LDL Cholesterol Calculator – the NIH Sampson Equation, Not the 1972 Shortcut
Most lab reports still estimate LDL with the Friedewald formula from 1972. This calculator uses the newer NIH Sampson equation instead — validated against direct measurement and accurate even when triglycerides run high — and shows you both numbers side by side, with exactly where your LDL sits on the standard scale.
Your lab numbers
🔒 Everything is calculated in your browser — your numbers never leave your device.
Your LDL result
| Estimate | Value |
|---|---|
| Sampson (NIH, 2020) — this page’s result | 123.4 mg/dL |
| Friedewald (1972) | 120 mg/dL |
| Difference between the two | +3.4 mg/dL |
| VLDL estimate (triglycerides ÷ 5) | 30 mg/dL |
🩺 A calculated LDL is an estimate, not a direct measurement of your blood. Triglycerides swing with fasting, recent meals and alcohol, and every estimate inherits that swing — if a number surprises you, a repeat test matters more than a second decimal.
- Three numbers off your report and the modern NIH Sampson equation: non-HDL comes first, then the equation models the triglyceride-carrying particles.
- The gap between Sampson and Friedewald grows exactly where the old formula is weakest, and the standard scale puts below 100 where it wants you.
- This is an estimate, not a diagnosis: where no formula is good enough, direct measurement takes over, and change over time only counts between comparable snapshots.
- Where your LDL sits
- Numbers that travel with your LDL
- How this calculator works
- The math behind the default example
- Two things worth knowing about your estimate
- Three panels, three stories the gap tells
- Reading your result against the scale, without a verdict
- When no formula is good enough: direct measurement
- Using the estimate over time, without over-reading it
- What to bring to the clinician, and what not to conclude
Where your LDL sits
The standard categories doctors read a lipid panel against (the NCEP scale). Your marker moves with your numbers — and notice how much of the scale the first band takes up: below 100 is where the scale wants you.
under 100
100–129
130–159
160–189
190+
Your LDL of 123.4 mg/dL lands in Near optimal (100–129) — the second band on the scale.
Two formulas, one blood sample
Sampson (NIH, 2020) — used here
Built by NIH researchers from modern, directly measured data. Instead of assuming your VLDL is always a fixed slice of triglycerides, it models how that share actually changes as triglycerides climb — so it holds up where the old formula falls apart.
LDL = TC÷0.948 − HDL÷0.971 − (TG÷8.56 + TG×non-HDL÷2140 − TG²÷16100) − 9.44
Validated for triglycerides up to 800 mg/dL.
Friedewald (1972) — the classic
The shortcut printed on most lab reports for fifty years. It assumes VLDL cholesterol is always exactly one fifth of triglycerides — a tidy guess that drifts badly once triglycerides pass about 200, and is considered unusable at 400 and above.
LDL = TC − HDL − TG÷5
Simple, familiar — and the reason this page shows both numbers.
Numbers that travel with your LDL
LDL is one panel among several — these read the rest of your metabolic picture.
How this calculator works
Three numbers off your report, one modern equation, and the workings shown — not hidden.
Copy your panel
Total cholesterol, HDL and triglycerides — the three lines every standard lipid panel prints. Use them exactly as reported, in mg/dL or mmol/L; there is no separate “LDL” line to copy, because on most reports that line is itself calculated.
Sampson does the estimating
The NIH equation first takes out HDL (that difference is your non-HDL), then models the cholesterol riding in VLDL particles from your triglycerides — flexibly, the way it actually behaves — instead of assuming it is always a flat fifth.
Read it against the scale
Your result lands on the standard category scale, next to the Friedewald figure your report probably printed. If the two disagree, the gap itself is information — it widens exactly where the old formula is weakest.
The math behind the default example
Step 1 — non-HDL first. Everything that is not HDL: 200 − 50 = 150 mg/dL. The equation needs this before it can model the triglyceride-carrying particles.
Step 2 — scale the cholesterols. 200 ÷ 0.948 = 210.97 and 50 ÷ 0.971 = 51.49. The divisors are the equation’s fitted corrections — they undo small systematic biases in how total and HDL cholesterol are measured.
Step 3 — the triglyceride term. 150 ÷ 8.56 + (150 × 150) ÷ 2140 − 150² ÷ 16100 = 17.52 + 10.51 − 1.40 = 26.64. This is the VLDL estimate — Friedewald would simply have used 150 ÷ 5 = 30. The curved terms are why Sampson stays accurate as triglycerides rise.
