DSIP dosage and evidence

What is DSIP?

How does DSIP work?

It is thought to affect sleep-related brain activity and stress hormones. Nobody knows exactly what it binds to.

What have studies in people found about DSIP?

DSIP has been given to people in small studies published between 1981 and 1998, always into a vein. The two teams that tested it for insomnia against a dummy, with nobody knowing who got what, both found the effect too weak to matter. The studies with big improvements come mostly from one author. He compared patients with themselves before treatment, or with a separate group of good sleepers. FDA reviewed the whole record in 2026 and found it was not enough.

• Withdrawal, Geneva hospital, 1984, 107 patients: 47 coming off alcohol and 60 coming off opioids got 25 nmol/kg into a vein, up to six times a day for up to six days. Of the 60 on opioids, the 40 who were assessed had clear drops, or complete loss, of visible withdrawal signs: racing heart, high blood pressure, shaking, sweating, goosebumps, yawning, watery eyes, diarrhoea and runny nose. Anxiety eased slowly. Within one to three days of stopping, it came back with severe sleeplessness in all but 13. Everyone got the drug and knew it, there was no comparison group, and no standard withdrawal scale was used. FDA notes the study did not explain well how improvement was checked.
• Insomnia, 1986, 18 patients in two age groups (29 to 59 and 60 to 83): six doses of 30 nmol/kg into a vein over one week, with no dummy group. Sleep was back to normal by the end of dosing in the younger group, and by the end of a week of follow-up in the older group. The worse their sleep at the start, the bigger the change.
• Insomnia, 1992, 16 patients in matched pairs (Bes and colleagues): each got 25 nmol/kg into a vein, or a sugar-water dummy, before three of five nights in a sleep lab, and nobody knew who got what. People on DSIP slept a little more efficiently and fell asleep faster. But the authors called these effects weak, and partly down to a chance shift in the dummy group. Nothing else changed, including how well people felt they slept. They concluded a short course of DSIP is not likely to be of much help for long-term insomnia.
• Insomnia, 1987, 14 middle-aged patients: seven nightly shots of 30 nmol/kg into a vein. Time to fall asleep dropped from 58.4 minutes at the start to 38.9 after the first shot and 27.6 after the last. Time awake after first falling asleep dropped from 120.8 minutes to 78, then 65.6. Total sleep rose from 313 minutes to 363.1, then 385, and sleep efficiency rose by 27 percent of the starting value. FDA's 2026 review notes there was no dummy group, so there were no dummy results. The comparison was a separate group of healthy sleepers.
• Insomnia, 1987, each person tried DSIP and a dummy on different nights, across four nights, and nobody knew which was which (Monti and colleagues): 25 nmol/kg into a vein. Waking in the night, time to reach deeper sleep, and time awake all fell on DSIP. None of it was clearly different from the starting nights or from the dummy nights. Where two sleep measures did differ from the dummy, they already differed before treatment. The authors concluded the sleep gain under DSIP was of little medical value.
• Opioid withdrawal, 1998, 7 people dependent on opioids: 35 nmol/kg into a vein on a fixed schedule, with everyone knowing what they got. Average withdrawal scores fell from 34.4 an hour before the first shot to 17.8 after it. Only 2 of the 7 got every planned shot and finished detox, and both needed the drug doxepin for sleeplessness on day 3. After the second and later shots, withdrawal symptoms kept coming back in everyone but one. The authors wrote that whether DSIP could help with detox is still an open question.
• Long-term pain, 1984, 7 patients with migraine, headache, long-standing ringing in the ears or pain attacks with no clear cause: shots into a vein on five days in a row, then five more every two to three days. Pain dropped clearly in 6 of the 7 compared with their own earlier readings, and low mood lifted at the same time. There was no comparison group, and chance did not decide anything.
• Insomnia, 1981, 6 middle-aged patients, each trying DSIP and a dummy in turn without knowing which (described in FDA's May 2026 review): 25 nmol/kg into a vein over 4 minutes at bedtime. It was slightly stimulating in the first hour, then helped sleep from the second hour for up to 6 hours. It did not change how long it took to fall asleep or when people woke up for good. Daytime sleepiness scores did not differ.
• Narcolepsy, a single case: a 35-year-old man who had had it for 10 years got 25 nmol/kg into a vein for a week in the evening, and a week in the morning. Sleep attacks fell from 6 to 3 on either schedule, and naps went from daily to every second or third day. FDA notes one person cannot show whether a drug works.
• Route: every published study in people put DSIP into a vein. FDA's 2026 review found no study of the shot under the skin, the way it is sold, for insomnia, narcolepsy or opioid withdrawal.
• At FDA's July 2026 advisory meeting on pharmacy compounding, emideltide, FDA's name for DSIP, was the one peptide up for review that the panel voted not to recommend for the compounding list.

