Oxytocin is FDA-approved as a labour drug. The nasal spray it once approved is discontinued.

Why this note exists

Most compounds pepmg tracks are short of human data. Oxytocin has the opposite problem: a ClinicalTrials.gov search on August 18, 2026 returned 921 registered studies, 853 of them interventional.[17] There is no shortage of trials. There is a shortage of trials that found anything, and the gap between the two is where most oxytocin marketing lives.

In the index generated on August 15, 2026, 48 of 112 vendors carried an oxytocin listing — 71 listings, twenty-seventh of 83 compounds by vendor coverage.[18] Sixty-nine of the 71 state a size in milligrams: 10 mg on 26 listings, 5 mg on 20, 2 mg on 11, with a long tail up to 100 mg.[18] Three listings are a nasal form; the other 68 are the standard powder vial for reconstitution.[18] Not one names its contents in international units — which is the only unit the human literature uses.[18]

A note on what kind of evidence this is: every efficacy and safety figure below is human data, and each carries its design, route and participant count in the same sentence as the number. Where a result comes from a secondary or exploratory endpoint, it says so. Where a comparison is not randomized, it says so. pepmg does not convert a published international-unit dose into a milligram figure, and does not convert a labelled obstetric dose into a protocol for a research vial.

The approval

What FDA approved, in the label's words

The indication is obstetric from the first line to the last. Antepartum, Pitocin "is indicated for the initiation or improvement of uterine contractions, where this is desirable and considered suitable for reasons of fetal or maternal concern, in order to achieve vaginal delivery," in patients with a medical indication such as Rh problems, maternal diabetes, preeclampsia at or near term, or prematurely ruptured membranes, and as adjunctive therapy in incomplete or inevitable abortion.[1] Postpartum, it "is indicated to produce uterine contractions during the third stage of labor and to control postpartum bleeding or hemorrhage."[1] That is the entire indication.

The label leads its Indications section with a caution rather than a claim. Under the heading IMPORTANT NOTICE: "Elective induction of labor is defined as the initiation of labor in a pregnant individual who has no medical indications for induction. Since the available data are inadequate to evaluate the benefits-to-risks considerations, Pitocin is not indicated for elective induction of labor."[1] The contraindications that follow are a list of obstetric emergencies — cephalopelvic disproportion, transverse lies, fetal distress where delivery is not imminent, a uterus already hypertonic, placenta previa.[1] This is a drug administered by clinicians to a monitored patient, on an infusion pump.

The route is the part that does not survive the trip to a product page. Queried through FDA's Drugs@FDA data on August 18, 2026, the oxytocin active ingredient returns nine applications. Five carry a prescription marketing status — Pitocin plus four other oxytocin injections — and every one of the five is an injection.[4][2] The remaining four are recorded as discontinued.[4]

The nasal spray

FDA approved an intranasal oxytocin once. It was for breastfeeding, and it is gone.

This is the fact most likely to be missing from a conversation about oxytocin nasal sprays, and it cuts both ways. Drugs@FDA carries NDA 012285: brand name SYNTOCINON, active ingredient oxytocin at 40 USP units per millilitre, dosage form SOLUTION, route NASAL, marketing status Discontinued.[3] An intranasal oxytocin product did once clear FDA. It is not on the market.

What it was approved for is the second half. The company that licensed the product for the United States stated in December 2013 that "Syntocinon™ Nasal Spray was approved in the U.S. in 1960 to assist initial postpartum milk ejection" and "was discontinued by Novartis in 1997 for commercial reasons."[5] That is a company statement about its own licensed asset rather than an FDA document, and it is cited here as such; the regulatory record independently confirms the nasal route, the strength and the discontinued status.[3][5]

Two consequences follow, and neither is rhetorical. First, an oxytocin nasal spray has never been approved in the US for social behaviour, bonding, anxiety, autism or anything adjacent to how a person feels — the one approval it had was for milk ejection during breastfeeding.[3][5] Second, every intranasal trial in the rest of this note used an unapproved formulation prepared for research, which is worth remembering before treating any of these results as evidence about a specific product in a bottle.

The pooled evidence

Forty-two randomized trials, 1,922 people, and a confidence interval that crosses zero

The most useful single document published on oxytocin this year is a meta-analysis in Neuroscience & Biobehavioral Reviews, August 2026. It extracted data from 42 double-blind randomized controlled trials comparing symptoms after intranasal oxytocin against placebo in autism spectrum disorder, schizophrenia spectrum disorders, substance use disorders and other mental disorders.[6]

Figures as reported in the 2026 meta-analysis. The two removed studies were outliers with very large treatment effects in substance use disorders; removing them both reduced the pooled estimate and eliminated the heterogeneity.[6]

Three things in that table are worth reading slowly. The overall effect was small and did not reach significance, and the between-trial heterogeneity was high enough (I² = 77.4%) that the pooled number was not describing one consistent phenomenon.[6] Removing two outlying substance-use trials dropped the estimate to g = 0.05 and collapsed the heterogeneity to zero — meaning the remaining 40 trials agreed with each other closely, on approximately no effect.[6] And the one signal that survived was in schizophrenia spectrum disorders, at g = 0.12, which is a real result and a small one.[6]

The authors also found no significant moderation by dose, by number of administrations, by whether a psychosocial intervention was attached, or by participant age, and reported that studies with more female participants showed greater effects.[6] Their own summary of the field's methods is the sentence to keep: "trial methods seldom aligned with optimal protocols."[6] This is a call for better trials, not a verdict that the molecule is inert — and it is not a basis for expecting an effect either.

