Abstract
Bleeding events result in morbidity and mortality in patients who underwent percutaneous coronary intervention (PCI). There are limited data on the predicting bleeding complications in patients who underwent stent implantation and subsequent dual antiplatelet therapy (PRECISE-DAPT) and Academic Research Consortium for High Bleeding Risk (ARC-HBR) scores’ ability to predict in-hospital outcomes in patients who underwent PCI. Consecutive patients who underwent PCI at tertiary centers from January 2016 to March 2018 were identified and the bleeding risk scores were calculated. The primary end point was the National Cardiovascular Data Registry-defined in-hospital bleeding stratified by low versus high predicted bleeding risk. The major and net adverse cardiovascular events were also examined. The discriminatory ability of the risk models was determined using receiver operating characteristic curves. Among 3,659 patients studied, the in-hospital major bleeding was 3.3% (n = 121). The patients characterized as high bleeding risk by either criterion had significantly higher bleeding rates than those meeting the low-risk criteria (ARC-HBR 5.4% vs 3.3%, p <0.001; PRECISE-DAPT 5.8% vs 2.4%, p <0.001), and higher major adverse cardiovascular events and net adverse clinical events. These risk estimates showed moderate and similar predictive ability (ARC-HBR high-risk area under the receiver operating characteristic curve [AUC] 0.62, PRECISE-DAPT ≥25 AUC 0.61, p = 0.49), with no incremental benefit to adding the estimates (AUC 0.60). The subgroup analysis revealed that women had higher bleeding rates than men (5.53% vs 2.39%, p <0.001); however, the predictive ability of the criteria were similar in women and men. The patients identified as having a high bleeding risk by the PRECISE-DAPT and the ARC-HBR criteria before PCI are at high risk for in-hospital bleeding and adverse outcomes independent of gender. The 2 scores have moderate predictive ability for bleeds. Further study is needed to determine strategies to reduce risk in this population.
Introduction
Bleeding in patients who underwent percutaneous coronary intervention (PCI) is recognized as an important independent predictor of morbidity and mortality1. In-hospital bleeding is associated with in-hospital death,2-4 longer length of stay, and higher 1-year mortality. Several clinical scores to predict bleeding have been developed and externally validated for in-hospital and 1-year bleeding risk.5-7 More recently, the predicting bleeding complications in patients who underwent stent implantation and subsequent dual antiplatelet therapy (PRECISE-DAPT) score, a simple 5-item risk score, proved to be more accurate in assessing the 1-year risk of out-of-hospital bleeding among patients initiated on dual antiplatelet therapy (DAPT) compared with the earlier scores.8 Recently, the Academic Research Consortium for High Bleeding Risk (ARC-HBR) bleeding risk definition was developed by expert consensus and incorporates major and minor criteria not included in previous scores to estimate a patient’s bleeding risk.9 Although the PRECISE-DAPT and ARC-HBR effectively predict bleeding after discharge, the clinical utility of these definitions alone or in combination for prediction of in-hospital outcomes, such as bleeding or ischemia, is unknown. We hypothesized that the 2 bleeding risk criteria, calculated before PCI, would be useful in assessing risk of in-hospital periprocedural bleeding and other adverse outcomes. Therefore, the primary objectives of this study were to compare in-hospital bleeding rates among patients who underwent PCI, characterized as high versus low bleeding risk using the PRECISE-DAPT and ARC-HBR scores and to evaluate their performance.
Methods
This study is a retrospective cohort registry using the Lifespan Cardiovascular Institute’s institutional National Cardiovascular Data Registry (NCDR) CathPCI registry (version 4.4) for all PCI procedures performed at 2 hospitals from January 2016 to March 2018. The data, extracted for quality purposes, are maintained on a secure server. The variables include demographic, co-morbidities, medications, preprocedural laboratory values, and procedural characteristics and in-hospital outcomes. Using the list of patients generated from this database, a deidentified Health Insurance Portability and Accountability Act-compliant REDCap database was created. The variables and criteria required for the calculation of the ARC-HBR criteria and PRECISE-DAPT that were not available in the institutional NCDR CathPCI database were extracted manually from the electronic health records by 3 authors (V.K., S.L, and C.S.) through chart review of the patients’ medical electronic medical records. The variables that were manually extracted through chart review included history of bleeding, anticipated use of long-term oral anticoagulation, chronic bleeding diathesis, history of cirrhosis, malignancy, Intracerebral hemorrhage, ischemic stroke, nondeferrable major surgery on DAPT, history of surgery or trauma, history of use of nonsteroidal anti-inflammatory drugs or steroids, using a combination of the variables obtained through the institutional NCDR database and the variables manually extracted, the ARC-HBR classification, and PRECISE-DAPT score before each procedure was determined.
The study population included consecutive patients who received PCI with a drug-eluting stent and in-hospital treatment with DAPT from January 2016 to March 2018 at Rhode Island and the Miriam Hospitals in Providence, Rhode Island. The procedural indication included both stable ischemic heart diseases and acute coronary syndrome (ST-elevation myocardial infarction (STEMI) and non-STEMI). Patients with out-of-hospital cardiac arrest and in-hospital coronary artery bypass graft (CABG) or who underwent concomitant in-hospital procedures (transcatheter aortic valve replacement, balloon aortic valvuloplasty, MitraClip, or WATCHMAN) were excluded.
