Authors: Pooja V. Selvam, Rahul Sharma, Peter Ganz, Roger S. Blumenthal, Martha Gulati
Categories: State-of-the-Art Review, Cardiovascular disease, Risk assessment, Preventive cardiology
Source: American Journal of Preventive Cardiology
Authors: Pooja V. Selvam, Rahul Sharma, Peter Ganz, Roger S. Blumenthal, Martha Gulati
Cardiovascular disease (CVD) is an umbrella term that includes coronary artery disease (CAD), heart failure (HF), and stroke. CVD may begin as early as the second decade of life and is driven in large part by traditional risk factors (e.g. dyslipidemia, hypertension, diabetes mellitus, smoking). Early identification of individuals at high risk for CVD is necessary for the timely initiation of pharmacotherapy and reduction in the risk of future major adverse cardiovascular events (MACE).
Both national and international primary prevention guidelines recommend the use of prediction models to estimate future disease risk. Over the past two decades, several risk scores have been developed including Framingham, Pooled Cohort Equation (PCE), SCORE, SCORE2, QRISK3, and more recently the Predicting Risk of Cardiovascular Disease Events (PREVENT) equation [[1], [2], [3]]. There has been a recent paradigm shift with the selective use of imaging with a coronary artery calcium (CAC) score, to refine ASCVD risk estimation and the initiation of preventive pharmacotherapies [4]. The widespread availability of multiple risk stratification tools whose results may differ significantly provides a challenge for clinicians to select the most appropriate tool for their patient population.
Current US guidelines recommend the use of the updated 2019 ACC/AHA ASCVD 10-yearr risk calculator (based on the PCE) as an initial prediction tool for 10-year risk of ASCVD events [5]. In 2023, the AHA introduced the PREVENT equation as an alternative to the PCE [6]. and it is anticipated that ACC/AHA guideline updates will endorse use of the PREVENT risk estimator. In this succinct review, we evaluate the PREVENT equation as a risk stratification tool in comparison to its widely used predecessor. Additionally, we discuss the consideration of coronary artery calcium (CAC) scoring for more personalized risk evaluation.
Although the 2019 ACC/AHA Prevention of CVD Guideline endorses the PCE to guide primary prevention practices, the PCE does not adequately capture recent changes in risk factors (less smoking, increased weight and greater development of chronic kidney disease), and may not be generalizable across certain populations.
Multiple key studies support the overestimation of risk by the PCE. Yadlowski et al. found that PCE consistently predicted a higher than observed ASCVD risk, especially among Black individuals. The authors developed revised PCEs with contemporary data and found that nearly 12 million adults who were previously labeled as high risk (10-year risk >7.5 %) would be relabeled to a lower risk category [7]. This discrepancy raised concerns about the appropriateness of treating with statins based on the PCE and the potential harm from unnecessary treatment if the actual risk was lower than the PCE’s prediction. These findings are consistent with a large meta-analysis of 86 prospective studies, where the PCE overestimated ASCVD risk on average by 41 % [8].
Current data indicates that the PCE often underpredicts CVD risk in individuals from lower socioeconomic status and in people with chronic inflammatory diseases, including HIV, sarcoid, and rheumatoid arthritis [[9], [10], [11], [12]]. Similarly, PCE may not accurately predict risk in ancestral groups not included in the derivation given that it only incorporated White and Black persons. For individuals from other racial groups, the 2019 Guideline suggests using risk estimates for White individuals of the same sex. However, several studies suggest that PCE may overestimate risk in East Asian and Hispanic populations, while underestimating risk in South Asian populations [[13], [14], [15]]. The omission of risk conferred by race can lead to overtreatment or undertreatment of individuals, potentially widening racial disparities in cardiovascular risk and outcomes.
To address these gaps, the PREVENT model was developed and validated using a significantly larger sample of > 6 million individuals aged 30 to 79, of diverse backgrounds, without known ASCVD (stroke, coronary heart disease) or heart failure at baseline [6]. This novel risk prediction equation is sex-specific, race-free, and assesses both short-term (10-year) and long-term (30-year) risk of CVD and its subtypes (MI, stroke, heart failure). The key risk factors used in this equation HDL-C, non-HDL-C, systolic blood pressure, diabetes mellitus, current smoking, body mass index (BMI), anti-hypertensive medication, lipid-lowering medication, and estimated glomerular filtration rate (eGFR). It provides the option of including cardiovascular kidney metabolic (CKM) syndrome modifiers, including hemoglobin A1C levels and urine albumin-creatinine ratios. This model also provides the option to include a unique feature - social deprivation index (SDI), which is assessed by the person’s zip code; it provides a basic estimation of cardiovascular disease risk attributable to social determinants of health [6].