What this does not tell you: The newest study in people here came out in 1998, and the newest insomnia study in 1992. Sleep studies ran from 1 to 18 people, and withdrawal studies from 7 to 107. Every study put the drug into a vein, and nothing has been published on the shots under the skin or nasal sprays that are sold. The good insomnia results come mostly from one author, comparing people with themselves or with a separate group. The two outside teams that used a dummy for comparison did not get the same results. There is no work on the best dose, no long-term data, and no check of whether any sleep change was big enough to matter by today's sleep medicine standards. FDA also notes that in human blood at body temperature, half of it was gone in 5 to 10 minutes. That fits poorly with effects said to last all night.

The state of the research

The evidence, in charts

The studies we cite

• Study in people1992
  Effects of delta sleep-inducing peptide on sleep of chronic insomniac patients. A double-blind study

  Sleep improved slightly, and the researchers said the improvement was not worth much. 16 people with long term sleeplessness were matched into pairs, and nobody knew who got DSIP and who got sugar water. Each had a vein injection before three of five nights in a sleep lab. Those on DSIP fell asleep faster and slept a little more efficiently. Nothing else changed, including how people said they slept. The authors traced part of the gain to a chance shift in the dummy group.

  16 chronic insomniacs, double-blind matched-pairs parallel groups, 25 nmol/kg intravenously before three of five laboratory nights. Sleep efficiency was higher and latency shorter on DSIP, but the authors judged the effects weak and partly attributable to a change in the placebo group and concluded that short-term treatment is not likely to be of major therapeutic benefit.

  PMID 1299794 Neuropsychobiology
• Study in people1987
  Study of delta sleep-inducing peptide efficacy in improving sleep on short-term administration to chronic insomniacs

  The sleep numbers moved, and none of the movement could be separated from where people started. Each person tried DSIP and a dummy in turn, across four nights, with the injection going into a vein. Nobody involved knew which night was which. Awakenings, time awake and time to reach deep sleep all fell on DSIP. None of those falls stood apart from the person's own baseline or from the dummy nights. The authors called the improvement of little clinical meaning.

  Double-blind crossover, 25 nmol/kg intravenously over four nights. Awakenings, NREM latency, total waking time and waking after sleep onset fell on DSIP but did not differ significantly from baseline or from placebo nights. The authors concluded sleep improvement under DSIP is of little clinical significance.

  PMID 3583493 International Journal of Clinical Pharmacology Research
• Study in people1987
  Effects of delta-sleep-inducing peptide on 24-hour sleep-wake behaviour in severe chronic insomnia

  The numbers are large, and there was no comparison group to make sense of them. 14 middle aged people with long term sleeplessness got seven nightly injections into a vein. Time to fall asleep dropped from about an hour to under half an hour, and the time spent awake at night roughly halved. Here is the problem. Nobody went without the drug. The comparison was a separate group of healthy sleepers, so there is no way to know what would have happened anyway.

  14 middle-aged chronic insomniacs given seven nightly intravenous injections of 30 nmol/kg. Sleep onset latency fell from 58.4 to 38.9 to 27.6 minutes and wake after sleep onset from 120.8 to 78 to 65.6 minutes, with daytime performance improved. FDA's 2026 review notes the study had no placebo arm and was compared against an external control group of healthy sleepers.

  PMID 3622582 European Neurology
• Study in people1986
  Efficacy of DSIP to normalize sleep in middle-aged and elderly chronic insomniacs

  Sleep was reported to return to normal, and everybody in the study got the drug. 18 people with long term sleeplessness had six injections into a vein over a week. Sleep normalised by the end of dosing in the younger group and by the end of a follow up week in the older one. There was no dummy group. With nothing to compare against, normal sleep cannot be separated from time passing and from being in a study.

  18 chronic psychophysiological insomniacs given six intravenous doses of 30 nmol/kg over a week with a one-week follow-up. Sleep normalized by the end of dosing in the 29 to 59 group and by the end of follow-up in the 60 to 83 group. No placebo arm.