A total of 4,081 procedures met the previously mentioned criteria. Of these, 94 were excluded because of missing data in the electronic medical records that were relevant to calculating the bleeding risk. This was usually due to missing preprocedural laboratory tests or past medical history not recorded in the chart, which occurred mostly in patients who underwent elective PCI. In patients that had more than 1 procedure within 12 months (n = 328), only the index PCI procedure was included in the analysis. In total, 3,659 procedures were included in our analysis (Figure 1).
Figure 1.
Inclusion and exclusion study flow diagram. DES = drug-eluting stent. OHCA = out-of-hospital cardiac arrest.
The primary outcome for this analysis was post-PCI bleeding. This was defined as any of the following, which occurred within 72 hours after PCI or before hospital discharge: (1) arterial access site bleeding, either external or leading to a hematoma >10 cm with femoral access, >5 cm with brachial access, or >2 cm with radial access; (2) retroperitoneal, gastrointestinal, or genitourinary bleeding; (3) intracranial hemorrhage; (4) cardiac tamponade; (5) postprocedural hemoglobin decrease of 3 g/100 ml in patients with a preprocedural hemoglobin level of ≤16 g/100 ml; or (6) postprocedural non-CABG-related blood transfusion in patients with preprocedural hemoglobin level of ≥8 g/100 ml.10
High bleeding risk was defined as PRECISE-DAPT score ≥25 or if the patient met 1 major criterion or 2 minor criteria for the ARC-HBR variables.9,11 Secondary outcomes included: (1) in-hospital major adverse cardiovascular events (MACE), which was a composite of death before discharge, myocardial infarction (MI), or ischemic or hemorrhagic stroke; (2) net adverse clinical events (NACEs), which was a composite of MACE + blood transfusion or any bleeding. All outcomes (MI, stroke) were site-determined and defined according to the NCDR version 4.4 criteria during the period of data collection.12
The univariate distribution of all variables is described. Patients with PRECISE-DAPT scores ≥25 were classified as high risk; whereas patients were classified as high risk if they met 1 major criterion or 2 minor criteria, as defined by the ARC-HBR definition.9 Categorical variables are reported as frequencies (counts and percentages), continuous variables are reported as means and standard deviations or medians and interquartile ranges. The bivariable comparisons between groups for categorical variables were conducted using Pearson chi-square or Fisher’s exact test, where appropriate. Continuous variables were compared using Student t test or Wilcoxon rank-sum if the data were not normally distributed. All tests are 2-sided; p <0.05 are considered statistically significant. The discriminatory ability of ARC-HBR and PRECISE-DAPT scores were assessed by comparing the area under the receiver operating characteristic curve (AUC). All relative risks were estimated using log-binomial regression models and are reported with 95% confidence intervals (CIs). To assess the effect modification of the bleeding risk scores by gender, we conducted stratified subgroup analyses, and including the interaction term between gender and bleeding risk scores in a multivariable logistic regression model. The analyses were performed using SAS 9.4 (Cary, North Carolina) and Stata/MP 16.1(College Station, Texas).
Results
Patient demographic, clinical, and procedural characteristics according to ARC-HBR and PRECISE-DAPT bleeding risk are presented in Table 1. With both bleeding risk criteria, the patients characterized as high bleeding risk were older; more often female; more likely to have had a previous MI, previous PCI, or previous CABG; more likely to be a current smoker; and more likely have peripheral arterial disease, chronic lung disease, diabetes, end stage renal disease on dialysis, have a lower preprocedural hemoglobin, and higher preprocedural creatinine than those patients classified as low risk (all p <0.001). There was no statistically significant difference in the race of the patients characterized as having a high bleeding risk by either bleeding risk criteria. Patients characterized as having a high bleeding risk by either criterion were significantly less likely to have radial access for PCI (p <0.001). Patients were more likely to present in cardiogenic shock and receive an intra-aortic balloon pump if they were characterized as having a high bleeding risk using either score.
Table 1.
