Predictive value of the thrombotic risk criteria proposed in the 2023 ESC guidelines for the management of ACS: insights from a large PCI registry

Abstract

Aim

To assess the value of the thrombotic risk criteria proposed in the 2023 guidelines of the European Society of Cardiology (ESC) for the management of acute coronary syndrome (ACS) to predict the ischaemic risk after percutaneous coronary intervention (PCI).

Methods and results

Consecutive patients with acute or chronic coronary syndrome undergoing PCI at a large tertiary-care center from 2014 to 2019 were included. Patients were stratified into low, moderate, or high thrombotic risk based on the ESC criteria. The primary endpoint was major adverse cardiovascular events (MACEs) at 1 year, a composite of all-cause death, myocardial infarction (MI), and stroke. Secondary endpoints included major bleeding. Among 11 787 patients, 2641 (22.4%) were at low-risk, 5286 (44.8%) at moderate risk, and 3860 (32.7%) at high-risk. There was an incremental risk of MACE at 1 year in patients at moderate (hazard ratios (HR) 2.53, 95% confidence interval (CI) 1.78–3.58) and high-risk (HR 3.39, 95% CI 2.39–4.80) as compared to those at low-risk, due to higher rates of all-cause death and MI. Major bleeding rates were increased in high-risk patients (HR 1.59, 95% CI 1.25–2.02), but similar between the moderate and low-risk group. The Harrell's C-index for MACE was 0.60.

Conclusion

The thrombotic risk criteria of the 2023 ESC guidelines for ACS enable to stratify patients undergoing PCI in categories with an incremental 1 year risk of MACE; however, their overall predictive ability for MACE is modest. Future studies should confirm the value of these criteria to identify patients benefiting from an extended treatment with a second antithrombotic agent.

Graphical Abstract

Open in new tabDownload slide

Introduction

The decision on antithrombotic regimen after percutaneous coronary intervention (PCI) is complex and requires an appropriate evaluation of the bleeding and ischaemic risk of the patient.^1,2 Occurrence of these complications may significantly vary across the wide spectrum of patients undergoing PCI and preprocedural risk assessment can facilitate the decision on antithrombotic therapy.^1,2 During the past years, several studies demonstrated the need to adopt bleeding avoidance strategies, including shorter or less potent antithrombotic regimen, especially in patients at high bleeding risk (HBR).² However, some patients not at HBR might derive benefit from an extended treatment with a second antithrombotic agent.^1,3,4 While a consensus was reached on the HBR criteria, there is no universally accepted definition or criteria of ischaemic or thrombotic risk for patients undergoing PCI.^4–6 Several tools have been developed for ischaemic risk stratification or to inform specific management decisions for patients undergoing PCI^7–11; however, only few aimed at identifying patients for whom an extended or more potent antithrombotic treatment is indicated.^7,12

In the 2020 guidelines of the European society of cardiology (ESC) for the management of patients with non-ST-segment elevation acute coronary syndrome (NSTE-ACS), a combination of clinical and procedural criteria has been proposed to identify patients with a moderate or high thrombotic risk for whom an extended treatment with a second antithrombotic agent might be considered.³ The same criteria have been adopted in the 2023 ESC guidelines for the management of ACS.⁵ The value of these criteria for risk stratification prior to PCI has, to date, not been established.

The aim of the current study is to evaluate the ability of these criteria to predict adverse events in patients with any clinical presentation undergoing PCI.

Methods

Study population

Consecutive patients who underwent PCI at a large-volume tertiary-care center (Mount Sinai Hospital, New York, NY, USA) and provided their informed consent were included in the institutional database. Baseline and procedural data were prospectively collected. Follow-up was carried out through in-person visits or telephone calls by dedicated research personnel at 30 days and at 1 year after PCI. Vital status was additionally assessed using hospital medical records, contacting primary care physicians, and through external death records such as the on-line obituary and Social Security Death Index.¹³ The modality of PCI treatment and antithrombotic management were at the treating physician's discretion, with dual antiplatelet therapy (DAPT) usually being usually prescribed for 12-month after ACS and for 6-month after an elective procedure in absence of concomitant long-term oral anticoagulation.