The largest single trial

SOARS-B: 290 children, 24 weeks, and a difference of two tenths of a point

The autism question got the best-resourced answer. SOARS-B was a 24-week, placebo-controlled phase 2 trial of intranasal oxytocin in children and adolescents aged 3 to 17 with autism spectrum disorder, funded by the National Institute of Child Health and Human Development and published in The New England Journal of Medicine in October 2021.[7][8] Of 355 screened, 290 were enrolled — 146 to oxytocin, 144 to placebo — randomized with stratification by age and verbal fluency, at a total target dose of 48 international units daily.[7]

The primary outcome was the least-squares mean change from baseline on the Aberrant Behavior Checklist modified Social Withdrawal subscale. It was −3.7 in the oxytocin group and −3.5 in the placebo group: a least-squares mean difference of −0.2, 95% confidence interval −1.5 to 1.0, P = 0.61.[7] Secondary outcomes, which included two further measures of social function and an abbreviated IQ measure, "generally did not differ between the trial groups," and the incidence and severity of adverse events were similar.[7] Both groups improved by roughly the same amount over six months.

Set against that is a 2025 dose-response meta-analysis in Frontiers in Psychiatry, which pooled 12 randomized trials in 498 patients with autism and found no significant effect on social impairments in its initial analysis, but reported a beneficial effect at a high dose of 48 IU per day and a dose-response trend suggesting higher doses might be more effective.[9] That is a genuine finding and it deserves stating alongside the null one — with the observation that 48 IU per day is precisely the dose SOARS-B targeted across 290 participants for six months, without separating from placebo.[7][9] The meta-analysis authors' own conclusion opens by conceding that "these findings show no consistent beneficial effects."[9]

Why anecdotes are unreliable here

Half a trial's worth of children improved before anyone got the drug

Oxytocin research has an unusually well-documented placebo problem, and one 2026 paper measures it directly. In the single-blind placebo lead-in phase of a multi-site randomized trial in autism, 87 children aged 3 to 12 received three weeks of placebo before any randomisation took place.[16] Forty-two of them — 48.3% — improved by 10 points or more on the Social Responsiveness Scale, 2nd Edition, a threshold the authors describe as a clinically significant degree of change.[16]

Caregiver guesses about which treatment a child was receiving did not significantly affect the placebo response (p = .534), and placebo response was associated with greater baseline symptom severity and higher cognitive ability.[16] This is a study about trial design, not about oxytocin's effect. But it establishes the number that matters for reading any uncontrolled account of oxytocin use: in this population, on a well-validated scale, nearly half of participants got clinically meaningfully better on nothing at all.

2026, indication by indication

What the last twelve months of human trials actually returned

Alcohol use disorder — the trial. A 12-week, double-blind, randomized multisite trial administered intranasal oxytocin at up to 70 IU per day, or placebo, to 100 individuals diagnosed with alcohol use disorder.[10] On the primary outcome, the weekly percentage of heavy drinking days over a 10-week maintenance phase, "no significant differences were observed."[10] Secondary drinking outcomes followed the same pattern, and there were no significant differences in AUD symptoms, alcohol-related consequences, craving, mood, sleep, pain or substance use; participants on oxytocin did score significantly lower on measures of anger and physical aggression at end of treatment.[10] The drug was well tolerated, with mild adverse events comparable across groups, the most common being hyposmia — reduced sense of smell — in both arms.[10]

Alcohol use disorder — the pooled answer. A September 2026 systematic review in Psychoneuroendocrinology ran frequentist, Bayesian and variability-based meta-analyses across six eligible randomized trials.[11] The pooled effect was not statistically significant (Hedges' g = 0.34, 95% CI −0.48 to 1.17, p = 0.47), the Bayesian analysis "confirmed the absence of a credible treatment effect" and provided moderate support for the null (BF₀₁ = 4.74), and the variability analysis found no evidence of increased outcome dispersion — which the authors read as arguing against a hidden subgroup of responders.[11]

Appetite and metabolism. A randomized, double-blind, placebo-controlled crossover trial gave 24 individuals with obesity a 4.5-hour intravenous oxytocin infusion reaching supraphysiological plasma concentrations of about 400 pg/mL, then measured food intake at an ad libitum meal.[14] Intake was 714 ± 382 kcal on oxytocin against 707 ± 385 kcal on saline, with no effect on appetite sensations, glucose, insulin, C-peptide, glucagon, GLP-1, GIP, cholecystokinin or gastric emptying.[14] Twenty-four participants is small, and the route was intravenous rather than intranasal — this is not a test of a nasal spray.[14]

