Abstract
Background: The Academic Research Consortium for High Bleeding Risk (ARC-HBR) defined a consensus clinical criterion for patients at HBR undergoing percutaneous coronary intervention (PCI). This study aimed to validate and compare the ARC-HBR criteria and the contemporary risk score for long-term bleeding outcomes using a cohort of patients undergoing PCI.
Methods and Results: This study analyzed 3,410 patients who underwent PCI between 2010 and 2013. The endpoint was defined as incidence of The Bleeding Academic Research Consortium 3 or 5 bleeding events. In addition to ARC-HBR, this study validated the predictability of the Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy (PRECISE-DAPT) score, Patterns of non-adherence to Anti-platelet Regimens In Stented patients (PARIS) bleeding score, and Coronary Revascularization Demonstrating Outcome Study in Kyoto (CREDO-Kyoto) bleeding scores for bleeding events. There was a trend toward an increase in bleeding events, as the risk score increased for all bleeding risk scores used in this study. The ARC-HBR criteria had higher diagnostic sensitivity for bleeding events than other bleeding risk scores.
Conclusions: Patients with a higher number of risk factors in each of the four bleeding risk scores had a higher risk of long-term bleeding events. In comparison to other contemporary risk scores, the ARC-HBR criteria were more sensitive in the identification of patients with bleeding events in the long-term.
Several studies have shown that dual antiplatelet therapy (DAPT) effectively reduced the ischemic events after percutaneous coronary intervention (PCI).1,2 Conversely, it was shown that prolonging the duration of DAPT increased the risk of major bleeding events after PCI.3–5 Therefore, for predicting bleeding events and determining the duration of DAPT, researchers proposed several risk scores such as the Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy (PRECISE-DAPT) score,6 Patterns of non-adherence to Anti-platelet Regimens In Stented patients (PARIS) bleeding risk score7 or Coronary Revascularization Demonstrating Outcome Study in Kyoto (CREDO-Kyoto) bleeding score.8 Furthermore, the Academic Research Consortium for High Bleeding Risk (ARC-HBR) recently provided a consensus document for the classification of high bleeding risk (HBR).9 Previous studies have shown that these risk scores were useful in the identification of bleeding events.10–12 One previous study revealed that the ARC-HBR criteria were more sensitive than the contemporary risk score for predicting bleeding events within 1 year,13 albeit the predictability of bleeding events in the long-term was unknown. Thus, we aimed to validate the ARC-HBR criteria and compare it with the contemporary risk score for long-term bleeding outcomes using a cohort of patients undergoing PCI with a second-generation drug-eluting stent.
Methods
Study Population
This was a retrospective single-center study involving 3,881 patients who underwent PCI using a second-generation drug-eluting stent in our hospital (Kurashiki Central Hospital) between January 2010 and December 2013. If multiple PCIs were performed on the same patient, the first PCI in the study period was included in the analysis. As a result, 3,453 patients were included in the analysis. After excluding 43 patients who lacked data for calculating risk scores, we enrolled 3,410 patients exclusively treated with second-generation drug-eluting stents (Endeavor; Medtronic Vascular, Santa Rosa, CA, USA; XIENCE V, Abbott Vascular, Santa Clara, CA, USA; NOBORI, Terumo, Tokyo, Japan; PROMUS, Boston Scientific, Natick, MA, USA; or Resolute Integrity, Medtronic Vascular, Santa Rosa, CA, USA) (Supplementary Figure 1). All therapies and data handling were performed in accordance with the Declaration of Helsinki. Informed consent was given by all patients. Consent for data collection and analysis for research purposes was ensured by opt-out methods, and the appropriate institutional ethics committee approved the study.
Data Collection
The baseline data were collected from medical records. Clinical follow-up data were collected from medical data and information from the primary care physician or patient by letter or telephone survey.
Antiplatelet Therapy
All patients were pretreated with aspirin (100 mg daily) and ticlopidine (200 mg daily)/clopidogrel (75 mg daily). In our center, patients continued dual antiplatelet therapy (DAPT) until the 8-month routine follow-up coronary angiography, although each physician ultimately decided the duration of DAPT. The status of antiplatelet therapy was evaluated throughout the follow-up period, and persistent discontinuation of DAPT was defined as the withdrawal of aspirin or thienopyridine for at least 2 office visits or telephone inquiries.
