How to calculate HOMA-IR from fasting insulin and glucose

HOMA stands for homeostasis model assessment. Matthews and colleagues introduced it in 1985 to estimate insulin resistance and beta cell function from a single fasting glucose and insulin pair [1].

• US units: HOMA-IR = glucose (mg/dL) x insulin (µIU/mL) / 405
• SI units: HOMA-IR = glucose (mmol/L) x insulin (µIU/mL) / 22.5

The 405 is simply 22.5 x 18, because 1 mmol/L of glucose equals about 18 mg/dL. Insulin must be in µIU/mL (the same as mU/L) in both formulas. If your lab reports pmol/L, divide by 6.0 first; some labs convert with 6.945 instead, which changes the result by about 15% [2].

These simple formulas are the original 1985 approximation. The HOMA developers later released a computer model, HOMA2, and stated that it gives the correct estimate [6]. HOMA2 values are not interchangeable with the simple formula, so use one method consistently.

Step 1. Check your insulin unit. Suppose your report shows 60 pmol/L. Divide by 6.0: 60 / 6.0 = 10 µIU/mL [2].

Step 2. Multiply glucose by insulin. 90 mg/dL x 10 = 900. Or 5.0 mmol/L x 10 = 50.

Step 3. Divide. 900 / 405 = 2.22. Or 50 / 22.5 = 2.22.

Step 4. Read it in context. A HOMA-IR of 2.2 is below the Japanese reference limit of 2.5 [7] and the Spanish 75th percentile of 2.6 [8], but above the 1.775 cut point used in Iran [3] and the 2.05 metabolic syndrome cut point in Spain [9]. That spread is why a single HOMA-IR number has no universal label.

A second example: glucose 100 mg/dL and insulin 20 µIU/mL gives 100 x 20 / 405 = 4.9, above almost every published cut point except the stand-alone 5.9 threshold from clamp studies [10].

The table shows why insulin usually drives HOMA-IR in people without diabetes. Moving from 80 to 110 mg/dL of glucose raises HOMA-IR by about a third, while doubling insulin doubles it.

The row for 125 mg/dL is the top of the ADA prediabetes range; 126 mg/dL or higher meets the fasting glucose criterion for diabetes [11]. In diabetes, HOMA-IR becomes harder to interpret because beta cell output is failing, and different cut points apply [3].

The differences come from three sources: who was studied, what the cut point was anchored to (a percentile, metabolic syndrome or a clamp test) and which insulin assay the lab used [12].

The Mexican American study found that 39.1% of adults were above its 3.80 cut point and noted that its result was very different from the popular clinical cut-off of 2.60 [4]. In Spain, the threshold also varied with age in women without diabetes [9].

For a person, the practical reading is relative: a value well above 3 on a standard assay usually means insulin resistance is likely, a value below about 1.8 makes it unlikely, and values in between depend on context such as waist size, triglycerides, HDL and family history.

• Imprecise for individuals. The original paper found HOMA-IR correlated with the clamp (Rs 0.88) but had a coefficient of variation of 31% [1]. A later study confirmed a strong group-level correlation (r = -0.82) and judged HOMA suitable for large studies [5].
• Insulin assay dependence. Assays disagree by a median of 24% [12], so the same blood can produce different HOMA-IR values on different labs.
• Single-sample noise. Fasting insulin can vary 25% to 50% between samples [13].
• Formula versions. The simple formula and the HOMA2 computer model give different numbers [6]. The developers warn that HOMA needs robust input data and careful interpretation [14].
• Not for insulin users. Injected insulin and insulin antibodies distort the insulin value [15].
• Not a diagnosis. HOMA-IR is not a diagnostic criterion in the ADA Standards of Care [11].

The reference method is the hyperinsulinemic euglycemic clamp, which is used in research rather than routine care [16].