DSIP Dosage: How Much Per Night & Protocol - Dosage Peptide

Search for DSIP peptide dosage and you will find charts specifying an exact number of micrograms per night, as though the figure were settled. This article asks a harder and more useful question: what dose of Delta Sleep-Inducing Peptide was actually administered in the published research, what route and timing were used, and does that literature support a per-night dosing protocol at all? The honest answer is that DSIP’s own pharmacology — a plasma half-life measured in minutes in every animal species tested, and reported effects that were explicitly not dose-proportional — is the reason no consensus dose ever emerged, and why the numbers circulating online bear little resemblance to the numbers in the studies.

What Does the Research Actually Say About DSIP Dosage?

The short version, stated plainly before the detail:

There is no validated DSIP dose. No dose-ranging trial was ever completed. No regulatory body has evaluated a DSIP dose for any indication. DSIP is not FDA-approved for any indication, is not approved in the EU, UK, or other major regulated markets, and is sold as a research chemical only.

The classic human studies used a weight-based intravenous dose of 25–30 nmol/kg, given by slow IV injection.[6][9] Converted to mass, that is roughly 1,485 mcg (about 1.5 mg) for a 70 kg adult — roughly fifteen times the ~100 mcg figure most commonly encountered in vendor and forum material, itself an unsourced convention with no published basis we could identify.

DSIP’s plasma half-life is reported on the order of minutes in animals — approximately 4 minutes in dogs, 2.9 minutes in a monkey, and 2.0 minutes in rats.[3] No human pharmacokinetic dataset exists. If human clearance resembles the animal figures, a compound cleared that fast could not hold a blood level across a night’s sleep.

The reported effects were not dose-dependent. Animal work described bell-shaped dose-response curves, where more peptide produced less effect.[5] Human work described a sleep effect with a latency of about an hour but a duration of up to 20 hours.[7] Neither pattern fits a “take X mcg at bedtime” model.

Independent replications largely failed. Two controlled studies concluded DSIP’s sleep effect was “of little clinical significance”[11] and “not likely to be of major therapeutic benefit.”[12]

That combination — minutes-long half-life, non-monotonic dose-response, delayed and persistent effects, and failed replication — is precisely why the field stopped. Any chart presenting a tidy DSIP dose per night is presenting a convention, not a finding. The reconstitution arithmetic below is given because people handling these vials deserve to compute concentrations correctly; it is not an endorsement that a correct number exists to compute.

Research Context

DSIP is a nonapeptide with the sequence Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu and a molecular weight of approximately 848.8 Da. Its origin story is unusual and worth understanding, because the circumstances of its discovery shaped — and arguably oversold — everything that followed.

Where DSIP came from

In the 1960s and 1970s, Marcel Monnier and Guido Schoenenberger at Basel pursued a then-popular hypothesis: that sleep is mediated by a circulating humoral factor. Their method was direct. They electrically stimulated the intralaminar thalamic area of rabbits to induce sleep, collected the cerebral venous blood via extracorporeal dialysis, and fractionated the dialysate hunting for the substance responsible. In 1978 they reported the isolation, amino-acid sequence, and synthesis of the nonapeptide, naming it Delta Sleep-Inducing Peptide.[1]

That original paper is more careful than the name suggests. The synthetic peptide was infused intraventricularly — directly into the brain ventricles — at 6 nmol/kg in rabbits, under double-blind conditions against control solution and eight related peptides. The reported result was a delta and spindle EEG-enhancing effect, with delta activity increased by a mean of about 35% relative to controls.[1] The authors described a “neurohumoral modulating and programming activity” — deliberately vague language, not a claim of sedation.

Two features of that founding experiment matter enormously and are almost always dropped in summaries. First, the route was intraventricular in rabbits, bypassing the blood-brain barrier entirely. Nothing about that establishes what a peripheral injection does in a human. Second, the endpoint was an EEG spectral change, not a report of feeling sleepy. The peptide was named for a change in a frequency band. The name has been doing promotional work ever since that the data never authorized.

The name is the single biggest source of confusion

“Delta Sleep-Inducing Peptide” is a hypothesis embedded in a label. Compare this to how compounds are usually named — had it been called, say, “nonapeptide 9” or given a neutral code, the modern research-chemical market for it would likely not exist. The subsequent literature never converged on DSIP being a sleep-inducing agent in any conventional sense, and its own principal investigators repeatedly noted effects that had nothing to do with sleep: stress tolerance, pain, withdrawal syndromes, chronobiological regulation.[7][2] By 1986, the major review of the field by Graf and Kastin concluded that the various physiological functions of DSIP and its mechanism of action “remain to be established.”[2] That sentence was written four decades ago and has not been superseded.

For the separate question of whether the sleep effect holds up at all — as distinct from this article’s question of what was dosed — see our companion review, does DSIP improve sleep in the research literature. This page deliberately does not relitigate efficacy; it examines the dosing question on its own terms.

Why the literature stopped in the early 1990s

DSIP research has a distinctive shape: a burst of activity from roughly 1977 to 1992, then near-silence. Scanning the indexed literature, the human clinical work is essentially confined to that window. The last substantive controlled human sleep study appeared in 1992 and was negative.[12] What little has appeared since is preclinical and mostly concerns engineered analogues rather than DSIP itself — for instance, fusion constructs designed to survive in circulation and cross the blood-brain barrier, tested in chemically-induced insomnia models in mice.[15]

That silence is informative. Fields do not abandon a compound that works. The pattern here — enthusiastic early reports from a small number of closely-related groups, followed by failed independent replication, followed by abandonment — is a recognizable one in pharmacology, and it is the single most important context for anyone reading a DSIP dosage chart in 2026.