Study Definition and Clinical Outcomes
Patients with HBR were defined as those who met at least 1 major or 2 minor criteria of the ARC-HBR. The ARC-HBR score was calculated by adding 1 point for any major criterion and 0.5 points for any minor criterion.13 Other contemporary scores were calculated by individual definitions of risk scores. The PRECISE-DAPT score was classified into 4 groups: very low: 0–10; low: 11–17; moderate: 18–24; and high: ≥25.8 The PARIS bleeding score was classified into 3 groups: low: 0–3; intermediate: 4–7; and high: ≥8).9 The CREDO-Kyoto bleeding score was classified into 3 groups: low: 0; intermediate: 1–2; and high: ≥3).10 The content of each bleeding score is shown in Supplementary Table. The endpoint of this study was defined as Bleeding Academic Research Consortium (BARC) type 3 or 5 bleeding14 events within 7 years. We defined drug-eluting stents other than Cypher (Cordis, Miami Lakes, FL, USA) and TAXUS (Boston Scientific, Natick, MN, USA) as second-generation drug-eluting stents.
Statistical Analysis
Continuous variables are expressed as mean±standard deviation and were compared using the Student’s t-test. Categorical variables were compared with the chi-squared test. Kaplan-Meier cumulative event curves were constructed for time-to-event variables and compared using the log-rank test. The predictability of the endpoints of each risk score were estimated by comparing the areas under the curve (AUC) of the receiver operating characteristic by C-statistics. In univariate analysis, the Cox regression model was used. In univariate analysis, risk factors of the ARC-HBR criteria and other contemporary bleeding risk scores, which were >0.1% frequency, were selected. Continuous variables were dichotomised at the median. We used IBM SPSS Statistics 23 (International Business Machines Corp., Armonk, NY, USA) for all statistical analyses.
Results
Baseline Characteristics
A study flow chart is shown in Supplementary Figure 1. In total, 3,410 patients were enrolled in the final analysis. The mean follow-up period was 2,757±590 days. The follow-up rate was as follows: after 1 year, 99.2%; after 2 years, 98.3%; and after 3 years, 97.0%. The clinical and procedural characteristics are shown in Table 1. The patients with bleeding events were older, had higher prevalence of chronic kidney disease (CKD), prior coronary artery bypass graft (CABG), and usage of DAPT and OAC. The prevalence of a high score in the 4 bleeding risk score categories was as follows: ARC-HBR score high (≥1): 52.9%; PRECISE-DAPT score high (≥25): 44.1%; PARIS bleeding score high (≥8): 32.6%; and CREDO-Kyoto bleeding score high (≥3): 15.1%. In comparison to other contemporary bleeding risk scores, more patients were classified as high risk using the ARC-HBR definition (vs. PRECISE-DAPT score, P<0.001; vs. Paris bleeding score, P<0.001; and vs. CREDO-Kyoto bleeding score, P<0.001).