Step 4 — the answer. 210.97 − 51.49 − 26.64 − 9.44 = 123.4 mg/dL. Friedewald’s shortcut on the same numbers gives 200 − 50 − 30 = 120 — close here, because at triglycerides of 150 the old assumption still roughly holds. Push triglycerides to 400 and the two methods disagree by more than 8 points.
Two things worth knowing about your estimate
Where the old formula quietly goes wrong — and what the lab does when no formula is good enough.
Friedewald’s whole shortcut rests on one assumption: that the cholesterol in your VLDL particles is always exactly one fifth of your triglycerides (TG ÷ 5). In reality that ratio drifts. When triglycerides are low, VLDL particles are relatively cholesterol-rich; when triglycerides climb, the particles swell with triglyceride and the cholesterol share falls well below a fifth. A fixed divisor cannot follow a moving ratio.
The error is small at everyday values — on this page’s default panel (triglycerides 150) the two formulas differ by 3.4 mg/dL. At triglycerides of 400 with the same total and HDL, Friedewald says 120 while Sampson says 128.4 — and direct measurement sides with Sampson. The gap matters most in exactly the patients doctors worry about: high triglycerides combined with a low LDL, where Friedewald can understate LDL enough to change a treatment decision. Sampson’s equation was published in JAMA Cardiology in 2020 precisely to close that gap, and it was validated against directly measured LDL for triglycerides up to 800 mg/dL.
Every formula on this page is still an estimate. When triglycerides run past roughly 800 mg/dL, when LDL is very low, or when a treatment decision hinges on the exact figure, laboratories can measure LDL directly instead: modern “direct LDL” assays isolate the LDL fraction chemically, and the reference method — beta quantification by ultracentrifuge — physically separates the lipoproteins. Many labs now refuse to report a Friedewald LDL at all when triglycerides are 400 or more, printing “unable to calculate” where the number used to be.
Between the two sits the Martin-Hopkins method (2013): it keeps Friedewald’s simple shape but replaces the fixed divisor 5 with an adjustable factor chosen from your triglyceride and non-HDL levels. It is more accurate than Friedewald and used by many large labs — but head-to-head studies generally find Sampson at least as accurate, especially at high triglycerides, which is why this calculator leads with Sampson and treats Friedewald as the comparison.
Three panels, three stories the gap tells
The gap between Sampson and Friedewald is not noise. It grows exactly where the old formula is weakest, and this calculator’s own engine shows the growth on the default panel as triglycerides rise.
On the default panel, total cholesterol 200, HDL 50 and triglycerides 150, Sampson gives 123.4 mg/dL and Friedewald gives 120 mg/dL, a gap of 3.4 mg/dL. The old formula’s shortcut, triglycerides divided by 5, gives a VLDL estimate of 30 mg/dL, against Sampson’s curved estimate of 26.64 mg/dL, and that 3.36 mg/dL difference in the VLDL term is essentially the whole gap between the two LDL figures. At everyday triglyceride levels the shortcut roughly holds, which is why the two formulas agree closely on ordinary panels and why most reports, which print Friedewald, look right most of the time.
Raise triglycerides to 400 on the same total and HDL, and the picture changes. Sampson gives 85.2 mg/dL and Friedewald gives 70 mg/dL, a gap of 15.2 mg/dL, more than four times the gap at triglycerides of 150. Friedewald’s VLDL shortcut now claims 80 mg/dL of VLDL cholesterol, while Sampson’s curved model, which lets the cholesterol share of VLDL particles fall as triglycerides rise, gives a much smaller figure. The engine on this page also flags the Friedewald figure as unreliable at triglycerides of 400 and above, which is the same threshold at which many laboratories refuse to report a Friedewald LDL at all, printing unable to calculate where the number used to be.
| Panel (TC 200, HDL 50) | Sampson (NIH) | Friedewald | Gap | Non-HDL |
|---|---|---|---|---|
| Triglycerides 150 mg/dL | 123.4 mg/dL | 120 mg/dL | 3.4 mg/dL | 150 mg/dL |
| Triglycerides 400 mg/dL | 85.2 mg/dL | 70 mg/dL | 15.2 mg/dL | 150 mg/dL |
Non-HDL cholesterol does not move in either row, because it is simply total minus HDL, 150 mg/dL in both cases, and it is the figure that stays trustworthy when the LDL estimate itself becomes shaky. That stability is why clinicians watch non-HDL alongside LDL: it counts all of the bad cholesterol together, it needs no triglyceride assumption, and it does not depend on which formula the laboratory used. When the two LDL formulas disagree by 15 points, the non-HDL figure is the anchor that did not move.
Reading your result against the scale, without a verdict
This calculator places your result on the standard category scale. The scale describes the number, not you, and it is not a diagnosis.
