HOMA-IR (Homeostatic Model Assessment of Insulin Resistance) combines fasting insulin and fasting glucose into a single number. It estimates how insulin-resistant your metabolism is right now, years before the problem shows up on a standard diabetes screen. It costs nothing extra — if you already have both values from a fasted draw, you already have your HOMA-IR.
The Verdict
What HOMA-IR physically represents
The model was built to describe a feedback loop. Your liver releases glucose overnight; insulin restrains it; the pancreas adjusts insulin until glucose settles. The product of the two fasting values estimates how far that loop has drifted from a healthy set point.
An important nuance: because the calculation uses fasting values, it reflects hepatic insulin resistance more than muscle insulin resistance. Muscle resistance shows up mainly after meals. That is why HOMA-IR correlates strongly with fatty liver and less tightly with post-meal glucose spikes.
The calculation
HOMA-IR = (fasting glucose in mg/dL × fasting insulin in μIU/mL) ÷ 405
Outside the US, where glucose is reported in mmol/L, the divisor is 22.5 instead. Both values must come from the same fasted draw. A newer computer model, HOMA2, corrects for non-linearity at higher glucose levels and is more accurate above roughly 180 mg/dL — but for screening in the normal-glucose range, the original formula is adequate.
Four worked examples
| Scenario | Fasting glucose | Fasting insulin | HOMA-IR | Interpretation |
|---|---|---|---|---|
| Lean, insulin-sensitive | 85 mg/dL | 3 μIU/mL | 0.63 | Optimal. Nothing to act on. |
| Looks normal on a standard panel | 95 mg/dL | 12 μIU/mL | 2.81 | Glucose passes. Insulin resistance is already established. |
| Prediabetic on paper | 110 mg/dL | 14 μIU/mL | 3.80 | Both inputs abnormal. Needs a clinical plan. |
| Late beta-cell failure | 150 mg/dL | 4 μIU/mL | 1.48 | A falsely reassuring score. Insulin production is failing, not sensitivity improving. |
The second row is the case that standard care misses. Every individual value passes; the composite does not. The fourth row is the case that fools people who track the score without understanding it.
Standard interpretation vs optimal targets
Most labs do not report HOMA-IR at all, so there is no printed reference range to compare against. The thresholds below come from research populations rather than from a single guideline, which is why they are stated as bands.
| Band | HOMA-IR | What it indicates |
|---|---|---|
| Optimal | Under 1.0 | Typical of lean, active adults with normal glucose tolerance. |
| Good | 1.0–1.5 | A reasonable working target for most adults. |
| Early insulin resistance | 1.5–2.5 | Glucose is usually still normal here. This is the reversible window. |
| Established insulin resistance | 2.5–2.9 | Commonly accompanied by high triglycerides and rising waist circumference. |
| Marked insulin resistance | Above 2.9 | Strongly associated with fatty liver and progression to type 2 diabetes. |
Why HOMA-IR beats fasting glucose alone
Fasting glucose is a lagging indicator by design. The pancreas can hold glucose inside the normal range for a decade by producing more insulin, and a glucose-only screen reports "normal" throughout that decade. HOMA-IR captures both the output and the effort behind it.
Its limits deserve equal billing. The reference method for insulin sensitivity is the hyperinsulinemic-euglycemic clamp, a research procedure taking several hours. HOMA-IR correlates well with the clamp at a population level but has meaningful individual scatter. Treat it as a screening estimate and a trend line, not a precise measurement.
How and when to test
- Fast 8–12 hours, water only. Both inputs must come from that single draw.
- Do not extend the fast past about 14 hours. Prolonged fasting suppresses insulin and deflates the score.
- Skip intense exercise for 24 hours beforehand. A hard session can improve insulin sensitivity enough to mask a resistant baseline.
- Use the same lab for repeat tests. Insulin assays are not fully standardized, so cross-lab comparisons introduce noise into the trend. Confirm an unexpected result on a second draw before changing anything, since the insulin term is the more variable half of the equation; our biomarker guides cover the standard practice of repeating an abnormal result before acting on it.
- Do not test during acute illness, after a poor night of sleep, or during a steroid course. Each inflates the result.
- Repeat every 3–6 months while actively changing something, and annually once stable.
What it means in combination with other markers
- HOMA-IR above 2.5 + triglyceride-to-HDL ratio above 3: two independent estimates agreeing. This is a confident read of insulin resistance.
- HOMA-IR above 2.5 + elevated ALT: points strongly toward fatty liver, which is the tissue-level counterpart of hepatic insulin resistance.
- HOMA-IR above 2.5 + elevated ApoB: metabolic and lipid risk stacked. Treating the insulin resistance often improves both.
