DSIP: why the ‘sleep peptide’ never became a sleep drug

Small Swiss studies through the 1980s reported that DSIP, a nine-amino-acid peptide, normalised sleep in chronic insomniacs, but two later independent placebo-controlled trials found effects too weak or inconsistent to matter clinically. DSIP was isolated in 1977 as a suspected natural sleep signal, and in animals it raised the slow, high-amplitude “delta” EEG activity that marks deep non-REM sleep — a specific claim, not a general sedative effect. Close to fifty years on, no gene, precursor or receptor has been identified, the human record rests on a handful of tiny studies, and we found no human trial published after 1995.

What DSIP is

Delta sleep-inducing peptide is nine amino acids long: Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu [1]. Guido Schoenenberger and Marcel Monnier in Basel isolated it from rabbits, and a later review describes the source as blood draining from the brain [1,2]. When they infused the synthetic version into the brain ventricles of rabbits, under double-blind conditions, it increased the slow “delta” waves and spindles on the EEG that mark deep non-REM sleep. Eight related peptides tested alongside it didn’t [1].

The name needs a word of explanation. “Delta” refers to delta waves, the slow, high-amplitude brain rhythm that dominates the deepest stage of non-REM sleep, sometimes called slow-wave sleep. So the original claim was specific: this peptide didn’t just make animals drowsy, it pushed the brain into deep sleep. That specificity is what set DSIP apart from a general sedative, and it is the exact claim the later human trials were built to test.

The 1977 work set a clear bar: a peptide that induces deep sleep. A 2006 review by two Russian sleep researchers, Vladimir Kovalzon and Tatyana Strekalova, mapped how much of the biology behind that bar was still missing — no gene for DSIP had been found, no precursor protein and no receptor, so it wasn’t clear DSIP was a natural signalling molecule at all [2]. Their own experiments in rabbits and rats had found that some synthetic DSIP analogues promoted slow-wave sleep but DSIP itself did not [2]. They suggested the antibodies used to “detect” DSIP in tissues might be picking up a different, related peptide.

There is also a stability problem. When Schoenenberger’s group incubated DSIP in human or rat blood, it broke down quickly, releasing free tryptophan; two modified analogues lasted longer [8]. That makes it hard to know how much intact DSIP reaches the brain after an injection.

The human evidence

Almost all the positive human data come from one researcher, Dieter Schneider-Helmert, working with Schoenenberger. All the studies used intravenous injections, usually 25 to 30 nanomoles per kg of body weight.

• 1981, six insomniacs. A single injection gave longer, less interrupted sleep, slightly more REM sleep and no daytime sedation. The effect started in the second hour; in the first hour there were hints of mild arousal [3].
• 1983, a summary of five studies. The team reported that DSIP not only improved sleep but also raised daytime alertness and, on psychological testing, improved “stress tolerance and coping ability” [12]. These broader claims were never tested by anyone else.
• 1986, 18 insomniacs. A week of injections (six doses of 30 nmol/kg) followed by a week off. Middle-aged patients’ sleep improved to normal values; older patients, whose sleep was worse at the start, took longer to normalise [4]. The abstract doesn’t describe a placebo group.
• 1987, 14 insomniacs. Seven nights of DSIP under placebo-controlled, double-blind conditions. The author reports that night-time sleep efficiency reached the level of normal sleepers and daytime alertness and performance improved [5].

Two other groups tested DSIP in insomnia, and neither was impressed. Jaime Monti’s team gave 25 nmol/kg or placebo over four nights in a double-blind crossover [6]. Total sleep time went up, driven by light stage 2 sleep rather than deep slow-wave sleep, and the apparent differences from placebo were already present at baseline. They concluded the improvement was “of little clinical significance” [6].

The University of Amsterdam trial by Bes and colleagues, summarised above, reached a similar verdict. The objective gains were weak, the authors thought they were partly explained by a chance shift in the placebo group, and the patients didn’t feel they slept better [7]. Their conclusion was that short-term DSIP “is not likely to be of major therapeutic benefit” [7].

Reading the two sets of results fairly

It would be easy to say the independent trials simply cancel out the Swiss ones, but the picture is messier. The Swiss team’s own 1983 summary describes five double-blind human studies, a delay of about an hour before sleepiness began, and effects lasting up to 20 hours; it also says slow injection was essential [12]. Those are claims the independent groups weren’t set up to test in the same way. Bes and colleagues injected in the afternoon, hours before bedtime [7], which may or may not matter for a peptide said to act after an hour and break down fast in blood [8].

