Atosiban for Threatened Preterm Birth: The APOSTEL 8 Trial and the Steroid Debate, Explained

Table of Contents

• Key Points
• Background: Why This Research Matters
• Study Methods: How the APOSTEL 8 Trial Was Conducted
• Key Findings: The Main Trial Results
• Expert Question 1: Do Steroids Still Help at 30 to 34 Weeks?
• Expert Question 2: Do Twin and Singleton Pregnancies Respond Differently?
• Expert Question 3: Could Lower or Fewer Steroid Doses Be Safer?
• Steroid Safety: What the Research Shows
• The Trial Authors' Reply
• Clinical Implications: What This Means for Patients
• Limitations: What These Letters Could Not Prove
• Recommendations for Patients and Families
• Frequently Asked Questions
• Source Information

Key Points

• Atosiban delayed birth beyond 48 hours in 78% versus 69% with placebo, but newborn outcomes were not improved.
• 98% of APOSTEL 8 participants received at least one steroid dose, making it hard to show added benefit from a full course.
• Experts question whether standard steroid dosing still helps at 30 to 34 weeks and call for reduced or individualized dose trials.
• Twin and singleton pregnancies may respond differently to atosiban and steroids, but the subgroup difference was not statistically reliable.
• A new trial called SNACS is testing whether a reduced steroid course is sufficient after 30 weeks; results are not yet available.

Background: Why This Research Matters

Preterm birth means a baby is born before 37 completed weeks of pregnancy. When labor threatens to start early, doctors face two urgent tasks. The first is to try to delay birth, often with a tocolytic (a drug that temporarily stops contractions). The second is to give the mother antenatal corticosteroids (steroid medicines given before birth) so the baby's lungs and other organs mature faster.

The drug at the center of this debate is atosiban, a tocolytic that works by blocking oxytocin, the hormone that drives labor contractions. Given as an intravenous infusion, it is used to buy time so that steroids can take effect.

Corticosteroids such as betamethasone are standard care for threatened preterm birth. The current dosing regimens were developed from trials conducted in the 1970s and 1980s. Those trials proved that steroids save lives for very early premature babies.

However, experts disagree about whether the same "one-size-fits-all" regimen still gives meaningful benefit at later gestational ages. Studies show that preterm-related morbidity and mortality (newborn illness and death linked to premature birth) decline steeply between 28 and 30 weeks of pregnancy. After 30 weeks, the balance between steroid benefits and steroid risks becomes less clear.

Study Methods: How the APOSTEL 8 Trial Was Conducted

APOSTEL 8 was a multicenter, randomized, placebo-controlled trial named after the Dutch research group that ran it. The full scientific citation is: van der Windt LI, Klumper J, Duijnhoven RG, et al., "Atosiban versus placebo for threatened preterm birth (APOSTEL 8)," published in The Lancet in 2025.

The trial enrolled 752 participants who arrived at hospitals with threatened preterm birth between 30 weeks and 0 days (30+0) and 33 weeks and 6 days (33+6) of gestation. Each participant was randomly assigned to receive either atosiban or an inactive placebo.

Researchers measured three main things. First, whether the pregnancy continued for more than 48 hours. Second, whether the participant completed the planned course of antenatal corticosteroids before delivery. Third, and most important, the health outcomes of the newborns, including illness and death.

Around one in five participants in the atosiban group (20%) had a twin pregnancy, compared with 15% in the placebo group. The study followed all babies for outcomes after birth.

Key Findings: The Main Trial Results

The trial produced three headline results, all of which are important for patients to understand:

• Atosiban delayed birth beyond 48 hours more often than placebo. In the atosiban group, 78% of pregnancies (about 78 in 100) continued for more than 48 hours. In the placebo group, 69% (about 69 in 100) did. This difference is expressed as a relative risk (RR) of 1.13, with a 95% confidence interval (CI) of 1.03 to 1.23. Because the entire confidence interval sits above 1.0, the delay is considered statistically significant. A confidence interval is the range within which the true effect most likely lies.
• More participants in the atosiban group completed their steroid course. The completion rate was 76% in the atosiban group versus 68% in the placebo group (RR 1.11; 95% CI 1.02–1.22). This makes sense: delaying labor gives the steroids time to work.
• Newborn outcomes were no better with atosiban. Despite the extra time gained and the higher rate of completed steroid courses, atosiban did not reduce neonatal morbidity and mortality (illness and death among newborns).

One number explains why the authors were puzzled: 98% of participants in both groups received at least one dose of antenatal corticosteroids, and nearly all received at least half a course. In other words, almost every baby in the study got some steroid protection before birth.

