Oxytocin

What oxytocin is and how it works

Oxytocin is a small peptide hormone made in the hypothalamus and released from the posterior pituitary. The approved US labeling for the synthetic injectable describes it as a nonapeptide and notes that it differs from vasopressin at only two amino acid positions — which is the labeling's own explanation for why oxytocin in pure form carries inherent pressor and antidiuretic properties that can become apparent when large amounts are administered. The same labeling states a plasma half-life of about 1 to 6 minutes. Those are descriptions of a regulated document, and they are stated here without a citation because a labeling statement is a regulatory fact rather than a research finding.

Peripherally, oxytocin acts at receptors in the myometrium whose concentration rises steeply through pregnancy, and that is the mechanism the approved drug is sold on. Centrally, oxytocin is also released within the brain, and it is that second role — as a signalling molecule involved in social behaviour, affiliation and stress — that generated the research literature almost every reader of this page arrived for.

The two halves of oxytocin are separated by a route, and the separation is the single most important thing on this page. The approved drug is given by intravenous infusion or intramuscular injection in a hospital. The social, bonding, trust, anxiety and sexual-function research is almost entirely intranasal, on the premise that a spray can deliver peptide along olfactory and trigeminal pathways and bypass the blood-brain barrier. Whether that premise holds is contested rather than settled: one influential review argues that very little of the large amounts applied intranasally appears to reach the cerebrospinal fluid, while peripheral concentrations are pushed to supraphysiologic levels with likely effects on the gastrointestinal tract, heart and reproductive tract.^[7] A randomized study that sampled both compartments did find significant rises in plasma and in cerebrospinal fluid after a 24 IU intranasal dose, in 11 men — but plasma peaked at 15 minutes while cerebrospinal fluid took up to 75 minutes, and the two did not correlate at all.^[8]

What the research actually shows

The approved use, by the approved route, has the strongest evidence here — and it is not evidence about anything a reader arrives asking about. A 2025 Cochrane network meta-analysis pooled 122 randomized trials in 121,931 women across 48 countries and names 10 IU intramuscularly or intravenously as the current World Health Organization standard for preventing postpartum haemorrhage. Ergometrine plus oxytocin reduced haemorrhage of at least 500 mL compared with oxytocin alone, risk ratio 0.76, 95% CI 0.64 to 0.90, high-certainty evidence — and the same review reports that all agents except carbetocin were associated with increased risks of some side effects compared with oxytocin, which is the favourable half and belongs in the same paragraph.^[12] For induction of labour, a Cochrane review of nine trials in 2,391 women found no significant difference between high-dose and low-dose intravenous regimens in vaginal delivery within 24 hours or caesarean section rate, while high-dose regimens significantly increased hyperstimulation without specified fetal heart rate changes, risk ratio 1.86, 95% CI 1.55 to 2.25. Those labels are the review's own selection criterion rather than a protocol anyone was prescribed: it defined a high-dose arm as at least 100 mU of oxytocin in the first 40 minutes with increments delivering at least 600 mU in the first two hours, a low-dose arm as anything below both figures, and its authors note that the review's conclusions are specific to the definitions it used.^[13]

The intranasal literature aimed at social outcomes is large and mostly null. The largest trial in the field, SOARS-B, randomized 290 children and adolescents aged 3 to 17 to intranasal oxytocin at a total target dose of 48 international units daily or placebo for 24 weeks; it MISSED its primary endpoint, with a least-squares mean difference on the ABC modified Social Withdrawal subscale of -0.2, 95% CI -1.5 to 1.0, P = 0.61, and secondary outcomes that generally did not differ.^[1] A Japanese multicentre trial in 106 enrolled adults also MISSED its primary endpoint — ADOS reciprocity fell in the oxytocin arm, but it fell in the placebo arm too, leaving a between-group effect size of -0.08, 95% CI -0.46 to 0.31, P = .69. Its authors report two secondary endpoints that did move: repetitive behaviour, effect size 0.44, P = .026, and duration of gaze fixation on socially relevant regions, effect size 0.55, P = .018. They conclude they cannot recommend continuous intranasal oxytocin treatment alone at that dose and duration, and both halves of that trial are reported here.^[2]