What both sides share is tiny samples: six, 14, 16 and 18 people. With numbers that small, a couple of unusually good or bad nights in one group can swing the result, and that is exactly what the Amsterdam team suspected had happened [7]. The honest summary is that DSIP never showed a benefit that was both reproducible and large enough to matter, and nobody went back to settle it with a bigger trial.

It is worth noticing that the one sleep variable that should rise if DSIP did what its name promises, deep slow-wave sleep, didn’t change in Monti’s trial [6].

Hormones and withdrawal

DSIP was also studied as a brake on the stress-hormone system. In a randomised, double-blind crossover in 11 healthy men, a single 25 nmol/kg injection lowered blood ACTH for at least 3 hours, though cortisol, the hormone ACTH controls, didn’t change [9]. A German team then tested it more directly: infusions of 3 to 4 mg had no effect on ACTH or cortisol responses to a hormone challenge or to a midday meal [10]. The picture that emerges is narrow: DSIP moved one upstream hormone in one small study, and larger infusions moved neither.

In the early 1980s, DSIP was given intravenously to 107 inpatients in alcohol or opiate withdrawal. The doctors reported that symptoms cleared or improved markedly in 97% of opiate and 87% of alcohol patients who could be assessed [11]. There was no control group, the assessment was clinical impression by staff who knew what patients were given, and withdrawal symptoms improve with time anyway. It is best read as a hypothesis that was never followed up with a proper trial.

What the animal work suggests

Apart from the original rabbit experiments [1], animal research on DSIP is scattered and inconsistent, which is exactly the point of the 2006 review [2]. Sleep effects in rabbits and rats have been hard to reproduce with DSIP itself, while several analogues seemed more active [2].

Safety and side effects reported

The Basel studies describe DSIP as well tolerated, with no daytime sedation [3]. In the withdrawal study, a few patients reported headaches [11]. That is the entire safety record, all from small, short studies of intravenous injections four decades ago. There is no safety data on repeated long-term use, on under-the-skin injection (the way it is used today by people buying it online), or on the quality of products sold as research chemicals.

Legal and regulatory status

DSIP has never been approved as a medicine by the UK’s MHRA, the European Medicines Agency or the US FDA, or, as far as we can find, by any other regulator. It is sold online only as a research chemical, which means it isn’t manufactured or tested to medicine standards and isn’t meant for human use. An approved medicine carries a batch-tested guarantee of identity and purity, sterility, endotoxin limits and heavy-metal limits; a vial sold as a research chemical carries none of those as a promise. If you are looking at one, the only thing worth asking for is an independent certificate of analysis, for that specific batch, covering those four things.

What is still open

• Whether DSIP exists naturally as a signalling molecule, since no gene, precursor or receptor has been identified.
• Why Schneider-Helmert’s results were so much stronger than the independent trials.
• How much intact DSIP reaches the brain, given how quickly it breaks down in blood.
• Whether it does anything at all when injected under the skin rather than into a vein.
• Anything about long-term safety.

What this means

HPR’s position is that DSIP is a hypothesis wearing a drug’s name, and it is better treated that way than dressed up as a sleep remedy. The peptide was named for a specific effect — pushing the sleeping brain into slow-wave sleep — and the human literature never delivered that effect at a size anyone could reproduce. Strip the name away and what remains is a nine-amino-acid chain with no identified gene, precursor or receptor behind it. A name is not a mechanism, and a mechanism is what a medicine needs.

That has a practical consequence for anyone reading a product page. When a compound is sold as a “sleep peptide” and its entire human record is a handful of small intravenous studies from the 1980s and 1990s, the buyer is not choosing something with a known risk-benefit profile — they are enrolling in an unfinished research question. There being no human trial published after 1995 tells you the field stopped asking, not that the answer came back yes. Someone injecting DSIP under the skin today is running an experiment with no protocol, no monitoring and no published safety record, on themselves.

We would change our view on clear terms: a properly powered, placebo-controlled trial of DSIP in chronic insomnia, using the timing and route the original work specified, measuring slow-wave sleep on polysomnography, and asking patients whether they feel different. If deep sleep improves reproducibly and the subjective benefit follows, DSIP earns its name back. Our position stands unless such a trial contradicts it. The burden of proof sits with whoever is selling it as a sleep aid, not with the people asking them to show their evidence.

Educational content only — not medical advice. Many peptides discussed on HPR are not approved for human use. Talk to a qualified clinician before making any decision about your health.