ASCVD vs PREVENT: which heart risk calculator to use in 2026

Calculators for this article: our ASCVD risk calculator (Pooled Cohort Equations) and our AHA PREVENT calculator.

Most of us learned primary prevention on one number: the 10-year ASCVD risk from the Pooled Cohort Equations, compared against a short list of thresholds. In 2026 that number has a successor, and the thresholds moved with it. The practical risk is not choosing the "wrong" calculator. It is pairing a number from one calculator with the cut-offs written for the other.

What the Pooled Cohort Equations estimate

The Pooled Cohort Equations (PCE) came with the 2013 ACC/AHA risk assessment guideline. They are sex- and race-specific equations for non-Hispanic White and non-Hispanic African American adults aged 40 to 79, and they estimate the 10-year risk of a first hard ASCVD event: nonfatal myocardial infarction, coronary heart disease death, or fatal or nonfatal stroke.^[1]

The 2018 cholesterol guideline built its primary-prevention decisions on that output. A 10-year risk of 5% to 7.5% was borderline, 7.5% to 19.9% was intermediate (where risk-enhancing factors favor starting a statin), and at 20% or more the goal was to lower LDL-C by at least 50%. Coronary artery calcium refined the uncertain middle: a CAC of zero allowed statin therapy to be withheld or delayed except in smokers, people with diabetes, or a strong family history, and a CAC of 100 or more meant a statin was indicated.^[2]

What PREVENT changes

The AHA PREVENT equations (Predicting Risk of cardiovascular disease EVENTs) were derived from 25 data sets totaling 3,281,919 adults and externally validated in 3,330,085 more from 21 data sets.^[3] The differences that matter at the bedside:

• Age range. Adults 30 to 79 without known cardiovascular disease, rather than 40 to 79.^[3]
• Race-free and sex-specific. Race is not an input.^[3]
• Kidney function and treatment. eGFR is a core predictor, as is statin use, alongside the traditional risk factors.^[3]
• More outcomes. The primary outcome is total CVD, defined as ASCVD plus heart failure, with separate equations for ASCVD and for heart failure.^[3] The equations give 10-year and 30-year estimates.^[4]
• Optional add-ons. Urine albumin-to-creatinine ratio, HbA1c and a ZIP-code social deprivation index can be added.^[3]
• Competing risk. Models were adjusted for the competing risk of non-cardiovascular death.^[3]

In external validation the median C-statistic for total CVD was 0.794 in women and 0.757 in men, with calibration slopes of 1.03 and 0.94.^[3] In the same validation data the PCE calibration slope for ASCVD ranged from 0.50 to 0.54, which the authors describe as overestimation of risk by about 50%.^[3]

Why the same patient gets a lower number

Because PCE overestimated in contemporary cohorts, PREVENT values run lower. Two national analyses put a size on it. Using NHANES data, mean 10-year ASCVD risk was 8.0% with the PCE and 4.3% with PREVENT; if the old thresholds were kept, the number of adults meeting statin criteria would fall from 45.4 million to 28.3 million, including 4.1 million people already taking a statin.^[5] A second analysis estimated that PREVENT would move about half of US adults (53.0%) into a lower ACC/AHA risk category and, with thresholds unchanged, reduce statin eligibility by an estimated 14.3 million people.^[6]

The point of those papers is not that PREVENT is wrong. It is that a better-calibrated number needs thresholds calibrated to it.

What the 2026 guideline did

The 2026 ACC/AHA multisociety guideline on the management of dyslipidemia retires and replaces the 2018 cholesterol guideline.^[7] I have not been able to read the full guideline text for this article, so the details below come from peer-reviewed analyses that describe it, and you should check them against the guideline itself before acting on them:

• It incorporates the PREVENT-ASCVD equations for 10-year ASCVD risk, replacing the Pooled Cohort Equations.^[8]
• Its 10-year PREVENT-ASCVD risk categories are less than 3% (low), 3% to less than 5% (borderline), 5% to less than 10% (intermediate) and 10% or more (high).^[8][9]
• It recommends selective use of coronary artery calcium scoring at a 10-year PREVENT-ASCVD risk of 3% to less than 10%.^[9]
• It adds a recommendation to consider statins in adults aged 30 to 59 whose 30-year ASCVD risk is 10% or more even when their 10-year risk is under 3%. One NHANES analysis estimates this single change accounts for 15.6 million additional eligible adults, and that overall eligibility rises from 25.7% under the 2018 guideline to 40.5% under the 2026 guideline among adults without ASCVD, with LDL-C of 70 mg/dL or more, not on lipid-lowering therapy.^[10]

So the 2026 framework did not simply swap calculators. It lowered the cut-offs to match the lower estimates and added a long-horizon trigger for younger adults.

Which one to use

The one combination to avoid is a PREVENT number judged against the 2018 thresholds of 5%, 7.5% and 20%. Our own PREVENT calculator says so in its output, because a value that is roughly half the PCE estimate will look reassuring against cut-offs that were never meant for it.

Practical notes on inputs

• eGFR matters now. PREVENT uses eGFR, so a current creatinine-based eGFR (CKD-EPI 2021) belongs in the visit, not just the lipid panel.^[3]
• Heart failure needs BMI. In our implementation the heart-failure estimates are withheld unless BMI is entered; the total-CVD and ASCVD equations have no BMI term.
• Statin use is an input. PREVENT takes current statin use into account, which the PCE did not.^[3]
• Add-ons are optional. Albuminuria, HbA1c and the social deprivation index refine the estimate. In the derivation paper the gain in discrimination from adding all three was small, but calibration improved when UACR was added among people with marked albuminuria.^[3]

Where decision support fits

The arithmetic is not the hard part. Remembering which equation a guideline expects, which inputs are missing from today's note, and which threshold set applies is the part that slips on a busy afternoon. Pogosh reads the note and, when a prevention question is in play, offers the relevant calculator with its inputs and its references. It never runs a calculator on its own; the clinician enters or confirms the values and decides what the result means for this patient.

Risk equations are population estimates. They inform a shared decision about a statin; they do not make it.