Mechanisms Studied

No receptor for DSIP has been identified. This is not a minor gap; it is the central unresolved fact about the molecule. There is no cloned DSIP receptor, no established binding site, and no accepted signal transduction pathway. Everything below is a proposed mechanism with varying and generally weak support.

Endogenous presence in the brain

DSIP-like immunoreactivity has been reported in mammalian brain tissue and in peripheral tissues, and its distribution was surveyed in Graf and Kastin’s review of the field.[2] An important caveat applies to all of this work: these studies measure “DSIP-like immunoreactivity” (DSIP-LI) using antisera, which detects material that cross-reacts with the antibody. That is not identical to detecting DSIP itself, and the distinction has never been fully resolved for this peptide. The consequence is easy to understate: a substantial fraction of the evidence that DSIP exists endogenously and varies with sleep rests on assays that cannot fully distinguish DSIP from related material. An immunoreactive signal that tracks slow-wave sleep is consistent with DSIP mattering; it is equally consistent with something the antiserum also recognizes mattering.

A study in humans measured cerebrospinal fluid DSIP-LI in 15 drug-free men with schizophrenia alongside polysomnography, and found DSIP-LI correlated with stage 3 and delta sleep measures (reported at p<0.05 for several comparisons).[14] The authors themselves framed this as a first report in a psychiatric population whose generalizability “awaits further study.” It is a correlation in a small, non-representative sample — supportive of DSIP-LI being associated with slow-wave sleep, silent on whether injecting DSIP causes it.

The adrenergic modulation hypothesis

Graf and Kastin’s review raised the possibility that DSIP acts by modulating adrenergic transmission rather than by any direct hypnotic action.[2] This would be consistent with the odd clinical picture reported — a compound that increased daytime alertness and performance in the same studies where it was said to improve night sleep[7] looks more like a modulator of state regulation than a sedative. The hypothesis was never established.

The opioid receptor hypothesis

A separate line of work proposed that DSIP possesses agonist activity at opiate receptors — a hypothesis explicit in the title of the principal human withdrawal paper, which describes DSIP as “a neuropeptide with potential agonistic activity on opiate receptors.”[13] Supporting animal evidence came from Scherschlicht’s group, where DSIP at 25.5 mcg/kg IV attenuated naloxone-precipitated withdrawal jumping in morphine-dependent mice.[5] This reasoning motivated the human withdrawal studies discussed later.

The hypothesis has a specific and unremedied weakness: it was inferred from functional resemblance — naloxone-sensitive effects — rather than from binding. No direct opioid receptor binding by DSIP was ever demonstrated. A compound whose effects are blocked by an antagonist may act at that receptor, or may act upstream of a pathway the antagonist also interrupts; distinguishing the two requires binding studies that were never published for DSIP. Graf and Kastin’s review left the mechanism unresolved,[2] and it has not been resolved since. The opioid hypothesis is best described as an interpretation that guided a research programme, not a finding that programme established.

Blood-brain barrier penetration

If a peripherally injected peptide is to affect sleep, it must reach the brain. Banks and colleagues injected DSIP and four analogues intravenously into dogs and measured DSIP-like material in cerebrospinal fluid sampled from the posterior fossa over 60 minutes. Peptide did enter CSF, and the degree of entry correlated with plasma concentration, plasma half-life, and lipophilicity (combined r = 0.813, p<0.00005), but not with plasma protein binding or molecular weight.[4]

Read carefully, this paper cuts both ways. It establishes that DSIP can cross into CSF — but it also identifies plasma half-life as one of the three determinants of how much gets in. Given that DSIP’s half-life is among the shortest reported for any studied peptide, its own penetration model predicts that very little DSIP reaches the central compartment after a peripheral injection. The 2024 fusion-peptide work exists precisely because researchers recognized that native DSIP’s delivery to the brain is the bottleneck, and engineered around it by fusing DSIP to a blood-brain-barrier-crossing TAT transduction domain (sequence GGGGYGRKKRRQRRR) via flexible GGGGS linkers, expressed and secreted from Pichia pastoris.[15] The resulting DSIP-CBBBP construct outperformed unmodified DSIP on neurotransmitter restoration in the mouse model, which is the clearest available statement of the problem: the peptide had to be rebuilt to get where it was supposed to act. That such engineering was considered necessary is itself a comment on native DSIP.

Non-sleep effects studied

For completeness, DSIP and its analogues have been examined in contexts unrelated to sleep. Graf and Kastin’s review catalogues a strikingly broad reported activity profile — effects on stress tolerance, pain, thermoregulation, endocrine secretion, and withdrawal states — while concluding that the physiological functions and mechanism “remain to be established.”[2] Schneider-Helmert and Schoenenberger similarly reported that DSIP raised alertness and performance during the waking state and appeared to modulate stress tolerance and coping.[7]

This breadth is worth pausing on, because it cuts against the compound rather than for it. A peptide credited with somnogenic, analgesic, anti-stress, endocrine, and anti-withdrawal actions, with no receptor and no established pathway linking any of them, is a peptide whose reported activity has never been localized. Broad, diffuse effect profiles are what one expects either from a genuine upstream regulator or from a literature reporting weak and inconsistent signals across many small studies. The DSIP literature does not adjudicate between those readings — and either way, an unlocalized effect profile is a further reason a clean dose-response was never established.

Why DSIP Resists a Simple Dose-Response Story

Most dosage articles can be written because the underlying compound behaves predictably: more drug produces more effect until a ceiling, the effect tracks the blood level, and the blood level decays on a known schedule. DSIP violates all three premises. This section is the most important on the page, because it explains why the chart you were looking for does not exist — and could not exist on the current evidence.

