Insulin Resistance Before Diabetes: Reading the Warning Signs

Drew, I had a patient this week — 46, fasting glucose 94, HbA1c 5.5, LDL fine. Everything in range. But his triglycerides have crept from 88 to 164 over four years, HDL is down, ALT is drifting up, and his waist is bigger. He asked if anything needs doing. I said probably not. Was I wrong?

You were wrong. That panel is describing compensated insulin resistance — a process that's been running, by the biology, probably for a decade.

But glucose is normal.

Glucose is the last thing to break. The beta cell compensates by secreting more insulin, and it can hold glucose in range for years. By the time fasting glucose crosses 126 or HbA1c crosses 6.5, you've already lost a substantial fraction of beta-cell function. The diagnosis arrives after the disease.

So what's actually happening earlier?

Subcutaneous fat — the designated safe storage depot — has reached that individual's personal capacity. Lipid then accumulates in the liver, skeletal muscle, and visceral compartment. Those sites are not inert: intrahepatic and intramyocellular lipid intermediates directly impair insulin signalling. Visceral adipose tissue also releases free fatty acids into the portal circulation. Insulin resistance follows, the beta cell compensates, and glucose looks fine.

What actually moves on labs during this compensated phase?

Fasting insulin rises — HOMA-IR, which is fasting insulin times fasting glucose divided by 405, above 2.5 is a widely used threshold, though it's not a guideline-endorsed diagnostic cut-off. Triglycerides rise, HDL falls, ALT drifts up reflecting hepatic fat. That's your patient's trajectory exactly.

How do I stage the liver without a biopsy?

FIB-4 — a calculated score using age, ALT, AST and platelet count — is guideline-endorsed by AASLD for fibrosis staging. Below 1.3 is low risk; above 2.67 warrants referral. For steatosis, hepatic ultrasound has limited sensitivity below thirty percent fat; vibration-controlled transient elastography with a controlled attenuation parameter is more sensitive and is guideline-supported for quantifying both steatosis and stiffness.

How strong is the evidence that treating this early actually changes outcomes?

For lifestyle — structured diet producing five to ten percent weight loss and one hundred fifty minutes of weekly moderate exercise — the Diabetes Prevention Program showed a fifty-eight percent reduction in progression to diabetes versus placebo. That's randomised trial data with clinical endpoints. Metformin reduced progression by thirty-one percent in the same trial, in higher-risk individuals.

And the newer agents — GLP-1 receptor agonists, SGLT2 inhibitors — what's the evidence for using them here?

Semaglutide has cardiovascular outcomes trial data in obesity without diabetes — SELECT trial — but it isn't approved or guideline-recommended for metabolic risk reduction in the compensated, normoglycaemic patient. You can tell a patient what the trial administered; you can't prescribe for that indication without acknowledging that's off-label and evidence is incomplete for this population.

What do I actually tell this patient?

Tell him his labs are describing a process, not a snapshot. Calculate his HOMA-IR, order a FIB-4, get a fasting insulin if you haven't. Then lead with lifestyle — it has the strongest evidence and the lowest risk. Sequence pharmacology after that if lifestyle is insufficient or he's high-risk.

If you had to give me one thing to hold onto from this.

Normal glucose is not normal metabolism. The window where almost everything is reversible is the decade before the diagnosis — and that's exactly where your patient is right now.