How to Assess Bleeding Risk in Patients Undergoing Percutaneous Coronary Interventions

Abstract

A relevant proportion of patients undergoing percutaneous coronary intervention (PCI) have a high risk of bleeding. The associated individual risk of ischaemia can be differentiated by an app-based approach and helps to determine the duration of intensified antithrombotic therapy.

Background and Aim

The choice of the optimal antithrombotic therapy and its duration after percutaneous coronary intervention (PCI) remains challenging [1]. From the perspective of ischemic risk, the type and duration of dual antithrombotic therapy (DAPT) depends on the clinical setting (e.g., acute coronary syndrome versus chronic coronary artery disease) and the complexity of the coronary anatomy and intervention (e.g., stent length and diameter, bifurcation treatment) [2,3]. Prolongation of intensified antithrombotic therapy, i.e., DAPT, may reduce the risk of ischemia, but is associated with an increased risk of bleeding and thus potentially increased mortality [1]. Current European and American guidelines recommend a DAPT duration of six months after PCI and implantation of drug-eluting stents for chronic coronary syndrome—an extension to twelve months for patients who remain free of bleeding complications and carry a high ischemic risk and a reduction to 1–3 months for high bleeding risk [3,4]. Both professional societies recommend a DAPT duration of twelve months for acute coronary syndrome with a possible extension for low bleeding risk and a reduction to six months for high bleeding risk [3,4]. In 2019, an initiative of an international expert panel named Bleeding Academic Research Consortium (BARC) on the Assessment of High Bleeding Risk (HBR) in PCI sought to define the available parameters for defining a high risk of bleeding after PCI based on the evidence to date [5]. In a subsequently published analysis, a first validation was performed and from this a two-stage scoring was modelled [6]. The aim of this brief overview is to present the importance of the standardized definition of high bleeding risk in the context of a subsequent validation that can be applied in daily clinical practice.

Underrepresentation of the High Bleeding Risk PCI Group in Previous Scores

Various scores derived from different population to assess the individual risk of bleeding and ischemia have been published [7,8,9,10,11,12,13,14]. As stated by the BARC and presented in detail in

[5], low to intermediate bleeding risk PCI populations were included. In many studies evaluating the safety of a shortened period of DAPT, patients with a high risk of bleeding were excluded and bleeding was defined differently in different studies. In these studies, the annual bleeding rate was found to range between 0.3 and 2.8% [5]. Therefore, the applicability of bleeding risk scores remains limited and due to the different collectives on which the models are based, they are most applicable to a low or medium risk collective [15].

Standardization of Criteria for High Bleeding Risk (HBR) Based on HBR PCI Population

As stated above, in 2019, the BARC sought to standardize the criteria for HBR [5], based on three prospective randomized trials (completed at the time), that investigated a short duration of DAPT in a patient population undergoing PCI that was assumed to be at high risk of bleeding [16,17,18]. The rate of BARC 3–5 bleeding [19] within one year after PCI ranged from 3.5 to 7.2% in these studies. The threshold HBR was set at an annual BARC 3–5 bleeding rate of more than 4% (for details see Urban et al. [5]). The definition of BARC 3 bleeding is an overt bleeding with a drop in hemoglobin (Hb) of at least 3 g/dl or an intracranial bleeding. BARC 4 bleeding is defined as a bleeding event within 48 h after coronary bypass grafting and BARC 5 refers to fatal bleeding.

Twenty different clinical risk factors were defined by consensus as major and minor risk factors for bleeding based on the findings of the above-mentioned studies [16,17,18]. Patients are considered at HBR if at least one major or two minor risk factors are present (

) [5].

The Arc-HBR Trade O Model [6]: A Two-Step Risk Stratification Approach to Individualize the Antithrombotic Management in PCI Patients with HBR

In the clinical context of percutaneous coronary interventions, the central question is how to find an adequate antithrombotic therapy in the individual patient situation to assess the risk of ischemic events in an individual HBR patient with an expected increased risk of bleeding. With the aim of improving the prognostic value for future ischemic or bleeding events, a group of researchers [6] undertook new modelling and compared the results with the prognostic performance for bleeding and ischemia of previous well-established scores, including DAPT [10], PRECISE-DAPT [12] and PARIS [11]. Patients from six PCI studies conducted between 2009 and 2017 served as the basis for a new model for ischemic risk assessment and a model for bleeding risk assessment after PCI. A total of 6641 HBR patients were identified and their actual incidence of bleeding and stent thrombosis (ST) or myocardial infarction (MI) within one year was recorded. Among these 6641 patients, 350 (5.3%) had either a MI and/or ST and 381 patients (5.7%) suffered BARC 3 to 5 bleedings [6]. Eight independent baseline predictors of risk of MI- and/or ST-elevation, eight predictors for risk of major bleeding (BARC 3 to 5) and four predictors of both MI/ST and major bleeding were identified (Table 3). The C-statistic of 0.74, expressing prognostic performance for bleeding, was better than for the PARIS score (C-statistic 0.63) and also better than for the PRECISE-DAPT score (C-statistic 0.64). Similar better prognostic performance was found for MI and/or ST [6]. This superior performance was probably primarily due to the different cohorts on which the models are based. The results of this study also show that the HBR collective is not homogeneous: in 23.4% the risk of BARC 3–5 bleeding is higher than that of ischemia, in 32.4% the risk of BARC 3–5 bleeding and ischemia are similar and in 44.1% the risk of ischemia is higher than that of BARC 3–5 bleeding [6].

