There is no established dose of DSIP, but there is something more unusual than a blank: a single figure that four separate research groups arrived at and used, over six years and in four different patient populations. That figure is 25 nanomoles per kilogram of body weight, given as an intravenous injection, and understanding why it is expressed that way explains most of what a dosage question about this peptide runs into.
What those studies found is a separate matter, covered in what the DSIP trials in people found.
The dose every study used
| Study | People | Dose and route | Schedule |
|---|---|---|---|
| Bes 1992 | 16 chronic insomniacs | 25 nmol/kg, intravenous | Afternoon before three consecutive nights |
| Monti 1987 | Chronic insomniacs, crossover | 25 nmol/kg, intravenous | Four nights |
| Dick 1983 | 67 patients in withdrawal | 25 nmol/kg, intravenous | Single injection, sole treatment |
| Bjartell 1989 | 11 healthy men | 25 nmol/kg, intravenous | Single dose, crossover against saline |
| Schneider-Helmert 1987 | 14 chronic insomniacs | Injection; dose not stated in the abstract | Seven successive nights |
A molar dose per kilogram is how a laboratory expresses a peptide it is studying as a signaling molecule rather than as a product. It counts molecules, not milligrams, which is the unit that matters when the question is whether a receptor is occupied.
It is also the reason a milligram figure is hard to find. Nobody in these studies was thinking about what a person would draw into a syringe at home, so nobody printed the number that question needs.
One study is missing from that column on purpose. The 1987 report on 24-hour sleep-wake behavior treated 14 middle-aged chronic insomniacs for seven successive nights under double-blind conditions, and its abstract gives no dose at all [8].
The route is not incidental
Every human dose above went into a vein, and in the two insomnia studies it went in during the afternoon rather than at bedtime [1][2].
That timing tells you what the investigators believed they were doing. They were not sedating anyone at the moment of sleep; they were testing whether an afternoon injection would shift the following night.
A pharmacodynamic effect at that dose was measurable. In eleven healthy men given 25 nmol/kg intravenously, ACTH-like immunoreactivity in plasma fell for at least three hours afterward, while cortisol followed its usual daily decline [4].
So the dose does something. Whether it does the thing people buy it for is the part the sleep studies were unable to establish.
What that dose produced
The 1992 study gave 25 nmol/kg before three consecutive nights and found higher sleep efficiency and shorter sleep latency than placebo [1].
Its own conclusion is the one worth quoting: the significant effects were weak, and in part could be due to an incidental change in the placebo group. Short-term treatment of chronic insomnia this way, the authors write, is not likely to be of major therapeutic benefit.
The 1987 crossover study used the same dose over four nights and reached a similar place [2]. Sleep improvement, it concludes, is of little clinical significance.
The largest human experience at this dose is not a sleep study at all. Sixty-seven patients withdrawing from alcohol or opiates received a single 25 nmol/kg injection as their only treatment, and 48 of the 49 evaluable patients improved [3]. That study was open, with no placebo arm.
Why a blood level is not a dose either
DSIP occurs naturally, which invites the idea that a dose could be set by topping up a deficit. Measurement does not support that framing.
A 1995 study compared morning plasma concentrations across sleep apnea patients, narcolepsy patients and normal controls, and found no significant differences between the three groups [5].
Its authors conclude that a single plasma measurement does not appear to work as a marker of disease activity. There is no deficiency to correct and no level to aim at.
No pharmacy can make it
DSIP is named on neither FDA compounding list: not in the 503A bulk drug substances categories updated May 14, 2026, nor in the 503B categories updated March 21, 2025 [6].
With no approved product and no compounding route, nothing establishes the strength, purity or sterility of material sold under the name, which is a separate problem from the dose itself.
That distinction matters when a schedule is quoted with confidence: a number that looks precise says nothing about what is in the vial it is measured out of. See whether peptides are safe.
Putting it together
The only human dose of DSIP on record is 25 nmol/kg intravenously, used in four studies between 1983 and 1992, and it was never converted to milligrams or tested under the skin.
Both sleep studies at that dose reported their own results as weak. Nothing published supports a nightly schedule, a milligram figure or a maintenance dose.
For what else has been tried against the same goal, see peptides for sleep and the sleep hub.