This was a retrospective observational study based on the CardioCHUVI (Cardiología del Complejo Hospitalario Universitario de VIgo) ACS registry (NCT03664388), in which the study participants were all patients with a primary and final diagnosis of ACS consecutively admitted to our department between January 2012 and September 2018. Secondary acute coronary syndromes during a hospital admission for another cause were not included in the registry. Demographic, clinical and angiographic variables, as well as those relating to management and complications were gathered prospectively by the department's cardiologists. Patients were treated according to their main physician's criteria.

All patients were included in administrative databases of the regional electronic medical records system at both hospital and outpatient level. This system allowed recording of clinical and laboratory data, pharmacological treatment, bleeding events, and vital status. Electronic medical records were meticulously reviewed by 2 independent clinicians at the adjudication of events of interest.

For this study, we comprehensively reviewed the variables related to the ARC-HBR criteria and retrospectively retrieved data on some ARC-HBR criteria which were not initially available in our database (table 1 of the supplementary data). This approach enabled us to apply the 20 ARC-HBR criteria (14 major and 6 minor) as originally described,11 except for the minor criterion “oral nonsteroidal anti-inflammatory drugs or steroid use”, which was defined as chronic instead of an expected daily intake for ≥ 4 weeks.11

The initial cohort of the present study comprised 5959 patients. For this study, patients were excluded for the following reasons: a) not undergoing in-hospital coronary stenting (1436 patients); b) not receiving DAPT (69 patients); c) missing data on the ARC-HBR criteria (18 patients) or for calculating the PRECISE-DAPT score (11 patients); and d) unavailable data on bleeding or vital status (13 patients). Heparin was used for procedural anticoagulation and periprocedural use of glycoprotein IIb/IIIa inhibitors was left to the discretion of the operator. DAPT duration was determined at the discretion of the treating clinicians.

The study conformed to the principles outlined in the Declaration of Helsinki, and was approved by the local ethics committee (Galicia; code HAC-2017-05, registry 2017/290).

The primary endpoints were the 1-year incidences of first BARC 3 or 5 bleeding and first ICH. Bleeding outcomes included peri-PCI, in-hospital bleeding and postdischarge bleeding. The secondary endpoint was the 1-year incidence of the composite of cardiovascular death, myocardial infarction, or ischemic stroke.

Cardiovascular death and myocardial infarction were defined according to Hicks et al.15 and the Third Universal Definition16), respectively. Diagnosis of ischemic stroke required an acute neurological deficit lasting ≥ 24hours with at least 1 imaging test and neurologist confirmation.

Patients were classified as being at HBR according to the ARC-HBR classification in the presence of ≥ 1 major criterion or ≥ 2 minor criteria. Since the 5-item (hemoglobin, age, leucocytes, serum creatine clearance, and previous bleeding) PRECISE-DAPT score is currently the recommended system (class IIb) for HBR classification by the European Society of Cardiology,2 we used the PRECISE-DAPT HBR classification (≥ 25 vs < 25 points) as a comparator to assess the relative performance of the new ARC-HBR classification.

The ascertainment of endpoints was performed by 2 trained independent cardiologists and was carried out between December and February 2020.

Quantitative data are presented as mean ± SD or median and interquartile range [IQR], as appropriate. Categorical variables are presented as numbers (%). Qualitative variables were compared by the chi-square test or Fisher exact test; quantitative variables were compared by the Student t-test or Mann-Whitney test. Both ARC and PRECISE-DAPT classifications were entered into separate Fine-Gray competing risk models (with death unrelated to bleeding as a competing risk) to test their association with the primary endpoints. The ability of each classification to separate HBR patients from non-HBR patients was checked by cumulative incidence curves and compared using the Fine-Gray test. The discrimination ability of the 2 classifications for predicting bleeding events was estimated via the c-statistic.

To evaluate the prognostic value of the ARC-HBR criteria, we generated and plotted the cumulative incidence for the primary endpoint by the number of the ARC-HBR criteria and examined if each ARC-HBR major or minor criterion, in isolation, met the ARC predefined cutoff for BARC 3 or 5 bleeding risk. Finally, the risk of the secondary endpoint by groups of HBR and non-HBR of ARC was assessed using the Fine-Gray method (with noncardiovascular death as a competing event). The c-statistic, as a discrimination measure at predicting the ischemic event with ARC-HBR, was also calculated.

All analyses were crudes since both the ARC-HBR binary classification and its individual criteria were established to be used without any statistical adjustments. The association between the HBR classifications with outcomes was expressed as subhazard ratios (sHR) with their 95% confidence intervals (95%CI). The bootstrap method (500 iterations) was used to estimate statistical significance, subhazard ratio (sHR), and c-statistic values as well as 95%CIs. Statistical analyses were 2-tailed and were mainly performed using STATA/MP 15.1. R 3.6.3 was used for calculating the c-statistic values via the “pec” package. Statistical significance was set at P < .05.

