Here are two facts most readers hold at the same time without noticing they cancel out.
Rapamycin is, reportedly, the undisputed champion of small molecules for extending lifespan in animal models, replicated across independent laboratories, effective even when started late in life.
And when a doctor remortgaged his house to fund the first rigorous human test of that drug paired with exercise, the older adults on rapamycin did, reportedly, about two fewer chair-stands, walked 4.87 meters less, and gripped 1.13 kilograms weaker than the people on placebo.
Every functional measure the trial recorded favored the sugar pill.
That is the shape of the problem the longevity field keeps stepping around: the mouse data is overwhelming, and the human data underneath it barely exists.
The mouse data is overwhelming, and the human data underneath it barely exists. That is the shape of the problem the longevity field keeps stepping around.
The trial is called RAPA-EX-01. It was led by Dr. Brad Stanfield with Dr. Matt Kaeberlein as co-author, funded entirely by public donations—$724,637 raised through Lifespan.io and VitaDAO, per Dr. Christin Glorioso—and published this April in the Journal of Cachexia, Sarcopenia and Muscle.
Forty sedentary adults aged 65 to 85 in Auckland, randomized 1:1 to once-weekly 6 mg rapamycin or matching placebo, all doing the same home program three days a week: 30-second chair-stands for strength, a stationary bike for endurance, for 13 weeks.
The design was not naive. Dosing landed on Day 6 of each training week, roughly 24 hours after the last workout, precisely to avoid stepping on the post-exercise anabolic signal while preserving rapamycin's autophagy benefits between sessions.
That is the "cycling hypothesis"—the idea that you can alternate mTORC1 on (exercise) and off (drug) across a week and get the best of both. It rests, reportedly, on real preclinical grounding, plus the counterintuitive finding that mTOR signaling is paradoxically elevated in aging muscle, which raised the possibility that inhibiting it might resolve, not worsen, age-related anabolic resistance.
"We hoped that weekly rapamycin dosed 24 hours after the last workout would preserve the autophagy benefits of mTORC1 inhibition while leaving room for post-exercise adaptation," Stanfield told Lifespan News. "It didn't."
His best explanation, per Lifespan News: rapamycin's roughly 62-hour half-life probably kept mTORC1 partially suppressed into the next training week. The timing window never actually opened.
The primary endpoint—change in chair-stands at 13 weeks—did not reach statistical significance in the intention-to-treat analysis (−2.13; 95% CI −4.61 to 0.34; p=0.089). But the two prespecified sensitivity analyses did: complete-case at p=0.045, per-protocol at p=0.007.
Read that carefully. Among the people who actually took the drug and did the exercise, the harm signal was statistically significant. The more faithfully you followed the protocol, the worse rapamycin looked.
The exploratory outcomes offered no rescue. Four epigenetic clocks—PCGrimAge, SystemsAge, OMICmAge, DunedinPACE—showed mixed, non-significant trends. C-reactive protein was higher in the rapamycin arm, contradicting the reflexive assumption that the drug lowers inflammation. HbA1c and LDL both drifted up. Adverse events ran, reportedly, 99 to 63, with one serious event—a pneumonia hospitalization—in the rapamycin arm.
So: fold the tent?
No. And it's worth being precise about why, because the temptation to over-read a negative is as much a failure of evidence discipline as over-reading a positive.
The trial enrolled 40 people. It was powered only to detect large effects, meaning real-but-modest effects in either direction would be expected to miss significance. It tested one narrow mechanism—muscle adaptation to training—under one specific protocol, in previously sedentary older adults, over 13 weeks.
That is not a longevity trial. It looks almost nothing like one. Rapamycin's animal case is built on lifespan and healthspan across the whole organism, not on how many times a 72-year-old can stand up from a chair in half a minute.
Peter Attia's framing is the correct one: this is "a small additional update in the negative direction on one narrow dimension," not a resolution of the broader question. Andrew Steele, watching a man who reportedly remortgaged his house to run it, put the honesty plainly—Stanfield "tells it straight," publishing a null primary endpoint he had every personal incentive to spin.
That is what makes RAPA-EX-01 valuable, and it has nothing to do with the result being negative.
A crowdfunded, prespecified, double-blind trial reported its data honestly against the interests of everyone who paid for it. That is the template. That is how a field with an evidence gap this wide is supposed to behave.
Now hold the template up against what the drug's evidence sheet actually says, because this is where the scandal lives.
Geroevidence, which grades the longevity pharmacopeia on a ladder from mechanism to hard outcomes, lists rapamycin as "Moderate" evidence—on surrogate endpoints only. Its hazard ratio field reads "HR —," reportedly with no mortality figure shown.
Compare the same dashboard's entry for GLP-1 receptor agonists: "Strong," HR 0.80, 95% CI 0.72–0.90, with 86 RCTs indexed. That is what hard human outcome data looks like when it exists.
Rapamycin has, per Geroevidence, 12 RCTs indexed and a dash where the mortality number should be. The champion of the mouse has no proven human outcome.
This is the gap the field talks over. It is not that rapamycin has been disproven in humans—it hasn't been tested in the way that would matter.
And the reason is structural, not scientific. As Steele lays out, a proper lifespan-and-healthspan trial for rapamycin would cost $50–100 million.
That is a rounding error against the potential upside, which he reportedly puts in the tens of trillions. It is also more than any drug company will spend on an out-of-patent molecule with no exclusivity to recoup, and more than any "heroic doctor/YouTuber" can crowdfund.
So the definitive trial does not get run. Instead a single remortgaged house and $724,637 in donations funded the one rigorous human study the field has, and it was aimed at chair-stands.
Meanwhile the market does not wait. The complementary and alternative medicine market for anti-ageing and longevity was valued around $63.6 billion in 2023 and is projected to grow more than 20% a year through 2030. Thousands of Americans already take rapamycin off-label. One widely circulated 2026 roundup reportedly files it under "Tier 2: Promising Evidence" and calls its human evidence level "Moderate"—the same word Geroevidence uses, sitting next to a hazard ratio that literally does not exist.
The word "Moderate" is doing enormous unspoken work. Moderate on surrogate endpoints, with no mortality data, is being read by a growing consumer base as moderate confidence that the drug extends life. Those are not the same claim.
This is the reflex RAPA-EX-01 should permanently retire—not rapamycin, but the habit of citing "consistent lifespan extension in mice across multiple independent laboratories" as though it settled the human question.
It doesn't. The animal case is a reason to run the trial. It is not a substitute for having run it.
The single most useful output of this study is therefore not the finding that weekly rapamycin blunted exercise gains in 40 older New Zealanders. It is the demonstration—by a self-funded volunteer effort—of how the work is supposed to be done, and the glaring evidence of who wasn't doing it.
A field valued in the tens of billions could not produce this. A man who put his house on the line did.
Until someone funds the $50–100 million trial, that dash in the hazard-ratio column is the most honest thing anyone can say about rapamycin and human lifespan. Everything else is mechanism, hope, and mice.