The short version

Plain English first. Everything below this section is the same material with the sources, the numbers and the caveats attached.

What DSIP is

DSIP is a nine amino acid peptide — Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu, or WAGGDASGE — isolated from rabbit brain in the 1970s and named for what it appeared to do to an EEG trace. Regulators now call it emideltide. It is a plain linear peptide with no disulfide bridge and no amidation, which makes it one of the simpler molecules we stock.

Both columns come from FDA's own 2026 evaluation. UNII for the free base is YN28Z5YZ73.

The supply chain cannot reliably identify this molecule

We check vendor chemistry as a matter of routine. DSIP produced the worst single example we have found for any compound in this library.

ProSpec lists a product titled "Delta Sleep Inducing Peptide | DSIP" whose own stated sequence is H-Gly-Glu-Pro-Pro-Pro-Gly-Lys-Pro-Ala-Asp-Asp-Ala-Gly-Leu-Val-OH, formula C₆₂H₉₈N₁₆O₂₂, molecular weight 1419.55 Da. That is a fifteen residue peptide. DSIP has nine residues and weighs 848.8 Da. Whatever is in that vial, it is not DSIP, and the listing says so in its own specification table. We confirmed it on two separate URLs.

Less dramatically but more commonly: Cayman Chemical and Molecular Depot both sell a product labelled acetate while printing 848.8 g/mol, which is the free base figure, with the acetate left as an unresolved multiplier. We could not find a single retail DSIP listing anywhere that states whether its milligram figure is peptide content or gross salt mass. On a 10 mg vial that difference is roughly 0.6 to 1.2 mg of actual peptide.

How it is said to work

Nobody knows. That is not us being difficult — it is the settled position of the peer-reviewed literature, and it is unusual to be able to say it this cleanly.

There is no DSIP receptor. There is no DSIP gene.

The standard review of this compound is Kovalzon and Strekalova, 2006, in the Journal of Neurochemistry, titled "Delta sleep-inducing peptide (DSIP): a still unresolved riddle." It lists, as bare statements of fact:

Fifty years after isolation there is no receptor, no gene, no precursor protein. One 1983 paper localised tritiated-DSIP binding sites in cultured rat brainstem neurons; binding sites are not a receptor, and nothing came of it in the four decades since.

The same review goes further and questions whether the molecule is real as an endogenous substance: "there is still no strong evidence of the natural occurrence of DSIP." Almost all the evidence for "DSIP in human plasma and CSF" is antibody immunoreactivity rather than isolation of the peptide — and anti-DSIP antisera are known to bind other things. A 1993 paper in European Journal of Biochemistry describes a seventy-seven residue porcine brain protein "recognized by an antiserum to delta-sleep-inducing peptide." The title alone makes the point.

What the founding experiment actually was

Worth knowing, because it is nothing like an injection. In the discoverer's own words:

In plain terms: rabbits were put to sleep by running current through their thalamus, blood was dialysed out of a vein in their skull, and that fluid was infused directly into the brain ventricles of other rabbits. Synthetic DSIP was later tested the same way — intraventricular infusion at 6 nmol/kg, producing a 35% rise in delta activity across 61 animals. Direct brain infusion in a rabbit is the origin of every claim on this page.

The endocrine claims are rat claims, and humans were tested

You will read that DSIP raises growth hormone and modulates cortisol. Both trace to rat work with hypothalamic and pituitary sites of action. Both were then tested in people:

The titles state the findings. When the rat endocrinology was checked in humans, it did not reproduce. We should say plainly that we have verified these two papers exist, in those journals, under those PMIDs, across several independent indexes, but we could not open the abstracts themselves — so we cannot tell you the sample sizes or doses.

Claims we could not source at all

Three mechanisms appear on vendor pages with no primary study behind them that we could find: NMDA or glutamate receptor modulation, GABA-A modulation, and blood pressure normalisation. On the last one the documented human signal runs the other way — see the safety section.

The opioid claim deserves its own note. It rests almost entirely on the title of a 1983 paper: "a neuropeptide with potential agonistic activity on opiate receptors." That activity was postulated from clinical impressions. The paper contains no binding data, no antagonist challenge and no pharmacological assay. A famous title is not a mechanism.

