Peter Attia on Peptides and Longevity Drugs, Fact-Checked
A source-checked look at where Peter Attia actually stands on NMN, rapamycin, metformin, GLP-1 drugs and BPC-157 — and where popular hype runs past the evidence.
Peter Attia is broadly where the evidence is: skeptical of oral NMN/NR for longevity, cautiously interested in rapamycin, a measured skeptic on metformin, and respectful of GLP-1 outcome data. Popular hype — NAD infusions, BPC-157 miracle-healing, metformin-for-everyone — generally runs ahead of his stated positions.
Peter Attia sits at an awkward spot in the longevity conversation: he is invoked as an authority for exactly the compounds he is most careful about. Search his name and you will find him quoted in the same breath as NAD infusions, “anti-aging” NMN, and peptide healing stacks — often by people selling them. The gap between what Attia has written and what the internet says he believes is the whole story here. This piece checks the popular claims against his own primary sources and against the trial data he cites. Everything below describes laboratory, literature, and clinical-trial findings, not use in people.
Where does Attia actually stand on NMN and NR?
The most repeated misattribution is that Attia endorses oral NAD precursors. He does not. In a 2021 write-up of the Yoshino nicotinamide mononucleotide trial he concluded plainly that he “remains a skeptic of this approach for life extension,” and noted that nicotinamide riboside had already failed to extend lifespan in the NIA Interventions Testing Program.2 Two years later, reviewing a newer NMN study, he praised the dose-dependent rise in circulating NAD it demonstrated but dismissed its “biological age” and six-minute-walk endpoints as smoke and mirrors, closing that “my skepticism remains intact.”1
His position is more precise than a flat rejection, and the precision matters. He concedes two things: oral NR raises the blood NAD metabolome in humans — the foundational pharmacokinetic work by Trammell and colleagues established that6 — and the compounds appear safe apart from their cost. What he rejects is the inferential leap from “NAD went up in blood” to “aging slowed in tissues.” That distinction is where most marketing quietly collapses.
n=25 The human NMN trial Attia treated as statistically-but-not-clinically meaningful enrolled twenty-five prediabetic women.
The mechanistic backbone of his skepticism comes from isotope-tracing work out of the Rabinowitz lab. Liu and colleagues traced NAD synthesis and breakdown fluxes and found that orally administered precursors are taken up and converted to NAD largely by the liver.5 If the liver intercepts most of an oral dose, the case that circulating precursors meaningfully raise NAD in muscle, brain, or other target tissues is unproven rather than disproven. Attia has also, separately, dismissed intravenous NAD+ infusions on first principles: cells do not import intact NAD+ across the membrane, so the pitch that an IV drip “refills” cellular NAD is mechanistically incoherent regardless of how the recipient feels afterward.
Is his rapamycin position “hype”?
No — and reading it as hype gets him backwards. Attia’s documented stance is “curious, cautiously interested, and nowhere near ready to close the book,” with explicit reservations about prescribing rapamycin indiscriminately and a warning that dosing, timing, and context “will matter a great deal.”3 The compound earns his interest because it clears a bar almost nothing else does: it extended median and maximal lifespan in genetically heterogeneous mice even when started at 600 days of age, in the multi-site ITP design built specifically to weed out lab-specific flukes.7 That same program later showed the effect is dose- and sex-dependent,8 which is itself a caution, not a selling point — it means the human regimen is far from settled.
A strong geroprotector candidate needs convergent lines: robust animal lifespan data and human outcome data. Rapamycin partly clears that bar; oral NR/NMN do not.
The human rapamycin evidence remains thin and short of longevity endpoints. The strongest healthspan signal is Mannick and colleagues’ trial of the analog RAD001 improving vaccine response in older adults9 — an immune-function readout, not a mortality one. The more recent PEARL trial, one of the few placebo-controlled rapamycin studies in healthy adults, reported broadly acceptable safety over a year but no robust functional or longevity efficacy signal.10 That is exactly the picture Attia describes: a compelling animal prior, an unresolved human question.
How do the drugs rank against each other on evidence?
