Why Your Standard Heart Risk Score Misses Nearly Half of First Heart Attacks

Two days before their heart attack, 45% of patients would have been told they were low risk.

That finding comes from a November 2025 study published in the Journal of the American College of Cardiology: Advances, led by researchers at Mount Sinai. The team examined 474 patients under age 66 who suffered their first heart attack between 2020 and 2025. None had known coronary artery disease beforehand. The researchers calculated what each patient's risk score would have been 48 hours before their cardiac event.

Using the widely deployed ASCVD Risk Estimator Plus, 45% of these patients would have been classified as low or borderline risk, categories that do not qualify for preventive treatment under current guidelines. When the researchers applied the newer PREVENT calculator, which includes additional variables intended to improve accuracy, the miss rate climbed to 61%.

Dr. Amir Ahmadi, Clinical Associate Professor of Medicine at the Icahn School of Medicine at Mount Sinai, stated the implication plainly: "If we had seen these patients just two days before their heart attack, nearly half would NOT have been recommended for further testing or preventive therapy."

Read that again. The tools most cardiologists use to decide whether you need help would have sent nearly half of these patients home with a clean bill of health. Two days before they had a heart attack.

The Problem Is Not the Math. It Is the Premise.

The instinct when confronting a 45% miss rate is to demand better calculators. Build more sophisticated algorithms. Add more variables.

This instinct misses the fundamental problem.

Risk calculators are statistical tools designed for populations. They estimate the probability that someone with your characteristics will experience an event within a defined time horizon, typically ten years. When applied to millions of people, these probabilities hold up reasonably well in aggregate.

But you are not a population. You are a single individual who will either have a heart attack or not have one. The 5% ten-year risk that qualifies as "low" provides cold comfort if you happen to be in that 5%.

Dr. Khurana articulates this problem directly: "The person in front of you is unique. He's not a statistic. While the guidelines are based on population statistics, that is an intellectual fallacy that not strictly thinking doctors don't realize."

The Mount Sinai study confirms what practitioners who think carefully about individual risk have long suspected: population-derived probabilities are the wrong tool for individual prevention decisions.

Symptoms Arrive Too Late

The Mount Sinai data revealed another troubling pattern. Approximately 60% of patients only noticed warning signs such as chest pain or shortness of breath less than 48 hours before their heart attack.

This finding demolishes the reassurance often given to patients with low risk scores: "Come back if you develop symptoms."

The atherosclerotic plaque that causes heart attacks develops silently over years or decades. It narrows arteries gradually, often without producing symptoms until a rupture event triggers acute clot formation. The first symptom may be the heart attack itself.

Waiting for symptoms is not a prevention strategy. It is a treatment strategy that has already failed at prevention.

Dr. Anna Mueller, first author of the study, summarized the clinical reality: "When we look at heart attacks and trace them backwards, most heart attacks occur in patients in the low or intermediate risk groups."

Look for the Disease, Not the Probability

If risk scores cannot reliably identify who will have a heart attack, and symptoms arrive too late for prevention, what remains?

Direct detection of the disease itself.

Atherosclerosis leaves physical evidence in the arteries long before it causes clinical events. Coronary artery calcium scoring uses low-dose CT imaging to detect calcified plaque deposits. This is not a probability estimate. It is direct observation of whether disease is present and, if so, how extensive it has become.

A patient with a zero calcium score has minimal subclinical atherosclerosis regardless of what any risk calculator predicts. A patient with a high calcium score has significant disease regardless of favorable traditional risk factors.

Dr. Ahmadi's conclusion aligns with this approach: "It may be time to fundamentally reconsider this model and move toward atherosclerosis imaging to identify the silent plaque before it has a chance to rupture."

At Healthy Heart Cardiology, this is already the standard of care. We do not rely on ten-year risk estimates to decide who deserves attention. We look directly at the arteries.

Dr. Khurana's framing is blunt: "I do not do guesswork on your ten-year heart attack risk.

We should be preventing the disease called atherosclerosis. If no disease, how can you have a heart attack? We should be treating the disease, not guessing and playing Russian roulette."

What the Risk Calculators Cannot See

Standard risk assessment tools incorporate age, sex, race, blood pressure, total and HDL cholesterol, diabetes status, and smoking history. These variables emerged from large epidemiological studies as statistically significant predictors at the population level.

But several potent risk factors are invisible to these calculators.

Lipoprotein(a) affects 20 to 30 percent of the population and dramatically amplifies cardiovascular risk through inflammatory, atherogenic, and prothrombotic mechanisms. It is genetically determined and largely unaffected by lifestyle or statin therapy. Standard risk calculators do not include it. Standard lipid panels do not measure it. A patient with dangerously elevated Lp(a) can receive a reassuring low-risk classification because the assessment tool is blind to one of their most significant risk factors.