Pharmacokinetics

There is no human pharmacokinetic study of DSIP. No half-life, no bioavailability, no clearance. FDA's 2026 briefing document has no pharmacokinetics section at all — which is itself the finding, because FDA would have reviewed one had it existed.

The "fifteen minute half-life" is not a half-life

Almost every page selling this compound quotes it. We followed the citation chain to the end, and it terminates in a 1984 narrative review by the peptide's own discoverer. Here is the sentence those pages are drawing on, verbatim:

That is the rate at which enzymes in ground-up brain tissue in a dish clip the tryptophan off the front of the molecule. It is not a plasma measurement, not an in vivo measurement, and not human. It has been laundered through secondary sources into "DSIP has a 15-minute half-life," and now everyone repeats it.

A competing figure of "roughly 30 to 40 minutes in humans, based on early pharmacokinetic work" also circulates. We could not find the work it refers to. No citation is given anywhere in the chain, and it contradicts the fifteen-minute number.

The same 1984 review contains a detail nobody selling DSIP mentions, and it cuts the opposite way: endogenous DSIP-like material in plasma, urine and CSF was found bound to a larger carrier protein and thereby protected from proteolysis. If that is right, free-peptide degradation rates may not describe the endogenous molecule at all.

One primary paper exists that would settle part of this — "Degradation and aggregation of delta sleep-inducing peptide (DSIP) and two analogs in plasma and serum," Peptides 1987, PMID 3628078. We could not retrieve it. We are telling you that rather than quoting a number we have not checked.

Blood-brain barrier

This one holds up better than most of the page. Passage has been shown in dogs and in a vascularly perfused guinea pig brain preparation, by more than one group, and FDA's briefing accepts that emideltide is "a peptide that crosses the blood-brain barrier." Two caveats worth keeping: some of that work used a radioiodinated analogue rather than DSIP itself, and the magnitude has never been quantified in a human.

What the evidence shows

This is the section that makes DSIP unusual, so we want to be exact about it. There are real human trials here — randomised, double-blind, placebo-controlled, with polysomnography. That is more than almost anything else in this library has. The trials came back negative.

The best-designed sleep study, and its own conclusion

Bes, Hofman, Schuur and Van Boxtel, 1992, in Neuropsychobiology. Sixteen chronic insomniac patients, double-blind, matched pairs, parallel groups, five consecutive nights in a sleep laboratory, intravenous DSIP at 25 nmol/kg against a glucose placebo. Their conclusion, verbatim:

The other genuinely controlled trial, Monti and colleagues in 1987 — six patients, randomised double-blind placebo-controlled crossover, IV 25 nmol/kg — found, in FDA's summary, "no significant differences compared to baseline or placebo."

Where the enthusiastic numbers come from

They come from the studies without controls. This is the whole pattern of the DSIP literature in one table:

The rule is clean enough to state as one: every study that controlled for placebo found little or nothing; every study that did not control found a great deal.

Two further things about the withdrawal figures, since they are the most quoted numbers attached to this compound. Dick 1984 excluded 13% of its alcohol patients and 22% of its opiate patients from the analysis, and outcomes were assessed by "clinical evaluation by the physician and the nursing staff" with no standardised scale. The 97% is a real quotation from a real paper. The design cannot carry it.

Registered trials

None. We searched ClinicalTrials.gov, and found no DSIP or emideltide entry — not completed, recruiting, terminated or withdrawn. FDA's review identifies none either. The most recent human study of any kind is from 1998.

On the animal side, Kovalzon and Strekalova are equally direct: "A number of studies failed to confirm SWS- or PS-promoting properties of DSIP... only minor effects were found in others, and even an impairment of sleep has been reported." They describe the sleep-factor hypothesis as "extremely poorly documented and is still weak."

Reconstitution

Read this section before you add water. DSIP has a genuine solubility problem that most compounds do not, and a 10 mg vial sits right in the middle of it.

The acetate salt is the way out of this, at roughly 10 mg/mL in buffered saline. Which means the single most useful question you can ask a DSIP supplier is which form they sent you — and, as noted above, we could not find one retail listing anywhere that answers it unprompted.

If your vial is acetate

Against the 100 to 300 mcg figures that circulate:

At 2 mL you are at 5 mg/mL, which is half the acetate's stated solubility and ten times the free base's. At 5 mL you are at 2 mg/mL — still four times the free base figure. There is no dilution of a 10 mg vial into a sane volume that clears the free base solubility limit.