On the default panel the Sampson result of 123.4 mg/dL lands in the near optimal band, 100 to 129 mg/dL, and the page shows that band alongside the figure. The total to HDL ratio on the same panel is 4.0, and the page notes that under 3.5 is considered good, so the ratio sits on the less favourable side of that line even while the LDL itself sits in near optimal territory. The two readings do not contradict each other: they measure different things, the LDL concentration and the balance between total and protective cholesterol, and a panel can look reassuring on one and less so on the other. Reading both, rather than cherry-picking the friendlier one, is the point of showing them together.
What the scale cannot do is decide anything about an individual. Categories are population bands, and treatment decisions depend on far more than a single LDL figure: overall risk, other conditions, medication, family history, and a clinician’s judgement. A result in the optimal band is not a clean bill of health, and a result in a higher band is not a diagnosis. This page is an information tool that shows how an estimate is calculated and where it sits on the standard scale. It does not diagnose, and it does not recommend treatment. Anyone whose result concerns them, or whose clinician has set a personal target, should discuss the figure with that clinician, bringing the full panel, not just the LDL line.
Units deserve the same care as categories. Panels are reported in mg/dL in the United States and in mmol/L in much of the rest of the world, and this calculator accepts both, converting between them before applying the equation. Entering a mmol/L figure as if it were mg/dL, or the reverse, produces a nonsense result that still lands somewhere on the scale, because the scale will label any number. Check the unit printed on the report before entering the three figures, and check that the result’s unit matches the report’s unit before comparing the two.
When no formula is good enough: direct measurement
Every formula on this page is an estimate, and there are panels where estimation should stop and measurement should take over.
The first such case is very high triglycerides. Sampson was validated against directly measured LDL for triglycerides up to 800 mg/dL, which is further than Friedewald can credibly go, but beyond roughly that level even Sampson’s curved model is extrapolating past its validation, and laboratories can measure LDL directly instead. Modern direct LDL assays isolate the LDL fraction chemically, and the reference method, beta quantification by ultracentrifuge, physically separates the lipoproteins. Direct measurement costs more and is not routine, which is why estimation remains the default, but it is the right step when a treatment decision hinges on the exact figure and no formula can be trusted to give it.
The second case is very low LDL, where small absolute errors become large relative errors, and where the difference between formulas can be as large as the figure itself. The third case is any panel where the laboratory itself refused to calculate: many labs now print unable to calculate for a Friedewald LDL when triglycerides are 400 or more, and that refusal is information, not an omission. It means the laboratory judged its own formula unreliable on that panel, and the correct response is a better method, Sampson or direct measurement, not a hand calculation with the same discredited divisor.
Between estimation and direct measurement sits the Martin-Hopkins method, published in 2013. It keeps Friedewald’s simple shape but replaces the fixed divisor 5 with an adjustable factor chosen from the triglyceride and non-HDL levels, and it is more accurate than Friedewald and used by many large laboratories. Head-to-head studies generally find Sampson at least as accurate, especially at high triglycerides, which is why this calculator leads with Sampson and shows Friedewald for comparison. The existence of three methods for one number is not a scandal. It is a reminder that LDL on most reports was never measured at all, and that the method matters most exactly where the number matters most.
Using the estimate over time, without over-reading it
A single panel is a snapshot. Change over time is the signal, but only if the snapshots are comparable.
Comparability starts with the method. A Friedewald LDL from last year’s report and a Sampson LDL from this calculator are not directly comparable, because the two formulas differ by an amount that grows with triglycerides: 3.4 mg/dL on the default panel, 15.2 mg/dL at triglycerides of 400. A panel whose triglycerides fell from 400 to 150 would show a Friedewald-to-Sampson change that mixes a real change in cholesterol with a change in method, and the method part could be larger than the real part. Compare Sampson with Sampson, Friedewald with Friedewald, and non-HDL with non-HDL, which needs no formula at all and is therefore the most comparable figure across years and laboratories.
Fasting is the second comparability condition. Triglycerides rise after eating, and because both formulas use triglycerides, a non-fasting panel will produce a different LDL estimate from a fasting panel on the same person, even if the underlying cholesterol has not changed. Whether fasting is required depends on the laboratory and the purpose of the test, and the report itself usually states the condition. Whatever the condition was, keep it the same across comparisons, or note the change when reading the trend.
Finally, treat small movements as noise. Laboratory measurement varies, triglycerides vary day to day, and the formulas amplify triglyceride variation into LDL variation. A move of a few mg/dL between panels, in either direction, is within the noise of the whole chain, and chasing it with changes in behaviour or treatment is over-reading. The movements worth attention are large, sustained, and visible in non-HDL as well as in LDL, because a change that appears in the formula-free figure as well as in the estimated one is a change in the cholesterol itself, not in the arithmetic.