The PREVENT score provides notable advancements, with improved generalizability across multiple fronts. Primarily, in comparison to prior risk prediction models, the PREVENT score was derived using a vastly larger sample size, including over 6 million individuals (approximately half for derivation and half for validation) from 46 different data sets in comparison to PCE which involved only 48,733 participants (24,626 for derivation and 24,107 for validation) [6]. The PREVENT model demonstrated excellent calibration in the overall population as well as among demographic groups (sex, race). The median slope of predicted versus observed risk for all CVD was 0.94 in men and 1.03 in women using the PREVENT score (consistent with excellent calibration) and the ASCVD risk calibration was 1.04 in men and 1.09 in women. Calibration improved significantly in individuals with albuminuria >300mg/g when urine albumin-to-creatinine ratio was added to the base model [6].
Additionally, the PREVENT model removed race from risk prediction, given that race is thought to be a predominantly social construct and does not necessarily reflect genetic differences. By decreasing the bias of race, this new model is a step in the right direction for promoting health equity across different populations. Given the increasing prevalence of cardiovascular disease among the younger population (age < 40), the PREVENT equation includes individuals age 30 to 39, enabling earlier risk identification and a greater opportunity for providing intensive advice about healthy lifestyle and possibly earlier use of pharmacotherapy. This model also includes eGFR as a risk modifier, which allows the personalization of risk among high-risk subgroups with CKM.
A major improvement in this model is the inclusion of heart failure (HF) as a primary outcome of interest. In patients who develop HF, over 60 % of patients have a history of ASCVD, which signifies a strong overlap between the factors that promote both disease processes [16]. Currently, the Pooled Cohort equation to Prevent HF (PCP-HF) model is rarely used to estimate race and sex-specific 10-year risk of HF using predictors including blood pressure, BMI, cholesterol, smoking status, glucose levels, and QRS duration [17]. The PREVENT algorithm effectively combined the risk estimation for incident ASCVD (via PCE) and incident heart failure (via PCP-HF), as both equations involve similar cardiovascular risk factors as their variables.
The prognostic performance of the PREVENT score can be evaluated using two critical metrics – calibration and discrimination. Above, we note that this model demonstrated excellent calibration. The slope of the observed to predicted risk for the PREVENT model was 1.09 in women and 1.04 in men at baseline. When urine albumin-to-creatinine ratio was added to the base model, there was a significant improvement in calibration from 1.05 to 1.39. In comparison, calibration of the PCE lies between 0.5 and 0.54, conferring an overestimation of risk by up to 50 % [6].
Over a follow up of five years, the PREVENT equation base model had good discrimination for total CVD in women (C-statistic 0.794) and men (C-statistic 0.757). However, when the three optional variables (urine albumin-to-creatinine ratio, hemoglobin A1c, and social deprivation index) were added to the base model for total CVD, there was only a modest improvement in discrimination (change in C-statistic 0.004 in females and 0.005 in males). When compared to the PCE (C-statistic 0.880), the PREVENT model (C-statistic of 0.890) showed only a slight improvement even with the addition of CKM-specific factors [18].
Additionally, while the PREVENT model is celebrated for its emphasis on incident heart failure, there is room for improvement in this realm. Overall, this model accurately captures risk predictors that are independently associated with heart failure and ASCVD. However, there is room for further research to identify the risk threshold for the initiation of heart failure prevention therapies.
In addition to healthy lifestyle modifications, statins remain the cornerstone of lipid-lowering therapy. The 2018 ACC/AHA/Multisociety guideline on the Management of Blood Cholesterol provides evidence-based recommendations to guide statin prescription for individuals based on 10-year risk using the PCE [19]. The current classification system stratifies risk into low (<5 %), borderline (5 to 7.5 %), intermediate (7.5 % to 19.9 %), and high risk (≥20 %) categories. Important risk enhancers, detailed in the following section, are also factored into the decision to promote the initiation and/or intensification of statin therapy. Recent studies shed light on how the PREVENT score would reclassify risk across these categories and thus affect treatment practices for statins and antihypertensive therapy, given that compared to PCE, improved calibration with PREVENT reclassifies many patients to lower risk. Improved calibration with the PREVENT model could mean more accurate risk assessment. However, its application in current practice requires further research to determine the new thresholds for treatment initiation based on the updated PREVENT data.
