Free clinical tool
HOMA-IR Calculator
Your HOMA-IR insulin resistance index and HOMA-B beta cell function from one fasting glucose and one fasting insulin — in mg/dL or mmol/L, µIU/mL or pmol/L.
- Reviewed by
- Alpha Health Finder Editorial Team
- Last reviewed
What HOMA-IR measures, and what it cannot
Insulin resistance means your tissues respond poorly to insulin, so the pancreas has to secrete more of it to keep fasting glucose where it belongs. The direct way to measure that is a hyperinsulinaemic-euglycaemic clamp: an insulin infusion, a glucose infusion adjusted minute by minute for several hours, and repeated blood sampling throughout. It is the reference standard and it is completely impractical outside a research unit.
HOMA is the shortcut. Matthews and colleagues published it in 1985 as a computer-solved model of the feedback loop between fasting glucose and fasting insulin: if the tissues are resistant, insulin will be high for whatever glucose you see, and the relationship between the two can be read backwards to estimate how resistant they are. Their estimate correlated with the euglycaemic clamp at Rs = 0.88. Alongside it the same model estimates beta-cell function — HOMA-B, sometimes written HOMA-%B — which asks the opposite question: for the insulin being secreted, how well is the glucose actually being held down.
The two indices move in opposite directions with glucose, and that is the reason to read them together. Rising insulin raises both. Rising fasting glucose raises HOMA-IR and lowers HOMA-B, because needing more insulin to hold a higher glucose is a sign the beta cells are losing ground rather than gaining it. A high HOMA-IR with a still-high HOMA-B is a different picture from a high HOMA-IR with a falling one.
What HOMA cannot do is diagnose anything. Its precision is poor by the authors' own account — the 1985 paper reports coefficients of variation of 31% for the insulin resistance estimate and 32% for the beta-cell one. Fasting insulin is not a standardised assay, and different immunoassays return results 20-40% apart on the same blood sample, which moves HOMA-IR by the same 20-40% before any biology is involved. And there is no agreed cut-off: the four published figures in the reference chart below range from 1.85 to 3.46, three of them from the same study. HOMA-IR is most useful as a trend in one person measured by one laboratory, and least useful as a single number compared against somebody else's threshold.
One more caveat worth knowing before you compare notes with anyone. What this page computes is HOMA1 — the closed-form approximation from the 1985 paper, which is what almost every published study and almost every calculator means by "HOMA-IR". HOMA2 is a later, revised, non-linear computer model. It is not an equation, so there is nothing to implement: it exists only as the Oxford HOMA2 Calculator. If a clinician has quoted you a HOMA2 figure, expect it to differ from the number here, particularly at the extremes.
The formula
HOMA-IR = fasting glucose (mmol/L) × fasting insulin (µIU/mL) / 22.5
= fasting glucose (mg/dL) × fasting insulin (µIU/mL) / 405
HOMA-B = 20 × fasting insulin (µIU/mL)
────────────────────────────────
fasting glucose (mmol/L) − 3.5 (as a percentage)
Units insulin pmol/L = µIU/mL × 6.00
glucose mmol/L = mg/dL × 0.0555
22.5 is a normalising constant, not an empirical fit: it is 5 µIU/mL of insulin multiplied by 4.5 mmol/L of glucose — the fasting values of an "ideal" insulin-sensitive person — so that person scores exactly 1.0. 405 is the same constant restated for mg/dL (22.5 × 18). HOMA-B has no mg/dL form at all; the 3.5 in its denominator is a glucose in mmol/L, which is why a mg/dL glucose must be converted before it can be used. Below 3.5 mmol/L (63.1 mg/dL) that denominator is zero or negative and HOMA-B has no meaning, so this calculator returns nothing rather than an infinity.
Worked example
Fasting glucose 100 mg/dL, fasting insulin 12 µIU/mL — a glucose most people would be told is fine, with an insulin nobody usually looks at:
- Convert the glucose: 100 mg/dL × 0.0555 = 5.55 mmol/L
- Convert the insulin for reference: 12 µIU/mL × 6.00 = 72 pmol/L
- HOMA-IR = 5.55 × 12 / 22.5 = 2.96
- Same answer the mg/dL way: 100 × 12 / 405 = 2.96
- HOMA-B = 20 × 12 / (5.55 − 3.5) = 240 / 2.05 = 117%
- A HOMA-IR of 2.96 sits above three of the four published reference figures below and beneath the fourth — which is exactly why this page prints all four rather than a verdict. The glucose alone said nothing; the insulin was doing the work.
