HOMA-IR (insulin resistance)

Homeostasis model assessment of insulin resistance from fasting glucose and insulin

Sources last checked: 2026-10-27

HOMA-IR (insulin resistance)

Homeostasis model assessment of insulin resistance from fasting glucose and insulin

Glucose unit
Fill in every field to see the result.

There is no universal cut-off: Matthews scaled the model so that normal-weight young adults average about 1, and published population cut-offs vary with ethnicity, age and the insulin assay. Not valid on insulin treatment or when β-cell function is severely reduced.

Source: Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28(7):412-419.Open

What is the HOMA-IR (insulin resistance)?

HOMA-IR is the homeostasis model assessment of insulin resistance, published by Matthews and colleagues in Oxford in 1985. It estimates insulin resistance from a single fasting glucose and insulin value; the constant 22.5 (or 405 with glucose in mg/dL) scales the result so that healthy, normal-weight young adults average about 1.

When to use it

  • Research and epidemiology on insulin resistance
  • Tracking insulin resistance over time in the same person with the same insulin assay
  • Context in metabolic syndrome, PCOS or MASLD work-up where a fasting insulin is already available

When not to use it

  • Patients on insulin or with markedly reduced β-cell function (e.g. long-standing type 2 or type 1 diabetes)
  • As a diagnostic test for diabetes — that rests on glucose and HbA1c criteria
  • Non-fasting samples

How it is calculated

HOMA-IR = fasting glucose (mmol/L) × fasting insulin (µU/mL) / 22.5 HOMA-IR = fasting glucose (mg/dL) × fasting insulin (µU/mL) / 405

Clinical pearls and limitations

  • There is no universal cut-off: Matthews scaled the model so that normal-weight young adults average about 1, and published population cut-offs vary with ethnicity, age and the insulin assay. Not valid on insulin treatment or when β-cell function is severely reduced.
  • There is no single validated cut-off; published thresholds range roughly from 1.7 to 3.0 depending on population and insulin assay.
  • Insulin assays differ substantially between laboratories, so values from different labs are not directly comparable.
  • The updated computer model HOMA2 gives different (non-linear) results; this calculator uses the original linear approximation.
  • Insulin pmol/L ÷ 6 ≈ µU/mL (some labs use 6.945).

Frequently asked questions

What is a normal HOMA-IR?

By construction, around 1 in healthy normal-weight young adults. Values above roughly 2–3 are often reported as insulin resistance, but the threshold depends on the population and the assay.

What is the HOMA-IR formula in mg/dL?

Fasting glucose (mg/dL) × fasting insulin (µU/mL) / 405.

Does HOMA-IR work in people with diabetes?

Only with caution in early type 2 diabetes not treated with insulin; it is not meaningful on insulin therapy.

Evidence and guidelines

  1. Primary source: Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28(7):412-419.

Have a follow-up question?

Medelix answers clinical questions from current guidelines and studies — every statement with its source.

Related calculators

Embed this calculatorFree to use on your website, intranet or teaching page. Copy the code below.
<iframe src="https://medelix.ai/en/embed/homa-ir" title="HOMA-IR (insulin resistance) – Medelix" width="100%" height="620" style="border:0;max-width:720px" loading="lazy"></iframe> <p><a href="https://medelix.ai/en/calculators/homa-ir-insulin-resistance">HOMA-IR (insulin resistance) calculator by Medelix</a></p>

For healthcare professionals and medical students. These tools support, and do not replace, clinical judgement. Always check results against the original source and your local protocols.