Insulin-resistance tests: fasting insulin, HOMA-IR, A1C, glucose, and CGM
There is no single universal routine-clinical cutoff that makes fasting insulin, HOMA-IR, A1C, glucose, OGTT, or CGM a standalone diagnosis of insulin resistance. They measure different parts of insulin and glucose physiology under different protocols.
Fasting insulin, HOMA-IR, A1C, fasting glucose, an oral glucose tolerance test, and continuous glucose monitoring do not measure the same thing. Validated glucose and A1C criteria are used to diagnose diabetes and prediabetes in defined contexts; fasting insulin and HOMA-IR estimate insulin dynamics but lack one universal clinical assay or cutoff; CGM measures interstitial glucose patterns, not insulin. No single wellness score independently establishes “insulin resistance” for every person.123
The most useful program starts with the decision a test is meant to change and names who interprets discordant results.
Put each measurement in its lane
| Test or estimate | What it measures | What it does not establish alone |
|---|---|---|
| Fasting plasma glucose | Blood glucose after a defined fast | Insulin level, daylong pattern or cause of an abnormal value |
| A1C | Glycation reflecting an approximate multiweek glucose average | Short-term swings or valid interpretation in every red-cell condition |
| Oral glucose tolerance test | Plasma glucose response at defined times after a standardized glucose load | Usual mixed-meal response or insulin resistance as one universal index |
| Fasting insulin | Immunoassay result for circulating insulin under stated fasting/preanalytic conditions | A device-independent universal normal range |
| HOMA-IR | Formula using fasting insulin and glucose as a model-based estimate | Direct clamp measurement or a universal diagnostic cutoff |
| CGM | Interstitial glucose estimates over time from a product-specific sensor | Insulin concentration, insulin action or a standalone diagnosis |
A panel can contain all six and still lack a coherent question.
Diabetes criteria and insulin-resistance estimates are not interchangeable
ADA standards use A1C and plasma-glucose criteria under defined conditions for diabetes and prediabetes diagnosis, often with confirmation when hyperglycemia is not unequivocal.1 NIDDK explains the fasting, random and oral-tolerance tests and factors that can affect them.2
Those diagnostic thresholds should not be relabeled as direct measurements of insulin resistance. A person can have compensatory high insulin while glucose remains within a reference range; another can have altered glucose for reasons not summarized by fasting insulin. Clinical context matters.
Conversely, a proprietary HOMA-IR flag does not replace validated assessment for diabetes, medication effects, pancreatic function, pregnancy-related testing or another condition.
HOMA-IR is a formula built on two assays
The common HOMA-IR calculation uses fasting glucose and fasting insulin. The result inherits every issue in both inputs: fasting duration, recent illness or exercise, collection and handling, glucose method, insulin assay, units and biological variation.
Insulin methods are not fully standardized across platforms, and published cutoffs differ by population, laboratory, assay, age, pubertal status, body composition and research purpose.34 A dashboard can display two decimal places without creating that precision in the underlying estimate.
Ask the laboratory and interpreting clinician for:
- specimen and draw conditions;
- glucose and insulin methods;
- units and formula version;
- reference or decision population;
- whether the value is validated for this clinical use;
- repeatability and action threshold; and
- what other findings must agree before the result changes care.
Do not compare a HOMA-IR number from one lab with an app cutoff built from another assay.
A1C is a time-integrated glucose marker with exceptions
A1C is convenient because fasting is not required, but red-cell lifespan, hemoglobin variants, anemia, blood loss or transfusion, kidney disease, pregnancy and some medicines or conditions can affect interpretation. A value is not “more accurate” merely because it summarizes more time.
Fasting glucose and OGTT capture different moments. The OGTT applies a standardized challenge and timed measurements, making it more sensitive to some patterns while requiring more preparation and time. A mixed meal, home glucose curve and OGTT are not interchangeable challenges.
If results conflict, ask who reviews preanalytics, assay limits and the need for confirmation rather than averaging the numbers.
CGM reveals patterns but does not measure insulin
CGM sensors estimate glucose in interstitial fluid with a lag and product-specific performance. They can show time patterns, post-meal excursions, overnight trends and response to activity or sleep. Those observations can generate hypotheses; they do not identify insulin concentration or prove why glucose changed.
FDA authorization is specific to the sensor, population, intended use, wear duration, calibration and labeling.5 An OTC wellness use cannot borrow diagnostic claims from a prescription system. The OTC-CGM guide covers sensor accuracy, compression lows, lag and result ownership.
Avoid informal “glucose-spike” rules without meal composition, starting value, sensor uncertainty, timing and validated endpoint. Lowering every visible rise is not a universal health target.
Reference ranges, percentiles and decision limits answer different questions
A laboratory reference interval describes results in a selected reference population. A research percentile ranks within a study. A diagnostic threshold is linked to evidence and a defined clinical decision. A coaching target may be proprietary. Label each one.
Terms such as “optimal insulin,” “metabolic age,” “insulin-resistance score” or “early prediabetes” can combine data in undocumented ways. Request the formula, source population, validation, units, uncertainty and action supported by the score.
The wellness-panel guide explains why more analytes create more opportunities for isolated flags.
The result needs an owner and a confirmation path
Before testing, decide who will:
- review medicines, illness, pregnancy and fasting conditions;
- interpret results with symptoms and history;
- resolve conflicting A1C, glucose, insulin or CGM patterns;
- order confirmatory testing when indicated;
- communicate urgent or unexpected findings;
- coordinate with primary, endocrine or weight-management care; and
- store and transfer the complete report.
A retail program that sells the panel but cannot manage an abnormal result has sold data without a care pathway.
- Name the decision State whether the question is diabetes screening, glucose pattern, research estimate, medication monitoring or another defined job.
- Verify the measurement Record specimen, fasting/challenge protocol, assay, units, sensor, software and preanalytic conditions.
- Label the threshold Distinguish diagnostic criterion, laboratory reference, research percentile and proprietary target.
- Keep metrics separate Do not treat A1C, glucose, insulin, HOMA-IR and CGM as interchangeable confirmations.
- Assign follow-up Name the clinician who resolves discordance, confirms findings and connects results to an appropriate plan.
Ask what a number can change before buying it
The decisive question is: “What does this exact assay or sensor measure, which validated threshold applies to my clinical question, and who owns confirmation when the results disagree?” A colorful insulin-resistance score is not a diagnosis by itself.
Sources
- American Diabetes Association. Standards of Care in Diabetes—2026: Diagnosis and Classification. Current professional criteria for diagnosing diabetes and prediabetes with plasma glucose and A1C; not a universal HOMA-IR cutoff. Accessed .
- National Institute of Diabetes and Digestive and Kidney Diseases. Diabetes tests and diagnosis. Federal overview of A1C, fasting plasma glucose, oral glucose tolerance and random glucose tests, confirmation and factors affecting interpretation. Accessed .
- PubMed. The current status of serum insulin measurements and the need for standardization. Current review of assay variability, reference intervals, preanalytics and why fasting-insulin or HOMA thresholds do not transfer universally. Accessed .
- PubMed. Standardization of insulin immunoassays: report of the American Diabetes Association Workgroup. Foundational professional evidence on lack of insulin-assay comparability and limits of universal decision points. Accessed .
- Food and Drug Administration. Medical devices that incorporate sensor-based digital health technology. Current FDA inventory and product-specific records for sensor-based devices, including continuous glucose monitors; used to distinguish interstitial glucose trends from insulin measurement or a standalone insulin-resistance diagnosis. Accessed .