What to bring to the clinician, and what not to conclude
This calculator’s output is most useful when it arrives at a consultation complete, with its method and its companions, rather than as a single number.
Bring the full panel, not just the LDL line: total cholesterol, HDL and triglycerides, the unit they were reported in, whether the sample was fasting, and the method the laboratory used, if the report states one. Bring the Sampson and Friedewald figures together, with the gap between them, because the gap itself is information about how much the estimate can be trusted on that panel. And bring the non-HDL figure, total minus HDL, because it needs no formula and is therefore the most comparable figure across years and laboratories. A clinician given all of that can see at a glance whether a change between panels is a change in cholesterol, a change in method, or a change in fasting state, and can judge the result against the patient’s overall risk rather than against a single band on a scale.
What not to conclude matters as much. Do not conclude that an optimal-band LDL is a clean bill of health, because the scale describes the number, not the person, and treatment decisions depend on far more than one figure. Do not conclude that a higher-band result is a diagnosis, for the same reason. Do not adjust medication, start or stop a treatment, or change a dose on the basis of a calculator result, because this page is an information tool that shows how an estimate is calculated, not a source of medical advice. And do not chase small movements between panels: laboratory measurement varies, triglycerides vary day to day, and a move of a few mg/dL is within the noise of the whole chain. The movements worth attention are large, sustained, and visible in non-HDL as well as in LDL.
Friedewald: three numbers in, one number out
Most LDL values on lab reports are calculated, not measured. The Friedewald equation — LDL = total cholesterol − HDL − (triglycerides ÷ 5), all in mg/dL — assumes VLDL cholesterol is roughly a fifth of triglycerides, and it is accurate enough for most people most of the time. Example: total 220, HDL 50, triglycerides 150 gives LDL 140 mg/dL, borderline high. The two errors to avoid are unit confusion (in mmol/L the divisor is 2.2, not 5) and trusting the formula when triglycerides exceed 400 mg/dL, where laboratories measure LDL directly instead.
Ranges are guides; targets are personal
Below 100 mg/dL is optimal in population terms, 130–159 borderline high, 190+ very high — but a personal target follows overall cardiovascular risk, not the LDL number alone. Someone with established heart disease may target below 70 or even 55 mg/dL, while a young person with no risk factors at 135 may simply be monitored. Non-HDL cholesterol (total minus HDL) is a useful secondary check when LDL and total seem to disagree. Any flagged value deserves a conversation with a clinician about the whole risk picture, not a verdict from a table.
LDL questions, answered
What the number is, when to trust it, and when to ask for a better measurement.
Why is the Sampson equation more accurate than Friedewald?
Friedewald (1972) assumes the cholesterol carried in VLDL particles is always exactly one fifth of your triglycerides. That ratio actually changes as triglycerides rise, so the old formula drifts — worst at high triglycerides and low LDL, where it can understate LDL by a clinically important margin. The Sampson equation (NIH, published in JAMA Cardiology in 2020) models that VLDL share with curved terms fitted to tens of thousands of directly measured lipid profiles, and it stays accurate for triglycerides up to 800 mg/dL — double Friedewald’s usable range.
What happens if my triglycerides are over 400?
Above 400 mg/dL the Friedewald formula is considered unreliable — many labs simply print “unable to calculate” instead of an LDL. This calculator flags the Friedewald figure with a warning badge in that range, but the Sampson result remains usable: it was validated up to triglycerides of 800. Beyond 800, treat any calculated figure as rough (this page says so when it happens) and ask your clinician about a direct LDL measurement instead.
Do I need to fast before a cholesterol test?
It depends on the purpose. Triglycerides rise for hours after a meal — especially a fatty one, or after alcohol — and because calculated LDL is built from triglycerides, a non-fasting sample can pull the estimate down. Many guidelines now accept non-fasting tests for general screening, but a fasting sample (usually 9–12 hours, water allowed) is still standard when triglycerides have been high, when results will guide treatment, or when your clinician asks for one. Follow the instruction on your own test order.
What is non-HDL cholesterol, and why do doctors watch it?
Non-HDL is simply total cholesterol minus HDL — everything left over: LDL, VLDL and the remnant particles, all of which can deposit cholesterol in artery walls. Because it captures all of those particles in one number, many clinicians treat it as a useful second target, especially when triglycerides are high and LDL is hard to estimate. A common benchmark is non-HDL under 130 mg/dL — roughly 30 points above the equivalent LDL goal.
