Clinical resources · Understanding the measurement
Two people can have the same fasting glucose, yet need different amounts of insulin to maintain it. HOMA-IR uses both measurements to estimate that relationship.
HOMA-IR means homeostasis model assessment of insulin resistance. It is an indirect estimate from fasting glucose and fasting insulin. A higher value can suggest lower insulin sensitivity in a comparable setting. It does not, by itself, diagnose diabetes, establish a cause, or tell someone which treatment to take. Original HOMA study; HOMA interpretation review.
Start with the glucose–insulin feedback loop
Insulin is a hormone released by pancreatic beta cells. In the fasting state, glucose and insulin participate in a feedback system involving the liver, pancreas and other tissues. HOMA uses a mathematical model of this basal balance to estimate insulin resistance and beta-cell function. A fasting blood sample is a snapshot of the balance the system has reached. Matthews and colleagues, 1985.
An analogy: room temperature and heating effort
Imagine two rooms at the same temperature. One needs little heating; the other needs much more. Reading the thermometers alone would miss that difference. In this analogy, glucose is the temperature and insulin is the heating effort. HOMA asks what the two readings together might reveal.
The analogy has limits: insulin is not a fuel gauge, and its blood concentration reflects more than secretion alone. The model needs assumptions. That is why a convincing explanation of the number is as important as the calculation.
Same glucose, different insulin: a worked example
These are invented teaching examples using the HOMA1 equation. They are not diagnostic categories or reference ranges.
| Example | Fasting glucose | Fasting insulin | HOMA1-IR |
|---|---|---|---|
| A | 90 mg/dL | 5 µU/mL | 1.11 |
| B | 90 mg/dL | 15 µU/mL | 3.33 |
Example B illustrates why looking only at glucose can miss a pattern compatible with compensatory insulin secretion. It does not prove that this person has exactly three times the biological insulin resistance, or three times the risk of disease. A model-derived ratio is not an individual prognosis.
The formula—and why the units matter
HOMA1-IR = fasting glucose × fasting insulin ÷ 405
Glucose in mmol/L
HOMA1-IR = fasting glucose × fasting insulin ÷ 22.5
In both equations, insulin must be in µU/mL (also written µIU/mL) or mIU/L, which are numerically equivalent. HOMA1-IR is an index, not a glucose concentration or a percentage. The 405 denominator follows from the glucose unit conversion: 22.5 × 18. Oxford HOMA FAQ.
For glucose 100 mg/dL and insulin 10 µU/mL: 100 × 10 ÷ 405 = 2.47, rounded. Use glucose and insulin from the same fasting draw. Follow the laboratory’s preparation instructions; do not stop medication simply to generate a score.
HOMA1 and HOMA2 are different models
The equations above are the original HOMA1 approximation. HOMA2 is an updated computer model with more detailed physiological relationships and assay considerations. Results and cutoffs should not be interchanged. Oxford cautions that the older equations are not appropriate for currently available insulin assays; the simple calculation here is presented to explain a widely reported research index, not as a validated individual diagnostic test. Oxford HOMA history.
What counts as a high result?
There is no universal cutoff that diagnoses insulin resistance in everyone. A study may choose a threshold to identify an upper portion of its population, or to distinguish people with a particular outcome. Those are different purposes. Age, sex and the population studied can alter the threshold. The EPIRCE study demonstrated that the choice of definition changed the cutoff even within a Spanish adult population. Gayoso-Diz and colleagues, 2013.
Before interpreting a quoted threshold, ask: Which HOMA model? Which assay? Which population? Which outcome? A value labelled “high” on a website may not have been validated for the person being tested.
Why a lower result is not always reassuring
The index depends on the pancreas being able to produce insulin. If insulin secretion is impaired, a low insulin concentration cannot automatically be interpreted as excellent insulin sensitivity. Insulin treatment, nonfasting samples and an unstable metabolic state also complicate interpretation. HOMA is a basal model, not a test of the response to a meal or a glucose challenge. Wallace, Levy and Matthews, 2004.
Why small changes may not mean improvement or deterioration
Laboratory methods matter. In a study of 97 fasting participants, HOMA estimates varied by as much as twofold across 11 insulin assays. That was a comparison of methods, not a twofold change in participants’ biology. Serial results are more interpretable when the laboratory, assay, model and sampling conditions are comparable. Avoid treating a small numerical change as a precise measure of biological change. Manley and colleagues, 2008.
For clinicians: identify the question before ordering the number
- Is the question glycemic status? Use established glucose and HbA1c pathways for diagnosing prediabetes and diabetes. HOMA-IR does not replace them. NIDDK diagnostic testing guidance.
- Is the question basal insulin resistance in a study or a carefully defined setting? Record the model, units, assay and sampling protocol. HOMA was developed and validated by comparison with other physiological methods, with limitations in individual precision. Original validation.
- Will the result change a clinical decision? Explain in advance what additional information it supplies and what it cannot resolve. A precise-looking decimal should not create an unsupported treatment target.
A useful report might read: “HOMA1-IR 2.47, calculated from fasting glucose 100 mg/dL and insulin 10 µU/mL; interpretation depends on sampling conditions, insulin assay and clinical context.” It is more informative than recording the score alone.
How this differs from glucose and HbA1c
Fasting glucose measures glucose at one time. HbA1c reflects average glucose exposure over about three months. HOMA-IR combines fasting insulin with glucose to estimate a different physiological relationship. A normal HbA1c therefore does not answer the same question as HOMA-IR. Diabetes diagnosis uses validated glycemic tests and, usually, confirmation—not a HOMA cutoff. NIDDK: tests and diagnosis.
Questions to take to an appointment
- Were both samples collected fasting, at the same time, and in the correct units?
- Is this HOMA1 or HOMA2, and does the laboratory provide an applicable interpretation?
- Could insulin treatment, impaired secretion or the testing circumstances make the result misleading?
- What decision, if any, would this result change alongside my other findings?
Glucose alone does not reveal the entire glucose–insulin relationship. HOMA-IR provides an estimate of that relationship, but its clinical meaning depends on how it was measured and why it was calculated.
Continue reading
Insulin resistance and its biology · Fasting glucose · HbA1c · TyG index · Clinical Resources directory
Educational information for patients and clinicians. This page does not provide individual diagnosis, treatment recommendations or laboratory interpretation. Sources are linked beside the statements they support.