Reynolds Risk Score Calculator

Use the Reynolds Risk Score Calculator

Free Reynolds risk score calculator: add hs-CRP and family history to lipids and blood pressure for a sex-specific 10-year cardiovascular risk estimate.

How to Use Reynolds Risk Score Calculator

1. Step 1: Choose Female or Male

   Set the Sex toggle first. It does not adjust a coefficient — it swaps the whole equation, from the Women Health Study model published in 2007 to the Physicians Health Study II model published in 2008.

2. Step 2: Enter age and systolic blood pressure

   Type your age in years and your usual systolic reading in mm Hg. Enter the pressure you actually run at, treated or untreated; the Reynolds equations have no separate term for blood pressure medication.

3. Step 3: Enter total cholesterol and HDL

   Copy both numbers straight off your lipid panel and set the units toggle to mg/dL or mmol/L to match it. LDL and triglycerides are not inputs to the Reynolds score.

4. Step 4: Enter your hs-CRP result

   Use the high-sensitivity CRP value, not a standard CRP. Switch the Reported as toggle to mg/dL if your lab prints it that way and the field converts it to mg/L for you.

5. Step 5: Tick the risk factors that apply

   Check Current smoker, Parent had MI before 60, and Diabetes. The parental cutoff is strictly before age 60, and a diabetic woman gets an extra HbA1c field because only the female equation carries an HbA1c term.

6. Step 6: Read the score and the hs-CRP sensitivity table

   Your 10-year risk and category appear immediately. Scroll to the sensitivity table to see the same profile scored at hs-CRP values from 0.5 to 10 mg/L, which shows how much of your result inflammation is actually driving.

Key Features

• Runs the published women (JAMA 2007) and men (Circulation 2008) equations separately
• hs-CRP sensitivity table showing your risk at 0.5, 1, 2, 3, 5 and 10 mg/L
• Isolates how many risk points come from inflammation alone
• Accepts mg/dL or mmol/L lipids and mg/L or mg/dL hs-CRP
• Flags hs-CRP above 10 mg/L, diabetic men and out-of-range ages
• What-if projections for smoking, blood pressure, HDL and cholesterol

Understanding Results

The Reynolds Risk Score Calculator Formula

Both Reynolds equations build a linear predictor called B, then push it through a Cox survival step. For women, B = 0.0799×age + 3.137×ln(SBP) + 0.180×ln(hs-CRP) + 1.382×ln(total cholesterol) − 1.172×ln(HDL), plus 0.818 for a current smoker, 0.438 for a parent who had a heart attack before 60, and 0.134×HbA1c for diabetics. Risk = [1 − 0.98634^exp(B − 22.325)] × 100. For men the coefficients and constants are different throughout: B = 4.385×ln(age) + 2.607×ln(SBP) + 0.963×ln(total cholesterol) − 0.772×ln(HDL) + 0.102×ln(hs-CRP), plus 0.405 for smoking and 0.541 for family history, with Risk = [1 − 0.8990^exp(B − 33.097)] × 100.

The detail that changes how you read your own number: hs-CRP enters as a natural logarithm. Doubling your CRP does not double its contribution, it adds a fixed 0.125 to B in women regardless of whether you go from 1 to 2 or from 8 to 16 mg/L. That compression is why the sensitivity table in the calculator matters more than the single headline percentage.

Reference Ranges & Interpretation

Reynolds reports four bands: under 5% is low, 5% to 9.9% is moderate, 10% to 19.9% is moderate-high, and 20% or more is high. The endpoint behind those percentages is a composite of myocardial infarction, ischemic stroke, coronary revascularization and cardiovascular death over ten years.

The hs-CRP input has its own reference bands, set by the 2003 AHA/CDC scientific statement: below 1.0 mg/L is low relative risk, 1.0 to 3.0 mg/L is average, and above 3.0 mg/L is high. Values above 10 mg/L are not interpretable as vascular risk and should be repeated after at least two weeks. Because within-person variability is wide, the same statement recommends averaging two separate draws taken while you are metabolically stable and free of infection.

