The Blind Spot in Secondary Prevention Now Has a Score

Key Takeaway: A newly developed risk calculator, SMART2-HF, can predict the 10-year and lifetime probability of developing heart failure in individuals with atherosclerotic cardiovascular disease. Validated in over 240,000 patients across six external datasets, the model uses routine clinical information and fills a critical gap in secondary prevention, where existing tools focus almost exclusively on recurrent heart attacks and strokes while overlooking heart failure.

A Missing Piece in Cardiovascular Risk

For the millions of patients already living with heart or vascular disease, a familiar ritual unfolds at every clinical visit: blood pressure is measured, cholesterol levels are reviewed, medications are adjusted, and risk scores are calculated to estimate the likelihood of the next heart attack or stroke. Yet, there is another outcome, at least as devastating, that these scores almost never address: heart failure. This is a remarkable blind spot. Heart failure affects over 64 million people globally, and patients with established atherosclerotic cardiovascular disease (ASCVD) are among the highest-risk group for developing the condition^[2]. Until now, however, clinicians have lacked a validated, practical tool to quantify this specific danger. A new prediction model, named SMART2-HF, is designed to close this very gap.

What Researchers Built and How They Tested It

The SMART2-HF model was developed using data from 7,698 patients with established ASCVD—meaning they had previously experienced events like a heart attack, stroke, or peripheral artery disease, or had documented blockages in their arteries. Over a median follow-up of 11.2 years, 13% of patients in the development cohort were newly diagnosed with heart failure^[1]. The model was engineered to predict both the 10-year and lifetime risk of this outcome using clinical variables routinely found in any standard medical record—no exotic biomarkers or advanced imaging required.

The model’s mathematical backbone was built upon the 13 standard clinical predictors included in the SMART2 algorithm: age, sex, smoking, diabetes, systolic blood pressure, non-HDL cholesterol, coronary artery disease, cerebrovascular disease, peripheral artery disease, abdominal aortic aneurysm, eGFR, CRP, and time since first ASCVD diagnosis. To this core set of parameters, researchers added just two critical variables that play a decisive role in heart failure pathophysiology: body mass index (BMI) and a history of atrial fibrillation. This 15-variable architecture aims to generate maximum clinical data without creating unnecessary testing costs, while incorporating non-HF (non-heart failure) deaths due to aging as a “competing risk”. This prevents the artificial overestimation of heart failure risk and potential treatment gains, especially in older patient groups.

What makes this study particularly robust is the scale of its external validation. The researchers tested the SMART2-HF model against six independent data sources encompassing 240,741 patients with ASCVD. The pooled C-statistic, a measure of how well a model distinguishes between patients who will and will not develop an outcome, was 0.696. To put this in context, a C-statistic of 0.5 means a model is no better than a coin flip, while 1.0 represents perfect prediction. A value near 0.70 is considered clinically useful and is comparable to many widely adopted cardiovascular risk calculators, including the original Framingham Risk Score^[3]. More importantly, the model performed consistently across demographic and clinical subgroups, suggesting it is broadly applicable, not just in narrow populations.

The discriminatory power of SMART2-HF was not limited to the general population; when compared with the current American Heart Association PREVENT-HF equations developed for the general population, the model exhibited significantly superior discrimination in this high-risk cohort with established vascular disease. Furthermore, the model was recalibrated using baseline risk multipliers for low- (CPRD/UK, HUNT3/Norway), medium- (SWEDEHEART/Sweden), and high-risk (ASCVD-Particles/Poland, Estonian Biobank) regions of the continent, in full compliance with European Society of Cardiology Cardiovascular Risk Collaboration (ESC CRC) guidelines. Decision curve analysis provided concrete proof that, at clinical treatment thresholds between 15% and 30%, decisions guided by SMART2-HF yield a net clinical benefit compared to “treat all” or “treat none” approaches.

Why Are ASCVD Patients Vulnerable to Heart Failure?

Understanding why people with existing cardiovascular disease are prone to heart failure requires a look beneath the surface of the arteries. Atherosclerosis, the progressive buildup of lipid-rich plaques in arterial walls, does not just pose a threat of acute blockage. It precipitates a chronic inflammatory state that reverberates throughout the cardiovascular system^[4]. Inflammatory cytokines like interleukin-6 and tumor necrosis factor-alpha promote fibrosis within the heart muscle tissue, gradually stiffening the ventricles and impairing their ability to relax and fill with blood. This is the pathological basis of heart failure with preserved ejection fraction (HFpEF), which now accounts for about half of all heart failure diagnoses^[5].

Meanwhile, patients who have had a heart attack may be left with areas of myocardium replaced by non-contractile scar tissue. Over months and years, the remaining healthy muscle tries to compensate by working harder—a process called adverse remodeling. The heart chamber dilates, wall stress increases, and neurohormonal systems, particularly the renin-angiotensin-aldosterone system, become chronically activated, accelerating the slide toward heart failure with reduced ejection fraction (HFrEF)^[6]. When you add the common comorbidities in ASCVD patients—hypertension, diabetes, chronic kidney disease, and obesity—the cumulative hemodynamic and metabolic burden on the heart becomes immense.

The SMART2-HF model captures these intersecting risk trajectories by using clinical data points that essentially serve as proxies for the underlying biology. By appropriately weighting these variables, it identifies which patients are accelerating toward heart failure and which, while high-risk, remain more stable.

