Advancing cardiovascular risk assessment

Cardiovascular disease remains the leading cause of morbidity and mortality globally despite significant strides in prevention and treatment. Accurate risk prediction is crucial for early intervention and effective management of cardiovascular conditions, which can ultimately reduce the burden of disease and healthcare costs.1 Traditional risk assessment models, while helpful, have limitations in their ability to accurately predict risk across diverse populations and at various stages of life.2 This gap highlights the need for more precise and individualized risk prediction tools. Historically, cardiovascular risk assessment has relied on clinical risk scores that integrate multiple risk factors. The Framingham Risk Score (FRS), introduced nearly 25 years ago, was one of the first tools to consolidate clinical variables into a single risk estimate for coronary artery disease (CAD) and stroke.3 The FRS was followed by the development of various additional scoring tools, which incorporate data from diverse cohorts to improve risk estimation for atherosclerotic cardiovascular disease (ASCVD). A summary of the most widely used scoring systems is presented in Table 1. Summary of the most common cardiovascular risk prediction tools SBP, systolic blood pressure; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; T2DM, type 2 diabetes mellitus; MI, myocardial infarction; CHD, coronary heart disease; CVD, cardiovascular disease; LDL-C, low-density lipoprotein cholesterol. At the same time, the advent of genetic research has introduced polygenic risk scores (PRSs) as a novel approach to risk prediction. PRSs aggregate the effects of numerous common genetic variants, each contributing a small amount to disease risk, to provide a more comprehensive assessment of an individual’s genetic predisposition. These scores have shown promise in reclassifying risk, particularly in cases where traditional clinical risk factors fall short. For instance, PRSs for CAD have been demonstrated to improve risk prediction and provide insights that may not be captured by traditional risk scores alone.4 Moreover, in a recent study, among patients with suspected stable CAD, polygenic risk was strongly associated with increased prevalence of high-risk plaque and accelerated plaque progression on coronary CTA scans over a median time of 10 years. Imaging also plays a crucial role in cardiovascular risk prediction, with coronary computed tomography angiography (CCTA) being the cornerstone modality. CCTA provides detailed anatomical visualization of coronary arteries, enabling the assessment of coronary artery calcium (CAC) scoring, which quantifies calcified plaque burden and correlates with future cardiovascular events. Higher CAC scores are strongly associated with increased cardiovascular risk, making it a valuable tool for risk stratification.5 Moreover, new technologies using perivascular adipose tissue phenotyping and radiotranscriptomics measure arterial inflammation from routine CCTA and have been shown to capture cardiovascular inflammatory risk, particularly in patients without obstructive CAD,6 and predict in-hospital adverse outcomes.7 Integration with PRSs further enhances the predictive power of CCTA, providing a more comprehensive risk assessment and improving the identification of individuals at higher risk for adverse cardiovascular events.8 Artificial intelligence (AI) and machine learning applications in cardiovascular research have recently gained significant traction. AI models, particularly those employing supervised learning, have shown considerable promise in advancing cardiovascular risk assessment. However, these models also face limitations, including overfitting and challenges related to interpretability. The ‘black box’ nature of some AI models can hinder our understanding of how predictions are made, which is crucial for clinical applications where transparency is essential.9 In a recent study published in JAMA, Neumann et al.10 present a comprehensive evaluation of the incremental value of several clinically available biomarkers—specifically high-sensitivity cardiac troponins, N-terminal pro-B-type natriuretic peptide (NT-proBNP), B-type natriuretic peptide, and high-sensitivity C-reactive protein—when added to established risk factors for long-term prediction of ASCVD events, heart failure, and both cardiovascular and all-cause mortality. By pooling data from 28 cohorts across 4 continents, the authors put together a sample size of over 100 000 individuals to showcase that each biomarker is associated with cardiovascular outcomes and contributes modest, yet statistically significant, improvements in the discriminatory power of existing prediction models. Combining high-sensitivity cardiac troponin I (hs-cTnI), NT-proBNP, and high-sensitivity C-reactive protein improved the C statistic from 0.81 to 0.82 for predicting 10-year incident ASCVD. Notably, NT-proBNP exerted the highest hazard ratios for all outcomes apart from myocardial infarction, solidifying the concept that impaired haemodynamic function is probably the best predictor for cardiovascular mortality and heart failure,11 while inflammation remains in the ‘heart’ of CAD. Indeed, high-sensitivity C-reactive protein has been well documented for more than a decade to be predictive of coronary revascularization or fatal and nonfatal myocardial infarction. This latest confirmation by Neumann et al.10 strengthens the imperative need for modifying inflammation in atherosclerosis by targeting inflammatory