Trust is where the popular story started, and a registered replication is where it stalled. A double-blind, placebo-controlled replication powered above 95% found no effect of oxytocin on trusting behaviour in the minimal-social- contact condition the original result depended on. The same paper's exploratory post hoc analysis suggested oxytocin might increase trust among people with a low disposition to trust in a no-social-contact condition — the authors label that finding exploratory and post hoc and say it requires confirmation, and it is reported that way here rather than promoted.^[3] Across psychiatry generally, a 2026 meta-analysis of 42 double-blind randomized trials, N = 1,922, found a small non-significant pooled effect, g = 0.17, 95% CI -0.05 to 0.38, with I-squared of 77.4%; removing two outlier substance-use trials reduced it to g = 0.05, 95% CI -0.03 to 0.12, and eliminated the heterogeneity entirely. It also found a small significant effect in schizophrenia-spectrum disorders, g = 0.12, 95% CI 0.01 to 0.23.^[6]

On sexual function specifically, the evidence is thin and unkind. A 22-week randomized, double-blind, placebo-controlled crossover trial in 30 pre- and postmenopausal women with sexual dysfunction, using 32 IU intranasally on demand before intercourse, found that oxytocin and placebo both improved sexual function, with no statistically significant treatment, sequence or interaction effect — and placebo's numerical improvement was the larger on several of the instruments.^[10] In 29 heterosexual couples given a single 24 IU intranasal dose in a naturalistic home setting, the authors state plainly that oxytocin did not alter classical parameters of sexual function such as drive, arousal, erection or lubrication, while orgasm intensity and contentment after intercourse did rise, with effect sizes the authors describe as small to moderate.^[11]

Where the evidence is weak

The field's own methodologists are its harshest critics. A systematic review of published studies of interactive intranasal oxytocin effects on psychosocial outcomes in healthy people concluded that tested effects were highly heterogeneous, that replication was rarely attempted and largely unsuccessful when it was, that statistical significance was unrelated to sample size, that power was critically low and unrelated to the rate of significant results, and that research practices were characteristic of an exploratory approach.^[5] That is a description of a literature in which a positive finding is hard to interpret, and it applies to the studies on this page as much as to any others.

"Not shown to work" is not the same claim as "shown not to work", and the distinction has been quantified. An equivalence-testing reanalysis of data from a meta-analysis synthesizing 32 intranasal oxytocin studies found that 26.1% of the non-significant meta-analytic effects were indicative of data insensitivity rather than statistical equivalence; in a separate set of 34 non-significant effects from previously unpublished laboratory data, 73.5% were due to data insensitivity.^[4] A null trial in this field often means the data could not tell.

The delivery question is unresolved in both directions. The randomized cerebrospinal fluid study above did detect a rise, in 11 men, with no plasma-to-cerebrospinal-fluid correlation.^[8] A first-in-human PET study in 6 healthy volunteers found high uptake in the nasal cavity, detectable but low and variable tracer distribution to trigeminal and brain regions, no clear dose-dependency, and concluded that the tracer is not presently well suited for central nervous system receptor imaging by the intranasal route.^[9] Neither result licenses the confident claim, common in consumer writing, that a nasal spray reliably delivers a behaviourally meaningful amount to the brain.

The specific claim most readers arrive with is close to unstudied. A PubMed search for "(intranasal oxytocin) AND (sexual desire OR libido) AND healthy AND randomized controlled trial[pt]" on 1 September 2026 returned exactly 1 record — the 29-couple study described above, in which desire and arousal did not change. A PubMed search for "intranasal oxytocin AND (erectile dysfunction) AND randomized controlled trial[pt]" on 1 September 2026 returned 0 records. Duration is the other gap: a PubMed search for "intranasal oxytocin AND randomized controlled trial[pt] AND (12 months OR 1 year OR 52 weeks OR long-term safety)" on 1 September 2026 returned 17 records, none of which administered intranasal oxytocin in a randomized trial for a year or longer. The longest randomized exposure identified anywhere for this page is the 24 weeks of SOARS-B.^[1]

Finally, the harms that are best documented belong to the route almost nobody reading this page is using. Uterine hyperstimulation on high-dose intravenous infusion is a measured effect with an unmeasured consequence,^[13] and hyponatraemia in labour is common enough to be worth studying — though the prospective study that looked hardest found it correlated with total fluid volume, P less than 0.001, and not with oxytocin administration.^[14] Reading those findings across onto a nasal spray, in either direction, is not supported.

