10-Years ASCVD Risk Calculator

The endpoint is narrower than heart disease

The 2013 ACC/AHA risk guideline built its pooled cohort equations around one specific list: a nonfatal heart attack, death from coronary heart disease, a fatal stroke, and a nonfatal stroke. Four outcomes. The guideline calls this hard ASCVD, and the word hard is doing real work. Angina that never becomes an attack, a transient ischemic attack that resolves, a stent, a bypass, heart failure, and narrowed leg arteries all stay outside the count.

That choice changes who scores low. A 62-year-old with a weakening heart and normal cholesterol can score 3% here even while heart failure is settling in, because the endpoint cannot see it. The Framingham general score on our CVD page covers that wider family; this equation does not. If a doctor quotes a number from the official ASCVD Risk Estimator Plus, it is this narrower count being described.

One more boundary matters: the equation predicts a first event only. Anyone who has already had a heart attack or stroke is in secondary prevention, where treatment is assumed and a risk percentage is beside the point.

Seven inputs and four equations

Every version of the equation reads the same seven inputs: age, total cholesterol, HDL cholesterol, systolic pressure, whether pressure medicine is taken, whether you smoke now, and whether you have diabetes. But the coefficients differ by sex and by race group, and the guideline publishes all four sets. Blood pressure switches coefficient depending on treatment, and three of the four versions carry interaction terms where age multiplies another input. Below is the full published table.

The arithmetic is always the same shape: add up coefficient times natural log for each input, subtract the group mean, raise e to that number, then use it as an exponent on the baseline survival. One minus the result is the ten-year risk. A dash means the term does not exist in that version of the equation.

Worked example: 55-year-old Black man, 7.8%

Take a 55-year-old Black man with total cholesterol 205 mg/dL, HDL 47 mg/dL, untreated systolic pressure 138 mm Hg, no smoking, no diabetes. The Black male equation has no interaction terms, so it fits in four lines.

The sum of 19.2331 sits below the group mean of 19.54, which is expected for a man with near-average numbers. Subtracting gives -0.3069, and e to that power is 0.7357. The baseline survival of 0.8954 raised to the 0.7357 power is 0.9219, so the risk is 7.8%. That lands just past the 7.5% line, in the intermediate band.

Two flags move him fast. Starting to smoke takes the same profile to 13.1%, and diabetes takes it to 14.4%. Both together reach 23.5%, past the 20% line. Untreated pressure pushed to 158 lands at 9.9%. Every one of those is the same man on paper; the equation just weighs each combination differently.

Same inputs, White equation: 6.9%

Swap only the race column. The White male equation carries a cholesterol coefficient of 11.853 against the Black male 0.302, offset by interactions, and its mean is 61.18 against 19.54. Feed the same 55-year-old profile through both and the White version returns 6.9%, a full point under the 7.8% from the Black version, enough to land one band lower.

The ordering is not stable across age, which surprises people: at 65 the White equation returns the higher number. That crossover is one reason external validation found the pooled cohort equations overestimate risk for some groups and underestimate it for others, and why the guideline restricted them to ages 40 to 79 in non-Hispanic White and African American adults. For people of other backgrounds the guideline says the White equations may be considered, with the caveat that they were not built for those populations.

The age 45 and 55 rows show how the bands work in practice: the identical profile crosses the 5% line in the mid-forties and the 7.5% line right around 54 to 55, before cholesterol or blood pressure moves at all. Age is the input that does the silent work.

The three decision lines: 5, 7.5 and 20

The 2018 cholesterol guideline kept the equations and split the output into four bands, which is where the 5%, 7.5% and 20% lines all come from. The table below is how those bands are meant to be used in a primary prevention visit for adults 40 to 75 without diabetes.

Two other thresholds orbit the same number. The 2022 USPSTF statin recommendation offers a statin at 10% or higher for adults 40 to 75 with at least one risk factor, and treats 7.5% to 10% as a selective zone. And the guideline expects a 30% LDL drop from moderate intensity treatment below 20% risk, a 50% drop at 20% or above. Note what the table does not do: no band issues a prescription. Every rung is phrased as a discussion.

Risk enhancers and the calcium tiebreaker

A percentage from seven inputs is blunt, so the 2018 guideline added risk enhancers: factors that tilt the statin conversation in the borderline and intermediate bands. The list includes a family history of early heart attack or stroke, persistently high LDL of 160 or more, chronic kidney disease, metabolic syndrome, inflammatory conditions such as rheumatoid arthritis or lupus, premature menopause or preeclampsia, high-risk ancestry, triglycerides persistently above 175, and elevated lipoprotein(a). None of these change the calculated percentage; they change what a clinician does with it.

When the decision stays tied, a coronary calcium scan acts as the tiebreaker. A score of zero in an intermediate-risk adult supports deferring a statin, while a score of 100 or above, or above the 75th percentile for age, favors starting one. The scan costs radiation and money, which is why the guideline keeps it for genuinely uncertain cases rather than ordering it for everyone.

The equations themselves have a blind spot here: they were built on cohorts from an era before statins were common, so they have no input for the medicine you already take. A person on a statin for years is scored as if untreated, which is one more reason to treat the output as a starting point.

Where the equations are known to miss

External validation since 2013 has been mixed, and the misses deserve a look before the hits are trusted. In a 2024 comparison run on NHANES data in JAMA Internal Medicine, mean 10-year ASCVD risk across US adults was 8.0% under the pooled cohort equations against 4.3% under the newer PREVENT equations, with the largest gap among Black participants. The pooled cohort equations tend to overestimate absolute risk in contemporary populations, which means some people past the 7.5% line are closer to it than the raw number suggests.

The structural limits are worth naming. The equations were derived from cohorts enrolled mostly in the 1980s and 1990s. Race is handled as a two-category input, which flattens enormous variation. The age window stops at 79 and starts at 40. Heart failure, peripheral artery disease, angina and TIA are absent from the endpoint. And statin use, the most common treatment in the population, is not an input at all.

The 2023 PREVENT equations from the American Heart Association are the acknowledged successor: they drop race, add kidney function and BMI, and produce lower numbers on the same people. They have not displaced the pooled cohort equations in the guidelines quoted above, so both are worth knowing.

Bringing the number into the room

A percentage is most useful as a conversation opener, so bring the inputs rather than just the result. Your latest lipid panel with the HDL value, recent home blood pressure readings, current medicines, smoking status, and any diabetes diagnosis cover all seven inputs. Add family history of early events, because the equation cannot see it and it is the enhancer clinicians weigh first.

Then ask three questions that the bands are built to answer. Where does the number sit against 5, 7.5 and 20? Does any risk enhancer on the list apply to me? And if the call is genuinely close, is a calcium scan the right tiebreaker? Those three questions turn a calculator output into the discussion the guideline sets out.

This page explains a published equation and is general health information only. It is not medical advice, it does not diagnose anything, and it does not recommend starting or stopping any medicine. The number is an estimate for a group of people who share your inputs, and the treatment decision belongs with a clinician who can see the rest of your history.