It helps to lay the compounds out on a single axis — how much of the enthusiasm is backed by hard human outcomes versus mechanism and mouse data.
| Compound | Strongest data | Human outcome evidence | Attia’s stance (his words) |
|---|---|---|---|
| GLP-1 agonists (semaglutide, tirzepatide) | Large RCTs on their indication | Yes — MACE reduction and weight loss | Treated as genuinely proven for indication |
| Rapamycin | Multi-site ITP mouse lifespan | Limited surrogate/short-term only | “Promising, not proven” |
| Metformin (for longevity) | Confounded epidemiology | No agnostic-of-diabetes RCT completed | Measured skeptic; backs TAME concept |
| Oral NR / NMN | Small mechanistic + PK studies | Thin; NR null in ITP | “Skepticism remains intact” |
| BPC-157 | Rodent healing studies | None (no completed human trials) | Not a documented endorsement |
“Strongest data” describes the highest-quality evidence that exists, not a recommendation. GLP-1 outcome trials are indication trials (cardiovascular risk, obesity), not lifespan trials. Rows for NR/NMN and BPC-157 rest heavily on in-vitro and animal work; nothing here is a dosing protocol or a statement about use in humans.
The GLP-1 row is worth being precise about, because it is the one place the hard data genuinely exists. The SELECT trial reported roughly a 20% reduction in major adverse cardiovascular events with semaglutide in non-diabetic overweight or obese adults,12 and SURMOUNT-1 reported up to about 20.9% mean weight loss with tirzepatide at 72 weeks.13 Those are large, pre-registered, outcome-based trials — a different evidentiary league from geroprotector mouse work. But they are trials of cardiovascular risk and obesity, not of aging biology or lifespan. Reading SELECT as proof that these drugs “extend lifespan” in the gerontological sense overstates it in the opposite direction, and Attia does not make that leap. These compounds appear here strictly as discussed literature, not as anything Condor Research merchandises.
What about metformin and BPC-157?
Metformin is where Attia is most easily caricatured. He is not anti-metformin; he is skeptical of the longevity case specifically. He discusses Konopka and colleagues’ RCT showing that adding metformin to aerobic training blunted improvements in skeletal-muscle mitochondrial respiration and cardiorespiratory fitness in older adults11 — a reason to hesitate before handing it to fit people who exercise. He also notes that the widely cited 2014 observational finding of lower mortality in metformin-treated diabetics failed to replicate in later work. Yet he still supports the TAME trial concept, because the only honest way to resolve the question is a properly designed RCT that tests metformin agnostic of diabetes — and no such trial has completed.
BPC-157 is the peptide most often laundered through Attia’s name, usually with claims he has never made. “Body protection compound-157” is a synthetic 15-amino-acid partial sequence derived from a human gastric juice protein. Its musculoskeletal and healing evidence is almost entirely rodent: an independent 2025 narrative review — notably from outside the originating laboratory — confirmed that the claims rest on animal studies with no completed human clinical trials.14 The regulatory status is unambiguous: the FDA placed BPC-157 in 503A Category 2, citing significant safety risks including immunogenicity and peptide-characterization concerns, which means it is not an approved drug and is not eligible for pharmacy compounding.15
An honest read of the evidence
The fair conclusion is that Attia is generally where the evidence is, and the honest weaknesses cut in every direction. On the pro-NMN side, the human data really are thin: the Yoshino trial was n=25, single-endpoint (muscle insulin sensitivity), showed no change in body composition or other metabolic markers, and leaned on an outlier responder4 — which is precisely why treating it as clinically meaningful is a reach. But Attia’s own prior deserves the same scrutiny. The liver-uptake argument comes from mouse and tracer work5; extrapolating “oral precursors only help the liver” into a definitive ruling against any human tissue benefit is stronger than the data strictly license. He frames it, correctly, as a strong prior rather than proof.
