Peptide Health Lab

Cognitive & neuro

DSIP

FOR RESEARCH PURPOSES ONLY

Also known as: Delta sleep-inducing peptide, Emideltide, Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu

Regulatory status
Research only
Evidence grade
Human RCT evidence

Last reviewed September 1, 2026 · 14 sources

What DSIP is and how it works

DSIP is a nonapeptide with the sequence Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu. It was isolated from the cerebral venous blood of rabbits by the Schoenenberger–Monnier group in Basel and characterized in 1977. In the originating work the peptide and eight related structures — five possible metabolic fragments, two substituted analogues and a related tripeptide — were synthesized and infused into the cerebral ventricles of a total of 58 rabbits under double-blind conditions, with neocortical and archicortical EEG analyzed by Fourier transform. Only the synthetic nonapeptide produced significant and specific enhancement of delta and spindle EEG patterns.[1] A companion paper the following year reported the amino-acid analysis, sequence, synthesis and activity of the nonapeptide in full.[2]

That is the entire foundation, and it is worth being precise about what it established: an infused synthetic peptide changed rabbit EEG. It did not establish a receptor, and nearly fifty years later no receptor has been identified. An independent 2006 review states plainly that no DSIP gene, no DSIP protein and no related receptor has ever been isolated, and that the peptide's natural occurrence and biological activity remain obscure.[9] There is no proposed mechanism on this page because the literature has not produced one that survived scrutiny.

A boundary that decides how most of this literature should be read. A substantial share of the papers carrying "DSIP" in their titles do not study the synthetic nonapeptide at all. They measure DSIP-like immunoreactivity — a signal detected by radioimmunoassay and immunohistochemistry in brain tissue, peripheral organs and the plasma of several mammals.[10] That signal comes from a substance that was never isolated or identified. The 2006 review goes further and hypothesizes that the immunoreactivity is produced by some other, still-unknown DSIP-like peptide, noting that the material is concentrated in neurosecretory hypothalamic nuclei that are not particularly relevant to sleep regulation.[9] Papers reporting DSIP-like immunoreactivity in a tissue are not papers about the molecule sold under this name, and nothing on this page treats them as such.

What the research actually shows

Unusually for a compound in this category, real double-blind human trials exist. They are small, they are forty years old, and their results do not point the same way.

The favourable results. The first published human application gave a slow intravenous infusion of 25 nanomoles per kilogram to six normal volunteers in the morning, under a double-blind crossover design. Subjects reported immediate sleep pressure; median total sleep time rose 59% over the following 130 minutes against placebo, and the following night showed shorter sleep onset, reduced stage 1 and better sleep efficiency — with EEG and behavioural analysis showing no sedation of the classical pharmacological kind.[4] The same dose in six middle-aged chronic insomniacs produced longer sleep duration, higher sleep quality with fewer interruptions and slightly more REM sleep, with the sleep-promoting effect appearing only in the second hour after administration and a slight arousing effect in the first.[5] A 1981 report in the Lancet is titled as an improvement in sleep in insomniacs and is the one record on this page that the National Library of Medicine indexes as a randomized controlled trial — but it was published as a two-page letter with no abstract, so the indexing is all there is: no cohort size, no dose, no endpoint and no result can be read from the record itself.[3] The longest course anywhere in the published human set was given in 1987: 14 middle-aged chronic insomniacs given the peptide under placebo-controlled, double-blind conditions for seven successive nights, with polysomnography at baseline, at the beginning and end of treatment and on a placebo post-treatment night. Night sleep improved substantially with the first dose and further with repetition, the improvement carried into the placebo post-treatment night, sleep efficiency and daytime rest reached the levels of normal controls, and daytime alertness and performance increased significantly. The authors concluded that the study demonstrates efficacy for the treatment of impaired sleep and daytime functions as well.[6]

The unsupportive results. Also in 1987, a double-blind crossover polysomnography study of the same 25 nanomoles per kilogram dose across four nights in insomniac patients — a report whose abstract gives no cohort size — found nocturnal awakenings, NREM sleep latency, total waking time and waking after sleep onset all reduced under the peptide, but none of those reductions significant against baseline or against double-blind placebo nights. Total sleep time and NREM sleep time were both increased by the peptide, the increase related to increases in stage 2, with stage 1, slow-wave and REM sleep unmodified; the significant peptide-versus-placebo differences were in NREM sleep time and stage 2 only, and the authors noted that the same differences existed already for the baseline values. They concluded that sleep improvement under treatment was of little clinical significance.[7] In 1992, 16 chronic insomniac patients were studied over five laboratory nights in a double-blind, matched-pairs, parallel-groups design, half receiving 25 nanomoles per kilogram intravenously before each of three nights and half a glucose solution. Objective sleep quality showed higher sleep efficiency and shorter sleep latency under the peptide, and one measure of subjective tiredness fell. The authors then reported that the significant effects were weak and in part could be due to an incidental change in the placebo group, that no other measure including subjective sleep quality changed, and concluded that short-term treatment of chronic insomnia with this peptide is not likely to be of major therapeutic benefit.[8]

