Does Emideltide or DSIP improve sleep?

Emideltide can change sleep measurements in some people.

That is not the same as showing that it treats insomnia.

The human record is a set of small intravenous studies, mostly from the 1980s. Some reported shorter sleep onset, fewer awakenings or more total sleep. Better-controlled studies did not establish a reliable or clinically meaningful benefit over placebo. No study tested the subcutaneous route later proposed for compounded products.

In July 2026, FDA staff proposed not adding both evaluated forms to the 503A Bulks List. The advisory committee then voted 6–7, with one abstention, against listing each form.

The evidence is mixed, preliminary and route-mismatched. Emideltide is not an established insomnia treatment.

DSIP and Emideltide are not a product specification

DSIP stands for delta sleep-inducing peptide. FDA used emideltide for the reported nine-amino-acid peptide behind that name.

But the label does not resolve the material in a vial.

FDA evaluated emideltide free base and emideltide acetate as two distinct bulk drug substances. Neither has a USP or National Formulary drug-substance monograph. Neither is a component of an FDA-approved drug.

The agency found inconsistent naming across nominations and commercial references. Some records did not make clear whether they referred to the free base, acetate or another derivative.

That uncertainty comes before any efficacy question. Evidence for one identified substance, route and formulation cannot silently validate another.

Six healthy volunteers showed a sleep signal

The most shareable positive number comes from a 1981 study.

Six healthy volunteers received slow intravenous synthetic DSIP and placebo in a double-blind crossover design. The abstract reported 59% more median total sleep time during a 130-minute observation interval after DSIP. On the subsequent night, sleep onset was shorter, stage 1 sleep fell and sleep efficiency improved.

The researchers did not observe classic pharmacological sedation. They also reported no psychological, physiological or biochemical side effects.

This is direct human evidence. It is also six healthy people observed over a short period.

It does not test chronic insomnia. It cannot establish uncommon harms, long-term safety or the effect of a subcutaneous compounded product.

The strongest positive insomnia report lacked a concurrent placebo arm

A 1987 report followed 14 middle-aged people with chronic insomnia through seven successive intravenous injections.

The within-person changes looked large:

• mean sleep onset fell from 58.4 minutes at baseline to 38.9 after the first injection and 27.6 after the last;
• wake after sleep onset fell from 120.8 minutes to 78.0 and then 65.6;
• total sleep time rose from 313 minutes to 363.1 and then 385.

The paper called the design placebo controlled. FDA did not.

The people with insomnia received placebo during baseline and after treatment, but the principal comparison used an external group of healthy sleepers. There was no simultaneous insomnia group randomized to remain on placebo.

Without that arm, improvement can reflect treatment, time in the laboratory, expectancy, baseline fluctuation or regression toward a person’s usual sleep. Comparing treated insomnia with healthy sleep also does not estimate the treatment effect against placebo.

The numbers are evidence of change inside the study. They are not a clean causal effect of emideltide.

The placebo comparisons were not convincing

Falk Bes and colleagues ran a smaller but more interpretable test.

Sixteen people with chronic insomnia entered a double-blind, placebo-controlled study. Eight received intravenous emideltide and eight received glucose placebo across three treatment nights.

Sleep efficiency was higher and sleep latency shorter with emideltide than with placebo. The authors judged those statistically significant effects weak and partly attributable to an incidental change in the placebo group. Other measures, including subjective sleep quality, did not change.

Another randomized double-blind placebo crossover study included six people with severe chronic insomnia. Non-REM sleep time and stage 2 differed significantly from placebo, but the same imbalance was already present at baseline. Most other outcomes did not differ significantly, and the authors judged the observed improvement of little clinical significance.

These trials are also tiny. They do not prove that emideltide has no effect. They do prevent the positive uncontrolled changes from becoming a reliable insomnia claim.

The route gap is absolute

Every human sleep study FDA identified used intravenous administration.

The compounding nomination concerned subcutaneous injection at a proposed concentration of 1,000 micrograms per milliliter.

FDA found no effectiveness study and no clinical safety data for that nominated subcutaneous route.

Changing the route can change exposure, formulation behavior and risk. For an injected peptide, impurities and aggregates can also alter immunogenicity.

An intravenous observation does not establish what a separately manufactured free-base or acetate product will do under the skin.

The adjacent claims are even thinner

The 2026 review also covered narcolepsy and opioid withdrawal.

For narcolepsy, FDA found one case report involving one 35-year-old man. There was no controlled trial and no long-term evidence for a chronic disorder.

For opioid withdrawal, one uncontrolled report included 60 people who used opioids. A second open-label study enrolled seven; only two completed every scheduled injection. Both lacked blinding and comparison groups.

Those reports cannot rescue the sleep claim. They raise different questions, use different outcomes and carry even weaker causal designs.

Relevant professional guidelines reviewed by FDA did not discuss emideltide for chronic insomnia, narcolepsy or opioid withdrawal.

Small studies cannot settle safety

The old intravenous insomnia studies reported no significant adverse effects in small numbers of participants.

FDA’s broader question was whether the proposed compounded use was adequately characterized. The agency identified:

• no clinical safety data for the nominated subcutaneous route;
• no nonclinical toxicity studies identified to inform safety considerations for potential clinical uses;
• incomplete information on impurities, aggregates and endotoxins;
• potential immunogenicity concerns for injected peptide products.

These are uncertainties and plausible risks. They are not proof that a specific marketed product caused harm.

They do show why six volunteers tolerated it is not a safety conclusion.

The 2026 vote was close—and still negative

FDA staff proposed not adding emideltide free base or emideltide acetate to the section 503A Bulks List.

On 24 July, the Pharmacy Compounding Advisory Committee considered both forms separately. Each received six yes votes, seven no votes and one abstention.

That produced advisory recommendations against listing.

The vote did not approve or ban an emideltide drug. It did not itself change the operative list. It did not add efficacy evidence. The committee’s advice is non-binding, and FDA can consider it with the staff reviews, public submissions and other information in later agency action.

As of 4 August 2026, the meeting record reviewed here did not show a later final rule adding either form.

Where the evidence actually lands

Primary sources

1. FDA 2026 Emideltide briefing document
2. FDA July 2026 PCAC meeting record
3. Schneider-Helmert et al., 1981
4. Schneider-Helmert and Schoenenberger, 1983
5. Schneider-Helmert, 1987
6. Bes et al., 1992
7. Monti et al., 1987