Rapamycin’s problem is the species gap. The compelling lifespan data are in mice78; the human record is limited to surrogate and short-term trials — immune function, one-year safety — with no longevity or mortality endpoint, and effects that are dose- and sex-dependent even in the animals.910 Metformin’s longevity case is worse: confounded observational epidemiology that failed to replicate, plus null or poorly powered RCTs, with the one trial designed to settle it still unfinished. BPC-157’s efficacy claims trace overwhelmingly to a single research lineage, which raises independent-replication and immunogenicity concerns that no human trial has yet addressed.14 And the GLP-1 trials, strong as they are, are indication trials — not evidence that these drugs slow aging. Overselling in either direction is the error to avoid.
All materials supplied by Condor Research are Research Use Only (RUO). Everything above summarizes in-vitro, animal, and clinical-trial literature and publicly documented commentary; none of it is a dosing protocol, clinical guidance, endorsement, or safety assessment for any organism. Nothing here should be read as a recommendation to acquire or administer any compound.
Condor Research · Scientific desk
Atrio Sciences s.r.o., IČO 57 669 651, Nitra (SK) · info@condorresearch.com
- Attia's oral NMN/NR skepticism is documented in his own words: he calls the longevity case unconvincing and treats the human data as statistically-but-not-clinically meaningful.
- His mechanistic argument leans on isotope-tracing work showing oral NAD precursors are taken up mainly by the liver, so tissue-wide benefit remains unproven; he separately calls IV NAD+ infusions implausible.
- On rapamycin he is 'promising, not proven' — he cites the multi-site ITP mouse lifespan data as the strongest signal while stressing that human dosing, timing and context are unsettled.
- On metformin he is a measured skeptic-for-longevity: he flags evidence that it blunts exercise adaptations and that the 2014 mortality claim failed to replicate, while still backing the TAME trial concept.
- GLP-1 agonists sit on the strong-evidence end for their approved indications — cardiovascular risk and obesity — with hard-outcome trials, not lifespan trials.
- BPC-157's musculoskeletal evidence is almost entirely rodent, with no completed human clinical trials, and the FDA placed it in 503A Category 2.
- Attia's evidentiary spine is convergent lines — robust animal lifespan data plus human outcomes — a bar rapamycin partly clears and oral NR/NMN do not.
Did Peter Attia say NAD IV drips don't work?
He has argued that intravenous NAD+ infusions are mechanistically implausible, because cells cannot import intact NAD+ across the membrane. His broader point is that raising a molecule in the bloodstream does not establish that it reaches or acts inside target cells — the same critique he applies to oral NAD precursors.
Does Attia take rapamycin himself?
His public writing describes an interest in the compound and its trial data, but frames it as "promising, not proven" and warns explicitly against indiscriminate prescribing. He treats human dosing, timing, and context as unresolved questions rather than settled practice, and does not present it as a recommendation for others.
Is the NMN trial data really that weak?
The most-cited human NMN trial enrolled twenty-five prediabetic women and improved muscle insulin sensitivity over ten weeks, with no change in body composition or other metabolic markers. A single positive surrogate endpoint in a small sample is a starting point, not proof of a longevity benefit — which is why Attia calls the effect statistically but not clearly clinically significant.
Why is BPC-157 so controversial if studies show it heals tissue?
Nearly all of those studies are in rodents, and a large share come from the single laboratory that first described the peptide, so independent replication is limited. There are no completed human clinical trials, and the FDA has flagged safety and characterization concerns by placing it in 503A Category 2. "Studies show" is doing a lot of work when the studies are preclinical and single-lineage.
Does Attia think metformin is bad?
Not as a diabetes drug. His skepticism is specific to the longevity claim in healthy people. He cites evidence that metformin can blunt exercise-induced mitochondrial and fitness gains in older adults and notes that the headline mortality finding did not replicate, while still supporting a proper trial to answer the question.
Are GLP-1 drugs longevity drugs?
Their strongest evidence is for cardiovascular risk and obesity, not aging. SELECT showed a roughly 20% cut in major adverse cardiovascular events with semaglutide and SURMOUNT-1 showed large weight loss with tirzepatide, but these are indication trials. Calling them "longevity drugs" overstates what the trials tested.