Atosiban

Atosiban is a modified oxytocin peptide that binds the myometrial oxytocin receptor without activating it, preventing the Gq-phospholipase C-inositol trisphosphate cascade that releases calcium from the sarcoplasmic reticulum. It also antagonises vasopressin V1a receptors in the uterus, which contribute to contractility in labour, and reduces prostaglandin release from decidua. Because it acts only on receptors concentrated in reproductive tissue, it produces far fewer maternal cardiovascular effects than beta-agonists such as ritodrine, which is its main clinical advantage. What it does not clearly do is improve neonatal outcomes: it buys 48 hours, which is the window that matters for antenatal corticosteroids and in-utero transfer, and recent placebo-controlled work has questioned whether it delays delivery meaningfully at all.

Targets: Oxytocin receptor, Vasopressin V1a receptor, Myometrial calcium signalling

• · Total course is capped at 48 hours and about 330 mg. Repeat courses are possible but rarely used more than three times.

Work out your exact syringe units →

• very commonNausea— The single most common complaint.
• commonHeadache and dizziness
• commonInjection-site reaction
• commonHyperglycaemia
• uncommonTachycardia and hypotension— Markedly less than with beta-agonist tocolytics.

• Gestational age below 24 or above 33 completed weeks
• Ruptured membranes after 30 weeks
• Abnormal fetal heart rate or intrauterine growth restriction with abnormal Dopplers
• Antepartum haemorrhage requiring immediate delivery
• Eclampsia or severe pre-eclampsia requiring delivery
• Intrauterine fetal death or suspected intrauterine infection
• Placenta praevia or placental abruption

• conflictoxytocin-obstetric — Direct antagonist and agonist at the same receptor.
• conflictcarbetocin — Opposing pharmacology at the oxytocin receptor.

• · Uterine contraction frequency
• · Continuous or intermittent fetal heart rate monitoring
• · Maternal blood pressure, pulse and blood glucose
• · Signs of intrauterine infection

Prescription hospital drug in the EU and many other countries; never approved in the US, where nifedipine and indomethacin are used instead.

• Tractocile EMA summary of product characteristics (label)
• Cochrane review of oxytocin receptor antagonists for inhibiting preterm labour (review)
• APOSTEL 8 randomised placebo-controlled trial of atosiban in threatened preterm birth (trial)

Atosiban occupies the myometrial oxytocin receptor without producing the conformational change that couples it to Gq, so phospholipase C is not activated, inositol trisphosphate is not generated, and calcium is not released from the sarcoplasmic reticulum. It is a competitive antagonist, which means it can be outcompeted by rising endogenous oxytocin, and that is part of why the effect is modest rather than absolute. It also antagonises vasopressin V1a receptors in the uterus, which contribute independently to myometrial contractility during labour, and it reduces prostaglandin release from decidua, which removes a second contractile drive. The clinical advantage over beta-agonists such as ritodrine is entirely about receptor distribution: oxytocin and V1a receptors relevant to labour are concentrated in reproductive tissue, whereas beta-2 receptors are everywhere, so atosiban does not produce the maternal tachycardia, pulmonary oedema, hyperglycaemia and tremor that made beta-agonist tocolysis dangerous. What it does not do is fix the underlying process. Preterm labour is usually the endpoint of infection, inflammation, decidual haemorrhage or uterine overdistension, and blocking the final contractile signal does not address any of them. That is the honest interpretation of APOSTEL 8, which found no significant delay in delivery against placebo, and of the wider evidence that tocolysis buys the 48 hours needed for antenatal corticosteroids and in-utero transfer without improving neonatal outcomes on its own.

The commonest failure is conceptual: treating tocolysis as a treatment for preterm labour rather than as a means of buying time for steroids and transfer. If steroids are already given and the patient is already in a tertiary centre, the marginal value is close to nothing, and APOSTEL 8 makes that harder to argue with. The second is tocolysing in the face of chorioamnionitis, abruption or a non-reassuring fetal heart rate, all of which are contraindications, because delaying delivery in those situations harms the baby. The third is gestational age errors: the licensed window is 24 to 33 completed weeks and the drug has no role outside it. The fourth is repeated courses without evidence. The fifth is a geographic reality worth stating: atosiban has never been approved in the United States, where nifedipine and indomethacin are used instead, so the entire American literature and practice on this question runs on different drugs.