HOMA-IR chart — fasting glucose against fasting insulin
Find your fasting glucose down the left and your fasting insulin across the top. The cell is your HOMA-IR. Both unit systems are given on every row and column.
| Fasting glucose | 3 µIU/mL18 pmol/L | 5 µIU/mL30 pmol/L | 8 µIU/mL48 pmol/L | 12 µIU/mL72 pmol/L | 20 µIU/mL120 pmol/L | 30 µIU/mL180 pmol/L |
|---|---|---|---|---|---|---|
| 70 mg/dL3.89 mmol/L | 0.52 | 0.86 | 1.38 | 2.07 | 3.45 | 5.18 |
| 80 mg/dL4.44 mmol/L | 0.59 | 0.99 | 1.58 | 2.37 | 3.95 | 5.92 |
| 85 mg/dL4.72 mmol/L | 0.63 | 1.05 | 1.68 | 2.52 | 4.19 | 6.29 |
| 90 mg/dL5.00 mmol/L | 0.67 | 1.11 | 1.78 | 2.66 | 4.44 | 6.66 |
| 95 mg/dL5.27 mmol/L | 0.70 | 1.17 | 1.87 | 2.81 | 4.69 | 7.03 |
| 100 mg/dL5.55 mmol/L | 0.74 | 1.23 | 1.97 | 2.96 | 4.93 | 7.40 |
| 110 mg/dL6.11 mmol/L | 0.81 | 1.36 | 2.17 | 3.26 | 5.43 | 8.14 |
| 120 mg/dL6.66 mmol/L | 0.89 | 1.48 | 2.37 | 3.55 | 5.92 | 8.88 |
| 126 mg/dL6.99 mmol/L | 0.93 | 1.55 | 2.49 | 3.73 | 6.22 | 9.33 |
| 140 mg/dL7.77 mmol/L | 1.04 | 1.73 | 2.76 | 4.14 | 6.91 | 10.36 |
| 160 mg/dL8.88 mmol/L | 1.18 | 1.97 | 3.16 | 4.74 | 7.89 | 11.84 |
| 180 mg/dL9.99 mmol/L | 1.33 | 2.22 | 3.55 | 5.33 | 8.88 | 13.32 |
HOMA-IR is linear in both inputs: double the insulin and the index doubles, double the glucose and it doubles again. That is why fasting insulin dominates the table — glucose varies over a narrow range in most people while insulin varies over a tenfold one. A perfectly insulin-sensitive person (glucose 4.5 mmol/L, insulin 5 µIU/mL) scores exactly 1.00 by construction. Every figure here is computed by the same function the calculator above uses.
HOMA-B chart — beta cell function by fasting glucose and insulin
The same two values read the other way: how much insulin is being secreted for the glucose it is holding. Note that HOMA-B falls as glucose rises, the opposite of HOMA-IR.
| Fasting glucose | 3 µIU/mL18 pmol/L | 5 µIU/mL30 pmol/L | 8 µIU/mL48 pmol/L | 12 µIU/mL72 pmol/L | 20 µIU/mL120 pmol/L | 30 µIU/mL180 pmol/L |
|---|---|---|---|---|---|---|
| 70 mg/dL3.89 mmol/L | 156% | 259% | 415% | 623% | 1038% | 1556% |
| 80 mg/dL4.44 mmol/L | 64% | 106% | 170% | 255% | 425% | 638% |
| 85 mg/dL4.72 mmol/L | 49% | 82% | 131% | 197% | 328% | 493% |
| 90 mg/dL5.00 mmol/L | 40% | 67% | 107% | 160% | 267% | 401% |
| 95 mg/dL5.27 mmol/L | 34% | 56% | 90% | 135% | 226% | 338% |
| 100 mg/dL5.55 mmol/L | 29% | 49% | 78% | 117% | 195% | 293% |
| 110 mg/dL6.11 mmol/L | 23% | 38% | 61% | 92% | 154% | 230% |
| 120 mg/dL6.66 mmol/L | 19% | 32% | 51% | 76% | 127% | 190% |
| 126 mg/dL6.99 mmol/L | 17% | 29% | 46% | 69% | 114% | 172% |
| 140 mg/dL7.77 mmol/L | 14% | 23% | 37% | 56% | 94% | 140% |
| 160 mg/dL8.88 mmol/L | 11% | 19% | 30% | 45% | 74% | 111% |
| 180 mg/dL9.99 mmol/L | 9% | 15% | 25% | 37% | 62% | 92% |
HOMA-B is expressed as a percentage of a reference beta-cell function, so 100% is the model's normal rather than a maximum — figures well above 100% are ordinary in insulin-resistant people whose pancreas is still compensating. The column of very large numbers at the top of the table is the equation's weak point, not a finding: at a fasting glucose of 70 mg/dL the denominator (glucose − 3.5) is only 0.39 mmol/L, so a trivial measurement error swings the answer enormously. Wallace and colleagues list measuring beta-cell function in isolation as an inappropriate use of HOMA; read this table next to the HOMA-IR one, never on its own.