Baseline demographic, clinical, and procedural characteristics
| ARC-HBR |
PRECISE-DAPT |
||||||
|---|---|---|---|---|---|---|---|
| Overall (n = 3659) | High risk (n = 1382) | Low risk (n = 2277) | E p-Value | High risk ≥25 (n = 980) | Low risk <25 (n = 2679) | p-Value | |
| Age (median, IQR) | 66.00 (58.00, 75.00) | 75.00 (66.00, 81.00) | 62.00 (55.00, 69.00) | < 0.001 | 78.00 (70.00, 84.00) | 63.00 (55.00, 70.00) | < 0.001 |
| Caucasian (%) | 3393 (92.7%) | 1292 (93.5%) | 2101 (92.3%) | 0.17 | 901 (91.9%) | 2492 (93.0%) | 0.26 |
| African-American (%) | 148 (4%) | 59 (4.3%) | 89 (3.9%) | 0.59 | 46 (4.7%) | 102 (3.8%) | 0.23 |
| Female, n (%) | 1067 (29.2%) | 501(36.3%) | 566 (24.9%) | < 0.001 | 408 (41.6%) | 495 (21.3%) | < 0.001 |
| Prior MI, n (%) | 1067 (29.2%) | 508 (36.8%) | 559 (24.5%) | < 0.001 | 373 (38.1%) | 694 (25.9%) | < 0.001 |
| Prior PCI, n (%) | 1201 (32.8%) | 516 (37.3%) | 685 (30.1%) | < 0.001 | 379 (38.9%) | 822 (30.7%) | < 0.001 |
| Prior CABG, n (%) | 404 (11.0%) | 237 (17.2%) | 167 (7.3%) | < 0.001 | 173 (17.7%) | 231 (8.6%) | < 0.001 |
| Current/Recent Smoker, n (%) | 871 (23.8%) | 201 (14.6%) | 670 (29.5%) | < 0.001 | 110 (11.2%) | 761 (28.4%) | < 0.001 |
| Cerebrovascular Disease, n (%) | 414 (11.3%) | 276 (20%) | 138 (6.1%) | < 0.001 | 191 (19.5%) | 223 (8.3%) | < 0.001 |
| PAD, n (%) | 373 (10.2%) | 221 (16%) | 152 (6.7%) | < 0.001 | 171 (17.4%) | 202 (7.5%) | < 0.001 |
| Lung disease, n (%) | 476 (13%) | 269 (19.5%) | 207 (9.1%) | < 0.001 | 187 (19.1%) | 289 (10.8%) | < 0.001 |
| Diabetes, n (%) | 1355 (37%) | 653 (47.3%) | 702 (30.8%) | < 0.001 | 452 (46.1%) | 903 (33.7%) | < 0.001 |
| Dialysis, n (%) | 75 (2.0%) | 74 (5.4%) | 1 (0.0%) | < 0.001 | 74 (5.4%) | 1 (0.0%) | < 0.001 |
| Ejection fraction (Median, IQR) | 55.00 (45.00, 60.00) | 53.00 (40.00, 60.00) | 55.00 (47.00, 60.00) | < 0.001 | 53.00 (40.00, 60.00) | 55.00 (45.00, 60.00) | < 0.001 |
| Pre-procedure Hemoglobin (Median, IQR) | 13.70 (12.20, 14.90) | 11.90 (10.60, 13.40) | 14.40 (13.40, 15.30) | < 0.001 | 11.60 (10.40, 13.20) | 14.20 (13.10, 15.20) | < 0.001 |
| Pre-procedure Creatinine (Median, IQR) | 0.90 (0.70, 1.10) | 1.10 (0.90, 1.50) | 0.80 (0.70, 1.00) | < 0.001 | 1.10 (0.90, 1.60) | 0.90 (0.70, 1.00) | < 0.001 |
| Radial access, n (%) | 2207 (60.3%) | 683 (49.4%) | 1524 (66.9%) | < 0.001 | 448 (45.7%) | 1759 (65.7%) | < 0.001 |
| IABP, n (%) | 45 (1.2%) | 32 (2.3%) | 13 (0.6%) | < 0.001 | 27 (2.8%) | 18 (0.7%) | < 0.001 |
| Cardiogenic shock, n (%) | 50 (1.4%) | 34 (2.5%) | 16 (0.7%) | < 0.001 | 20 (2.0%) | 30 (1.1%) | 0.03 |
CABG = coronary artery bypass grafting; MI = myocardial infarction; PAD = peripheral arterial disease; PCI = percutaneous coronary intervention.
The variables associated with a bleeding event are demonstrated in Table 2. Bleeds were more statistically more likely to occur in older patients; women; and patients with a past medical history of cerebrovascular disease, peripheral vascular disease, lung disease, and heart failure with reduced ejection fraction (all p <0.05). Bleeds were more common when the femoral access was used, when an intra-aortic balloon pump was used, or when the patient was in cardiogenic shock (all p <0.001). The patients who experienced a bleed stayed in the hospital for longer periods of time (5 vs 2 days, p <0.001). There were 80 bleeding events; 61.3% were due to bleeding at the access site (n = 49), 21.3% were hematomas (n = 17), 13.8% were gastrointestinal (n = 11), 2.5% were retroperitoneal (n = 2), 1.3% were genitourinary (n = 1), and 21.3% were classified as other (n = 17).
Table 2.