1. 6.75 mgStep 1, minute 0 — 6.75 mg as a single intravenous bolus over 1 minute.
2. 18 mgStep 2, hours 0-3 — 18 mg per hour by continuous infusion for 3 hours.
3. 6 mgStep 3, hours 3-48 — 6 mg per hour by continuous infusion for up to 45 hours. Total course capped at 48 hours and about 330 mg. The three steps are fixed and are not titrated against contraction frequency.
4. —Repeat courses — Repeat courses are possible but rarely used more than three times, and the evidence for repeating is weak.

• Oxytocin receptor on myometrium — Competitive antagonist; numeric Ki not resolved this session.

  Blocks Gq-phospholipase C coupling, preventing inositol trisphosphate generation and sarcoplasmic calcium release, so contractions weaken and space out. Competitive, so it can be overwhelmed by high endogenous oxytocin.

• Vasopressin V1a receptor on myometrium — Antagonist

  Removes a second Gq-mediated contractile drive that becomes relevant in labour.

• Decidual prostaglandin release

  Reduced, removing a further contractile and cervical-ripening stimulus.

• Oxytocin receptor on breast myoepithelium — Antagonist

  Theoretically inhibits milk ejection, but the drug is stopped long before lactation is established so it is clinically irrelevant.

• APOSTEL 8 Multicentre randomised placebo-controlled trial · 2025

  Atosiban versus placebo for threatened preterm birth. The trial did not demonstrate a significant benefit for atosiban over placebo, which has prompted genuine reconsideration of routine tocolysis in this population.

Contractions typically reduce within about 10 minutes of the bolus. The initial distribution half-life is around 18 minutes with a terminal half-life near 1.7 hours, so the effect is fully established within the first hour and fades within a few hours of stopping. The total course is 48 hours by design, matched to the time antenatal corticosteroids need to achieve full effect. There is no delayed benefit and nothing accumulates.

Atosiban is never the therapy; it is the 48-hour window in which the actual therapy happens. Antenatal corticosteroids, betamethasone or dexamethasone, are the intervention with a proven neonatal mortality and morbidity benefit, and magnesium sulphate for fetal neuroprotection below 32 weeks is the second. In-utero transfer to a hospital with a neonatal intensive care unit is the third. If none of those three are going to happen, the case for tocolysis largely evaporates. It is not combined with other tocolytics; nifedipine and indomethacin are alternatives rather than partners. It is directly antagonistic to oxytocin and carbetocin. Antibiotics are given for group B streptococcus prophylaxis and for confirmed infection, but not routinely to prolong pregnancy, since the ORACLE II follow-up showed harm from co-amoxiclav in intact membranes.

Against beta-agonists such as ritodrine and terbutaline: atosiban is clearly better tolerated, with far less maternal tachycardia, pulmonary oedema and metabolic disturbance, and that is the comparison that got it licensed. Against nifedipine: broadly similar tocolytic efficacy, oral administration, and a fraction of the cost, which is why much of the world uses nifedipine and why the US never needed atosiban. Nifedipine causes more maternal hypotension and flushing. Against indomethacin: effective but limited by fetal ductal constriction and oligohydramnios, so it is generally restricted to under 32 weeks and short courses. Against doing nothing, which is now the relevant comparison after APOSTEL 8: the honest answer is that the evidence for meaningful delay is weaker than a decade of routine practice assumed, and the 48 hours it was supposed to buy may not reliably materialise.

Not a monthly drug. Priced per 48-hour hospital course, and substantially more expensive than oral nifedipine, which is the main practical argument against it in health systems that have both. I did not source figures this session.

• ?Volume of distribution, clearance and protein binding figures here come from general European product literature rather than a label I resolved this session, so treat them as approximate.
• ?The three-letter structure is the standard published description; I did not resolve it from a primary source, so verified is false.
• ?APOSTEL 8's enrolment and follow-up duration were not confirmed against the paper, so both are null.
• ?The active metabolite's contribution to overall effect is described in product literature but I did not verify its potency relative to the parent this session.
• ?No cost figures were sourced.

• Atosiban versus placebo for threatened preterm birth (APOSTEL 8): a multicentre, randomised controlled trial van der Windt LI, et al., Lancet, 2025 · PMID 40049187

  The placebo-controlled trial that undercut the routine use of this drug. If you read one paper on atosiban, read this one rather than the older active-comparator studies.

• Study protocol for a randomised trial for atosiban versus placebo in threatened preterm birth: the APOSTEL 8 study Klumper J, et al., BMJ Open, 2019 · PMID 31772083

  The protocol, useful for understanding exactly what population was studied and what the endpoints were before the result was known.