On supply: a 10 mg vial is a lot of DSIP against a 250 mcg dose — forty doses, which is well past the 28 days bacteriostatic water's preservative is validated for. Reconstitute part of the vial at a time. Cayman's own datasheet for the acetate goes further and says an aqueous solution should not be stored more than one day. The supply planner will work the dates out.

Reported protocols

The circulating figure is 100 to 300 mcg subcutaneous, thirty to sixty minutes before bed. We traced it. It has no derivation from any study, and the most candid vendor page says so itself.

What the trials actually used, and how far it is from what is sold

Every human study used intravenous dosing at 25 to 35 nmol/kg. At 848.8 g/mol, 1 nmol is 0.8488 mcg, so for a 70 kg adult:

So the clinical dose is around 1.5 mg intravenously. The marketed dose is 0.1 to 0.3 mg subcutaneously. That is five to fifteen times lower, by a different route, with no bioavailability data anywhere to bridge the two.

Put on this vial, the gap is easy to see: at the clinical intravenous dose a 10 mg vial holds about six doses. At the marketed subcutaneous dose it holds thirty-three to a hundred. The marketed protocol is not a scaled translation of the clinical evidence. It is a different regimen with no evidentiary connection to it.

One vendor claim to set aside: a "flat dose-response curve above roughly 500 mcg." Not only is it unsourced, it inverts the only primary statement on the subject — Schoenenberger reported a parabolic dose-response with "different effective optima," meaning more is not simply the same, and may be less.

Storage

Lyophilised, sealed, cold, dark and dry. Our storage guide covers the general conventions.

FDA's record gives −20 °C for both forms, with the free base stable up to three years and the acetate four years or more. That is colder than the fridge temperature some peptides tolerate, and it is worth honouring.

Reported effects

A few hundred patients, all intravenous, all short-term, mostly unblinded, from the 1980s and 1990s. Nothing since 1998. And for the route being sold, nothing at all.

What was actually reported in humans

The hypotension finding deserves emphasis, because it is the clearest case we have seen of a paper and its regulator reading the same data differently. Dick and colleagues summarised their own tolerability as good, "aside from headaches reported by a few patients." FDA, reading the same study, counted three serious adverse events. Note also that "normalises blood pressure" appears on vendor pages with no source at all — while the one documented human cardiovascular signal is a blood pressure drop severe enough to be logged as serious.

The toxicology package is empty in all five domains

FDA, verbatim: "the nominators did not submit, and FDA did not identify, in the literature nonclinical toxicity studies to inform safety considerations." No acute toxicity, no repeat-dose, no genotoxicity, no developmental or reproductive toxicity, no carcinogenicity. In any species.

FDA added a specific gap that follows from the opioid claim: "stimulation of the opioid system by emideltide could lead to development of addiction, FDA did not identify in the literature nonclinical studies to inform the potential addictive liability." The mechanism vendors advertise as a benefit is the one a regulator flagged as an unstudied risk.

The certificate distinction, which is our kind of problem

FDA reviewed the certificates submitted with the nomination and made a criticism worth carrying across to how you read any peptide certificate, ours included:

"However, there is no information on the specified impurity limits/testing results as attribute control in the withdrawn nomination CoA to demonstrate quality control of impurity profile." It also noted that "no bioburden/endotoxin test is mentioned in the CoA" — and that endotoxin testing is a critical attribute for anything intended for injection — and that residual solvent testing was absent.

Three different questions, routinely collapsed into one number. Purity says what share of the peptide present is the right peptide. Impurity profile says what else is in there. Endotoxin says whether it is safe to put in a body. A certificate showing 99% purity has answered the first and said nothing about the other two.

Regulatory status

Approval. None, anywhere. FDA found no monograph in the United States Pharmacopeia or National Formulary, none in the European Pharmacopoeia (11th edition), none in the Japanese Pharmacopoeia (18th), none in the International Pharmacopoeia. Its GlobalEdge search concluded that emideltide free base and acetate "are not registered drugs in any country searched."

