A 55-year-old man has a total cholesterol of 213 mg/dL, HDL of 50 mg/dL, and an untreated systolic pressure of 120 mmHg. He doesn't smoke, doesn't have diabetes and takes no statin. This is the worked example printed in the 2013 risk assessment guideline, which gave him a 10-year risk of 5.3% on the pooled cohort equations (PCE), using the equations for white men.4 On the PCE scale, that is the borderline range (5% to <7.5%).3
Run through the PREVENT™ equations, with an eGFR of 90, the same man's 10-year ASCVD risk is 2.8%: low risk under the 2026 categories. His 30-year ASCVD risk is 14.9%.
Two things happened. The number dropped by roughly half, which is what the 2026 guideline says to expect. And the 30-year figure, which the PCE never offered, crosses the one 30-year threshold the guideline names. This guide covers why PREVENT runs lower, why the old cutoffs don't apply to it, the inputs worth checking, and how to read the two time horizons.
Illustrative case: the 2013 guideline's worked example, extended with an assumed eGFR of 90 and no statin; PREVENT figures from the Curie Calc implementation.
At a glance
- PREVENT estimates 10-year (ages 30 to 79) and 30-year (ages 30 to 59) risk of a first ASCVD event. The 2026 ACC/AHA dyslipidemia guideline uses it in place of the PCE.
- For the same patient, PREVENT's 10-year estimates run about 40% to 50% lower than the PCE's. The 2026 guideline lowered its risk categories to match: 3%, 5% and 10% instead of 5%, 7.5% and 20%.
- A PCE threshold such as 7.5% does not apply to a PREVENT number.
What PREVENT measures
PREVENT estimates the probability of a first atherosclerotic cardiovascular event: myocardial infarction, fatal coronary heart disease, or fatal or nonfatal stroke.1 The American Heart Association published the equations in 2023 and 2024, derived and validated in large contemporary US datasets.12 They predict several outcomes; Curie Calc reports the ASCVD outcome, the one the cholesterol guideline uses.
The 10-year estimate covers adults aged 30 to 79 without established cardiovascular disease. The 30-year estimate covers ages 30 to 59 only.13 The equations are sex-specific. Race and ethnicity are not predictors: the authors describe them as social constructs and kept them only to check calibration.1
Compared with the PCE, PREVENT adds kidney function (eGFR) and statin use as inputs, drops race, and estimates risk with an adjustment for the competing risk of non-cardiovascular death.13 The base model uses ten inputs; optional add-ons (urine albumin-to-creatinine ratio, HbA1c and a social deprivation index) form extended models, which Curie Calc does not offer.
Source: Khan 20241; Khan 20232; 2026 ACC/AHA dyslipidemia guideline3.
Why PREVENT runs lower than the PCE
The 2026 guideline states that PREVENT's 10-year estimates are about 40% to 50% lower than the PCE's for the same risk factor profile, and attributes the gap to the PCE overestimating risk.3 The guideline notes that the PCE were derived in about 25,000 people from older birth cohorts, while PREVENT was derived in about 3.3 million contemporary US adults and validated in a similar number.3
The derivation paper puts a number on the gap at the familiar statin threshold. At a PCE estimate of 7.5%, the median PREVENT-ASCVD estimate was 4.9% in women and 3.7% in men.1 In a weighted national sample of US adults aged 40 to 75 without known ASCVD, mean predicted 10-year ASCVD risk was 8.0% on the PCE and 4.3% on PREVENT.5
The gain is in calibration, not ranking. PREVENT places patients in roughly the same order as the PCE: its C-statistics for ASCVD were 0.774 in women and 0.736 in men, an improvement of about 0.005 to 0.007.1 What changed is how closely the predicted percentages match observed event rates in the derivation and validation data.3 Independent validation in four US health systems found that calibration varied substantially by site.9
Why the old cutoffs don't carry over
Because the numbers moved down, the thresholds moved with them. The 2026 guideline publishes a crosswalk of approximately equivalent ranges:3
| Risk group | PCE (10-year) | PREVENT-ASCVD (10-year) |
|---|---|---|
| Low | <5% | <3% |
| Borderline | 5% to <7.5% | 3% to <5% |
| Intermediate | 7.5% to <20% | 5% to <10% |
| High | ≥20% | ≥10% |
2026 ACC/AHA dyslipidemia guideline, Table 12, adapted from Khan et al.3
Applying a PCE cutoff to a PREVENT number shifts patients down. Two 2024 studies of national survey data applied PREVENT to the thresholds then in use, before the 2026 guideline: one estimated that 17.3 million adults aged 40 to 75 would lose statin eligibility,5 and another that about half of adults aged 30 to 79 would move to a lower risk category and statin eligibility would fall by 14.3 million.6 Neither tested the 2026 thresholds. With the new thresholds, the guideline's own estimate is that about 25 million adults would be identified at a PREVENT risk of 3% or more, against 26 million at a PCE risk of 5% or more.3
The case above shows that the crosswalk is approximate for any one patient. His PCE estimate of 5.3% sits in the PCE borderline range; his PREVENT estimate of 2.8% sits just under PREVENT's.
