Fasting Insulin: The Test That Can Pre-determine Diabetes
Not long ago a 21-year-old patient came to see me because she didn’t feel right. Nothing dramatic. Tired in a way that sleep wasn’t fixing, and her monthly cycle had gone irregular. We ran a full panel, fasting insulin included, and everything on it was gorgeous. Fasting glucose, hemoglobin A1c at 5.1, LDL cholesterol, liver enzymes, inflammatory markers. Spot on, all of it.
One marker was different. Her fasting insulin came back at 9.1. The laboratory’s reference range for that test is 3 to 24, so 9.1 sits on the low side of normal. Nothing gets flagged. Nobody calls you about a 9.1.
That number is the reason we kept looking. We ran a test that tracks glucose and insulin together over several hours after a sugar load, and it showed she was releasing three and four times more insulin than she should have needed to keep her blood sugar looking that good. Her panel wasn’t beautiful because she was metabolically healthy. It was beautiful because her pancreas was working overtime to make it look that way.
She’s 21. Under the way we currently screen people, nobody would have raised any of this with her until her early thirties, when the word prediabetes finally turned up on a report.
What is fasting insulin, and why isn’t it on a standard blood panel?
Fasting insulin measures how much insulin your pancreas is releasing after eight to twelve hours without food. It isn’t on a standard panel because standard panels exist to diagnose diabetes, and you don’t need insulin to diagnose diabetes. Glucose and A1c already do that job.
The difference between those tests is that glucose and A1c report the result, while fasting insulin tells you what your body had to spend to get there. Two people can both show a fasting glucose of 88, and one of them is producing a trickle of insulin to hold it there while the other is producing a flood. On the panel your doctor ordered, those two people are indistinguishable.
I want to be clear that none of this is obscure science and I am not the one doctor who happened to figure it out. There are outstanding endocrinologists who understand insulin far better than I do. What almost none of them do is apply that knowledge to somebody who looks healthy, because inside the disease model there is no reason to. My 21-year-old presented nothing that required investigating.
What is a normal fasting insulin level?
A typical lab reference range for fasting insulin runs from about 3 to 24 µIU/mL. That range is normal in the statistical sense: it describes the people who show up at the lab. It is a description of your neighbours, and it says nothing whatsoever about whether your neighbours are well.
Normal doesn’t necessarily mean healthy. It means average.
Picture yourself standing in line with a hundred other people waiting to get blood drawn. Look around at who is in that line. When researchers at the University of North Carolina applied the standard criteria to a nationally representative sample of American adults, only 12.2% qualified as metabolically healthy. That puts roughly 88% on the other side of it. Now go back to the line. Whatever counts as normal on your report is a description of the population it came from, and that population is not a well one.
So I read lab markers in three zones instead of two. Red is what the lab flags high or low, and red is where medicine starts paying attention. Green is the sweet spot, the number a body produces when it is doing its job easily. Between them sits a yellow band that nobody ever discusses with you. When a marker crosses out of green into yellow, that is where my hand starts waving. By the time it reaches red, you have had the problem for years.
I am not going to hand you one cutoff number and call it the border between health and disease, because that is not how the test gets used in practice. What I will tell you is what happened with her. A 9.1 sitting in the lower half of a 3 to 24 range does not get flagged by any lab in the country, and on paper there was nothing to discuss. She was also 21, she felt unwell, and her cycle had stopped behaving. That is what sent us looking.
Can your insulin be high while your A1c and glucose look normal?
Yes, and that is the usual sequence. Insulin rises first, and it rises precisely in order to keep glucose normal. Blood sugar only starts to drift once the pancreas can no longer compensate, and by that point the pattern has been running for a long time.
Look at how A1c actually gets read. I want to see 5.2 or below. If yours comes back 4.7 or 4.6, that is fine; lower is better. From 5.3 to 5.6 is the band I would call insulin resistance, and I promise you it will never come up at a routine appointment. Then at 5.7 you cross a line and receive a diagnosis of prediabetes.
I dislike that name. It sounds like an early warning, something caught in good time. It is not early. It is late.
The Whitehall II researchers followed more than 6,000 British civil servants, and when they looked back at the 505 who went on to develop type 2 diabetes, insulin sensitivity was already falling roughly five years ahead of the diagnosis. That study tracked fasting glucose and glucose tolerance testing, so it says nothing about A1c one way or the other. What I will tell you from thirty years of practice is that by the time an A1c reads 5.7, the pattern underneath it has usually been building for five to ten years.
Blood sugar and A1c are lagging indicators. You will not catch anything early by watching them. My 21-year-old had hyperinsulinemia and an A1c of 5.1, and there was no reason on earth to flag it. (If your A1c has climbed since you cut carbohydrates, that is a separate and much less alarming story, and I have written about why A1c can read higher on a low-carb diet.)
To be fair to your doctor: at an A1c of 5.4 there is nothing to prescribe. That is a correct assessment. Nobody puts a 5.4 on medication. But nothing to prescribe isn’t the same as nothing going on, and the guidelines only ask your doctor the first of those two questions.
What does a fasting insulin of 9 actually mean?
On its own, less than you would think. A 9 in a symptomatic 21-year-old means something quite different from a 9 in a 68-year-old who lifts three times a week. The number is a reason to keep looking. Think of it as a position on the map, and treat it that way.
