HOMA-IR Index (Insulin Resistance)

What HOMA-IR estimates and when it is used

HOMA-IR (Homeostasis Model Assessment of Insulin Resistance) is an index that estimates insulin resistance from glucose and insulin measured in the same fasting sample. It was described by Matthews and colleagues in the journal Diabetologia in 1985.

The model's reasoning is physiological. In the fasting state, glucose and insulin are kept in balance by a feedback loop between the liver and the pancreas. When tissues respond poorly to insulin, the pancreas needs more insulin to maintain the same glucose level, and the product of the two rises.

In the original study, HOMA-IR correlated at 0.88 with the euglycemic clamp, a reference method that is expensive and laborious. Because it requires only one sample, the index spread through epidemiological studies and began to appear in routine test requests. The same sample allows calculation of the HOMA-B index, which estimates beta-cell function, with the same two inputs.

How to calculate HOMA-IR in the calculator: inputs and units

The calculator asks for two values from the same fasting sample, and both must be greater than zero:

1. Fasting glucose, in mg/dL, as it appears on most Brazilian laboratory reports.
2. Fasting insulin, in µU/mL. The unit mU/L is numerically identical (1 µU/mL = 1 mU/L), so the value is entered without conversion.

The tool multiplies glucose by insulin, divides by 405 and shows the result with two decimal places. It then compares the value with the 2.7 cutoff from the Brazilian BRAMS study, explained in the interpretation section. Glucose below 20 triggers a warning that the value appears to be in mmol/L.

Before entering values, check three points on the report: whether glucose is in mg/dL and not mmol/L, whether insulin is in µU/mL or mU/L and not pmol/L, and whether both measurements came from the same sample. To follow long-term glycemic control, which HOMA-IR does not measure, use the estimated average glucose from HbA1c.

HOMA-IR formula: 1985 version and symbol definitions

The calculator uses the linear approximation of the original HOMA model (HOMA1), described by Matthews and colleagues in 1985.

HOMA-IR = G_fasting (mg/dL) × I_fasting (µU/mL) ÷ 405
HOMA-IR = G_fasting (mmol/L) × I_fasting (µU/mL) ÷ 22.5

• G_fasting: fasting plasma glucose.
• I_fasting: fasting plasma insulin.
• 405 and 22.5: normalization constants. The value 405 is 22.5 × 18, because 1 mmol/L of glucose corresponds to about 18 mg/dL.

The result is dimensionless. Given how the model was calibrated, values close to 1 correspond to the sensitivity of young adults of normal weight, and higher values indicate more insulin for the same glucose level.

There is also HOMA2, an updated computational model that does not fit into a simple formula. Levy, Matthews and Hermans argued in 1998 that correct assessment uses the computer program. This calculator does not apply HOMA2, and the two results are not interchangeable.

Worked HOMA-IR example, step by step

Consider a hypothetical case, with no clinical interpretation: fasting glucose of 100 mg/dL and fasting insulin of 12 µU/mL, measured in the same sample.

1. Multiply glucose by insulin: 100 × 12 = 1,200.
2. Divide by 405: 1,200 ÷ 405 = 2.962…
3. Round to two decimal places: HOMA-IR = 2.96.

Checking with the mmol/L version leads to the same number. Glucose of 100 mg/dL is equivalent to 5.556 mmol/L (100 ÷ 18); 5.556 × 12 ÷ 22.5 = 2.96.

The calculator returns 2.96 for these data and, because the value is greater than 2.7, reports that it is above the insulin resistance cutoff of the Brazilian BRAMS study. With insulin of 10 µU/mL and glucose of 90 mg/dL, the index would be 2.22, below the cutoff. With the 2.3 cutoff that the same study associated with metabolic syndrome, however, the reading would be different. The example shows why the number alone does not establish a diagnosis. If the same patient is overweight, the BMI calculator helps provide context.

How to understand the HOMA-IR result

The higher the HOMA-IR, the more insulin is needed to maintain fasting glucose, which the model interprets as greater insulin resistance. The index is continuous and has no universal normal value.

The calculator uses the Brazilian cutoff from the Brazilian Metabolic Syndrome Study (BRAMS), by Geloneze and colleagues (2009), which studied 1,203 adults without diabetes. The cutoff for insulin resistance, defined by the 90th percentile of a healthy group, was HOMA1-IR above 2.7. Above it, the tool reports that the value exceeds that cutoff; up to 2.7, that it is within it.

The same study proposed a cutoff of 2.3 to identify metabolic syndrome, which the tool does not apply. Both values apply to the insulin assay and population of that study, and do not apply to children or to people who use insulin.

The result helps describe the metabolic profile, but does not diagnose diabetes, prediabetes or metabolic syndrome. Interpretation is up to the professional who knows the patient, the other tests and the laboratory method.

Common errors, units and verification checklist

Most HOMA-IR errors come from mixed-up units or an inappropriate sample, not from the calculation.

• Glucose in mmol/L with the 405 divisor: the result comes out about 18 times smaller. Convert first (mmol/L × 18 ≈ mg/dL) or use the 22.5 divisor.
• Insulin in pmol/L: the value needs to be converted to µU/mL using the factor the laboratory provides for its assay; entering it in pmol/L overestimates the index.
• Non-fasting collection: the model assumes a basal steady state; after a meal, glucose and insulin rise and the index loses meaning.
• Use of exogenous insulin: the measured insulin no longer reflects pancreatic secretion.
• Rounding midway through the calculation: converting glucose, rounding and only then multiplying can change the second decimal place; the calculator uses the values entered without rounding.

Checklist before releasing or recording the result:

1. Are both measurements from the same fasting sample?
2. Is glucose in mg/dL and insulin in µU/mL or mU/L?
3. Was the formula used HOMA1 (÷ 405), and not HOMA2?
4. Is the cutoff used for comparison compatible with the population and the assay?
5. Does the value make sense given the glucose level and the clinical picture?

Limitations of HOMA-IR and comparison with other methods

HOMA-IR has low individual precision: in the original study, the coefficient of variation of the estimate was 31%. A single measurement can vary considerably between samples, and for that reason the index works better in groups than for classifying an individual.

Situations in which the estimate is less reliable:

• people using insulin or with very reduced secretion, as in type 1 diabetes;
• marked hyperglycemia, outside the range in which the linear approximation was described;
• comparison of values obtained with different insulin assays, which are not standardized between manufacturers.

Wallace, Levy and Matthews (2004) reviewed the use of the model and recommended caution: HOMA is appropriate for cohorts and epidemiological studies, and the input data need to be robust.

The lipid profile often accompanies insulin resistance; see LDL cholesterol by the Friedewald formula.

Frequently asked questions about the HOMA-IR index

References

1. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412-9. View
2. Geloneze B, Vasques AC, Stabe CF, Pareja JC, Rosado LE, Queiroz EC, et al. HOMA1-IR and HOMA2-IR indexes in identifying insulin resistance and metabolic syndrome: Brazilian Metabolic Syndrome Study (BRAMS). Arq Bras Endocrinol Metabol. 2009;53(2):281-7. View
3. Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004;27(6):1487-95. View
4. Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21(12):2191-2. View
5. Rodacki M, Cobas RA, Zajdenverg L, et al. Diagnóstico de diabetes mellitus. Diretriz Oficial da Sociedade Brasileira de Diabetes. 2026 edition (revised July 9, 2024). View