The most striking clinical reality revealed by the data is “discordance” in risk. A full 25% (1 in 4) of male and female patients in the lowest quartile for 10-year risk of recurrent vascular events (heart attack and stroke), as calculated by the classic SMART2 score, landed in the high-risk category for heart failure when assessed with SMART2-HF. In other words, a significant portion of atherosclerosis patients who appear “safe and stable” in the clinician’s eyes because their cholesterol and blood pressure are near target are actually being driven toward the precipice of heart failure by underlying factors like progressive obesity, atrial fibrillation, silent diastolic dysfunction, or microvascular endothelial damage.

What This Means for Patients Tomorrow

The practical significance of SMART2-HF lies in its potential to reshape how clinicians approach secondary prevention. Current guidelines for ASCVD patients are largely focused on preventing recurrent ischemic events, for which statins, antiplatelet agents, and blood pressure medications are prescribed. The companion SMART2 model already helps quantify this recurrent risk. SMART2-HF now adds a parallel dimension: the risk of developing heart failure.

This matters because the preventive strategies for heart failure, while overlapping, are not identical to those for recurrent atherosclerotic events. SGLT2 inhibitors, originally developed for diabetes, have shown dramatic reductions in heart failure hospitalizations even in patients without diabetes^[7]. Aggressive blood pressure control, weight management, and treatment of atrial fibrillation take on special urgency when heart failure, not another heart attack, is the primary threat. A patient identified by SMART2-HF as being on a high-risk trajectory can be prioritized for these interventions, transforming secondary prevention from a one-size-fits-all protocol into a truly personalized strategy.

The most revolutionary dimension offered by this integration is the ability to simulate the “HF-free life expectancy” gained from treatment for each individual. For an individual between 40 and 80 years old, the model calculates not just 10-year probabilities but the lifetime event chain up to age 90. For example, in a 60-year-old with coronary artery disease, atrial fibrillation, and mild obesity, SMART2-HF reveals how many additional years the patient could live free of heart failure by adding cardiometabolic therapies like SGLT2 inhibitors or GLP-1 receptor agonists, or by achieving intensive systolic blood pressure control. In consultations where the physician and patient are looking at the same screen, these concrete years become the most powerful persuasive tool, breaking down barriers of abstract percentages and medication non-adherence.

Noteworthy Limitations

No single study rewrites clinical practice overnight. A C-statistic of 0.696, while clinically useful, leaves meaningful room for misclassification: some patients flagged as high-risk will never develop heart failure, while some at lower risk will. The model was developed and validated primarily in cohorts from high-income countries; its performance in more diverse global populations remains to be confirmed. Furthermore, prediction is not prevention; showing that acting on SMART2-HF scores actually improves patient outcomes will require prospective interventional trials. Finally, heart failure is a heterogeneous syndrome, and whether the model predicts HFpEF and HFrEF with equal accuracy is an important question for future research.

Another methodological limitation is that the primary endpoint was defined by acute events requiring hospitalization and HF-related deaths. Early-stage heart failure cases diagnosed with mild-to-moderate symptoms in an outpatient setting, which do not necessitate hospitalization, may not be fully captured by this radar, leading to a slight underestimation of absolute risk rates. Additionally, to maintain clinical accessibility and practicality, the scoring system excluded advanced biomarkers like natriuretic peptides (BNP/NT-proBNP) or echocardiographic strain parameters. Therefore, SMART2-HF is not a standalone, definitive clinical diagnostic tool, but rather a powerful secondary triage guide indicating who should be urgently referred for further cardiac imaging and aggressive organ-protective therapies.

Still, in a landscape where secondary prevention has long been one-sided, laser-focused on the next heart attack while heart failure progresses silently, the SMART2-HF model represents a meaningful course correction. For the clinician sitting across from a patient with established cardiovascular disease, it offers something that has been glaringly absent: a number to ground the conversation about heart failure risk, and a reason to act on it.

Scientific Sources

1. Reitsma TH, et al. Prediction of incident heart failure in established atherosclerotic cardiovascular disease: the SMART2-HF model. European heart journal. 2026;47(34):4743-4757. PubMed: https://pubmed.ncbi.nlm.nih.gov/41810961/
2. Savarese G, et al. Global burden of heart failure: a comprehensive and updated review of epidemiology. Cardiovasc Res. 2023. DOI: 10.1093/cvr/cvac013
3. D’Agostino RB Sr, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008. DOI: 10.1161/CIRCULATIONAHA.107.699579
4. Libby P, et al. Inflammation and atherosclerosis. Circulation. 2002. DOI: 10.1161/hc0902.104353
5. Dunlay SM, et al. Epidemiology of heart failure with preserved ejection fraction. Nat Rev Cardiol. 2017. DOI: 10.1038/nrcardio.2017.65
6. Cohn JN, et al. Cardiac remodeling — concepts and clinical implications: a consensus paper from an international forum on cardiac remodeling. J Am Coll Cardiol. 2000. DOI: 10.1016/s0735-1097(99)00630-0
7. McMurray JJV, et al. Dapagliflozin in patients with heart failure and reduced ejection fraction. N Engl J Med. 2019. DOI: 10.1056/NEJMoa1911303