Your fasting glucose came back at 92 mg/dL. Your HbA1c is 5.3%. The lab flagged nothing, the doctor said "looks great," and you went home. Meanwhile, your pancreas may have been quietly working overtime for years to produce those normal-looking numbers.
That hidden effort has a name — insulin resistance — and there is a cheap blood test that catches it long before glucose budges. It is called fasting insulin, and most standard panels leave it out.
Why normal glucose can hide an abnormal metabolism
Insulin is the hormone that moves glucose out of your blood and into muscle, liver, and fat cells. When those cells stop responding well, your pancreas does the obvious thing: it makes more insulin. More insulin keeps glucose in range. So glucose stays normal — not because your metabolism is healthy, but because your beta cells are compensating.
That compensation can run for years, sometimes a decade or more. Fasting glucose and HbA1c only start to drift once the pancreas can no longer keep up. By the time those two markers look bad, a lot of the damage is already done. Measuring insulin itself skips the wait.
What fasting insulin and HOMA-IR actually measure
Fasting insulin is a single number: how much insulin is circulating after 8–12 hours without food. Reported in μIU/mL (sometimes pmol/L).
HOMA-IR combines insulin with glucose into one score. The formula:
- HOMA-IR = (fasting insulin in μIU/mL × fasting glucose in mg/dL) ÷ 405
- If your glucose is in mmol/L, divide by 22.5 instead of 405.
An example: insulin 12, glucose 95. That is (12 × 95) ÷ 405 = 2.8. Both inputs would pass a standard lab reference range. The combined score would not.
How to read the numbers
Reference ranges here are worth treating with care. Lab "normal" ranges for fasting insulin often stretch to 25 μIU/mL, because they are built from a population in which insulin resistance is common. Normal is not the same as optimal. The ranges used across metabolic-health practice and the research literature look roughly like this:
- Fasting insulin under 5–7 μIU/mL — good insulin sensitivity.
- 7–12 μIU/mL — a grey zone. Worth watching and re-testing.
- Above 12–15 μIU/mL — your pancreas is working hard.
And for HOMA-IR:
- Under 1.0 — optimal.
- 1.0–1.9 — normal by most references; mild resistance possible at the top of that band.
- 2.0–2.9 — early insulin resistance.
- 3.0 and above — the threshold most research studies use for significant insulin resistance.
An honest caveat: HOMA-IR cut-offs are not internationally standardised. They shift by population, age, sex, and — importantly — by which insulin assay the lab runs. A 15-year prospective study in a Chinese cohort landed on different optimal cut-offs than the values commonly quoted in US practice. This is a reason to track your own trend on one lab rather than obsess over crossing a line.
What the outcome data shows — and what it doesn't
Here is where a lot of longevity content overstates the case, so let's be precise.
The strong claim, well supported: insulin resistance measured by HOMA-IR is independently associated with higher cardiovascular and all-cause mortality in adults without diabetes. A meta-analysis of non-diabetic adults found HOMA-IR predicted mortality even after adjusting for the usual risk factors. That is a meaningful signal, because these were people whose glucose panels looked fine.
The weaker claim: fasting insulin on its own is a less consistent predictor than HOMA-IR. The same meta-analysis found the association held for HOMA-IR but not reliably for fasting insulin alone. Older cohorts disagree with each other — the Paris Prospective Study found fasting insulin independently predicted cardiovascular death over 11 years, while the Rancho Bernardo Study found no increase in fatal cardiovascular disease from high insulin in older adults.
What this means for you: order both, and read the combined score. Fasting insulin alone is useful as an early warning and as something to trend, but HOMA-IR is the number with the better outcome evidence behind it. Anyone telling you a single insulin value predicts your lifespan is ahead of the data.
One more nuance worth knowing: fasting C-peptide, which reflects how much insulin your pancreas actually produced, has outperformed other measures for predicting cardiovascular and overall death in at least one large non-diabetic cohort. It is not yet a routine test, but it is a reasonable add-on if your insulin numbers are borderline and you want a second read.
