Insulin resistance tests: HOMA-IR, fasting insulin and the rest compared
Reviewed 16 sourcesPublished by PepVault
Which insulin resistance tests you can get, what each measures, whose cut-offs apply, the HOMA-IR formula worked in mg/dL and mmol/L, and what comes next.

- Last reviewed
- Sources
- 16
- Questions
- 8
- Compounds
- 1
The short answer
No routine blood test diagnoses insulin resistance. The practical one is HOMA-IR: fasting insulin (µIU/mL) times fasting glucose (mg/dL), divided by 405. It has no agreed cut-off; studies put the line between about 1.5 and 3.8. Fasting glucose, HbA1c and the glucose tolerance test diagnose prediabetes and diabetes with fixed ADA cut-offs. The clamp is the research standard.
- Glucose and HbA1c can stay normal while the pancreas makes extra insulin to keep them there, so they show insulin resistance late.
- HOMA-IR needs one fasting draw with insulin and glucose; a single result is rough (the model's own coefficient of variation is 31%).
- Published HOMA-IR cut-offs run from 1.55 to 3.8 depending on the population, and insulin assays disagree between labs.
- The triglyceride-to-HDL ratio and the TyG index come free with a lipid panel, but each misses part of the picture.
Evidence: Diagnostic cut-offs for glucose, HbA1c and the OGTT are the ADA's 2026 standards; HOMA-IR, TyG and the lipid ratio are checked against the clamp or other direct tests in cross-sectional studies of a few hundred people, with no agreed cut-off; the prevention numbers come from a large randomized trial.

Why can your glucose be normal while you are insulin resistant?
Insulin resistance means your muscle, liver and fat respond less to insulin, the hormone that moves glucose out of the blood. The pancreas answers by making more insulin. As long as its beta cells (the insulin-making cells) keep up, blood glucose stays normal; type 2 diabetes starts when that extra output fails1. So a normal fasting glucose or HbA1c says the system is coping, not that it is working easily.
That is why no standard panel names it. The NIH's diabetes institute says health care professionals may not test for insulin resistance and that the direct test is used mainly in research2. What clinicians diagnose instead is prediabetes and diabetes, with glucose tests and fixed cut-offs3. The tests below cover both: the ones that estimate insulin resistance itself, and the ones that catch its effect on glucose.

PepVault for iPhone
Keep your protocol in PepVault
PepVault's bloodwork log has fasting glucose, HbA1c, fasting insulin, HOMA-IR, triglycerides and HDL built in. You enter HOMA-IR as a number, since the app does not calculate it; tap any marker to see it against its range and its trend across panels.
7 days free on the yearly plan for new subscribers.
Which insulin resistance tests can you actually get?
Most need one fasting blood draw. Two need a sweet drink and two to three hours, and one needs a research unit.
| Test | What it needs | What it tells you | Cut-off, and whose |
|---|---|---|---|
| Fasting glucose | Fasting draw (8 hours or more) | Glucose after the night | ADA: 100–125 mg/dL prediabetes, 126 or more diabetes3 |
| HbA1c | Draw at any time, no fasting | Average glucose over about three months | ADA: 5.7–6.4% prediabetes, 6.5% or more diabetes3 |
| Fasting insulin | Fasting draw | How hard the pancreas works to hold glucose | No standard; each lab sets its own range6 |
| HOMA-IR | Fasting insulin and glucose, then a formula | An estimate of insulin resistance | No agreed cut-off: 1.55 to 3.8 across studies5 |
| Triglyceride-to-HDL ratio | Standard lipid panel | A lipid sign of insulin resistance | 3.0 or more in overweight adults, mg/dL (Stanford study)8 |
| TyG index | Fasting triglycerides and glucose, then a formula | An estimate without insulin | Study-specific; 4.68 against the clamp in one study10 |
| Oral glucose tolerance test (OGTT) | 75 g glucose drink, draws at 1 and 2 hours | How fast glucose clears | ADA 2 hours: 140–199 mg/dL prediabetes, 200 or more diabetes3; IDF 1 hour: 155 or more12 |
| Hyperinsulinemic-euglycemic clamp | Two IV lines, insulin and glucose drips, hours in a lab | Insulin sensitivity measured directly | Research standard, not a clinic test114 |
PepVault's bloodwork log has fasting glucose, HbA1c, fasting insulin, HOMA-IR, triglycerides and HDL built in. You enter HOMA-IR as a number, since the app does not calculate it; tap any marker to see it against its range and its trend across panels. Log it in PepVault
How do you calculate HOMA-IR?
HOMA (homeostasis model assessment) was built in Oxford in 1985 from a model of how fasting glucose and insulin balance each other4. The short formula most labs and studies use is fasting insulin (µIU/mL) times fasting glucose (mmol/L), divided by 22.55. With glucose in mg/dL, the divisor is 405, because 22.5 × 18 = 405 and about 18 mg/dL of glucose is 1 mmol/L (ADA's 126 mg/dL is its 7.0 mmol/L)3.