Published HOMA-IR reference figures — and why they disagree
Four cut-offs from the peer-reviewed literature, lowest to highest. Three of them come from the same 2,459-person study and differ only in how the threshold was chosen.
| HOMA-IR figure | Population | How it was chosen | Source |
|---|---|---|---|
| 1.85 | Non-diabetic men, Spanish general adult population (EPIRCE, n=2,459) | Cut-off best separating men with and without metabolic syndrome (ATP III), by Youden index | Gayoso-Diz 2013 |
| 2.05 | Spanish general adult population (EPIRCE, n=2,459) | Threshold once metabolic syndrome components are taken into account rather than a percentile | Gayoso-Diz 2013 |
| 2.81 | Men in a large Brazilian clinical laboratory database | Upper limit of the 95% reference interval (interval 0.38-2.81) | Brazilian reference-interval proposal 2024 |
| 3.46 | Spanish general adult population (EPIRCE, n=2,459) | 90th percentile of the observed distribution | Gayoso-Diz 2013 |
There is no universally agreed HOMA-IR cut-off, and this table is here to show that rather than to hide it — the highest figure is 1.9 times the lowest. Two reasons it will not settle. Fasting insulin immunoassays return results 20-40% apart on the same blood sample, and the distribution genuinely differs between populations and ethnicities. No major endocrine society currently recommends HOMA-IR as a standardised clinical screening test. Treat these as landmarks for a conversation with your clinician, not as a diagnosis.
Fasting insulin unit conversion chart — µIU/mL to pmol/L
Insulin is reported in µIU/mL on US labs and pmol/L on many European ones. Multiply by 6.00 to go one way, divide to go the other. The last two columns show what each value does to the indices at a fixed fasting glucose of 90 mg/dL.
| Fasting insulin | pmol/L | HOMA-IRglucose 90 mg/dL | HOMA-Bglucose 90 mg/dL |
|---|---|---|---|
| 2 µIU/mL | 12 | 0.44 | 27% |
| 3 µIU/mL | 18 | 0.67 | 40% |
| 5 µIU/mL | 30 | 1.11 | 67% |
| 7 µIU/mL | 42 | 1.55 | 94% |
| 10 µIU/mL | 60 | 2.22 | 134% |
| 12 µIU/mL | 72 | 2.66 | 160% |
| 15 µIU/mL | 90 | 3.33 | 201% |
| 20 µIU/mL | 120 | 4.44 | 267% |
| 25 µIU/mL | 150 | 5.55 | 334% |
| 30 µIU/mL | 180 | 6.66 | 401% |
| 40 µIU/mL | 240 | 8.88 | 535% |
| 60 µIU/mL | 360 | 13.32 | 802% |
1 µIU/mL = 6.00 pmol/L, the factor attributed to the WHO and the American Diabetes Association. Entering a pmol/L insulin into a calculator expecting µIU/mL inflates HOMA-IR six-fold, which is the single commonest way this index gets misread — a perfectly ordinary 60 pmol/L becomes a HOMA-IR of 13.32 instead of 2.22. The calculator above takes either unit, so the mistake cannot be made.
Fasting glucose unit conversion chart — mg/dL to mmol/L
Glucose is reported in mg/dL in the US and mmol/L almost everywhere else. The last two columns show each value's effect on the indices at a fixed fasting insulin of 10 µIU/mL.