The present study cohort comprised 4412 ACS patients treated with DAPT and coronary stenting. Table 1 summarizes the differences in clinical and procedural characteristics between patient groups according to the ARC-HBR-classified HBR status.

There were 1303 patients (29.5%) in the ARC-HBR group and 3109 in the ARC non-HBR group. More patients were classified as being at HBR by PRECISE-DAPT (n = 1,774; 40.2%) (McNemar test P-value < .001).

Compared with the ARC non-HBR group, patients in the ARC-HBR group were older, had a higher prevalence of women and cardiovascular as well as noncardiovascular comorbidities. They also had less frequent use of the transradial approach, a higher proportion of multivessel disease, and less often received ticagrelor or prasugrel. Median DAPT duration was shorter in the ARC-HBR group (10.2 months) than in the ARC non-HBR group (12.0 months); P < .001 (table 1).

Table 2 shows the prevalence of the ARC-HBR criteria. The most common major criteria were hemoglobin < 11g/dL (5.0%), oral anticoagulant use (4.6%), spontaneous bleeding (1.5%), active malignancy (1.3%), and estimated glomerular filtration rate < 30mL/min/1.73 m2 (1.3%). The remaining 11 major criteria were rarely observed (≤ 0.5%).

The most common minor criteria were age ≥ 75 years (27.0%), estimated glomerular filtration rate 30 to 59mL/min/1.73 m2 (19.6%), mild anemia (10.7%), oral nonsteroidal anti-inflammatory drug or steroid use (8.5%), and prior ischemic stroke (3.4%). Spontaneous bleeding was rarely observed (0.3%).

Throughout 11.6 ± 1.7 months, 162 patients had a BARC 3 or 5 bleeding event; of them, 70 had periprocedural in-hospital bleeding.

By ARC groups, the overall incidence of BARC 3 or 5 bleeding was significantly higher in the HBR group than in the non-HBR group (9.4% vs 1.3%; P < .001; sHR, 7.3; 95%CI, 5.1-10.4) (figure 1). The rate of periprocedural in-hospital bleeding was 4.3% in the HBR group (vs 0.5%, P < .001; sHR, 6.6; 95%CI, 3.6-12.0), whereas postdischarge bleeding rate was 5.3% (vs 0.9%, P < .001; sHR, 6.1; 95%CI, 3.9-9.5) (figure 2).

Figure 1.

Cumulative incidence for BARC 3 or 5 bleeding by high bleeding risk groups according to the ARC-HBR and the PRECISE-DAPT classifications. 95%CI, 95% confidence interval; ARC-HBR, Academy Research Consortium for High Bleeding Risk; BARC, Bleeding Academy Research Consortium; C =, c-statistic; P-D, PRECISE-DAPT (Predicting Bleeding Complication in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy); sHR, subhazard ratio.

(0.27MB).

Figure 2.

Cumulative incidence for periprocedural in-hospital and postdischarge bleeding by high bleeding risk groups according to the ARC-HBR and the PRECISE-DAPT classifications. 95%CI, 95% confidence interval; ARC, Academy Research Consortium; BARC, Bleeding Academy Research Consortium; C =, c-statistic; HBR, High Bleeding Risk; P-D, PRECISE-DAPT (Predicting Bleeding Complication in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy); sHR, subhazard ratio.

(0.18MB).

By PRECISE-DAPT groups, the overall incidence of BARC 3 or 5 bleeding was significantly higher in the HBR than in the non-HBR group (7.2% vs 1.4%; P < .001; sHR, 5.4; 95%CI, 3.8-7.9) (figure 1). The rate of periprocedural in-hospital bleeding was 3.3% in the HBR group (vs 0.5%; P < .001; sHR, 4.7; 95%CI, 2.5-8.8), while postdischarge bleeding rate was 4.1% (vs 0.9%; P < .001; sHR, 4.6; 95%CI, 2.9-7.3) (figure 2).

Both classifications were significantly associated with BARC 3 or 5 bleeding, albeit ARC showed a more pronounced association overall (sHR = 7.3 vs 5.4), with periprocedural in-hospital bleeding (sHR = 6.6 vs 4.7), and with postdischarge bleeding (6.1 vs 4.6). The c-statistic values also favored ARC overall (C = 0.74 vs 0.70), for periprocedural bleeding (0.73 vs 0.68), and for postdischarge bleeding 0.72 vs 0.68, respectively).