Problem one: the half-life is measured in minutes

Kato and colleagues developed an enzyme immunoassay for DSIP and used it to measure clearance in animals given intravenous DSIP. The results were unambiguous:

Dog

4.0 ± 0.7 minutes

n = 4 anesthetized dogs; 1–2 mg IV; mean metabolic clearance rate 30.7 ± 2.5 mL/kg/min

Monkey

2.9 minutes

Single animal

Rat

2.0 ± 0.54 minutes

n = 3

Source: Kato et al., Neuroendocrinology 1984.[3] No equivalent human pharmacokinetic dataset appears in the indexed literature.

An important qualification belongs here rather than in a footnote: no human half-life for DSIP has ever been measured. The reasoning that follows extrapolates from dog, monkey and rat, and species extrapolation of clearance is an inference, not a demonstration. Peptide clearance does not scale identically across species, and the animal work used anesthetized preparations and doses far above the human protocol. What follows is the best available reasoning, not proof.

With that stated, sit with what a 2–4 minute half-life would mean. After roughly 20 minutes — five half-lives — approximately 97% of an injected dose is gone from plasma. A person administering DSIP at bedtime and attributing their 3 a.m. sleep quality to a circulating peptide level is describing something difficult to reconcile with the animal pharmacokinetics, which are the only pharmacokinetics that exist for this peptide. For that account to hold, human clearance would have to differ from every species measured by orders of magnitude — possible in principle, but nobody has looked.

This weakens the entire logic of “dose per night.” The phrase presupposes that the dose is what carries you through the night. For DSIP, the animal data give no reason to think it is — and no human data exist either way. Any real effect would more plausibly be a triggered downstream cascade — the peptide acting as a brief signal that sets something else in motion — and the size of a trigger does not scale with effect the way a sedative’s blood level does.

Problem two: the dose-response curve was reported as bell-shaped

Scherschlicht and colleagues at Roche studied synthetic DSIP in rabbits, cats, and morphine-dependent mice. Their findings included sleep effects at 25 mcg/kg IV in rabbits and preferential REM augmentation at 25 mcg/kg IV in cats. But the sentence that matters most for dosing is their conclusion that “in most experimental situations, indications for bell-shaped dose-response curves of DSIP were found.”[5]

A bell-shaped (non-monotonic) dose-response means the effect rises to a peak and then declines as dose increases further. The practical implications are severe:

“More” is not conservative. Under a bell-shaped curve, increasing the dose can move you away from the reported effect, not toward it. The usual intuition that a higher dose is at least a stronger version of the same thing is simply false here.

The peak location is unknown in humans. The animal data suggest a peak exists somewhere; nobody mapped it in people. Without a human dose-ranging study, the location of the maximum is a guess.

Self-titration cannot work. The standard folk method — start low, increase until something happens — assumes monotonicity. On a bell curve, an increase that produces less effect is ambiguous: you cannot tell whether you are below the peak or past it.

Schneider-Helmert’s human work echoed the same theme from a different angle: repeated morning injections still had a strong positive effect on night sleep, while two doses daily did not produce that effect.[8] More total peptide did not yield more benefit. The paper reports the absence of the positive effect under twice-daily dosing rather than a measured decrement, and no head-to-head statistical comparison of the two schedules is given — but even read at its most conservative, that is not how a dose-dependent hypnotic behaves.

Problem three: the reported timing is backwards

Schneider-Helmert and Schoenenberger summarized five human studies and reported a latency to sleep induction of about 1 hour, but a duration of effect of up to 20 hours.[7] An earlier study noted that sleep-promoting effects appeared only in the second hour after injection, and that in the first hour a slight arousing effect was indicated.[6]

Set that beside the 2–4 minute half-life and the mismatch is stark: a compound essentially cleared from plasma within 20 minutes, reported to begin acting at 60 minutes and to persist for up to 20 hours. The effect, if real, is entirely decoupled from the presence of the drug. That decoupling is also why the studies reported carry-over: effects were maintained into placebo nights after treatment stopped.[10][9]

A skeptical reading is available and should be stated: an intervention whose reported benefits appear late, persist for a day, and continue after discontinuation is an intervention whose reported benefits are difficult to distinguish from expectation, regression to the mean, and the substantial placebo response that characterizes insomnia research generally. The original authors interpreted the carry-over as evidence of “programming” a chronobiological system. That is one reading. It is not the only one.

What Doses Were Actually Used in the Human Studies?

Here is where the gap between the literature and the marketplace becomes concrete — and it is much larger than most readers expect.

The studies dosed by body weight, in nanomoles

The human DSIP studies did not use flat microgram doses. They used molar doses scaled to body weight, overwhelmingly 25 nmol/kg, occasionally 30 nmol/kg:

Schneider-Helmert & Schoenenberger 1981[6]

6 middle-aged chronic insomniacs

25 nmol/kg

Acute IV

Longer sleep duration, fewer interruptions; effect only from 2nd hour

Schneider-Helmert 1984[8]

Insomniacs, multiple studies

IV before sleep

Normalization after ~4 administrations; twice-daily dosing did not produce the effect

Schneider-Helmert 1986[9]

18 chronic psychophysiological insomniacs

6 × 30 nmol/kg over 1 week

IV

Improvement; effects correlated with baseline severity

Schneider-Helmert 1987[10]

14 middle-aged chronic insomniacs

Not specified in abstract

IV — 7 successive nights (6 injections)

Improved sleep and daytime performance; maintained into placebo night

Monti et al. 1987[11]

Chronic insomniacs

25 nmol/kg, 4 nights

IV, double-blind crossover

Negative — “of little clinical significance”

Bes et al. 1992[12]

16 chronic insomniacs

25 nmol/kg × 3

IV, afternoon

Negative — “not likely to be of major therapeutic benefit”

Dick et al. 1983[13]

67 patients in withdrawal

Open-label; symptom improvement reported

Converting nmol/kg into micrograms — and the 15× surprise

Because DSIP’s molecular weight is approximately 848.8 g/mol, one nanomole weighs about 848.8 nanograms, or 0.8488 mcg. The conversion is straightforward and worth showing in full so it can be checked:

25 nmol/kg × 848.8 ng/nmol = 21,220 ng/kg = 21.2 mcg/kg

(Molecular weight is conventionally written as 848.8 g/mol; expressing the same quantity as 848.8 ng/nmol is what makes the units cancel correctly against a dose in nmol/kg. The 30 nmol/kg figure works the same way: 30 × 848.8 = 25,464 ng/kg = 25.46 mcg/kg.)