Table 1. Baseline Patient Characteristics
| |
Overall
(N=3,410) |
Bleeding (−)
(N=3,154) |
Bleeding (+)
(N=256) |
P value |
| Patient characteristics |
| Age (years) |
70.0±11.0 |
69.8±11.1 |
72.4±9.98 |
<0.001 |
| Male sex |
2,553 (74.9) |
2,371 (75.2) |
182 (71.1) |
0.148 |
| Body mass index, kg/m2 |
24.0±3.6 |
24.1±3.6 |
22.9±3.5 |
<0.001 |
| Hypertension |
2,452 (71.9) |
2,265 (71.8) |
187 (73.0) |
0.673 |
| Diabetes mellitus |
1,356 (39.8) |
1,255 (39.8) |
101 (39.5) |
0.915 |
| Insulin-treated diabetes |
336 (9.9) |
308 (9.8) |
28 (10.9) |
0.545 |
| Dyslipidemia |
2,077 (60.9) |
1,935 (61.4) |
142 (55.5) |
0.064 |
| History of smoking |
517 (15.2) |
483 (15.3) |
34 (13.3) |
0.383 |
| eGFR (mL/min/1.73 m2) |
62.0±26.4 |
62.5±26.3 |
55.8±26.9 |
<0.001 |
| eGFR ≥30, <60 |
1,171 (34.3) |
1,069 (33.9) |
102 (39.8) |
0.054 |
| eGFR <30 |
362 (10.6) |
320 (10.1) |
42 (16.4) |
0.002 |
| Hemodialysis |
189 (5.5) |
168 (5.3) |
21 (8.2) |
<0.001 |
| Previous PCI |
1,250 (36.7) |
1,157 (36.7) |
93 (36.3) |
0.910 |
| Previous MI |
1,143 (33.5) |
1,048 (33.2) |
95 (37.1) |
0.206 |
| Previous CABG |
151 (4.4) |
131 (4.2) |
20 (7.8) |
0.006 |
| Peripheral arterial disease |
252 (7.4) |
232 (7.4) |
20 (7.8) |
0.788 |
| Clinical diagnosis |
| Chronic coronary syndrome |
2,097 (61.5) |
1,949 (61.8) |
148 (57.8) |
0.208 |
| Acute coronary syndrome |
– |
– |
– |
0.348 |
| Unstable angina |
413 (12.1) |
375 (11.9) |
38 (14.8) |
|
| NSTEMI |
206 (6.0) |
194 (6.2) |
12 (4.7) |
|
| STEMI |
694 (20.4) |
636 (20.2) |
58 (22.7) |
|
| Access site, n (%) |
|
|
|
0.446 |
| Radial |
967 (28.4) |
902 (28.6) |
65 (25.4) |
|
| Brachial |
457 (13.4) |
418 (15.4) |
39 (11.2) |
|
| Femoral |
1,986 (58.2) |
1,834 (58.1) |
152 (59.4) |
|
| Target lesion |
|
|
|
0.106 |
| Left main trunk |
527 (15.5) |
474 (15.0) |
53 (20.7) |
|
| Left anterior descending |
1,954 (57.3) |
1,804 (57.2) |
150 (58.6) |
|
| Left circumflex |
979 (28.7) |
910 (28.9) |
69 (27.0) |
|
| Right coronary artery |
1,512 (44.3) |
1,410 (44.7) |
102 (39.8) |
|
| Graft |
13 (0.4) |
13 (0.6) |
0 (0) |
|
| Lesion type |
| Ostium lesion |
656 (19.2) |
589 (18.7) |
67 (26.2) |
0.003 |
| In-stent restenosis |
442 (13.0) |
415 (13.2) |
27 (10.5) |
0.232 |
| Chronic total occlusion |
551 (16.2) |
517 (16.4) |
34 (13.3) |
0.193 |
| Bifurcation |
1,765 (51.8) |
1,632 (51.7) |
133 (52.0) |
0.949 |
| ACC/AHA B2 or C |
2,756 (80.8) |
2,550 (80.8) |
206 (80.5) |
0.882 |
| Procedural |
| Rotablator use |
150 (4.4) |
138 (4.4) |
12 (4.7) |
0.815 |
| IVUS assessment |
1,674 (49.1) |
1,552 (49.2) |
122 (47.7) |
0.633 |
| Medication at discharge |
| Aspirin |
3,391 (99.4) |
3,136 (99.4) |
255 (99.6) |
0.710 |
| Thienopyridine |
3,396 (99.6) |
3,142 (99.6) |
254 (99.2) |
0.335 |
| DAPT |
3,379 (99.1) |
1,782 (98.7) |
1,597 (99.5) |
0.017 |
| OAC |
410 (12.0) |
352 (11.2) |
58 (22.7) |
<0.001 |
| DAPT+OAC |
401 (11.8) |
344 (22.3) |
57 (10.9) |
<0.001 |
| Risk score |
| ARC-HBR |
1,805 (52.9) |
1,620 (51.4) |
185 (72.3) |
<0.001 |
| PRECISE-DAPT score |
23.6±14.3 |
23.1±14.2 |
29.3±14.8 |
<0.001 |
| PRECISE-DAPT score ≥25 |
1,504 (44.1) |
1,350 (42.8) |
154 (60.2) |
<0.001 |
| PARIS bleeding score |
6.6±2.7 |
5.9±3.0 |
7.2±3.0 |
<0.001 |
| PARIS bleeding score ≥8 |
