What it is
DSIP is a short chain of nine amino acids (a peptide). The name stands for “delta sleep-inducing peptide.” Swiss researchers found it in rabbits. They put it into the rabbits’ brains, and the brain waves changed. They moved towards “delta” waves. These are the slow brain waves of deep sleep. Its official generic name is emideltide.
Nearly fifty years later, its job in the body is still unclear. No one has found the gene that makes it. No one has found the receptor (the docking spot on cells) it works through. It is not proven that the human body makes it on its own. The DSIP used in studies today is made in a lab. It is not an approved medicine in the United States.
DSIP (emideltide) is H-Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu-OH (C35H48N10O15, 848.8 g/mol, CAS 62568-57-4). Schoenenberger and Monnier (1977) reported that intraventricular infusion of the synthetic nonapeptide in rabbits enhanced delta and spindle EEG activity, while five fragments, two nonapeptide analogues and a related tripeptide did not.
No precursor gene, protein or receptor has been identified (Kovalzon and Strekalova 2006). DSIP-like immunoreactivity is found in hypothalamic neurosecretory nuclei and elsewhere, and those authors propose that an as-yet-unidentified DSIP-like peptide may account for it. Sleep-promoting activity in later rodent and rabbit work was more consistent for certain synthetic analogues than for DSIP itself. Other actions reported in animal and in-vitro studies are broad and inconsistent, including a proposed interaction with opioid systems that prompted the withdrawal studies.
Pharmacokinetics: half-life is about 2–3 minutes in rats and a monkey and 3–6 minutes in dogs after intravenous dosing (Kato et al. 1984), 5–10 minutes in human serum in vitro (Graf et al. 1987), and about 8 minutes in plasma as cited by FDA, driven by aminopeptidase cleavage from the N-terminus. A saturable transport route across the choroid plexus has been described in animals. No data exist for subcutaneous dosing.
Who made it and when
DSIP came from a lab in Basel, Switzerland. Since the 1960s, Marcel Monnier’s team had studied the blood of sleeping rabbits. The rabbits were kept asleep by electric pulses to a deep part of the brain called the thalamus. In 1977, Guido Schoenenberger and Monnier published the structure of DSIP. They also showed that a lab-made copy changed brain waves in the same way.
In the late 1970s and 1980s, researchers in Basel and Geneva tested it in people. They gave it into a vein. They tested it in people with insomnia, narcolepsy (a sleep disorder), and drug or alcohol withdrawal. No drug company ever sold it as a medicine.
Monnier’s group first described a sleep-inducing factor in rabbit cerebral venous blood collected during thalamic stimulation in the 1960s; Schoenenberger and Monnier characterised the nonapeptide and its synthetic analogue in PNAS in 1977. Kovalzon and Strekalova (2006) credit the “Schoenenberger–Monnier group from Basel” with the 1977 isolation.
Clinical work followed quickly: Schneider-Helmert, Graf and Schoenenberger reported in the Lancet (1981) that synthetic DSIP improved sleep in insomniacs, followed by further Basel studies in chronic insomnia and a single narcolepsy case (Schneider-Helmert 1984–1987). Tissot’s group in Geneva studied it in alcohol and opiate withdrawal (Dick et al. 1983 and 1984), and a German group reported an open study in opioid detoxification (Backmund et al. 1998), with study drug supplied by Strathmann AG of Hamburg. WHO assigned the INN emideltide in 1993. No marketing authorisation was found in any country, and no developer patent was verified for this article.
What the data say
In animals, the early rabbit tests reported deeper, slow-wave sleep after DSIP. But later animal studies often could not show a reliable sleep effect from DSIP itself.
In people, the studies are small and old. Adults with long-term insomnia got short courses of DSIP into a vein. At best, their sleep changed only a little. One study of 16 patients was double-blind (neither patients nor researchers knew who got what). Its authors said the effects were weak. They said it was unlikely to be of much use as a treatment.
Other open studies looked at drug and alcohol withdrawal. They reported fewer symptoms. But they had no comparison group, and everyone knew who got DSIP. No trial is listed on ClinicalTrials.gov. In 2026 the FDA reviewed the studies. It found the evidence was not enough for insomnia, narcolepsy or opioid withdrawal.
FDA (2026) counted about 209 people given intravenous DSIP (25–150 nmol/kg, 1–15 days) across published studies; none received it subcutaneously.
Chronic insomnia: Schneider-Helmert and Schoenenberger (1981) ran a randomised, double-blind, within-subject crossover in six patients (25 nmol/kg IV before bedtime), reporting fewer awakenings from the second hour onward but no change in sleep-onset latency. Schneider-Helmert (1987) gave seven nightly injections of 30 nmol/kg to 14 patients with placebo only on the first and last nights. Bes et al. (1992), in a double-blind matched-pairs parallel study of 16 patients (25 nmol/kg IV on three afternoons), found higher sleep efficiency and shorter latency, but the authors judged the effects weak, possibly driven by a change in the placebo group, and concluded that short-term treatment was unlikely to be of major therapeutic value.
Withdrawal: Dick et al. (1984) gave 25 nmol/kg IV, up to six injections daily, to 107 inpatients (60 opiate, 47 alcohol) in an uncontrolled open study and reported marked improvement in most evaluable patients, with anxiety slower to resolve and insomnia often recurring. Backmund et al. (1998) treated seven patients (35 nmol/kg IV); only two completed the schedule. Narcolepsy data consist of a single case report.
Regulatory picture
Approval: no medicine with DSIP is approved by the FDA.
Compounding (pharmacies mixing their own versions): in September 2023, the FDA put DSIP in “Category 2.” This is a group of substances pharmacies should not compound from. The FDA was worried about possible immune reactions and impurities. It also pointed to a lack of safety information. By April 2026, the FDA listed it as withdrawn from that group. This was because the people who had put it forward pulled it back. FDA staff then reviewed it for insomnia, narcolepsy and opioid withdrawal.
In July 2026, an FDA advisory committee voted against adding it to the list of allowed compounding ingredients. The vote was close. The committee agreed with FDA staff. The meeting discussed seven peptides. DSIP was the only one the committee turned down. In October 2026, the US health department (HHS) published a draft policy easing enforcement for the six peptides the committee had backed. DSIP was not included.
Enforcement: in 2020, the FDA sent a warning letter to a compounding pharmacy. The letter named DSIP as an ingredient the pharmacy should not have compounded.
Approval: none, and no US IND-stage programme could be verified.
Compounding: “Emideltide (DSIP)” entered Category 2 of the interim 503A policy on 2023-09-29, with FDA citing immunogenicity risk, peptide-related impurities and API-characterisation complexity. FDA’s safety-risk page (updated 2026-04-22) lists it among substances withdrawn by the nominators. It is not on the 503A bulks list (21 CFR 216.23), so this article records compounding status as unknown.
FDA’s 2026 briefing document evaluated emideltide free base and acetate for opioid withdrawal, chronic insomnia and narcolepsy. It judged neither form well characterised (non-standard naming, missing impurity, aggregate, endotoxin and bioburden data), noted the very short half-life and absence of subcutaneous data, and found the clinical studies too small, unblinded or inconsistently reported to establish effectiveness; FDA-approved therapies exist for all three conditions. On 2026-07-24 the Pharmacy Compounding Advisory Committee voted against inclusion; published accounts give the tally as 6 yes, 7 no and 1 abstention, and official minutes had not been posted when this was written. The HHS draft enforcement-discretion policy of 2026-10-30 (docket HHS-OASH-2026-0496) covers the six peptides the committee recommended and does not include emideltide.
