PREVENT Risk Calculator

Use the PREVENT Risk Calculator

Free PREVENT risk calculator using the AHA 2023 equations to estimate 10-year and 30-year cardiovascular, heart failure and stroke risk for ages 30 to 79.

How to Use PREVENT Risk Calculator

1. Step 1: Pick the outcome you care about

   Tap one of the five outcome chips at the top: Total CVD, ASCVD, Heart failure, Coronary heart disease or Stroke. Total CVD is selected by default because it is the outcome the AHA highlights.

2. Step 2: Enter age, sex, cholesterol and blood pressure

   Type your age (30 to 79) and select sex at birth, then enter total cholesterol and HDL. PREVENT works from non-HDL, so it subtracts HDL from total for you. Add systolic blood pressure and flag whether you take blood pressure medication or a statin.

3. Step 3: Add weight, height and kidney function

   Enter weight and height in metric or imperial; the calculator derives BMI. For kidney function, either type a known eGFR or switch to the creatinine tab and enter a serum creatinine in mg/dL, which is converted using the race-free CKD-EPI 2021 equation.

4. Step 4: Optionally add UACR and HbA1c

   Press Add labs and enter a urine albumin-to-creatinine ratio in mg/g and an HbA1c percentage if you have them. The badge under the results changes from Base model to + UACR, + HbA1c or Full model so you can see which PREVENT equation produced the number.

5. Step 5: Read the 10-year and 30-year panels

   The two cards show absolute risk for the selected outcome with its risk band. The 30-year card stays greyed out above age 59 because the 30-year equations were only fitted to ages 30 to 59.

6. Step 6: Check what each change is worth

   Scroll to the levers panel to see how many percentage points quitting smoking, dropping systolic BP to 110 mmHg, cutting non-HDL to 120 mg/dL or reaching BMI 24 would remove from your risk. Use Copy all 10 results to save every outcome at both horizons.

Key Features

• All five PREVENT outcomes: total CVD, ASCVD, heart failure, CHD and stroke
• 10-year and 30-year absolute risk from the AHA 2023 equations
• Optional UACR and HbA1c switch to the enhanced CKM equations
• Serum creatinine converted to eGFR with CKD-EPI 2021
• Ranked panel showing what each modifiable factor is worth in percentage points
• mg/dL and mmol/L, metric and imperial unit toggles

Understanding Your PREVENT Results

The Equation Behind This PREVENT Risk Calculator

PREVENT is a sex-specific logistic model. Each predictor is centred and rescaled before it is multiplied by its coefficient: age enters as (age − 55) / 10, non-HDL cholesterol as mmol/L minus 3.5, HDL as (mmol/L − 1.3) / 0.3, and systolic blood pressure is split at 110 mmHg into two linear segments so that low and high pressure can bend the curve differently. eGFR is split the same way at 60 mL/min/1.73m², and BMI at 30 kg/m². Eight interaction terms then let age modify the effect of cholesterol, blood pressure, diabetes, smoking, obesity and kidney function. The weighted sum plus a constant goes through the logistic transform risk = e^x / (1 + e^x) to produce absolute risk.

There is not one PREVENT equation but fifty: five outcomes × two sexes × two time horizons × several optional-predictor variants. The 30-year models add an age-squared term the 10-year models do not have. This PREVENT risk calculator carries the published base, HbA1c, UACR and full coefficient tables and picks the right one from what you enter.

Reference Ranges & Interpretation

Ten-year risk is read in four bands: under 5% is low, 5% to 7.4% borderline, 7.5% to 19.9% intermediate and 20% or more high. The 7.5% line is the one with teeth — the 2025 AHA/ACC high blood pressure guideline recommends starting antihypertensive medication in stage 1 hypertension once 10-year total CVD risk by PREVENT reaches 7.5%. Thirty-year risk uses wider bands, with 10%, 20% and 30% as the dividing lines, because a longer horizon accumulates more events by construction.