Legal and regulatory status

Oxytocin is FDA-approved, and the approval is narrow in three dimensions at once: indication, route and population. How narrow depends on reading the whole record rather than one field of it. A Drugs@FDA query on the openfda.generic_name field for "oxytocin" on 1 September 2026 returned three applications; a query of the same database on the same date against the products.active_ingredients.name field for "OXYTOCIN" returned nine, and the three are a strict subset of the nine. The six that the narrower field misses are not absent from the record — the indexed generic-name block is unpopulated on those applications, which is a property of the index rather than a regulatory fact. A same-session semaglutide control returned records on the generic-name path, so neither search was malformed; what separates them is the field, not the query.

Read against the ingredient-name field, five oxytocin applications are currently marked Prescription: NDA 018261 (PITOCIN, Par Health), NDA 018248 (Fresenius Kabi), NDA 018243 (Hikma), ANDA 200219 (Hikma Farmaceutica) and ANDA 091676 (Sagent). Four are marked Discontinued: ANDA 077453, NDA 018245 (SYNTOCINON injection, Novartis), NDA 019185 (oxytocin in dextrose, Abbott) and NDA 012285 (SYNTOCINON nasal solution). Every product under all five current applications is an injectable given by injection. The only nasal product anywhere in the record is discontinued, and there is no approved oral product at all.

The approved indications, in the labeling's own terms, are antepartum initiation or improvement of uterine contractions where that is desirable for fetal or maternal reasons in order to achieve vaginal delivery — covering medically indicated induction, stimulation of labor in uterine inertia, and adjunctive therapy in incomplete or inevitable abortion — and postpartum production of uterine contractions during the third stage of labor and control of postpartum bleeding or hemorrhage. Nothing about bonding, trust, social cognition, autism, anxiety, depression, desire or orgasm appears anywhere in that list.

A boxed warning sits on this drug, and it is not a brand-only feature. A query of the openFDA SPL labeling endpoint on the openfda.substance_name field for "OXYTOCIN" on 1 September 2026 returned 26 labels, ten of which belong to approved drug applications. The coded boxed_warning field is populated on five of those ten: the four Pitocin labels under NDA 018261, and the West-Ward/Hikma generic label under ANDA 200219. It is not populated on the other five, all of them NDA 018248 labels from Fresenius Kabi and its repackagers. But every one of the ten carries the IMPORTANT NOTICE text inside its INDICATIONS AND USAGE section, including the five that leave the boxed-warning field empty — so what varies across the approved labels is field coding and section placement, not whether the notice is in the label. Saying the generics carry no boxed warning would be wrong twice over: one generic populates the field, and the rest carry the notice anyway.

The two populated wordings differ. The Pitocin boxed warning, headed IMPORTANT NOTICE, reads: "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." The ANDA 200219 boxed warning, under the same heading, reads: "Oxytocin Injection, USP (synthetic) is indicated for the medical rather than the elective induction of labor. Available data and information are inadequate to define the benefits to risks considerations in the use of the drug product for elective induction. Elective induction of labor is defined as the initiation of labor for convenience in an individual with a term pregnancy who is free of medical indications."

The WARNINGS section adds that the drug, when given for induction of labor or augmentation of uterine activity, should be administered only by the intravenous route and with adequate medical supervision in a hospital. Its OVERDOSAGE section attributes water intoxication with convulsions to the drug's inherent antidiuretic effect when large amounts are infused for long periods, and its labeled dosing directions cap total amount at 30 international units in a 12-hour period for one indication, for that reason. All of that is labeling text and is stated here uncited.

No intranasal oxytocin product is currently approved in the United States. Route-filtered queries on 1 September 2026 against Drugs@FDA, the NDC directory and the SPL labeling endpoint each returned HTTP 404 NOT_FOUND for oxytocin by the nasal route, while an identically shaped desmopressin acetate nasal control returned records from all three — so the null is a real absence rather than a broken query. One nasal product was approved historically: NDA 012285, SYNTOCINON nasal solution, 40 USP units per mL, which Drugs@FDA carries as Discontinued, and which a Federal Register notice of 7 August 1997 lists among applications whose approvals were withdrawn, effective 8 September 1997, after their holders told the agency the products were no longer marketed and requested withdrawal. The historical nasal label's indication wording could not be retrieved in this session and is therefore not quoted anywhere on this page.