Expert Question 1: Do Steroids Still Help at 30 to 34 Weeks?

Two groups of obstetric researchers wrote letters to The Lancet after the APOSTEL 8 results appeared. The first group, led by Lola Loussert (Toulouse, France) and Paul Guerby, with colleagues from Lille, France, and Québec City, Canada, congratulated the trial team. Then they raised a pointed question.

If atosiban helped more women complete their full steroid course, the experts reasoned, then the atosiban group should have had healthier babies. It did not. The two groups had identical newborn outcomes. The most likely explanation, they argue, is that a full course of steroids may not add meaningful benefit beyond 30 weeks of pregnancy.

The letter cites a secondary analysis of a separate trial called BETADOSE, which compared a half course of corticosteroids with a full course. For babies born before 32 weeks, the study found no difference in survival without severe illness: 66.5% in the half-course group versus 65.8% in the full-course group. The risk difference was only 0.7%, with a 95% CI of −5.6 to +7.1. That confidence interval crosses zero, meaning the result is fully compatible with no difference at all.

A second group of experts, led by Ruben Ramirez Zegarra (Basel, Switzerland) with Beatrice Valentini and Tullio Ghi (Rome, Italy), made a similar argument with additional evidence. They pointed out that robust data on steroids given between 30+0 and 33+6 weeks are "sparse" (very limited). One large observational study of more than 13,000 infants born between 23 and 32 weeks found no survival or breathing benefit for infants born after 29 weeks. Other randomized trials give inconsistent answers: some show benefits between 30 and 34 weeks, and others show no benefit.

The experts stress that the current steroid regimen rests on trials from the 1970s and 1980s. They ask whether it still offers meaningful benefit at 30 to 34 weeks, or whether reduced or gestational age-tailored dosing (doses adjusted to the baby's exact week of development) could be both safer and equally effective.

Expert Question 2: Do Twin and Singleton Pregnancies Respond Differently?

The second concern from the Loussert group focuses on twin pregnancies. In APOSTEL 8, the effect of atosiban appeared to go in opposite directions depending on whether the pregnancy carried one baby or two:

• In twin pregnancies, atosiban was associated with a relative risk of 1.70 (95% CI 0.64–4.51). A relative risk above 1 means a higher chance of the outcome being measured. However, the confidence interval is very wide and includes 1.0, so this result is not statistically reliable on its own.
• In singleton pregnancies, atosiban was associated with a relative risk of 0.75 (95% CI 0.46–1.23). A relative risk below 1 means a lower chance of the outcome. Again, the confidence interval crosses 1.0.

Statistically, the formal test for whether these two effects truly differ — the interaction test — was not significant (p=0.14). A p-value above 0.05 means the difference between twins and singletons could easily be due to chance. But the experts caution that subgroup analyses like this often lack statistical power, meaning the study was not large enough to detect a real difference if one exists.

The two groups were also not strictly similar. Twin pregnancies made up 20% of the atosiban group but only 15% of the placebo group. That imbalance matters because twins have different pregnancy risks than singletons.

There is very little published evidence on atosiban or corticosteroids in twin pregnancies, the letter notes. Twins also handle steroid medicines differently at a chemical level. A 2002 study (Ballabh and colleagues) showed that the pharmacokinetics (how the body absorbs, distributes, and clears a drug) of betamethasone differ between twin and singleton pregnancies. Combining twin and singleton results into one analysis, the experts warn, may have introduced bias into the study's conclusions.

Expert Question 3: Could Lower or Fewer Steroid Doses Be Safer?

The Basel and Rome group drew attention to a clue inside the APOSTEL 8 data. Since nearly all babies received at least half a course of steroids and outcomes were good, the trial "hinted" that a partial course (fewer than the standard two doses) might be safe after 30 weeks.

One major trial has already tried this idea. A non-inferiority study involving 3,244 participants tested a single dose of 11.4 mg of betamethasone against the standard two-dose regimen. The goal of a non-inferiority trial is to prove that the simpler treatment is "not worse" than the standard one. In this case, the single dose failed to show non-inferiority, meaning researchers could not prove that one dose was as good as two. The experts note, however, that the trial had methodological limitations, especially in how outcomes were chosen and measured.

A new trial called SNACS (NCT05114096) is currently running. It is expected to provide additional evidence on whether a reduced course of antenatal corticosteroids is sufficient, particularly for the 30-to-34-week window.

Steroid Safety: What the Research Shows

The letters emphasize that concerns about steroid safety are not new. The risks of antenatal corticosteroids appear to be dose-dependent, meaning higher or more frequent doses bring greater risk. Repeated courses of steroids have been linked to increased short-term and long-term adverse outcomes in children.