Do not add the five outcomes together. Total CVD, ASCVD, heart failure, coronary heart disease and stroke are five separately fitted models over overlapping event definitions, so for the same person they might read 14.7%, 9.2%, 8.1%, 4.4% and 5.4% respectively. Total CVD is the composite the AHA highlights and the outcome the hypertension guideline threshold refers to.

Assumptions & Limitations

PREVENT is a primary-prevention instrument. It assumes you have not already had a heart attack, stroke or heart failure diagnosis; if you have, your risk is already established and secondary prevention targets apply instead. Ten-year estimates are valid for ages 30 to 79 and 30-year estimates only for 30 to 59, because those were the age windows in the derivation cohorts of 3,281,919 adults across 25 datasets.

Two limits are worth naming explicitly. First, this implementation leaves the Social Deprivation Index as missing, because deriving it needs a US ZIP-code lookup; the published equations handle that with a dedicated missing-SDI indicator, which is what the calculator uses. Second, PREVENT was derived overwhelmingly on US data, so calibration outside the United States is not guaranteed — external validations have generally found good discrimination but variable calibration. Discrimination for total CVD was a C-statistic of 0.794 in women and 0.757 in men, which is good but far from deterministic: many people at 5% risk have events and most people at 20% risk do not.

Complete Guide: PREVENT Risk Calculator

Run a 55-year-old man through this PREVENT risk calculator and then through the tool his doctor was using in 2023, and you can get two answers that point in opposite directions. Total cholesterol 210, HDL 38, systolic 135 untreated, a current smoker, BMI 31, eGFR 60. The Pooled Cohort Equations return 15.1% — comfortably past the 7.5% line that has triggered statin conversations for a decade. PREVENT returns 6.3%. Same man, same labs, same afternoon, and one model calls him a clear candidate for treatment while the other files him under borderline.

That is not a bug in either calculator. It is the whole point of the 2023 rewrite, and it is why anyone who has tracked their cardiovascular risk for years is about to see their number move without anything about their health changing. This guide walks the two models side by side: where they diverge, what got added, what got deleted, and which of the ten numbers PREVENT produces you should actually carry into an appointment.

Where PREVENT and the Pooled Cohort Equations Split

The divergence is not uniform. At low risk the two models agree almost exactly; the gap opens as risk rises. The table below runs five profiles through both published coefficient sets, comparing 10-year ASCVD risk. All five assume no prior cardiovascular disease and no statin. Because the Pooled Cohort Equations require a race input that PREVENT deliberately removed, the White/Other coefficients are used for the older model.

Read the third row again, because that is the one with consequences. The 55-year-old smoker crosses the 7.5% threshold in one direction under the old model and the other direction under the new one. A 13.7-point drop for the 60-year-old is even larger in absolute terms, but he stays above 7.5% either way, so his treatment plan probably does not change. The people whose care actually shifts are clustered in a band roughly between 7% and 20% on the old scale. If you want to see the old model's answer for yourself, our Pooled Cohort Equations calculator implements the 2013 equations unchanged.

What Actually Changed Inside the Equation

Four structural changes separate PREVENT from its predecessor, and each one has a visible effect on the output.

Race came out. The Pooled Cohort Equations carried separate coefficient sets for White and African American adults, which meant two people with identical biology received different risk estimates. The AHA removed the variable on the reasoning that race is a social rather than biological construct, and that the coefficient had been standing in for unmeasured social and environmental exposure. In its place PREVENT offers an optional Social Deprivation Index derived from US ZIP code.

Kidney and metabolic health came in. eGFR, BMI, HbA1c and urine albumin-to-creatinine ratio are all predictors now. This is the cardiovascular-kidney-metabolic framing the AHA has pushed since 2023: kidney disease and cardiovascular disease are treated as one syndrome rather than two specialties. If you do not know your eGFR, the calculator above will derive it from a serum creatinine using CKD-EPI 2021, or you can work it out separately with our eGFR calculator.

Non-HDL replaced the total-and-HDL pair. PREVENT still asks for total cholesterol and HDL, but only so it can subtract one from the other. The model is fitted on non-HDL cholesterol in mmol/L, centred at 3.5. Notably it does not use LDL at all, which trips people up when they bring a lipid panel where LDL is the number in bold. Our cholesterol calculator will break a full panel into its components if you need the arithmetic done.