| Fasting glucose | mmol/L | HOMA-IRinsulin 10 µIU/mL | HOMA-Binsulin 10 µIU/mL |
|---|---|---|---|
| 63 mg/dL | 3.50 | 1.55 | not defined |
| 70 mg/dL | 3.89 | 1.73 | 519% |
| 80 mg/dL | 4.44 | 1.97 | 213% |
| 90 mg/dL | 5.00 | 2.22 | 134% |
| 100 mg/dL | 5.55 | 2.47 | 98% |
| 110 mg/dL | 6.11 | 2.71 | 77% |
| 126 mg/dL | 6.99 | 3.11 | 57% |
| 140 mg/dL | 7.77 | 3.45 | 47% |
| 160 mg/dL | 8.88 | 3.95 | 37% |
| 180 mg/dL | 9.99 | 4.44 | 31% |
| 200 mg/dL | 11.10 | 4.93 | 26% |
| 250 mg/dL | 13.88 | 6.17 | 19% |
Glucose mg/dL × 0.0555 = mmol/L, derived from the molecular weight of glucose (180.156 g/mol). The 63 mg/dL row is there deliberately: at 3.497 mmol/L it falls just under the 3.5 mmol/L floor in the HOMA-B denominator, so HOMA-B has no value there while HOMA-IR is still perfectly well defined. Notice also how much less the glucose column moves HOMA-IR than the insulin column in the chart above — across this entire range HOMA-IR only rises from 1.55 to 6.17.
Frequently asked questions
- What is the HOMA-IR formula?
- HOMA-IR = fasting glucose (mmol/L) × fasting insulin (µIU/mL) ÷ 22.5, or equivalently fasting glucose (mg/dL) × fasting insulin (µIU/mL) ÷ 405. The 22.5 is not an empirical fit — it is 5 µIU/mL multiplied by 4.5 mmol/L, the fasting values of a notionally ideal insulin-sensitive person, so that person scores exactly 1.0. Both forms come from Matthews and colleagues, Diabetologia 1985.
- What is a normal or good HOMA-IR?
- There is no universally agreed cut-off, and anyone quoting you a single number is overstating the evidence. Published figures include 1.85 for non-diabetic men and 2.05 for adults generally when chosen against metabolic syndrome, 2.81 as the upper limit of a laboratory reference interval in men, and 3.46 as a simple 90th percentile — three of those four from the same 2,459-person study. Fasting insulin assays also differ by 20–40% on one sample, so the trend in your own results from one laboratory is worth far more than any threshold.
- How do I calculate HOMA-IR with beta cell function?
- Beta cell function uses the same two blood values: HOMA-B = (20 × fasting insulin in µIU/mL) ÷ (fasting glucose in mmol/L − 3.5), reported as a percentage. At a fasting glucose of 100 mg/dL (5.55 mmol/L) with insulin 12 µIU/mL, that is 240 ÷ 2.05 = 117%. There is no mg/dL version of this equation, so a US glucose has to be converted first — multiply by 0.0555. The calculator above returns both indices at once.
- How do I convert fasting insulin from pmol/L to µIU/mL?
- Divide pmol/L by 6.00, the conversion factor attributed to the WHO and the American Diabetes Association — so 60 pmol/L is 10 µIU/mL and 90 pmol/L is 15 µIU/mL. Going the other way, multiply µIU/mL by 6.00. Getting this backwards is the commonest HOMA-IR error there is: entering 60 pmol/L into a calculator expecting µIU/mL turns a HOMA-IR of 1.23 into 7.40 at the same glucose.
- Can HOMA-IR be high when my fasting glucose is normal?
- Yes, and that is much of the point of running it. Fasting glucose is the last thing to move, because the pancreas compensates by secreting more insulin for years before glucose drifts. A glucose of 90 mg/dL with insulin 20 µIU/mL gives HOMA-IR 4.44, while the same 90 mg/dL with insulin 5 gives 1.11 — identical glucose, a fourfold difference in the index. This is why fasting insulin is worth ordering alongside glucose rather than instead of it.
- What is the difference between HOMA1 and HOMA2?
- HOMA1 is the closed-form approximation published in 1985 and is what virtually every study, every lab report and this calculator mean by "HOMA-IR". HOMA2 is a later revised non-linear computer model that accounts for effects the simple equation cannot, and it is not an equation at all — it exists only as the Oxford HOMA2 Calculator software, so no web page can reproduce it with arithmetic. The two agree reasonably in the middle of the range and diverge at the extremes, so a HOMA2 figure from your clinician will not match the number here.
- Why does HOMA-B go down when my fasting glucose goes up?
- Because HOMA-B asks how much insulin is being secreted relative to the glucose it is failing to hold down. Needing 10 µIU/mL to sit at 126 mg/dL is worse beta-cell performance than needing 10 µIU/mL to sit at 85 mg/dL, so the index falls as glucose climbs — the mirror image of HOMA-IR, which rises. That inverse relationship is why the two indices are meant to be read together; Wallace and colleagues specifically list measuring beta-cell function in isolation as an inappropriate use of HOMA.
