Why doctors don't test fasting insulin

Your fasting glucose comes back at 88 mg/dL and your HbA1c at 5.4%. Both print “normal,” and your doctor moves on. Nowhere on that sheet is the one number that would have told you how hard your body is working to keep those two so calm: fasting insulin. It isn’t flagged, it isn’t borderline, it’s just absent. The lab never measured it, because the order form never asked, and almost no doctor will recommend it.

That’s not your doctor being lazy or hiding the ball. It’s the default in nearly every clinic, and the reasons are more boring and more defensible than the conspiracy you might suspect. (What fasting insulin actually is and why it’s one of the early-warning markers a standard panel skips lives in the biomarkers your doctor isn’t tracking; this one is the “why isn’t it standard” argument.)

Why don’t doctors recommend a fasting insulin test?

Diabetes gets diagnosed off glucose. The American Diabetes Association anchors the whole pathway to it: a fasting plasma glucose of 126 mg/dL or higher, an HbA1c of 6.5% or higher, or a 2-hour glucose of 200 mg/dL or higher on an oral glucose tolerance test. Prediabetes sits just below, HbA1c 5.7 to 6.4% and fasting glucose 100 to 125. Insulin appears nowhere in that list.

There’s a real reason for that, and it isn’t an oversight. Diabetes is defined by glucose because glucose is the thing that does the damage. High blood sugar is what scars vessels and wears out kidneys over years. The DCCT (1993) and UKPDS (1998), the trials that showed lowering glucose prevents that damage, were run on glucose and HbA1c. So glucose is both the agent of injury and the variable the evidence is built on. Insulin is the system straining to hold glucose down. Useful to know, but it isn’t the thing that blinds you or fails your kidneys.

Then there’s the cutoff problem. There is no guideline-endorsed insulin number that triggers a diagnosis or unlocks a covered treatment. A test result with no decision attached to it doesn’t earn a slot in a 15-minute visit. Say it honestly: fasting insulin is informative, not diagnostic. And reimbursement follows the diagnostic pathway. A test outside the guideline-defined workflow is frequently not covered, so it doesn’t get ordered by default. That’s the system paying to diagnose and treat defined disease, and on that narrow job, glucose and HbA1c genuinely do the work.

For most people whose glucose and HbA1c are clean and steady, the standard pathway is a reasonable bet. It comes apart in a specific group: the people whose insulin has been quietly compensating for years while every glucose reading stayed perfect.

Why fasting insulin has no standard cutoff

Here’s the quietest reason, and the most legitimate one. Fasting insulin assays are not standardized the way glucose is. The American Diabetes Association and the IFCC have flagged this for years: run the same blood sample through two different immunoassay platforms and you can get materially different insulin values back. A “normal” cutoff that means one thing at one lab can mean something else at another, because the labs aren’t measuring on a shared, harmonized scale.

This is the technical reason guideline bodies are reluctant to set a hard insulin threshold. You can’t write a universal cutoff onto a measurement that isn’t comparable across labs. It would mean one thing in Mumbai and another in Bangalore, and a guideline that fails that way is worse than no number at all.

It bites your own result too. A single insulin value, read in isolation against someone else’s reference range, is weak. The fix is comparability: run it on the same lab and the same platform each time, and read the direction your number moves, not the absolute figure measured against a stranger’s range. One number against the wrong yardstick can mislead you in either direction.

Why fasting insulin is still worth knowing

Hyperinsulinemia can precede measurable glucose dysfunction by years. Long before fasting glucose or HbA1c drift upward, the pancreas can be quietly raising insulin output to hold glucose in the normal band. The calm glucose is bought with rising insulin.

This is what the absent number would have shown. A fasting insulin in the high single digits to low teens (µIU/mL), against an optimal often cited around 2 to 5, under a flawless 88 mg/dL glucose is compensation: the pancreas working harder to hold the line. The glucose looks paid-up because insulin is covering the bill.

There’s a cleaner way to read the two together, and it has a name. Take your fasting insulin and your fasting glucose and combine them into a single estimate of how insulin-resistant you are: HOMA-IR, calculated as (fasting insulin µIU/mL × fasting glucose mg/dL) / 405. Neither value alone tells you much. A normal glucose held up by a high insulin and the same glucose on a low insulin are two very different metabolic states, and reading the pair together is what separates them. This is the move Depth is built on, putting two signals side by side to surface what neither shows alone.

Right-size it, though. This is an early signal, not a diagnosis, and it’s most useful read as a direction over repeated draws rather than as a single pass or fail. One insulin value is a hint; you confirm it by watching where the number goes over a year.

If you want fasting insulin tested, here’s the ask

Add fasting insulin alongside fasting glucose, both fasted, on the same lab each time, so HOMA-IR is computable from the pair. Asking for the two together is the whole ask. This is the kind of panel Depth puts in front of you by default, fasting insulin and glucose drawn at home, rolled into HOMA-IR, and read across your HbA1c, ApoB, and wearable signals, so getting early access is the simplest way to actually order it.

Read it next to the rest of the picture, not alone. Insulin beside glucose, HbA1c, ApoB, and triglycerides is where the metabolic story actually shows up.

And follow the slope. Two or three draws across a year tell you whether insulin is climbing under a steady glucose, the finding that matters, and the one the assay problem can’t ruin because you’re reading your own direction on your own lab. That’s the case trends beat thresholds makes in full.

Ordering it yourself is how a motivated person sees the thing the pathway was never built to catch: the years of quiet compensation that run underneath a glucose that still reads 88.