In the current analysis, patients with any clinical presentation who underwent PCI between January 2014 and December 2019 with implantation of a drug eluting stent were included. Based on the criteria of the 2023 ESC guidelines for ACS management to identify patients with increased thrombotic risk who may benefit from an extended treatment with a second antithrombotic agent, the study population was stratified into 3 groups: low, moderate, or high risk.³ The high-risk group consisted of patients meeting at least one of the clinical risk criteria [diabetes mellitus, chronic kidney disease (CKD), defined as glomerular filtration rate (GFR) between 15 and 60 mL/min, prior myocardial infarction (MI), peripheral artery disease (PAD), age <45 years, multivessel coronary artery disease (CAD)] and at least one of the procedural risk criteria (≥3 stents implanted, ≥3 lesions treated, total stent length >60 mm, left main PCI, chronic total occlusion (CTO), bifurcation lesion treated with 2 stents). Patients in the moderate-risk group met at least one of the following clinical risk criteria: diabetes mellitus, CKD, prior MI, or PAD. The low-risk group included patients that did not fulfil the criteria for the high or moderate risk group (Central Figure). Three thrombotic risk criteria of the 2023 ESC guidelines for ACS (accelerated CAD, concomitant systemic inflammatory disease, and history of stent thrombosis) were not available in the database.

Patients with GFR <15 mL/min, with missing information on the ESC criteria and those lost to follow-up without experiencing adverse events in the first 30 days after PCI were excluded.

Study outcomes and definition

The primary outcome was major adverse cardiovascular events (MACEs), a composite of all-cause death, MI or any stroke occurring between the index procedure and 1 year after PCI. Secondary outcomes included the individual component of the primary outcome, target vessel revascularization (TVR) and major bleeding. MI was defined according to the Third Universal Definition and only spontaneous events were considered (i.e. MI related to PCI or coronary artery bypass graft [CABG] were excluded).¹⁴ Major bleeding included in-hospital and post-discharge bleeding (Supplementary material online, Methods).¹⁵ HBR was defined according to the Academic Research Consortium (ARC) definition.^6,16

Statistical analysis

Variables were presented as numbers and percentages if categorical and compared using the chi-square test. Continuous data were presented as mean ± standard deviation and compared using the Student's t-test. Clinical event rates at 1 year were obtained using the Kaplan—Meier method and compared using the log-rank test for the time to the first event. Crude risks estimates were calculated with a Cox regression model, using the low-risk group as reference. Risks were expressed as hazard ratios (HR) and 95% confidence intervals (CIs). The Harrell's concordance statistic was calculated to assess the discrimination of the criteria to predict the primary and secondary outcomes at 1 year in patients with low, moderate, and high thrombotic risk. Moreover, the risks of the primary and secondary outcomes and the Harrell's concordance statistic were assessed in patients presenting with ACS, CCS and non-HBR. In addition, the Harrell's concordance statistic with respect to MACE was calculated for the complex PCI criteria proposed by Giustino et al.,¹⁷ for the PARIS thrombotic score ,⁷ and the DAPT score.¹⁸ All probability testing was 2-sided, and a P-value of <0.05 was considered statistically significant for all tests. All data were analysed using Stata version 16 (StataCorp, College Station, TX, USA).

Results

Population characteristics

From 2014 to 2019, 21 476 PCI were performed at Mount Sinai Hospital (Supplementary material online, Figure S1). After exclusion of duplicate procedures or patients meeting the exclusion criteria, the final study population consisted of 11 787 patients, of whom 2641 (22.4%), 5286 (44.8%), and 3860 (32.7%) patients were identified as low, moderate, and high-risk, respectively (Figure 1). The most frequently met criteria were diabetes and CKD in the moderate risk group, multivessel disease, total stent length >60 mm and diabetes in the high-risk group (Figure 2).

Besides the ESC criteria for thrombotic risk, other clinical and procedural characteristics significantly differed between the 3 risk groups. Patients at moderate risk and at high-risk were older, more likely to have hypertension, dyslipidaemia, prior PCI, prior CABG, prior stroke, atrial fibrillation, anaemia, and lung disease (Table 1). The prevalence of HBR patients was higher in the moderate risk (49.3%) and high-risk (43.7%) than in the low-risk (23.6%) categories.