Prader-Willi syndrome — where a signal did appear. A European double-blind randomized placebo-controlled study enrolled 52 infants with Prader-Willi syndrome, median age 2.2 months, assigning 4 IU per day of intranasal oxytocin or placebo for a four-week efficacy evaluation.[12] The primary endpoint did not separate: normalization rates on the Neonatal Oral-Motor Scale were similar between groups.[12] A secondary measure did — videofluoroscopy of swallowing produced a higher responder rate on oxytocin (53.3% vs 16.7%, p = 0.05) and a greater reduction in total score (least-squares mean difference −1.55, 95% CI −2.9 to −0.2, p = 0.03).[12] The paper also compares the treated cohort at about age three against an unexposed cohort of 24 children and reports better motor, adaptive and behavioural outcomes; that three-year comparison is between cohorts, not a randomized comparison, and the senior author discloses a patent licensed to a company in this field.[12]

Wound healing, and how a p-value of .048 should be read. A randomized, double-blind, placebo-controlled trial in JAMA Psychiatry gave 80 healthy heterosexual couples — 160 participants — seven days of twice-daily intranasal oxytocin or placebo after four suction-blister wounds were applied to the forearm, crossed with a structured partner-appreciation task.[13] Couples doing the task who received oxytocin showed improved wound healing (b = −0.125, P = .048) — and the paper states plainly that "these effects were not consistently robust in sensitivity analyses" (b = −0.090, P = .10).[13] The data were collected between 2011 and 2013 and analysed a decade later, and PubMed's record for the article lists a published erratum.[13]

Safety, on the narrow question that was asked. Sixty-one adults with obesity in a randomized trial received eight weeks of intranasal oxytocin at 24 IU four times daily, or placebo, with fasting oestradiol, testosterone and prolactin measured before and after.[15] No differences were found between groups (all p ≥ 0.140), nor in on-treatment menstrual cycle length (p = 0.234).[15] The authors describe this as "preliminary support for reproductive safety" in that population — a null result on three hormones over eight weeks in 61 people, which is what it is and no more.[15]

The market

The literature dosed in IU. The vial is sold in milligrams.

Every human study above states its dose in international units: 48 IU per day in SOARS-B, up to 70 IU per day in the alcohol trial, 24 IU four times daily in the obesity study, 4 IU per day in the Prader-Willi infants.[7][10][15][12] The FDA label works the same way — it doses labour induction as an infusion rate in milliunits per minute, and its one plain-amount instruction is "[t]en (10) units of Pitocin can be given after the delivery of the placenta."[1]

None of the 71 listings in pepmg's index names its contents in international units. Sixty-nine of them state milligrams.[18] An international unit is a measure of biological activity, not of mass, and the factor relating the two is specific to the preparation being measured. pepmg therefore does not convert an IU dose into milligrams or a milligram figure into IU, on this compound or any other — a published dose is reported in the unit its source published, or it is not reported.

The form mismatch runs the same direction. Sixty-eight of the 71 listings are a standard powder vial for reconstitution and injection; three are a nasal preparation.[18] Nearly all of the human evidence discussed above is intranasal.[6] The approved product is an injection for use in a delivery room.[1] There is no configuration of those three facts in which the research market's default product matches either the approved medicine or the studied intervention, and pepmg makes no claim about the contents of any vendor's vial.

Three things this note is not saying

First, it is not saying oxytocin does nothing. The 2026 meta-analysis found a small, statistically significant effect in schizophrenia spectrum disorders, and reported that studies with more female participants showed greater effects — a moderation finding worth testing properly rather than dismissing.[6] The Prader-Willi infant study found a real difference on a secondary swallowing measure.[12] Those are signals. They are also narrow, in specific clinical populations, at doses and routes stated in the papers.

Second, it is not saying the approval is thin. Pitocin is a decades-old obstetric medicine with a defined role in induction and in controlling postpartum haemorrhage, and its label's opening caution about elective induction is a sign of a regulator holding a line, not of a weak file.[1][2]

Third, it is not saying an oxytocin nasal spray is dangerous. Across these trials it was generally well tolerated, with mild adverse events, and the most common one reported in the alcohol trial — reduced sense of smell — occurred in the placebo arm too.[10] The honest problem is efficacy, not acute harm, and the honest gap is that no intranasal oxytocin product is approved in the United States, so nothing tells a buyer what is in the bottle.[3][4]

What it is saying is narrower: "FDA-approved" is a true statement about oxytocin, and it means an injection given in a delivery room to start contractions or stop bleeding.[1] The nasal route had exactly one approval, for milk let-down, and that product is discontinued.[3][5] Everything else is 42 pooled randomized trials whose confidence interval crosses zero.[6]

Questions people are asking