Research · 9 min read · Published September 8, 2026

Peptides for Sleep: DSIP, Epitalon and Pinealon Against the Evidence

What the 1980s and 1992 DSIP insomnia trials actually found, why the FDA advisory committee voted against DSIP (emideltide) and for Epitalon in July 2026, what Khavinson's single-lab melatonin data can and cannot support, how ghrelin and GHRH change slow-wave sleep, and how all of it compares with approved insomnia drugs.

Do peptides work for sleep?

Not on current evidence. DSIP has intravenous trials from 1981 to 1992 that disagree, and the FDA advisory committee voted 6 to 7 against it on July 24, 2026. Epitalon has single-laboratory melatonin data and no insomnia efficacy trial, though the committee voted 7 to 4 for it. Pinealon has no human sleep data. All three are unapproved.

Regulatory status first. DSIP (delta sleep-inducing peptide, now formally named emideltide), Epitalon and Pinealon are not approved drugs in the United States, the United Kingdom, the European Union or Australia. In the United States they cannot be legally compounded either: the July 2026 committee votes were advisory, FDA scientists opposed every nominee, and nothing changes until the agency finishes rulemaking, as our FDA status tracker explains. Products sold 'for research' are unapproved drugs. What follows reads the sleep evidence for these three peptides against sleep physiology and against the drugs that did get approved.

What did the DSIP human trials find?

Between 1981 and 1992, about 209 people received intravenous DSIP in short studies. Schneider-Helmert reported better sleep in 6 chronic insomniacs (1981) and in healthy volunteers. Monti's 1987 double-blind study found no significant difference from placebo, and Bes's 1992 study of 16 insomniacs found effects that were 'weak' and partly due to 'an incidental change in the placebo group'.

DSIP was isolated from rabbit blood in Basel in 1977 as a candidate sleep factor. Schneider-Helmert and Schoenenberger gave 6 middle-aged chronic insomniacs a single intravenous dose of 25 nmol/kg and reported longer sleep with fewer interruptions (Experientia, 1981); in 6 healthy volunteers dosed in the morning, sleep in the following 130 minutes rose 59 percent against placebo (Int J Clin Pharmacol Ther Toxicol, 1981).

The replications did not hold up. Monti and colleagues gave chronic insomniacs the same 25 nmol/kg intravenously in a double-blind placebo-controlled design and found that awakenings, sleep latency and time awake fell on DSIP but 'no significant differences were found in comparison to baseline or to double-blind placebo nights'; the one significant difference, in stage 2 sleep, was already present at baseline (Int J Clin Pharmacol Res, 1987). Bes and colleagues ran 16 chronic insomniacs through a matched-pairs parallel-group design over five nights, saw higher sleep efficiency and shorter latency on DSIP, and then wrote in their own abstract that the effects were weak and partly attributable to a change in the placebo group (Neuropsychobiology, 1992). In 2006 a review in the Journal of Neurochemistry titled DSIP 'a still unresolved riddle', noting that no gene, precursor protein or receptor has ever been found and that the sleep hypothesis 'is extremely poorly documented and still weak' (Kovalzon and Strekalova).

Why did the FDA committee vote against DSIP and for Epitalon?

On July 24, 2026 the Pharmacy Compounding Advisory Committee voted 6 to 7 against emideltide (DSIP), nominated for insomnia, opioid withdrawal and narcolepsy, and 7 to 4 for Epitalon, nominated for insomnia. FDA staff briefing documents concluded both were poorly characterized chemically and lacked adequate effectiveness data; the Epitalon vote went against that advice.

The FDA's DSIP briefing document, dated May 11, 2026, is the most thorough independent reading of this literature. Its conclusion on effectiveness: there was no study of the nominated subcutaneous route at all, the intravenous results were 'inconclusive and at best preliminary', and the positive studies mostly used within-subject comparisons rather than a separate placebo arm. On safety, the FDA adverse event system held zero reports through March 2024, the intravenous trials recorded no significant adverse events, and a 1984 study of 107 people in alcohol or opiate withdrawal noted transient sweating, headache, nausea and vertigo in nine and hypotension at first injection in two. Reporting from STAT quoted a panelist who voted yes on other peptides saying the committee was not being asked to judge dosing, efficacy and safety in the usual sense, which tells you how to weigh the 7 to 4 Epitalon vote.