Variables by ACC defined bleeding
| n = 3,659 | Total cohort (n = 3,659) | No bleeding event (n = 3,538) | Bleeding event (n = 121) | p Value |
|---|---|---|---|---|
| Length of hospital stay (days) | ||||
| Median (IQR) | 2 (1–3) | 2 (1–3) | 5 (3–9) | <0.001* |
| Age (years) | ||||
| Median (IQR) | 66 (58–75) | 66 (57–74) | 73 (64–80) | <0.001* |
| Caucasian, n(%) | 117 (96.7) | 0.11† | ||
| Yes | 3,393 (92.7) | 3,276 (92.6) | ||
| African-American, n(%) | 148 (4.0) | 145 (4.1) | 3 (2.5) | 0.49† |
| Yes | ||||
| Female, n(%) | 1,067 (29.2) | 1,008 (28.5) | 59 (48.8) | <0.001† |
| Yes | ||||
| Prior MI, n(%) | 1,067 (29.2) | 1,025 (28.9) | 42 (34.7) | 0.19† |
| Yes | ||||
| Prior PCI, n(%) | 1,201 (32.8) | 1,165 (32.9) | 36 (29.8) | 0.49† |
| Yes | ||||
| Prior CABG, n(%) | 404 (11.0) | 393 (11.1) | 11 (9.1) | |
| Yes | 0.56† | |||
| Current/recent smoker, n(%) | 871 (23.8) | 846 (23.9) | 25 (20.7) | |
| Yes | 0.45† | |||
| Cerebrovascular disease, n(%) | 414 (11.3) | 391 (11.1) | 23 (19.0) | 0.01† |
| Yes | ||||
| PAD, n(%) | 373 (10.2) | 347 (9.8) | 26 (21.5) | <0.001† |
| Yes | ||||
| Lung disease, n(%) | 476 (13.0) | 452 (12.8) | 24 (19.8) | 0.03† |
| Yes | ||||
| Diabetes, n(%) | 1,355 (37.0) | 1,309 (37.0) | 46 (38.0) | 0.85† |
| Yes | ||||
| Dialysis, n(%) | 75 (2.1) | 70 (1.9) | 5 (4.1) | 0.10† |
| Yes | ||||
| Pre-PCI ejection fraction | 55 (45–60) | 55 (45–60) | 42 (33–60) | |
| Median (IQR) | <0.001* | |||
| Pre-procedure creatinine | 0.90 (0.70–1.1) | 0.90 (0.70–1.1) | 1.0 (0.80–1.2) | |
| Median (IQR) | 0.14* | |||
| Radial access, n(%) | ||||
| Yes | 2,207 (60.3) | 2,169 (61.3) | 38 (31.4) | <0.001† |
| IABP, n(%) | <0.001† | |||
| Yes | 45 (1.2) | 32 (0.9) | 13 (10.7) | |
| Cardiogenic shock | ||||
| Yes | 50 (14) | 36 (10) | 14 (116) | <0001† |
CABG = coronary artery bypass grafting; MI = myocardial infarction; PAD = peripheral arterial disease; PCI = percutaneous coronary intervention.
Wilcoxon rank-sum
Fisher’s exact test
The patient outcomes stratified by ARC-HBR and PRECISE-DAPT bleeding risk (high vs low) are shown in Table 3. Of the 3,659 patients studied, 121 had bleeding events (3.3%). The ARC-HBR criteria characterized 1,382 patients (37.7%) as having a high bleeding risk and the PRECISE-DAPT score characterized 980 patients (26.7%) as having a high bleeding risk. Patients characterized as having a high bleeding risk by either criterion had higher in-hospital bleeding rates than those who were at low risk (ARC-HBR 5.4% vs 3.3%, p <0.001; PRECISE-DAPT 5.8% vs 2.4%, p <0.001). Patients at high bleeding risk according to both criteria also experienced significantly higher rates of the secondary end points; MACE and NACE; and of individual end points, including mortality, CVA/stroke, and postprocedural MI (all p <0.05).
Table 3.
In-hospital outcomes according to bleeding risk defined by ARC-HBR and PRECISE-DAPT
| ARC-HBR |
PRECISE-DAPT |
||||||
|---|---|---|---|---|---|---|---|
| Overall (n = 3,659) | High risk (n =1,382) | Low risk (n=2,277) | p Value | High risk ≥25 (n=980) | Low risk <25 (n=2,679) | p Value | |
| Bleeding, n (%) | 121 (3.3%) | 74 (5.4%) | 47 (2.1%) | <0.001 | 57 (5.8%) | 64 (2.4%) | <0.001 |
| MACE, n (%) | 89 (2.4%) | 47 (3.4%) | 42 (1.8%) | 0.004 | 31 (3.2%) | 58 (2.2%) | 0.09 |
| NACE, n (%) | 159 (4.4%) | 87 (6.3%) | 73 (3.2%) | <0.001 | 62 (6.3%) | 97 (3.6%) | 0.001 |
| Stroke, n (%) | 20 (0.5%) | 13 (0.9%) | 7 (0.3%) | 0.011 | 9 (0.9%) | 11 (0.4%) | 0.07 |
| MI, n (%) | 51 (1.4%) | 20 (1.4%) | 31 (1.4%) | 0.83 | 10 (1.0%) | 41 (1.5%) | 0.24 |
| Mortality, n (%) | 20 (0.5%) | 15 (1.1%) | 5 (0.2%) | <0.001 | 13 (1.3%) | 7 (0.3%) | <0.001 |
Abbreviations: MACE = major adverse cardiovascular events (composite of death before discharge, myocardial infarction (MI), or stroke); NACE = net adverse clinical events (composite of MACE+blood transfusion or any bleeding).