A key study by Diao used nationally representative data from the National Health and Nutrition Examination Surveys (NHANES) to estimate the number of adults who would undergo a change in risk categorization based on the PREVENT equation. In a cohort of 7765 adults between age 30 to 79 years, the PREVENT equation would reclassify 53 % of adults into lower risk categories and only 0.41 % would be reclassified into higher risk categories. If the statin benefit threshold of 7.5 % 10-year risk were to be carried forward unchanged into future ACC/AHA Guideline Updates, this would reduce the number of people eligible for receiving statins by 14 million and the number of people eligible for anti-hypertensive medications by 2.6 million. If the current risk threshold for statin initiation is maintained, over a 10-year span, the profound decrease in treatment numbers has been estimated to result in 107,000 potential episodes of myocardial infarction or stroke [20].
Shetty et al. performed a study involving 4342 participants and compared 10-year ASCVD risk and mortality outcomes estimated by PCE and PREVENT, in addition to the inclusion of cardiac biomarkers (NT pro-BNP and cardiac troponin) on mortality estimates. Compared to PCE, the PREVENT equation estimated a lower 10-year ASCVD risk for 81 % of individuals in this cohort. Interestingly, the sex-stratified analysis revealed a higher proportion of males compared to females that were re-classified with lower risk estimates using the PREVENT equation. The inclusion of biomarkers did not affect risk prediction significantly. This study found that using the ASCVD risk threshold >5 % for treatment with statins, replacement of the PCE with the PREVENT tool would reduce statin eligibility in 11 million individuals if future US cholesterol guidelines do not adjust the level of net statin benefit from 7.5 % 10-year risk to a lower value [21]. A similar study by the same group compared the 10-year ASCVD risk using the PCE v PREVENT equations in individuals from the Systolic Blood Pressure Intervention (SPRINT) trial. As seen in the previous study, 70 % of participants were reclassified to a lower risk category and sex-stratification revealed a relative underestimation of 10-yr risk in males compared to females when using the PREVENT equation compared to the PCE [22].
All three of these studies demonstrate a reduction of treatment eligibility (statins, antihypertensive medications) with the use of the PREVENT model. This data is reflected in the recently updated AHA/ACC 2025 Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults, which recommends lowering the 10-year cardiovascular disease risk threshold from 10 % to 7.5 % using the PREVENT model to determine treatment eligibility for individuals with blood pressures ranging from 130–139/80–89 mm Hg. The reclassification of risk to lower values without changes in current thresholds for lipid-lowering treatment may lead to the reversal of overall improvement in total cholesterol and LDL-C levels that was largely due to the widespread use of current guideline-recommended statin therapy.
The PREVENT model has shown great promise with its improved discrimination and calibration to accurately predict future CVD. The studies above highlight drastic changes in current preventive treatment guidelines and provide impetus for further research to determine appropriate treatment thresholds that can inform future guidelines. Additionally, further studies evaluating the discrimination and calibration of the PREVENT score in under-represented populations, such as Asians and Hispanics, can help enhance health equity and reduce racial disparities. Furthermore, guidance on how to incorporate the current risk enhancers that are currently considered with the PCE model can help guide future management with the PREVENT model.
Risk calculators provide a simple, statistical approach for risk estimates based on the results of population studies. However, at the individual level, the decision to treat goes beyond population studies, and is influenced by additional factors not included in the risk models, such as family history, lifestyle, and the presence of CVD risk enhancers. In addition, a CAC score provides a unique individualized, phenotypic assessment of cardiovascular risk that can guide treatment decisions in asymptomatic males > 40 years of age and females > 45 years.
The AHA currently recommends quantitative risk assessment with the PCE as an important first step in assessing a person’s ASCVD risk. The next step involves the consideration of risk enhancing factors to facilitate clinician-patient decision-making in adults with borderline (5 % to 7.5 % 10-year ASCVD risk) or intermediate (7.5 % to 20 %) risk. These risk enhancers family history of premature ASCVD, chronic kidney disease, persistent LDL-C ≥ 160 mg/dL or triglycerides ≥175 mg/dL, elevated lipoprotein (a), high-risk racial/ethnic groups, factors unique to women (pre-eclampsia, premature menopause), and elevated inflammatory conditions among others. If there remains uncertainty in the use of preventive interventions, a CAC score serves as the best arbiter to up-risking select individuals and therefore justifying the initiation or intensification of statins [4]. Data suggests that the use of CAC testing after initial quantitative risk assessment noticeably improves discrimination and reclassification of ASCVD risk [23].
