A volunteer on our team — call him R., a 41-year-old with no diagnosed sleep disorder — spent six nights injecting a growth-hormone-releasing peptide sold online as a recovery and "deep sleep" aid. We picked it because it is among the most common compounds circulating in the recovery-and-longevity corner of wellness culture, and because its marketing makes a testable claim: that peptides like it deepen slow-wave sleep and raise the biomarkers associated with overnight repair. We ran the test not to endorse or condemn the practice, but to ask a narrower question: when someone takes one of these compounds and reports sleeping better, what is actually being measured, and does the measurement mean what the seller says it means?
The verdict, in one sentence: the peptide produced a real, measurable shift in one overnight biomarker, no detectable change in the metric most users actually care about, and nothing that resolves whether any of it improves health.
How we ran the test
This was a single-subject protocol, and we will say plainly up front: n=1 proves nothing on its own. We designed it as a demonstration of the measurement gap, not as evidence of efficacy.
R. ran a four-night baseline (no compound), then six nights on a GH secretagogue at the dose printed on the vial from a compounding source, injected roughly 30 minutes before bed. Across all ten nights we collected three categories of data:
- A wearable. A wrist device estimating sleep stages from heart rate, movement, and skin temperature — the kind most readers already own and the kind influencer claims usually cite.
- A home EEG headband. A consumer device that records actual brain electrical activity, our closest available proxy for the slow-wave sleep the compound is supposed to increase. It is not clinical polysomnography, and it underestimates deep sleep relative to a sleep lab, but it measures the right signal rather than inferring it.
- A morning blood draw on two baseline mornings and two on-compound mornings, assayed for IGF-1, the downstream marker that growth-hormone secretagogues are designed to raise.
We had no polysomnography reference, no placebo, and no blinding — R. knew what he was taking. Those are real limitations, and they cap how far any of this generalizes. What the design does isolate cleanly is the relationship between three different things we are all tempted to treat as one.
What we found
The blood result was unambiguous. R.'s IGF-1 rose from a baseline mean of 148 ng/mL to 211 ng/mL on the compound — a 43 percent increase, well outside assay noise. The peptide did the biochemical thing it is built to do. If the goal were "move this biomarker," the trial succeeded.
The EEG headband showed a smaller, messier signal: slow-wave sleep in the first two cycles rose by an estimated 9 to 14 minutes per night on three of six nights, and was flat or slightly down on the other three. Night-to-night variance in the baseline period was nearly as large. We would not call that a clean effect.
The wearable — the device most people would use to judge whether the compound "worked" — showed essentially nothing. Its "deep sleep" estimate moved within its own baseline range. R.'s subjective rating ("I slept hard") tracked the wearable on some nights and the EEG on others, with no consistent pattern.
So: a clear biomarker change, an ambiguous brain-signal change, no consumer-metric change, and a subjective sense of improvement that did not reliably correspond to any of them. That divergence is the entire story.
Why a moving biomarker is not a working drug
The published evidence on therapeutic peptides splits sharply, and the split matters more than any single result. Some peptides are among the best-validated drugs in medicine: insulin replaced a death sentence, and GLP-1 agonists like semaglutide have large, randomized, hard-outcome trials behind them — weight, cardiovascular events, measured in tens of thousands of people. These are peptides too. The category is not the problem.
The recovery-and-sleep compounds circulating now are a different evidentiary animal. For the growth-hormone secretagogues, the human literature is thin and mostly old, concentrated in small Phase I and II safety studies rather than trials powered for sleep or longevity endpoints. The compounds frequently cited alongside them for tissue repair rest largely on animal work — rodents, and in some cases racehorses — which establishes a mechanism worth investigating, not a result you can bank on in a human body.
Here is the trap, and it caught our own trial. Raising IGF-1 is a surrogate endpoint — a number that stands in for a hoped-for outcome. The history of medicine is a graveyard of surrogates that moved beautifully while patients did no better, or did worse. Drugs have raised "good" markers and increased mortality. A 43 percent IGF-1 jump tells you the compound is biologically active. It does not tell you whether that activity makes you healthier, and chronically elevated IGF-1 carries its own debated associations with cancer risk that no recovery influencer is measuring on their followers.
The three things people conflate
| What's measured | What it actually is | What it can't tell you |
|---|---|---|
| Wearable "deep sleep" minutes | An algorithm's estimate from heart rate and motion | Whether real slow-wave sleep changed at all |
| Lab biomarker (e.g. IGF-1) | A genuine biochemical effect of the compound | Whether the effect improves any health outcome |
| Clinical outcome | Fewer events, longer life, measured function | Available only from large, controlled, long trials — which don't yet exist here |
Most of the confidence in these compounds lives in the left two columns and is spoken about as though it came from the right one. A wearable telling you that you got more deep sleep is the weakest of the three signals — it is an inference layered on an inference — yet it is the one that drives the most purchases, because it is the one people see every morning.
Who this is for, and who it isn't
This experiment, and this skepticism, is for the reader who has already heard the recovery-peptide pitch from a podcast or a clinic and wants to know what the numbers underneath actually support. The honest answer is: a real biochemical effect, a real safety question, and an absence of the outcome data that would justify the longevity framing.
It is not for someone hoping we will green-light self-injection. We can't, and not out of institutional reflexivity. The distrust that pushes people toward these compounds — the slow timelines, the dismissals, the sense that conventional care optimizes for billing over healing — is often earned. But the response to a system that under-measures you is not a compound that measures the wrong thing. If you pursue this, do it with physician oversight, a sourced product rather than a gray-market vial, and at minimum a before-and-after blood panel, so you are running something closer to a test and less close to a guess.
Evidence grade
For the central claim — that recovery-and-sleep peptides like the one we tested improve sleep in a way that benefits health — we grade the evidence Weak. The mechanism is plausible and the biomarker effect is real; the human outcome data are absent. Weak is not the same as wrong. It means the question is open and the marketing has run far ahead of it.
Back to R.
By the second week off the compound, R.'s IGF-1 had drifted back toward baseline and his wearable looked exactly as it had before he started. He said he felt no different either way, which is its own kind of data point. The biomarker moved; the man, as far as we could measure, did not.
Which leaves the question our instruments couldn't answer, and neither can the current literature: if a compound reliably shifts the chemistry of overnight repair but changes nothing you can feel and nothing we can yet tie to living longer or better — is that a medicine waiting for its trial, or a number waiting to disappoint us?