Table 4 examines the bleeding event rates based on the classification by both the ARC-HBR and PRECISE-DAPT. The majority of patients were classified as having a low bleeding risk by both criteria (2,112 of 2,659, 57.9%). Among the rest of the patients, 23.1% (825 of 3,659) were characterized as having a high bleeding risk by both criteria, whereas 15.6% and 4.3% were characterized as having a high bleeding risk by only the ARC-HBR criteria or the PRECISE-DAPT criteria, respectively. Patients classified as high bleeding risk by both ARC-HBR and PRECISE-DAPT were more likely to have a bleeding event (51 of 825, 6.2%) than those classified as having a low bleeding risk by both criteria (41 of 2,112, 1.93%). Patients with discordant classifications (where 1 score stratified as a high bleeding risk, whereas the other as low bleeding risk) had an intermediate rate of bleeding (ARC-HBR = high and PRECISE-DAPT <25 = 23 of 557, 4.1% and ARC-HBR = low and PRECISE-DAPT ≥25 = 6 of 155, 3.9%).
Table 4.
Rates of in-hospital bleeding stratified by high versus low bleeding risk using PRECISE-DAPT and ARC-HBR scores
| N (3,659) | Bleeding (n = 121) | p Value | |
|---|---|---|---|
| PRECIS–DAPT<25, ARC-HBR=low | 2,122 (57.9%) | 41 (1.93%) | <0.001* |
| PRECISE-DAPT <25, ARC-HBR=high | 557 (15.6%) | 23 (4.1%) | |
| PRECISE-DAPT ≥25, ARC-HBR=low | 155 (4.3%) | 6 (3.9%) | |
| PRECISE-DAPT ≥25, ARC-HBR=high | 825 (23.11%) | 51(6.2%) |
Fisher’s exact test.
Data are (n, row %).
The test characteristics were calculated for both the ARC-HBR and PRECISE-DAPT. The ARC-HBR criteria exhibited a higher sensitivity but lower specificity for predicting in-hospital bleeding than the PRECISE-DAPT score (sensitivity % [95% CI] 61.2% [51.9% to 70.0%] vs 47.1% [38.0% to 56.4%]; specificity % [95% CI] 63.0% [61.4% to 64.6%] vs 73.9% [72.4% to 75.4%]). When the c-statistics or AUCs were used to compare the diagnostic value of the criteria, there was no significant difference in the ARC-HBR and PRECISE-DAPT AUC in the prediction of in-hospital bleeding, p = 0.49 (Figure 2). The combined ARC-HBR high bleeding risk and PRECISE-DAPT high bleeding risk AUC was also statistically similar to the AUC of the individual curves.
Figure 2.
ROC curve comparison of high bleeding risk defined by PRECISE-DAPT, ARC-HBR, and a composite score and the primary outcome of in-hospital bleeding p = 0.49. ROC = receiver operating characteristic.
The relative risk of in-hospital bleeding given classification by the 2 bleeding risk criteria is shown in Table 5. The relative risk if classified as high bleeding risk by ARC-HBR was 2.59 (95% CI 1.81 to 3.72), while that of patients classified as high bleeding risk by PRECISE-DAPT had a relative risk of 2.43 (95% CI 1.72 to 3.45).
Table 5.
Relative risk of in-hospital bleeding stratified by high versus low bleeding risk using PRECISE-DAPT and ARC-HBR scores
| RR (95% CI) | |
|---|---|
| PRECISE-DAPT ≥25 | 2.43 (1.72, 3.45) |
| ARC-HBR High Risk | 2.59 (1.81 3.72) |
| PRECISE-DAPT & ARC-HBR High Risk | 2.50 (1.76, 3.56) |
| PRECISE-DAPT or ARC-HBR High Risk | 2.69 (1.86, 3.90) |
CI = confidence interval.
Women comprised 29.2% (n = 1,067) of the patients and had a significantly higher bleeding rates than men (5.53% vs at 2.39%, p <0.001). In both women and men, however, the bleeding event rates among those stratified as having a high bleeding risk was statistically significantly higher than the rates among those stratified as having a lower bleeding risk (Table 6). We assessed the interaction between gender and bleeding risk criteria and the interactions were not found to be statistically significant (p >0.05).
Table 6.
Bleeding rates in women and men stratified by according to high versus low ARC-HBR and PRECISE-DAPT risk scores
| ARC-HBR |
PRECISE-DAPT |
|||||||
|---|---|---|---|---|---|---|---|---|
| High risk n (%) | Low risk n (%) | p Value | Interaction term p Value | High risk n (%) | Low risk n (%) | p Value | Interaction term p Value | |
| Male | 38/881 (4.30%) | 24/1,711 (1.40%) | <0.001* | 0.13 | 27/572 (4.72%) | 35/2,020 (1.73%) | <0.001* | 0.17 |
| Female | 36/501 (7.20%) | 23/566 (4.10%) | 0.03* | 30/408 (7.35%) | 29/659 (4.40%) | 0.04* | ||
Fisher’s exact test.
