The DSIP 1992 insomnia study found weak changes on sleep recordings, without people reporting better sleep. Chronic insomnia means ongoing trouble falling or staying asleep. Sixteen people took part, and treatment lasted only three nights. The authors judged a major short-term benefit unlikely.
DSIP stands for delta sleep-inducing peptide, a short chain of protein building blocks. Its name sounds more definite than this experiment's results. Bes and colleagues' paper in Neuropsychobiology, 1992 is useful precisely because it separates a measurable change from a meaningful improvement.
What the DSIP 1992 insomnia study asked
The question was whether DSIP changed sleep in people with chronic insomnia. Researchers compared it with placebo, an inactive glucose (sugar) solution, in a sleep laboratory.
The study was double-blind: participants and the researchers assessing them did not know which treatment they received. That helps keep expectations from shaping the results. It doesn't remove the uncertainty that comes with a very small group.
Who took part, and what happened each night?
The 16 participants were divided between DSIP and placebo, eight in each group. The design used matched pairs, meaning participants were paired to make the groups more comparable. They stayed in parallel groups: each person received one assigned treatment rather than switching between both.
The public abstract doesn't specify an age range or detailed exclusion criteria. Its findings therefore cannot establish how DSIP would affect children, pregnancy, or people with a different sleep disorder.
| Laboratory night | What happened |
|---|---|
| 1 | Participants became accustomed to sleeping in the laboratory |
| 2 | Researchers recorded baseline sleep, the starting measurements before treatment |
| 3–5 | DSIP or glucose placebo was given in the afternoon before each night's recording |
The experimental dose was 25 nanomoles per kilogram of body weight, an amount adjusted for body size. It was given intravenously, meaning into a vein. That describes this monitored study; it does not establish a dosing schedule for another preparation.
What did the recordings and morning ratings measure?
Researchers used polysomnography, overnight recordings of brain waves and other sleep signals. They measured sleep efficiency, the share of time in bed spent asleep, and sleep latency, the time needed to fall asleep. They also examined how sleep was divided among its stages.
Participants supplied the other half of the picture: their own ratings of sleep quality and tiredness. A recording can show a change while the person still feels they slept poorly.
The abstract describes several measurements rather than identifying one primary endpoint, the main outcome chosen beforehand. That makes it especially important to consider the overall pattern rather than selecting the most favorable finding.
What changed compared with placebo?
The headline result was weak objective improvement without better self-rated sleep. The abstract reports higher sleep efficiency and shorter time to fall asleep with DSIP than placebo. FDA's reviewers, summarizing the full paper, read it more cautiously: they described DSIP's shorter time to fall asleep as only numerical and found no significant difference between the groups in objective sleep measures or sleep quality. It also reports a decrease in one tiredness rating within the DSIP group.
That last comparison is easy to overread. A change inside the DSIP group alone doesn't show that DSIP beat placebo.
The authors cautioned that the statistically significant findings were weak and might partly reflect a chance change in the placebo group. Statistical significance means a finding passed the study's test for separating a signal from chance. It doesn't establish a change large enough to matter in daily life.
FDA, the Food and Drug Administration, revisited the trial in its briefing document for the July 2026 Pharmacy Compounding Advisory Committee meeting, pages 27–28. The placebo group's time to fall asleep was unusually high on one treatment night, while the DSIP group's corresponding measure did not change. A gap between groups can grow because the comparison group worsens, even without improvement in the treated group.
The unchanged sleep-quality ratings matter alongside those recordings. The study did not establish that participants felt they were sleeping better.
What does the trial say about side effects?
Its safety reporting cannot support a claim of zero reactions. In Appendix 6, FDA's review marks adverse events for Bes's 1992 study as “Not addressed.” Adverse events are unwanted health changes recorded during a study.
| Safety question | What the trial-specific FDA table documents |
|---|---|
| Were adverse events described? | “Not addressed” |
| Is there an event-by-event DSIP versus placebo comparison? | None supplied in that summary |
| Does this establish that nobody had a reaction? | It does not; missing reporting is different from zero events |
Even careful reporting from eight DSIP recipients would struggle to characterize uncommon reactions. Three treatment nights also cannot answer questions about months of use. The trial's sleep measurements and its safety reporting are separate limits.
How does it compare with the 1987 crossover study?
The earlier Monti and colleagues study, International Journal of Clinical Pharmacology Research, 1987 used a crossover design: participants received DSIP and placebo during separate periods.
Its abstract reports no significant improvement over baseline or placebo for several measures of disrupted sleep. Some differences in a lighter stage of sleep were already present before treatment. The authors judged the overall improvement to have little clinical significance.
Those are different experiments, but both show why the comparison matters as much as the favorable finding. FDA's later review called the overall intravenous insomnia evidence inconclusive and preliminary. It also found no effectiveness study for administration under the skin.
The broader DSIP benefits overview puts these findings beside the other early experiments. Neither study settles the evidence for a current compounded preparation.
What can this study mean for a prescription today?
This was a tiny human experiment, rather than a larger program confirming a treatment's benefits and risks. The phases of clinical trials explain that difference. Where DSIP stands in FDA's separate review of compounding ingredients is covered in the current DSIP update.
As of November 5, 2026, the day this article was written, an October 30 Federal Register notice announces a draft policy from the Department of Health and Human Services (HHS). It describes conditions under which the department does not intend to take action against certain peptide compounders. DSIP is absent from its six-peptide appendix. The draft is nonbinding; the HHS peptide policy explainer covers its conditions.
Compounded DSIP is not FDA-approved. A licensed provider may still prescribe — that decision is between you and your doctor. If prescribed through Promise, a licensed U.S. compounding pharmacy prepares and dispenses the medication.
At Promise, a licensed provider reviews every request, and not everyone qualifies. The prescriber sets any individual dose and timing. The useful question is whether the evidence fits the person's sleep problem, medical history and current medicines, rather than whether one old result sounds encouraging.
