BPC-157 has zero completed human trials, and Tony Huge uses it anyway. When he eats wheat his gut normally takes five days or more to recover; on BPC-157 it takes one or two, every time, across roughly 60 repetitions. This video is about holding both facts at once: the rat literature that built the peptide’s reputation, the very short list of human evidence, the regulatory gray zone, and how to read your own results without fooling yourself.
Video: BPC-157 Explained: The Rat Studies, the Human Data, and the Trade-Offs. Published August 3, 2026. Watch on YouTube.
What the video covers
- 00:00 The two true things: Tony’s gut repairs in one to two days on it, and there are no completed human trials.
- 00:34 How NSAIDs and cortisone work: They block the inflammation signal, which is step one of repair.
- 01:20 How BPC-157 is proposed to work: VEGFR2, new blood vessels, and supply lines to tissue that has almost none.
- 03:50 The NSAID trade-off: Journal of Applied Physiology on impaired tendon-to-bone healing, and why cortisone injections get capped.
- 05:05 The study that built the reputation: The 2006 rat Achilles tendon-to-bone model.
- 07:56 Every dose on one chart: Effective from 10 picograms to 10 micrograms per kilo, a million-fold spread.
- 09:58 Half-life and route: Under 30 minutes in rats, about five in a dog, and no validated rat-to-human conversion.
- 10:35 The complete human evidence: A cancelled phase 1, a 16-patient case series, a two-person safety pilot.
- 13:49 Legal status: Off the FDA’s Category 2 list in April 2026, not on Category 1, not approved anywhere.
- 14:22 WADA: Prohibited at all times, in and out of competition.
- 15:24 Reading anecdotes: When an observation is weak, when it carries weight, and the own-control design.
- 21:06 The parachute paper: Why “no RCT” is not the same as “no effect.”
- 23:31 Grading every claim: Nothing reaches A. Most of it sits at C.
Two ways to treat an injury
Tony starts with the two most common injury treatments on the planet, ibuprofen and cortisone, and the trade-off he thinks people miss. Both work by shutting down inflammation. But inflammation is step one of the body’s healing process, the repair crew arriving. Block the signal and pain drops, but so does the rate of repair. You feel better while the rebuild slows down, and that is how the drugs are designed to work.
The BPC-157 proposal runs the other way. Damaged tissue needs supplies to rebuild, protein and immune cells among them, and supplies travel through blood. Tendons and ligaments heal slowly compared to muscle precisely because they are so poorly supplied. BPC-157 is proposed to act on VEGFR2, a receptor on the cells lining blood vessels, and signal for new vessels to grow in. Instead of turning the inflammation off, it turns the supply on. Tony’s side-by-side is blunt: on the left, 60 years of human data and FDA approval; on the right, rat studies.
He gives the anti-inflammatories their due, and then the trade-off. He cites the Journal of Applied Physiology on NSAID exposure impairing tendon-to-bone and soft tissue healing, and notes that repeated cortisone injections are associated with tendon weakening over time. The reason clinicians cap how often they inject is not that it stopped working; it is that the tendon cannot take it. In rats, BPC-157 reversed corticosteroid-impaired tendon-to-bone healing.
Why rats, and why that is not the weakness
The study that built the reputation is Krivic and colleagues, 2006, Journal of Orthopaedic Research, a rat Achilles tendon-to-bone model with dosing started 30 minutes after surgery. Treated rats showed higher failure load, greater stiffness, better collagen organization, and more capillary density than saline controls. Four separate measures moving the same direction, and the capillary result is exactly what the proposed mechanism predicts.
Tony spends time on “it’s just a rat study” because he thinks it gets used to wave away evidence by people who never asked why rats were used. Tendon-to-bone healing, angiogenesis, and collagen repair run the same way in rats and humans. Rats let you standardize the injury, and they let you take the tendon out afterward and pull it until it breaks, which no human trial will ever do. He is fair about the limits: animal data carries well for tissue repair, wound healing, angiogenesis, and gastric ulcer models, and badly for the brain, immune system, sepsis, and cancer, where drugs have a long history of curing the mouse and doing nothing for the patient. The BPC-157 literature sits almost entirely in the first group.