Applied across body weights:

60 kg (132 lb)

~1,273 mcg (1.27 mg)

~1,528 mcg (1.53 mg)

70 kg (154 lb)

~1,485 mcg (1.49 mg)

~1,783 mcg (1.78 mg)

80 kg (176 lb)

~1,698 mcg (1.70 mg)

~2,037 mcg (2.04 mg)

90 kg (198 lb)

~1,910 mcg (1.91 mg)

~2,292 mcg (2.29 mg)

This deserves emphasis because it is rarely stated anywhere: the dose used in the classic human DSIP studies was on the order of 1.3 to 2.3 milligrams — not micrograms. A 70 kg adult in Schneider-Helmert’s protocol received approximately 1,485 mcg intravenously. The figure of ~100 mcg subcutaneously most commonly encountered in vendor listings and forum protocols is roughly one-fifteenth of the studied dose, by a different route.

One caveat about that comparator, stated because this page holds every other claim on it to a citation: the ~100 mcg figure is itself unsourced. It has no published basis, no paper behind it, and no traceable origin that we were able to identify; it is a market convention repeated until it acquired the appearance of a standard. We describe it here as the prevailing convention because it demonstrably circulates, not because any source establishes it. That is exactly the point — the number most readers arrive with is the one number on this page that cannot be cited to anything.

The implication is genuinely awkward for both camps. If you believe the positive DSIP studies, then the conventional ~100 mcg is far below anything ever shown to do anything — a single 5 mg vial would contain barely three studied doses. If you disbelieve the studies (which the failed replications support), then the dose question is moot. There is no coherent position from which ~100 mcg per night is the number the research points to. It appears to be a convention that emerged from the research-chemical market, not from the literature.

One further caveat on the bell-shaped curve: none of the above should be read as an argument that larger doses are therefore better. Under a non-monotonic dose-response, a 15× increase toward the studied figure could equally overshoot a peak. The honest statement is that the relationship between DSIP dose and effect in humans is unmapped, and the studied doses are simply the only numbers with any published basis at all.

How Is DSIP Reconstituted? The Arithmetic

Lyophilized DSIP is supplied as a powder that must be reconstituted with bacteriostatic water (BAC) before any volume can be measured. The arithmetic is the same for every peptide and is worth doing by hand once so the calculator output can be sanity-checked. The site catalog lists DSIP in 5 mg and 10 mg vials, each conventionally reconstituted with 3 mL BAC.

Step 1: concentration

Concentration is total peptide mass divided by diluent volume:

5 mg vial ÷ 3 mL = 1.667 mg/mL = 1,667 mcg/mL

10 mg vial ÷ 3 mL = 3.333 mg/mL = 3,333 mcg/mL

Step 2: volume for a target amount

Volume equals target dose divided by concentration. For a 100 mcg reference amount:

5 mg vial: 100 mcg ÷ 1,667 mcg/mL = 0.06 mL

10 mg vial: 100 mcg ÷ 3,333 mcg/mL = 0.03 mL

Step 3: converting to insulin syringe units

A U-100 insulin syringe divides 1 mL into 100 units, so 1 unit = 0.01 mL. Multiply millilitres by 100:

5 mg vial: 0.06 mL × 100 = 6 units

10 mg vial: 0.03 mL × 100 = 3 units

The full reference table, both vial sizes, at 3 mL BAC:

100 mcg

0.06 mL — 6 units

0.03 mL — 3 units

250 mcg

0.15 mL — 15 units

0.075 mL — 7.5 units

500 mcg

0.30 mL — 30 units

The table stops at 500 mcg deliberately. The studied figure of ~1,485 mcg for a 70 kg adult was a slow intravenous injection given in a sleep laboratory under polygraphic monitoring — it has never been administered subcutaneously to a human in any published study, and rendering it as a number of units on an insulin syringe would convert a monitored hospital procedure into an at-home target it was never validated as. The conversion arithmetic above is fully general: anyone with a legitimate laboratory reason to compute that volume can do so from the concentration. This page will not print it as a draw instruction.

One observation the table does support: the 3-unit measurement required for 100 mcg from a 10 mg vial sits near the practical accuracy floor of a U-100 syringe, where a one-unit error is a 33% dosing error. The 5 mg vial gives better resolution at small amounts, which is the main practical argument for choosing it.

For the vial-specific breakdowns, see the DSIP 5 mg vial dosage protocol and the DSIP 10 mg vial dosage protocol. To check any of the arithmetic above against different diluent volumes, use the peptide dosage calculator, and for the mechanics of the reconstitution itself, the peptide reconstitution guide covers technique. Readers unfamiliar with reading U-100 markings should start with the insulin syringe units guide.

Why diluent volume changes nothing biologically

A recurring confusion deserves direct treatment: changing the amount of BAC does not change the dose. It changes only the concentration, and therefore the volume you draw. Reconstituting the same 5 mg vial with 1 mL instead of 3 mL produces 5,000 mcg/mL, and 100 mcg then occupies 0.02 mL (2 units) rather than 0.06 mL (6 units). Identical peptide, identical amount, different mark on the barrel. More diluent is generally preferable for small target amounts precisely because it spreads the same mass across more units and reduces the proportional impact of a measurement error.