Moderate and high-risk patients more frequently presented with chronic coronary syndrome or unstable angina and less frequently with an acute MI as compared to low-risk patients. Femoral access, multivessel disease, and moderate/severe calcified lesions were more frequent in moderate and high-risk patients (Supplementary material online, Table S1). Conversely, other complex PCI features, such as B2/C lesions, bifurcation, CTO, and longer total stent length were slightly more prevalent in the low than moderate risk group and the most prevalent in the high-risk group.

Medications prescribed at discharge were similar between the 3 groups, with most of the patients being discharged on P2Y₁₂ inhibitor (>99%), and with clopidogrel being more frequently prescribed in low- and moderate- than in high-risk patients (Table 1).

Clinical outcomes at 1 year

At 1 year, the primary outcome of death, MI, or stroke occurred in 13 patients (1.0%) in the low-risk group, 266 patients (4.4%) in the moderate risk group, and 217 patients (6.1%) in the high-risk group (Figure 2). The risk for the primary outcome was significantly increased in the moderate risk group (HR 2.53; 95% CI, 1.78–3.58) and the high-risk group (HR 3.39; 95% CI, 2.39–4.80) relative to the low-risk group (Figure 3).

There was a stepwise risk increase for all-cause death in the moderate risk (2.0%; HR 3.08; 95% CI, 1.79–4.32) and high-risk (2.9%; HR 4.63; 95% CI, 2.69–7.96) groups as compared to the low-risk group (0.6%). MI occurred significantly more frequently in the moderate risk (2.3%; HR 2.59; 95% CI, 1.60–4.18) and high-risk (2.7%; HR 3.16; 95% CI 1.94–5.13) groups than in the low-risk group (0.9%). Stroke rates were generally low and similar in the low-risk (0.2%), moderate risk (0.3%; HR 1.53; 95% CI 0.56–4.21) and high-risk (0.4%; HR 1.69; 95% CI 0.60–4.80) groups.

The risk of TVR was significantly higher in the moderate (6.4%, HR 1.57; 95% CI, 1.24–1.99) and high (10.3%, HR 2.61; 95% CI 2.07–3.29) risk groups as compared to the low-risk group (4.1%) (Figure 3).

At 1 year, major bleeding events were observed in 96 patients (3.8%) in the low-risk group, 231 patients (4.6%) in the moderate risk group, and 220 (5.9%) patients in the high-risk group. Relative to the low-risk group, the risk for major bleeding was similar in the moderate risk group (HR 1.21; 95% CI 0.95–1.54), while it was significantly increased in the high-risk group (HR 1.59; 95% CI 1.25–2.02).

Overall, the rates of ischaemic and bleeding events were higher in patients presenting with ACS than in CCS or in non-HBR patients (Supplementary material online, Tables 2–4). Among ACS patients, those at moderate and high-risk were significantly more likely to experience adverse events as compared to the low-risk group, while among CCS patients only the high-risk group had an increased hazard of complications. In the non-HBR group, patients with moderate or high-thrombotic risk were at higher risk of all-cause death, MI or stroke, mostly due to an excess of MI.

Predictive ability of adverse events at 1 year

In the overall population, the Harrell's concordance statistic for MACE at 1 year was 0.60 (95% CI 0.57–0.62), while for the secondary outcomes ranged between 0.55 and 0.61 (Figure 4). The predictive ability of the ESC thrombotic risk criteria for MACE, its individual components and major bleeding was similar in patients with ACS, CCS, but poorer in non-HBR patients (Supplementary material online, Table S5); all-cause death was associated with the highest Harrell's C-index (0.61–0.64).

All-cause death, MI or stroke was poorly predicted by the application of the complex PCI criteria of Giustino et al. (Harrell's C-index 0.54, 95% CI 0.51–0.56), and well predicted by the PARIS thrombotic score (Harrell's C-index 0.65, 95% CI 0.62–0.68), and by the DAPT score (Harrell's C-index 0.66, 95% CI 0.64–0.69).