A million-fold dosing window
On Tony’s chart, with each grid line ten times the last, the rat tendon and ligament doses reported effective were 10 picograms, 10 nanograms, and 10 micrograms per kilo. The rat gastric ulcer range runs 200 nanograms to 50 micrograms per kilo. What people actually take, 250 to 500 micrograms a day, works out to roughly 3 to 6 micrograms per kilo for an 80 kg adult, which lands at the high end of what worked in animals. The market did not pull that number from thin air, and Tony calls that the most defensible thing about consumer dosing.
“The same reported effect at 10 picograms and 10 micrograms per kilo. How can that be true?”
Tony Huge, 09:05
His two explanations are a wide therapeutic window, where a signaling peptide saturates its target at a tiny dose and everything above is excess, or measurements noisier than the papers suggest. Nobody has run the study to separate them. What is not disputed is the half-life, under 30 minutes in rats and about five minutes in a dog, so the peptide acts as a brief signal rather than circulating all day. Route matters too: most rat work is intraperitoneal, most human use is subcutaneous or oral, and there is no agreed conversion. In Tony’s words, the consumer dose was a guess.
The complete human evidence, on one slide
No completed, published, randomized controlled trial, for tendons, gut, or anything. A phase 1 in 42 healthy volunteers registered in 2015, cancelled in 2016, no results and no explanation ten years later. A retrospective, uncontrolled knee pain series of 16 patients with 87.5% reporting improvement. A safety pilot in two people at 20 mg with no adverse events. Tony adds a second problem underneath the animal data: the large majority of it comes from a single research group, and independent replication is how a finding becomes reliable.
His grading uses A for proven in people, B for limited human data, C for animal or lab, and D for mechanism only. Tendon, ligament, muscle, ulcer and gut, wound, and nerve healing all get C. So do gut protection from NSAID damage, blood pressure effects, and angiogenesis as a measured outcome. Joint pain relief, interstitial cystitis in 12 patients, and short-term IV safety get B. Nothing reaches A, and Tony is direct that B here means the weakest form of human evidence there is.
Cheat Sheet Pivot
What Tony reported and what the studies used, not instructions for anyone else.
- Tony uses BPC-157 for gut repair after wheat exposure, with a known five-day baseline before he ever started the peptide, a one-to-two-day recovery on it, and roughly 60 repetitions of the same challenge. He calls this the strongest shape an observation can have and still not a trial.
- The consumer range he charted, 250 to 500 micrograms a day, sits at the top of the effective rat range; he historically used higher doses because the compound is cheap and he has never seen side effects at any dose, while wondering what the minimum effective dose is.
- Reported side effects on forums include injection-site bumps, redness, nausea, dizziness, and drowsiness. Tony does not usually see them and speculates about product quality or mild allergy.
- Regulatory status as of the video: removed from FDA Category 2 in April 2026, not added to Category 1, not approved for any indication anywhere. Prohibited by WADA at all times.
- The largest body of animal work is gut and ulcer healing, not injury, and that matches where the anecdotal reports cluster.
BPC-157’s entry alongside the rest of the peptide catalog is in the Miracle Molecules Cheat Sheet. For how Tony has run it over the long term, read BPC-157 Complete Protocol Guide: My 9-Year Real-World Experience, and for a lab’s view of the same literature, see 5 Critical BPC-157 Research Insights.
Where the evidence stops
Tony’s closing argument is about how to think, not what to take. An anecdote is weak when the injury would have healed anyway, when several things changed at once, when the change was slow and subtle, or when it happened once. Flip every one of those and it carries weight: a stable baseline, a stuck injury that moves, the same challenge and the same response repeated. Clinicians have a name for using prior history as the control arm, the own-control design, and it still cannot rule out expectation. He cites the BMJ parachute paper, Smith and Pell 2003, for the point that demanding an RCT for a large, fast effect is a category error rather than rigor, and he adds that trials report averages, which can describe nobody in the room.
None of that changes the status. There is no human trial. The animal data comes mostly from one group. The dose was extrapolated without a conversion. Tony leaves it as a hypothesis, stated as a hypothesis: that BPC-157 is exceptional at supporting human tissue repair, pending the studies nobody has run.
Keep going
For why the compounding status keeps moving, read The FDA Peptide Crackdown: Regulatory Capture, Not Patient Safety. New companion articles reach the tonyhuge.is email list first, and the peptide archive is on tonyhuge.is.