Why the badge on this page reads "Human RCT evidence," and what it is standing on. The evidence grade on this site describes the quality of the evidence testing a claim, not the direction the evidence points. Controlled, double-blind, placebo-comparator human trials of this compound against a polysomnographic sleep endpoint were conducted and published, which is the top of a three-value scale whose other options — "animal studies only" and "anecdotal reports only" — would both misdescribe that record. The grade is worth stating precisely rather than leaving as a badge. Of the five published human trials, exactly two carry randomized-controlled-trial indexing: the 1992 double-blind parallel-groups study in 16 patients, which is the one with a full abstract and which concluded against major therapeutic benefit, and the 1981 Lancet letter, whose design is not readable from the record at all.[8][3] The remaining three are indexed as controlled clinical trials and describe themselves as double-blind and placebo-controlled without naming a randomization method. So the badge rests on a genuinely randomized, genuinely placebo-controlled human trial of the sleep claim — and that trial is one of the ones that came out against the compound. A high grade on this page means the question was asked properly. It does not mean the answer was yes.

Outside sleep, the record is thin. A 1998 open clinical trial in opioid detoxification was published as a brief report and was, by its own description, neither randomized nor blinded.[13] A contemporary review records that the peptide had been evaluated for insomnia, pain and withdrawal, and that its physiological functions and a possible mechanism involving adrenergic modulation remained to be established.[11] The most recent preclinical work is a 2021 rat study: intranasal administration at 120 micrograms per kilogram before and for seven days after middle cerebral artery occlusion significantly improved rotarod motor recovery, while the smaller infarct volume in treated animals did not reach statistical significance.[14]

Where the evidence is weak

The founding hypothesis failed, and the field said so. The 2006 independent review is the most important citation on this page. It states that the link between this peptide and sleep was never further characterized, that no gene, protein or receptor was isolated, and that the hypothesis of DSIP as a sleep factor is extremely poorly documented and still weak. It also reports that in the reviewers' own earlier work, certain artificial structural analogues of the peptide — but not the peptide itself — showed significant slow-wave-sleep promoting activity in rabbits and rats.[9] A 1970s pedigree in a top journal is often read as validation. Here the correct reading is the opposite: a compound characterized in 1977 that nobody could subsequently tie to a receptor or a gene is a compound whose central premise did not hold up.

The human trials are tiny and old. Six, six, fourteen, sixteen — and two of the six published human records state no cohort size at all, the 1987 crossover study and the 1981 Lancet letter, so their contribution to that list is unknown.[7] Every human result on this page comes from a sleep laboratory between 1981 and 1992, in populations of single or low double digits, in chronic insomniacs or a handful of volunteers. Nothing published in the last thirty years re-tested the sleep question in humans at any size.

Pharmacology makes the trials hard to interpret even on their own terms. Incubating the peptide in human or rat blood released degradation products, with formation dependent on temperature, time and species; the authors concluded that the rapid disappearance of injected peptide was degradation, while labelled analogs degraded more slowly and instead formed complexes that unlabelled excess did not displace, consistent with non-specific binding or aggregation.[12] A compound that is destroyed and aggregated in circulation, with no known receptor and no measured brain exposure, is one where a null result and a delivery failure are not distinguishable from each other.

Route, population and duration all differ from the way it is used now. Every human study whose abstract states a route states a slow intravenous infusion or injection given in a laboratory.[5][7][8] The populations were middle-aged chronic insomniacs, six healthy volunteers under psychophysiological and EEG observation, and — in a 1998 brief report indexed to heroin dependence — patients undergoing opioid detoxification.[4][13] The longest course any published abstract describes is seven successive nights.[6] There is no published human pharmacokinetic study and no bioavailability figure for any route, and nothing published follows anyone using it outside a monitored setting. The gap between the published evidence and current use is a gap in route, dose, population, duration and supervision simultaneously.

Breadth of reported activity is a caution, not a feature. The older literature reports effects on electrophysiological activity, brain neurotransmitter levels, circadian and locomotor patterns, hormonal levels, psychological performance, and the activity of other neuropharmacological drugs including their withdrawal.[10] For a molecule with no identified target, a long list of reported effects is a description of an uncharacterized pharmacology rather than a description of versatility.

