Metabolic Health and Glucose: How Insulin Resistance Begins

Bottom line

Insulin resistance usually begins many years before glucose rises, and the first tissue to fail is often muscle ²⁴. A fasting glucose, read alone, is often a late signal.

Truly favorable metabolic health is rarer than people think. On U.S. national nutrition-survey data for 2017–2018, only 6.8 percent of U.S. adults were in a favorable range on body fat, glucose, lipids, and blood pressure, and free of cardiovascular disease ²³.

The strongest levers are not in a pill bottle. In a randomized trial of more than three thousand people, a lifestyle intervention cut progression from prediabetes to type 2 diabetes by 58 percent. Metformin cut it by 31 percent ²⁵.

The views broadly agree in treating metabolic health as the foundation under other chronic diseases. They split on three things: whether to watch every glucose spike, how much meal timing matters, and how large a place supplements can take.

What it is

Insulin resistance means the body has become sluggish in its response to insulin.

Insulin, from the pancreas, lets muscle, liver, and fat cells take glucose in to use or store. When that response rusts, the pancreas secretes more. Glucose can still look acceptable while insulin has been high for a long time. That stage is hyperinsulinemia.

One account describes metabolic disease as a continuum: from hyperinsulinemia, to insulin resistance, to fatty liver, and on to type 2 diabetes ². It sums the picture up as a disorder in fuel processing ².

Metabolic syndrome is a cluster of signals from that process: higher blood pressure, higher triglycerides, higher glucose, abdominal obesity, and lower HDL cholesterol ².

The test clinicians use most often is HbA1c. It is the share of hemoglobin in red cells that has been "pickled" by sugar, and it reflects average glucose over the past two or three months. On the lines used by the U.S. Centers for Disease Control and Prevention (CDC), an HbA1c from 5.7 percent to 6.4 percent is prediabetes, and 6.5 percent and above is diabetes ²¹.

Why it matters

The CDC's latest figures show that in 2023 about 40.1 million people in the United States had diabetes, 12.0 percent of the population. About another 115.2 million adults had prediabetes. Among adults with diabetes, about 27.6 percent had not been diagnosed ²⁰.

A 2019 study, using newer guideline lines, found only 12.2 percent of U.S. adults metabolically healthy. Even among people whose body-mass index was not high, fewer than one third met the bar ²². Being lean is not the same thing.

The cost is mostly indirect. One comparison sets heart disease, cancer, and dementia as pillars standing on metabolic disease, which itself rarely kills directly ².

The liver is another early casualty. A systematic review of 92 studies estimates that fatty-liver prevalence in adults worldwide rose from about 25 percent in 1990–2006 to 38 percent in 2016–2019 ³⁰.

Why does the "beginning" often sit in muscle? Under insulin stimulation, about 80 percent of the glucose the body takes up enters skeletal muscle ²⁴. Muscle is the largest warehouse for glucose. A review by diabetes researcher Ralph DeFronzo argues that muscle insulin resistance is the first defect in type 2 diabetes, measurable decades before pancreatic beta cells (the cells that make insulin) fail and glucose rises in a clear way ²⁴. The same review notes that lean children of two parents with type 2 diabetes have about a 40 percent lifetime risk of the disease ²⁴.

Different perspectives

Main positions

One view puts metabolic health at the foundation of a longevity strategy and argues for finding trouble early. It lists five tools: nutrition, exercise, stress management, sleep, and drugs or supplements, and moves nutrition, sleep, and exercise first, before reaching for a drug ³. Carbohydrate tolerance differs by person and moves with training volume. A large carbohydrate intake can cause no glucose trouble during heavy training and still not work once training volume falls ³. Two weeks of four hours' sleep a night cut glucose handling by about half ³. A continuous glucose monitor, a sensor on the arm, can serve as a probe of a carbohydrate edge ³.

That view sets a high bar for nutrition evidence. A small study offered for the carbohydrate–insulin theory of obesity did not measure insulin, so the conclusion does not hold ⁴. A large observational link between meat and diabetes risk is not cause, and the extra risk roughly halved after body-mass index was adjusted ⁶. On the brain, a mouse study found that blocking the receptor for a molecule called IL-10 cut clogged capillaries in diabetic brains by 63 percent. That account notes there is no human evidence yet ⁵.

Another view comes at glucose from the brain and from behavior. Deeper metabolic-disease material is mostly from guests ⁷⁸. Wanting sugar has two paths: sweetness on the tongue, and a nutrient signal in the gut. Both trigger dopamine, the signal tied to wanting ⁹. A food's glycemic index, how fast it raises glucose, is usually measured eaten alone. With fiber or fat, the meal's index falls. Fructose does not press the hunger signal down as effectively as glucose ⁹. Sugar is not treated as uniformly bad: eating very sweet fruit and starch after high-intensity training can help refill muscle glycogen ⁹. Eating inside two hours of bedtime, that account holds, suppresses nighttime growth hormone, so glucose and insulin are better kept relatively low at sleep — and it grants that being too hungry makes sleep impossible ¹⁰.