What Route Was Studied, and How Does It Compare to Common Practice?

This is the second large and under-discussed divergence between the DSIP literature and contemporary practice, and it compounds the dose gap described above.

The human studies used intravenous administration

Every human DSIP sleep study identified in the indexed literature administered the peptide intravenously.[6][9][11][12] The withdrawal studies were likewise IV.[13] Not a single controlled human subcutaneous DSIP sleep study appears in the record.

Moreover, the route was not incidental — it was reported as critical. Schneider-Helmert and Schoenenberger stated explicitly that “slow injection proved essential.”[7] The human studies were conducted with polygraphic recording in sleep-laboratory settings,[10] not in circumstances resembling self-administration. Whatever the original investigators believed they were doing, they believed the rate and route of delivery mattered to the result, and they delivered it under monitoring.

Why the IV-to-subcutaneous gap is not a technicality

For many peptides, moving from IV to subcutaneous changes the pharmacokinetics modestly — a slower rise, a lower peak, decent bioavailability. For DSIP specifically, the substitution is far more consequential, and the reason follows directly from the half-life:

Subcutaneous absorption competes with a 2–4 minute clearance. A subcutaneous depot releases peptide gradually into circulation. When the elimination half-life is minutes, peptide is being cleared roughly as fast as it is absorbed, and the peak plasma concentration achievable is far below what a bolus IV injection produces. The area under the curve is not merely reduced; the shape of the exposure is fundamentally different.

Banks’ model predicts less CNS entry. Since CSF entry correlated with plasma concentration and plasma half-life,[4] a route that lowers the plasma peak should lower central exposure further still.

Subcutaneous tissue contains peptidases. Small linear peptides with no protective modifications are vulnerable to degradation at the injection site before absorption. DSIP is a short, unmodified linear nonapeptide with no such protection.

Scherschlicht’s animal data offer a rough calibration. In rabbits, the reported IV effective dose was 25 mcg/kg while the subcutaneous dose was 1 mg/kg — a 40-fold difference. In cats, 25 mcg/kg IV versus 100 mcg/kg subcutaneous, a 4-fold difference. In morphine-dependent mice, 25.5 mcg/kg IV versus doses beyond 85 mcg/kg subcutaneous.[5] The ratios are inconsistent across species, which is itself telling, but in every case the subcutaneous dose required was substantially higher than the intravenous one.

Stack the two gaps together and the picture for the conventional protocol is unflattering: approximately one-fifteenth of the studied dose, delivered by a route that animal data suggest requires several-fold more peptide, targeting a compound whose dose-response may not be monotonic anyway. Whatever that protocol is, it is not a reproduction of the research.

What About Timing — the “Per Night” Question?

The phrase “per night” in this article’s title reflects how people search, so it deserves a direct answer rather than a dodge: the research does not establish a per-night timing protocol, and parts of it actively undercut the premise.

What the studies did regarding timing

Pre-sleep (evening)

Single IV injections of 25 nmol/kg before sleep[8]

Improvement reported; effects began only in the 2nd hour[6]

Morning

Repeated morning injections[8]

Increased daytime activity and still had “a strong positive effect on night sleep”

Twice daily

Two doses daily[8]

Positive night-sleep effect not obtained — more peptide, no more benefit

Afternoon

25 nmol/kg IV in the afternoon before the night[12]

Weak, possibly artefactual effects; negative conclusion

The morning finding is the one that should stop a reader. If a compound injected in the morning still improves that night’s sleep — despite being cleared from plasma before breakfast — then “dose per night, taken at bedtime” is not describing the mechanism at all. And if twice-daily dosing failed to produce the effect that once-daily morning dosing did, the intuition that timing failures can be solved by adding another dose finds no support in the only data available.

Evening administration is therefore best described as a research convention — the natural thing to try when investigating a compound named for sleep — rather than an optimized or validated schedule. It is presented here as a description of what investigators did, not as a recommendation for what anyone should do.

Cycle Length: What the Repeated-Dosing Studies Did

“Cycle length” is a framework imported from anabolic and growth-hormone-secretagogue practice, where receptor downregulation and axis suppression make it meaningful. For DSIP — with no identified receptor and no established axis — the concept has no mechanistic anchor. What the literature offers instead is a handful of fixed-duration administration schedules:

7 successive nights, placebo-controlled and double-blind, in 14 middle-aged chronic insomniacs.[10]

6 injections of 30 nmol/kg over one week, with a one-week follow-up, in 18 chronic psychophysiological insomniacs.[9]

4 consecutive injections, after which normalization of sleep structure was reported.[8][7]

3 injections across three nights in the negative Bes study.[12]

The common thread is roughly one week, but that figure reflects sleep-laboratory logistics — how long you can keep subjects instrumented and funded — far more than any biological rationale. No study compared cycle durations. No study established a maximum safe duration. No study examined tolerance or dependence with extended use, and no study followed a repeated-dosing cohort beyond a few months.

The reported carry-over effects complicate the notion further: if benefits persist into post-treatment placebo nights,[10] then continuous administration has no coherent justification even within the framework the original authors proposed. Anyone presenting a DSIP “cycle” of a defined number of weeks followed by a defined break is inventing structure that the literature does not contain.

Current Evidence Level

Stating the evidence tier precisely, without softening:

FDA-approved for any indication?

No. DSIP is not FDA-approved for any indication, at any dose, by any route, and is not approved in the EU, UK, or other major regulated markets. No approval by the FDA or a comparable regulator has been identified.

Currently in clinical development?

No. No active clinical development programme is evident. Human study effectively ceased in the early 1990s.

