What is insulin resistance?
Insulin lets muscle, liver and fat cells take up and store glucose. In insulin resistance, these tissues respond less well, so more insulin is needed to keep blood glucose in range. As long as the pancreas can keep up, glucose stays normal. Over time, if insulin output can no longer compensate, glucose rises into the prediabetes and then type 2 diabetes range.
Insulin resistance is closely linked with abdominal obesity, physical inactivity, high triglycerides, low HDL cholesterol, high blood pressure, fatty liver disease and polycystic ovary syndrome (see our PCOS blood tests guide).
Fasting insulin
Fasting insulin is measured after an overnight fast, usually alongside fasting glucose. A raised fasting insulin with a normal glucose suggests the pancreas is working harder than usual to maintain that normal glucose. Results are reported in µIU/mL (also written mU/L) or pmol/L; 1 µIU/mL is approximately 6 pmol/L, though the exact factor used varies slightly between labs.
There is no internationally standardised insulin assay, so reference intervals differ considerably between labs. Use the range printed on your report. Fasting insulin is not useful in people who take insulin injections, and certain medicines and recent food intake will affect it.
HOMA-IR
HOMA-IR (Homeostatic Model Assessment of Insulin Resistance) combines fasting glucose and fasting insulin into one number:
- US units: glucose (mg/dL) × insulin (µIU/mL) ÷ 405
- SI units: glucose (mmol/L) × insulin (µIU/mL) ÷ 22.5
Higher values suggest more insulin resistance. The diagram below shows why HOMA-IR can pick up something glucose alone misses.
What is a "normal" HOMA-IR?
HOMA-IR was designed so that a value of about 1 corresponds to a healthy young reference population. Beyond that, published thresholds for "insulin resistant" vary widely — studies have used figures from roughly 2 to 3 or higher, depending on the population, ethnicity, age and insulin assay. Because of this, many labs do not report HOMA-IR at all, and a value of, say, 2.6 should be read as "worth looking at the rest of the picture", not as a diagnosis.
Triglyceride/HDL ratio
Insulin resistance tends to raise triglycerides and lower HDL cholesterol, so the ratio between them is a cheap, indirect clue that can be calculated from a standard lipid panel. Divide triglycerides by HDL using the same units. A ratio above about 3 (in mg/dL), or about 1.3 (in mmol/L), is often quoted as suggestive, but performance differs by ethnicity and sex, and non-fasting samples, alcohol and recent meals can raise triglycerides.
TyG index
The triglyceride–glucose (TyG) index uses two routine fasting results: ln[fasting triglycerides (mg/dL) × fasting glucose (mg/dL) ÷ 2]. It does not need an insulin measurement, which makes it attractive where insulin testing is unavailable. Research links higher TyG values with insulin resistance and future diabetes risk, but there is no agreed clinical cut-off and it is mainly used in research.
Comparing the markers
| Marker | How it is calculated | Less suggestive | More suggestive of insulin resistance |
|---|---|---|---|
| Fasting insulin | Measured directly (µIU/mL or pmol/L; 1 µIU/mL ≈ 6 pmol/L) | Within the lab's reference interval, toward the lower end | Above the lab's interval, or high relative to fasting glucose |
| HOMA-IR | Glucose (mg/dL) × insulin (µIU/mL) ÷ 405, or glucose (mmol/L) × insulin ÷ 22.5 | Around 1 is typical of insulin-sensitive adults | Studies often use thresholds somewhere between about 2 and 3; no agreed cut-off |
| Triglyceride/HDL ratio | Triglycerides ÷ HDL cholesterol (same units) | Lower ratios | Often quoted: > 3 in mg/dL (≈ > 1.3 in mmol/L); varies by ethnicity and sex |
| TyG index | ln[triglycerides (mg/dL) × glucose (mg/dL) ÷ 2] | Lower values | Proposed cut-offs differ between studies and populations |
These are general research-based patterns, not diagnostic ranges. Reference intervals vary by lab — use the range printed on your report.
Limitations to keep in mind
- No universal cut-offs — thresholds were derived in specific populations and do not transfer neatly to everyone.
- Assay variability — insulin immunoassays are not standardised, so the same blood can give different insulin (and HOMA-IR) values at different labs.
- Day-to-day variation — fasting insulin fluctuates; a single result is a snapshot.
- Fasting matters — all four markers assume a proper overnight fast.
- Not valid in some situations — insulin treatment, advanced diabetes with failing insulin production, acute illness, or some medicines (e.g. corticosteroids) distort the results.
What the results can (and cannot) change
Clinical guidelines do not currently recommend routine insulin or HOMA-IR testing for screening. Decisions are usually based on established tests — fasting glucose, the OGTT and HbA1c — plus your overall cardiometabolic risk. Where insulin resistance markers are raised, the usual response is the same as for prediabetes: attention to weight, physical activity, diet, sleep, and regular monitoring. Abnormal liver enzymes may also prompt assessment for fatty liver disease, which often accompanies insulin resistance.
See your metabolic markers together
Upload your blood test report to our AI analyzer — it reads glucose, insulin, triglycerides and HDL against your lab's reference ranges and explains the pattern in plain language.
Analyze my blood test →