Guests disagree. Casey Means cites 93.2 percent of U.S. adults not in a favorable metabolic state and argues for CGM as awareness ¹¹. She compares GLP-1 weight-loss drugs, which mimic a gut hormone and lower appetite, to a bandage that does not change the environment ¹¹. Robert Lustig thinks fructose has a distinct harm and that a calorie is not a calorie ¹². Layne Norton treats a single fasting glucose as a snapshot, weighs HbA1c and HOMA-IR more, and thinks early versus late eating matters little when calories match ¹³.

Another view worries more about glucose that stays high, plus sitting, than about one spike. The main tool described there is exercise that raises heart rate around meals. In people who rarely move, muscle is insensitive to glucose, so sugar is more easily taken into fat or left to injure vessels ¹⁷. After training, GLUT4 "doors" on the muscle surface stay more open for about 48 hours, so frequent trainers already spike less. That view holds that a heart-rate-raising move, such as bodyweight squats, beats a post-meal walk ¹⁷. A meta-analysis of eight crossover trials found post-meal exercise lowered glucose more than pre-meal exercise or none, and sooner was better ¹⁷.

A sugary drink has no food matrix, the fiber in fruit that slows absorption, so glucose rises fast ¹⁴. One estimate: about 133,000 diabetes deaths worldwide in 2010 were tied to sugary drinks ¹⁶. In the large prevention trial, exercise had about twice metformin's effect ¹⁸. The same site also notes that metformin may blunt the mitochondrial adaptation from exercise, and that any benefit for healthy people who train is unclear ¹⁸.

Berberine, a plant alkaloid, is studied mainly as an activator of AMPK, a cellular energy-sensing switch ¹⁵. In one trial it lowered fasting glucose 26 percent and HbA1c 18 percent, in metformin's range. Early diabetes trials were poor in quality, and berberine markedly changes body levels of metformin and of several other drugs ¹⁵. On time-restricted eating, compressing the day's food into a shorter window, that site treats several studies as pointing to metabolic-marker changes beyond weight loss ¹⁹.

Where views agree and differ

Agreement first. Muscle is the largest destination for glucose, and exercise is the core tool ³¹⁷. Sugary drinks are the easiest change to start ¹⁴¹². Sleep affects glucose directly ³¹⁰. That metabolic trouble precedes a glucose diagnosis is in DeFronzo's review, cited above ²⁴.

Disagreement one: whether to watch every spike. Casey Means treats a CGM as an awareness tool for ordinary people ¹¹. Another account uses one as a time-limited probe ³. A third says glucose that stays high is more worth worrying about than the spikes people talk about ¹⁷. One side weighs "people change what they can see." The other weighs harm from long exposure.

Disagreement two: how much meal timing matters. One site leans toward early time-restricted eating ¹⁹. In an interview, Norton thinks that when total calories and protein match, early versus late makes little difference ¹³. A 12-month randomized trial of 139 people found that, with calories restricted to the same degree, adding time-restricted eating brought no significant extra benefit on weight or metabolic markers ³¹.

Disagreement three: is fructose distinctly harmful? Lustig thinks yes ¹². A study in children swapped sugar in the diet for starch without changing calories, and after 9 days glucose tolerance and hyperinsulinemia had improved ²⁹. Another view is more cautious about weak evidence that blames a complex metabolic problem on a single nutrient ⁴⁶.

Disagreement four: where supplements sit. One account spends a lot of space on the berberine evidence ¹⁵. Another lists drugs and supplements last among the five tools ³.

PearHealth's synthesis

We take the continuum. Insulin resistance is a road that can be spotted early, so signals before fasting glucose, such as fasting insulin, are worth attention. Muscle and exercise come first: the firmest evidence is lifestyle and weight-loss trials ²⁵²⁶, and mechanism work also points to muscle as the main destination for glucose ²⁴.

We agree that sugary drinks are a high-leverage change. We also grant that most of the related evidence is observational ²⁷²⁸, and that it has to be read next to the confounding one analysis flags ⁶.

We hold three things back. First, whether healthy people wearing a CGM long term and chasing every spike improves long-term outcomes has no direct evidence yet. Second, an independent effect of time-restricted eating is unstable in the larger randomized trials ³¹. Third, berberine has trials showing short-term improvement in glucose markers ³², but trial quality is uneven and drug interactions exist ¹⁵. We do not treat it as the first step.

How strong is the evidence

"Lifestyle can prevent, and even reverse, early type 2 diabetes" is the firmest claim. "Exactly what to eat, and when" is much weaker.