1,111 (32.6) |
992 (31.5) |
119 (46.5) |
<0.001 |
| CREDO-Kyoto score |
1.1±1.4 |
1.0±1.3 |
1.4±1.6 |
<0.001 |
| CREDO-Kyoto score ≥3 |
514 (15.1) |
454 (14.4) |
60 (23.4) |
<0.001 |
Data are expressed as numbers (%) or mean±standard deviation. ACC, American College of Cardiology; AHA, American Hospital Association; ARC-HBR, Academic Research Consortium for High Bleeding Risk; CABG, coronary artery bypass grafting; CAD, coronary artery disease; CREDO-Kyoto, Coronary Revascularization Demonstrating Outcome Study in Kyoto; DAPT, dual antiplatelet therapy; eGFR, estimated glomerular filtration rate; IVUS, intravascular ultrasound; MI, myocardial infarction; NSTEMI, non-ST elevation myocardial infarction; OAC, oral anticoagulant; PARIS, Patterns of non-adherence to Anti-platelet Regimens In Stented patients; PCI, percutaneous coronary intervention; PRECISE-DAPT, Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy; STEMI, ST elevation myocardial infarction.
The prevalence of ARC-HBR is shown in Supplementary Figure 2. There was a high frequency of moderate CKD (n=1,171: 34.3%), older age ≥75 (n=1,164: 34.1%), moderate anemia (n=848: 24.9%), severe anemia (n=580: 17.0%), oral anticoagulation (n=410: 12.0%), and severe CKD (n=362: 10.6%). Chronic bleeding diathesis (n=1: 0.03%) and major surgery on DAPT (n=1: 0.03%) were rarely observed. The prevalence of other contemporary bleeding scores is shown in Supplementary Figure 2. There was a high frequency of body mass index (BMI) <25 (n=2,172: 63.7%), prior myocardial infarction (n=1,143: 33.5%), and current smoking (n=517: 15.2%).
Clinical Outcomes
The cumulative incidence of bleeding endpoint for the ARC-HBR score and other contemporary bleeding risk scores are shown in Figure 1. In any bleeding risk score, the cumulative incidence of BARC 3 or 5 bleeding tended to increase as the score increased (Figure 1). The Kaplan-Meier curves of the ARC-HBR score and other contemporary bleeding risk scores for bleeding endpoint are shown in Figure 2. Supplementary Figure 3 shows the DAPT duration for each risk score. The CREDO-Kyoto score became higher as the DAPT period became shorter. The same trend was not seen in the other risk scores.
Predictor of BARC 3 or 5 Bleeding
In a univariate analysis, 15 criteria of ARC-HBR, excluding chronic bleeding diathesis and planned non-deferrable, non-cardiac major surgery on DAPT because of low frequency, and 8 criteria of other contemporary bleeding risk scores (total number of white blood cells, BMI <25, ≥35, current smoker, triple antithrombotic therapy, heart failure, peripheral artery disease, prior myocardial infarction, and atrial fibrillation) were used to analyse the predictors of BARC 3 or 5 bleeding (Figure 3). The 7-year post-procedural predictors for BARC 3 or 5 bleeding were 13 factors (severe and moderate CKD, severe and moderate anemia, thrombocytopenia, age ≥75 years, oral anticoagulation, greater WBC, BMI <25, triple antithrombotic therapy, heart failure, and atrial fibrillation). Of these, 8 risk factors were included in the ARC-HBR (Figure 3, Table 2). The risk factors that overlap in many bleeding scores were more likely to be associated with the risk of bleeding.