Based on this data, the study group developed an app-based staged assessment of HBR patients to determine the risk of BARC 3–5 bleeding on the one hand and the risk of ischemia on the other. For illustration purposes, three exemplary HBR patient cases were presented to illustrate [20].

Table 3. Multivariate predictors of BARC Type 3–5 bleeding and MI and/or ST at one year [6].

	BARC type 3–5 bleeding		MI and/or ST
Predictor	HR (95% CI)	p-value	HR (95% CI)	p-value
Age > 65	1.50 (1.08–2.08)	0.01	NA	NA
Diabetes	NA	NA	1.56 (1.26–1.93)	<0.001
Prior MI	NA	NA	1.89 (1.52–2.35)	<0.001
Liver disease, cancer or surgery	1.63 (1.27–2.09)	0.0001	NA	NA
COPD	1.39 (1.05–1.83)	0.02	NA	NA
Current smoker	1.47 (1.08–1.99)	0.01	1.48 (1.09–2.01)	0.009
NSTEMI or STEMI presentation	NA	NA	1.82 (1.46–2.25)	<0.001
Haemoglobin, g/dl
>13	1 (Reference)		1 (Reference)
11–12,9	1.69 (1.30–2.20)	<0.001	1.27 (.99–1.63)	0.005
<11	3.99 (3.06–5.20)		1.50 (1.12–1.99)
eGFR, ml/min/1.73 m2
>60	1 (Reference)		1 (Reference)
30–59	0.99 (0.79–1.24)	0.02	1.30 (1.03–1.66)	0.001
<30	1.43 (1.04–1.96)		1.69 (1.20–2.37)
Complex procedureb	1.32 (1.07–1.61)	0.008	1.50 (1.21–1.85)	<0.001
Bare metal stentc	NA	NA	1.53 (1.23–1.89)	<0.001
OAC at discharge	2.00 (1.62–2.48)	<0.001	NA	NA
C statistic	0.68	NA	0.69	NA

BARC, Bleeding Academic Research Consortium; CI, Confidence interval; COPD, Chronic obstructive pulmonary disease; eGFR, Estimated glomerular filtration rate; HR, Hazard ration; MI, Myocardial infarction; NA, Not applicable; NSTEMI, Non-ST-segment elevation MI; OAC, Oral anticoagulants; ST, Definite or probable stent thrombosis; STEMI, ST-segment elevation MI; SI, conversion factor: to convert hemoglobin to grams per liter multiply by 10.0. a At least 1 of 3 modified ARC high bleeding risk criteria (cancer, severe liver disease, and planned major surgery; b As defined by Giustino et al. [21]; c Compared with drug eluting stents or drug coated balloons.

Table 3. Multivariate predictors of BARC Type 3–5 bleeding and MI and/or ST at one year [6].

	BARC type 3–5 bleeding		MI and/or ST
Predictor	HR (95% CI)	p-value	HR (95% CI)	p-value
Age > 65	1.50 (1.08–2.08)	0.01	NA	NA
Diabetes	NA	NA	1.56 (1.26–1.93)	<0.001
Prior MI	NA	NA	1.89 (1.52–2.35)	<0.001
Liver disease, cancer or surgery	1.63 (1.27–2.09)	0.0001	NA	NA
COPD	1.39 (1.05–1.83)	0.02	NA	NA
Current smoker	1.47 (1.08–1.99)	0.01	1.48 (1.09–2.01)	0.009
NSTEMI or STEMI presentation	NA	NA	1.82 (1.46–2.25)	<0.001
Haemoglobin, g/dl
>13	1 (Reference)		1 (Reference)
11–12,9	1.69 (1.30–2.20)	<0.001	1.27 (.99–1.63)	0.005
<11	3.99 (3.06–5.20)		1.50 (1.12–1.99)
eGFR, ml/min/1.73 m2
>60	1 (Reference)		1 (Reference)
30–59	0.99 (0.79–1.24)	0.02	1.30 (1.03–1.66)	0.001
<30	1.43 (1.04–1.96)		1.69 (1.20–2.37)
Complex procedureb	1.32 (1.07–1.61)	0.008	1.50 (1.21–1.85)	<0.001
Bare metal stentc	NA	NA	1.53 (1.23–1.89)	<0.001
OAC at discharge	2.00 (1.62–2.48)	<0.001	NA	NA
C statistic	0.68	NA	0.69	NA

