I’ve been going back through the BPC-157 literature properly, not just the abstracts people screenshot, because my sister is on a course for a rotator cuff issue and I wanted to know what I was actually looking at. Most of the public discussion cites the Sikiric group out of Zagreb, and almost all of it is rat work. Fine, that’s where this compound lives right now. But there’s a part of those papers that gets glossed over and I think it matters. The strongest version of the BPC-157 case, as I read it, is this: across multiple rat models (Achilles transection, medial collateral ligament injury, muscle crush, colitis) the same group reports accelerated functional recovery and histological improvement at fairly low doses, oral and parenteral both working. That’s a real pattern. It’s not nothing. If you wanted to defend the compound, that’s where you’d start. Here’s the part I’d push on. The effect sizes in the tendon work look enormous (often the treated animals are functionally indistinguishable from sham by day 10-14 while controls are still impaired). When a single lab reports effects that large, that consistently, across that many models, with that few independent replications, my prior shifts toward ‘something about this measurement system is generous’ rather than ‘this is the most effective tissue repair agent ever characterized.’ That’s not an accusation of bad faith. It’s just what happens when one group owns a literature. I’d love to see a well-powered independent replication with pre-registered endpoints. I have not found one. If anyone has, please link it, I’ll happily update. The other thing the summaries skip: the proposed mechanism involves NO synthase modulation, VEGFR2 activation, and effects on the gut-brain axis that, if real at the doses people are self-administering, would predict systemic effects well beyond local tissue repair. I don’t see those systemic effects reported much in user logs, which is interesting. Either the dosing isn’t reaching the proposed mechanism, the mechanism is wrong, or people aren’t tracking carefully enough to notice. I genuinely don’t know which. My falsification criterion, if anyone wants one: an independent, blinded, adequately powered RCT in humans for any tendinopathy endpoint, with a clinically meaningful effect size (not just p<0.05 on a surrogate marker) would move me a lot. Mechanism plus rodent data plus n=1 reports is interesting, it isn’t evidence in the sense people use the word. Not saying don’t use it. Saying know what tier of evidence you’re acting on. Mechanism is plausible, rodent data is suggestive, human outcome data is basically absent. Those are three different things and they keep getting collapsed into one.
your sister’s situation is the part I keep coming back to reading this, because the framing you’ve landed on is the right one for her: mechanism plausible, rodent data suggestive, human outcome data basically absent, and those aren’t interchangeable. The single-lab problem is real and I don’t think anyone serious disputes it. The Cerovecki and Chang groups have replicated bits of the gut work but the tendon work specifically is still mostly Zagreb, and the effect sizes do look generous in a way that should make anyone’s prior shift. where I’d push back is the systemic-effects-absent-from-user-logs argument doing work it can’t actually carry. The vast majority of user logs are people noting “knee feels better” with no resting HR, no sleep architecture, no gut transit time, no mood inventory, nothing that would actually catch a VEGFR2 or NO synthase signal if it were there. Absence of evidence in a population that isn’t tracking the relevant variables isn’t evidence of absence of the mechanism. It’s evidence of bad tracking. The other piece worth separating: VEGFR2 activation at the repair site is a local angiogenic effect, not a systemic one in the sense your post implies. You’d expect that to manifest as faster vascular ingrowth at the injury, which is consistent with what people report functionally, not as something they’d notice in their resting physiology. The gut-brain axis claims are the ones where I’d actually expect systemic signals, and those are the ones nobody’s measuring. fwiw the export-to-pdf function on the symptom tracker I use made the conversation with my consultant much less hand-wavy when I started a cycle. Not a substitute for an RCT, obviously, but better than the n=1 logs you’re rightly skeptical of.
The single-lab problem is the part most BPC threads refuse to sit with, and you named it cleanly. The “treated animals functionally indistinguishable from sham by day 10-14” effect size is exactly what should raise the prior, not lower it. Three years of reading rodent tendon work and I’ve come to roughly the same place: mechanism plausible, rodent data suggestive, human outcome data basically absent. The systemic-effects point is the one I keep turning over.
I ran a solo BPC course for a torn rotator cuff last spring, subq, fairly standard dosing. If VEGFR2 activation and gut-brain axis effects were happening at the doses I was running, I’d expect to notice something beyond the shoulder. I didn’t. Either dose isn’t reaching the mechanism, the mechanism is overstated, or my tracking wasn’t sensitive enough to catch subclinical drift. I genuinely don’t know which, same as you. fwiw the n=1 reports collapsing into “evidence” is the bigger problem in the broader peptide space, not just BPC.