Inputs worth checking
Are the cholesterol values in mg/dL or mmol/L?
The equations work internally in mmol/L, while US labs report mg/dL. A value of 200 read as mmol/L instead of mg/dL is off by a factor of about 39, so the unit a tool assumes decides the result. Curie Calc opens in mg/dL and converts internally.
Which eGFR?
PREVENT was derived with eGFR from the 2021 CKD-EPI creatinine equation, without race.13 A Cockcroft-Gault creatinine clearance is a different measurement (mL/min, not indexed to body surface area) and is not the specified predictor. Curie Calc's CKD-EPI 2021 calculator gives the matching value.
Does "statin use" include other lipid-lowering drugs?
The derivation identified statin use specifically, and Curie Calc counts statins only. The specifications don't address ezetimibe or a PCSK9 inhibitor used alone.1
Does diabetes mean any diabetes?
Yes. The PREVENT item is any history of diabetes. That differs from some other scores: RCRI, for example, counts only insulin-treated diabetes.
What counts as current smoking and as blood pressure?
Current smoking means cigarette use in the past 30 days.1 Systolic pressure is the measured value, and antihypertensive treatment is a separate input, so a treated patient with a normal reading has both recorded.
Why is there no BMI?
BMI enters PREVENT only through the heart failure equations. It has no coefficient in the ASCVD equations, so it cannot change an ASCVD estimate.1 Curie Calc's ASCVD calculator does not ask for it.
Reading the result: 10-year and 30-year risk
The 10-year estimate carries the guideline's four risk categories. The 2026 guideline recommends (Class 1, Level B-NR) that in adults aged 30 to 79 without ASCVD or subclinical atherosclerosis and with an LDL-C between 70 and 189 mg/dL, the PREVENT-ASCVD equations be used to estimate 10-year risk, categorised as low (<3%), borderline (3% to <5%), intermediate (5% to <10%) or high (≥10%).3 The guideline explains the 3% boundary as the point where, in primary prevention statin trials, the expected benefit of a moderate-intensity statin begins to outweigh the risk of new diabetes.3
The 30-year estimate carries no categories. The guideline's four categories apply to 10-year risk only. It names one 30-year threshold: in adults aged 30 to 59 with a low 10-year risk, a 30-year ASCVD risk of 10% or more is among the findings it lists when discussing lipid-lowering therapy.3 It gives the reason as the long gap between exposure to atherogenic lipids and events, which a 10-year window misses.3
Two cautions on the 30-year figure. It is produced only for ages 30 to 59, so it is absent above 59. And it is a looser approximation than the 10-year one: the published equations fit the underlying model with an R² of 0.976 to 0.9945 at 30 years, against 0.9987 to 0.9997 at 10 years.1
Both figures are group estimates: the share of people with similar measured risk factors expected to have a first event. They are not a forecast for one patient.
Worked examples
Synthetic cases unless noted. Any input not mentioned is "no". Figures computed with the Curie Calc PREVENT implementation.
| Case | 10-year ASCVD risk | Category | 30-year ASCVD risk | Learning point |
|---|---|---|---|---|
| AHA's own published example: woman, 50, TC 200, HDL 45, treated SBP 160, diabetes, eGFR 90 | 9.2% | Intermediate | 35.4% | Matches the equations' published worked example. |
| Woman, 45, TC 230, HDL 48, untreated SBP 135, eGFR 95 | 1.5% | Low | 10.1% | Low 10-year risk, 30-year risk just over 10%. |
| Man, 40, smoker, TC 240, HDL 40, untreated SBP 132, eGFR 100 | 2.9% | Low | 16.4% | A young smoker can sit just under the 3% boundary with a substantial 30-year risk. |
What the score doesn't tell you
PREVENT estimates a first event. It does not apply to people with established ASCVD or heart failure, to ages outside 30 to 79, or, for the 30-year figure, above 59.1 The equations are sex-specific and no coefficient set exists for a patient outside a binary sex. No estimate is produced when a required input is missing or outside the supported range: total cholesterol 130 to 320 mg/dL, HDL 20 to 100 mg/dL, systolic pressure 90 to 200 mmHg, eGFR 15 to 150 mL/min/1.73 m².