The firmest layer is large randomized trials. The Diabetes Prevention Program enrolled 3,234 adults with prediabetes. Over a mean 2.8 years, a lifestyle intervention cut incidence by 58 percent and metformin by 31 percent ²⁵. In the British DiRECT trial, 46 percent of the intensive weight-management group were in remission at 12 months, against 4 percent of controls. Among people who lost 15 kilograms or more, 86 percent reached remission ²⁶.

The middle layer is large cohorts. A meta-analysis of more than 310,000 people found that those who drank the most sugary drinks (often one or two servings a day) had a 26 percent higher risk of type 2 diabetes than those who almost never drank them ²⁷. In the Framingham Heart Study, people who regularly drank sugary drinks over the long term had a 52 percent higher risk of high triglycerides ²⁸. Studies of this kind can show association. They cannot prove cause on their own.

Mechanism research is quite solid, including muscle data from the clamp method, a reference approach that infuses insulin and glucose together ²⁴. It answers how something happens, not what to do.

The weakest layer is animal research and small short trials. The IL-10 mechanism linking diabetes and dementia is, so far, only in mice ⁵. Berberine trials are mostly short and use glucose markers as the endpoint ¹⁵³². Time-restricted eating does not agree across trials ¹⁹³¹.

What you can do

The direction is plain. Make muscle work often, sleep enough, swap out sugary drinks, keep an eye on waist size, then use blood work to see whether the direction is right.

• Give muscle more work. Strength training and aerobic training both help. A common pattern is to stand and walk after a meal, or to do a few sets of bodyweight squats ¹⁷.
• Deal with liquid sugar first. Swapping sugary drinks for water, tea, or an unsweetened drink is a small change with a wide effect ¹⁴²⁷.
• Protect sleep. A run of short nights directly weakens the body's ability to handle glucose ³.
• Change how a meal is built. Refined carbohydrate eaten with vegetables, protein, and fat raises glucose more gently ⁹.
• Watch waist size, not only weight. Abdominal fat is one signal of metabolic syndrome ², and the amount of weight lost tracks clearly with the share of people who reach diabetes remission ²⁶.

Questions worth raising with a clinician:

1. Besides fasting glucose and HbA1c, is fasting insulin a fit, or a HOMA-IR calculation?
2. If a parent has type 2 diabetes, is an oral glucose tolerance test worth doing, to catch the problem earlier ²⁴?
3. If liver enzymes or triglycerides run high, does that point toward fatty liver and a further look ³⁰?
4. If a supplement such as berberine is under consideration, can it interact with drugs already in use, especially glucose-lowering drugs ¹⁵?
5. At the prediabetes stage, how do the benefits and harms of medication look in this case ²⁵?

Common misconceptions

Misconception one: if fasting glucose has not crossed the line, nothing is wrong. Insulin resistance can exist for decades before glucose rises ²⁴, and a large share of people with diabetes have not been diagnosed ²⁰.

Misconception two: a lean person is metabolically healthy. Among U.S. adults whose body-mass index is not high, fewer than one third fully meet the metabolic bar ²².

Misconception three: every glucose spike is injuring the body. One view holds that glucose that stays high, plus sitting, is the combination more worth worrying about ¹⁷.

Misconception four: switching to a sugar-free drink solves it. Framingham found that people who had recently been regular drinkers of low-calorie sweetened drinks had higher risk on some lipid markers, while long-run average intake was not associated with disordered lipids ²⁸. One review also treats sweetener research as quite mixed ¹⁴.

Misconception five: one pill or one supplement can stand in for exercise. In the DPP trial, the lifestyle intervention clearly outperformed metformin ²⁵.

How it connects to your blood work

Insulin resistance is one of the few problems that can show an early hint on a blood test. Related markers in PearLab:

• fasting-glucose: the most used, and often a signal that rises late ²⁴.
• hba1c: two or three months of average glucose, and a main basis for calling prediabetes and diabetes ²¹.
• fasting-insulin: the extra secretion the pancreas makes to push glucose into cells. One continuum puts hyperinsulinemia at the front ².
• homa-ir (the HOMA-IR insulin-resistance index): an estimate from fasting glucose and fasting insulin together. Norton thinks it is more informative than a single fasting glucose ¹³.
• c-peptide: for every portion of insulin the pancreas releases, it releases a portion of C-peptide, a reflection of actual pancreatic output.
• ogtt-2h-glucose: glucose two hours after a standard glucose drink, which can find a problem that fasting does not show ²¹.
• triglycerides and hdl-c: two of the five signals of metabolic syndrome ², and also tied to sugary-drink intake ²⁸.
• alt (alanine aminotransferase) and ggt (gamma-glutamyl transferase): clues to liver injury, often used in a first look for fatty liver ³⁰.