10-Year CVD Risk Calculator

A risk score built around your cholesterol numbers

The pooled cohort equations are the risk formula that shipped with the 2013 ACC/AHA cholesterol guideline, published in Circulation by Goff and colleagues. They take nine inputs: age, sex, race group, total cholesterol, HDL cholesterol, systolic blood pressure, whether you take pressure medicine, whether you smoke now, and whether you have diabetes. Four versions of the formula exist, one for each sex and race group combination.

The output is the chance of a first hard atherosclerotic event within ten years. Hard means four outcomes only: a nonfatal heart attack, death from coronary heart disease, a nonfatal stroke, and death from a stroke. Angina, a stent, bypass surgery, heart failure and peripheral artery disease are outside the definition. A person can score 4% here and still end up with chest pain in the next decade.

This page is built for the lipid question. Two of the nine inputs are cholesterol values, and both are measured on every standard panel, so the score is the one place where a total cholesterol of 245 and an HDL of 38 turn into a number you can compare against a treatment threshold. The 2022 USPSTF statin recommendation uses exactly this kind of estimate: for adults 40 to 75 with at least one risk factor, a 10-year risk of 10% or higher earns a grade B recommendation to start a statin, and 7.5% to just under 10% earns a grade C recommendation to offer one selectively.

Read the result as a screening reference, not a diagnosis. It describes a group of people who share your numbers, not your personal future, and the decision that follows belongs in an exam room with your history in view.

How total cholesterol and HDL enter the formula

Every input is converted to a natural log, multiplied by a weight from the published table, and added to a running sum. Cholesterol is not a flat addition. In the White female equation the total cholesterol term is 13.540 times ln(total cholesterol), and a correction of minus 3.114 times ln(age) times ln(total cholesterol) is applied on top. At age 55 that leaves a net weight of about 1.06 per log unit, so a 10% higher total cholesterol multiplies the underlying odds by roughly 1.11. HDL carries a negative weight, minus 13.578 plus 3.149 times ln(age), which at 55 works out to about minus 0.96 per log unit. A 10% higher HDL multiplies the odds by roughly 0.91.

That structure explains why the same panel means different things at different ages. At 45 the age correction is smaller, so total cholesterol counts for a little more; at 70 it is larger, so the same total cholesterol counts for less, and age itself dominates. The formula is telling you something clinically familiar: a high total cholesterol at 45 is a long exposure, and a high total cholesterol at 75 is a shorter one on top of a decade count that already carries most of the risk.

Two practical consequences follow. First, the ratio of total cholesterol to HDL matters more than either number alone, because the two coefficients push the sum in opposite directions. A total of 220 with HDL 65 can score lower than a total of 190 with HDL 35. Second, because the equation uses logs, a change from 240 to 210 is worth more than a change from 210 to 180. The first 30 points buy a bigger drop in the log term than the second 30.

Systolic pressure, smoking and diabetes enter the same way, with their own weights. Smoking is the largest single term in the male equation, and it is also the only one that flips completely on a yes or no answer. Diabetes adds a fixed amount, around 0.66 in the White equations, which is why a person with diabetes and clean lipids can still land in a treatment band.

Two profiles, worked end to end

Take a 55-year-old woman, White, total cholesterol 245, HDL 48, systolic pressure 128, no pressure medicine, not a smoker, no diabetes. The weighted sum comes to -29.36 against a cohort mean of -29.18, a difference of -0.18. Raising e to that power gives 0.83, and the baseline survival of 0.9665 raised to 0.83 gives 0.972. One minus 0.972 is 2.8%, in the low band. Out of 1,000 women with this profile, about 28 would have a first event in ten years.

Now take a 62-year-old man, White, total cholesterol 210, HDL 40, systolic 132, no pressure medicine, not a smoker, no diabetes. His total cholesterol is 35 points lower than hers, yet he scores 12.8%. The weighted sum is 61.61 against a mean of 61.18. The difference is age and the low HDL: at 62 the age term alone carries most of the total, and an HDL of 40 adds weight in the wrong direction. His score sits above the 10% line where USPSTF 2022 recommends starting a moderate intensity statin, and about 128 of 1,000 men with his profile would have a first event within ten years.

That pair is the whole argument for running the numbers instead of eyeballing the lipid panel. The woman has the worse cholesterol and the lower risk. The man has the better cholesterol and the higher risk, and his HDL is the part of his panel that pushed him over the line.

The last row is the stacking problem. One treated condition and one diagnosis move the same 62-year-old from 12.8% to 26.7%, more than double, without a single cholesterol value changing. Lipids are the lever this page is built around, but they are not the only lever, and a lipid-only reading of the score will understate risk whenever pressure or glucose is in play.

Matching the result to statin thresholds

Two guideline systems quote different cutoffs, and both are in use. The 2013 ACC/AHA guideline that produced the equation puts the statin conversation at 7.5% for adults 40 to 75 with LDL between 70 and 189 mg/dL. The 2022 USPSTF recommendation splits the same range into a firm recommendation at 10% and a selective offer at 7.5% to 10%, and it requires at least one risk factor: dyslipidemia, diabetes, hypertension or smoking.

