Metabolic health
What Is HOMA-IR, and How Do You Calculate It?
A HOMA-IR score can read 'optimal' on one calculator and 'resistant' on another from the exact same blood draw. The formula isn't the confusing part — the missing agreement on what the number means is.
The short answer
HOMA-IR is not a direct measurement of insulin resistance but a calculation — (fasting insulin in µU/mL × fasting glucose in mg/dL) ÷ 405 — built from a mathematical model of how the liver and pancreas interact. There is no single official 'normal' cutoff: research commonly reads scores under about 2 as insulin-sensitive and above roughly 2.5 as resistant, but that line shifts with the insulin assay used and the population studied, and clinicians rarely order it to diagnose anything.
What this actually measures
Nothing, by itself. HOMA-IR is arithmetic performed on two things that are measured directly — fasting blood glucose and fasting blood insulin — run through a formula derived from a mathematical model of the feedback loop between the liver and the pancreas, rather than a direct physical measurement of insulin resistance.
| Band | Value | What it depends on |
|---|---|---|
| Commonly read as insulin-sensitive | below ~2.0 | A widely used research convention, not an official cutoff; shifts with the insulin assay and the population used to derive it. |
| Commonly read as insulin-resistant | ~2.5 and above | Frequent in NHANES-based U.S. analyses, but the equivalent threshold in the EPIRCE Spanish cohort ranged from 1.85 to 3.46 depending on sex, age and which criteria were used. |
| Diagnostic status | not a diagnosis on its own | NIDDK states insulin-resistance testing is used mainly in research; prediabetes and diabetes are diagnosed via A1C, fasting glucose or an oral glucose tolerance test instead. |
Run the same fasting blood draw through three different online HOMA-IR calculators and you can get three different verdicts — “optimal,” “borderline,” and “insulin resistant” — out of the same two numbers. That isn’t a website bug. It reflects something true about the score itself: HOMA-IR was never built to hand back one agreed threshold, and most of the confusion around it comes from treating a research index as if it were a lab-defined cutoff.
What the calculation actually is
HOMA-IR stands for Homeostatic Model Assessment of Insulin Resistance. It isn’t something a machine measures directly, the way a thermometer measures temperature. It’s arithmetic performed on two things that are measured directly — your fasting blood glucose and your fasting blood insulin — run through a formula built from a mathematical model of how the liver and pancreas communicate.
The model dates to 1985, when researchers Matthews and colleagues published equations describing glucose regulation as a feedback loop, then showed that a simplified version of those equations, using only a fasting glucose and a fasting insulin value, tracked reasonably with more elaborate laboratory measures of insulin sensitivity. HOMA-IR is the simplified version that stuck, mostly because it needs one blood draw instead of an hours-long clamp procedure.
The formula, worked through
The standard version, using the units most U.S. labs report, is:
HOMA-IR = (fasting insulin in µU/mL × fasting glucose in mg/dL) ÷ 405
If your lab reports glucose in mmol/L instead of mg/dL, the divisor changes to 22.5. Mixing the two — a mg/dL glucose value run against the mmol/L constant, or the reverse — is the most common calculation error, and it doesn’t produce an obviously broken number. It produces a wrong one that still looks plausible.
Worked through with round numbers: a fasting insulin of 12 µU/mL and a fasting glucose of 95 mg/dL gives (12 × 95) ÷ 405, which comes out to roughly 2.8.
| You have | Formula | Divide by |
|---|---|---|
| Glucose in mg/dL | insulin (µU/mL) × glucose (mg/dL) | 405 |
| Glucose in mmol/L | insulin (µU/mL) × glucose (mmol/L) | 22.5 |
Both values have to come from the same fasting draw, after roughly 8 to 12 hours without food. A HOMA-IR built from a post-meal glucose and a separately drawn fasting insulin isn’t measuring what the formula assumes it’s measuring, no matter how carefully you do the multiplication.
Why “normal” doesn’t have one answer
Search for a HOMA-IR reference range and you’ll find a number that looks authoritative — usually somewhere under 2 called normal, somewhere over 2.5 or 3 called resistant. What you’re actually looking at is a research convention, not a line drawn by a diagnostic body the way a fasting-glucose threshold for diabetes is.