- Does HOMA-IR need a fasting blood sample?
- Yes — both glucose and insulin must come from the same fasting draw, conventionally after 8 to 12 hours without food, because the entire model is built on the steady-state relationship between the two at rest. A post-meal insulin can be ten times the fasting value, so feeding a non-fasting sample into the equation produces a number that means nothing at all. If your report does not say fasting, the HOMA-IR calculated from it should be discarded rather than interpreted.
- How accurate is HOMA-IR compared with a glucose clamp?
- The 1985 validation found the HOMA estimate of insulin resistance correlated with the euglycaemic clamp at Rs = 0.88, which is strong for a two-value shortcut against a multi-hour reference procedure. But precision at the level of one person is poor: the same paper reports coefficients of variation of 31% for insulin resistance and 32% for the beta-cell estimate. That combination — good correlation across a group, wide scatter within an individual — is why HOMA is well suited to cohort studies and poorly suited to labelling one person.
Sources
- [1]Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–419
- [2]Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004;27(6):1487–1495 — on HOMA1 versus the HOMA2 computer model, and on the inappropriate uses of the index
- [3]Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21(12):2191–2192 — the introduction of HOMA2
- [4]Gayoso-Diz P, Otero-González A, Rodriguez-Alvarez MX, et al. Insulin resistance (HOMA-IR) cut-off values and the metabolic syndrome in a general adult population: effect of gender and age. BMC Endocr Disord. 2013;13:47
- [5]Proposal for fasting insulin and HOMA-IR reference intervals based on an extensive Brazilian laboratory database — source of the 1 µIU/mL = 6.00 pmol/L conversion and the reference interval in men
- [6]The case for broader use of HOMA-IR in metabolic assessment: a narrative review — source of the mg/dL (÷405) form and of the 20–40% between-assay variation in fasting insulin
Embed this calculator
Free to use on your own site, no permission needed and no account. The widget resizes itself to fit your layout. Keeping the credit line underneath is appreciated, but the calculator works exactly the same if you delete it.
<!-- HOMA-IR Calculator — by Alpha Health Finder. Free to embed; keeping the credit line below is appreciated but not required. -->
<iframe src="https://alphahealthfinder.com/embed/tools/homa-ir-calculator"
title="HOMA-IR Calculator"
width="100%" height="2080" loading="lazy"
style="border:1px solid #e2e8f0;border-radius:12px;max-width:680px;width:100%"
data-ahf-embed="homa-ir-calculator"></iframe>
<p style="max-width:680px;margin:8px 0 0;font:13px/1.5 system-ui,-apple-system,sans-serif;color:#64748b">
<a href="https://alphahealthfinder.com/tools/homa-ir-calculator">HOMA-IR Calculator</a> by <a href="https://alphahealthfinder.com">Alpha Health Finder</a>
</p>
<script src="https://alphahealthfinder.com/embed/resize.js" async></script>Next steps
Related calculators
- LDL Cholesterol CalculatorLDL-C by all three equations — Friedewald, Martin-Hopkins and Sampson — side by side, in mg/dL or mmol/L.
- A1c to Average Glucose ConverterHbA1c to estimated average glucose in mg/dL or mmol/L, plus NGSP % to IFCC mmol/mol — with the real spread around each estimate.
- Testosterone Reference Range by AgeYour total testosterone as a percentile among men your own age, from the published Travison centiles — plus what that does and does not mean.
- eGFR CalculatorEstimated GFR by the 2021 CKD-EPI equations, which carry no race coefficient, in mg/dL or µmol/L, with cystatin C confirmation.
- Cholesterol Unit ConverterA whole lipid panel between mg/dL and mmol/L, with the triglyceride factor kept separate from the cholesterol one — the commonest error on a lipid report.
- Free Testosterone CalculatorFree and bioavailable testosterone from total T, SHBG and albumin. Vermeulen method, with the full conversion chart.
Educational reference, not medical advice. HOMA-IR and HOMA-B are calculated indices, not diagnoses, and not measurements of insulin resistance itself. Their precision in one person is poor — the original paper reports coefficients of variation above 30% — and fasting insulin assays differ by 20–40% between laboratories. There is no universally agreed HOMA-IR cut-off. Interpretation, and any decision about treatment, belongs to your clinician.