Apolipoprotein B provides a more accurate measure of atherogenic particle burden than LDL cholesterol. Two patients with identical LDL values can have markedly different ApoB concentrations, and it is the particle count that drives plaque formation. Standard screening does not include this measurement.

Insulin resistance can be present and metabolically significant long before A1C reaches the threshold for prediabetes diagnosis. Patients with normal A1C can already have the metabolic dysfunction that accelerates atherosclerosis. Standard risk assessment relies on diabetes diagnosis, missing years of subclinical damage.

Family history receives only crude treatment in standard calculators. The algorithms do not distinguish between a father who had a heart attack at 48 and a grandfather who had one at 82. They cannot incorporate the specific patterns of lipid abnormalities and metabolic dysfunction that run through families.

At Healthy Heart Cardiology, we check all of these because the information changes clinical decisions. A patient with a "low risk" ASCVD score but elevated Lp(a), high ApoB, and a family history of premature coronary disease is not actually low risk. The calculator simply cannot see what we can measure directly.

A Different Model

The distinction comes down to philosophy. Standard practice asks: "What is your probability of having a heart attack in the next ten years?" We ask: "Do you have the disease that causes heart attacks?"

The first question generates a number. The second generates an answer.

50 to 70% of people will have some atherosclerotic plaque before they die. It is the most common disease in the modern world and the number one killer. We are not appropriately addressing it because we start treatment too little, too late, and not enough.

No one should arrive at a hospital with a heart attack that was preventable. If you come to us five to ten years before you were destined to have one, we can change that trajectory. But that requires looking for the disease, not waiting for a risk calculator to give you permission to worry.

Don't rely on a "low risk" score that might be incomplete. Schedule a personalized risk assessment

Sources

• Mueller A, Ahmadi A, et al. Prevalence and Distribution of ASCVD Risk Scores Before First Heart Attack. JACC: Advances. November 2025.

• Medscape: Current Risk Scores Miss Nearly Half of First Heart Attacks

• Healthy Heart Cardiology

Author Disclosures and Professional Affiliations

Sandeep Khurana, MD, FACC

The views expressed in this blog are those of the author, Sandeep Khurana, MD, FACC, and reflect the clinical approach utilized at Healthy Heart Cardiology, PLC. Dr. Khurana maintains an active medical license in Michigan and is an Attending Physician at Corewell Health in Grand Rapids. He is board certified in Clinical Lipidology, Cardiovascular Disease, and Interventional Cardiology.

As the practice of cardiovascular medicine relies heavily on translational research, Dr. Khurana serves as a Principal Investigator (PI) for numerous ongoing and past clinical trials. These research activities ensure that our practice remains at the forefront of medical advancements discussed in this blog post, which focuses on personalized prevention.

Clinical Research Disclosures:

The content of this blog discusses advancements in treating risk factors such as elevated Lp(a), inflammation (hs-CRP), and high LDL cholesterol. Dr. Khurana has served as a Principal Investigator for trials studying therapeutic agents related to these targets, including, but not limited to:

• Lp(a) Inhibition: Serving as PI for the OCEAN(a)-Outcomes and OCEAN(a)-PREVENT studies, which assess Olpasiran (a siRNA inhibiting Lp(a) production).

• Inflammation (hs-CRP): Serving as PI for the ZEUS study, which investigates Ziltevekemab (a MoAb that binds IL-6 Ligand) in ASCVD subjects with elevated hs-CRP.

• LDL Cholesterol Reduction (PCSK9 and others): Serving as PI for trials involving various lipid-lowering agents, including Inclisiran (VICTORION and ORION trials), Evolocumab (FOURIER and VESALIUS trials), Oral PCSK9 antagonists (CORAL REEF OUTCOMES), and Bempedoic Acid (CLEAR-CV Outcomes trial).

• Heart Failure Prevention: Serving as PI for the EMPACT-AMI trial, studying Empagliflozin (an SGLT2 inhibitor) in post-MI patients at risk of developing CHF.

These research activities are conducted across the practice locations in Grand Rapids and Rochester Hills, Michigan. This involvement is crucial for advancing the field of preventive cardiology and bringing the newest data into patient care.

The practice of medicine and the field of cardiovascular risk assessment are continually evolving based on new research, such as the UK Biobank data discussed (published in 2025). Readers are encouraged to consult with their healthcare provider for medical advice personalized to their specific situation.