Radial access was used for access in the majority (2,207 of 3,659, 60.3%; Table 1) of patients. However, the rates of radial access were lower among patients characterized as having a high versus low bleeding risk by either the ARC-HBR criteria (49.4% vs 66.9%, p <0.001) or the PRECISE-DAPT score (45.7% vs 65.7%, p <0.001). The bleeding rates were lower among patients with radial access than femoral access (1.7% vs 5.7%, p <0.001; Table 7). Procedures with radial access were less likely to experience a bleed than those with femoral access, regardless of whether patients were characterized as having a high versus low risk by either PRECISE-DAPT or ARC-HBR (all p <0.001; Table 7). The comparison of bleeding rates in elective (11 of 482) versus nonelective (110 of 3,177) procedures found that the difference in the bleeding rates was not statistically significant (2.28% vs 3.46%, p = 0.22).
Table 7.
Bleeding rates by access site according to high versus low-risk scores
| n = 3,659 | Overall Bleed Rate |
ARC-HBR |
PRECISE-DAPT |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Bleed rate n (%) | p Value | Low risk n(%) (n = 2,277) |
p value | High risk n(%) (n = 1,382) |
p Value | Low risk n(%) (n=2,679) |
p Value | high Risk n(%) (n=980) |
p Value | |
| Femoral (753) | 83 (5.7%) | <0.001* | 27 (3.6%) | 0.001* | 56 (8.0%) | <0.001* | 38 (4.1%) | <0.001* | 45 (8.5%) | <0.001* |
| Radial (2,207) | 38 (1.7%) | 20 (1.3%) | 18 (2.6%) | 26 (1.5%) | 12 (2.7%) | <0.001* | ||||
Fisher’s exact test.
Discussion
The association of high bleeding risk, defined by PRECISE-DAPT and ARC-HRB, with in-hospital outcomes was investigated in a contemporary cohort of patients who underwent PCI with a drug-eluting stent. This main findings of this study were that (1) a higher bleeding risk by both criteria was statistically correlated with the primary outcome of in-hospital bleeding, and MACE and NACE, (2) the ARC-HBR criteria have a higher sensitivity but lower specificity for in-hospital bleeding than the PRECISE-DAPT score, but the AUC of the receiver operating characteristic curves shows no statistically significant difference in performance, (3) the risk of in-hospital bleeding was the lowest among those predicted to be at low risk by both criteria and the highest in those predicted to be at high risk by either PRECISE-DAPT or ARC-HBR, and (4) women have a higher in-hospital bleeding than men but both criteria predict risk equally well in women and men.
Because bleeding has been increasingly recognized as a driver of morbidity and mortality, preprocedural stratification of bleeding risk has become increasingly important. The PRECISE-DAPT score uses 5 easily collected variables to risk stratify patients into low or high bleeding risk groups.11 The ARC-HBR criteria share many of these same variables but also includes risk factors associated with bleeding, such as chronic kidney disease, thrombocytopenia, chronic bleeding diathesis, history of cirrhosis, malignancy, long-term nonsteroidal anti-inflammatory drugs/steroid use, and so on, which have been shown to be the risk factors for bleeding.9 Both definitions have previously been independently validated as the predictors of bleeding at 1 year after PCI.8,13 To the best of our knowledge, this is the first study to compare the performance of the 2 criteria in the prediction of in-hospital outcomes. The results of this study suggest that both scores are useful predictors of bleeding and ischemic outcomes, especially if both scores stratify a patient as having a low bleeding risk. Patients stratified as having a high risk by PRECISE-DAPT alone or ARC-HBR alone have similar risk with those found to be at high risk by both definitions.
Adding the 2 scores together did not result in an associated change in the AUC. We surmise that this is likely due to the fact that both ARC-HBR and PRECISE-DAPT scores are measuring similar outcomes in a similar cohort. The scores on their own had very similar AUCs; therefore, it is not unexpected that the combined score also had a similar AUC.
An important finding in our study was that women had a higher rate of in-hospital bleeding than men, echoing the findings of previous studies.14,15 Indeed, previous in-hospital bleeding risk scores, such as CRUSADE, specifically include the female gender as an independent variable associated with bleeding.7 However, our study demonstrates that despite the fact that neither the PRECISE-DAPT nor the ARC-HBR criteria includes gender as a variable, women were more often classified as having a high bleeding risk using both criteria. A nonsignificant p value for interaction suggests that the predictive ability of the scores is similar and equally clinically useful in women and men. This is important because neither ARC-HBR nor PRECISE-DAPT includes gender to calculate the bleeding risk; thus, they may more appropriately account for variables driving increased bleeding risk in women.
This study suggests that the ARC-HBR and PRECISE-DAPT bleeding criteria can be used to risk stratify patients and guide the therapeutic selection to lower the bleeding risk. For instance, we found that radial access was used at lower rates in patients stratified as having a higher bleeding risk, despite the lower rates of bleeding seen with radial access.16,17 Other strategies to lower bleeding risk include the choice of less potent P2Y12 inhibitors or minimizing use of upstream P2Y12 inhibitor in patients without STEMI and the choice of intraprocedural anticoagulation. A combination of these strategies may help lower the risk of bleeding among high-risk patients who underwent PCI.
Our study enrolled patients who underwent both urgent PCI, and elective PCI. We found that bleeding rates were numerically lower in elective PCI, but it was not a statistically significant difference. In previous studies of bleeding after PCI, urgent PCI for non-STEMI or STEMI was associated with a significantly higher rate of bleed.4,18,19 An explanation for this difference could be the low rate of radial access in older studies compared with our patient population.