Human dose-ranging trial completed?

No. Studies used a fixed 25–30 nmol/kg without systematic dose comparison.

Human pharmacokinetic data?

No published dataset identified. Half-life figures derive from dogs, a monkey, and rats.[3]

Subcutaneous route studied in humans?

No controlled study identified. Human work was intravenous.

Efficacy replicated independently?

Largely no. Positive results cluster around one research group; independent studies were negative.[11][12]

Controlled safety database?

No. Tolerability was described as “good” in small short studies. That is not a safety database.

Mechanism established?

No. No receptor identified; mechanism “remains to be established.”[2]

Regulatory category

Research chemical. Sold for laboratory research use only; not a medicine, not a supplement.

Weighing the positive and negative studies

The positive human evidence is real but structurally weak, and the weakness is specific: nearly all of it comes from a single investigator group. Schneider-Helmert, frequently with Schoenenberger — the peptide’s co-discoverer — authored the great majority of positive reports.[6][7][8][9][10] When the discoverers of a molecule are also the near-exclusive source of evidence that it works, the standard scientific response is to weight independent replication heavily.

The independent replications are the negative ones. Monti and colleagues in Uruguay ran a double-blind crossover of 25 nmol/kg over four nights and found that although several sleep measures moved in the expected direction, none differed significantly from baseline or placebo, and where differences did reach significance for NREM and stage 2 sleep, the same differences already existed at baseline. Their conclusion: “sleep improvement under DSIP treatment is of little clinical significance.”[11] Bes and colleagues in Amsterdam ran a double-blind matched-pairs parallel-groups design in 16 chronic insomniacs and found the statistically significant effects were “weak and in part could be due to an incidental change in the placebo group,” with subjective sleep quality unchanged. Their conclusion: short-term DSIP treatment of chronic insomnia “is not likely to be of major therapeutic benefit.”[12]

Both negative studies used the same dose as the positive ones. This is the crux of the dosing question: the failure to replicate cannot be attributed to underdosing relative to the original protocol, because the dose was matched. That removes the most common escape hatch — “they just didn’t use enough” — and leaves the more uncomfortable possibility that the original effect was smaller than reported, or absent.

The withdrawal literature, briefly

The most dramatic-sounding DSIP claims come from the withdrawal studies. Dick and colleagues administered 25 nmol/kg IV to 67 patients in alcohol or opiate withdrawal and reported a beneficial effect in 48 of 49 evaluable patients — a figure that sounds close to miraculous and circulates widely on that basis.[13] It requires heavy qualification. The study was open-label with no control group; outcomes were assessed by unblinded clinical impression rather than instruments; and 27% of patients were lost or unsuitable for evaluation, meaning the headline denominator is 49, not 67. Withdrawal syndromes also resolve on their own over exactly the timescale studied.

Those features interact badly. When an unblinded clinician rates a self-limiting condition in a cohort that has already lost a quarter of its patients, a near-universal response rate is close to the expected result whether or not the drug does anything. That is not a reason to conclude the peptide is inert; it is a reason to conclude the study cannot tell us. An uncontrolled study of a self-limiting condition cannot establish efficacy, and this work was never replicated under controlled conditions — which, four decades on, is itself the most informative fact about it.

Handling, Storage, and Measurement Accuracy

Even where a dose is uncertain, the physical handling of a lyophilized peptide follows general principles that determine whether the amount in the syringe corresponds to the amount on the label. These points are process, not protocol.

Stability considerations specific to DSIP

DSIP is a short, linear, unmodified peptide with no cyclization, no D-amino acid substitutions, and no C-terminal amidation — the modifications that confer stability in engineered peptides. Two structural features are worth noting. The original characterization found that only the pure alpha-aspartyl peptide is highly active, in contrast to its beta-Asp isomer.[1] Aspartyl residues in peptides are known to be susceptible to isomerization, and DSIP contains both an Asp residue and adjacent Gly — the sequence context most associated with that rearrangement. Additionally, the N-terminal tryptophan is among the more oxidation-prone and light-sensitive residues.

The practical consequence is that identity and activity can diverge for this molecule: material could assay as present while having partially converted to a form the founding paper reported as far less active. This is a further, independent reason why dose figures for DSIP carry unusual uncertainty, and why analytical certificates should be read as a statement about the material at the time of testing.

General handling principles

Reconstitute gently. Direct the diluent stream against the vial wall rather than onto the powder, and swirl rather than shake. Agitation shears peptides.

Protect from light. Relevant for any tryptophan-containing peptide.

Cold chain. Lyophilized powder is markedly more stable than solution; once reconstituted, refrigeration and a bounded in-use window apply. Our guide to storing peptides before and after reconstitution covers the general framework.

Diluent choice. Bacteriostatic water contains benzyl alcohol as a preservative; sterile water does not and offers no protection against microbial growth across repeated vial entries.

Vial capacity. Confirm the vial physically holds the intended diluent volume before adding it. The 5 mg and 10 mg DSIP vials in the site catalog are specified for 3 mL, so the worked examples above fit; the caution matters when adapting this arithmetic to smaller vials from other sources, where 3 mL can exceed the available headspace. If a vial will not take 3 mL, recompute the concentration for the volume it does take rather than forcing the standard figure — the method in the section above is what carries over, not the specific numbers.