Legal and regulatory status

DSIP is not approved by the FDA for any indication. Nothing containing it holds an approved US application, an approved label, or an assigned indication. Material sold under the name is designated for laboratory research use, which is what the research-only framing on this page reflects. As far as the published literature shows, the compound has no marketing approval in any jurisdiction; it was characterized in Switzerland in 1977 and never completed a development programme anywhere.

The compounding position is specific and worth stating precisely. Under the nonproprietary name emideltide, the substance was nominated as a bulk drug substance for use in pharmacy compounding under section 503A of the Federal Food, Drug, and Cosmetic Act. The agency sorted that nomination into category 2, the bin reserved for substances it judged potentially to present significant safety risks, and the nominator later withdrew it, leaving the evaluation unfinished. What FDA recorded as its concerns were the risk of immunogenicity for certain routes of administration, complexities with regard to peptide-related impurities and active-ingredient characterization, and the fact that it had identified no safety-related information for the proposed route of administration and so lacked sufficient information to know whether the drug would cause harm if administered to humans. The net position is that the substance is absent from the 503A bulks list and no lawful compounding pathway exists for it in the United States.

The identity question is unusually sharp for this compound and it is a research problem before it is a regulatory one. With no approved manufacturer there is no release specification and no lot testing, and the published chemistry establishes that this peptide degrades and aggregates in biological fluid in ways that depend on temperature, time and species — so the difference between intact material and its fragments is not something a label can be relied on to settle.[12] PHL does not sell peptides and takes no position on how unapproved material reaches anyone.

Questions to bring to a provider

The productive conversation is about the sleep problem, not the peptide. Questions worth raising:

  • Has the sleep complaint been characterized — onset, maintenance, early waking, circadian phase, apnoea, restless legs, medication, alcohol, mood — and does that change which treatments have evidence behind them?
  • Five controlled human trials of this compound reached opposite conclusions, and two of them — including the only one that is both randomized-indexed and readable — concluded the benefit was of little clinical significance and not likely to be of major therapeutic benefit. How should a split literature that small be read?[7][8]
  • Every published human dose whose route is stated was given intravenously in a sleep laboratory. What is known about any other route in anyone?[4]
  • Given that no receptor, gene or endogenous protein has ever been identified for this peptide, what would either of us be monitoring for?[9]
  • Would cognitive behavioural therapy for insomnia, which has a large controlled evidence base, be a more reasonable first step than an unapproved compound with a forty-year-old one?
  • For an unapproved compound that degrades and aggregates in plasma, how would anyone confirm that a vial holds intact material at the concentration claimed?[12]

A clinician who answers "this was tested properly in the 1980s, in trials too small and too divided to settle the question, and nobody has retested it since" is describing the literature accurately, not being dismissive.

Evidence by claim

Grades describe how strong the evidence is; the line under each grade describes what kind of studies it is. How we grade evidence.

Sleep in chronic insomnia
Human RCT evidence human RCT · human observational

This claim carries a human randomized-trial grade because controlled, double-blind human trials genuinely tested it, not because they agreed. Five published human trials exist and they divide. On the supportive side: a 1981 Lancet record, which the National Library of Medicine indexes as a randomized controlled trial but which was published as a two-page letter carrying no abstract, so its design cannot be read from the record; a 1981 study in six middle-aged chronic insomniacs reporting longer sleep duration, higher sleep quality and fewer interruptions, with the sleep-promoting effect confined to the second hour after administration; and a 1987 study in 14 middle-aged chronic insomniacs, placebo-controlled and double-blind across seven successive nights, reporting substantially improved night sleep, persistence into a placebo post-treatment night, and significantly increased daytime alertness and performance, whose authors concluded that the study demonstrates efficacy for impaired sleep and daytime functions. On the unsupportive side: a 1987 double-blind crossover polysomnography study, which reports no cohort size, found reductions in awakenings and waking time that were not significant against baseline or placebo, found its significant differences in NREM sleep time and stage 2 already present at baseline, and concluded that sleep improvement was of little clinical significance; and a 1992 double-blind matched-pairs parallel-groups study in 16 patients found higher sleep efficiency and shorter sleep latency but judged the effects weak and partly attributable to an incidental change in the placebo group, concluding that short-term treatment of chronic insomnia is not likely to be of major therapeutic benefit. Of the five, only the abstract-less Lancet letter and the 1992 study carry randomized-controlled-trial indexing; the other three are indexed as controlled clinical trials and describe themselves as double-blind and placebo-controlled without stating a randomization method. Nothing in the record ranks one of these trials above another on quality, and this page does not.