United States — the one the committee agreed with FDA on

DSIP, as emideltide, went before FDA's Pharmacy Compounding Advisory Committee on 24 July 2026 under docket FDA-2025-N-6895. FDA's own conclusion, verbatim:

What happened next is what makes DSIP the odd one out of that meeting. Across seven peptides, FDA's review staff recommended against all seven. The committee overruled its own scientists on six of them — and agreed with them on DSIP.

Both the free base and the acetate were voted on separately, and both came back 6–7–1. Of every peptide reviewed that week, this is the only one where a committee inclined to be generous to compounders looked at the evidence and declined.

Sourcing note, because it matters. FDA's written conclusion and the voting questions are from FDA's own published documents. The numeric tallies come from independent legal and industry analyses — four of them, in agreement — because FDA had not posted a transcript or minutes carrying the counts when we checked. The direction of the vote is confirmed several times over; the exact numbers rest on secondary reporting.

None of this changes the law. A committee recommendation is advice, and FDA has issued no final rule. DSIP's formal status is unchanged.

Canada

Health Canada's public advisory of 9 April 2026 names DSIP directly, in its list of unauthorized injectable peptides seized by Health Canada. The advisory states that peptides function as prescription drugs in Canada and that research-use-only labelling "does not make these products legal or exempt from regulatory requirements." Sleep is one of the marketing categories it names specifically.

Sport

DSIP is not named on the 2026 WADA Prohibited List — we searched it for "DSIP", "delta sleep", "delta-sleep-inducing" and "emideltide". That is not permission. Section S0 prohibits at all times "any pharmacological substance which is not addressed by any of the subsequent sections of the List and with no current approval by any governmental regulatory health authority for human therapeutic use." FDA has established the predicate: not a registered drug in any country searched. An athlete should treat DSIP as prohibited.

Sentient Biogenix supplies DSIP for laboratory research use only. It is not for human or veterinary use.

Questions we actually get

Does DSIP actually help you sleep?

On the controlled evidence, no. Two trials were properly blinded and placebo-controlled. One found no difference from placebo. The other found weak effects its own authors suspected were an artefact of the placebo group, no change in how patients rated their own sleep, and concluded DSIP "is not likely to be of major therapeutic benefit." The strongly positive results all come from studies with no control group.

Then why is it called delta sleep-inducing peptide?

Because of what it did to a rabbit's EEG in the 1970s when infused directly into the brain ventricles. The name records a hypothesis from the founding experiment, not a result from a clinical trial. The standard review of the compound calls the sleep-factor hypothesis "extremely poorly documented and is still weak."

Is the fifteen-minute half-life real?

Not as a half-life. The sentence it comes from describes how fast brain tissue in a dish clips tryptophan off the molecule. There is no human pharmacokinetic study of DSIP at all — FDA's 2026 review has no pharmacokinetics section, because there was nothing to review.

Does it raise growth hormone?

In rats, apparently yes. In humans it was tested and the paper is titled "Delta sleep-inducing peptide administration does not influence growth hormone and prolactin secretion in normal women." The same happened with the cortisol claim.

My vial won't dissolve properly. What did I do wrong?

Possibly nothing. The free base dissolves at only 0.5 mg/mL, so a 10 mg vial would need 20 mL of water to fully dissolve — more than anyone puts in a vial. The acetate goes to about 10 mg/mL. If it will not clear, add more water, do not shake it, and ask your supplier which form it is. That question is worth asking before you buy.

Where does 100 to 300 mcg before bed come from?

Nowhere we could find. No page that states it cites anything, and the most candid one says outright that the community doses come from "extrapolation." For comparison, the dose used in every human trial works out to about 1,500 mcg — intravenously.

Should the FDA vote change how I think about this?

It is worth knowing, mostly for what it says about the peptides around it. A committee that was willing to overrule FDA's own scientists on six compounds in two days looked at DSIP and did not. It changes nothing legally. But if you were reading that meeting as a signal, this is the one compound where the signal points the other way.