Table 2. Predictor of 7-Year BARC 3 or 5 Bleeding (Univariate Analysis)
| |
Incidence of bleeding |
Univariate analysis |
Bleeding (−)
(N=3,107) |
Bleeding (+)
(N=303) |
Hazard ratio |
P value |
| ARC-HBR major criteria |
| Oral anticoagulation |
343 (11.0) |
67 (22.1) |
2.29 |
<0.01 |
| Severe CKD |
315 (10.5) |
47 (15.5) |
1.63 |
<0.01 |
| Severe anemia |
503 (16.2) |
77 (25.4) |
1.76 |
<0.01 |
| Bleeding or transfusion major |
26 (0.8) |
5 (1.7) |
1.99 |
0.19 |
| Thrombocytopenia |
67 (2.2) |
14 (4.6) |
2.20 |
<0.01 |
| Liver cirrhosis |
8 (0.3) |
2 (0.7) |
2.58 |
0.22 |
| Active malignancy |
67 (2.2) |
5 (1.7) |
0.76 |
0.56 |
| Stroke major |
60 (1.9) |
7 (2.3) |
1.20 |
0.65 |
| Prior trauma or surgery |
29 (0.9) |
2 (0.7) |
0.71 |
1.00 |
| ARC-HBR minor criteria |
| Age ≥75 years |
10.37 (33.4) |
127 (41.9) |
1.44 |
<0.01 |
| Moderate CKD |
1,049 (33.8) |
122 (40.3) |
1.32 |
0.02 |
| Moderate anemia |
754 (24.3) |
94 (31.09) |
1.40 |
<0.01 |
| Bleeding or transfusion minor |
11 (0.4) |
2 (0.7) |
1.87 |
0.41 |
| Use of oral NSAIDs or steroid |
130 (4.2) |
14 (4.6) |
1.11 |
0.72 |
| Stroke minor |
268 (8.6) |
31 (10.2) |
1.21 |
0.35 |
| Other bleeding risk score factors |
| Total number of WBC |
1,325 (42.6) |
107 (35.3) |
0.77 |
<0.01 |
| BMI <25 |
1,957 (63.0) |
215 (71.0) |
1.44 |
<0.01 |
| BMI >35 |
23 (0.7) |
1 (0.3) |
0.44 |
0.72 |
| Current smoker |
476 (15.3) |
41 (13.5) |
0.87 |
0.41 |
| TAPT |
335 (10.8) |
66 (21.8) |
2.42 |
<0.01 |
| Heart failure |
298 (9.6) |
38 (12.5) |
1.35 |
0.10 |
| Peripheral artery disease |
228 (7.3) |
24 (7.99) |
1.09 |
0.71 |
| Prior MI |
10.36 (33.3) |
107 (35.3) |
1.09 |
0.49 |
| Atrial fibrillation |
272 (8.8) |
42 (13.9) |
1.72 |
<0.01 |
Data are expressed as numbers (%) or mean±standard deviation. BMI, body mass index; CKD, chronic kidney disease; NSAIDs, non-steroidal anti-inflammatory drugs; TAPT, triple anti-thrombotic therapy; WBC, white blood cell. Other abbreviations as in Table 1.
The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the ARC-HBR positive for BARC 3 or 5 bleeding at 7 years were 72.9%, 49.0%, 12.2%, 94.9%, and 51.1%, respectively (Table 3). ARC-HBR had a higher sensitivity and lower specificity compared with other contemporary bleeding risk scores (PRECISE-DAPT score, PARIS bleeding score, and CREDO-Kyoto bleeding score).
Table 3. Diagnostic Ability of Bleeding Prediction Systems to Detect 7-Year BARC 3 or 5 Bleeding
| |
Sn (%) |
Sp (%) |
PPV (%) |
NPV (%) |
Accuracy (%) |
| ARC-HBR (≥1) |
72.9 |
49.0 |
12.2 |
94.9 |
51.1 |
| PRECISE-DAPT score (≥25) |
57.4 |
57.2 |
11.6 |
93.2 |
57.2 |
| PARIS bleeding score (≥8) |
44.9 |
68.6 |
12.2 |
92.7 |
66.5 |
| CREDO-Kyoto bleeding score (≥3) |
22.1 |
85.6 |
13.0 |
91.9 |
80.0 |
BARC, The Bleeding Academic Research Consortium; NPV, negative positive value; PPV, positive predictive value; Sn, Sensitivity; Sp, Specificity. Other abbreviations as in Table 1.