A Spanish population study, the EPIRCE cross-sectional survey, shows why that line moves. Using a 90th-percentile method, its own threshold for insulin resistance came out at 3.46. Using a different, clinically anchored method — tied to the presence of metabolic syndrome components rather than a percentile — the equivalent threshold in the same dataset dropped to 2.05. In non-diabetic men, the cutoff that best separated people with and without metabolic syndrome was 1.85; in women aged 50, it was 2.07. The same paper found HOMA-IR’s accuracy at flagging metabolic risk in women declining with age far more than it did in men — a detail that matters only if you know which cutoff, in which population, at which age, produced the chart you’re comparing yourself against.
| What’s commonly cited | Where it tends to come from | What it actually depends on |
|---|---|---|
| Below ~2.0 | General research convention | Which insulin assay, which population |
| ≥2.5 | Common in NHANES-based U.S. analyses | Cohort, age band, sex |
| 1.85–3.46 | EPIRCE study, one single population | Percentile method vs. metabolic-syndrome method |
None of these numbers is wrong. They’re answers to slightly different questions, computed on different groups of people, and none of them is an officially adopted diagnostic threshold.
Why two labs can disagree on the same person
Part of the instability sits inside the insulin measurement itself, before the formula even runs. Fasting insulin is measured by immunoassay, and different manufacturers’ assays are not fully standardized against one another, even when each claims traceability to the same reference material. Research comparing several commercial insulin assays against a mass-spectrometry reference method found some assays running high at low insulin concentrations and low at high concentrations, with meaningful gaps between methods on identical samples.
Practically, that means you can send the same fasting blood draw to two labs running different insulin assays and get two different HOMA-IR scores without anything about your metabolism having changed between the two tubes. That’s a real limitation of the underlying insulin test, not a flaw specific to the HOMA-IR formula — but it lands on HOMA-IR directly, since the formula has no way to correct for it.
Why your doctor probably isn’t ordering this
If you’ve asked a clinician for a HOMA-IR test and been redirected toward a fasting glucose or an A1C instead, that isn’t a brush-off. The U.S. National Institute of Diabetes and Digestive and Kidney Diseases states plainly that health care professionals may not test for insulin resistance at all, because the test is used mainly in research settings. Prediabetes and type 2 diabetes are diagnosed instead from A1C, fasting plasma glucose, or an oral glucose tolerance test — tests with published, agreed cutoffs that HOMA-IR doesn’t have.
That doesn’t make HOMA-IR useless. It makes it a research and self-tracking tool rather than a diagnostic one: good at showing a trend in one person tested consistently, weaker at telling you exactly where you sit against an official line, because no such line exists yet.
What actually moves your number
A few things shift HOMA-IR that have nothing to do with long-term metabolic health, worth ruling out before treating a single score as meaningful:
- How long you actually fasted. Coffee with milk, “just a bite” the night before, or a fast shorter than the assumed 8–12 hours all push fasting insulin upward.
- Acute illness or a bad night’s sleep beforehand. Both raise cortisol, which raises fasting glucose and insulin independent of your baseline metabolic state.
- Which assay your lab happens to use, as above — a genuinely different number, not noise you’re imagining.
- Time of day and recent exercise, both of which shift fasting insulin on a scale of hours, not months.
What moves the number over months, more durably, is visceral fat and how readily your muscle tissue takes up glucose in response to insulin — the underlying biology the formula is trying to approximate in the first place.
Where the popular advice overreaches
A fair amount of wellness content treats “get your HOMA-IR under 1” as a target with the same authority as a blood-pressure guideline. It isn’t. No major diabetes or endocrinology society has published an official HOMA-IR treatment target, for the reasons above: the number moves with the assay and the population, so a universal target would be anchored to nothing stable. Chasing a specific decimal point from a single lab visit is the wrong use of the score.
What genuinely holds up
Set against a single external cutoff, HOMA-IR is on shaky ground. Set against your own trend — same lab, same assay, similar fasting window, checked periodically over months — a rising or falling HOMA-IR is a reasonable thing to track, because most of the instability described above cancels out when you’re comparing yourself with yourself rather than with an internet chart built on someone else’s population. That’s also closer to how the model was validated in the first place: against changes within a modeled system, not against a single universal line.
When to stop reading and see someone
A single HOMA-IR score is not itself an emergency, but book a clinical appointment rather than re-testing at home if a high result comes together with acanthosis nigricans (dark, thickened skin at the neck, armpits or knuckles), a fasting glucose already at or above 100 mg/dL, rapid unexplained weight change, or irregular periods alongside excess hair growth — those combinations point toward a diagnosis that needs managing, not a lab value that needs monitoring.
Questions we get
What counts as a normal HOMA-IR score?