References

1. US Food and Drug Administration. FDA Briefing Document, Pharmacy Compounding Advisory Committee — Emideltide (DSIP), July 2026. fda.gov — chemistry and solubility of both forms, the formulation feasibility criticism, the absence of toxicology in all five domains, the FAERS and compounding-history searches, the certificate-of-analysis findings, monograph and registration status, and FDA's conclusion against listing.
2. US Food and Drug Administration. Presentation slides, Pharmacy Compounding Advisory Committee, 24 July 2026. fda.gov — "no data to support use via the nominated SC route", "SC route: no safety data whatsoever", the FAERS result, and the serious adverse events from Dick 1984.
3. US Food and Drug Administration. Voting questions, PCAC July 2026. fda.gov — the separate free base and acetate votes.
4. Bes F, Hofman W, Schuur J, Van Boxtel C. Effects of delta sleep-inducing peptide on sleep of chronic insomniac patients. A double-blind study. Neuropsychobiology 1992;26(4):193–197. PMID 1299794. pubmed.ncbi.nlm.nih.gov — the best-designed trial, and the source of the quoted conclusion.
5. Kovalzon VM, Strekalova TV. Delta sleep-inducing peptide (DSIP): a still unresolved riddle. Journal of Neurochemistry 2006;97(2):303–309. onlinelibrary.wiley.com — the absence of receptor, gene and precursor; the doubt over endogenous occurrence; the failures to replicate in animals.
6. Schoenenberger GA. Characterization, properties and multivariate functions of delta-sleep inducing peptide (DSIP). European Neurology 1984;23(5):321–345. PMID 6548966. karger.com — the founding rabbit experiment in the discoverer's own words, the "15 min" sentence in its actual context, the carrier-protein binding, and the parabolic dose-response.
7. Schoenenberger GA, Maier PF, Tobler HJ, Wilson K, Monnier M. The delta EEG (sleep)-inducing peptide (DSIP). XI. Amino-acid analysis, sequence, synthesis and activity of the nonapeptide. Pflügers Archiv 1978;376(2):119–129. PMID 568769. pubmed.ncbi.nlm.nih.gov
8. Kaeser HE. A clinical trial with DSIP. European Neurology 1984;23(5):386–388. PMID 6391926. pubmed.ncbi.nlm.nih.gov — the open-label n=7 study behind "sleep was normalized".
9. Dick P, Costa C, Fayolle K, Grandjean ME, Khoshbeen A, Tissot R. DSIP in the treatment of withdrawal syndromes from alcohol and opiates. European Neurology 1984;23(5):364–371. PMID 6548969. karger.com — the source of the "97% and 87%" figures, its exclusions and its assessment method.
10. Backmund M, Meyer K, Rothenhaeusler HB, Soyka M. Opioid detoxification with delta sleep-inducing peptide: results of an open clinical trial. Journal of Clinical Psychopharmacology 1998;18(3):246–249. PMID 9617990. pubmed.ncbi.nlm.nih.gov — the most recent human study of DSIP by any route.
11. Delta sleep-inducing peptide administration does not influence growth hormone and prolactin secretion in normal women. Psychoneuroendocrinology. PMID 8475226. pubmed.ncbi.nlm.nih.gov
12. Delta-sleep-inducing peptide does not affect CRH and meal-induced ACTH and cortisol secretion. Psychoneuroendocrinology. PMID 7777652. pubmed.ncbi.nlm.nih.gov
13. Health Canada. Think twice before injecting peptides bought online: unauthorized products can seriously harm you. Public advisory, 9 April 2026. recalls-rappels.canada.ca
14. World Anti-Doping Agency. The 2026 Prohibited List, in force 1 January 2026, section S0.
15. Cayman Chemical. Emideltide (acetate) product insert 34426. caymanchem.com — the acetate solubility figure and the one-day limit on aqueous storage.

Sources we could not open, listed rather than hidden. The PCAC vote tallies come from four independent legal and industry analyses rather than an FDA transcript, which had not been posted when we checked. We could not retrieve the abstracts of the two negative endocrine studies, of Monti 1987, or of "Degradation and aggregation of delta sleep-inducing peptide (DSIP) and two analogs in plasma and serum" (PMID 3628078) — which is why no plasma stability number appears on this page. FDA describes Schneider-Helmert 1981 as randomised, double-blind and placebo-controlled; the paper's own summary as we retrieved it does not say so, and we could not reach the methods section to settle it, so we have not relied on that study either way. The Kovalzon and Strekalova quotations were read from a third-party PDF rather than the publisher's copy. Our ClinicalTrials.gov finding rests on a domain-restricted search plus two independent confirmations, not a native registry query.