The gap between 7.5% and 10% is not a contradiction. USPSTF graded the evidence separately for the two slices and found the net benefit smaller in the lower one, so the wording changes from start to offer. In practice a score of 8% or 9% is a decision, not a prescription. Risk enhancers usually settle it: a parent with a heart attack before 55, LDL of 160 or higher, chronic kidney disease, metabolic syndrome, an inflammatory condition such as rheumatoid arthritis or psoriasis, or South Asian ancestry.

A coronary artery calcium scan is the other tiebreaker. It costs roughly $100 to $400 without insurance in most markets, and a score of zero reclassifies a large share of intermediate-risk adults downward, often enough to defer a statin. A score above zero argues the other way. The scan is most useful in exactly the 7.5% to 20% zone where the conversation is genuinely open, and least useful in the high band, where treatment is already indicated.

The lipid levers, measured

The calculator shows two lipid scenarios next to your result, and the size of each move is worth understanding before you treat either as a target.

For the 55-year-old woman at 2.8%, dropping total cholesterol by 30 mg/dL takes her to 2.4%, a fall of 0.4 points. Raising HDL by 10 mg/dL takes her to 2.3%, a fall of 0.5 points. Both moves are small in absolute terms because her baseline is low; there is not much risk left to remove. For the 62-year-old man at 12.8%, the same 30 point cholesterol drop takes him to 11.3% and the HDL rise takes him to 11.1%, moves of 1.5 and 1.7 points. The same lipid change is worth three to four times more to him, in absolute points, because he has more risk on the table.

That asymmetry is the main reason a low score is not a reason to ignore a bad panel. The percentage drop is similar in relative terms, roughly 10% to 15% of the score in both cases, but the clinical value of the drop scales with the starting risk.

How much of this a person can achieve without medicine depends on the starting lipid profile. Diet and exercise move LDL by roughly 10% to 20% in most people, with larger drops in people who change several habits at once. A moderate intensity statin, by the standard definition used in the ACC/AHA guidelines, lowers LDL by 30% to 49%. If a statin cuts LDL by a third and LDL is about 60% of total cholesterol, the total cholesterol number falls by roughly 20%, which is more than half of the lever this page models. That is why the score and the treatment decision are linked: the same panel that produces the number also predicts how far the number can fall.

Two cautions. The equation does not model LDL directly, so do not translate an LDL target into a score change by hand; rerun the calculator with the new total cholesterol instead. And the score does not fall as fast as the panel does. A statin starts lowering LDL within weeks, while the risk reduction in trials accumulates over years, which is one reason guidelines talk about treating for decades rather than for a single lab cycle.

Where the equation stops being useful

The score is a good ranking tool and an imperfect absolute forecast. Both halves of that sentence matter.

• Age dominates everything. The same 62-year-old man at 72, with every other input frozen, scores 23.9%. The woman at 55 scores 6.6% at 65. Ten years of calendar time outweighs most realistic lipid changes, which is why a low score at 45 should not be read as a permanent verdict.
• Race is a two-way switch. The equations separate African American from White and fold everyone else into the White version. They tend to underestimate risk in South Asian Americans and overestimate it in East Asian Americans.
• The training cohorts are old. The participants were followed in the 1990s, before statins were routine. Validation studies in modern populations have repeatedly found the equations run high, sometimes by a fifth or more. Your doctor may quote a lower number from a newer tool and both can be correct.
• PREVENT is the 2023 replacement. The American Heart Association's Predicting Risk of cardiovascular disease EVENTs equations add kidney function, urine protein, HbA1c and social factors, cover ages 30 to 79, and generally produce lower estimates. Many clinics still print the pooled cohort score because the 2013 and 2022 treatment thresholds were written around it.
• It is a primary prevention tool. Anyone who has had a heart attack, stroke, stent or bypass is in secondary prevention, where treatment is already indicated and the score adds nothing. The same applies to LDL of 190 mg/dL or higher, and to familial hypercholesterolemia, which the guidelines handle as a separate diagnosis.
• It ignores several real risk factors. Family history, waist circumference, kidney function, inflammatory disease, sleep apnea and social determinants are all outside the formula. That is what risk enhancers are for.

None of this makes the number useless. A tool that ranks people correctly is exactly what a treatment threshold needs, and the 7.5% and 10% lines are built on that ranking. It does mean the number should start a conversation, not end one.

What to bring to the appointment

A printout of the score is most useful when the raw inputs sit next to it. Bring four things: the lipid panel with total cholesterol and HDL and the date it was drawn, an average of several home blood pressure readings rather than one office number, an honest smoking answer that includes occasional use, and the family history detail that matters, meaning which relative, which event, and at what age.

Then ask three questions. Does this score change what you would recommend today? If the answer involves a statin, what intensity, and what LDL drop would you expect? And if the score is in the 7.5% to 20% range, would a calcium scan change the plan before I commit to a daily pill?

For background reading between visits, the NHLBI blood cholesterol guide explains what each number on the panel means, and the CDC heart disease overview puts the population scale behind the risk. Neither replaces the score, and the score does not replace the visit. Used together, they make the ten minutes with a clinician more specific than a general talk about eating better.

One last note on rhythm. If your numbers are stable and your score is low, rechecking every four to six years is reasonable. If you have diabetes, kidney disease or a strong family history, or if you just started or stopped a statin, the interval should be shorter, because the inputs that drive the score can change faster than the calendar suggests.