAD. There are 4 major findings in this study. First, as continuous variables in univariate analysis, the CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scores were all significant predictors of the mid-term cumulative incidence of TLF after elective DESs placement (Table 3). Second, as categorical variables, cut-off predictive values of 2, 5, and 7 for the CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scores, respectively, were established for cumulative TLF on univariate and multivariate analyses (Tables 3,4). Third, all 3 categorical scores were strong predictors of the mid-term cumulative incidence of TLF after elective DESs placement on multivariate analysis with the consistent predictors of DES failure (Tables 2–4). Fourth, the prognostic impacts of CHADS2 score ≥2, CHA2DS2-VASc score ≥5, and CHA2DS2-VASc-HS score ≥7 for TLF in DES-treated lesions were statistically equivalent in pairwise comparisons of ROC curves (Table 5; Figure).
In the JCS/JHRS 2020 guidelines on pharmacotherapy for cardiac arrhythmias, the CHADS2 score, the original and simple risk score of cardiocerebrovascular thromboembolism, was defined as Class I for the risk stratification of cardiocerebrovascular thromboembolism in Japanese patients with atrial fibrillation instead of the CHA2DS2-VASc score (Class IIa).11 Therefore, a simple CHADS2 score incorporating only a few patient characteristics has been widely used in daily practice not only by primary care physicians, but also by specialized cardiologists.7–11 In contrast, the CHA2DS2-VASc-HS score, established by modifying the CHADS2
and CHA2DS2-VASc scores, is the most accurate of the 3 cardiocerebrovascular thromboembolic risk scores for the prediction of coronary severity.12 Risk stratification scores, initially developed for some risk assessments, have been proven to be accurate diagnostic and prognostic scores for major cardiovascular events in all-comer DES-treated cohorts.2–7 Therefore, we retrospectively compared the prognostic impact of the CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS scores for DES-treated lesions. All 3 scores were significant predictors of cumulative TLF, showing high significance when considered as either continuous or categorical variables (Tables 3,4). The accuracy of the prognostic impact of CHADS2 score ≥2, CHA2DS2-VASc score ≥5, and CHA2DS2-VASc-HS score ≥7 for DES-treated lesions was statistically equivalent (Table 5), although the respective predictive accuracy for the incidence of TLF was considered mild to moderate, with AUCs of approximately 0.56–0.58 (Figure). On multivariate analyses, the hazard ratio of the CHADS2 score ≥2 was approximately 1.6- to 1.8-fold higher than that of the CHA2DS2-VASc score ≥5 and CHA2DS2-VASc-HS score ≥7 (Table 4). Thus, we propose that a CHADS2 score≥2 could be used as a simple prognostic cardiocerebrovascular thromboembolic risk score not only for cardiocerebrovascular thromboembolism risk in patients with atrial fibrillation,11 but also for cumulative TLF risk in patients with CAD treated with advanced DESs in Japan.
The present study is the first to show the CHA2DS2-VASc-HS score, among 24 PCI-related variables, as a strong predictor of the cumulative mid-term incidence of TLF after elective DESs placement with a cut-off value of 7 (Tables 3,4). In the present cohort, all 4 significant variables related to patient characteristics receiving coronary DESs. A CHA2DS2-VASc-HS score ≥7 remained significant compared with variables that were significant in univariate analysis, namely CKD Stage 3b to 5 exp 5D,17 calcified lesion,20 diffuse lesion,19 use of a debulking device,22 total DES length ≥60 mm,21 and post-procedural %DS ≥20.23 All these variables are consistent predictors of DES failure in the initial and late phases. Through the 3 analyses with individual scores, hemodialysis (CKD5D)15,16 and anemia2,17 were the common predictors of TLF (Table 4), overlapping various risk scores.2,4–6 Thus, although the accuracy of the 3 categorical scores for the prediction of TLF by logistic regression analyses was mild to moderate, the greater significance of the 3 variables for the cumulative mid-term incidence of TLF in multivariate Cox proportional hazard models showed the validity of the 3 scores for the prediction of TLF in patients with CAD treated with the recent advanced DESs. The cumulative incidence of TLF slightly but gradually increases after DESs placement because of late TLR and even later stent thrombosis.16,24 Predictors of late TLF are known to be mostly similar to those of short- to mid-term TLF.24 Thus, all 3 scores (CHADS2
≥2, CHA2DS2-VASc ≥5, and CHA2DS2-VASc-HS ≥7) would be valid for predicting the long-term incidence of TLF. The accuracy of risk stratification after DESs placement among cardiocerebrovascular thromboembolic risk score families, such as CHADS2, CHA2DS2-VA,25 CHA2DS2-VASc, CHA2DS2-VASc-HS, and CHA2DS2-VASc-HSF, needs to be examined in prospective multicenter cohorts.