The evidence cited includes several distinct findings:

• Animal studies found that steroid exposure caused delayed fetal brain growth and impaired cortical development (slower development of the brain's outer layer, which handles thinking, language, and sensation).
• Human cohort studies have reported higher rates of mental and behavioral disorders in children who were born preterm and exposed to antenatal corticosteroids before birth.
• A meta-analysis of over 1.25 million children found elevated risks of mental and behavioral disorders not only in preterm infants, but also in late-preterm and term infants (babies born closer to or at full term) who had been exposed to antenatal corticosteroids.

The researchers behind these letters are not arguing that steroids are useless. They are arguing that the risk-benefit calculation may shift after 30 weeks, when a baby's baseline risk of severe complications is already much lower than at 24 to 28 weeks.

The Trial Authors' Reply

Larissa I. van der Windt and colleagues, the APOSTEL 8 trial team, published a brief reply alongside the letters. They thanked both groups for their "thoughtful responses" to the study. They agreed that both correspondences raise important questions about the standard regimen of antenatal corticosteroids.

In scientific publishing, an authors' reply that concedes the importance of the critique is significant. It signals that even the trial team recognizes the need for further research on steroid dosing, timing, and the specific needs of the 30-to-34-week gestational window.

Clinical Implications: What This Means for Patients

For a pregnant person experiencing threatened preterm birth between 30 and 34 weeks, these letters do not change current medical practice overnight. The APOSTEL 8 trial's core finding stands: atosiban delays birth but does not, by itself, improve newborn outcomes. Steroids remain standard care because their benefits are well established for early preterm birth.

The deeper implication is that the medical community is now openly questioning whether the same steroid dose should be used for every pregnancy between 24 and 34 weeks. The correspondents call for a shift toward more personalized approaches — for example, reduced or individualized dosing for babies at 30+0 to 33+6 weeks, who start with a lower risk of severe complications than babies born at 24 to 28 weeks.

For patients, this means two things. First, the care you receive today is based on the best current evidence, which still supports steroids for threatened preterm birth. Second, the scientific debate is moving toward tailoring treatment to each pregnancy. That includes paying closer attention to whether the pregnancy involves twins, what the mother's blood levels of the drug look like, and exactly how many weeks pregnant the patient is.

The call for urgent research applies to both atosiban and corticosteroids. The experts specifically note there is an urgent need for well-designed trials evaluating alternative regimens of antenatal corticosteroids in the 30-to-34-week window, and for better data on twins.

Limitations: What These Letters Could Not Prove

It is important to recognize what this article is and is not. The APOSTEL 8 trial itself was a well-designed randomized controlled trial, the gold standard for medical evidence. The letters that follow, however, are expert opinions based on secondary analyses and reviews of earlier studies. They are hypothesis-generating, not definitive.

The subgroup finding in twin pregnancies, for example, had a non-significant interaction test (p=0.14). That means the apparent difference between twins and singletons could be a statistical fluke. The experts acknowledge this limitation directly, noting that such analyses often lack power. The baseline imbalance between groups (20% twins in the atosiban group versus 15% in the placebo group) further weakens any subgroup conclusion.

The observational studies cited for steroid risks cannot prove that steroids caused the mental and behavioral disorders seen in children. Confounding factors — such as the reason the mother needed steroids in the first place — could explain part of the association. The animal studies, while concerning, do not always translate directly to humans.

Finally, the single-dose betamethasone trial failed to prove non-inferiority, which means we do not currently know whether a reduced dose is safe enough to recommend. The ongoing SNACS trial may provide answers, but results are not yet available.

Recommendations for Patients and Families

If you or someone you love is facing threatened preterm birth, here is what you should know and ask:

1. Do not refuse corticosteroids based on this debate. The proven benefit of steroids for babies born before 34 weeks remains the medical standard. The question is about optimal dosing, not about whether steroids have value.
2. Ask your obstetrics team whether your situation is singleton or multifetal. The APOSTEL 8 analysis suggests that twin and singleton pregnancies may respond differently to both atosiban and corticosteroids, although this remains unproven. If you are carrying twins, ask specifically how the evidence applies to you.
3. Ask about gestational age. If you are at 30+0 to 33+6 weeks, the balance of benefits and risks may differ from someone at 24 to 28 weeks. Your care team can explain why they recommend a full, half, or tailored steroid course for your exact week of pregnancy.
4. Ask about tocolysis. Atosiban can delay birth for more than 48 hours, but the APOSTEL 8 trial found no improvement in newborn outcomes. Understand that the purpose of a tocolytic is to provide a window for steroids and for transfer to a hospital with a neonatal intensive care unit (NICU), if needed.
5. Ask about long-term follow-up. Researchers are still investigating the long-term effects of both atosiban and corticosteroids on child development. If your baby is born preterm, ask your pediatrician about appropriate developmental monitoring.
6. Stay informed about research updates. The SNACS trial (NCT05114096) and other studies now underway will help clarify whether reduced steroid dosing is safe after 30 weeks. This is an actively evolving area of obstetrics.