Comparisons of the performance of bleeding risk models are shown in Figure 4. The predictability of ARC-HBR for the primary endpoint (AUC: 0.63) was equivalent to or higher than that of other contemporary bleeding risk scores (PRECISE-DAPT score: 0.61; P=0.17; PARIS bleeding score: 0.61; P=0.11; CREDO-Kyoto bleeding score: 0.56; P<0.01).
Discussion
The main findings of the present study were as follows. First, the 7-year cumulative incidence of BARC 3 or 5 bleeding events for the ARC-HBR score and other contemporary bleeding risk scores tended to increase as each bleeding risk score increased. Second, risk factors that overlapped with multiple risk scores tended to be more predictive of bleeding events. Third, the ARC-HBR criteria had a high sensitivity for the prediction of bleeding endpoints at 7 years.
Previous studies have shown that by stratifying the scores, bleeding risk scores are useful in predicting bleeding events.8–10,12,15 As for the ARC-HBR criteria, it has been shown that stratification among HBR patients can also stratify the bleeding risk.13,16,17 Therefore, in this study, the ARC-HBR score was determined according to the Bern registry.13 There was a trend toward an increase in bleeding events as the risk score increased for the bleeding risk scores used in this study. This trend has been shown in short-term results of previous studies, but it was also shown to be same for long-term bleeding events.8–10,13,15,16 The ARC-HBR criteria and other risk scores were found to be even more useful in predicting long-term bleeding events when stratification occurred within each score.
Each bleeding risk score had a specific criterion and factor for evaluating bleeding risk.8–11 Among the risk factors for the bleeding risk scores used in the present study, there was no overlapped risk factor for all risk scores, except for severe CKD. Although there were several different risk factors of each bleeding risk scores, determining the most important predictor for bleeding events was controversial. In the present study, not only factors of ARC-HBR criteria but also those of other contemporary bleeding risk scores were analysed to validate predictors of BARC 3 or 5 bleeding events at 7 years. In the univariate analysis, 13 risk factors were predictors of BARC 3 or 5 bleeding. Of these, 8 risk factors overlapped with one of these risk scores. More overlapping risk factors tended to be more predictive of bleeding events. This suggests that although various risk factors for bleeding have been identified, it may be possible to simplify the list by narrowing it down to important factors that overlap.
The sensitivity of the ARC-HBR criteria for BARC 3 or 5 bleeding in the present study was higher than that of other contemporary bleeding risk scores. In contrast, the specificity of the ARC-HBR criteria for BARC 3 or 5 bleeding in the present study was lower than that of other contemporary bleeding risk scores. A previous report from the Bern PCI Registry, which compared bleeding risk scores at 1 year, also showed that the ARC-HBR criteria had higher sensitivity compared with other bleeding risk scores at the cost of lower specificity.13 Eight of 13 risk factors for predicting BARC 3 or 5 bleeding were included in the ARC-HBR criteria. Conversely, 7, 6, and 3 risk factors were included in the PRECISE-DAPT score, PARIS bleeding score, and CREDO-Kyoto bleeding score, respectively. These might be the reasons why the ARC-HBR score had a high sensitivity for predicting bleeding events. The sensitivity of these clinical bleeding risk scores increased with the inclusion of more risk items, and there was a trade-off between sensitivity and specificity. As a result, c-statistics were comparable among the 4 bleeding risk scores. The low incidence of bleeding events may be the reason for the low positive predictive rate. This may be a clinical limitation of the bleeding risk score. However, given the high sensitivity of the ARC-HBR score, patients with multiple bleeding risk factors and a low risk of ischemic events may need to keep bleeding prevention in mind even in the long term.
