METS-IR estimates how insulin resistant you are from four routine numbers — fasting glucose, fasting triglycerides, HDL cholesterol and body mass index — so it needs a fasting blood draw plus your height and weight, and no insulin assay at all. That last point is the reason the score exists. The reference method for insulin resistance is the euglycemic-hyperinsulinemic clamp, which takes hours in a research unit, and the usual shortcut, HOMA-IR, still needs a fasting insulin measurement that many clinics do not order and that assays report inconsistently.
METS-IR was published in 2018 and validated against clamp-measured insulin sensitivity. It is a continuous score, not a yes-or-no test.
What you need
- Fasting plasma glucose, drawn after roughly 8 to 12 hours without food.
- Fasting triglycerides, from the same draw.
- HDL cholesterol, which comes on the same standard lipid panel as the triglycerides.
- Your height and weight, to give body mass index.
- Consistent units. Glucose, triglycerides and HDL can each be reported in mg/dL or mmol/L, and mixing conventions gives a number that looks plausible and is wrong.
Nothing else is used: no fasting insulin, no C-peptide, no glucose tolerance test, no scan. This is a lab measurement, not a home one, but both tests are cheap, widely available and often drawn together at a routine check-up.
How the measurement works
Insulin resistance cannot be read off a single blood value, so the score infers it from downstream traces. When muscle and liver stop responding well to insulin, fasting glucose creeps up, the liver exports more triglycerides, HDL falls, and fat accumulates. METS-IR folds those signals into one number.
Each input pushes in a known direction. Higher fasting glucose raises the score, as do higher triglycerides and a higher body mass index; higher HDL lowers it. The glucose and triglyceride terms are compressed logarithmically before being combined with body mass index, so very high lipid values do not run away with the result, while body mass index enters more directly and carries much of the weight.
Doing this by hand invites unit errors, so put your four values into the METS-IR: metabolic score for insulin resistance and let it do the conversions and the arithmetic. If you do not know your body mass index, the Body Mass Index (BMI) gives you that input first.
How to read the result
The output is a unitless number: higher means less insulin sensitive, lower means more. Beyond that, be careful, because there is no single internationally agreed threshold. The cut-offs in the literature were derived by ROC analysis inside particular cohorts and differ by population, by sex, and by whether the study predicted insulin resistance now or diabetes later. A decimal quoted without its source is not a threshold to act on.
Read it as a direction of travel. The most informative comparison is with your own earlier score, measured the same way, fasting, at the same laboratory. A score drifting upward over a year says something; a single value near someone else’s borderline says much less. If your real question is how likely you are to develop type 2 diabetes, a dedicated tool such as Risk of developing diabetes asks for age, family history and other factors METS-IR never sees.
What it does not tell you
METS-IR does not measure insulin. It is a statistical stand-in for a clamp study, and it cannot say where the resistance sits, since liver, muscle and fat tissue behave differently and one number cannot separate them. It is also not a diagnosis: diabetes and prediabetes are defined by fasting glucose, HbA1c or a glucose tolerance test, and a reassuring METS-IR rules out neither.
It also inherits the weaknesses of its inputs. Body mass index misreads muscular people as higher-risk and older adults with little muscle as lower-risk, and its relationship to body fat varies by ancestry. Triglycerides move with a late meal, alcohol the night before, acute illness or untreated thyroid disease. Fibrates, high-dose omega-3 and statins change lipids directly, so treatment can pull the score down without changing how your tissues handle insulin.
Note who it was built in. The score was derived and first validated in Mexican adults and later tested in East Asian and European cohorts; it holds up reasonably, but thresholds do not transfer. It has not been established for children, for pregnancy, or for people with type 1 diabetes or on insulin.
Related reading
Disclaimer. This article is for information only and does not replace medical advice. Talk to a qualified clinician before changing anything about your health.