[3] [5] [6] [7] [8]

Acute sleep induction in healthy adults
Anecdotal reports only human observational

One study, six people. In four men and two women given a slow intravenous infusion in the morning under a double-blind crossover design, subjects reported immediate sleep pressure and median total sleep time rose 59% over the following 130 minutes compared with placebo, with shorter sleep onset, reduced stage 1 and better sleep efficiency on the following night; EEG and behavioural analysis showed no sedation of the classical pharmacological kind. It has not been replicated in a larger healthy cohort in the forty-five years since, and a cohort of six with an unstated randomization method does not support a higher grade than this one.

[4]

DSIP as an endogenous sleep factor
Anecdotal reports only review

The founding hypothesis is the weakest claim on the page and the field says so itself. An independent 2006 review states that the link between the peptide and sleep was never further characterized, that no DSIP gene, protein or receptor was isolated, and that the hypothesis of DSIP as a sleep factor is extremely poorly documented and still weak. That review notes that the immunoreactive material distributed in neurosecretory hypothalamic nuclei sits in regions not particularly relevant to sleep regulation, and that in the authors' own earlier work certain artificial structural analogues — but not the peptide itself — promoted slow-wave sleep in rabbits and rats.

[9]

Opioid withdrawal and detoxification
Anecdotal reports only human observational · review

A 1998 open clinical trial of the peptide in opioid detoxification was published as a two-page brief report carrying no abstract, and is indexed to heroin dependence and substance withdrawal syndrome; by its title it was open, meaning neither randomized nor blinded, and it has not been followed by a controlled trial. A contemporary review records that considerable work had been carried out evaluating the peptide's potential therapeutic use in insomnia, pain and withdrawal, and that its various physiological functions and a possible mechanism of action involving modulation of adrenergic transmission remained to be established. No published randomized trial supports this use.

[13] [11]

Motor recovery after experimental stroke
Animal studies only animal

In Sprague-Dawley rats given intranasal peptide at 120 micrograms per kilogram before and for seven days after middle cerebral artery occlusion, motor performance on the rotarod recovered significantly compared with vehicle. Infarct volume was smaller in treated animals but the difference was not statistically significant, and the report should be read with both halves intact. This is a 2021 rat study and there is no human counterpart.

[14]

Any outcome from self-administration outside a laboratory
Anecdotal reports only human observational · human RCT · in vitro

The published human record for this compound is narrow in a specific way. Every human study whose abstract states a route states intravenous administration in a monitored setting; the populations were middle-aged chronic insomniacs, six healthy volunteers under psychophysiological and EEG observation, and heroin-dependent patients in an open detoxification report; and the longest course any published abstract describes is seven successive nights. No published human study reports a pharmacokinetic characterization or a bioavailability figure for any route. The plasma chemistry compounds the gap rather than closing it: incubation in human or rat blood degraded the peptide rapidly, so an amount administered does not indicate an amount delivered by any route. Everything claimed about the compound as it is currently used rests on report rather than on published evidence.

[4] [6] [8] [13] [12]

FOR RESEARCH PURPOSES ONLY

Typical protocol range in the research

  • First reported human application: six normal volunteers, four men and two women, in a double-blind crossover design against placebo, with morning administration and psychophysiological and EEG measurement

    25 nanomoles per kilogram of body weight, given as a slow intravenous infusion [4]

  • Six middle-aged chronic insomniacs given a single acute administration, with sleep duration, quality and interruptions measured

    25 nanomoles per kilogram of body weight, intravenously [5]

  • Sixteen chronic insomniac patients in a double-blind, matched-pairs, parallel-groups design across five laboratory nights, with polysomnography and subjective sleep quality as endpoints; the comparator arm received a glucose solution

    25 nanomoles per kilogram of body weight intravenously, in the afternoon before each of three consecutive nights [8]

  • Sprague-Dawley rats given intranasal administration for 8 days around an intraluminal middle cerebral artery occlusion, with rotarod motor performance and infarct volume as endpoints

    120 micrograms per kilogram of body weight, once 60 (±15) minutes before occlusion and daily for 7 days after reperfusion [14]

Every published human figure for this compound is the same intravenous dose — 25 nanomoles per kilogram — given in a sleep laboratory between 1981 and 1992, in cohorts of six to sixteen people where the report states a cohort size at all. No published trial establishes a dose by any other route in humans, and none establishes a dose for use outside a monitored laboratory setting. The milligram figures that circulate in community discussion have no published trial behind them and are not reported here.

Ranges are what published research reports, not a recommendation. PHL does not prescribe, and nothing here is personalized to you.