- Take the fasting insulin and fasting glucose from the same morning draw.
- Check the insulin unit. µIU/mL and mU/L are the same number. If your lab reports pmol/L, divide by 6.0, the factor the American Diabetes Association publishes (72 pmol/L is 12 µIU/mL). Some labs and calculators divide by 6.945, which gives a number about 14% lower7.
- In mg/dL: multiply insulin by glucose, then divide by 405. Example: 12 µIU/mL × 90 mg/dL = 1,080; 1,080 ÷ 405 = 2.67.
- In mmol/L: 90 mg/dL ÷ 18 = 5.0 mmol/L. Then 12 × 5.0 = 60; 60 ÷ 22.5 = 2.67. Both routes give the same answer.
- Write the result with the date and the lab: insulin tests differ from lab to lab, so a trend means most at one lab6.
Read one result loosely. In the 1985 paper, HOMA's insulin resistance estimate tracked the clamp well (Spearman correlation 0.88), but its coefficient of variation was 31%, which the authors said limits its use4. Its makers' later review calls it a sound clinical and epidemiological tool only when the fasting samples are robust and the result is read with care11.
What HOMA-IR is too high?
There is no agreed number. Studies have set the line either at a percentile of their own population or at the value that best picked out people with metabolic syndrome, and they land in different places. A 2026 review puts the published range at 1.37 to 3.631; the Spanish EPIRCE study's table of earlier studies runs from 1.55 to 3.85.
| Population (study) | Cut-off | How it was set |
|---|---|---|
| Thailand, 738 adults with normal weight and glucose | 1.55 | 90th percentile5 |
| Japan, 161 healthy adults | 1.7 | 90th percentile5 |
| Spain, adults without diabetes (EPIRCE, 2,459 adults in all) | 1.85 men; 2.07 women at age 50 | Best at finding metabolic syndrome5 |
| US NHANES, 2,804 adults with normal weight and glucose | 2.73 | 66th percentile5 |
| France, 1,153 adults aged 35–64 | 3.8 | 75th percentile5 |
The insulin test adds its own spread. When an American Diabetes Association work group ran 39 serum samples through 10 commercial insulin methods, 7 of the 10 were off from the reference method by more than 15.5% in 36% to 100% of samples6. The same blood can give a different HOMA-IR at two labs. Your clinician reads your number against the lab's range and the rest of your results.
Commonly reported: fasting insulin is not on the standard panel and has to be asked for by name, and most lab reports print insulin and glucose but no HOMA-IR, so people work it out themselves.
What do the triglyceride-to-HDL ratio and the TyG index add?
Both come from a fasting lipid panel you may already have, with no insulin test. In 258 overweight adults without diabetes at Stanford, a triglyceride-to-HDL ratio of 3.0 or more (both in mg/dL) picked out the most insulin-resistant third with 64% sensitivity and 68% specificity: it caught about two in three and wrongly flagged about one in three of the rest8. Example: triglycerides 150 and HDL 40 give 3.75. If your lab reports mmol/L, the ratio comes out smaller: multiply it by 2.29 before comparing (the conversion factors are 88.57 for triglycerides and 38.67 for HDL), so the 3.0 line is about 1.3 in mmol/L.
It does not work for everyone. In 98 overweight African American adults at the NIH, the ratio did little better than chance (area under the curve 0.56, where 0.5 is a coin toss); fasting insulin did far better (0.85)9.
The TyG index adds glucose to triglycerides. Its Mexican developers wrote it as ln[triglycerides (mg/dL) × glucose (mg/dL) / 2], ln being the natural log10. Against the clamp in 99 people, it correlated at −0.681 with measured insulin sensitivity, and a cut-off of 4.68 gave 96.5% sensitivity and 85.0% specificity10. That 4.68 is not on the formula's printed scale. With triglycerides 150 and glucose 90, ln(150 × 90 ÷ 2) gives 8.82, while half the log of the product, ln(150 × 90) ÷ 2, gives 4.76; a value as low as 4.68 exists only on the second scale. TyG cut-offs also vary from study to study1, so compare a TyG number only with a cut-off computed the same way.
When is a glucose tolerance test worth doing?
The OGTT shows what fasting numbers can't: how your body handles a load. You drink 75 g of glucose and blood is drawn at set times3. The ADA reads the 2-hour value. The International Diabetes Federation's 2024 position statement adds the 1-hour value: 155 mg/dL (8.6 mmol/L) or more in someone with normal glucose tolerance marks intermediate hyperglycemia, and 209 mg/dL (11.6 mmol/L) or more is diagnostic of type 2 diabetes. It calls the 1-hour value a stronger predictor of diabetes than fasting glucose, HbA1c or the 2-hour value, and it is reached earlier12.