In the meantime, the central message for patients is this: the standard treatments for threatened preterm birth are being carefully re-examined by experts. That re-examination is a sign of a healthy, self-correcting medical field — and it is precisely why evidence-based care continues to improve for mothers and babies.

Frequently Asked Questions

What is atosiban and why is it used?

Atosiban is a labor-stopping drug given by intravenous infusion. It blocks oxytocin, the hormone that drives contractions. In threatened preterm birth, doctors use it to delay delivery for more than 48 hours. This delay buys time for steroid medicines to mature the baby's lungs before birth.

Does atosiban improve the baby's health?

In the APOSTEL 8 trial of 752 patients, atosiban delayed birth beyond 48 hours in more patients than placebo, but newborn illness and death rates were the same in both groups. Extra time did not lead to better newborn outcomes. Almost all participants received at least one dose of steroid shots.

What is the debate about steroids at 30 to 34 weeks?

Experts question whether the standard full course of steroid shots still helps babies born after 30 weeks. A full course may not add meaningful benefit beyond what at least half a course already provides. They call for research on reduced or individualized steroid dosing for this later preterm window.

Do twin and singleton pregnancies respond differently?

In APOSTEL 8, twin pregnancies were 20% of the atosiban group and 15% of the placebo group. Atosiban's effect on newborns appeared different for twins compared to singletons, but the difference could have been due to chance. The study was not large enough to give a reliable answer.

Could a lower steroid dose be safer?

Some experts think a partial course of steroids might be safe after 30 weeks because almost all APOSTEL 8 babies received at least half a course and had good outcomes. One large trial could not prove one dose was as good as two. A new trial called SNACS is studying reduced dosing now.

What are the risks of antenatal corticosteroids?

The risks appear to increase with higher or more frequent doses. Repeated steroid courses have been linked to delayed fetal brain growth in animals and higher rates of mental and behavioral disorders in children from a large meta-analysis. However, these studies cannot prove that steroids caused the disorders.

Should I refuse corticosteroids based on this debate?

No. The proven benefit of steroids for babies born before 34 weeks remains the medical standard. The debate is about optimal dosing, not about whether steroids have value. Your care team will recommend a course tailored to your exact week of pregnancy and whether you are carrying twins.

Should I get a second opinion about receiving corticosteroids and atosiban for threatened preterm birth between 30 and 34 weeks?

In APOSTEL 8, atosiban delayed birth beyond 48 hours more often than placebo (78% vs 69%), but newborn outcomes were no better. Nearly all participants received steroids, and experts have questioned whether the standard full steroid course adds meaningful benefit after 30 weeks, calling for research on reduced or individualized dosing. If you are 30 to 34 weeks pregnant or carrying twins, a second opinion can help clarify whether the recommended steroid and tocolytic regimen fits your specific situation. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

This patient-friendly article is based on peer-reviewed research published as correspondence in The Lancet, Volume 406, September 27, 2025, pages 1338–1339.

The original correspondence includes:

• Original trial: van der Windt LI, Klumper J, Duijnhoven RG, et al. Atosiban versus placebo for threatened preterm birth (APOSTEL 8): a multicentre, randomised controlled trial. Lancet 2025; 405: 1004–13.
• Letter 1: Loussert L, Garabedian C, Dupuis N, Bujold E, Guerby P. Atosiban for threatened preterm birth: the APOSTEL 8 trial. Departments of Obstetrics in Toulouse and Lille, France, and Québec City, Canada.
• Letter 2: Ramirez Zegarra R, Valentini B, Ghi T. Correspondence from the University Hospital Basel, Switzerland, and Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy.
• Authors' reply: van der Windt LI and colleagues, the APOSTEL 8 trial investigators.

The letters cite numerous supporting studies, including the BETADOSE trial analysis (Baud O, et al., Am J Obstet Gynecol 2024), a half-dose versus full-dose betamethasone trial (Schmitz T, et al., Lancet 2022), a cohort study of over 13,000 infants (Manktelow BN, et al., 2010), a Cochrane systematic review (McGoldrick E, et al., 2020), long-term outcome research after repeat steroid doses (Wapner RJ, et al., N Engl J Med