BARC, Bleeding Academic Research Consortium; CI, Confidence interval; COPD, Chronic obstructive pulmonary disease; eGFR, Estimated glomerular filtration rate; HR, Hazard ration; MI, Myocardial infarction; NA, Not applicable; NSTEMI, Non-ST-segment elevation MI; OAC, Oral anticoagulants; ST, Definite or probable stent thrombosis; STEMI, ST-segment elevation MI; SI, conversion factor: to convert hemoglobin to grams per liter multiply by 10.0. a At least 1 of 3 modified ARC high bleeding risk criteria (cancer, severe liver disease, and planned major surgery; b As defined by Giustino et al. [21]; c Compared with drug eluting stents or drug coated balloons.

Three major patient profiles can be distinguished: First, the risk of MI/ST exceeds the risk of major bleeding. Second, the risk of major bleeding exceeds the risk of MI/ST. Third, the risks of MI/ST and major bleeding are similar.

Figure 1 shows three of our own patient examples, which were calculated according to the ARC HBR app as follows:

Patient 1 is a 63-year-old man with an ischemic stroke one year ago, persistent atrial fibrillation, mild anemia (Hb 11.9 g/dl), chronic renal failure (estimated glomerular filtration rate [eGFR] 29 mL/min/1.73 m²), need for long-term oral anticoagulation (novel oral anticoagulant [NOAC] apixaban twice daily 2.5 mg), diabetes, non-ST segment elevation MI presentation and very complex PCI. In addition to apixaban twice daily 2.5 mg, antithrombotic therapy was supplemented after PCI with clopidogrel 75 mg per day for one year and acetylsalicylic acid 100 mg daily for one week. The estimated BARC 3–5 bleeding risk after PCI (day 3–365) is 8.6% and the estimated risk of MI or ST (day 3–365) is 21.6% according to the ARC HBR app. The risk of MI/ST clearly exceeds the risk of bleeding, although the patient obviously has important bleeding risk factors. In principle, an extension of dual antithrombotic treatment beyond the twelve-month period should be considered, as the net benefit outweighs the increased risk of bleeding.

Patient 2 is a 77-year-old woman with stable angina pectoris CCS class II. An uncomplicated PCI with implantation of a drug-eluting stent is performed. Comorbidities include mild anemia (Hb 10.7 g/dl), diabetes and mild chronic renal failure (eGFR 47 mL/min/1.73 m2). The estimated BARC 3–5 bleeding risk after PCI (day 3–365) is 8.6% and the estimated risk of MI or ST (day 3–365) is 2.1% according to the ARC HBR app. Therefore the risk of bleeding is greater than the risk of MI or ST. The risk of bleeding increased significantly and at the same time the risk of ischemia remains low. A shortening of the usual six-month dual antithrombotic therapy in stable coronary artery disease and stent implantation can be considered.

Patient 3 is a 51-year-old man with NSTEMI, straightforward single drug-eluting stent PCI of a first marginal branch, mild anemia (Hb 12.3 g/dl), liver cirrhosis and currently smoking. The estimated BARC 3–5 bleeding risk after PCI (day 3–365) is 5.5% and the estimated risk of MI or ST (day 3–365) is 3.4% according to the ARC HBR app. In this patient, the risks of bleeding and MI/ST can be considered equivalent, thus a normal twelve months of dual antithrombotic therapy should be given.

These examples demonstrate that the estimation of the absolute and relative risk of bleeding and MI and/or ST following PCI with the help of the ARC-HBR score can help to improve clinical decision-making for individual patients.

In summary, this defines three collectives (highlighted by the three patient examples) where a) a prolongation/intensification of antithrombotic therapy should be considered if there is a disproportionate risk of ischemia, b) a shortening/simplification of antithrombotic therapy must be considered if there is a disproportionate risk of bleeding or c) the risk of ischemia and bleeding are more or less balanced. Accordingly, in this latter group DAPT can be given according to the guidelines, twelve months for acute coronary syndrome and 3–6 months after elective implantation of drug-eluting stents [6].

Keypoints

• Approximately 40% of an all-comers percutaneous coronary intervention population meet the criteria for high post-interventional bleeding risk (HBR) with dual antiplatelet therapy.
• The balance between bleeding and thrombotic risk varies from HBR patient to HBR patient.
• In contrast to previous scores, a currently available model with stepwise app-based scoring (ARC HBR App) for bleeding and ischemia is based on a HBR population.
• This scoring should be applied individually to patients at high risk of bleeding and can contribute to improved risk stratification.