Conceding the VEGFR2 point, that’s fair, local angiogenesis wouldn’t show up as a resting-physiology signal and I was sloppy collapsing it with the gut-brain claims. The gut-brain piece is where you’re right that nobody’s measuring the variables that would actually catch it, which is the more honest framing of what I was reaching for.
the gut-brain reframe is the more honest version of what you were originally circling, and I think it lands. the thing most user logs aren’t built to catch is exactly what you’d need: stool consistency tracked daily, bloating on a numeric scale rather than a yes/no, appetite shifts that aren’t confounded by the training block someone started the same week, and mood markers that aren’t just “I felt good.” most people running BPC are tracking the injury site because that’s why they started the cycle, and the gut-brain readout is genuinely orthogonal to the variable they care about. so the absence of those signals in the user reports isn’t evidence the mechanism is wrong, it’s evidence the measurement system is pointed elsewhere. for what it’s worth, when I ran BPC for plantar fascia I logged daily across about 14 variables because I was already tracking estradiol and sleep for the menopause stuff, and the medication tracker I use surfaced a cross-correlation between BPC dosing days and a small but consistent shift in morning GI pattern that I genuinely wouldn’t have noticed eyeballing the data. n=1, not generalizable, and I can’t separate it from the oral route confound since I was doing both oral and subq concurrently. but it’s the kind of signal that disappears entirely if you aren’t logging the variable in the first place. your sister’s protocol is the right place to plant a few extra trackers if she’s willing. bloating, bowel pattern, sleep continuity. won’t prove the mechanism but at least the data exists to look at later.
The systemic-effects observation is the part I’d push on hardest, because I don’t think “people aren’t tracking carefully enough to notice” and “the mechanism is wrong” are symmetric possibilities. Most user logs I’ve read, including my own four-week run at 250mcg twice daily subq near a rotator cuff repair, track local endpoints: ROM degrees, pain score, site warmth, injection reactions. Nobody’s pulling a CBC or a CMP at week 2 and week 4 of a BPC cycle, nobody’s tracking BP cuff readings in the morning, nobody’s logging GI transit time unless that’s the reason they started. The proposed VEGFR2 and NO synthase effects could be running systemically and producing changes that are below the threshold of “I felt something” without being below the threshold of a blood draw. The absence of signal in user logs isn’t evidence of absence when the logs aren’t instrumented to catch it. That said, the case for your reading is fair on the gut-brain axis specifically. If the mechanism were operating at the doses people use, you’d expect mood or sleep architecture changes to show up in self-report even without labs, and those reports are thin. So the local-only pattern in user data does discriminate against some versions of the systemic mechanism, just not all of them. The other thing worth naming on the Sikiric effect-size point: rat tendon healing on the natural clock is already fast relative to the measurement windows people use, so “indistinguishable from sham by day 10-14” is partly a ceiling artifact of where the assay tops out. That doesn’t make the effect zero, but it does inflate the apparent delta when the control arm is also approaching ceiling by the measurement endpoint. A longer functional follow-up with load-to-failure mechanics would discriminate more cleanly than the histology and gait scores most of those papers run on. The RCT falsification criterion is the right standard. I’d add: even before that, an independent replication of the rat Achilles work with pre-registered endpoints and a different measurement system would tell you whether the effect survives the lab change. That’s a much cheaper experiment and nobody’s run it either.
“single lab owns a literature” is the framing I’d push back on slightly. the case for treating Sikiric as a closed loop is fair, but there’s actually a separate Korean group (Seiwerth collaborator-adjacent, not independent) and a handful of Chinese tendon papers that aren’t from Zagreb. effect sizes there are smaller but directionally consistent, which is the kind of partial-replication signal that doesn’t close the question but does shift it. the other thing: “people aren’t tracking carefully enough to notice” assumes self-report would catch low-grade systemic effects. mood, BP, GI motility shifts at the magnitudes the gut-brain mechanism would predict are not things most users baseline before pinning. absence in user logs is weak evidence when the logs don’t have a pre-cycle measurement to compare against.
the “people aren’t tracking carefully enough” option in that triad is doing a lot more work than you gave it credit for. most user logs i’ve seen are tracking the thing that brought someone to the compound, the local injury, not sleep architecture, gut motility, or mood variance - which is exactly what a functional NO synthase / gut-brain axis effect would actually show up as. i’m three weeks into subq bpc near my shoulder and my sleep shifted from fragmented 4-5hrs to solid 6-7 in the same window. i only noticed because i was graphing it in careclinic alongside pain scores and the trendline was hard to argue with. still can’t untangle whether that’s the compound or just inflammation settling. but “no systemic effects in user reports” and “no one was tracking the right endpoints” are not the same claim, and the absence-of-evidence part of your argument leans on one being true when it might just be the other.
