HOMA-IR Calculator
Calculate HOMA-IR, the homeostasis model assessment of insulin resistance, from a fasting glucose and a fasting insulin taken at the same time. This HOMA-IR calculator accepts glucose in mg/dL or mmol/L and insulin in µIU/mL or pmol/L, uses the original formula of 1985[1] and shows how your score compares with published reference values.
For information only — not medical advice.
How to use
- Enter your fasting glucose from the lab report and pick its unit.
- Enter your fasting insulin from the same blood sample and pick its unit: µIU/mL (also written mU/L or mIU/L) or pmol/L.
- Read your HOMA-IR under the form.
Example: a fasting glucose of 90 mg/dL with a fasting insulin of 10 µIU/mL gives a HOMA-IR of 2.22[1]. That is below the cut-off for insulin resistance used in the Bruneck Study[2].
How it is calculated
HOMA is a model of how fasting glucose and fasting insulin hold each other in balance: the pancreas releases insulin, and insulin keeps glucose from rising[1]. Its authors also gave a simple approximation, which most studies use instead of the full computer model, and the calculator uses it too[3]:
- HOMA-IR (homeostasis model assessment of insulin resistance)
- HOMA-IR = fasting glucose (mmol/L) × fasting insulin (µIU/mL) ÷ 22.5
- = fasting glucose (mg/dL) × fasting insulin (µIU/mL) ÷ 405
- Insulin and glucose units
- Insulin: 1 µIU/mL = 6.00 pmol/L
- Glucose: 1 mmol/L = 18 mg/dL
The score is scaled so that 1.0 is the model's reference for healthy, normal-weight adults under 35[3]. In the original study, HOMA-IR agreed well with the glucose clamp, the reference method for insulin resistance (rank correlation 0.88), but single estimates were imprecise, with a coefficient of variation of 31 percent[1].
What the result means
A higher HOMA-IR means that more insulin is needed to hold fasting glucose at its level — that is, more insulin resistance. There is no single normal range.
In the Bruneck Study, a random sample of 888 people aged 40 to 79, the cut-off for insulin resistance was 2.77: the lower edge of the top fifth of HOMA-IR among people who were not obese and had no metabolic disorders[2]. Other studies and laboratories use other cut-offs, because insulin tests and populations differ.
Limits
- Both values must come from one fasting blood sample[3].
- HOMA-IR is not meaningful for people who take insulin.
- A single result varies from test to test; the model's own estimate had a coefficient of variation of 31 percent[1].
- Insulin in pmol/L is converted at 6.00 pmol/L per µIU/mL[4]. Some laboratories use 6.945; a 2019 commentary on insulin units calls that factor wrong and says it underreports insulin by about 15 percent[4]. With it, HOMA-IR comes out about 14 percent lower[4]. If your report gives µIU/mL or mU/L, enter that.
- The calculator does not diagnose insulin resistance, prediabetes or diabetes.
Frequently asked questions
What is a normal HOMA-IR?
There is no single normal value. The model's reference is 1.0 for healthy, normal-weight adults under 35[3], and the Bruneck Study used 2.77 as the cut-off for insulin resistance[2]. Other populations and laboratories use other cut-offs.
Can I calculate HOMA-IR from a non-fasting sample?
No. The model describes the fasting state, so glucose and insulin must both come from a fasting sample[3].
Why does my lab report a different HOMA-IR?
Laboratories may round differently or convert insulin from pmol/L with another factor. With 6.945 instead of 6.00, the score comes out about 14 percent lower[4].
Sources
- Matthews et al., 1985. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. doi:10.1007/bf00280883
- Bonora et al., 1998. Prevalence of insulin resistance in metabolic disorders: the Bruneck Study. Diabetes. doi:10.2337/diabetes.47.10.1643
- Wallace et al., 2004. Use and Abuse of HOMA Modeling. Diabetes Care. doi:10.2337/diacare.27.6.1487
- Knopp et al., 2019. Insulin Units and Conversion Factors: A Story of Truth, Boots, and Faster Half-Truths. J Diabetes Sci Technol. doi:10.1177/1932296818805074