# Four peptides, four kinds of missing evidence

> Four peptides, four kinds of missing evidence - Research Peptide Fundamentals research peptides - Checked Peptides — A side-by-side gap comparison of four Research Peptide Fundamentals research peptides: BPC-157, retatrutide, NAD+ and PT-141, sorted by what their evidence does not establish rather than what it does.

**THE LADDER**

The same phrase, limited evidence, means four completely different things across these compounds. This page sorts them by what is absent rather than by what is claimed.

## The short version

When a compound is described as having limited evidence, the sentence hides more than it says. Evidence can be limited because nobody has studied it in people at all. It can be limited because the studies were good but measured a stand-in rather than the thing that matters. It can be limited because the measurement was taken in the blood while the claim is about muscle or brain. Or it can be limited because the studies were large and rigorous and enrolled a different group of people from the ones now using the compound.

All four of those are true of something on this desk, and each one calls for a different response. The first means waiting. The second means waiting for a specific trial to read out. The third means a different measurement, not a bigger one. The fourth means the evidence exists but does not transfer.

The table below sorts the four compounds by what they are missing. The sections under it explain why the distinction changes what a reader should do with a claim.

## The evidence ladder, side by side

| Compound | Status | Strongest human evidence | Longest published exposure | What was measured | The gap in one line |
| --- | --- | --- | --- | --- | --- |
| [BPC-157](/bpc-157) | Not approved anywhere; prohibited in sport | A two-adult intravenous safety pilot [1] | A single administration [1] | Tolerability and safety biomarkers [1] | No human efficacy endpoint has ever been measured, and no human half-life exists [2][3] |
| [Retatrutide](/retatrutide) | Investigational; Phase 3 running | Phase 2 randomised trial, 338 adults [11] | 48 weeks [10][11] | Body weight, HbA1c, liver fat [10][11][12] | Every endpoint is a surrogate, no outcome trial has reported, and nothing published follows people after they stop [8] |
| [NAD+](/nad) | Sold as a supplement; NMN status contested | Randomised precursor trials of eight to ten weeks [14][15][17] | Ten weeks [15] | Blood NAD+, muscle insulin sensitivity, walking distance [14][15][17] | The rise is measured in blood while the hypothesis is about tissue, and tissue data remain sparse [13] |
| [PT-141](/pt-141) | FDA-approved 2019 for one narrow indication | Two Phase 3 randomised trials, 1,267 participants [20] | 52 weeks, open-label [21] | Sexual-desire and distress scores [20] | The studied population is narrow, the effect sizes are small, and two models disagree about the mechanism [18][19] |

Read down the fourth column and the ladder is obvious. Read across the last one and the point of this desk is: the rungs are not the same kind of rung.

## The species gap, and why animal results do not carry

BPC-157 is the clean example. Its rodent literature is broad and internally consistent: a fully transected Achilles tendon healing faster [6], gastric ulcers shrinking with an inhibition ratio of 45.7 to 65.6 percent [5], a coherent blood-vessel mechanism running through VEGFR2 [4]. Nothing about that work is sloppy.

The problem is the jump. A 2025 review states that only three pilot studies have examined BPC-157 in humans and that rigorous large-scale trials are lacking [2]. The pharmacokinetic work that would tell anyone how to design a human study was done in rats and beagle dogs, and the two species disagree about intramuscular bioavailability by roughly threefold, at 14 to 19 percent against 45 to 51 percent [3]. When two animal species disagree that much, extrapolating to a third is not conservative, it is arbitrary.

The practical consequence is worth stating plainly. Every schedule circulating in research-use communities is arithmetic performed on an elimination half-life of under 30 minutes measured in animals [3]. There is no human number to check it against.

## The endpoint gap, and the difference between a stand-in and an outcome

Retatrutide's human trials are the sort BPC-157 does not have: randomised, placebo-controlled, hundreds of participants, published in major journals [11][12]. Its gap is at the other end of the study design.

Weight, HbA1c and liver fat are surrogate endpoints. They are measured because they are fast and reliable and because long observation suggests they track outcomes that matter. But a surrogate can move in the right direction while the outcome does not follow, which is exactly why regulators require dedicated outcome trials, and why the 2025 review describing the programme keeps using the word ongoing [8].

Time is the second half of this gap and it is easy to miss. The longest published retatrutide exposure is 48 weeks [10][11], and the dose-dependent heart-rate rise peaked at around week 24, inside that window [11]. A signal that peaks halfway through the observation period leaves the rest of its shape unobserved. Nor does any published trial follow participants after treatment ends, which leaves the question people ask most often, whether the change holds, without a published answer.

NAD+ is a different case that reads similarly at a glance. Its trials also measure a stand-in, but the stand-in sits in a different bodily compartment from the claim rather than earlier on the same causal chain.

## The compartment gap, and the population gap

NAD+'s intermediate step is the best-established fact on this desk: nicotinamide riboside raised whole-blood NAD+ by 22, 51 and 142 percent across three ascending amounts over eight weeks [17], and NMN raised it significantly at every amount tested over 60 days [14]. Nobody disputes it.

What the 2025 review of the human evidence concluded is that clinical efficacy has been limited, that age-related NAD+ decline has been consistently observed only in a limited number of human studies, and that data on tissue-specific NAD+ remain sparse [13]. The measurement is real and it is in the wrong compartment for the claim. The fix is not a larger trial of the same design; it is a different measurement.

PT-141's gap is different again, and it is the one most likely to be misread as strength. It is approved. It has two Phase 3 trials with 1,267 participants [20] and a 52-week open-label extension in 684 [21]. But the approval covers acquired, generalised low sexual desire in premenopausal women, and the pharmacokinetics in the label, including the terminal half-life of about 2.7 hours, were established in that group [22]. Off-label use in men or in postmenopausal women is not covered by any of it. Evidence does not transfer between populations for free, and a strong result in one group is not a weak result in another; it is no result in another.

The effect sizes underneath the approval are also smaller than the word approved suggests, at an integrated +0.35 on the desire scale and -0.33 on the distress item [20], and whether that magnitude is noticeable to a person has been contested rather than settled.

## What the four have in common

Three things run across all four entries, and they are the reason a gap register is more useful than a benefits list.

First, in every case the popular claim is more specific than the evidence behind it. The literature says a rat tendon healed [6]; the claim says tendons heal. The literature says liver fat fell on a scan [10]; the claim says the liver was repaired. The literature says blood NAD+ rose [17]; the claim says cells were rejuvenated. The literature says desire scores improved in premenopausal women [20]; the claim says desire improves. Each step is small and each one is unearned.

Second, three of the four have an unregulated supply channel attached to them, and it is a separate risk from the pharmacology. Material sold as research-grade BPC-157, retatrutide or PT-141 has unverified identity, concentration and sterility, and a compounded NAD+ injection has already been recalled at the most serious level for bacterial endotoxin. None of that is a property of the molecules; all of it is a property of the market.

Third, and most usefully, every one of these gaps is closable by a study that could be run. That is what distinguishes an honest gap from a mystery. A human pharmacokinetic study would close the BPC-157 question [3]. An outcome trial would close the retatrutide one [8]. A tissue-level measurement would close the NAD+ one [13]. A trial in the off-label population would close the PT-141 one [22]. Naming which study is missing is a far more useful thing to hand a reader than the phrase more research is needed.

---

Checked Peptides is an independent reading desk that keeps a written record of what peptide research has not established yet: it sells nothing, prescribes nothing, and treats a missing study as a fact worth printing.