Assumptions & Limitations

The women's equation was fitted on 24,558 women aged 45 and over in the Women's Health Study; the men's on 10,724 men aged 50 and over in Physicians' Health Study II. Both cohorts were overwhelmingly white US health professionals, so calibration elsewhere is not guaranteed. Diabetic men were excluded from the male derivation entirely, which is why that model carries no HbA1c term and will under-read for them.

Three further cautions. The Reynolds endpoint includes revascularization while the 2013 Pooled Cohort Equations do not, so a Reynolds percentage is systematically more generous than an ASCVD percentage and treatment thresholds published for one do not transfer to the other. Statin therapy suppresses hs-CRP by roughly 15% to 40% independently of LDL, so a treated patient's score reads low. And the score is a primary-prevention tool only: if you have already had a heart attack, stroke, stent or bypass, a ten-year first-event estimate does not apply to you. Discuss any result with a qualified clinician rather than acting on it.

Complete Guide: Reynolds Risk Score Calculator

A Reynolds risk score calculator hangs on a single term the older cardiovascular models never had: 0.180 × ln(hs-CRP) in women, 0.102 × ln(hs-CRP)in men. That one coefficient is the entire argument for the score. Paul Ridker's team added high-sensitivity C-reactive protein and a parent's early heart attack to age, blood pressure and lipids, then asked whether the combination sorted people better than Framingham did.^[1]It did — but the natural log in that term does something most write-ups skip over, and once you see it you will read your own result differently. This guide runs one profile through the equation by hand, re-runs it as a man, and then translates every coefficient into the hs-CRP value it would take to match.

The two Reynolds equations, written out

There is no single Reynolds Risk Score. There are two, derived in two different cohorts, and they do not share a shape. The women's model came out of the Women's Health Study — 24,558 initially healthy women aged 45 and up, followed a median of 10.2 years.^[1]The men's model came a year later from the Physicians' Health Study II, 10,724 men aged 50 and up.^[2] Different populations, separately fitted coefficients, separate baseline survival constants.

Notice what is missing from the men's version. No HbA1c term, because the Physicians' Health Study II derivation excluded diabetic men outright — diabetes was treated as a coronary risk equivalent, so there was nothing to fit. Feed a diabetic man into the male equation and it will quietly under-read. Notice too that age enters linearly for women (0.0799 per year) and logarithmically for men (4.385×ln(age)). These are not stylistic choices; they are what the two datasets supported.

LDL cholesterol is absent from both. Reynolds runs on total cholesterol and HDL, which is why your lipid panel needs only two numbers off it. If you only have LDL and triglycerides on hand, our cholesterol calculator will back out total cholesterol and the TC:HDL ratio for you.

A 60-year-old woman, start to finish

Take the profile the calculator loads by default: a 60-year-old woman, systolic 135 mm Hg, total cholesterol 225 mg/dL, HDL 50 mg/dL, hs-CRP 3.0 mg/L, non-smoker, and a father who had a heart attack at 57. Term by term:

Now the survival step. B − 22.325 = 1.393. exp(1.393) = 4.026. Raise the baseline 0.98634 to that power and you get 0.9461, so the risk is (1 − 0.9461) × 100 = 5.4%. Moderate band. Roughly 5 women in 100 with this exact profile would have a heart attack, ischemic stroke, coronary revascularization or cardiovascular death within ten years.

Look again at the hs-CRP row: 0.198 out of a total of 23.718. Under one percent of B. That is not a rounding artefact, it is the log. Because the term is 0.180×ln(CRP) rather than 0.180×CRP, doubling your CRP from 3 to 6 adds only 0.125 to B. Push it all the way from 0.5 to 10 mg/L — a twentyfold move spanning the entire clinically reported range — and this woman goes from 3.9% to 6.6%. Real, but 2.7 percentage points, not a category-jumping transformation. Anyone who tells you an hs-CRP result will "change everything" has not run the arithmetic.

The same numbers score 16% in a man

Change one input on that profile — sex — and the answer becomes 16.0%. Not 5.4%. A threefold difference from flipping a toggle, because you have not adjusted a coefficient, you have swapped equations entirely. The male baseline survival constant is 0.8990 against the female 0.98634, and that gap alone encodes the much higher underlying event rate in middle-aged men.