Discussion

To date, there are no universal criteria to identify patients with a higher risk of ischaemic complications, for whom a prolonged or more potent antithrombotic therapy after PCI might be beneficial.^1,4,5 In the current study, we assessed for the first time the predictive value of the thrombotic risk criteria of the 2023 ESC guidelines for the management of ACS in a large all-comers PCI-registry. The main findings can be summarized as follows:

• 1) Patients at moderate or high thrombotic risk were older, more likely to have cardiovascular risk factors and comorbidities in addition to the ESC thrombotic risk criteria, as compared to low-risk patients

• 2) The 1 year risk of MACE, all-cause death, MI or TVR was significantly higher in moderate or high-risk patients than in the low-risk group

• 3) Patients at high-risk, but not those at moderate risk more frequently suffered from major bleeding

Patients with low, moderate and high-risk according to the ESC thrombotic risk criteria represented 22.4, 44.8, and 32.7% of the study population. The most frequently met thrombotic risk criteria were diabetes, CKD, multivessel disease, and total stent length >60 mm. At 1 year after PCI, moderate risk patients had a 2–3 times higher risk of MACE, all-cause death, MI and a nearly 1.6 times increased hazard of TVR; the high-risk group had an about 3–4 times higher risk of MACE, all-cause death or MI and were about 2.6 times more likely to suffer from TVR. However, the ability of the ESC criteria to predict MACE or any individual ischaemic event at 1 year was modest (Harrell's c-index between 0.55 and 0.61), and consistent when assessed separately in ACS and CCS patients, but poorer in non-HBR patients.

The ESC thrombotic risk criteria were chosen mostly based on the results of previous large RCTs that assessed the effect of an extended treatment with a second antithrombotic agent (i.e. P2Y12 inhibitor or low-dose rivaroxaban), namely the DAPT, COMPASS and PEGASUS-TIMI 54 trial,^18–22 and some observational studies.^23–25 One limitation of the ESC definition is that multivessel disease or other procedural risk criteria (i.e. complex PCI features), which were identified as independent predictors of adverse outcomes in previous analyses, are considered high-risk criteria only in combination with clinical risk factors, such as CKD or diabetes.^17,26–28 Moreover, severely calcified lesion requiring atherectomy or PCI of saphenous vein graft were not included among the procedural risk criteria, even though these features are associated with worse outcomes.^29–31 Furthermore, the risk of CKD patients with eGFR <15 mL/min has not been specified in the ESC definition. These limitations might explain at least in part the modest discrimination of this subset of criteria for adverse events prediction (Harrell's C-index 0.60 for MACE).

Of note, other scores conceived to assess the ischaemic risk and to inform DAPT duration showed a slight better predictive ability. The DAPT score, which was derived from the results of the DAPT trial to identify patients deriving a benefit from a 30- rather than 12-month DAPT after PCI using 9 variables had a moderate discrimination in our study (C-index 0.66), value that was consistent with previous reports (C-index 0.64).^12,18,32 Similarly, the PARIS ischaemic score, whose aim was to predict the composite outcome of stent thrombosis or MI and to facilitate clinical decisions on DAPT duration using 6 variables showed a moderate predictive value (C-index 0.65), which was slightly higher than in a previous study (C-index 0.61).⁷ The application of the complex PCI criteria proposed by Giustino et al.¹⁷ had a poor predictive ability.

Other risk assessment tools to predict mortality or ischaemic outcomes but not inform the type of antithrombotic regimen after PCI are limited by the absence of some important clinical or procedural variables, the need of complex calculation, failure to predict all relevant fatal or non-fatal cardiovascular events, or missing external validation.^7–11,33 A recent score developed from the British Cardiovascular Intervention Society (BCIS) database to identify complex high-risk indicated PCI (CHIP) and, including 7 clinical and 6 procedural factors had a better predictive ability for events in hospital (0.65) and at 1 year (0.70).^13,29 However, as for the great majority of the above-mentioned risk assessment tools, the ability of the BCIS-CHIP score to inform DAPT duration remains uncertain.