The predictors bend instead of running straight.The old model logged everything and left it linear. PREVENT splits systolic blood pressure at 110 mmHg, eGFR at 60 mL/min/1.73m² and BMI at 30 kg/m² into two segments each, then adds eight age-interaction terms so that the same risk factor can weigh differently at 40 than at 70. That is why the effect of a variable is not constant across its range, which the next two sections make concrete.

Why Your Number Dropped by Half

The systematic downward shift is a calibration story, not a biology story. The Pooled Cohort Equations were fitted on cohorts recruited between the 1960s and the 1990s — ARIC, CHS, CARDIA, Framingham. In those decades roughly a third of American adults smoked, statins were unavailable for most of the follow-up, and hypertension frequently went untreated for years. Event rates were simply higher.

PREVENT was derived on 3,281,919 adults across 25 contemporary datasets, in a population where statin therapy, blood pressure control and lower smoking prevalence had already suppressed events. A model learns the event rate of the era it was trained on. Applying a 1990s event rate to a 2026 patient inflates the estimate, which is exactly the overestimation that repeated external validations of the Pooled Cohort Equations had been reporting for a decade.

The uncomfortable corollary: the 7.5% statin threshold in the 2018 cholesterol guideline was calibrated against Pooled Cohort numbers. Feeding a systematically lower PREVENT estimate into a threshold that assumed higher inputs would shrink the treated population substantially, and that recalibration has not been settled. The 2025 AHA/ACC high blood pressure guideline is the first major guideline to build PREVENT in explicitly, and it uses 10-year total CVD of 7.5% or more as one trigger for starting antihypertensive therapy in stage 1 hypertension. Note that it names total CVD, not ASCVD — a distinction the last section returns to.

The Kidney Curve Is Flat Until It Is Not

Because eGFR is split at 60, its effect is deliberately mild in the normal range and steep below it. Holding a 55-year-old man fixed at total cholesterol 200, HDL 45, systolic 130 untreated, non-smoker, no diabetes, BMI 27, and varying only kidney function:

From 90 down to 60, ASCVD risk moves two-tenths of a percentage point — effectively nothing. Below the knot it accelerates hard: 45 gives 5.0% and 30 gives 7.0%, more than doubling from the starting value. Total CVD nearly triples over the same span, from 5.0% to 13.7%, and heart failure almost quadruples. The practical reading is that a mildly reduced eGFR in the 60s is not worth anxiety, while a result under 45 changes the picture materially and belongs in a conversation about staging, which our CKD staging calculator handles using the KDIGO categories.

BMI Moves Heart Failure and Nothing Else

This is the finding that surprises people who read the coefficient tables closely. BMI is advertised as a PREVENT predictor, and it is — but only for one outcome. In the published base model the BMI terms are exactly zero for ASCVD, coronary heart disease, stroke and total CVD. They are non-zero only in the heart failure equations, where the BMI-above-30 coefficient is 0.2998 for women and 0.3727 for men per 5 kg/m².

Hold a 55-year-old woman fixed and move only her BMI, and the arithmetic is stark:

Fifteen BMI points and the heart attack number does not move at all, while heart failure risk rises by half. This is not the model claiming obesity is harmless to arteries; it is the model saying that once you already know someone's blood pressure, lipids, glucose and kidney function, BMI adds no further information about atherosclerotic events — but it still adds information about the myocardium. If you are entering weight and height into the calculator above and watching the ASCVD figure sit still, that is the coefficient table behaving as published, not a broken input. Our adult BMI calculator covers what the index does and does not capture in more depth.

A Worked Example: 40 Years Old, Two Horizons

The 30-year output is the capability the Pooled Cohort Equations never had, and it matters most for exactly the people the old model was worst at serving: adults in their thirties and forties whose 10-year risk is trivially small. Take two 40-year-olds, both with total cholesterol 195, HDL 48, BMI 27, eGFR 95, no diabetes, no medication.