The measurement accuracy problem at small volumes

The arithmetic above produces some very small volumes, and small volumes carry disproportionate error:

100 mcg from 10 mg / 3 mL

0.03 mL

3 units

±33%

100 mcg from 5 mg / 3 mL

0.06 mL

6 units

±17%

250 mcg from 5 mg / 3 mL

0.15 mL

15 units

±7%

500 mcg from 5 mg / 3 mL

0.30 mL

30 units

±3%

A 3-unit draw is close to the resolution limit of the instrument, compounded by hub dead space and meniscus reading. One caution about how to read this table: proportional measurement error falls as the drawn volume rises, but that is a statement about the syringe, not about the peptide. Better measurement precision at a larger amount is not an argument for using a larger amount — under a non-monotonic dose-response, a precisely measured wrong dose is still a wrong dose. Accuracy and appropriateness are separate questions, and only the first is settled by arithmetic. For readers encountering terms such as lyophilized, bacteriostatic, nonapeptide, or bell-shaped dose-response for the first time, our peptide research glossary defines the vocabulary used throughout this library.

Limitations

The limitations of the DSIP dosing literature are not peripheral caveats — they are the main finding. Set out plainly:

Limitations of the evidence base

No human dose-ranging study exists. The 25–30 nmol/kg figure was chosen early and repeated, never optimized. It is a historical artefact, not an established dose.

No human pharmacokinetic data. Every half-life figure quoted for DSIP, on this page and everywhere else, comes from dogs, one monkey, and rats.[3] Human clearance is genuinely unknown.

Tiny samples. The human sleep studies enrolled roughly 6 to 18 subjects each. These sample sizes cannot detect modest effects or uncommon harms.

Single-group provenance. Positive findings cluster around the discoverers; independent groups reported negative results at the same dose.

Old methodology. This work predates modern trial registration, pre-specified endpoints, and publication-bias safeguards. Selective reporting cannot be excluded or assessed.

Non-representative populations. The dosed subjects were chronic insomniacs and patients in alcohol or opiate withdrawal; separately, one observational CSF study measured DSIP in 15 men with schizophrenia rather than administering it.[14] No published human study administered DSIP to healthy adults for sleep, which is precisely the population buying it. None of this literature speaks to that group.

Publication bias by abandonment. A field that stops publishing may be a field where negative results went unpublished. The absence of post-1992 human data is not neutral.

Limitations of any dosing guidance, including this page

The reconstitution math is arithmetic, not validation. Every calculation here is checkable and correct as arithmetic. That says nothing about whether the resulting amount does anything.

The bell-shaped curve makes extrapolation unsafe. Interpolating between doses assumes monotonicity, which the animal data contradict.[5]

Route substitution is unvalidated. Converting an IV protocol to subcutaneous has no evidential basis for this peptide.

Research-chemical material is of unverified identity and purity. Compounds sold for research use are not manufactured to pharmaceutical standards. Content, purity, isomeric composition, and endotoxin burden are not assured, and third-party certificates vary in rigour.

No controlled safety database exists. “Compatibility of DSIP was good” in a 1983 summary of five small human studies[7] is not a safety profile — that paper reports no total subject count, and tolerability described in a narrative summary is not systematically collected adverse-event data. Long-term effects, interactions, and effects in any special population are entirely uncharacterized.

What DSIP is not

A closing clarification, because the name invites the error: DSIP is not a benzodiazepine-like sedative and does not behave like one. It has no established GABAergic activity, no demonstrated receptor, no reliable acute sedative effect, and no dose at which sedation predictably occurs. The reported human profile — onset around an hour, duration up to 20 hours, increased daytime alertness, effects persisting after discontinuation — describes something categorically unlike a hypnotic. Reasoning about DSIP by analogy to sleep medications imports assumptions that its own literature does not support.

Frequently Asked Questions

What is the standard DSIP dosage per night?

There is no standard or validated DSIP dosage. No dose-ranging trial was ever performed and DSIP is not approved for any indication. The classic human studies used 25–30 nmol/kg intravenously — approximately 1,485 mcg for a 70 kg adult at 25 nmol/kg — administered in sleep laboratories. The ~100 mcg figure widely quoted online is roughly one-fifteenth of that, by a different route, and has no published basis.

How do I reconstitute a 5 mg DSIP vial?

With 3 mL of bacteriostatic water, a 5 mg vial yields 5 ÷ 3 = 1.667 mg/mL, or 1,667 mcg/mL. A 100 mcg reference amount then equals 100 ÷ 1,667 = 0.06 mL, which is 6 units on a U-100 insulin syringe. A 10 mg vial in 3 mL yields 3,333 mcg/mL, so 100 mcg equals 0.03 mL, or 3 units. Adding more or less diluent changes the concentration and the volume drawn, never the amount of peptide.

Why is DSIP dosing so uncertain compared to other peptides?

Three reasons compound. Its plasma half-life is roughly 2–4 minutes in animals, so it cannot maintain a level overnight. Animal work reported bell-shaped dose-response curves, meaning more peptide can produce less effect. And human reports described onset around an hour with duration up to 20 hours — an effect decoupled from drug presence. No consensus dose emerged because the pharmacology never supported one.

Is DSIP taken subcutaneously or intravenously in the research?

Every identified human DSIP sleep study used intravenous administration, and investigators reported that slow injection was essential. No controlled human subcutaneous sleep study exists. Animal data suggest subcutaneous routes required substantially higher doses than intravenous — in rabbits, 1 mg/kg subcutaneous versus 25 mcg/kg IV, a 40-fold difference. The common subcutaneous practice is not a reproduction of the studied protocol.

How long is a DSIP cycle in the research?

The repeated-dosing studies ran roughly one week, typically 4 to 7 injections. No study compared cycle lengths, established a maximum duration, or examined tolerance. The one-week figure reflects sleep-laboratory logistics rather than biology. Any defined DSIP “cycle with a break” circulating online is invented structure, not a research finding.

Does DSIP actually improve sleep?