Ask for insulin at each draw and the same test yields the Matsuda index, which correlated at 0.73 with the clamp in 153 adults13. The clamp itself holds insulin at about 100 µU/mL by drip while glucose is infused to keep blood sugar steady; the glucose rate needed is the measure of insulin sensitivity14. It needs two IV lines and trained staff, so it stays in research1.
How do you get tested?
- Ask for fasting glucose, HbA1c, fasting insulin and a lipid panel together, so HOMA-IR, the ratio and TyG all come from one draw.
- Fast for at least 8 hours before the draw: no calories, the ADA's definition of fasting3. HbA1c needs no fasting, but the others do.
- Note each unit on the report (mg/dL or mmol/L, µIU/mL or pmol/L) before you work out any index.
- Retest at the same lab when you compare over time, since insulin methods differ6.
- Expect a confirming test after an abnormal glucose result: unless the result is unmistakable, the ADA needs two abnormal results to diagnose diabetes, from the same test on two days or from two different tests, which may be drawn at the same time3.
What does your clinician do with the results?
The decisions follow the glucose diagnosis, not a HOMA-IR number. Normal results are repeated at least every 3 years from age 35; prediabetes is retested every year3. For adults with overweight at high risk, the ADA recommends a diabetes prevention program aiming for at least 5% to 7% weight loss and 150 minutes of moderate activity a week, and says metformin should be considered, especially at ages 25 to 59 with a BMI of 35 or more and higher glucose, or after gestational diabetes15.
The evidence behind that is strong. In the Diabetes Prevention Program, 3,234 adults with raised glucose were followed for an average of 2.8 years. The lifestyle program cut new diabetes by 58% and metformin by 31%, against placebo16. No drug other than metformin has long-term data for preventing diabetes on its own, the ADA adds, but it calls GLP-1-based therapy for weight management in people with overweight or obesity highly beneficial15. Insulin resistance markers help sort who is at risk early; a 2026 review finds no case yet for testing everyone, and no trial yet showing that care guided by them does better1.
Log each panel in PepVault's bloodwork log and your weigh-ins in Body, so the markers your clinician follows and the weight their plan aims to move are kept in one place, each as its own trend. Track your weight in PepVault
When to get help A fasting glucose of 126 mg/dL or more, or an HbA1c of 6.5% or more, is in the ADA's diabetes range: book a visit with your clinician rather than waiting for the next routine check3.
Sources
- García AH, De Sanctis JB, Celli Kubler GA et al. Before the storm: insulin resistance as a sentinel biomarker across the cardiometabolic continuum. A comprehensive narrative review. Front Endocrinol (Lausanne). 2026. PubMed 43054313
- National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). Insulin resistance and prediabetes. niddk.nih.gov
- American Diabetes Association Professional Practice Committee. 2. Diagnosis and classification of diabetes: Standards of Care in Diabetes-2026. Diabetes Care. 2026. PubMed 41358893
- Matthews DR, Hosker JP, Rudenski AS et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985. PubMed 3899825
- Gayoso-Diz P, Otero-González A, Rodriguez-Alvarez MX et al. Insulin resistance (HOMA-IR) cut-off values and the metabolic syndrome in a general adult population: effect of gender and age: EPIRCE cross-sectional study. BMC Endocr Disord. 2013. PubMed 24131857
- Miller WG, Thienpont LM, Van Uytfanghe K et al. Toward standardization of insulin immunoassays. Clin Chem. 2009. PubMed 19325009
- Knopp JL, Holder-Pearson L, Chase JG. Insulin units and conversion factors: a story of truth, boots, and faster half-truths. J Diabetes Sci Technol. 2019. PubMed 30318910
- McLaughlin T, Abbasi F, Cheal K et al. Use of metabolic markers to identify overweight individuals who are insulin resistant. Ann Intern Med. 2003. PubMed 14623617
- Sumner AE, Finley KB, Genovese DJ et al. Fasting triglyceride and the triglyceride-HDL cholesterol ratio are not markers of insulin resistance in African Americans. Arch Intern Med. 2005. PubMed 15983289
- Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M et al. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp. J Clin Endocrinol Metab. 2010. PubMed 20484475
- Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004. PubMed 15161807
- Bergman M, Manco M, Satman I et al. International Diabetes Federation position statement on the 1-hour post-load plasma glucose for the diagnosis of intermediate hyperglycaemia and type 2 diabetes. Diabetes Res Clin Pract. 2024. PubMed 38458916
- Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care. 1999. PubMed 10480510
- DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979. PubMed 382871
- American Diabetes Association Professional Practice Committee. 3. Prevention or delay of diabetes and associated comorbidities: Standards of Care in Diabetes-2026. Diabetes Care. 2026. PubMed 41358891
- Knowler WC, Barrett-Connor E, Fowler SE et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002. PubMed 11832527
Newsletter · Weekly
The weekly peptide briefing.
New guides, the week’s peptide news, one tool. Free, one email a week, unsubscribe anytime.