Side effects & safety signals

  • Rapid degradation and non-specific aggregation in blood

    Demonstrated in human and rat blood · Incubation in human or rat blood released degradation products, with formation dependent on temperature, time and species; the authors concluded that the rapid disappearance of injected peptide was degradation, while labelled analogs degraded more slowly and instead formed complexes consistent with non-specific binding or aggregation. This is not a toxicity finding, and it is the single most important pharmacological fact on this page: it is a large part of why the human trials are hard to interpret, and it means the amount that reaches any target is not knowable from the amount administered.

    [12]

  • Tolerability reported only in single-digit cohorts over hours to days

    No adverse events reported in the small published trials · In six healthy volunteers the compound was described as well tolerated with no psychological, physiological or biochemical side effects observed, and in six chronic insomniacs there was no daytime sedation or other side effect reported. Those are absences of findings in cohorts of six, measured over a single day and a single night respectively. They cannot speak to uncommon harms, and nothing published follows anyone over months.

    [4] [5]

  • No identified receptor, gene, or endogenous protein

    Standing state of the field · An independent 2006 review states that no DSIP gene, protein, or related receptor has been isolated, and that the peptide's natural occurrence and biological activity remain obscure. A compound whose target is unknown offers nothing against which an off-target effect could be anticipated, which is a different kind of uncertainty from a drug with a known receptor and an unmeasured long-term record.

    [9]

  • Broad, uncharacterized effects outside sleep

    Reported across the older literature · Beyond sleep, the peptide has been reported to affect electrophysiological activity, brain neurotransmitter levels, circadian and locomotor patterns, hormonal levels, psychological performance, and the activity of neuropharmacological drugs including their withdrawal. A later review adds that a possible mechanism involving modulation of adrenergic transmission remained to be established. Breadth of reported activity in a compound with no established mechanism is a reason for caution around any concurrent medication, and a reason to raise it with a clinician rather than a reason to expect a particular effect.

    [10] [11]

Published lists are never exhaustive, so report anything unexpected to a licensed provider.

Storage, handling & reconstitution

Lyophilized storage
Lyophilized DSIP is conventionally kept refrigerated at 2–8 °C and protected from light, with freezing used for longer holding. No published stability study addresses shelf life of the powder in a vial. What the published chemistry does establish is that this particular peptide is unstable once it meets blood, degrading rapidly in human and rat plasma, which is a fact about the molecule in circulation rather than about a sealed container.
Reconstituted storage
Once in solution, peptides are conventionally refrigerated at 2–8 °C, protected from light, and treated as short-dated rather than indefinitely stable. No published study establishes a solution shelf life for DSIP, and the peptide's documented tendency to aggregate and form complexes in biological fluid is a reason to treat any solution of it as short-dated rather than a reason to treat it as robust.
Reconstitution diluent
The conventional diluent for lyophilized research peptides, including DSIP, is bacteriostatic water: sterile water preserved with 0.9% benzyl alcohol. Sterile water without a preservative is used where a preservative is contraindicated and is treated as single-use.
Reconstitution & handling
Conventional handling: sanitize the stopper before piercing it, add the diluent slowly down the inside wall of the vial rather than directly onto the powder, and let the powder dissolve on its own. Swirl gently if needed; never shake. A properly reconstituted solution is clear, and anything cloudy, discoloured, or carrying visible particulate is discarded. PHL does not publish volume or unit calculations and does not describe how any preparation is administered, because those are dosing and administration decisions that belong with a licensed clinician.
Handling notes
DSIP has no approved manufacturer in the United States, so there is no label, no assigned beyond-use date, and no lot-level stability testing standing behind any storage claim here. FDA has separately identified peptide-related impurities and active-ingredient characterization as concerns for material carrying this name. Treat the conventions above as conventions, and note that nothing guarantees a vial holds the molecule its label names at the concentration claimed.

The how-to guides cover the conventions behind these fields, and what the published stability data does and does not establish: Storage, Reconstitution, Handling.

Questions for your provider

Bring this page to a licensed provider. A provider can order labs, review your medications and history, and tell you whether anything here is relevant to your situation. This page can't. Peptide Health Lab does not prescribe and does not sell peptides.

Want a structured starting point for that conversation? The Stack Builder turns your goals into a research-cited outline you can review together.

Citations

14 sources · every identifier checked against PubMed

Before you act on any of this

This page is educational only and is not medical advice. It does not diagnose, treat, or prescribe. Review it with a licensed provider before making any health decision. Peptide Health Lab does not sell peptides.

How we grade and verify evidence · Terms of Use · Privacy Policy

Open the Stack Builder