  PMID 3792404 European Neurology
• Study in people1984
  DSIP in the treatment of withdrawal syndromes from alcohol and opiates

  Withdrawal signs eased, and the study cannot prove the drug did it. 107 hospital patients in Geneva coming off alcohol or opioids got vein injections up to six times a day. Most of those assessed had a clear drop in visible signs: racing heart, tremor, sweating and watering eyes. Within one to three days of stopping, anxiety and severe sleeplessness returned in all but 13. Everybody got the drug, nobody was kept in the dark, and no standard scale was used.

  107 inpatients in Geneva, 47 with alcohol and 60 with opioid withdrawal, given 25 nmol/kg intravenously up to six times a day for up to six days. Objective withdrawal signs improved markedly in most assessed patients, but anxiety and insomnia relapsed within 24 to 72 hours of stopping in all but 13. Open label, no blinding, no comparison group, no standard scale.

  PMID 6548969 European Neurology
• Study in people1984
  Therapeutic effects of delta-sleep-inducing peptide (DSIP) in patients with chronic, pronounced pain episodes. A clinical pilot study

  Pain fell in six of seven people, and there was nobody to compare them with. Seven patients with migraine, headache, long standing tinnitus or unexplained pain attacks got DSIP into a vein, daily for five days and then every few days. Their pain scores fell against their own earlier readings, and their low mood lifted at the same time. No control group. No sorting by chance. This is a first look and not an answer.

  7 patients with migraine, vasomotor headache, chronic tinnitus or psychogenic pain given intravenous DSIP on five consecutive days then five injections every 48 to 72 hours. Pain fell significantly in 6 of 7 against their own baseline period, with depressive symptoms also falling. No control group.

  PMID 6548970 European Neurology
• Review of studies2006
  Delta sleep-inducing peptide (DSIP): a still unresolved riddle

  A review. Decades on, nobody has found DSIP's gene, its parent protein or its receptor. So the idea that it drives sleep is still weak.

  The DSIP-sleep link was never characterized further; no gene, precursor or receptor identified, so the sleep hypothesis remains weak.

  PMID 16539679 Journal of Neurochemistry
• Review of studies2026
  July 23-24, 2026: Meeting of the Pharmacy Compounding Advisory Committee

  The official FDA meeting record. The panel reviewed emideltide (DSIP) for the compounding list, for opioid withdrawal, insomnia and narcolepsy. It voted against recommending it.

  Emideltide (DSIP) reviewed for the 503A bulks list for opioid withdrawal, insomnia and narcolepsy; the panel voted against recommending it.

  Source FDA
• Review of studies2026
  FDA Evaluation of Emideltide-Related Bulk Drug Substances (Emideltide (Free Base) and Emideltide acetate)

  A regulator read the whole DSIP file at once and judged it not enough. FDA went back over every study in people and how each was set up. It concluded the evidence does not show DSIP works for long-term sleeplessness, narcolepsy or opioid withdrawal, even given into a vein as in those studies. For the shot under the skin that people buy, it found no study at all. One study described as having a dummy group turned out to have none.

  FDA's full inventory of the human DSIP record, with each study re-described and its design reassessed. Concludes there is insufficient evidence on effectiveness by the intravenous route for chronic insomnia, narcolepsy and opioid withdrawal, and no study at all supporting the nominated subcutaneous route. Notes that the positive insomnia studies lacked adequate control groups and that one described as placebo controlled had no placebo arm.

  Source U.S. Food and Drug Administration, CDER briefing document, 2026
• Review of studies2026
  Certain Bulk Drug Substances for Use in Compounding that May Present Significant Safety Risks

  This is a regulator's manufacturing and safety notice, not a result. FDA lists emideltide, its name for DSIP, for the risk that an injected peptide provokes the immune system and for the difficulty of proving what is in the vial. It states it has found no safety information for the route the product is sold for. The page says nothing about whether DSIP helps anyone sleep.

  Emideltide (DSIP) is listed for immunogenicity risk by certain routes and for peptide-related impurity and characterization complexities, with FDA stating it has not identified safety-related information for the proposed route of administration. Updated 22 April 2026.

  Source U.S. Food and Drug Administration
• Animal study1977
  The delta sleep inducing peptide (DSIP). Comparative properties of the original and synthetic nonapeptide

  In animals. The original team isolated this nine-part peptide and made it in the lab.

  Original isolation and synthesis of the nonapeptide by the Schoenenberger and Monnier group.

  PMID 862769 Experientia