• Woman, systolic 118, non-smoker. 10-year total CVD 0.9%. Thirty-year total CVD 7.1%, of which heart failure accounts for 2.6%.
• Man, systolic 132, current smoker. 10-year total CVD 2.9%. Thirty-year total CVD 18.0%, with heart failure at 8.2%.

On a 10-year view the second person looks fine — 2.9% is nowhere near any treatment threshold, and under the old model he would have been told to come back in five years. On a 30-year view he is carrying nearly one-in-five odds of a cardiovascular event before he turns 70, and roughly a one-in-twelve chance of heart failure specifically. That is a different conversation, and it is the conversation that actually changes outcomes, because the smoking and the blood pressure are both still reversible at 40. The 30-year models stop at age 59 by design, since projecting three decades from 65 would run past the follow-up available in the derivation cohorts.

When Adding a Lab Result Lowers Your Risk

Here is a counterintuitive property of the optional inputs. The base model does not assume your urine albumin is normal; it prices in the average risk of a population whose albumin status is unknown, using a dedicated missing-value indicator. Supply a genuinely normal result and the estimate goes down. Take a 55-year-old diabetic man, total cholesterol 200, HDL 42, systolic 138 on treatment, on a statin, BMI 31, eGFR 70:

A normal albumin result buys this man 2.1 percentage points off his total CVD estimate; a severely increased one adds nearly 11. The spread from 12.5% to 25.4% is wider than anything his cholesterol could do, which is the argument for ordering the test rather than leaving the field blank. HbA1c behaves the same way, with separate coefficients depending on whether you have diabetes, and both together switch the calculator to the full model. If your glucose control is the open question, our diabetes risk calculator works through screening thresholds in more detail.

Which Number to Take to Which Appointment

PREVENT hands you ten figures. A short decision framework for which one is load-bearing:

• Blood pressure conversation: 10-year total CVD. This is the outcome the 2025 AHA/ACC hypertension guideline threshold of 7.5% refers to. Pair it with our blood pressure calculator to confirm which stage your readings fall into first.
• Statin conversation: 10-year ASCVD, with the caveat that the familiar 7.5% cut-point was calibrated on Pooled Cohort numbers and has not been formally re-derived for PREVENT. Bring both estimates if you have them.
• You are under 50 and everything looks fine: 30-year total CVD. The 10-year figure is almost guaranteed to be reassuring at that age and tells you very little.
• Obesity, diabetes or reduced eGFR in the picture: the heart failure row. It is the only outcome BMI touches and it moves faster than ASCVD as kidney function declines.

One last framing worth keeping. A C-statistic of 0.794 in women and 0.757 in men is good discrimination for a risk model, and it is still a long way from prophecy: plenty of people estimated at 5% have events, and most people estimated at 20% do not. The number's job is to size a decision, not to predict your life. If you want a second reading from a differently constructed model, the Framingham cardiac risk calculator and the Reynolds risk score, which adds high-sensitivity CRP, weight the same risk factors differently and are useful for seeing how much of your estimate is model choice rather than biology.

References

1. 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. ahajournals.org
2. Ndumele CE, Rangaswami J, Chow SL, et al. Cardiovascular-Kidney-Metabolic Health: A Presidential Advisory From the American Heart Association. Circulation. 2023;148(20):1606-1635. ahajournals.org
3. 2025 AHA/ACC Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults. Journal of the American College of Cardiology. 2025. jacc.org
4. Goff DC, Lloyd-Jones DM, Bennett G, et al. 2013 ACC/AHA Guideline on the Assessment of Cardiovascular Risk. Circulation. 2014;129(25 Suppl 2):S49-S73. ahajournals.org
5. Inker LA, Eneanya ND, Coresh J, et al. New Creatinine- and Cystatin C-Based Equations to Estimate GFR without Race. New England Journal of Medicine. 2021;385:1737-1749. nejm.org

Risk figures in the tables above were computed with the published PREVENT and Pooled Cohort coefficient sets for the stated profiles and are illustrative, not clinical advice.

Frequently Asked Questions