pathways such as the NLRP3 inflammasome, inhibiting key cytokines, and promoting pro-resolution mechanisms. Another crucial observation was the sustained improvement in risk prediction over a long-term period, with a median follow-up duration of nearly 12 years. The addition of biomarkers to conventional risk factors consistently enhanced the C statistic across various time horizons, including beyond 10 years. This long-term predictive value underscores the potential role of these biomarkers in primary prevention strategies aimed at addressing cardiovascular risk over extended periods. The findings are supported by post hoc analyses from major clinical trials, such as JUPITER, WOSCOPS, and SPRINT, which have shown that individuals with elevated biomarker levels, particularly hs-cTnI and NT-proBNP, not only have higher cardiovascular risk but also derive greater benefit from preventive treatments like statins and intensive blood pressure management. Moreover, the Neumann et al.10 study revealed that the incremental value of biomarkers was particularly pronounced in older individuals (aged ≥65 years) compared to younger cohorts (aged <65 years). Conventional risk factors tend to have diminished predictive power with increasing age; however, in this study, the overall net reclassification index was higher in older people compared with younger people after the inclusion of hs-cTnI, NT-proBNP, and high-sensitivity C-reactive protein. Despite the impressive scale of the pooled data, the study’s sample size remains relatively small compared to newer mega-cohorts, such as the UK Biobank and the ‘All of Us’ research programme, which include hundreds of thousands of participants. Additionally, the study’s diverse population still underrepresented certain key ancestry groups, limiting the generalizability of the findings to those populations. Furthermore, while the biomarkers showed incremental improvements in prediction, the absolute changes in the C statistic were modest, typically in the second or third decimal place. This modest improvement raises questions about the clinical significance of these improvements and whether they translate into meaningful differences in patient outcomes. The study also did not address the cost-effectiveness of adding these biomarkers to routine clinical practice, an important consideration given the potential for increased healthcare costs associated with more widespread biomarker testing. Future research should focus on validating these findings in larger, more diverse cohorts, exploring the utility of these biomarkers in specific subgroups, and assessing their impact on clinical decision-making and patient outcomes. In conclusion, the study by Neumann et al. marks a significant advancement in the field of cardiovascular risk prediction by demonstrating the incremental value of incorporating serum biomarkers such as hs-cTnI, NT-proBNP, and high-sensitivity C-reactive protein into established risk models. The findings highlight that readily available blood biomarkers, representing distinct pathophysiological pathways, enhance the predictive accuracy of long-term cardiovascular outcomes and maintain their utility across extended follow-up periods. More recently, another prospective cohort study involving 27 939 initially healthy US women followed participants over a 30-year period and found that a single combined measure of high-sensitivity C-reactive protein, LDL cholesterol, and lipoprotein(a) levels provided a robust prediction of long-term cardiovascular risk, highlighting the importance of a comprehensive biomarker-based approach to cardiovascular risk assessment.12 While these types of studies offer promising evidence for the utility of circulating biomarkers in cardiovascular risk prediction, they also open the door for further research. Future studies should aim to explore these findings in even larger and more diverse populations, address the cost-effectiveness of implementing biomarker testing on a broader scale, and assess the impact on clinical outcomes. In doing so, we can continue to refine our approach to cardiovascular risk prediction, moving closer to more precise and individualized healthcare solutions that benefit patients across all demographics. This work was supported by the National Institutes of Health K23 HL159276-01 (B.W.) and AHA 21CDA851511 (B.W.). Not applicable. Biography: Christos P. Kotanidis is a Postdoctoral Researcher at the Brigham and Women’s Hospital, Harvard Medical School, and an Editorial Fellow at the New England Journal of Medicine, both in Boston, MA, USA. His research focuses on vascular inflammation, translational data science, and cardiovascular imaging. He obtained his medical degree in Greece and his DPhil (PhD) from the University of Oxford. He trains clinically in the UK as an NIHR-funded Academic Clinical Fellow in Cardiology. Biography: Dr. Weber is the Director of the Cardio-Rheumatology Clinic, an Assistant Professor of Medicine at Harvard Medicine School, and an Associate Physician in Prevention Cardiology and Cardiovascular Imaging at Brigham and Women’s Hospital. Her clinical interests include cardiovascular risk stratification and prevention with patients with systemic inflammatory disorders, including integration of multi-modality imaging, specific anti-inflammatory medications and cardiovascular effects, pericardial disease, and novel anti-inflammatory treatments. Her research is broadly focused on translational research in systemic inflammatory conditions harnessing advanced cardiovascular imaging and immunology.