The AUC values were lower than those reported in past studies.6–8 This may be due to the following 2 reasons. One is that the risk scores used in this study are a tool to predict bleeding in 1 to 2 years; however, this study evaluated long-term events. Therefore, bleeding events increased over time, and the cumulative incidence of bleeding in non-HBR patients gradually increased. As the negative predictive value decreased, the predictability also decreased. The other is that there is a possibility of missing some events due to the long follow-up period. The decrease in the positive predictive value may have reduced the predictability.
Even though the CREDO-Kyoto bleeding risk score data were provided from a Japanese registry, the predictability was lower in this cohort. We believe that there were some reasons for these results. First, although anemia and old age were the main predictors of bleeding events in this cohort, the CREDO-Kyoto bleeding risk score did not include these factors. Second, almost half of all patients were classified into the low-risk group (Score 0) of the CREDO-Kyoto score. The higher incidence of bleeding in this group compared to that of the other bleeding risk scores may be a reason for the lower AUC. Third, the differences in bleeding endpoints might have affected the outcomes. Finally, in the CREDO-Kyoto score, higher risk patients tended to have a shorter DAPT period. This may have been a negative bias in the predictability of the CREDO-Kyoto score.
Study Limitations
First, this was a single-center retrospective study. Second, in this study, almost all of included patients were Japanese; therefore, these results might be not simply be adapted to other countries’ populations. Third, because the crude model was used in the univariate analysis, confounding between risk factors could not be eliminated. Fourth, the definitions of bleeding and patient backgrounds used to define each bleeding risk score are different, which may affect the predictability in this cohort. Fifth, patients who underwent PCI between 2010 and 2013 were included in the study, and the clinical practice is different from the current practice, which may not be applicable to the current clinical practice. Sixth, in this cohort, there was insufficient power to evaluate the impact of rare risk factors for bleeding such as prior bleeding and liver cirrhosis. Seventh, because the definition of bleeding was BARC 3 or 5, there may have been a bias in favor of the ARC-HBR and PARIS scores, which use BARC as the definition of bleeding. Finally, due to the retrospective nature of this study, there was a possibility of underestimation in the calculation of bleeding risk scores.
Conclusions
In the ARC-HBR and other bleeding risk scores, patients with a higher number of each risk factor had a higher risk of long-term bleeding events. In particular, the ARC-HBR criteria were more sensitive for identifying patients with bleeding events in the long-term than other contemporary risk scores. Screening, weighting, and stratification of known risk factors may be necessary for a more accurate prediction of bleeding risk.
Acknowledgments
The authors appreciate the staff members of the cardiac catheterization laboratory, and Miho Kobayashi, Makiko Kanaike, and Takako Yukiyoshi for their assistance with the manuscript.
Sources of Funding
This study did not receive any specific funding.
Disclosures
The authors declare no conflicts of interest.
IRB Information
The institutional review board of Kurashiki Central Hospital (Ref. No: 3537) approved this study.
Supplementary Files
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-21-0901
References
- 1.
Leon MB, Baim DS, Popma JJ, Gordon PC, Cutlip DE, Ho KK, et al. A clinical trial comparing three antithrombotic-drug regimens after coronary-artery stenting: Stent Anticoagulation Restenosis Study Investigators. N Engl J Med 1998; 339: 1665–1671.
- 2.
Mauri L, Kereiakes DJ, Yeh RW, Driscoll-Shempp P, Cutlip DE, Steg PG, et al; DAPT Study Investigators. Twelve or 30 months of dual antiplatelet therapy after drug-eluting stents. N Engl J Med 2014; 371: 2155–2166.
- 3.
Palmerini T, Benedetto U, Bacchi-Reggiani L, Della Riva D, Biondi-Zoccai G, Feres F, et al. Mortality in patients treated with extended duration dual antiplatelet therapy after drug-eluting stent implantation: A pairwise and Bayesian network meta-analysis of randomised trials. Lancet 2015; 385: 2371–2382.
- 4.