This is the practical reason the Reynolds score earned its reputation as a women's tool. Framingham-derived models tended to place almost every woman under 60 in a low-risk bucket where no treatment decision follows, which made them clinically useless for exactly the group clinicians were least sure about. Reynolds did not fix that by inflating women's risk. It fixed it by re-sorting within the low band, so that a woman at 5.4% and a woman at 1.8% stop looking identical. If you want the same profile scored against the 2013 Pooled Cohort Equations for comparison, run it through the ASCVD risk calculator, and against the older model in our Framingham risk calculator.

What hs-CRP would it take to equal smoking?

Here is a translation I have not seen published anywhere, and it is the fastest way to calibrate how much weight to give an inflammation marker. Every flat term in the equation can be converted into the hs-CRP value that would produce the same change in B. Set 0.180×ln(CRP) equal to the coefficient, solve for CRP, and you get a direct exchange rate. Baseline is hs-CRP 1.0 mg/L, where the log term is zero.

Read the smoking row twice. In the women's equation, being a current smoker moves B by as much as an hs-CRP of 94 mg/L would — a value you would only see in sepsis. Real hs-CRP results almost never exceed 10 mg/L in a stable outpatient, and above that the AHA and CDC tell you to discard the result and retest.^[3] So within the range the test can actually produce, smoking outweighs inflammation by roughly an order of magnitude. On the default profile above, ticking the smoker box adds 6.4 percentage points; sweeping hs-CRP across its entire plausible range adds 2.7.

The men's column is stranger still. A parental heart attack before 60 carries a coefficient of 0.541, equivalent to an hs-CRP of 201 mg/L. Family history, not inflammation, is what the male Reynolds model added over Framingham — the CRP term is nearly decorative by comparison. Both factors sit in the score's name-recognition, but only one of them is doing heavy work in men.

Reynolds counts a stent as an event. ASCVD doesn't.

This is the single most consequential thing to understand before comparing your Reynolds number to any other risk percentage, and it is buried in the methods sections. The Reynolds endpoint is a composite of myocardial infarction, ischemic stroke, coronary revascularization and cardiovascular death.^[1][2] The 2013 Pooled Cohort Equations count first hard ASCVD only: non-fatal MI, coronary death, fatal or non-fatal stroke.^[7] Revascularization is not on that list.

Why it matters: a stent or a bypass is a treatment decision made by a cardiologist, not a biological event that happened to an artery. Its rate depends on how aggressively your health system catheterizes people with stable angina. Including it inflates the Reynolds event count relative to ASCVD by a margin that varies with local practice. So a 12% Reynolds score and a 12% ASCVD score are not the same statement, and the Reynolds figure is the more generous of the two. Treatment thresholds published for one model — the 7.5% statin discussion trigger, for instance — were calibrated on that model's endpoint and do not transfer.

Where the 0.017 c-statistic gain actually hides

Discrimination barely improved. The women's Reynolds model posted a c-statistic of 0.808 against 0.791 for the ATP III comparator, a gain of 0.017.^[1] In men it was 0.708 versus 0.700, a gain of 0.008.^[2] If discrimination were the whole story, nobody would have adopted this score.

Reclassification is the story. In women, 40% to 50% of those the older model parked in the intermediate band moved into a higher or lower category once CRP and family history were added.^[1] In men, 17.8% of the whole cohort was reclassified.^[2] Those movements were largely correct, in the sense that reclassified-upward people went on to have more events. A model can sort a population no better on average and still be substantially more useful at the one decision point where the answer is genuinely in doubt.

That is also why the score is only worth running on borderline cases. If your heart disease risk calculator result is already 2% or already 30%, adding hs-CRP will not change what happens next. The band between roughly 5% and 20% is where an extra input can actually flip a decision, and it is where the JUPITER trial found its signal: 17,802 people with LDL under 130 mg/dL but hs-CRP at or above 2.0 mg/L, randomized to rosuvastatin, had a 44% reduction in the primary endpoint.^[4] The 2018 AHA/ACC cholesterol guideline turned that into a rule of thumb, listing hs-CRP of 2.0 mg/L or more as a risk-enhancing factor for people in the borderline and intermediate bands.^[5]

Five things that move hs-CRP and have nothing to do with your arteries

CRP is an acute-phase protein made by the liver in response to interleukin-6. Anything that raises IL-6 raises CRP, and most of those things are not atherosclerosis.

