Tirzepatide "Prevents Diabetes." Prevented, Delayed, or Masked?

The argument always arrives the same way. Two options in the cart, a real price gap between them, and somebody in the chat drops the line that ends the discussion: tirzepatide prevents diabetes, semaglutide just makes you lose weight.

I've heard that sentence enough times to recite the rest of the thread from memory. And I should say up front what I have riding on it: I run a research-peptide company, incretin attention is the single biggest driver of questions we get, and we sell nothing containing tirzepatide. Not a vial, not a variant, nothing. So I have no cart to protect here, just a stake in whether this market can read its own evidence.

The word doing all the work in that sentence is "prevents." Almost nobody forwarding the screenshot can tell you what it was counting.

Prevented, delayed, or masked

I run every prevention claim through three questions, in this order, and I have never once needed a fourth.

One: what event was counted, a disease or a threshold? Type 2 diabetes, as a trial endpoint, is a line on a lab report. Cross the glucose or HbA1c cutoff, you're a case. Stay under it, you're not. That's how it has to be operationalized, and there's nothing dishonest about it. But it means a drug that lowers blood glucose is being scored on whether it lowered blood glucose. Fewer people crossing the line is a real, clinically meaningful thing. It is not the same act as removing the disease from the body.

Two: who got pre-selected into the population? Prevention trials in this space enroll people with prediabetes and elevated BMI, because that's where enough events happen fast enough to measure anything. Which means the result belongs to those people. It does not travel to someone with normal glucose tolerance, and it does not travel to a lean 26-year-old in a Discord asking whether he should run it prophylactically. That extrapolation happens constantly and it is not supported by anything.

Three: what happened after the drug stopped? This is the question that decides which of the three words you're allowed to use. Prevented means the risk is gone. Delayed means the clock restarted. Masked means the number moved while the drug was in the system and the underlying state didn't. If you cannot say what happened off drug, you cannot say "prevent" without an asterisk, and the asterisk is the whole story.

Why I'm not repeating the number

The screenshot leads with a big relative risk reduction. I'm not going to reprint it here, and not because I think it's fake.

Because a relative risk reduction with no event counts beside it is a receipt with the total torn off. Ninety-something percent fewer of what, out of how many, over how long, those four numbers are one fact, and the market keeps circulating one-quarter of it. If the person handing you the percent can't hand you the two absolute event rates from each arm, they don't have the finding. They have a ratio.

Go get the paper. Read the enrollment criteria. Find the off-drug follow-up. Then you can say the sentence out loud.

The three tiers people keep blending

Most of the confusion in these threads isn't about tirzepatide at all. It's that three completely different grades of evidence get stacked into one word"research", and then argued with equal confidence.

At the top: randomized trials with prespecified outcome endpoints. That's where the prevention discussion lives, and it's why the prevention discussion deserves to be taken seriously in the first place.

In the middle: post-hoc analyses and mechanistic signals. Real work. Hypothesis-generating. Not the same thing as a result the trial was built to produce.

At the bottom: the single case report. And here's where I'll get specific, because I actually read this one and it's a perfect illustration of how a tier gets misused.

One woman. Severe hypertriglyceridemia, type 2 diabetes, recurrent hypertriglyceridemia-mediated acute pancreatitis. Tirzepatide started. Triglycerides at initiation were 335 mg/dL; 137 at three months; 85 at twelve. HbA1c went from 7.0% to 5.5%. Insulin was discontinued entirely. Inflammatory markers normalized, and across twelve months of follow-up she had no further pancreatitis episodes.

That is a striking chart. It is also n=1, with no control, no randomization, and no way to separate the drug from everything else happening in that woman's life over a year.

And here's the irony I can't get past: I have watched "there's a pancreatitis case report" get passed around as a red flag on this compound. The published case report I sat down and read describes a woman who stopped having attacks. Same tier, opposite direction, and neither version proves a damn thing about population risk. A case report cannot establish that tirzepatide causes pancreatitis. It cannot establish that it treats it either. That's what the bottom tier means.

The semaglutide comparison, stated honestly

There is no head-to-head trial of tirzepatide against semaglutide on a diabetes-prevention endpoint. None. So every "tirzepatide is evidence-superior for prevention" claim in this market is an inference from weight-loss magnitude, more weight lost, therefore more threshold-crossing avoided.

That inference might be right. It is still an inference, not a result, and this market spends the first as though it were the second every single day.

What this costs us

Two commitments, since I'm raising the bar. For every compound we list, our page states whether a randomized human outcome trial exists, what population it enrolled, and whether the supporting data is randomized, post-hoc, animal, or case-level, the same tiers I just used. And we do not borrow tirzepatide's clinical evidence to describe structurally related compounds we sell. Our incretin-adjacent listings are being audited right now for exactly that kind of borrowed credibility, because it is the easiest sentence in this industry to write and the hardest one to defend.

The honest version of the whole thing fits in one line: in higher-BMI adults with prediabetes, randomized treatment sharply reduced crossing the diabetes diagnostic threshold while treatment continued, and nobody has run the comparison against semaglutide.

Say that instead. You'll be the most credible person in the thread.

Read the Research Translated evidence-tier index, how we separate randomized outcomes from post-hoc signals and case reports before we write a word about any compound.