DSIP: Named for Sleep, Then Matched by Saline in Its Own Discovery Species

01What it is

DSIP, short for delta sleep inducing peptide, is a chain of nine amino acids linked end to end, sequence Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu, molecular weight 848.8. The American regulator's name for it is emideltide.

Guido Schoenenberger and Marcel Monnier, in Basel, isolated it in 1977 from the blood leaving the brains of rabbits put to sleep by electrical stimulation, and named it for what it did to the rabbit EEG, the electrical trace of the brain's activity, when infused into the fluid filled chambers inside the brain, a route called intraventricular. (PMID 265572, PMID 16539679) Nine synthetic peptides went into the ventricles of 58 rabbits under double blind conditions, and only the intact nine amino acid version enhanced the slow delta waves and spindle bursts of deep sleep.

There is no established mechanism. Kovalzon and Strekalova's 2006 review says the link between DSIP and sleep has never been further characterised, in part because no DSIP gene, protein or related receptor has been isolated, and calls the sleep factor hypothesis extremely poorly documented and still weak. (PMID 16539679)

02What the trials found

Every human administration whose route I could establish went into a vein, apart from one study given up the nose, which PubMed carries no abstract for. (PMID 11763019)

Nine papers come from the discovering laboratory and its circle, Schoenenberger or Monnier named on six, and two of the nine are a letter and a summary paper rather than studies with their own data. The 1981 opener: 6 healthy volunteers, 25 nmol/kg by slow infusion into a vein in the morning, double blind and crossed over, total sleep time up 59 percent as a median over the next 130 minutes against placebo. (PMID 6895513) Schneider-Helmert 1987 gave 14 middle aged chronic insomniacs seven successive injections into a vein of 30 nmol/kg and reported night sleep efficiency and daytime rest at the levels of normal controls, with daytime alertness and performance up significantly. (PMID 3622582) FDA's 2026 reviewers read that full paper and report that, although the authors called it placebo controlled, there was no placebo arm and the comparison was an externally matched control group.

Two double blind laboratory trials came from groups with no author in common with Basel, both at that same 25 nmol/kg into a vein. Monti and colleagues treated 6 patients with severe chronic insomnia across four nights, crossed over. Waking time and time to fall asleep fell under DSIP, none of it separated from baseline or from the placebo nights, and where DSIP did beat placebo the same gap was there before treatment. Monti's group read that as a sleep improvement too small to matter clinically. (PMID 3583493) Bes and colleagues in Amsterdam ran 16 chronic insomniacs for five nights in parallel groups, 8 on DSIP, 8 on a glucose placebo. Sleep efficiency was higher and time to fall asleep shorter, and the authors then called those effects weak, possibly an accident in the placebo group, noted that subjective sleep quality did not move, and concluded that short term treatment is not likely to be of major therapeutic benefit. (PMID 1299794)

The one randomised controlled trial I found from this century was not about insomnia. Pomfrett and colleagues in Manchester randomised 24 women having surgery, 12 to saline and 12 to DSIP into a vein at 25, 50 or 100 nmol/kg. It raised heart rate, lowered heart rate variability, and at the lowest of those three doses, 25 nmol/kg, raised the bispectral index, a monitor reading that rises as anaesthesia gets lighter. (PMID 19142086)

03What the animals actually got

The steadiest sleep effects in the animal record belong to molecules that are not DSIP.

Kovalzon's group in Moscow implanted electrodes in rabbits, injected DSIP and thirteen synthetic analogues, meaning versions of the molecule with one piece swapped out, into the lateral ventricle of the brain, and monitored sleep for 7 to 12 daytime hours. DSIP and most of the analogues had no statistically significant effect on sleep against saline given to the same animals. Two analogues did work, raising slow wave sleep by 10 to 15 percent, and a third significantly suppressed it. (PMID 11525128) The same group's 1986 paper, on an analogue built to resist the enzymes that chop peptides apart, is titled in translation "Active analog of the inactive 'sleep peptide'". (PMID 3730559)