The ACC 2022 Expert Consensus Decision Pathway Guidelines provide a graded framework for the initiation of LDL-C lowering therapies based on CAC score. For individuals with CAC score 1–99, treatment with moderate to high-intensity statins are advised. For CAC >100 and CAC >1000, the appropriate addition of alternative lipid-lowering therapies, such as ezetimibe and PCSK-9 inhibitors is recommended [24].
The PREVENT equations provide updated risk models in the assessment of cardiovascular disease risk. Studies comparing the PCE and PREVENT equations have consistently shown lower ASCVD 10-year risk estimates using the PREVENT equation, in large part due to improved calibration of absolute risk levels. This has repercussions in the realm of statin and anti-hypertensive medication prescription, potentially leading to undertreatment and increase in MACE, unless the risk thresholds for antihypertensive and statin therapy are lowered. We strongly favor the selective use of non-invasive imaging using CAC scores in conjunction with risk prediction equations to help guide the primary prevention of ASCVD. Since the PREVENT equation tends to estimate a risk that is 30 to 50 % lower than that of the PCE, further guidance is required to help determine a suitable threshold for pharmacotherapy of hypertension and hyperlipidemia.Central Legend.Unlabelled image
M.G. is supported by contracts from the National Heart, Lung, and Blood Institutes nos. N01- HV-068,161, N01-HV-068,162, N01-HV-068,163, N01-HV-068,164, grants U01 HL064829, U01 HL649141, U01 HL649241, K23 HL105787,K23 HL125941, K23 HL127262, K23HL151867, T32 HL069751, R01 HL090957, R03 AG032631, R01 HL146158, R01 HL146158–04S1, R01 HL124649, R01 HL153500, U54 AG065141, General Clinical Research Center grant MO1- RR00425 from the National Center for Research Resources, the National Center for Advancing Translational Sciences Grant UL1TR000124, Department of Defense grant PR161603 (CDMRP- DoD), and grants from the Gustavus and Louis Pfeiffer Research Foundation, Danville, NJ, The Women’s Guild of Cedars-Sinai Medical Center, Los Angeles, CA, The Ladies Hospital Aid Society of Western Pennsylvania, Pittsburgh, PA, and QMED, Inc., Laurence Harbor, NJ, the Edythe L. Broad and the Constance Austin Women’s Heart Research Fellowships, Cedars-Sinai Medical Center, Los Angeles, CA, the Barbra Streisand Women’s Cardiovascular Research and Education Program, Cedars-Sinai Medical Center, Los Angeles, CA, The Society for Women’s Health Research, Washington, D.C., the Linda Joy Pollin Women’s Heart Health Program, the Erika Glazer Women’s Heart Health Project, the Adelson Family Foundation, Cedars-Sinai Medical Center, Los Angeles, CA, Robert NA. Winn Diversity in Clinical Trials Career Development Award (Winn CDA), and the Anita Dann Friedman Endowment in Women’s Cardiovascular Medicine & Research. This work is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute, the National Institutes of Health, or the U.S. Department of Health and Human Services. Fig. 1Fig. 1Comparison between variables and outcomes of PCE v PREVENT score.Fig 1
None.
All authors have seen and approved the final version of the article, Cardiovascular Disease Risk Estimates Using the New PREVENT Equation: The Good, Bad, and the Ugly. The article is the authors’ original work and hasn’t received prior publication. The article is not under consideration for publication elsewhere.
Pooja V. Selvam: Writing – original draft, Visualization, Data curation, Conceptualization. Rahul Sharma: Writing – original draft, Data curation, Conceptualization. Peter Ganz: Writing – review & editing, Formal analysis, Data curation. Roger S. Blumenthal: Writing – review & editing, Formal analysis, Data curation. Martha Gulati: Writing – review & editing, Visualization, Project administration, Formal analysis, Data curation, Conceptualization.
The authors declare the following financial interests/personal relationships which may be considered as potential competing