the “dosing isn’t reaching the proposed mechanism, the mechanism is wrong, or people aren’t tracking carefully enough” trichotomy is the most honest framing i’ve seen in one of these threads. from six weeks of wrist tracking i’d guess it’s mostly the third one - symptom diaries on reddit aren’t capturing gut motility changes or vascular response, they’re capturing “does my wrist hurt less today,” which is a pretty coarse filter for detecting VEGFR2 or NO-synthase effects. the single-lab problem you’re flagging is real and imo undersold. effect sizes that large with that few independent replications in any other field would have people asking hard questions about pre-registration and outcome selection. the functional equivalence to sham at day 14 numbers in particular look like they’re either genuinely extraordinary or the measurement is doing a lot of work. probably worth knowing which before your sister spends much on it.
the falsification framing is solid and I’d sign it, but “I don’t see those systemic effects reported much in user logs” is doing more work than it can carry. you correctly discount the rat measurement system as possibly generous. user logs are a far worse instrument for the opposite reason: they’re unblinded, uncontrolled, and almost nobody is tracking the variables the mechanism would actually predict. absence of systemic effects in that dataset isn’t a null result, it’s a null instrument. you can’t read signal out of a measurement system with no sensitivity in the first place. and there’s a candidate systemic signal people just don’t label as systemic. the subq-versus-oral sleep delta keeps coming up, and the most plausible non-placebo route there isn’t gut, it’s vagal: subq abdominal delivery hits vagal afferents differently than an oral cap that gets chewed up by gastric proteases first. that’s exactly the gut-brain axis arm of the mechanism you flagged, showing up as “slept better,” not logged as a mechanism observation. your third option, “people aren’t tracking carefully enough to notice,” is the live one ime. i log ROM in degrees weekly plus a free-text note tied to each dose so i can reconstruct what else was going on, and even that misses systemic stuff because i’m not looking for it. nobody’s running the panel the mechanism would demand. agree on the tiers though.
Six years of daily logs says the “not tracking carefully enough” option deserves more scrutiny than the OP gives it. Gut motility shifts noticeably by week two on bpc at 500mcg/day - I see it consistently across multiple courses. Sleep architecture changes that HRV captures. On two separate runs a mood stabilization that doesn’t reduce cleanly to “injury hurts less so I’m less stressed.” That’s not absence of systemic signal. It’s signal that doesn’t show up in forum posts bc people ask “did your tendon heal,” not “did your gut change” - the question shapes what gets reported. The single-lab replication problem is legitimate and the three-tier evidence breakdown is the clearest framing I’ve seen written out. No argument there. But the falsification criterion - an RCT with a clinically meaningful effect size - is a clinical standard. It’s not the right bar for a personal decision made on a specific documented injury with tracked outcomes. The OP criticizes people for collapsing three evidence tiers into one, then conflates two different decision contexts (clinical practice vs. individual risk management) without acknowledging it. Those are also not the same thing.
The systemic-effects gap you flagged is the part I can actually speak to from my own logs. Two BPC cycles for a shoulder, one running now for a wrist, all local subq, and across that whole stretch I’ve got nothing systemic worth reporting. No gut changes, no mood shift, nothing I’d have missed because I track daily. So fwiw your middle option, dosing not reaching the proposed mechanism at the amounts we self-administer, fits my n=1 better than “people aren’t watching.” The repair stayed stubbornly local.
your three-way fork (“dosing isn’t reaching the proposed mechanism, the mechanism is wrong, or people aren’t tracking carefully enough”) is the cleanest framing of this I’ve seen, but I’d weight the first branch heavier than you might. Most of those user logs are small-dose local subq right at the injury site, which is the exact protocol least likely to hit a systemic VEGFR2 or gut-brain threshold.
So the missing systemic signal isn’t surprising, it’s what you’d predict from the route. fwiw I track daily and have logged nothing systemic across multiple local cycles, no gut, no mood shift, nothing I’d have missed. that doesn’t rescue the mechanism, to be clear. it just means the absence of systemic effects in n=1 logs is weak evidence against it, since nobody’s dosing in a way that would surface it anyway.
