Pillar 02
Cells stop responding to insulin, and the body compensates by making more of it. It can hide behind normal blood sugar for a decade or more.
| Test | What it measures | The number | What it misses |
|---|---|---|---|
| Fasting glucose | Blood sugar after an overnight fast. The standard test. | American Diabetes Association: 100 to 125 mg/dL is prediabetes, 126 and above is diabetes. | It moves late, by design. It is measuring the thing the extra insulin is successfully holding down. |
| HbA1c | Roughly the average blood sugar over the previous two to three months, read off sugar stuck to red blood cells. | ADA: 5.7 to 6.4 percent is prediabetes, 6.5 percent and above is diabetes. | It is an average of the same late signal, so it inherits the same lateness. Anything that changes how long red cells live, including anaemia, distorts it. |
| Two-hour glucose tolerance test | Blood sugar two hours after drinking 75 grams of glucose. A challenge rather than a snapshot. | ADA: 140 to 199 mg/dL is prediabetes, 200 and above is diabetes. | It repeats poorly. The same person tested twice in the same month can land in different categories. |
| One-hour glucose, same test | The reading an hour in, which most labs throw away. | The International Diabetes Federation moved on this in 2024: 155 mg/dL or above is intermediate hyperglycaemia, 209 and above is type 2 diabetes (PMID 38537890). | Newer, and the ADA has not adopted it. You have to ask for the one-hour draw to be kept. |
| Fasting insulin, HOMA-IR, TyG index | Attempts to measure the compensation itself rather than the sugar. | No cut-off, for any of the three. I could not find a threshold set by any guideline body. The numbers quoted online come from individual population studies. | Insulin assays are not standardised against each other, so the same blood can return different values at different labs. A TyG number quoted without its units is uninterpretable. |
| Triglyceride to HDL ratio | Two numbers already on a standard lipid panel, divided. | A 2003 research proposal, roughly 3.0 when both are in mg/dL as US labs report them (PMID 14623617). Not a guideline. | Mediocre even in the group it was built in, and it behaves differently across ethnic groups. |
| Waist to height ratio | Body shape, measured with a tape. | The UK's NICE guideline NG246 puts it plainly: keep your waist under half your height, so a ratio below 0.5. | It measures shape, not insulin. Plenty of people cross 0.5 and are insulin sensitive, and plenty of thin people are not. |
The gold standard is not something you can buy. The reference method is a hyperinsulinaemic-euglycaemic clamp: two intravenous lines, several hours, and a technician adjusting a glucose drip at the bedside to hold your blood sugar flat while insulin runs in (PMID 382871). It lives in research units. Every marker in the table above is a stand-in for it, and every one of them is worse than it.
Three of the most-quoted numbers have no agreed threshold. Fasting insulin, HOMA-IR and the TyG index are all over social media with confident cut-offs attached. Those cut-offs come from individual studies in particular populations, not from any guideline body. That does not make the tests useless, but a number presented as a diagnosis when nobody has agreed what it means is doing something other than informing you.
15 compounds covered here are marketed or taken on this rationale. Each entry says where its dose came from and what the published record supports, which is usually less than the pitch.
Five hours of infusion, at roughly ten times the blood carnitine the label calls normal, moved muscle carnitine by nothing unless insulin was also pinned high by a second drip. The only protocol I found that raised it outside that laboratory setup was 24 weeks of capsules taken with 160 grams of sugar a day.
It came out of a Parkinson and Alzheimer programme that produced no drug, then took 10.6 percent off body weight in 24 weeks. The phase 3 that should have followed was never run, and the Mexican approval you will see quoted is contradicted by the sponsor's own releases.
Sold on a 20 to 30 hour half-life, and cleared out of a virgin rat about eleven times faster than the hormone it copies.
The first GLP-1 and glucagon dual agonist approved anywhere. It is neither retatrutide nor tirzepatide, and the weight loss figure vendors quote came from a press release rather than the journal.
I could not find a human study of this peptide. The trials and the fat loss claims belong to AOD-9604, one oxygen atom away.
The single agent has its own trials and its own numbers, and they are about half the size of the ones quoted for it.
502 adults, 24 weeks, three doses against placebo, no separation on the primary endpoint. It is hGH 177-191 with a tyrosine on the front, one oxygen away from the fragment sold as 176-191, and neither of those records is this one's.
One indication, one population, and a label that calls it weight neutral. The dose everyone repeats belongs to a product that is no longer sold.
Geref left the American market in 2009 because the company stopped selling it. FDA put that in writing four years later, and that one sentence is what the compounding industry stands on.
An approved drug with placebo controlled phase 3 evidence in tens of thousands of people, and a ClinicalTrials.gov record that reproduces the biggest of those trials character for character, down to Eli Lilly's internal protocol code.
Three compounds sold as one system. I could find no study of any two of them together, and the confident version of each one rests on something that does not support it.
Merck's MK-0677 is Lumos Pharma's LUM-201, recruiting children right now at a dose worked out per kilogram. Everyone else takes a flat 25 mg, from one study in 1996.
The evidence here is large and mostly first rate. One ClinicalTrials.gov entry for it is a word for word copy of a study that ended in 2022, and it is the only record in its cluster with nothing on its face to give it away.
This is the one category on this site where the drugs demonstrably work and the trials are enormous. Which is exactly why the risk has moved off the molecule and into the vial.
One of the most studied drugs of the past decade. The real questions are what is in the vial you bought, and whether you are allowed to have it.
Free. One email a week at most. Unsubscribe in one click.
Double opt-in · No spam · Unsubscribe anytime