On 26 July 2026, a University of Michigan team published work that reframes a question people often ask after a routine blood panel: can you develop type 2 diabetes if your weight is normal? Their answer is yes, and the reason is that fat tissue can fail even when there is very little of it.
The study does not introduce a new test. What it changes is how to read the numbers already on your lab report. Below: what the researchers found, why a normal body weight does not rule out insulin resistance, and which markers deserve a second look. For the fundamentals of glucose and insulin testing, our guide to diabeteksen verikoetestit remains the reference.
What the Michigan study found
The researchers studied familial partial lipodystrophy type 2 (FPLD2), a rare inherited condition in which fat is lost from the limbs and redistributed elsewhere. Working with tissue donated by patients and a matching mouse model, they found that the fat cells were not simply missing. They were sick.
Three things went wrong at once. The gene activity that normally lets a fat cell take up and store lipids was switched off. The mitochondria, the tiny power plants inside each cell, stopped working properly. And both the fat cells and the immune cells living among them shifted into an inflamed state. The tissue then wasted away.
When healthy fat is lost, lipids have nowhere safe to go. They accumulate in the liver and the muscles instead, and the hormone signals that fat tissue normally sends break down. As lead investigator Elif Oral put it, type 2 diabetes is usually described as a disease of the insulin-producing beta cells in the pancreas, but it is also a disease of fat cells.
Why thin does not mean metabolically healthy
FPLD2 itself is rare. The principle it illustrates is not. Researchers describe what they call a personal fat threshold: each of us has a limited capacity to store fat safely under the skin, and that capacity varies enormously from one person to the next. Once the threshold is crossed, at any body weight, fat starts spilling into the liver, the pancreas and the bloodstream.
Someone with a low threshold can cross it while still looking slim. Someone with a high threshold may carry considerable weight and stay metabolically stable for decades. This is why a normal number on the scale is weak reassurance on its own, and why fatty liver can turn up in people who are not overweight. The scale measures how much tissue you carry. It says nothing about whether that tissue is doing its job.
Which blood markers actually flag insulin resistance
Here is the honest starting point, and it comes from the National Institute of Diabetes and Digestive and Kidney Diseases: there is no routine clinical test for insulin resistance itself. Direct measurement is a research procedure. What clinicians rely on instead is a group of ordinary markers that drift in a recognisable pattern, often years before blood sugar becomes abnormal.
| Merkki | What a shift can suggest | Reference points |
|---|---|---|
| Paastoverensokeri | The pancreas is no longer fully compensating | 100 to 125 mg/dL indicates prediabetes |
| HbA1c | Average blood sugar over roughly three months is creeping up | 5.7 to 6.4 percent indicates prediabetes |
| Triglyseridit | Fat is being pushed into the blood rather than stored safely | Often the earliest marker to move |
| HDL-kolesteroli | Tends to fall as triglycerides climb | Read together with triglycerides, not alone |
| Paastoinsuliini | The pancreas is working harder to hold glucose steady | Not on standard panels; must be requested |
| ALT | The liver may be storing fat it should not be storing | A slow, quiet rise matters more than one spike |
Two of these are best read as a pair. Triglycerides and HDL move in opposite directions once insulin resistance sets in, so a rising triglyceride level alongside a falling HDL-kolesteroli tells you more than either number on its own. A quietly climbing ALT points to a liver taking on fat it was never meant to hold.
The good news is that most of this is already in front of you. Glucose and liver enzymes sit on the same request form, so a kattava aineenvaihduntapaneeli combined with a lipid profile covers five of the six rows above. Only fasting insulin has to be asked for specifically.
Uusimmat tieteelliset edistysaskeleet
Because measuring insulin is awkward and expensive, researchers have spent the last few years testing simpler substitutes built from tests every lab already runs.
The one attracting the most attention combines fasting triglycerides and fasting glucose into a single number, known as the triglyceride-glucose index. Two 2025 studies, one in adults aged 18 to 60 and one in young adults, found that it tracks insulin resistance reasonably well and is far easier to obtain than an insulin-based calculation, which makes it useful where insulin testing is not readily available. In the same work, glycated albumin, another candidate, performed poorly for this particular purpose.
A word of caution is warranted, though, and it comes from the population this article is about. An eight-year follow-up of non-obese adults published in 2024 found that neither the triglyceride-glucose index nor the insulin-based calculation reliably predicted who would go on to develop type 2 diabetes in that group. In other words, the simple indices are useful screening signals, not verdicts, and they are weakest precisely in people of normal weight.