The Epitalon briefing document is shorter on clinical evidence because there is less of it. The nomination cited one publication mentioning human use, and the FDA's own literature search 'did not retrieve publications discussing efficacy' in insomnia. The committee voted for it anyway; under the statute that still means nothing until rulemaking.

What does Khavinson's melatonin data show, and what is its limitation?

Vladimir Khavinson's St Petersburg group has reported since the 1990s that epithalamin (a pineal extract) and Epitalon (its synthetic tetrapeptide) restore night-time melatonin release and circadian rhythm in elderly people and old monkeys with reduced pineal function. The limitation is that essentially all of it comes from one laboratory, in Russian-language journals, without independent replication or a controlled insomnia outcome.

The representative paper is a 2007 Advances in Gerontology report that in aging monkeys and people night-time and average melatonin fall, and that epithalamin and Epitalon 'recover night release of endogenous melatonin' in those with pineal insufficiency (Khavinson group, 2007). Restoring melatonin is an interesting endpoint, but it is a biomarker, not sleep. Low melatonin has not been consistently associated with insomnia in the elderly, a point the FDA's Epitalon review makes, so even if the finding replicated it would not establish that Epitalon treats insomnia. Pinealon, a Glu-Asp-Arg tripeptide from the same school, has animal neuroprotection studies and small Russian clinical reports but no human sleep trial; its evidence tier for sleep is none.

What growth hormone secretagogues do to slow-wave sleep

Unlike DSIP, the GH axis has replicated controlled human sleep data. Four hourly GHRH boluses raised slow-wave sleep from 14 to 20 percent of the night in 7 men (Steiger, 1992). Ghrelin boluses increased slow-wave sleep in 7 men (Weikel, 2003). Oral MK-677 for 7 days raised stage IV sleep about 50 percent in 8 young adults (Copinschi, 1997).

This is the literature that 'sleep peptide' marketing borrows from without saying so. Slow-wave sleep and the nocturnal growth hormone surge coincide, and Axel Steiger's Munich group showed that intravenous GHRH given at 22:00, 23:00, 24:00 and 01:00 to 7 healthy men increased the share of the night spent in slow-wave sleep from 14.0 to 20.2 percent while raising GH and blunting cortisol (Neuroendocrinology, 1992). The same group found that ghrelin, the endogenous ligand of the receptor that ipamorelin, the GHRPs and MK-677 act on, given as four 50 microgram boluses, increased slow-wave sleep across the night and delta activity in the second half, while reducing REM in the middle third and raising cortisol early (Am J Physiol, 2003). Copinschi's double-blind crossover gave 8 young adults MK-677 at 5 and 25 mg for 7 days and 6 older adults 14-day courses, and found roughly 50 percent more stage IV sleep and over 20 percent more REM in the young group at the high dose (Neuroendocrinology, 1997).

How approved sleep drugs compare

For context, the FDA approved daridorexant (Quviviq) on January 7, 2022 for insomnia with difficulty falling or staying asleep, on a development program of 1,854 patients across more than 160 sites, joining suvorexant and lemborexant in the orexin-antagonist class; the FDA's 2022 novel approvals list records it. Those drugs reached the market with polysomnography-measured endpoints, replicated across trials, in the population they were meant for. The American Academy of Sleep Medicine's guideline, which the FDA cited in both peptide reviews, does not mention DSIP or Epitalon, and it places cognitive behavioral therapy for insomnia ahead of any drug. Against that standard, the three peptides here have four small intravenous trials from before 1993 that disagree, one lab's melatonin biomarker, and no controlled insomnia trial.

The bottom line

DSIP is a 49-year-old hypothesis without a receptor, a precursor or a consistent human effect, and in July 2026 the FDA committee agreed by one vote. Epitalon's melatonin story is single-source. Pinealon has nothing for sleep. The peptides that do change sleep architecture in controlled studies are growth hormone secretagogues, which carry metabolic costs and were never trialed in insomnia. As of September 2026 the evidence tier for 'peptides for sleep' is human pilot at best, and none of it is approved.