This study also shows that both the PRECISE-DAPT and ARC-HBR criteria can be used to predict in-hospital bleeding. Previous analysis from the GLOBAL-LEADERS trial has shown that PRECISE-DAPT score can be used to predict Bleeding Academic Research Consortium (BARC) 3 to 5 bleeding within 30 days of PCI20,21 However, as far as the authors are aware, this is the first study to examine the performance of the ARC-HBR criteria in the prediction of in-hospital bleeding. Further research using larger data sets is needed, but this suggests that the ARC-HBR criteria can also be used to predict in-hospital bleeding.
Some limitations of our study should be considered. Because this was a retrospective study, the quality of data collected relied on the fidelity of the patient medical record, especially in the collection of the variables of the ARC-HBR criteria. In addition, the results of this study are limited to in-hospital outcomes because of the nature of the outcomes data from the CathPCI database. Finally, bleeding events may be under-reported, with possible under-reporting of nonfatal complications.
Disclosures
Dr. Aronow is a consultant for Philips, Silk Road Medical; Dr. Abbott receives research grants form Boston Scientific and MicroPort and is a consultant for Abbott, Medtronic, Philips, and Recor. The remaining authors have no conflicts of interest to declare.
References
- 1.Kinnaird TD, Stabile E, Mintz GS, Lee CW, Canos DA, Gevorkian N, Pinnow EE, Kent KM, Pichard AD, Satler LF, Weissman NJ, Lindsay J, Fuchs S. Incidence, predictors, and prognostic implications of bleeding and blood transfusion following percutaneous coronary interventions. Am J Cardiol 2003;92:930–935. [DOI] [PubMed] [Google Scholar]
- 2.Rao SV. Bleeding as a predictor of mortality risk. Rev Cardiovasc Med 2006;7(suppl 3):S12–S18. [PubMed] [Google Scholar]
- 3.Rao SV, O’Grady K, Pieper KS, Granger CB, Newby LK, Van de Werf F, Mahaffey KW, Califf RM, Harrington RA. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol 2005;96:1200–1206. [DOI] [PubMed] [Google Scholar]
- 4.Chhatriwalla AK, Amin AP, Kennedy KF, House JA, Cohen DJ, Rao SV, Messenger JC, Marso SP. National Cardiovascular Data Registry. Association between bleeding events and in-hospital mortality after percutaneous coronary intervention [published correction appears in JAMA 2013;310(6):647] JAMA 2013;309:1022–1029. [DOI] [PubMed] [Google Scholar]
- 5.Mehran R, Pocock SJ, Nikolsky E, Clayton T, Dangas GD, Kirtane AJ, Parise H, Fahy M, Manoukian SV, Feit F, Ohman ME, Witzenbichler B, Guagliumi G, Lansky AJ, Stone GW. A risk score to predict bleeding in patients with acute coronary syndromes. J Am Coll Cardiol 2010;55:2556–2566. [DOI] [PubMed] [Google Scholar]
- 6.Stone GW, McLaurin BT, Cox DA, Bertrand ME, Lincoff AM, Moses JW, White HD, Pocock SJ, Ware JH, Feit F, Colombo A, Aylward PE, Cequier AR, Darius H, Desmet W, Ebrahimi R, Hamon M, Rasmussen LH, Rupprecht HJ, Hoekstra J, Mehran R, Ohman EM, Investigators ACUITY. Bivalirudin for patients with acute coronary syndromes. N Engl J Med 2006;355:2203–2216. [DOI] [PubMed] [Google Scholar]
- 7.Subherwal S, Bach RG, Chen AY, Gage BF, Rao SV, Newby LK, Wang TY, Gibler WB, Ohman EM, Roe MT, Pollack CV Jr, Peterson ED, Alexander KP. Baseline risk of major bleeding in non-ST-segment-elevation myocardial infarction: the CRUSADE (Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA Guidelines) Bleeding Score. Circulation 2009;119:1873–1882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Choi SY, Kim MH, Cho YR, Sung Park J, Min Lee K, Park TH, Yun SC. Performance of PRECISE-DAPT score for predicting bleeding complication during dual antiplatelet therapy. Circ Cardiovasc Interv 2018;11:e006837. [DOI] [PubMed] [Google Scholar]
- 9.Urban P, Mehran R, Colleran R, Angiolillo DJ, Byrne RA, Capodanno D, Cuisset T, Cutlip D, Eerdmans P, Eikelboom J, Farb A, Gibson CM, Gregson J, Haude M, James SK, Kim HS, Kimura T, Konishi A, Laschinger J, Leon MB, Magee PFA, Mitsutake Y, Mylotte D, Pocock S, Price MJ, Rao SV, Spitzer E, Stockbridge N, Valgimigli M, Varenne O, Windhoevel U, Yeh RW, Krucoff MW, Morice MC. Defining high bleeding risk in patients undergoing percutaneous coronary intervention. Circulation 2019;140:240–261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rao SV, McCoy LA, Spertus JA, Krone RJ, Singh M, Fitzgerald S, Peterson ED. An updated bleeding model to predict the risk of postprocedure bleeding among patients undergoing percutaneous coronary intervention: a report using an expanded bleeding definition from the National Cardiovascular Data Registry CathPCI Registry. JACC Cardiovasc Interv 2013;6:897–904. [DOI] [PubMed] [Google Scholar]
