saw this come up again and the thread split into “this changes everything” vs “snake oil,” which is the usual flattening. neither is the honest answer. a few things i’d separate out before anyone gets excited or dismissive. 1. “stem cells repair tissue” and “stem cells reverse aging” are not the same claim. the first has real mechanism behind it in specific contexts (some hematologic, some orthopedic). the second is a much bigger jump and the data doesn’t carry it yet. a compelling mechanism is not a proven outcome. i keep saying this about peptides and it applies here identically. 2. what’s the actual metric. “reverses aging” measured how. an epigenetic clock reading, a biomarker panel, grip strength, subjective energy, a mortality curve? those diverge. a clinic showing improved inflammatory markers in 8 patients is showing you improved inflammatory markers in 8 patients, not a reversed aging process. n matters, and in this space n is usually tiny and the people in it are desperate. 3. the clinic vs trial gap is the whole story. the registered trials are narrow and slow on purpose. the offshore clinics charging 20k are not running the trials, they’re selling the headline the trials haven’t earned yet. so: promising in defined repair contexts. genuinely unsettled as a longevity intervention. the gap between feeling better and knowing why is exactly where the marketing lives. ymmv, not my field, just the same epistemology i apply to everything.
the part I’d push on is buried in your point 2, where you list “an epigenetic clock reading” alongside grip strength and a mortality curve as if they sit on the same shelf. they don’t, and the clock is the one that flatters itself the most. those clocks were trained to predict chronological age (or in the second-gen ones, time-to-death in a specific cohort), so they’re calibrated surrogates, not a readout of aging itself. which means a clinic that “lowers your epigenetic age by 4 years” has moved a proxy that was never validated to move in the direction interventions push it. the regression weights don’t know what your stem cell infusion did, they only know what correlated with age in the training set. so the conflation you’re flagging at the clinic-vs-trial level also lives one layer up, inside the metric you’d reach for to settle it. agreed on everything else. the honest version of the question isn’t “did the clock move,” it’s “is this clock causal or descriptive,” and for every clock I’ve read the methods on the answer is still descriptive. ymmv.
edit: forgot to add
the lumping in “for every clock i’ve read the methods on the answer is still descriptive” is where i’d push. horvath/hannum, sure, trained on chronological age, pure surrogate. but grimage and phenoage weren’t, they were trained on mortality and clinical-biomarker panels, so they’re outcome-anchored in a way the first-gen clocks aren’t. that doesn’t make them causal, agreed, but it makes “descriptive” do a lot of work across objects that were built differently. the bigger thing is i don’t think causal vs descriptive is even the axis that settles it. a clock could be perfectly causal for the endpoint it was trained on and still tell you nothing about lifespan, bc the training endpoint isn’t the thing you actually care about. so “is the clock causal” and “did the intervention extend healthspan” can both resolve and still not touch each other. the gap doesn’t close, it just moves again. ymmv.
most infused MSCs never reach the tissue people imagine them repairing. the biodistribution data on IV mesenchymal stem cells is pretty unkind here: a large fraction gets trapped in the pulmonary capillary bed on first pass and the circulating signal is mostly gone within hours to a couple days, depending on the model. so whatever benefit shows up at week 8 in those small clinic cohorts, it almost certainly isn’t the injected cells sitting in the knee or the heart doing structural work that long. the leading mechanistic story has quietly shifted to paracrine signaling, the cells dump exosomes and cytokines, briefly nudge the local immune environment, then clear. which matters for this thread bc it splits a conflation nobody’s named yet: “stem cells engraft and rebuild tissue” vs “stem cells transiently signal and then die.” those are different products with different durability, and the marketing leans on the first while the kinetics support the second. the reason it’s worth separating is the dosing implication. if the effect is a short paracrine pulse rather than durable engraftment, then a single 20k infusion is pharmacologically closer to a one-time anti-inflammatory hit than to a regenerative reset, and you’d expect the readout to decay on whatever clock the secreted factors run on, not hold. that’s testable, btw, you’d look at whether the inflammatory-marker improvement in those 8-patient cohorts persists past the point where the cells are demonstrably cleared, or reverts. i haven’t seen a clinic publish that follow-up curve, which tells you something about what they expect it to show. none of this touches the clock debate you two are having, it’s a layer below it. before you even get to “did the intervention move grimage,” there’s “was the intervention present long enough to plausibly cause a durable change,” and for IV stem cells the honest answer is mostly no, it was a transient exposure dressed up as a permanent one. not my field on the longevity endpoints, but the biodistribution work is solid and it’s the part that never makes the brochure. ymmv.