A separate long-term analysis of more than eight thousand middle-aged and older adults, published in Cardiovascular Diabetology, adds a practical nuance: insulin resistance and abdominal fat interact in ways that are not simply additive. Neither measure alone captures the risk, which is why clinicians look at waist size alongside the blood work rather than choosing between them.
What this changes for you
Nothing here means a slim person should start ordering extra tests. It means three specific things.
First, a normal weight is not, by itself, a reason to skip screening. The CDC recommends an A1C test from age 45, and earlier if you have other risk factors such as a family history of diabetes, high blood pressure, a history of gestational diabetes or polycystic ovary syndrome. Body weight is one risk factor among several, not the gatekeeper.
Second, look at the pattern, not the single line. One triglyceride value slightly above range is noise. Triglycerides rising, HDL falling and ALT drifting upward across two or three annual panels is a trend worth discussing with your doctor.
Third, remember what a normal result does and does not rule out. Because the pancreas compensates for years, fasting glucose ja HbA1c can both stay in range while insulin resistance is well established. If you have symptoms or a strong family history, say so rather than assuming a clean panel settles the question.
Usein kysytyt kysymykset
Can I ask my doctor for an insulin resistance test?
Not as a standard test, because the definitive measurement is a research procedure. What you can reasonably discuss is a fasting insulin level added to your usual fasting glucose, plus a lipid profile and liver enzymes, which together give a good practical picture.
I have a normal BMI. Should I be screened for diabetes?
Screening is driven by age and risk factors, not weight alone. Family history, ethnicity, blood pressure, a previous gestational diabetes diagnosis and polycystic ovary syndrome all count, and any of them can justify testing at a normal BMI.
Which single marker moves first?
There is no universal answer, but triglycerides and fasting insulin usually shift before fasting glucose or HbA1c, because the body compensates on the glucose side for a long time before the numbers give way.
Does this study mean losing weight is bad?
No. It describes a rare genetic disease in which fat tissue is destroyed by a faulty gene, not weight loss achieved through diet or exercise. The transferable lesson is about fat tissue quality, not quantity.
Sanasto
- Adipose tissue: body fat, understood as an active organ that stores energy and releases hormones, not as inert padding.
- Insulin resistance: a state in which cells respond poorly to insulin, so the pancreas has to produce more of it to keep blood sugar normal.
- Lipodystrophy: a group of rare conditions involving abnormal loss or distribution of body fat.
- Mitochondria: the structures inside cells that generate energy; when they fail, the whole cell struggles.
- Personal fat threshold: the individual limit beyond which a person can no longer store fat safely under the skin.
- Triglyceride-glucose index: a calculation combining fasting triglycerides and fasting glucose, used as a simple stand-in for insulin resistance.
Lähteet
- Michigan Medicine, University of Michigan: Why losing the wrong fat can trigger diabetes (26 July 2026)
- Maung JN et al., Altered lipid metabolism and inflammatory programs associate with adipocyte loss in familial partial lipodystrophy 2, Journal of Clinical Investigation
- National Institute of Diabetes and Digestive and Kidney Diseases: Insulin Resistance and Prediabetes
- Centers for Disease Control and Prevention: A1C Test for Diabetes and Prediabetes
- Yang Y et al., The interaction between triglyceride-glucose index and visceral adiposity in cardiovascular disease risk, Cardiovascular Diabetology (PubMed)
- Kundu SK et al., Triglyceride-Glucose Index Versus Glycated Albumin: Diagnostic Tools for Early Detection of Insulin Resistance and Pre-Diabetes (2025)
- Reynoso-Roa AS et al., Performance of the Triglyceride-Glucose Index for Early Detection of Insulin Resistance in Young Adults, Diabetology (2025)
- Jayashankar CA et al., Triglyceride-Fasting Glucose Index and HOMA-IR as Predictors of Type 2 Diabetes in Individuals With Normal Body Mass Index, Cureus (2024)
Lisälukemista
- Lipid panel explained: cholesterol, LDL, HDL and triglycerides
- Diabetes: symptoms, causes, diagnosis and treatment
- AST/ALT ratio: what it means and how to read it
- CRP blood test: how to read your inflammation results
- Advanced lipid panel: what it tests and who needs it
Interpret your lab results with AI DiagMe
Insulin resistance rarely shows up as one dramatic number. It shows up as a pattern across glucose, triglycerides, HDL and liver enzymes, spread over several pages of a report that is not written for patients. AI DiagMe turns those pages into plain language in a few minutes, so you arrive at your appointment knowing which lines to ask about.
Get your results explained in minutes