What made 9.1 interesting in her case was everything sitting next to it. She felt unwell and her monthly cycle had become irregular, and those two things together are why I ran a second test instead of filing the panel under normal and moving on.
A mid-range fasting insulin also reads differently when it arrives alongside the signs that commonly travel with high insulin:
- Waist thickening while the scale barely moves.
- The three o’clock collapse, the one that needs food or caffeine to get through.
- Triglycerides drifting up while HDL drifts down, both still sitting comfortably inside their ranges.
- Skin tags, or darkened velvety skin at the back of the neck and under the arms.
- A family history you already know about and have been quietly waiting for.
If you want to go a step further with the number, fasting insulin feeds into HOMA-IR, a calculation combining fasting insulin and fasting glucose that has been used since 1985 to estimate insulin resistance. Any clinician can work it out from those two values. In her case the more revealing test was the dynamic one, watching what insulin actually did over several hours, because that is where the threefold and fourfold response showed itself. The fasting number simply told me to go and look.
How do you ask your doctor to run it?
Ask for a fasting insulin drawn at the same time as your fasting glucose, after eight to twelve hours with nothing but water. It is inexpensive, every major lab offers it, and it needs no preparation beyond the fast itself.
If it helps to have the sentence ready, use a specific one. “I’d like a fasting insulin added to my next draw, alongside fasting glucose, because I want to track my metabolic health over time instead of waiting for a diagnosis.” You’re asking for information about yourself. Nobody has to do anything with it.
You may be told it isn’t indicated. Under standard diabetes screening guidelines that is a perfectly accurate answer, because those guidelines exist to find diabetes and you don’t have diabetes. Across much of the US you can also order the test yourself through a direct-to-consumer lab if you would rather skip the negotiation.
Two practical points. Use the same lab and the same fasting window every time, because assays differ between labs and you want your own numbers comparable to each other. And get the first one before you change anything, so whatever you do next has a baseline to be measured against. A single number with nothing to compare it to is close to useless.
What actually brings fasting insulin down?
Nothing in a bottle. I use nutritional supplements in practice, sometimes very aggressively, and yes, there are supplements that help. But you don’t get your health from what’s inside a plastic container. You get it from your diet, your lifestyle, and the genetics you were born with, and insulin in particular answers to how you eat, how you move, how you sleep, and how much stress you are carrying.
Everything I have done clinically for thirty years sorts into two questions. What is interfering with my health, and what am I missing? If there is interference, remove it. If there is insufficiency, supply it. Almost every health question you have fits into one of those two buckets, and once it is in the right bucket you can finally act on it.
Think of your body as the city of New York. When all the traffic lights are working, traffic moves beautifully. Break one light and midtown backs up for thirty blocks in every direction. You can spend your energy sending cars down side streets, or you can go and fix the light. High insulin is a broken light, and a great deal of what gets prescribed downstream of it is car diversion.
So here is the assignment, and it is deliberately small. Write down one interference you know you could remove, and one insufficiency you know you could supply. You do not need me to identify either of them. Most people can name both in about ten seconds: the sixty-hour work weeks that mean no gym and no cooking, the vegetables you never eat, the vitamin E you are short on. The other vitamin E. Vitamin Exercise. Take those two actions for fourteen days and pay attention to what changes.
Sleep belongs in that list more often than people expect, and I have written separately about what poor sleep does to insulin and about the wider picture of insulin resistance. Then reassess and go around again. That loop is the entire method. Locate where you are, sort what you find into the two buckets, act, measure, repeat.
One caution about copying anybody’s protocol, mine included. The diet I follow is probably right for fewer than 20% of people. The diet my wife follows is probably right for 65 to 70% of them. We are all human and all fairly similar, and we still do not respond the same way to the same intervention. Which is exactly why you want a number you can track. Rules you’re told to obey don’t survive contact with an actual body.
My 21-year-old wasn’t a candidate for medication and wasn’t anywhere near a diagnosis. What she needed was to know where she was standing on the map, roughly ten years before the map would have told her by itself. That is all fasting insulin does. It is a cheap test that answers a question the standard panel never asks, and if you are going to add one marker to your next blood draw, make it that one.
This is general education and not medical advice about your particular situation. Bring any number you find to a clinician who knows your history, and don’t change anything you have been prescribed on the strength of a blog post.
References
- Araújo, J., Cai, J., & Stevens, J. (2019). Prevalence of Optimal Metabolic Health in American Adults: NHANES 2009–2016. Metabolic Syndrome and Related Disorders, 17(1), 46–52. https://doi.org/10.1089/met.2018.0105
- Tabák, A. G., Jokela, M., Akbaraly, T. N., Brunner, E. J., Kivimäki, M., & Witte, D. R. (2009). Trajectories of glycaemia, insulin sensitivity, and insulin secretion before diagnosis of type 2 diabetes: an analysis from the Whitehall II study. The Lancet, 373(9682), 2215–2221. https://doi.org/10.1016/S0140-6736(09)60619-X
- Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., & Turner, R. C. (1985). Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia, 28(7), 412–419. https://doi.org/10.1007/BF00280883
Want to go deeper on the markers that actually predict how the next decade goes? Start with the health markers worth tracking for longevity, or browse the rest of the blog.