This study has several limitations. The study was a retrospective, non-randomized, single-center, lesion-based, small cohort study. Patient-based analyses to predict long-term outcomes could not be fully estimated. The validity of these cardiocerebrovascular thromboembolic risk scores needs to be ascertained in multicenter cohorts and in large-scale prospective analyses by comparing them with the widely used scores.2–7 Owing to the small cohort size, we could not further examine their prognostic impact by dividing them into tertiles or quartiles. As mentioned above, because the accuracy of the 3 scores for the prediction of TLF was mild to moderate (Table 5; Figure), their validity for TLF needs to be confirmed by establishing a new risk score with the significant variables presented in Table 4. However, owing to the small cohort, a new predictive score could not be explored, as done previously.26 Although all procedures were conducted under IVUS guidance, intravascular findings related to long-term outcomes were not included. The regimens and durations of SAPT and DAPT after DES placement, which are closely related to major bleeding,2–7 have also not been fully estimated.
In conclusion, all 3 cardiocerebrovascular thromboembolism risk scores (i.e., CHADS2, CHA2DS2-VASc, and CHA2DS2-VASc-HS) were strong, statistically equivalent predictors of the cumulative mid-term incidence of TLF after elective DESs placement, with cut-off values of 2, 5, and 7, respectively. Thus, a CHADS2 score ≥2 could be widely used as a simple prognostic cardiocerebrovascular thromboembolic risk score not only for cardiocerebrovascular thromboembolism risk in patients with atrial fibrillation,11 but also for the cumulative risk of TLF in patients with CAD treated with DESs in Japan.
Sources of Funding
This study was supported by funding from a Dokkyo Medical University Young Investigator Award (No. 2021-18) to T.U.
Disclosures
I.T. is a member of Circulation Reports’ Editorial Team. The remaining authors have no conflicts of interest to declare.
IRB Information
This study was approved by the Ethics Committee of Dokkyo Medical University Saitama Medical Center in March 2020 (Accepted no. 19105).
Data Availability
Deidentified participant data will not be shared.
References
- 1.
Nakamura M, Kimura K, Kimura T, Ishihara M, Otsuka F, Kozuma K, et al. JCS 2020 guideline focused update on antithrombotic therapy in patients with coronary artery disease. Circ J 2020; 84: 831–865.
- 2.
Wykrzykowska JJ, Garg S, Onuma Y, de Vries T, Goedhart D, Morel MA, et al. Value of age, creatinine, and ejection fraction (ACEF score) in assessing risk in patients undergoing percutaneous coronary interventions in the “all-comers” LEADERS trial. Circ Cardiovasc Interv 2011; 4: 47–56.
- 3.
Wykrzykowska J, Garg S, Girasis C, de Vries T, Morel MA, van Es GA, et al. Value of the SYNTAX score (SX) for risk assessment in the “all-comers” population of the randomized multicenter LEADERS trial. J Am Coll Cardiol 2010; 56: 272–277.
- 4.
Natsuaki M, Morimoto T, Yamaji K, Watanabe H, Yoshikawa Y, Shiomi H, et al. Prediction of thrombotic and bleeding events after percutaneous coronary intervention: CREDO-Kyoto thrombotic and bleeding risk scores. J Am Heart Assoc 2018; 7: e008708.
- 5.
Urban P, Mehran R, Colleran R, Angiolillo DJ, Byrne RA, Capodanno D, et al. Defining high bleeding risk in patients undergoing percutaneous coronary intervention. A consensus document from the Academic Research Consortium for High Bleeding Risk. Circulation 2019; 140: 240–261.
- 6.
Okabe K, Miura K, Shima Y, Ikuta A, Taguchi Y, Takahashi K, et al. Comparison and validation of long-term bleeding events for academic bleeding risk (ARC-HBR) criteria and contemporary risk scores for percutaneous coronary intervention with a second-generation drug eluting stent. Circ J 2022; 86: 1379–1387.
- 7.
Tabata N, Yamamoto E, Hokimoto S, Yamashita T, Sueta D, Takashio S, et al. Prognostic value of the CHADS2 score for adverse cardiovascular events in coronary artery disease patients without atrial fibrillation: A multi-center observational cohort study. J Am Heart Assoc 2017; 6: e006355.
- 8.
Pilgrim T, Kalesan B, Zanchin T, Pulver C, Jung S, Mattle H, et al. Impact of atrial fibrillation on clinical outcomes among patients with coronary artery disease undergoing revascularisation with drug-eluting stents. EuroIntervention 2013; 8: 1061–1071.
- 9.
Akao M, Ogawa H, Masunaga N, Minami K, Ishigami K, Ikeda S, et al. 10-year trends of antithrombotic therapy status and outcomes in Japanese atrial fibrillation patients: The Fushimi AF Registry. Circ J 2022; 86: 726–736.
- 10.
Ide Y, Ogawa H, Ishigami K, Ikeda S, Doi K, Hamatani Y, et al. Clinical characteristics and outcomes of very elderly patients with atrial fibrillation at high bleeding risk: The Fushimi AF Registry. Circ Rep 2021; 3: 629–638.
- 11.
Ono K, Iwasaki YK, Akao M, Ikeda T, Ishii K, Inden Y, et al. JCS/JHRS 2020 guideline on pharmacotherapy of cardiac arrhythmias. Circ J 2022; 86: 1790–1924.
