Homeostatic Model Assessment for insulin resistance from fasting glucose and insulin.

What it is and when to use it

The HOMA-IR index estimates insulin resistance from a single pair of fasting values: plasma glucose and plasma insulin. Matthews and colleagues at the Radcliffe Infirmary in Oxford described it in Diabetologia in 1985, based on a mathematical model of the feedback loop between the liver and the pancreatic β-cell. The paper's approximation is HOMA-IR = glucose (mmol/L) × insulin (µU/mL) / 22.5; this calculator takes glucose in mg/dL and divides by 405 (22.5 × 18). The scale is calibrated so that healthy normal-weight adults under 35 years of age have an insulin resistance of 1. In the original study it correlated with the euglycaemic clamp (Rs = 0.88) and with fasting insulin (Rs = 0.81).

How to interpret it

The 1985 paper sets no cut-offs: it treats HOMA-IR as a relative value, where 1 is a healthy young adult of normal weight. The bands on this page (<1.0 optimal sensitivity; 1.0–1.9 mild resistance; ≥2.0 significant resistance) do not come from the paper and are indicative only. In the study, the median was 1.21 in subjects without diabetes (aged 23 to 67, weighing 90% to 142% of ideal body weight) and 2.89 in patients with type 2 diabetes, so at least half of the healthy controls were above 1.0. The authors warn that its low precision limits clinical use from a single sample.

Limitations and when not to use it

The coefficient of variation was 31%, and the authors judged that one fasting sample drawn in clinic does not reliably reflect an individual's insulin resistance. They advised measuring insulin with the subject rested and noted that averaging three samples taken 5 minutes apart improves precision, because of pulsatile insulin secretion and the effect of stress. The result depends on the insulin immunoassay (including possible detection of proinsulin), so it is not directly comparable between laboratories. Exercise also alters it. The patients with diabetes in the study were on diet alone; with exogenous insulin the index cannot be interpreted.

Frequently asked questions

How do you calculate HOMA-IR?

Multiply fasting glucose by fasting insulin (µU/mL) and divide by 22.5 if glucose is in mmol/L, or by 405 if it is in mg/dL. Example: 90 mg/dL and 4 µU/mL give 90 × 4 / 405 = 0.89.

What is a normal HOMA-IR value?

The original paper sets no universal normal value or cut-off. It defines 1 as the resistance of a healthy normal-weight adult under 35, and in its sample the median in subjects without diabetes was 1.21. Cut-offs used in practice vary with the population and the insulin assay.

What is the difference between HOMA-IR and HOMA-β?

Both come from the same model. HOMA-β estimates β-cell function as a percentage: 20 × insulin / (glucose in mmol/L − 3.5), where 100% corresponds to a healthy young adult of normal weight. This calculator reports HOMA-IR only.

Can HOMA-IR be used in patients who inject insulin?

It is not advisable. The model reads fasting insulin as the patient's own β-cell response to glucose, and injected insulin breaks that relationship. The patients with diabetes in the original study were on diet alone, so there is no validation in people treated with insulin.