Sharma A, Agrawal S, Garg A, Vallakati A, Lavie CJ, Helft G. Duration of dual antiplatelet therapy following drug-eluting stent implantation: A systemic review and meta-analysis of randomized controlled trials with longer follow up. Catheter Cardiovasc Interv 2017; 90: 31–37.
- 5.
Toyota T, Shiomi H, Morimoto T, Natsuaki M, Kimura T. Short versus prolonged dual antiplatelet therapy (DAPT) duration after coronary stent implantation: A comparison between the DAPT study and 9 other trials evaluating DAPT duration. PLoS One 2017; 12: e0174502.
- 6.
Costa F, van Klaveren D, James S, Heg D, Räber L, Feres F, et al; PRECISE-DAPT Study Investigators. Derivation and validation of the predicting bleeding complications in patients undergoing stent implantation and subsequent dual antiplatelet therapy (PRECISE-DAPT) score: A pooled analysis of individual-patient datasets from clinical trials. Lancet 2017; 389: 1025–1034.
- 7.
Baber U, Mehran R, Giustino G, Cohen DJ, Henry TD, Sartori S, et al. Coronary thrombosis and major bleeding after PCI with drug-eluting stents: Risk scores from PARIS. J Am Coll Cardiol 2016; 67: 2224–2234.
- 8.
Natsuaki M, Morimoto T, Yamaji K, Watanabe H, Yoshikawa Y, Shiomi H, et al; CREDO-Kyoto PCI/CABG Registry Cohort 2, RESET, and NEXT trial investigators. Prediction of thrombotic and bleeding vents after percutaneous coronary intervention: CREDO-Kyoto thrombotic and bleeding risk scores. J Am Heart Assoc 2018; 7: e008708.
- 9.
Urban P, Mehran R, Colleran R, Angiolillo DJ, Byrne RA, Capodanno D, et al. Defining high bleeding risk in patients undergoing percutaneous coronary intervention. Circulation 2019; 140: 240–261.
- 10.
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. Eur Heart J 2019; 40: 2632–2653.
- 11.
Cao D, Mehran R, Dangas G, Baber U, Sartori S, Chandiramani R, et al. Validation of the Academic Research Consortium High Bleeding Risk Definition in contemporary PCI patients. J Am Coll Cardiol 2020; 75: 2711–2722.
- 12.
Natsuaki M, Morimoto T, Shiomi H, Yamaji K, Watanabe H, Shizuta S, et al. Application of the academic research consortium high bleeding risk criteria in an all-comers registry of percutaneous coronary intervention. Circ Cardiovasc Interv 2019; 12: e008307.
- 13.
Ueki Y, Bar S, Losdat S, Otsuka T, Zanchin C, Zanchin T, et al. Validation of Academic Research Consortium for Hight Bleeding Risk (ARC-HBR) criteria in patients undergoing percutaneous coronary intervention and comparison with contemporary bleeding risk scores. EuroIntervention 2020; 16: 371–379.
- 14.
Mehran R, Rao SV, Bhatt DL, Gibson CM, Caixeta A, Eikelboom J, et al. Standardized bleeding definitions for cardiovascular clinical trials: A consensus report from the Bleeding Academic Research Consortium. Circulation 2011; 123: 2736–2747.
- 15.
Choi KH, Song YB, Lee JM, Park TK, Yang JH, Choi JH, et al. Clinical usefulness of PRECISE-DAPT score for predicting bleeding events in patients with acute coronary syndrome undergoing percutaneous coronary intervention: An analysis from the SMART-DATE randomized trial. Circ Cardiovasc Interv 2020; 13: e008530.
- 16.
Sorrentino S, Claessen BE, Chandiramani R, Guedeney P, Vogel B, Baber U, et al. Long-term safety and efficacy of durable polymer cobalt-chromium everolimus-eluting stents in patients at high bleeding risk: A patient-level stratified analysis from four postapproval studies. Circulation 2020; 141: 891–901.
- 17.
Cao D, Mehran R, Dangas G, Baber U, Sartori S, Chandiramani R, et al. Validation of the Academic Research Consortium High Bleeding Risk definition in contemporary PCI patients. J Am Coll Cardiol 2020; 75: 2711–2722.