The evidence is mixed and weak. Positive reports came almost entirely from the group that discovered the peptide. Two independent controlled studies using the same dose concluded the effect was “of little clinical significance” and “not likely to be of major therapeutic benefit.” Because the negative studies matched the positive studies’ dose, the failure cannot be explained by underdosing. Our companion article reviews this question in depth.

Is DSIP FDA-approved or available by prescription?

No. DSIP is not FDA-approved for any indication, and is not approved in the EU, UK, or other major regulated markets. It is not a prescription medicine, not a supplement, and is not in active clinical development. Material sold under this name is a research chemical intended for laboratory research use only, manufactured outside pharmaceutical quality standards.

What is the difference between 5 mg and 10 mg DSIP vials?

Only the mass of powder. Reconstituted with the same 3 mL, the 10 mg vial is exactly twice as concentrated (3,333 versus 1,667 mcg/mL), so any given amount occupies half the volume. This matters for measurement accuracy: a 100 mcg amount is 3 units from a 10 mg vial versus 6 units from a 5 mg vial, and at 3 units a single-unit misread is a 33% error.

Can DSIP dosing be titrated upward if nothing happens?

Standard titration logic assumes a monotonic dose-response — that more produces more. Animal studies reported bell-shaped curves for DSIP, and in the human work twice-daily dosing did not produce the night-sleep effect that once-daily morning injections did. Under a non-monotonic curve, an increase producing less effect is ambiguous: it cannot distinguish being below the peak from having passed it. The relationship between DSIP dose and effect in humans has never been mapped.

References

Schoenenberger GA, Maier PF, Tobler HJ, Wilson K, Monnier M. The delta EEG (sleep)-inducing peptide (DSIP). XI. Amino-acid analysis, sequence, synthesis and activity of the nonapeptide. Pflugers Arch. 1978;376(2):119–29. doi:10.1007/BF00581575

Graf MV, Kastin AJ. Delta-sleep-inducing peptide (DSIP): an update. Peptides. 1986;7(6):1165–87. doi:10.1016/0196-9781(86)90148-8

Kato N, Honda Y, Ebihara S, Naruse H, Takahashi Y. Development of an enzyme immunoassay for delta sleep-inducing peptide (DSIP) and its use in the determination of the metabolic clearance rate of DSIP administered to dogs. Neuroendocrinology. 1984;39(1):39–44. doi:10.1159/000123952

Banks WA, Kastin AJ, Coy DH, Angulo E. Entry of DSIP peptides into dog CSF: role of physicochemical and pharmacokinetic parameters. Brain Res Bull. 1986;17(2):155–8. doi:10.1016/0361-9230(86)90111-5

Scherschlicht R, Aeppli L, Polc P, Haefely W. Some pharmacological effects of delta-sleep-inducing peptide (DSIP). Eur Neurol. 1984;23(5):346–52. doi:10.1159/000115712

Schneider-Helmert D, Schoenenberger GA. The influence of synthetic DSIP (delta-sleep-inducing-peptide) on disturbed human sleep. Experientia. 1981;37(9):913–7. doi:10.1007/BF01971753

Schneider-Helmert D, Schoenenberger GA. Effects of DSIP in man. Multifunctional psychophysiological properties besides induction of natural sleep. Neuropsychobiology. 1983;9(4):197–206. doi:10.1159/000117964

Schneider-Helmert D. DSIP in insomnia. Eur Neurol. 1984;23(5):358–63. doi:10.1159/000115714

Schneider-Helmert D. Efficacy of DSIP to normalize sleep in middle-aged and elderly chronic insomniacs. Eur Neurol. 1986;25(6):448–53. doi:10.1159/000116050

Schneider-Helmert D. Effects of delta-sleep-inducing peptide on 24-hour sleep-wake behaviour in severe chronic insomnia. Eur Neurol. 1987;27(2):120–9. doi:10.1159/000116143

Monti JM, Debellis J, Alterwain P, Pellejero T, Monti D. Study of delta sleep-inducing peptide efficacy in improving sleep on short-term administration to chronic insomniacs. Int J Clin Pharmacol Res. 1987;7(2):105–10. PMID: 3583493

Bes F, Hofman W, Schuur J, Van Boxtel C. Effects of delta sleep-inducing peptide on sleep of chronic insomniac patients. A double-blind study. Neuropsychobiology. 1992;26(4):193–7. doi:10.1159/000118919

Dick P, Grandjean ME, Tissot R. Successful treatment of withdrawal symptoms with delta sleep-inducing peptide, a neuropeptide with potential agonistic activity on opiate receptors. Neuropsychobiology. 1983;10(4):205–8. doi:10.1159/000118012

van Kammen DP, Widerlöv E, Neylan TC, Ekman R, Kelley ME, Mouton A, Peters JL. Delta sleep-inducing-peptide-like immunoreactivity (DSIP-LI) and delta sleep in schizophrenic volunteers. Sleep. 1992;15(6):519–25. doi:10.1093/sleep/15.6.519

Mu X, Qu L, Yin L, Wang L, Liu X, Liu D. Pichia pastoris secreted peptides crossing the blood-brain barrier and DSIP fusion peptide efficacy in PCPA-induced insomnia mouse models. Front Pharmacol. 2024;15:1439536. doi:10.3389/fphar.2024.1439536

Bibliographic records for the above were verified against PubMed.

Research use only. This article is an educational review of published research and is not medical advice. DSIP is not FDA-approved for any indication and is not approved in the EU, UK, or other major regulated markets. It is not a medicine or a supplement, and no validated human dose, route, timing, or cycle exists for it. Nothing here recommends, endorses, or instructs human use of DSIP, and the doses described are reported solely as a record of what was administered in published research settings. dosagepeptide.com is an independent reference library; it does not sell peptides and has no clinical relationship with readers. Anyone with concerns about sleep should consult a qualified healthcare professional about evidence-based options.