The mouse life span result that circulates as a DSIP finding is the same substitution: it used Deltaran, which FDA describes as a freeze dried mixture of DSIP and glycine at one part to ten by weight, given under the skin to 54 female SHR mice per group, a Swiss derived strain, at about 100 mcg/kg for five days a month for life. Mean life span did not move. What did move was the tail: the last 10 percent of survivors lived 17.1 percent longer, maximum life span rose 24.1 percent, and total spontaneous tumour incidence fell 2.6 fold, all of it in mice given that two component preparation rather than DSIP alone. (PMID 12782416)

Two studies I found gave DSIP under the skin and scored sleep, and they disagree about REM, the rapid eye movement stage. Susić 1987 gave 8 cats 120 nmol/kg, about 102 mcg/kg, and recorded 8 hours of brainwave tracing: significantly more delta sleep, no change in the amount of REM sleep, and reductions in waking time and sleep onset that missed significance against a Ringer control, which is a plain salt solution. (PMID 3671519) Scherschlicht and colleagues gave cats 100 mcg/kg under the skin and reported REM preferentially augmented, and rabbits 1 mg/kg the same way, with more spindle dominated light non REM sleep. (PMID 6548967)

04How it compares

Three peptides went before the same advisory committee on 24 July 2026.

A committee vote is not an approval, and none of the three has an approved product I could find.

05Where the dose came from

Most of the human sleep trials used 25 nmol/kg into a vein, and Schneider-Helmert's two used 30 nmol/kg, higher still. At a molecular weight of 848.8 the 25 nmol/kg figure is 21.2 mcg/kg, about 1,485 mcg for a 70 kg adult. What circulates is 100 to 300 mcg under the skin once in the evening, usually 1 to 3 hours before intended sleep, and that is community convention reported here, not a recommendation. For the same adult it is 1.4 to 4.3 mcg/kg, roughly 5 to 15 times smaller per kilogram than the trial dose, and it goes in by a route neither FDA's reviewers nor I found a human trial for. Against the two cat experiments, both near 100 mcg/kg under the skin, it is 24 to 71 times smaller.

The 25 nmol/kg figure was measured going into a vein. The 100 to 300 mcg figure was measured going onto a web page. The dosing page I opened describes those numbers as coming from research write ups rather than clinical trials. I could find no published derivation for them, and no study in any species comparing the 5 to 7 night, 4 week and 6 to 8 week cycles it attaches.

FDA states a short plasma half life of 8 minutes, citing a 2001 review rather than a primary measurement, and I could not trace it back to a primary human study. Kato and colleagues, measuring disappearance after an injection into a vein, got 4.0 minutes in 4 anaesthetised dogs, 2.9 in a monkey and 2.0 in 3 rats. (PMID 6379493) Fast clearance from a vein does not by itself sink a nightly injection under the skin, since absorption out of the tissue can be the slow step, and neither FDA nor I found a study measuring that in any species. Untested, not refuted.

06What could go wrong

The safety record is thin, old, almost entirely by vein, and its largest item is not in the abstracts.

Dick and colleagues gave DSIP into a vein to 107 inpatients in alcohol or opiate withdrawal, 47 and 60, judged by physicians and nursing staff, with no control group or blinding. The published abstract says tolerance was good aside from headaches reported by a few patients. (PMID 6548969) FDA's 2026 reviewers read the full paper and report that 95 of the 107 had no significant adverse events, 9 had minor transient effects including perspiration, headache, nausea and vertigo, and 3 had serious adverse effects, two of them low blood pressure at the start of the first injection. One of those patients had a second injection and experienced worsening low blood pressure.

The other signal is Pomfrett above: heart rate up, variability down, anaesthesia lighter, in the 12 women who received it. For the route people actually use, FDA's conclusion is flat: no nonclinical or clinical safety data and no effectiveness data for chronic insomnia, narcolepsy or opioid withdrawal by the proposed route under the skin.

07What nobody knows

Whether it is a gene product at all. A 2011 search of mammalian gene and protein databases reported that nothing is known of any precursor protein or encoding gene. (PMID 21262293)

What the antibodies have been measuring. Every human study of DSIP in blood that I read measures DSIP-like immunoreactivity, whatever the antibody stuck to, and it tracks body temperature and sits lower in REM sleep than in waking (PMID 8175965), falling at the moment of falling asleep (PMID 8745061). The 2006 review proposes that a different, unidentified peptide is at least partly responsible for both.

