"I Made Myself Older By Mistake": What Rapamycin Reveals About Biological Age Tests
Bryan Johnson stopped rapamycin after an epigenetic-clock study suggested it was speeding up his aging. Dr Brad Stanfield — a family doctor running his own rapamycin trial — argues the conclusion outruns the data, because biological age kits have never been validated against real outcomes. His counter-proposal: judge interventions by hard outcomes and randomized human trials, not by a number on a test kit.
Overview
This video is a physician's line-by-line reaction to Bryan Johnson's announcement that rapamycin, taken for roughly five years as a longevity intervention, appeared to be accelerating his aging. Dr Brad Stanfield opens by explaining his own starting point: he is a practising family doctor who builds protocols from randomized human trials and clinical guidelines, and he is currently running a human trial of rapamycin himself.
He walks through the compound's history — discovered in Easter Island soil, developed as an antifungal, then approved in 1999 as a powerful immunosuppressant for organ transplants — and explains why that immunosuppressant reputation, plus the fact that rapamycin cannot be patented, made his own trial take two and a half years to crowdfund. He credits the preclinical evidence as genuinely intriguing, especially the Interventions Testing Program results replicated across three separate labs.
The heart of his objection is measurement: Bryan's conclusion rests on biological age test kits and a preprint covering sixteen epigenetic clocks, and no long-term study has yet shown that lowering one of those numbers translates into fewer cancers, less cardiovascular disease or lower mortality. He also flags the confounding problem — Blueprint changes many variables at once, so attributing mouth ulcers, cholesterol shifts and a raised resting heart rate to one compound is hard, though Bryan's stop-and-observe test is a reasonable move.
Stanfield then places the whole discussion on the evidence pyramid: case reports sit at the bottom, randomized trials and systematic reviews at the top, and a compelling n-of-1 story should not outrank a trial. He closes by describing his own study — older adults, exercise plus weekly everolimus or placebo, with the 30-second chair stand test as the primary outcome — as the kind of evidence that could actually move the field.
Key quotes
4The way that we figure out what the best health protocol is, is by looking at what the randomized human clinical trials show.
They're interesting research tools, but I don't think anyone should be using them to make their own clinical judgments.
You can't figure out what exactly those biological age results are actually meaning.
Right now I think it's best to focus on hard outcomes — death rates, cancer rates, cardiovascular disease rates.
Key ideas
9Two different ways of building a protocol
One path collects randomized trial data and clinical guidelines, then has clinicians agree on recommendations. The other runs experiments on a single willing body. Both produce conclusions, but they carry very different weight.
A molecule with two reputations
Rapamycin came out of Easter Island soil as an antifungal, then turned out to be a powerful immunosuppressant and was approved in 1999 for transplant patients. That second identity is why the longevity conversation around it has always been cautious.
Unpatentable means underfunded
Because the compound cannot be patented, there is little commercial incentive to run trials. Stanfield spent two and a half years crowdfunding his own study from individual supporters on the internet.
Transplant doses are not longevity doses
The side effect profile people fear — infections, poor wound healing, metabolic disruption — comes from transplant patients on high doses plus other immunosuppressants. Healthy-volunteer trials use far lower, intermittent dosing.
The animal data is unusually solid
The 2009 lifespan result came from the Interventions Testing Program, which runs the same protocol in three separate labs at once specifically to beat the reproducibility problem. Repeated runs kept showing the effect, and combining it with acarbose extended it further.
mTORC1, mTORC2 and which version you take
The longevity interest sits with mTORC1 inhibition, while mTORC2 is tied to the metabolic side effects. Different analogues differ in selectivity — Stanfield chose everolimus for his trial because it leans more towards mTORC1.
The validation gap in biological age kits
If one 60-year-old scores 50 and another scores 70, nobody has yet run the multi-year follow-up showing the second person actually faces more disease. Until that exists, moving the number is not the same as moving your health.
Isolating one variable inside a hundred-variable protocol
Mouth ulcers, slow wound healing, cholesterol and glucose shifts and a raised resting heart rate all appeared. Stopping the compound and watching them reverse is the cleanest signal available in an n-of-1 setting — but it is still one person.
The evidence pyramid, and where a case report sits
Case reports and case series form the base; cohort studies sit above them, then randomized trials, then systematic reviews. A vivid personal story does not outrank a trial, however well produced the video is.
Practical takeaways
6- 1
Treat clock results as curiosity, not verdict 12:40
Log a biological age result if you enjoy tracking it, but don't start or stop anything because the number moved. It has not been shown to predict what happens to you.
- 2
Change one thing at a time 14:20
If you are running several new habits at once and something shifts, you have no idea which one did it. Stagger changes far enough apart to read the signal.
- 3
The stop test is the most honest test you have 16:20
When you suspect something is causing a side effect, remove it and watch whether the effect reverses. That is what turned suspicion into a conclusion here.
- 4
Ask where a claim sits on the evidence pyramid 19:30
Before adopting anything, check whether it rests on one person's report, an observational study, or a randomized trial. The label changes how much weight it deserves.
- 5
Anchor on hard outcomes 20:40
Mortality, cancer incidence and cardiovascular events are what actually matter. Prefer measures with that kind of track record over scores that are still being validated.
- 6
Build the foundations before the experiments 21:30
The strongest returns still come from the basics medicine already recommends. Exotic interventions with thin human evidence carry real downside risk.
Topics & chapters
15A doctor sets up the reaction
Stanfield introduces the video and explains that he is running a human trial of rapamycin himself, which is why the story interests him.
Protocols from trials vs protocols from self-experiments
He describes how clinical guidelines are built from randomized data, and how different that is from assembling a personal protocol.
Easter Island soil to antifungal
The origin story of rapamycin, discovered in the 1960s in a soil sample from a cave, and first seen as a fungal treatment.
Why nobody funds these trials
The compound can't be patented, so there's no commercial return — Stanfield crowdfunded his study over two and a half years.
1999 approval and the immunosuppressant problem
FDA approval for transplant patients gave rapamycin a difficult reputation, and made trials in healthy people harder to launch.
The 2009 mouse result and the ITP
Lifespan extension in older mice, replicated across three labs, plus larger effects when combined with acarbose.
How it works: mTORC1
The growth pathway behind the longevity hypothesis, and how exercise and protein activate it while the drug inhibits it directly.
The problem with biological age kits
Stanfield explains why unvalidated clock results shouldn't drive personal decisions about supplements, diet or drugs.
mTORC2, selectivity and drug forms
Why the analogue matters: different versions hit mTORC1 and mTORC2 differently, which affects side effects.
The dosing experiments
Weekly schedules, blood levels measured at 90 minutes and out to 96 hours, and questions about how high the doses actually were.
Side effects and the decision to stop
Mouth ulcers, slow healing, cholesterol and glucose changes, a raised resting heart rate — and what happened after discontinuation.
The preprint and sixteen epigenetic clocks
The study that triggered the 'accelerated aging' headline, and the confounders that make it hard to interpret.
The evidence pyramid
Case reports at the base, randomized trials and systematic reviews at the top — and why the order matters.
Stanfield's own trial
Older adults, exercise plus weekly everolimus or placebo, with the 30-second chair stand test as the primary outcome.
Closing: hard outcomes and shared results
A plea to measure aging by mortality and disease rates, and appreciation for publishing results whether they work or not.