The “I don’t see those systemic effects in user logs” line is the one I’d hold loosely. Most of those logs are running BPC alongside PT, dose changes, sleep, three other compounds, so a systemic signal would be buried under the noise, not absent from it. Absence in an unblinded n=1 that never isolated anything tells you about the tracking, not the mechanism.
edit: realized I said that wrong
the single-lab replication problem is real and the evidence tier point is the strongest thing in this post, no argument there. but the mechanism section is doing something worth pushing back on. “NO synthase modulation, VEGFR2 activation, and gut-brain axis effects” get listed like they’re parallel concerns that should all produce the same kinda detectable systemic signal. they’re not. imo if angiogenesis is actually the bottleneck in tendon and ligament repair (and there’s decent structural biology reason to think it’s in connective tissue specifically), then VEGFR2 activation in that context wouldn’t necessarily produce the kinda obvious systemic readout a self-experimenter would notice. the gut-brain axis stuff is a different mechanism class entirely, with different predictions. treating “why aren’t users reporting systemic effects” as a single falsifying question collapses two separate claims before you’ve tested either one
one slice of the Sikiric body of work that almost never makes it into the tendon-focused summaries is the CNS stuff: haloperidol catalepsy reversal, amphetamine sensitization blunting, forced swim test effects. all rodent obviously, same lab caveat applies, but it’s a totally separate axis from tissue repair and it predicts self-reportable CNS effects at the doses people are running. which actually loops back to your “I don’t see those systemic effects reported much in user logs” point from a different direction. if the dopaminergic and serotonergic modulation is real at human-relevant doses, mood and sleep deltas should be the easiest systemic readout to catch, way easier than asking someone to notice NO-mediated vascular changes from a self-report. that they’re not consistently reported is either a dose issue (oral vs subq exposure profile to the brain target is very different), the rodent assays don’t translate, or people aren’t tracking mood and sleep against injection days at all. my bet is mostly the third. most logs I see are pain, ROM, and lift numbers, and the day-by-day mood trace isn’t even captured. fwiw the weekly trend summary in CareClinic is what made the sleep-on-subq pattern visible to me on my own cycle, because eyeballing a daily journal you miss it but a 7-day rolling view shows the dip clean. that’s exactly the kind of read that would let you actually answer the question you’re raising on the CNS axis without needing a Zagreb RCT. but yeah, until someone runs registered CNS endpoints in humans, the rodent dopamine work sits at the same tier as the tendon work, just less talked about because it doesn’t sell as a “healing peptide.”
the three options you give for missing systemic effects aren’t equally weighted. most logs track the presenting injury: rom, pain score, swelling. but if the NO synthase and gut-brain axis effects are real, they’d show up in sleep architecture, gut motility, mood variance, none of which most people are logging. the absence of evidence reads differently when the logs are pointed at the wrong signal.
the “one lab owns this literature” caution is the right instinct and I won’t argue the evidence tier with you. where I’d push is the systemic-effects step. you frame the absence in user logs as a three-way fork: dosing not reaching mechanism, mechanism wrong, or people not tracking carefully. there’s a fourth. subq BPC stays put. it’s hypovascular tissue and it doesn’t distribute far from the pin, so a local shoulder injection isn’t dosing the gut-brain axis the way the rat oral models were. “reaching the proposed mechanism” systemically and reaching it at the injury site are two different questions, and the fork collapses them. the people running oral for gut symptoms are the ones who’d actually test the systemic prediction, not the subq tendon crowd. fwiw even the careful loggers mostly only chart the injury site. I’ve got pain and ROM graphed by dose-week, nothing systemic on the same view, so absence there is me not looking, not me looking and finding nothing. that’s a measurement gap, not a mechanism result. ymmv.
The three-tier framing at the end is the part most forum discussions collapse. Mechanism plus rodent data plus n=1 reports do keep getting treated as mutually reinforcing when they’re actually independent evidence streams that would each need to hold up separately. What I’d add from my own logs (six years, daily entries covering gut motility, HRV, and pain scale) is that the systemic effects are real and measurable if you’re tracking secondary markers properly before you start. Gut motility shifts by week two at 500mcg/day subcu, consistently, across multiple separate courses. HRV catches sleep architecture changes that precede any subjective improvement by several days. That’s not placebo territory if it’s reproducible across courses with off-gaps in between. But it’s also not the independent replication you’re asking for. Your falsification criterion is fair. n=1 logs, even careful ones over years, are a different tier than what you’d need to actually move the prior, and I think it’s important to say that plainly rather than paper over it.
the “people aren’t tracking carefully enough to notice” framing is where i’d push back. it’s not a granularity issue, it’s a category problem. reddit symptom diaries don’t capture NO synthase modulation or vascular response, they capture “does my wrist hurt less today.” so the absence of systemic effect reports in user logs isn’t evidence for any of your three options, it’s just the wrong instrument entirely, measuring the wrong thing in the wrong way to detect what you’re asking about. your single-lab ownership point lands harder imo. that’s the actual structural problem here, not the self-report data. six weeks logging my own wrist and i can’t tell if bpc is doing anything, partly bc i changed sleep and caffeine at the same time. which is almost exactly the confound you’re naming in the rodent work, just at a much smaller scale with no sham group.