New Biomarker-Based Risk Score Performs Well in ACS

A new biomarker-based risk score accurately predicts recurrent events in patients discharged after ACS and does so better than currently used risk-prediction tools, such as GRACE 2.0, according to a new study published this week.

The ABC-ACS risk score, which is calculated using the patient’s age (A), biomarkers (B), and clinical characteristics (C), showed good discriminatory ability for the 1-year risk of CV death or MI and performed slightly better when it came to predicting the 1-year risk of CV death than MI.

“We’ve been working on combinations of clinical characteristics and measurements of biomarkers as indicators of underlying biology that could be useful and improve risk prediction and provide decision support tools,” senior investigator Lars Wallentin, MD, PhD (Uppsala University, Sweden), told TCTMD, noting that the best predictors of clinical events in the risk score are increased concentrations of the biomarkers. 

“These are stronger than any clinical characteristics, and that’s why we combined them,” continued Wallentin. “Currently, in the guidelines, there is a recommendation about balancing the intensity and duration of antithrombotic treatment after ACS based on the estimated risk of a new ischemic event and the risk of bleeding, but there are very few reliable tools.”

Risk stratification is particularly needed after coronary revascularization, rather than before it, because most ACS patients undergo invasive angiography and PCI, the researchers say. For that reason, they wanted to develop a tool to help tailor secondary prevention based on information gleaned from angiography and laboratory testing. To date, the group have already developed similar biomarker-based ABC-AF scores to predict the risk of stroke, major bleeding, and mortality in atrial fibrillation (AF) patients, as well as an ABC risk-prediction calculator developed for patients with stable CAD.

The new study is published online October 24, 2022, in the Journal of the American College of Cardiology. 

Based on PLATO, Validated in TRACER

The risk score was developed using data from the PLATO trial and based on 8,785 person-years of follow-up. A wide range of possible variables for the risk score were entered into a multivariable Cox regression model, with an algorithm culling the least important variables to leave age, two biomarkers—growth differentiation factor 15 (GDF-15) and N-terminal pro-brain natriuretic peptide (NT-proBNP)—and several clinical factors, such as extent of coronary artery disease, previous vascular disease, and Killip class, as well as type of ACS and type of P2Y12 inhibitor.

Overall, ABC-ACS had a C-index of 0.71 (95% CI 0.69-0.74), reflecting the probability that the risk model predicted patients who would have an event, and had good calibration. The model performed well in patients with STEMI and NSTE ACS, with C-indices of 0.73 (95% CU 0.71-0.76) and 0.69 (95% CI 0.65-0.72), respectively. It also performed well regardless of the type of P2Y12 inhibitor (clopidogrel or ticagrelor) used for treatment. The ABC-ACS ischemic score classes (low < 5%, intermediate 5% to 10%, and high > 10%) also had good discrimination for predicting the 1-year incidence of CV death and MI. 

Investigators validated the risk score using the TRACER cohort, which included 4,896 years of follow-up. In the external cohort, the C-index was 0.72 (95% CI 0.68-0.76). Like in the PLATO cohort, the ABC-ACS ischemia score showed good discrimination for predicting CV death/MI across patients at low, intermediate, and high risk of events. The ABC-ACS ischemia score outperformed the GRACE 2.0 score for the evaluated clinical outcomes, with investigators reporting a significantly higher C-index in the TRACER validation cohort.

“We have very good estimates of ischemic risk, with good discrimination, and above all with good calibration,” said Wallentin. “With calibration, it’s more stable with biomarkers than with using clinical history.” 

Researchers also performed a “decision-curve” analysis to assess the usefulness of ABC-ACS when compared with GRACE 2.0. This analysis is used to calculate the net benefit of using one prediction model for risk stratification compared to the default strategy. Compared with GRACE 2.0, the ABC-ACS ischemia score had a higher net benefit across different decision threshold probabilities (ie, the range of ischemic risk). There was a consistent relative reduction in risk of CV death/MI with clopidogrel and ticagrelor, but there was a larger absolute benefit with ticagrelor in patients with higher ABC-ACS ischemia scores.

Clinicians have a number of ways to treat patients based on the risk of ischemic events, said Wallentin. There are several different antiplatelet agents that can be combined with aspirin—prasugrel, ticagrelor, and clopidogrel—and patients can be treated with dual antiplatelet therapy anywhere from 1 month to 1 year, depending on risk.

“We think [ABC-ACS] is a useful tool because it provides you with an estimate of risk over a period of time,” and that could help to tailor antiplatelet intensity, he said.

What About Bleeding Risk?

In an editorial, Serge Korjian, MD, Robert W. Yeh, MD, and Donald E. Cutlip, MD (all Beth Israel Deaconess Medical School, Boston, MA), point out that several risk scores have been developed to predict residual risk after ACS, including the GRACE and TIMI scores. As PCI became the standard of care, other scores were developed to identify patients at higher risk of recurrent events, as well as to help balance the risks of bleeding and ischemic complications. These scores, which include GRACE 2.0, DAPT, and PRAISE, also include biomarker data, procedural characteristics, and machine-learning algorithms to improve risk prediction but their use in routine practice is limited.   

The editorialists say the ABC-ACS ischemia score adds to greater understanding about risk prediction after ACS, but note that it has limitations. It needs to still be validated in a real-world population before it’s adopted more broadly. Also, the score used a biomarker—GDF-15—that isn’t widely available in clinical laboratories, which they say could limit usability.

To TCTMD, Wallentin said they discovered GDF-15 roughly 15 years ago and have previously shown that it is strongly associated with bleeding. It has also been shown to be associated with a risk of major CV events in patients with ACS and chronic coronary syndrome. GDF-15 is a cytokine related to cellular stress and aging that the researchers say is only weakly expressed in healthy adults. However, it is released into the bloodstream in response to hypoxia, reactive oxygen species, or inflammation.  

“It seems that GDF-15 might be a marker of biological age,” said Wallentin. “It appears to be a better marker of biological rather than chronological age.” He acknowledged, however, that GDF-15 is not routinely measured because there is no CV indication for its use, but said it is being tested in clinical trials.

One other drawback, according to the editorialists, is that the ABC-ACS has not been validated for bleeding prediction; Wallentin said, however, they have already developed an ABC score to predict bleeding in ACS patients, which they have submitted it to a medical journal for publication.