How to get tested, and how to not waste the test
Ask for "fasting insulin" alongside your standard fasting glucose. Most labs run it for a modest fee, and many direct-to-consumer panels include it. Then protect the result:
- Fast properly. 10–12 hours, water only. A "small" coffee with milk can move insulin.
- Do not train hard the day before. A brutal session acutely changes insulin sensitivity in both directions depending on timing. Keep the 24 hours before the draw ordinary.
- Sleep normally. One night of short sleep measurably reduces insulin sensitivity. Testing after a red-eye gives you a number about your travel, not your metabolism.
- Use the same lab. Insulin assays vary between labs enough to make cross-lab comparison unreliable. Trend within one assay.
- Re-test in 3–6 months if you change something. Insulin responds faster than most markers.
What actually moves the number
Ranked by strength of evidence, not by how interesting it sounds.
Proven
Resistance training. A 2025 meta-analysis of randomised trials in middle-aged and older adults with type 2 diabetes found resistance training improved HOMA-IR by a mean difference of −1.15, alongside better fasting glucose, HbA1c, muscle mass, and strength. Subgroup analysis pointed to higher-intensity programmes run longer than 12 weeks for the largest HOMA-IR effect. Muscle is your largest glucose sink; more of it, used harder, changes the maths.
Aerobic training — with enough weeks behind it. A 2025 randomised trial in young women found eight weeks of aerobic exercise improved insulin sensitivity, but four weeks did not. Adaptations here are real but not fast.
Combined training plus modest calorie restriction. In a 12-week randomised trial in women with obesity, concurrent strength and aerobic training with calorie restriction beat aerobic training alone on HOMA-IR.
Walking after meals. Meta-analyses of postprandial exercise show clear reductions in post-meal glucose, with larger 24-hour effects at longer durations. Light walking that breaks up sitting beats standing. Ten to fifteen minutes after your largest meal is the cheapest intervention on this list.
Losing visceral fat, and sleeping enough. Both act directly on insulin sensitivity, and both are consistently supported.
Promising but not settled
Berberine and myo-inositol both have randomised trials showing improvements in insulin sensitivity markers, but the trials are mostly small, short, and heavily weighted toward specific groups (inositol in PCOS in particular). Treat them as reasonable experiments to run and measure, not as substitutes for training. Berberine also interacts with a long list of medications.
Overstated
Chasing every CGM spike. Continuous glucose monitors are genuinely useful for seeing how you respond to specific meals. But a healthy person's glucose is supposed to rise after eating. Flattening every curve is not a validated goal, and glucose variability in non-diabetic people has not been shown to predict outcomes the way HOMA-IR does. Use a CGM to find your worst three meals, then go back to training.
Supplement stacks marketed as "insulin mimetics." Almost none have outcome data. Some have a signal on a surrogate marker in one small trial. That is a starting point for an experiment, not a reason to spend money every month indefinitely.
Where this fits in your panel
Fasting insulin and HOMA-IR are early-warning markers. They pair naturally with three others: HbA1c (your 3-month glucose average), triglyceride-to-HDL ratio (a rough, free proxy for insulin resistance already sitting in your lipid panel), and hs-CRP (inflammation, which travels with metabolic dysfunction). Read together, those four tell a much clearer story than any one of them alone.
This is educational content, not medical advice — discuss your results and any changes to training, supplements, or medication with a qualified clinician who knows your history.
The practical takeaway is unglamorous: order the test, note where you sit, change one thing that has real evidence behind it, and re-test in three to six months. Insulin is one of the few markers that will show you an answer that quickly.
VitalNexa tracks fasting insulin, HOMA-IR, HbA1c, and your full lipid and inflammation panel over time, folds them into your VitalNexa Biological Age, and syncs your Oura, WHOOP, or CGM data alongside so you can see which changes actually moved the number. Get started free and find out what your metabolism has been hiding.