The NDC directory makes the badge-scope point in one line. Queried on 1 September 2026 on the generic_name field for "oxytocin", it returned 58 listing records — and the active_ingredients.name field for "OXYTOCIN" on the same date returned the identical 58, so on this endpoint the two paths agree rather than diverging as they do in Drugs@FDA. The 58 break down as 27 registered as bulk ingredient, 16 as unapproved homeopathic human OTC drugs taken orally, 3 as bulk ingredient for human prescription compounding, 1 as a drug for further processing, and 11 as approved prescription drugs under an NDA or ANDA — all eleven of which are injections. Under a fifth of the directory's oxytocin rows are approved drugs, and not one of them is a spray or a pill. On the compounding lists, oxytocin appears on none of FDA's 503A nomination categories (checked with three in-document positive controls), appears on the 503B Category 1 "Bulk Drug Substances Under Evaluation" list carrying the marker whose own footnote reads "Designates bulk drug substances that are components of FDA approved drugs" — a statement about the approved injection and not an endorsement of any compounded nasal preparation — and appears on neither table of FDA's bulk drug substances safety risks page, whose content is current as of 22 April 2026.

Anti-doping. A full-text search of the 2026 WADA International Standard Prohibited List, in force 1 January 2026, and of the 2026 Monitoring Program — both retrieved on 1 September 2026 from the WADA-hosted PDF assets linked by USADA's prohibited-list page — returned no occurrence of oxytocin in either document. Section S0, the non-approved-substances catch-all, reaches only a pharmacological substance "with no current approval by any governmental regulatory health authority for human therapeutic use"; oxytocin has a current US approval, so S0's own trigger condition does not reach it, and no substance section on the 2026 List names it. The only 2026 provision that bears on its approved route at all is M2.2, a prohibited-method rule about intravenous infusions and injections above a stated total volume per 12-hour period that excepts those legitimately received in the course of hospital treatment — a rule about a route rather than about this molecule. Absence from the List is never a statement that a substance is permitted, and the List is reissued annually.

Oxytocin is prescription-only, and legitimate access runs through a licensed prescriber in a clinical setting. Peptide Health Lab does not sell peptides, does not prescribe, and does not tell anyone where to obtain anything. What the published pharmacology adds to that regulatory boundary is a reason the boundary matters: the same review that questions central delivery notes that peripheral concentrations after intranasal dosing are raised to supraphysiologic levels, with likely effects on the gastrointestinal tract, heart and reproductive tract.^[7]

A note on the evidence badge above. This page is graded human_rct because randomized controlled trials in humans exist for the claim readers typically arrive with — the intranasal, social and sexual one — and exist in quantity. The grade records what kind of evidence has been generated. It does not record that the evidence was favourable, and mostly it was not: the pooled estimates are null and the largest trials missed their primary endpoints.

Questions to bring to a provider

A provider is the right person to ask what an approved indication actually covers, and oxytocin is an unusually clean example of the gap between an approval and a reputation. Useful questions include: what is the approved product, by what route, and for whom — and does anything I have read about apply to that product at all? If a preparation is not the approved injection, what is known about its identity and its behaviour, given that the published data on this page were generated with research-grade material under trial conditions?^[7]

Where a specific outcome is the goal, it is worth asking what the trials in that outcome actually found rather than what the category is said to do — the largest autism trial, the largest trust replication and the two sexual-function trials each reported a null primary result, and a clinician can put that beside whatever else is on the table.^[1][10] Questions about interactions and existing conditions belong in the same conversation: a history of hyponatraemia, cardiac disease, or any condition where fluid balance is already fragile is a reason to raise this with a clinician before considering anything, given that the best-documented systemic effects of oxytocin come from the parenteral route and include effects on fluid balance.^[14] Finally, it is reasonable to ask what would count as evidence of benefit and over what timeframe. The randomized intranasal trials described above ran 6, 12, 22 and 24 weeks; the 12-week one was a double-blind, placebo-controlled multisite trial in 100 adults given up to 70 IU per day, which found no significant difference on its primary drinking outcome and reported adverse events as mild and comparable across groups.^[15] What a longer horizon would require, and whether it has been studied at all, is a fair question to put to a clinician alongside the durations that have been.