Performance varies by setting. A validation across 6.4 million people in 44 observational cohorts and 18 trials reported generally good performance across regions.7 Within the US, a national study found close calibration,8 while a study across four health systems found calibration varied by site.9
Family history, lipoprotein(a), coronary artery calcium and inflammatory conditions are not inputs. The 2026 guideline treats them as factors to personalise the estimate after it is calculated, not as part of it.3
About PREVENT™
The PREVENT™ equations were developed by select members of the American Heart Association Cardiovascular-Kidney-Metabolic Scientific Advisory Group. The risk equations were derived and validated in a large, diverse sample of over 6 million individuals.[1], [2]
[1] Khan SS, Matsushita K, Sang Y, et al. Development and Validation of the American Heart Association Predicting Risk of Cardiovascular Disease EVENTs (PREVENT™) Equations. Circulation 2023. DOI: 10.1161/CIRCULATIONAHA.123.067626.
[2] Khan SS, Coresh J, Pencina MJ, et al. Novel Prediction Equations for Absolute Risk Assessment of Total Cardiovascular Disease Incorporating Cardiovascular-Kidney-Metabolic Health: A Scientific Statement From the American Heart Association. Circulation 2023;148(24):1982-2004. DOI: 10.1161/CIR.0000000000001191.
The information provided herein is based in part on PREVENT™, which is the American Heart Association's (AHA) online risk calculator for primary prevention patients. The use of the AHA's material is limited to certain aspects of PREVENT™.
The information derived from the use of PREVENT™ is based on PREVENT™ 1.0.0. Updates and future adaptations of PREVENT™ may yield different results and conclusions.
Frequently asked questions
Does the 7.5% statin threshold apply to PREVENT?
No. That threshold was defined against the PCE. The 2026 guideline uses 3%, 5% and 10% for 10-year PREVENT-ASCVD risk, and publishes a crosswalk placing PCE 7.5% to <20% alongside PREVENT 5% to <10%.3
Why is my patient's PREVENT risk so much lower than their old ASCVD risk?
Which eGFR does PREVENT use?
The 2021 CKD-EPI creatinine equation, without race.1
Does PREVENT include race?
No. The authors excluded race and ethnicity as predictors.1
Sources
- Khan SS, Matsushita K, Sang Y, et al. Development and validation of the American Heart Association's PREVENT equations. Circulation. 2024;149(6):430-449. Corrections: Circulation. 2024;149(11):e956 (PMID 38466792); Circulation. 2026;154(17):e665 (PMID 42872321). (opens in new tab)
- Khan SS, Coresh J, Pencina MJ, et al. Novel prediction equations for absolute risk assessment of total cardiovascular disease incorporating cardiovascular-kidney-metabolic health: a scientific statement from the American Heart Association. Circulation. 2023;148(24):1982-2004. (opens in new tab)
- Blumenthal RS, Morris PB, et al. 2026 ACC/AHA/AACVPR/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of dyslipidemia. Circulation. 2026;153(17):e1154-e1276. Corrections: 2026;153(25):e1447 (PMID 42330109); 2026;154(10):e393 (PMID 42709919). (opens in new tab)
- Goff DC Jr, Lloyd-Jones DM, Bennett G, et al. 2013 ACC/AHA guideline on the assessment of cardiovascular risk. Circulation. 2014;129(25 Suppl 2):S49-73. (opens in new tab)
- Anderson TS, Wilson LM, Sussman JB. Atherosclerotic cardiovascular disease risk estimates using the Predicting Risk of Cardiovascular Disease Events equations. JAMA Intern Med. 2024;184(8):963-970. (opens in new tab)
- Diao JA, Shi I, Murthy VL, et al. Projected changes in statin and antihypertensive therapy eligibility with the AHA PREVENT cardiovascular risk equations. JAMA. 2024;332(12):989-1000. (opens in new tab)
- Neuen BL, Major RW, Grams ME, et al. Multinational validation of the PREVENT and SCORE2 cardiovascular risk equations across 6.4 million individuals. Nat Med. 2026;32(7):2629-2638. (opens in new tab)
- Khan SS, Sang Y, Huang X, et al. External validation of the PREVENT equations in a national sample of US adults. Circ Popul Health Outcomes. 2026;9:e012988. (opens in new tab)
- Cho SMJ, Levin M, Chen R, et al. AHA PREVENT equations and cardiovascular disease risk in diverse health care populations. J Am Coll Cardiol. 2025;86(3):181-192. (opens in new tab)
This article is educational reference material for healthcare professionals. It does not provide medical advice, diagnosis, or treatment recommendations, and it does not decide statin or other preventive therapy for any individual patient. The Curie Calc PREVENT™ calculator implements the AHA PREVENT™ equations and was checked against the AHA's published code; this article is drawn from that implementation and the cited guidelines.