08My take

The founding result was retested and it did not come back. Someone went back to the same species and route, ran the parent alongside thirteen chemical relatives, and two of the relatives worked while the thing with the name on it did not. A third made sleep worse. That sits in a Russian journal from 2001 that almost nobody selling this has read.

The smaller point is more useful. FDA's reviewers pulled the full text of a 1984 paper and found three serious adverse events, two of them low blood pressure, where the abstract says headaches. Everyone quoting DSIP safety, me included until I read that document, has been quoting the abstract, because the abstract is the part that is free.

If you have bought DSIP, I want to know what the vendor said the dose was based on, and whether the figure was in micrograms per kilogram or just micrograms.

Frequently asked

Is DSIP approved anywhere?+

There is no approved DSIP product in the United States, and I found none in any other market I checked. On 24 July 2026 an FDA advisory committee voted 6 in favour and 7 against adding it, under the name emideltide, to the list of bulk substances compounding pharmacies may lawfully work with. That is not an approval, and emideltide was the only one of seven peptides that week the committee declined to recommend.

Does DSIP actually help you sleep?+

The human record splits by who ran the study. Nine papers come from the discovering laboratory and its circle and report sleep improving, and every one whose route I could establish gave DSIP into a vein, though two of the nine are a letter and a summary of five earlier studies rather than studies with their own data. The two double blind sleep laboratory trials run by groups with no author in common with that laboratory, using the same 25 nmol/kg into a vein, concluded separately that the improvement was of little clinical significance and that short term treatment of chronic insomnia was not likely to be of major therapeutic benefit. There are also no records for DSIP on the United States government trials registry: I queried it for the full phrase and for the abbreviation and both returned zero, so any identifier beginning NCT offered for a DSIP trial can be checked in seconds.

How much DSIP do people use, and where does that number come from?+

What circulates is 100 to 300 micrograms injected under the skin once in the evening, and that is a community convention reported here rather than a recommendation to use it. I could find no published derivation for it. Most of the human sleep trials used 25 nanomoles per kilogram into a vein, which is about 21 micrograms per kilogram, or roughly 1,485 micrograms for a 70 kilogram adult, and Schneider-Helmert's two studies used 30 nanomoles per kilogram, higher still. So the circulating figure is about 5 to 15 times smaller per kilogram and goes in by a route neither FDA's reviewers nor I found a human trial for.

Has anyone studied DSIP injected under the skin?+

Not in a human being, in my searches or in FDA's. FDA's 2026 reviewers state there was no study evaluating effectiveness by the nominated route under the skin, and nothing on how much of it gets into the blood, or how fast it leaves, by that route either. In animals I found two studies that gave it under the skin and scored sleep, and they disagree: 8 cats given 120 nanomoles per kilogram showed significantly more delta sleep and no change in REM sleep, the rapid eye movement stage, while a second paper gave cats 100 micrograms per kilogram and reported REM sleep preferentially increased.

What is known about DSIP side effects?+

Very little, and what is known comes from doses into a vein and is decades old. The largest series gave DSIP into a vein to 107 inpatients in alcohol or opiate withdrawal, and its published abstract records only headaches in a few patients. FDA's 2026 reviewers, reading the full paper, report 9 patients with minor transient effects including perspiration, headache, nausea and vertigo, and 3 with serious adverse effects, two of them low blood pressure at the start of the first injection, with one of those going on to worsening low blood pressure after a second injection. Separately, a 2009 randomised trial in 24 women having surgery found that the 12 who received DSIP had raised heart rate, lowered heart rate variability and lighter rather than deeper anaesthesia. For injection under the skin, FDA reports no nonclinical or clinical safety data at all.

Was DSIP not proven in rabbits?+

It was named on a rabbit experiment in 1977, in which nine synthetic peptides were infused into the brain ventricles of 58 rabbits under double blind conditions and only the intact nine amino acid version enhanced delta and spindle patterns on the brainwave trace. A later Moscow series put DSIP and 13 synthetic analogues back into the brain ventricles of rabbits with implanted electrodes and reported that DSIP and most of the analogues had no statistically significant effect on sleep compared with saline given to the same animals, while two analogues raised slow wave sleep by 10 to 15 percent and one suppressed sleep.

---

This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.