ICH occurred in 26 patients out of the total population. The incidence of ICH was significantly higher in the ARC-HBR group than in the non-HBR group (1.2% vs 0.4%; P = .003; sHR, 3.3; 95%CI, 1.5-7.1). The same pattern was observed using PRECISE-DAPT (1.0% vs 0.3%; P = .005; sHR, 3.3; 95%CI, 1.4-7.6) (figure 3).

Figure 3.

Cumulative incidence for intracranial hemorrhage by high bleeding risk groups according to the ARC-HBR and the PRECISE-DAPT classification systems. 95%CI, 95% confidence interval; ARC, Academy Research Consortium; HBR, high bleeding risk; P-D, PRECISE-DAPT (Predicting Bleeding Complication in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy); sHR, subhazard ratio.

(0.26MB).

Both classification systems were significantly associated with ICH risk to a similar extent (sHR = 3.3 for both), and displayed similar c-statistics (C = 0.65).

The risk of bleeding outcomes increased as the number of minor and major criteria was higher (figure 4): for BARC 3 or 5 bleeding, the risk was 1.8%, 5.0%, 9.4%, 16.8%, 22.5%, and 29.5% for 1 isolated minor criterion, ≥ 2 isolated minor criteria, 1 major criterion (isolated or plus 1 minor criterion), 1 major criterion plus ≥ 2 minor criteria, ≥ 2 major criteria (isolated or plus 1 minor criterion), and ≥ 2 major criteria plus ≥ 2 minor criteria, respectively.

Figure 4.

Cumulative incidence at 1 year for BARC 3 or 5 bleeding and intracranial hemorrhage by the number of the ARC-HBR criteria. ARC-HBR, Academy Research Consortium for High Bleeding Risk; BARC, Bleeding Academy Research Consortium.

(0.31MB).

For ICH, the risk was 0.3%, 0.5%, 0.7%, 1.5%, 2.2%, and 3.9% for 1 isolated minor criterion, ≥ 2 isolated minor criteria, 1 major criterion (isolated or plus 1 minor criterion), 1 major criterion plus ≥ 2 minor criteria, and ≥ 2 major criteria plus ≥ 2 minor criteria, respectively. No ICH events were observed in patients with 2 major criteria (isolated or plus 1 minor criterion).

The 1-year rates of BARC 3 or 5 bleeding associated with each ARC-HBR criteria, in isolation, are shown in figure 5. Sixteen out of 20 ARC criteria (80%) met the predefined cutoffs. That is, among the 14 major criteria, 10 (71.4%) were associated, each in isolation, with a BARC 3 or 5 bleeding risk ≥ 4%. Prior traumatic ICH, recent major surgery/trauma, prior moderate-severe ischemic stroke, and brain arteriovenous malformation were not associated, in isolation, with a BARC 3 or 5 bleeding risk ≥ 4%. All 6 ARC minor criteria were associated, each in isolation, with a BARC 3 or 5 bleeding risk < 4%.

Figure 5.

Cumulative incidence for BARC 3 or 5 bleeding at 1 year by each of the ARC-HBR criteria, in isolation*. ARC-HBR, Academy Research Consortium for High Bleeding Risk; BARC, Bleeding Academy Research Consortium; bAMV, brain arteriovenous malformation; eGFR, estimated glomerular filtration rate; ICH, intracranial hemorrhage; NSAIDs, nonsteroidal anti-inflammatory drugs.

*A complete description for each criterion is provided in table 2.

(0.23MB).

The ischemic endpoint was observed in 432 patients (329 myocardial infarctions, 56 ischemic strokes, and 47 cardiovascular deaths). The incidence of ischemic events was 2-fold higher in the ARC-HBR group than in the ARC non-HBR group (14.3% vs 8.3%; P < .001; sHR, 1.8; 95%CI, 1.5-2.1) (figure 6). The c-statistic for predicting the ischemic endpoint with ARC-HBR was 0.57 (95%CI, 0.55-0.59).

Figure 6.

Cumulative incidence for the ischemic endpoint (cardiovascular death, myocardial infarction, or ischemic stroke) by the ARC-HBR groups. 95%CI, 95% confidence interval; ARC, Academy Research Consortium; HBR, high bleeding risk; sHR, subhazard ratio.

(0.19MB).

The major strengths of this study include the use of the ARC-HBR criteria as originally conceived (excepting oral nonsteroidal anti-inflammatory drugs/steroids use), the concomitant assessment of BARC 3 or 5 bleeding and ICH, and the independent assessment of outcomes. Important limitations, however, need to be addressed. This was a retrospective study with the limitations inherent to this type of design. The single-center design of this study may limit the generalizability of our findings, and the study power was limited to adequately exploring ICH by all individual ARC-HBR criteria.

The findings and conclusions of this study do not apply to the subgroup of patients treated without DAPT (n = 69), as they were excluded from the study to reduce the heterogeneity of the population studied and minimize possible biases.