GHRP-2 has two human dose figures in the published record. They differ by purpose, by route, by schedule and by population — and the figures circulating on retail pages match neither of them.
It also has zero registered clinical trials, while being in routine hospital use. That contradiction turns out to be the most useful thing on this page, because it is a fault in the method rather than a fact about the compound.
What the compound is
GHRP-2 — also pralmorelin, also KP-102 — is a synthetic growth hormone secretagogue. It stimulates the pituitary to release growth hormone by acting at the ghrelin receptor, a mechanism distinct from growth hormone-releasing hormone, which is why the two are sometimes given together to test whether the pituitary responds to both routes.
The registry returns nothing, and that is a fact about the registry
A ClinicalTrials.gov search on 20 September 2026:
| Search term | Studies returned |
|---|---|
GHRP-2 |
0 |
GHRP2 |
0 |
pralmorelin |
0 |
KP-102 |
0 |
growth hormone releasing peptide-2 |
0 |
GHRP-6 |
0 |
hexarelin |
0 |
Seven searches, nothing. On the compounds elsewhere in this catalogue, that pattern has usually meant what it appears to mean — dihexa returns zero under three names and has genuinely never been given to a person in a registered trial.
Here it means something different, and the difference matters.
GHRP-2 is used in Japanese hospitals as a pituitary stimulation test. A 2026 paper in Endocrine Journal from St. Marianna University School of Medicine Hospital in Kanagawa analysed 69 patients who underwent testing between August 2015 and December 2024, and reports that a peak cortisol below 12 μg/dL during the GHRP-2 stimulation test was specific to hypothalamic secondary adrenal insufficiency among patients with normal cortisol responses to the corticotropin-releasing hormone test.
That is not a fringe use. It is a diagnostic test running in a university hospital for the better part of a decade, published in a national endocrinology journal — and it is invisible to the registry this site normally treats as the human record.
The explanation is mundane: Japanese clinical research registers with Japanese registries. A ClinicalTrials.gov census is a census of ClinicalTrials.gov. Nothing on this site should read a zero from it as "never studied in humans" without asking where the studies for that compound would have been filed, and this page is the correction.
The diagnostic dose
From the analytical literature: a single intravenous dose of 100 micrograms.
The source is a doping-control method paper — Reverter-Branchat, Segura and Pozo, Drug Testing and Analysis 2021;13(3):510–522 (PMID 33197153) — which validated a dried-blood-spot detection method and applied it to samples from an administration study at that dose. The compound was detectable up to four hours after injection.
That four-hour detection window is the closest thing to a duration figure this compound has. It is a detection limit rather than a half-life, and it appears on this site's half-life chart in exactly those terms.
The therapeutic dose
One trial, and it is small.
Mericq et al., J Clin Endocrinol Metab 1998;83(7):2355–60 — Effects of eight months treatment with graded doses of a growth hormone (GH)-releasing peptide in GH-deficient children:
| Period | Duration | Dose |
|---|---|---|
| 1 | 2 months | 0.3 μg/kg/day GHRP-2, subcutaneous |
| 2 | 2 months | 1.0 μg/kg/day |
| 3 | 2 months | 3.0 μg/kg/day |
| 4 | 2 months | 3 μg/kg GHRP-2 plus 3 μg/kg GHRH, subcutaneous |
Six prepubertal children with growth hormone deficiency and growth failure. Overnight 12-hour episodic GH secretion and toxicity measures were monitored at the end of each period, with IGF-I, IGFBP-3 and stadiometer height recorded.
Two things must be said about this trial together, because neither alone is honest.
It is the best human therapeutic evidence for this compound — real doses, real people, eight months, supervised, peer-reviewed. And six participants cannot establish efficacy or safety in anyone, least of all in the adults who buy the compound, who differ from the trial population in age, in pituitary function and in having no deficiency to correct. The trial tested whether a GH response could be sustained over time in children whose own GHRH signalling was inadequate. That is not the question a healthy adult is asking.
A citation trap this site has already been caught by
PMID 9661608 is the GHRP-2 trial above. It is not a GHRP-6 trial.
The two compounds are cross-cited constantly, and this site's compound library caught the same misattribution on 11 September 2026, which is why the warning sits in the keyword map. It is worth stating in public because the error propagates: a retail page cites a "GHRP-6 study" that is a GHRP-2 study, the next page copies the citation, and within a few links the dose figure has changed compound without anybody changing a number.
The general rule, which applies beyond these two: open the paper, do not trust the name attached to the citation.
Why there is no half-life
No human pharmacokinetic study of GHRP-2 was found in this site's sources, and for this class the absence has a published explanation rather than being an open question.
Esposito et al., J Pept Sci 2015;21(1):1–9 (PMID 25469748), from the Ghent University doping control laboratory, set out to evaluate in vitro models — human liver microsomes and S9 fraction — for predicting the metabolism of peptide doping agents including GHRP-2, GHRP-6, hexarelin and TB-500. The paper's stated reason for using those models is that pharmacokinetic studies with human volunteers are often not allowed because of ethical constraints.
One qualification, added on 21 September 2026 while building this compound's sibling record, because the sentence above reads more broadly than it should. That constraint is real where it applies, and it is not a property of the class. GHRP-6 has a full human pharmacokinetic study — nine male healthy volunteers, single intravenous boluses at three dose levels, a validated LC-MS assay with a stable-isotope internal standard — run at the Center for Genetic Engineering and Biotechnology in Havana and published in 2013. The figures and their limitations are on the GHRP-6 dosage record. What a European doping laboratory is permitted to do is a fact about a jurisdiction, not about a molecule, and this page previously invited the wrong inference.
So for this compound the metabolic literature is a test-tube literature by design. A chart that prints a confident half-life for GHRP-2 is not reporting a measurement in a person, and the reason is on the record.
What the two dose records do not add up to
Set them side by side:
| Diagnostic | Therapeutic | |
|---|---|---|
| Dose | 100 μg | 0.3–3.0 μg/kg/day |
| Route | Intravenous | Subcutaneous |
| Frequency | Once | Daily for 8 months |
| Population | Adults under investigation | 6 prepubertal children with GH deficiency |
| Purpose | Provoke a response to measure it | Attempt growth |
| Source | Drug Test Anal 2021 | J Clin Endocrinol Metab 1998 |
Neither is a dose for a healthy adult taking the compound to raise growth hormone, and no study in the record tested one. A 100 μg bolus given once is a measurement instrument. A microgram-per-kilogram daily ladder in six deficient children is a pilot.
The honest summary is that this compound has a real clinical use and a real trial history, and neither of them is the use it is sold for.
Method
The registry census was run against the ClinicalTrials.gov v2 API on 20 September 2026 under seven search terms. The literature was searched on PubMed on the same date; each figure above is cited to the paper it was read from, by PMID or by journal, volume and page. Where a paper is a doping-control method study rather than a clinical trial, that is stated, because the dose in it is an administration protocol for an analytical experiment and not a therapeutic recommendation.
Related records: the dosage chart grades every compound here by the tier of source behind its figure, and how to read a dosing claim sets out what a six-participant trial can and cannot support. Nothing on this page is a dosing instruction.