Insulin resistance
When muscle, liver and fat respond weakly to insulin, the pancreas makes more to keep glucose normal. It can run for years before glucose itself rises.
Insulin moves glucose out of the blood into muscle and fat, and tells the liver to stop releasing glucose. In insulin resistance those tissues respond less than they should. The pancreas compensates by releasing more insulin, so blood glucose often stays normal for a long time. The first lab sign is frequently a high fasting insulin with a normal fasting glucose. When the pancreas can no longer keep up, glucose rises: first into the prediabetes range, then into type 2 diabetes.
Insulin resistance travels with excess visceral fat, fatty liver, high triglycerides, low HDL cholesterol and high blood pressure. Physical inactivity and poor sleep make it worse. Exercise and weight loss improve it.
How it is estimated
The research standard is the euglycaemic clamp, which infuses insulin and measures how much glucose is needed to hold blood sugar steady. It is not a routine test. In practice, insulin resistance is estimated from fasting blood:
- HOMA-IR. From the homeostasis model assessment published by Matthews and colleagues in 1985: fasting glucose (mmol/L) times fasting insulin (microunits per mL), divided by 22.5. In US units, glucose in mg/dL times insulin, divided by 405. In that paper it correlated well with the clamp (Rs 0.88), but its coefficient of variation was 31 percent, so a single value is imprecise.
- Triglycerides and HDL. High triglycerides with low HDL is a common pattern in insulin resistance, and both are on a standard lipid panel.
- Fasting glucose and HbA1c. These show when compensation is failing.
Both values must come from the same fasting draw. The ADA defines fasting as no caloric intake for at least 8 hours, and insulin in particular rises sharply after food, so a non-fasting sample makes HOMA-IR meaningless.
There is no single agreed HOMA-IR cut-off; laboratories and studies use different ones, and insulin assays differ between labs. Tracking the same test at the same lab over time is more useful than comparing one value with a threshold from elsewhere.
The diagnostic thresholds for glucose
| Test | Normal | Prediabetes | Diabetes |
|---|---|---|---|
| Fasting glucose | Below 100 mg/dL | 100 to 125 mg/dL | 126 mg/dL or higher |
| HbA1c | Below 5.7% | 5.7% to 6.4% | 6.5% or higher |
These are the American Diabetes Association criteria. A diagnosis of diabetes normally needs confirmation on a repeat test.
What GLP-1 drugs do to it
GLP-1 receptor agonists improve glucose control directly and, through weight loss, improve insulin sensitivity. In the three-year SURMOUNT-1 analysis of 1,032 adults with prediabetes, type 2 diabetes was diagnosed in 1.3 percent on tirzepatide against 13.3 percent on placebo over 176 weeks, a hazard ratio of 0.07. The GLP-1 blood work page lists which markers tend to move and when.
The app's bloodwork tracking reads fasting glucose, fasting insulin, HbA1c, triglycerides and HDL from a lab PDF or manual entry, groups them in the metabolic panel with a 0 to 100 score, flags values worth taking to a physician promptly, and plots each against your dose timeline.
Related terms: surrogate endpoint, DEXA scan, pharmacodynamics, confidence interval. Back to the glossary.
Questions people ask
What is insulin resistance?
A state in which muscle, liver and fat respond weakly to insulin, so the pancreas has to release more to keep blood glucose normal.
How do you calculate HOMA-IR?
Fasting glucose in mmol/L times fasting insulin in microunits per mL, divided by 22.5. With glucose in mg/dL, divide by 405 instead.
What fasting glucose is prediabetes?
100 to 125 mg/dL, under American Diabetes Association criteria. An HbA1c of 5.7 to 6.4 percent is also in the prediabetes range.
Can insulin resistance exist with normal glucose?
Yes. The pancreas can compensate for years by releasing more insulin, so a high fasting insulin may appear before glucose rises.
Do GLP-1 drugs prevent type 2 diabetes?
In SURMOUNT-1 participants with prediabetes, 1.3 percent on tirzepatide and 13.3 percent on placebo developed type 2 diabetes over 176 weeks, a hazard ratio of 0.07.
Sources
- Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man (Matthews et al.), Diabetologia, 1985
- Diagnosis and classification of diabetes: Standards of Care in Diabetes 2026, Diabetes Care, American Diabetes Association, 2026
- Tirzepatide for obesity treatment and diabetes prevention (SURMOUNT-1, 176 weeks), New England Journal of Medicine, 2025
- Mechanisms of action and therapeutic application of glucagon-like peptide-1 (Drucker), Cell Metabolism, 2018