Calculated or directly measured LDL — which should I trust?
For most routine panels, a calculated LDL — especially by Sampson — is accurate enough, and it is what nearly every report uses. A direct measurement is worth asking about when triglycerides are very high (over 400–800), when LDL is very low, when two results disagree sharply, or when the exact figure would change a treatment decision. If your report’s LDL was calculated with Friedewald and your triglycerides were high, re-running the same panel through this page’s Sampson equation is a reasonable first check before repeating the blood draw.
What LDL number should I be aiming for?
There is no single target that fits everyone. The categories on this page (optimal under 100, very high at 190+) are population bands from the NCEP scale, not personal goals. Actual targets are set by your overall risk: someone who has already had a heart attack or stroke is often treated to LDL under 70, while a younger person with no other risk factors may simply be advised on lifestyle first. Only a clinician who knows your full history can set your number — treat this page’s scale as context for that conversation, not a substitute for it.
Do I need to fast before a cholesterol test?
It depends on the laboratory and the purpose of the test, and the report itself usually states the condition. Triglycerides rise after eating, and because both the Sampson and Friedewald formulas use triglycerides, a non-fasting panel will produce a different LDL estimate from a fasting panel on the same person, even if the underlying cholesterol has not changed. For comparing panels over time, the important thing is consistency: keep the fasting condition the same across comparisons, or note the change when reading the trend. If the report does not state the condition, ask the laboratory before comparing its figures with an earlier panel.
Why does my report’s LDL differ from this calculator?
Almost always because the report used a different method. Most reports print a Friedewald LDL, while this calculator leads with Sampson, and the two differ by 3.4 mg/dL on the default panel and by 15.2 mg/dL at triglycerides of 400. Check which method the report used, if it states one, and compare like with like: this page shows both figures side by side precisely so the difference is visible rather than hidden. Units are the other common cause: a figure entered in mmol/L as if it were mg/dL, or the reverse, produces a nonsense result that still lands somewhere on the scale. Check the unit printed on the report before entering the three figures.
What is non-HDL cholesterol, and why does it stay steady?
Non-HDL cholesterol is total cholesterol minus HDL cholesterol: everything that is not protective, counted together. On the default panel it is 200 minus 50, or 150 mg/dL, and the page notes that under 130 is desirable. It stays steady when the LDL formulas disagree because it needs no triglyceride assumption and no formula at all: it is a subtraction, not an estimate. That is why clinicians watch it alongside LDL, and why it is the most comparable figure across years and laboratories. When Sampson and Friedewald disagree by 15 points, the non-HDL figure is the anchor that did not move, and a change that appears in non-HDL as well as in LDL is a change in the cholesterol itself, not in the arithmetic.
What LDL number should I be aiming for?
This page does not set a personal target, and no single number on it should be read as one. The category scale describes the number: on the default panel, 123.4 mg/dL lands in the near optimal band of 100 to 129 mg/dL, and the total to HDL ratio of 4.0 sits on the less favourable side of the under-3.5 line the page notes. But a personal target depends on overall risk, other conditions, medication, family history, and a clinician’s judgement, and it differs between people with the same LDL figure. Anyone whose result concerns them, or whose clinician has set a personal target, should discuss the figure with that clinician, bringing the full panel, not just the LDL line. This calculator is an information tool, not a source of medical advice.
How is LDL calculated?
Usually by the Friedewald formula: total cholesterol − HDL − (triglycerides ÷ 5) in mg/dL. Example: 220 − 50 − 30 = 140 mg/dL.
What LDL level is optimal?
Below 100 mg/dL is optimal as a population guide, but personal targets depend on overall cardiovascular risk and may be below 70 or 55 mg/dL for high-risk people.
When is calculated LDL unreliable?
When triglycerides exceed 400 mg/dL, after eating (non-fasting samples), or in certain metabolic conditions — then direct measurement or the Martin-Hopkins equation is used.
How do I convert mg/dL to mmol/L?
Divide by 38.67: 100 mg/dL ≈ 2.59 mmol/L. Remember the Friedewald divisor becomes 2.2 in mmol/L units.
More health calculators
The rest of the dashboard: weight, energy and daily burn.
BMI Calculator
Your body mass index and its healthy range.
📏Body Fat Calculator
Body fat percentage from tape measurements.
🔥BMR Calculator
Basal metabolic rate — calories burned at rest.
⚡TDEE Calculator
Total daily energy expenditure, explained.
🍽️Calorie Calculator
Daily calorie targets for your goal.
📖LDL Cholesterol Explained
Normal range, Friedewald formula, and when it fails.