There isn't one number every lab or guideline agrees on. Many research papers treat a score under about 2.0 as insulin-sensitive, and analyses built on U.S. NHANES data commonly use 2.5 as a working line for resistance. But a Spanish population study (EPIRCE) found the equivalent threshold ranged from 1.85 in non-diabetic men to 3.46 using a 90th-percentile method in the same cohort, depending on which criteria and which age group were used to derive it. Treat any single cutoff you read online as a research convention, not a diagnostic line drawn by a guideline body.
Is a high HOMA-IR score the same as being diagnosed with insulin resistance?
No. HOMA-IR is a research index, and the U.S. National Institute of Diabetes and Digestive and Kidney Diseases notes that health care professionals may not test for insulin resistance directly at all, because the test is used mainly in research studies. In clinical practice, prediabetes and type 2 diabetes are diagnosed from fasting plasma glucose, an oral glucose tolerance test, or A1C — not from a calculated insulin-resistance score. A high HOMA-IR is a signal worth discussing with a clinician, not a diagnosis by itself.
Why might two HOMA-IR calculators give different results from the same blood draw?
Because the insulin half of the equation is measured by immunoassays that are not fully standardized across manufacturers and laboratories, even when each method claims traceability to the same reference standard. Research comparing commercial insulin assays against a mass-spectrometry reference method found some assays running high at low insulin concentrations and low at high ones, with meaningful gaps between methods on identical samples. Send one fasting blood draw to two labs using different insulin assays and you can get two different HOMA-IR numbers without anything about your metabolism changing.
How do you actually calculate HOMA-IR from lab results?
Multiply your fasting insulin in microunits per millilitre (µU/mL) by your fasting glucose in milligrams per decilitre (mg/dL), then divide by 405. A fasting insulin of 12 µU/mL with a fasting glucose of 95 mg/dL gives (12 × 95) ÷ 405, or about 2.8. If your glucose is reported in mmol/L instead, the constant changes to 22.5 rather than 405 — using the wrong constant for your units is the most common home-calculation error, and it produces a number that looks plausible rather than obviously wrong.
Where the figures came from
- Matthews DR et al., 'Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man,' Diabetologia (1985) — HOMA-IR's original formula and the model it is derived from — fasting glucose and insulin used as a proxy for insulin resistance rather than a direct measurement
- NIDDK — Insulin Resistance & Prediabetes — Health care professionals may not test for insulin resistance directly, since the test is used mainly in research; prediabetes and diabetes are instead diagnosed via A1C, fasting plasma glucose or an oral glucose tolerance test
- MedlinePlus — Insulin in Blood test — What a fasting insulin blood test measures, and why high insulin alongside normal glucose is read as a sign of insulin resistance
- EPIRCE cross-sectional study on HOMA-IR cut-off values and metabolic syndrome — HOMA-IR cutoff thresholds for insulin resistance vary substantially by sex, age and the diagnostic criteria used to derive them within the same population
- The Current Status of Serum Insulin Measurements and the Need for Standardization — Insulin immunoassays from different manufacturers are not fully standardized and can produce meaningfully different results on the same sample despite shared traceability claims
Ines Calderon
Editor responsible for the metabolic health section
Ines edits the metabolic health section and the tools section. Most of her work sits in the gap between what a number on a lab report means and what a reader can actually do about it on a Tuesday. She is not a clinician and holds no medical qualification; what she does is read the primary sources, write down what they say rather than what they are usually reported to say, and mark clearly where a question stops being answerable by an article.
More in Metabolic health
Dawn Phenomenon: Why Blood Sugar Is Highest in the Morning
A high reading on waking looks like last night's food catching up with you. Usually it isn't — it's a hormone surge that starts hours before you're awake, and it is not the only mechanism that can produce the same number.
What Is a Glucose Tolerance Test, and What Do the Results Mean?
A fasting glucose gives you one number at rest, and an A1C averages three months of it. This test asks something neither can: how fast your body actually clears glucose once it has real work to do.
Does Magnesium Deficiency Cause Insulin Resistance?
Every wellness roundup calls magnesium the fix for insulin resistance. The primary sources describe something messier: a nutrient that supports insulin signalling, and a mineral that resistance itself burns through faster, moving in both directions at once.
The 5 Metabolic Syndrome Criteria, Explained
Meeting three of the five checks earns the label, but which three you meet, and whether a number you already treated with medication still counts, changes what the diagnosis is actually telling you.