What Insulin Resistance Actually Is, and How It Is Assessed
Quick answer
Insulin resistance is a physiological mechanism rather than a diagnosis, and the UK has no routine clinical test that confirms or excludes it. Treat an "insulin resistance score" as one piece of information, not a label, and have it read alongside your HbA1c, lipids, liver tests and waist measurement before any treatment is started on the strength of it.
Insulin resistance is a mechanism, not a diagnosis. It describes a state in which the liver, skeletal muscle and fat tissue respond less readily to insulin than they should, so more insulin is needed to achieve the same effect on glucose. That is a real and measurable physiological phenomenon in research settings. It is not something a UK laboratory can report back as a yes or no.
This matters because insulin resistance is now sold widely as a diagnosis. Private panels offer an "insulin resistance score", and patients arrive having been told they have a condition, often with a number attached and a treatment recommendation to match. The underlying biology is genuine. The certainty attached to the number usually is not.
What follows is an account of what the term actually means, why there is no routine clinical test for it, what HOMA-IR can and cannot tell you, and what clinicians use in practice when the question is whether someone is at metabolic risk.
What insulin resistance means physiologically
Insulin has several jobs. In muscle and fat it promotes glucose uptake from the blood. In the liver it suppresses the production of new glucose. In adipose tissue it restrains the release of fatty acids. Insulin resistance means that one or more of these actions is blunted.
The pancreas responds by secreting more insulin. For a time this compensation works, and blood glucose stays within the normal range despite insulin levels being high. This is why insulin resistance and raised glucose are not the same thing, and why a normal HbA1c does not exclude insulin resistance.
Resistance is also not uniform across tissues. Hepatic and peripheral insulin resistance can differ within the same person, and a single number cannot distinguish between them. That is one reason why summary indices are harder to interpret than they first appear.
Why there is no routine clinical test
The reference method is the hyperinsulinaemic euglycaemic clamp, in which insulin and glucose are infused under controlled conditions and the glucose required to maintain a steady level is measured. It is accurate and it is also expensive, lengthy and dependent on specialist infrastructure, so its use is largely confined to research.
There is a second obstacle, which gets less attention. Insulin assays are not interchangeable between laboratories. A 2025 comparison of twelve commercial immunoassays against isotope dilution mass spectrometry found substantial between-method differences, with several assays overestimating insulin at low concentrations and underestimating it at high ones. Repeatability within a single assay was excellent; agreement between assays was not. A fasting insulin result therefore carries meaning relative to the laboratory that produced it, and any index calculated from it inherits that problem.
What HOMA-IR is, and what it cannot do
HOMA-IR is the homeostasis model assessment of insulin resistance. It is calculated from fasting glucose and fasting insulin, conventionally as fasting glucose in mmol/l multiplied by fasting insulin in mIU/l, divided by 22.5.
It was designed as an epidemiological tool and it performs reasonably well in that role, distinguishing groups in population studies. Applied to one person in a clinic room, its limitations become more visible:
- There is no universally accepted cut-off. Reported thresholds in the literature range from roughly 1.7 to 3.9, and they vary with population and laboratory method.
- It reflects the fasting state only, so it says nothing about how glucose and insulin behave after eating.
- The same value can arise from different underlying states. Because it depends on fasting insulin, a given result may reflect greater beta-cell secretion or reduced insulin clearance, which are not the same problem.
- It has been shown to disagree with clamp measurements in specific settings, including comparisons between ethnic groups and before and after some drug treatments.
None of this makes HOMA-IR useless. It makes it a rough indicator rather than a diagnosis, and it does not support treating a single value as a threshold that has been crossed.
What clinicians actually use as proxies
UK practice does not depend on measuring insulin. NICE guidance on identifying people at high risk of type 2 diabetes uses a two-stage approach: a validated risk assessment first, then a blood test where indicated, using HbA1c or fasting plasma glucose. The markers clinicians weigh alongside that are mostly ordinary ones:
- Central adiposity. NICE advises keeping waist measurement to less than half of height, used alongside BMI in adults with a BMI below 35.
- HbA1c and fasting glucose, which show whether compensation is beginning to fail.
- Triglycerides and HDL cholesterol, since the pattern of high triglycerides with low HDL tends to track with insulin resistance.
- Liver enzymes and evidence of hepatic steatosis, which commonly accompany it.
- Blood pressure, and in women features suggesting polycystic ovary syndrome.
- Occasionally clinical signs, such as acanthosis nigricans, the velvety darkening of skin in flexures that can accompany marked insulin resistance.
Composite indices such as the triglyceride-glucose index exist and correlate with clamp measurements in research. They do not currently have agreed UK diagnostic thresholds, and we would not use one to label someone.
Mechanism, not diagnosis
The practical consequence is this. Insulin resistance is worth understanding because it explains why weight is harder to lose for some people, why PCOS behaves as it does, and why fatty liver and type 2 diabetes cluster together. It is not worth paying for a number that cannot be compared between laboratories and has no agreed threshold.
If you have been given an insulin resistance result, bring it. We will read it alongside your HbA1c, lipids, liver tests, waist measurement, medicines and history, which together say considerably more than the index does alone. And if the proposed treatment depends entirely on that one figure, that is a reasonable thing to question.
Sources & further reading
- NICE PH38: type 2 diabetes prevention in people at high risk — glossary and risk thresholds
- Rohlfing et al., Clinical Chemistry and Laboratory Medicine (2025): the current status of serum insulin measurements and the need for standardization
- Reaven, Diabetologia (2013): what do we learn from measurements of HOMA-IR?
- Rodzeń et al., Nutrients (2026): insulin resistance — current state of knowledge and clinical implications
- NICE: keep the size of your waist to less than half of your height
Reviewed by
Dr Chinnadorai Rajeswaran
Specialist Endocrinologist · MBBS MRCP FRCP
Last Reviewed: November 2026
Read about diabetes and related care.
Related insights
Type 2 Diabetes and Weight: What Remission Really Means
Type 2 diabetes remission means glucose levels below the diabetes range without glucose-lowering medication for a sustained period; it is not a cure, and diabetes can return. Weight loss can support remission for some people, especially earlier in the condition, but medication changes and ongoing HbA1c monitoring should be clinician-supervised.
Do artificial sweeteners affect insulin sensitivity?
Evidence does not show that non-sugar sweeteners as a group uniformly impair insulin sensitivity or cause obesity; effects and long-term evidence vary. Replacing sugary drinks with low- or no-calorie alternatives may reduce sugar intake, but this is not a guaranteed health outcome and water is also an option.
Prediabetes: What the Diagnosis Means, and What Changes It
An HbA1c of 42 to 47 mmol/mol or a fasting plasma glucose of 5.5 to 6.9 mmol/l identifies raised risk of type 2 diabetes rather than a disease, and progression is not inevitable — in pooled cohort data, reversion to normal glucose levels was more common than progression. Ask your GP practice about referral to the NHS Diabetes Prevention Programme, and keep to annual HbA1c or fasting glucose testing rather than acting on one result alone.