- 12.
Cetin M, Cakici M, Zencir C, Tasolar H, Baysal E, Balli M, et al. Prediction of coronary artery disease severity using CHADS2 and CHA2DS2-VASc scores and a newly defined CHA2DS2-VASc-HS score. Am J Cardiol 2014; 113: 950–956.
- 13.
Mizutani Y, Ishikawa T, Nakamura H, Yamada K, Ukaji T, Kondo Y, et al. A propensity score-matched comparison of midterm outcomes between drug-coated balloons and drug-eluting stents for patients with acute coronary syndrome: A single-center study. Int Heart J 2022; 63: 217–225.
- 14.
Yamada K, Ishikawa T, Nakamura H, Mizutani Y, Ukaji T, Shimura M, et al. Midterm safety and efficacy of elective drug-coated balloon angioplasty in comparison to drug-eluting stents for unrestrictive de novo coronary lesions: A single center retrospective study. J Cardiol 2022, doi:10.1016/j.jjcc.2022.11.014.
- 15.
Okuno S, Ishihara T, Iida O, Asai M, Masuda M, Okamoto S, et al. Two-year clinical outcomes of biodegradable polymer vs. durable polymer drug-eluting stent implantation in patients with end-stage renal disease on dialysis. Circ Rep 2020; 3: 18–25.
- 16.
Aoki J, Kozuma K, Awata M, Nanasato M, Shiode N, Tanabe K, et al. Five-year clinical outcomes of everolimus-eluting stents from the post marketing study of CoCr-EES (XIENCE V/PROMUS) in Japan. Cardiovasc Interv Ther 2019; 34: 40–46.
- 17.
Landes U, Kornowski R, Assali A, Vaknin-Assa H, Greenberg G, Lev EI, et al. Predictors of long term outcomes in 11,441 consecutive patients following percutaneous coronary interventions. Am J Cardiol 2015; 115: 855–859.
- 18.
Won KB, Shin ES, Kang J, Yang HM, Park KW, Han KR, et al. Body mass index and major adverse events during chronic antiplatelet monotherapy after percutaneous coronary intervention with drug-eluting stents: Results from the HOST-EXAM Trial. Circ J 2023; 87: 268–276, doi:10.1253/circj.CJ-22-0344.
- 19.
Claessen BW, Smits PC, Kereiakes DJ, Parise H, Fahy M, Kedhi E, et al. Impact of lesion length and vessel size on clinical outcomes after percutaneous coronary intervention with everolimus- versus paclitaxel-eluting stents pooled analysis from the SPIRIT (Clinical Evaluation of the XIENCE V Everolimus Eluting Coronary Stent System) and COMPARE (Second-generation everolimus-eluting and paclitaxel-eluting stents in real-life practice) randomized trials. JACC Cardiovasc Interv 2011; 4: 1209–1215.
- 20.
Shiode N, Kozuma K, Aoki J, Awata M, Nanasato M, Tanabe K, et al. The impact of coronary calcification on angiographic and 3-year clinical outcomes of everolimus-eluting stents: Results of a XIENCE V/PROMUS post-marketing surveillance study. Cardiovasc Interv Ther 2018; 33: 313–320.
- 21.
Sharp AS, Latib A, Ielasi A, Larosa C, Godino C, Saolini M, et al. Long-term follow-up on a large cohort of “full-metal jacket” percutaneous coronary intervention procedures. Circ Cardiovasc Interv 2009; 2: 416–422.
- 22.
Sakakura K, Ito Y, Shibata Y, Okamura A, Kashima Y, Nakamura S, et al. Clinical expert consensus document on rotational atherectomy from the Japanese Association of Cardiovascular Intervention and Therapeutics. Cardiovasc Interv Ther 2021; 36: 1–18.
- 23.
Watanabe H, Morimoto T, Shiomi H, Natsuaki M, Kawai K, Kozuma K, et al. Impact of angiographic residual stenosis on clinical outcomes after new-generation drug-eluting stents implantation: Insights from a pooled analysis of the RESET and NEXT trials. J Am Heart Assoc 2018; 7: e008718.
- 24.
Kimura T, Morimoto T, Nakagawa Y, Kawai K, Miyazaki S, Muramatsu T, et al. Very late stent thrombosis and late target lesion revascularization after sirolimus-eluting stent implantation: Five-year outcome of the j-Cypher Registry. Circulation 2012; 125: 584–591.
- 25.
Choi SY, Kim MH, Kim HB, Kang SY, Lee KM, Hyun KY, et al. Validation of the CHA2DS2-VA score (excluding female sex) in nonvalvular atrial fibrillation patients: A nationwide population-based study. J Clin Med 2022; 11: 1823.
- 26.
Gohbara M, Iwahashi N, Okada K, Ogino Y, Hanajima Y, Kirigaya J, et al. A simple risk score to differentiate between coronary artery obstruction and coronary artery spasm of patients with acute coronary syndrome without persistent ST-segment elevation. Circ J 2022; 86: 1509–1518.