- HOMA-IR above 2.5 + elevated hsCRP: inflammation and metabolic dysfunction reinforce each other. Visceral fat is the usual common cause.
- HOMA-IR above 2.5 + low SHBG: a recognized pairing. Low SHBG is one of the earlier laboratory signs of hyperinsulinemia.
- Normal HOMA-IR + high post-meal glucose on a monitor: muscle insulin resistance that a fasting measure cannot see. An oral glucose tolerance test with insulin levels is the next step. A panel belongs before a sensor rather than after it, because a sensor shows the glucose outcome without showing the insulin effort holding it there; sensors can now be bought over the counter without a subscription, which makes a short intermittent wear a reasonable follow-on once a panel has flagged something.
What lowers HOMA-IR
- Losing 5–10% of body weight. The most reliable lever when body fat is elevated, and the one with the largest effect size.
- Resistance training two to three times weekly. Builds glucose-disposing tissue; the benefit persists only while training continues.
- Reducing refined carbohydrate and added sugar. Acts mainly through the insulin term, which is the more responsive half of the equation.
- Restoring 7+ hours of sleep. One of the fastest-acting changes available.
- Metformin, where clinically indicated, particularly in prediabetes with additional risk factors.
- GLP-1 receptor agonists, which produce large reductions largely through weight loss.
- Berberine, at 500mg two to three times daily, with a modest but replicated effect and frequent GI side effects.
When a result warrants seeing a physician
- HOMA-IR above 2.9, especially with a waist circumference above 40 inches in men or 35 inches in women.
- Any fasting glucose at or above 100 mg/dL, or HbA1c at or above 5.7%.
- An elevated score alongside raised liver enzymes, which warrants a fatty-liver assessment.
- A score that falls while glucose or HbA1c rises — the beta-cell failure pattern.
- Elevated scores in a woman with irregular cycles, acne, or excess hair growth, which should prompt a PCOS evaluation.
Frequently Asked Questions
What is a good HOMA-IR score?
Under 1.0 is optimal and typical of lean, active adults. Under 1.5 is a reasonable working target. Scores of 1.5–2.5 suggest early insulin resistance, and above 2.9 is a strong marker of established resistance. The exact cut-point is debated — published thresholds for "insulin resistant" range from 1.8 to 2.9 depending on the population studied and the insulin assay used.
How is HOMA-IR calculated?
With glucose in mg/dL: HOMA-IR = (fasting glucose × fasting insulin) ÷ 405. With glucose in mmol/L, divide by 22.5 instead. Both inputs must come from the same fasted blood draw. Using a glucose from one visit and an insulin from another produces a meaningless number.
Is HOMA-IR better than fasting glucose alone?
For catching insulin resistance early, yes. Fasting glucose stays normal for years while the pancreas compensates. HOMA-IR captures both the glucose and the insulin effort required to hold it there. It is not better than an oral glucose tolerance test with insulin levels, which detects post-meal dysfunction that a fasting measure misses entirely.
When is HOMA-IR invalid?
It is uninterpretable in anyone taking exogenous insulin, because the assay cannot distinguish injected insulin from pancreatic output. It is unreliable in advanced type 2 diabetes, where failing insulin secretion pulls the score down while the patient gets worse. It is also invalid on a non-fasting sample, and it should not be calculated from a hemolyzed tube, which falsely lowers insulin.
Does HOMA-IR mean the same thing in every population?
No. South Asian populations develop insulin resistance and fatty liver at lower BMI and lower waist circumference than European populations, and studies in those groups have proposed lower diagnostic cut-points. East Asian populations also tend to show metabolic dysfunction at lower body weight. A HOMA-IR of 2.2 carries different weight in a lean South Asian adult than in a heavier European adult of the same age.
How quickly can HOMA-IR improve?
Quickly. Measurable improvement typically appears within 2 to 6 weeks of consistent change, driven mostly by the insulin term rather than the glucose term. Very low-calorie interventions can drop hepatic insulin resistance within days. Re-test at 8 to 12 weeks for a stable value, using the same lab both times.
Is HOMA-IR still meaningful if I take metformin or a GLP-1?
Yes, but it describes your treated state rather than your untreated baseline. The reason the score is uninterpretable on exogenous insulin does not carry over, because that problem is an assay one, where injected insulin cannot be separated from pancreatic output, and neither of these agents puts insulin into the sample. Both are listed on this page among the things that lower HOMA-IR, so a score measured on stable therapy is a reasonable trend line for whether the treatment is working. Treat any dose change as the start of a new baseline rather than a continuation of the old one, and read the result with the clinician managing the prescription.