DSIP

DSIP has been given to people, and the results are on record. Two sleep-laboratory studies from the 1980s infused it into chronic insomniacs. The more favourable one, in 18 patients, had no placebo group described in its summary. The controlled one, a double-blind crossover against placebo with full sleep recordings, found that slow wave sleep did not change and concluded that any improvement was of little clinical significance.

The name says delta sleep-inducing peptide. That name was a proposal made at the 1977 isolation, and it has never been confirmed. No gene, no precursor protein and no receptor for DSIP have been found. When researchers tracked the body's own DSIP-like material across a full day in 12 healthy volunteers, it peaked in mid-afternoon, bottomed out around 1 a.m., and was lower during deep sleep than during waking.

DSIP is a nonapeptide, a chain of nine amino acids with a sequence unlike any known peptide family. A group in Basel isolated it in 1977 from the cerebral venous blood of rabbits whose thalamus was being electrically stimulated to induce sleep, and they regarded it as a candidate sleep-promoting factor.

Even its identity is less settled than the name suggests. A 2006 review in the Journal of Neurochemistry notes that the sequence was determined only once, never rechecked with modern methods, and that there is some doubt the synthetic DSIP used in later research is identical to the natural material from the rabbit samples. FDA lists the compound as Emideltide (DSIP).

A Molecule Without Its Machinery

A signalling peptide normally comes with a gene that encodes it, a precursor it is cut from, and a receptor it acts on. For DSIP, none of the three has been identified. The 2006 review lists the gaps plainly: no precursor protein, no precursor gene, no specific receptor, no receptor gene, and no confirmed DSIP sequence in the material that reacts with DSIP antibodies.

That last point matters for every measurement in humans. Antibody tests detect DSIP-like immunoreactivity, and the review's authors suspect a different, still unidentified peptide may be responsible for much of it.

What Has Been Proposed

Animal and laboratory studies have attributed a long list of actions to DSIP: lowering corticotropin, raising luteinizing hormone and growth hormone release, effects on stress, body temperature, pain, seizures and opioid or alcohol dependence. The same review calls these data separate and contradictory, and reports that its direct effects on noradrenergic, serotonergic and GABA-ergic transmission are not consistent.

Stability is part of the problem. Given from outside, DSIP is broken down quickly by an aminopeptidase-like enzyme, and its half-life in the body is no more than several minutes. Some modified analogues built to resist that enzyme did promote slow wave sleep in rabbits and rats. DSIP itself did not, in the same research group's tests.

The overreach on DSIP starts with the name and is carried by it.

Reading the name as a finding. Delta sleep-inducing was the hypothesis at the 1977 isolation. The controlled human trial found no change in slow wave sleep, the delta sleep the name refers to, and the 2006 review calls the sleep-factor hypothesis extremely poorly documented and still weak.

Quoting the favourable trial alone. The 18-patient study reporting normalised sleep is the one usually cited. Its summary describes no placebo arm. The study that did use placebo, with blinding and sleep recordings, found the gains were mostly in light stage 2 sleep and partly present before treatment.

Treating the body's own DSIP as evidence of a sleep role. The 1994 plasma study measured a marker, not a treatment, and the marker ran the wrong way: highest in the afternoon, lowest at night, lower in REM and slow wave sleep than in waking. Even that marker may not be DSIP, since the antibody-reactive material has never been sequenced as DSIP.

Borrowing animal claims for people. Stress, pain, seizure and withdrawal effects come from animal and laboratory work that the 2006 review itself calls separate and contradictory. None of them was an outcome in the human sleep trials.

Human exposure in the sleep studies was short, a week of infusions in one and four nights in the other, given under laboratory supervision. Their published summaries do not report adverse events either way, so there is no tolerability rate to quote and no data on repeated use beyond those windows.

FDA lists the compound as Emideltide (DSIP). The agency states that compounded emideltide may pose a risk of immunogenicity for certain routes and may involve complexities with peptide-related impurities, and that it has not identified safety-related information for the proposed route, so it lacks sufficient information to know whether the drug would cause harm in humans. It is not FDA approved and sits on neither 503A list: the nominators withdrew the nomination on or about 22 April 2026, which did not resolve those concerns. FDA's Pharmacy Compounding Advisory Committee reviewed Emideltide (DSIP) at its 23 to 24 July 2026 meeting, for opioid withdrawal, chronic insomnia and narcolepsy FDA has published no minutes, transcript or written vote tally, so no outcome is stated here. A committee recommendation would not change any legal status in any case; rulemaking on the full record decides that.

WADA does not name DSIP on the 2026 Prohibited List. Section S0 prohibits at all times any pharmacological substance without current approval from a governmental health authority for human therapeutic use, which covers it.

DSIP sits in an unusual spot. The human sleep question was tested directly, with the right instrument, and the controlled answer was small and confined to light sleep. The uncontrolled study that looked better cannot rule out that patients would have improved anyway.

What the evidence supports: DSIP was given to insomniacs in short supervised courses, and the placebo-controlled result was described by its authors as of little clinical significance. What it does not support: any claim that DSIP deepens sleep, that the body uses it to regulate sleep, or that it treats stress, pain or withdrawal in people. Nearly fifty years after the isolation, the molecule still lacks a gene, a receptor and a confirmed natural form.

Not shown to increase deep sleep. In the double-blind crossover in insomniacs, slow wave sleep and REM sleep were unchanged; the gain in total sleep came from stage 2.

Not an established hormone. No gene, precursor protein or receptor has been identified, and the natural material that reacts with DSIP antibodies has not been confirmed to be DSIP.

Not higher during deep sleep. In 12 healthy volunteers, circulating DSIP-like material was lowest around 1 a.m. and lower in slow wave and REM sleep than in waking.

Not approved or cleared for compounding. FDA lists it as Emideltide (DSIP), it is on neither 503A list, and no outcome of its July 2026 advisory committee review has been published by FDA.

Not permitted in tested sport. It is unnamed on the WADA list and caught by S0 as an unapproved substance.

DSIP is the compound I use to show what a name can do. Most of what I cover has too little human data to judge. DSIP had the right experiment run on it decades ago, a blinded, placebo-controlled sleep study with recordings, and the deep sleep it is named for did not change.

I do not think that makes DSIP worthless as biology. The question of what the antibody-reactive material in human blood actually is remains open, and the stabilised analogues that did affect animal sleep are a real lead for someone. But a lead in a rabbit is not a sleep aid in a person, and I would rather say clearly what was tested and what came back than let three words from 1977 keep doing the work.