A before-and-after is a comparison, and a comparison is only as good as what was measured at each end. For BPC-157 the published human record is three uncontrolled reports, and two of the three measured nothing but what patients said afterward.
What those reports contain is worth setting out in full, because the numbers circulating online come from them. The trial-by-trial evidence is covered separately in what human evidence for BPC-157 exists.
What the three human reports recorded
The first is a retrospective chart review at a private clinic in Orlando, covering injections given between 2019 and 2020 [1]. Seventeen patients had a peptide injected into a knee; sixteen were reached by phone afterward.
Twelve of those sixteen had received BPC 157 alone. Eleven of the twelve, or 91.6%, told the caller their knee pain had improved significantly. Across all sixteen, 87.5% reported relief.
Within that group of twelve the split was 91.6% versus 8.3%, which is eleven people and one person. The four who also received a second peptide split 75% versus 25%, which is three people and one.
Arms that size cannot support a comparison, and the paper does not offer one. The percentages are the study restating its counts.
The paper is unusually candid about its own instrument. It states that no specific tools were used to measure improvement in function, quality of life, stiffness or activities of daily living.
| Report | People | What was given | What was measured |
|---|---|---|---|
| Knee pain, 2021 | 16 of 17 reached | Intra-articular injection, 6 to 12 months earlier | A phone question about pain, rated from memory |
| Interstitial cystitis, 2024 | 12 women, ages 39 to 76 | 10 mg total, injected around the bladder in one procedure | Global Response Assessment questionnaire |
| Intravenous safety, 2025 | 2 adults | 10 mg, then 20 mg, each infused over an hour | Blood chemistry and vital signs |
The bladder study is the strongest of the three, and it is still a single-arm pilot [2]. Ten of twelve women rated their result at 100% and two at 80%, on a five-point global questionnaire.
Every woman in it had already failed the approved treatment, which makes a favorable response more striking and a placebo response no less likely.
The third report is not an efficacy study at all [3]. Two people received 10 mg and then 20 mg by infusion, and the authors looked for changes in heart, liver, kidney, thyroid and glucose markers.
One older claim sits outside that set. A 2006 review states that the peptide was safe in clinical trials for inflammatory bowel disease run under the codes PL-10, PLD-116 and PL 14736 [8].
Those trials have no retrievable primary report. What survives is a sentence about them inside a review whose own experiments are in rats, so there is nothing to read about who was treated or what improved.
They found none, and they say so as a safety finding in two people. It is not a before-and-after of a healed injury.
What a photograph cannot settle
A pair of images or a pair of pain scores taken months apart carries several explanations at once, and the three reports above cannot separate them.
Most soft-tissue injuries improve on their own. A patient who seeks an injection is usually at their worst on the day they seek it, so the next measurement tends to be better whatever was injected.
People also change what they do. Somebody paying for a peptide often rests differently, trains differently and watches the joint more closely for the following month.
And the question itself shapes the answer. Asking someone six to twelve months later how much an injection helped is not the same as measuring them at the time.
The measured before-and-afters are in rats
The imaging and tissue evidence people expect from a before-and-after does exist, in animals. A 2026 rat study divided Achilles tendon repair into groups and compared them on histopathology and biomechanics [6].
That design produces a genuine paired measurement, because the tendon can be loaded to failure and sectioned at the end. No human study of BPC-157 has done anything comparable.
Animal results do not carry over on their own. Dose, route and healing timelines differ, and a rat tendon is not a human one, which is why the dose question has no human answer either. That is set out in BPC-157 dosage.
The other unknown is what was in the vial
A before-and-after also assumes the person took what they think they took. For a substance with no approved product, that assumption is doing real work.
A 2026 review in Missouri Medicine covers injectable peptides sold outside the regulated supply chain, and the hazards that follow from unverified contents [5].
FDA lists BPC-157 in 503A Category 2, the group of bulk substances that raise significant safety risks[7]. There is no lawful compounding route to it, so there is no batch record to compare either.
The legal picture around peptides sold this way is its own subject, and it changes more often than the science does. See whether peptides are legal and the related evidence for TB-500, which is often injected alongside it.
What would actually settle it
A randomized, blinded trial with a placebo arm and an imaging endpoint would settle it, and none has reported[4].
Until one does, the honest description of the human record is three small uncontrolled reports from a single clinic, two of which rest on what patients said.
That is not the same as saying the compound does nothing. It is saying nobody has measured whether it does.
For the wider question of how these compounds are studied, see the recovery hub.