• Statins. They cut hs-CRP by roughly 15% to 40% independently of their LDL effect. If you are already on one, your Reynolds score reads artificially low, and the score was never validated as a treatment-monitoring tool.
• Estrogen. Oral hormone therapy and combined oral contraceptives can double or triple hs-CRP through first-pass hepatic metabolism. Transdermal estrogen largely does not. Same hormone, different route, different CRP.
• Adiposity. Visceral fat is an IL-6 factory. A large share of the CRP–risk association washes out after adjusting for body mass index, which is one reason the marker's independence has been argued over for twenty years.
• Any recent infection or flare. A chest infection, a dental abscess or an active rheumatoid flare pushes hs-CRP into double digits. Values above 10 mg/L are not vascular signal, and the standing advice is to retest at least two weeks later.^[3]
• Ordinary biological variability. Within-person variation in hs-CRP is wide enough that a single draw is a poor estimate. The AHA/CDC statement recommends averaging two measurements taken at least two weeks apart, in a metabolically stable person, free of infection.^[3]

There is a deeper caveat. Mendelian randomization work using CRP gene variants found no association between genetically raised CRP and coronary disease, which argues that CRP is a marker riding along with vascular inflammation rather than a cause of it.^[6]That does not make it useless for prediction — thermometers do not cause fevers and are still worth reading — but it does explain why lowering CRP is not itself a treatment target.

When Reynolds is the wrong instrument

The score has hard boundaries, and pushing past them produces confident-looking numbers that mean nothing:

• Diabetic men. No HbA1c term exists in the male model because diabetic men were excluded from the derivation. Use the Pooled Cohort Equations.
• Anyone with established disease. A prior MI, stroke, stent or bypass puts you in secondary prevention, where a 10-year primary-prevention estimate is meaningless.
• Ages outside 45–80 for women or 50–80 for men. The cohorts did not contain those people. A Reynolds result for a 35-year-old is extrapolation.
• hs-CRP above 10 mg/L. Not a valid input. Treat and retest.
• Familial hypercholesterolemia. Lifelong LDL exposure is not captured by a 10-year model that does not even take LDL as an input.
• Cohorts that look nothing like the derivation samples. The Women's Health Study was overwhelmingly white US health professionals; the Physicians' Health Study II was US male physicians. Calibration in other populations is not guaranteed.

Run the score, then run it again with hs-CRP set to 1.0 mg/L. If your band does not change between those two runs — and for most people it will not — then the inflammation marker was never going to alter your management, and the conversation to have with your clinician is about blood pressure, smoking and lipids instead.

References

1. Ridker PM, Buring JE, Rifai N, Cook NR. Development and validation of improved algorithms for the assessment of global cardiovascular risk in women: the Reynolds Risk Score. JAMA. 2007;297(6):611–619. PubMed
2. Ridker PM, Paynter NP, Rifai N, Gaziano JM, Cook NR. C-reactive protein and parental history improve global cardiovascular risk prediction: the Reynolds Risk Score for men. Circulation. 2008;118(22):2243–2251. PubMed
3. Pearson TA, Mensah GA, Alexander RW, et al. Markers of inflammation and cardiovascular disease: AHA/CDC scientific statement. Circulation. 2003;107(3):499–511. AHA Journals
4. Ridker PM, Danielson E, Fonseca FAH, et al. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein (JUPITER). N Engl J Med. 2008;359(21):2195–2207. NEJM
5. Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC multisociety guideline on the management of blood cholesterol. Circulation. 2019;139(25):e1082–e1143. AHA Journals
6. C Reactive Protein Coronary Heart Disease Genetics Collaboration. Association between C reactive protein and coronary heart disease: Mendelian randomisation analysis. BMJ. 2011;342:d548. BMJ
7. 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–S73. AHA Journals

Frequently Asked Questions