Interestingly, in the current study we observed that the 1 year rate of major bleeding was relatively high in the 3 risk groups (ranging between 4 and 6%) and significantly increased in high-risk patients due to the high prevalence of ARC-HBR criteria in this group (43.7%).

In our study, around 75% of patients was at moderate or high thrombotic risk, but nearly half of them met the ARC-HBR criteria; that means that roughly one third of the study population might have been eligible for an extended treatment with a second antithrombotic agent. Indeed, in presence of HBR, shorter or less potent antithrombotic regimens are recommended irrespective of concomitant ischaemic risk.^1,2,4,5

An effective and safe treatment option for patient at high risk for both ischaemic and bleeding events is P2Y₁₂ monotherapy.³⁴ The TWILIGHT trial showed that in such patients DAPT continuation beyond 3 months after PCI does not decrease ischaemic events and causes bleeding harm as compared to ticagrelor monotherapy.³⁵ In addition, according to data from randomized controlled trials, a chronic maintenance with P2Y₁₂ monotherapy might be safer and more effective than long-term aspirin.^36–39

Additional studies are needed to validate tools aiming at identifying patients without HBR features but high ischaemic risk, and to compare the different antithrombotic options available for this vulnerable population.

Limitations

Several limitations should be considered when interpreting the results of the current study. First, the study might suffer from limited generalizability, since it was conducted at a single albeit large volume center. Given the observational nature of the study, underestimation of adverse events during follow-up cannot be excluded, even though data were prospectively collected. The ESC thrombotic risk criteria were developed to stratify the risk of patients beyond 1 year after PCI; in the present study, follow-up was limited to 1 year after PCI; therefore, the ability of the criteria to predict adverse outcomes after 1 year could not be assessed. Three less frequent ESC thrombotic risk criteria were not available (i.e. accelerated CAD, systemic inflammatory disease, history of stent thrombosis on DAPT treatment). Moreover, since data on DAPT duration or adherence over time were not available, no conclusion can be drawn on the appropriateness of the ESC thrombotic risk criteria to inform DAPT duration. Finally, causes of death are not available.

Conclusions

The thrombotic risk criteria of the 2023 ESC guidelines for ACS enabled to stratify patients undergoing PCI in categories with an incremental risk of all-cause death, MI or stroke at 1 year; however, their predictive value with respect to these events was modest. Future studies should confirm the value of these criteria to identify patients deriving a net benefit from an extended treatment with a second antithrombotic agent.

Funding

This study did not receive any external funding and was supported by the Zena and Michael A. Wiener Cardiovascular Institute Icahn School of Medicine at Mount Sinai New York NY. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Conflict of interest

A.S. received a research grant from the Swiss National Science Foundation. G.D. has received consulting fees and advisory board fees from AstraZeneca; has received consulting fees from Biosensors; and previously held stock in Medtronic. A.N.K. is a post-doctoral fellow with funding from the Heart Foundation of Australia. R.M. reports institutional research payments from Abbott, Abiomed, Alleviant Medical, Amgen, AM-Pharma, Arena, AstraZeneca, Biosensors, Biotronik, Boston Scientific, Bristol Myers Squibb, CardiaWave, CeloNova, Chiesi, Concept Medical, Cytosorbents, Daiichi Sankyo, Element Science, Faraday, Humacyte, Idorsia Pharmaceuticals, Janssen, Magenta, Mediasphere, Medtelligence, Medtronic, Novartis, OrbusNeich, Penumbra, PhaseBio, Philips, Pi-Cardia, PLx Pharma, Protembis, RenalPro, RM Global, Shockwave, Vivasure, Zoll; personal fees from AstraZeneca, Ionis Pharmaceuticals, Novartis, Novo Nordisk, Vectura, WebMD; Equity <1% in Applied Therapeutics, Elixir Medical, Stel, ControlRad (spouse); AMA (Scientific Advisory Board, JAMA Cardiology Associate Editor), ACC (BOT Member, SC Member CTR Program), SCAI (Women in Innovations Committee Member); Faculty CRF (no fee).

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References