- 11.Costa F, van Klaveren D, James S, Heg D, Räber L, Feres F, Pilgrim T, Hong MK, Kim HS, Colombo A, Steg PG, Zanchin T, Palmerini T, Wallentin L, Bhatt DL, Stone GW, Windecker S, Steyerberg EW, Valgimigli M. 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. [DOI] [PubMed] [Google Scholar]
- 12.NCDR. NCDR Registry, CathPCI Registry Coder’s Data, Dictionary, 4 (4). https://www.ncdr.com/WebNCDR/docs/default-source/public-data-collection-documents/cathpci_v4_codersdictionary_4-4.pdf?sfvrsn=b84d368e_2. Accessed January 9, 2022. [Google Scholar]
- 13.Cao D, Mehran R, Dangas G, Baber U, Sartori S, Chandiramani R, Stefanini GG, Angiolillo DJ, Capodanno D, Urban P, Morice MC, Krucoff M, Goel R, Roumeliotis A, Sweeny J, Sharma SK, Kini A. Validation of the Academic Research Consortium high bleeding risk definition in contemporary PCI patients. J Am Coll Cardiol 2020;75:2711–2722. [DOI] [PubMed] [Google Scholar]
- 14.Ahmed B, Dauerman HL. Women, bleeding, and coronary intervention. Circulation 2013;127:641–649. [DOI] [PubMed] [Google Scholar]
- 15.Chichareon P, Modolo R, Kerkmeijer L, Tomaniak M, Kogame N, Takahashi K, Chang CC, Komiyama H, Moccetti T, Talwar S, Colombo A, Maillard L, Barlis P, Wykrzykowska J, Piek JJ, Garg S, Hamm C, Steg PG, Jüni P, Valgimigli M, Windecker S, Onuma Y, Mehran R, Serruys PW. Association of sex with outcomes in patients undergoing percutaneous coronary intervention: A subgroup analysis of the GLOBAL LEADERS randomized clinical trial. JAMA Cardiol 2020;5:21–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Baklanov DV, Kaltenbach LA, Marso SP, Subherwal SS, Feldman DN, Garratt KN, Curtis JP, Messenger JC, Rao SV. The prevalence and outcomes of transradial percutaneous coronary intervention for ST-segment elevation myocardial infarction: analysis from the National Cardiovascular Data Registry (2007 to 2011) [published correction appears in J Am Coll Cardiol 2013;61:1469] J Am Coll Cardiol 2013;61:420–426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Feldman DN, Swaminathan RV, Kaltenbach LA, Baklanov DV, Kim LK, Wong SC, Minutello RM, Messenger JC, Moussa I, Garratt KN, Piana RN, Hillegass WB, Cohen MG, Gilchrist IC, Rao SV. Adoption of radial access and comparison of outcomes to femoral access in percutaneous coronary intervention: an updated report from the national cardiovascular data registry (2007–2012). Circulation 2013;127:2295–2306. [DOI] [PubMed] [Google Scholar]
- 18.Subherwal S, Peterson ED, Dai D, Thomas L, Messenger JC, Xian Y, Brindis RG, Feldman DN, Senter S, Klein LW, Marso SP, Roe MT, Rao SV. Temporal trends in and factors associated with bleeding complications among patients undergoing percutaneous coronary intervention: a report from the National Cardiovascular Data CathPCI Registry. J Am Coll Cardiol 2012;59:1861–1869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Stone GW, Witzenbichler B, Guagliumi G, Peruga JZ, Brodie BR, Dudek D, Kornowski R, Hartmann F, Gersh BJ, Pocock SJ, Dangas G, Wong SC, Kirtane AJ, Parise H, Mehran R, Investigators HORIZONS-AMI Trial. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med 2008;358:2218–2230. [DOI] [PubMed] [Google Scholar]
- 20.Vranckx P, Valgimigli M, Odutayo A, Serruys PW, Hamm C, Steg PG, Heg D, Mc Fadden EP, Onuma Y, Benit E, Janssens L, Diletti R, Ferrario M, Huber K, Räber L, Windecker S, Jüni P. GLOBAL LEADERS Investigators. Efficacy and safety of ticagrelor monotherapy by clinical presentation: pre-specified analysis of the GLOBAL LEADERS trial. J Am Heart Assoc 2021;10:e015560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gao C, Buszman P, Buszman P, Chichareon P, Modolo R, Garg S, Takahashi K, Kawashima H, Wang R, Chang CC, Kogame N, Tomaniak M, Ono M, Hara H, Slagboom T, Aminian A, Naber CK, Carrie D, Hamm C, Steg PG, Onuma Y, Geuns RV, Serruys PW, Zurakowski A. Influence of bleeding risk on outcomes of radial and femoral access for percutaneous coronary intervention: an analysis from the GLOBAL LEADERS trial. Can J Cardiol 2021;37:122–130. [DOI] [PubMed] [Google Scholar]