The current guideline recommends that in case of a threatened preterm birth < 34+0 weeks where delivery is expected within 2-10 days, tocolysis should be applied to allow prenatal corticosteroids to take effect. While the benefits of corticosteroids has been extensively demonstrated, there is no evidence supporting the added value of tocolysis.

Atosiban

Nifedipine

Description of studies

Hawkins (2021) performed a randomized, double-blind, placebo-controlled trial to determine the effect of nifedipine in preventing preterm birth. Pregnant women between 28+0 and 33+6 weeks of gestation with a singleton pregnancy, uterine activity, intact membranes and cervical dilation of 2 to 4 cm were included. Exclusion criteria were women with pregnancy-associated hypertension, chronic hypertension that required medication, placental complications such as suspected abruption and placenta previa, chorioamnionitis, fetal death, fetal malformations, and any condition considered a contraindication to the use of tocolytic agents. Besides, women enrolled in a concurrent study involving progesterone therapy for a history of preterm birth were also excluded. In total, 46 women received nifedipine (20 mg orally followed by repeat dose after 90 minutes if contractions persisted; continued every 4 hours for 48 hours) and 42 women received placebo. Groups were comparable at baseline. The outcome of interest was perinatal mortality.

Nijman (2016) performed a double-blinded multicenter randomized placebo-controlled trial to assess the effect of nifedipine in women with PPROM without contractions on perinatal outcomes and prolongation of pregnancy. Pregnant women (≥18 years) between 24+0 and 33+6 weeks of gestation with ruptured membranes without signs of active labor were included. Exclusion criteria were ≥3 contractions per 10 minutes, previous treatment with tocolysis in the last 7 days (tocolysis for <6 h for transportation was allowed), symptoms justifying start of tocolysis, ruptured membranes ≥72 h, signs of chorioamnionitis or intra uterine infection, signs of fetal distress, fetal major congenital anomaly, contraindication for the use of nifedipine and maternal disease as reason for delivery (such as hypertension, HELLP syndrome or preeclampsia). The study was ended earlier due to slow recruitment. In total, 25 women received nifedipine (20 mg orally every 6 hours until start of active labor for a maximum of 18 days or until 34+0 weeks gestation) and 25 women received placebo. Groups were comparable at baseline, except for gestational age at study entry. Women who received nifedipine had a higher gestational age than women who received placebo (median gestational age of 29.9 weeks as compared to 27.0 weeks, respectively). Outcomes of interest were a composite of adverse perinatal outcome (perinatal death, bronchopulmonary dysplasia, periventricular leukomalacia > grade 1, intraventricular haemorrhage > grade 2, necrotizing enterocolitis > stage 1 and culture proven sepsis), perinatal mortality, periventricular leukomalacia, intraventricular haemorrhage, necrotizing enterocolitis and sepsis.

Romero (2000) performed a multicenter, double-blind, placebo-controlled trial to determine the efficacy and safety of atosiban in the treatment of preterm labor. Women between 20 and 33+6 weeks gestation, in preterm labor with intact membranes, a cervical dilatation of ≤3 cm, live fetus(es) and who provided written informed consent were included. Since only a small percent of the participants had a gestational age below 26 weeks (10% for atosiban and 5% for placebo), it was decided to include this study in the literature analysis. Exclusion criteria were fetal or placental abnormalities, fetal distress, suspected chorioamnionitis, maternal indications for delivery, urinary tract infection and overt clinical manifestations of substance abuse. In total, 246 women received atosiban (bolus with 6.75 mg atosiban over 1 minute followed by infusion of 300 mg/min for 3 hours and 100 mg/min for up to 45 hours) and 254 women received placebo. Groups were comparable at baseline, except for gestational age at admission. Women in the atosiban group had a lower gestational age than women in the placebo group (mean gestational age of 31.0 weeks as compared to 30.3 weeks, respectively). Outcomes of interest were perinatal mortality, respiratory distress syndrome, intraventricular haemorrhage and necrotizing enterocolitis.

Van der Windt (2025) performed an international, double-blind, randomized, placebo-controlled, superiority trial (APOSTEL-8) to evaluate whether tocolysis with atosiban is superior to a placebo in reducing neonatal morbidity and mortality. Women aged 18 years or older with a singleton or twin pregnancy and a threatened preterm birth between 30+0 and 33+6 weeks of gestation were included. Exclusion criteria were the inability to provide informed consent, contraindications to tocolysis, prior corticosteroid treatment for threatened preterm birth, signs of fetal distress, evident intrauterine infection (such as fever ≥38°C, fetal tachycardia >160 bpm, uterine pain, purulent fluid, and elevated plasma C-reactive protein or leukocytosis), and known fetal chromosomal or severe structural congenital abnormalities. In total, 377 women received atosiban and 377 women received placebo (0.9% saline). Treatment with atosiban consisted of a bolus injection of 6.75 mg/0.9 mL in 1 minute, followed by 3 hours continuous infusion of 18 mg/h and continuous infusion of 6 mg/h for 45 hours. Groups were comparable at baseline. Outcomes of interest were a composite outcome of adverse neonatal outcomes (including perinatal mortality (stillbirth and death until 28 days postpartum), bronchopulmonary dysplasia, periventricular leukomalacia above grade 1, intraventricular haemorrhage above grade 2, necrotizing enterocolitis above Bell’s stage 1, retinopathy of prematurity above grade 2 or needing laser therapy, and culture-proven sepsis), perinatal mortality, respiratory distress syndrome, neonatal sepsis, necrotizing enterocolitis, intraventricular haemorrhage, periventricular leukomalacia and gestational age at birth.

Vis (2014) performed a randomized controlled trial to assess whether tocolysis with nifedipine can be omitted in women with symptoms of preterm labor, a shortened cervix and negative fetal fibronectin test. Women with a gestational age between 24 and 34 weeks with symptoms of preterm labor, intact membranes, a negative fetal fibronectin test and a cervical length between 10 and 30 mm were included. Exclusion criteria were placenta previa, contraindications to tocolysis such as lethal congenital abnormality, suspected intrauterine infection, ruptured membranes or suspected fetal distress. Besides, women who had been treated with tocolytics in the previous 7 days (except for transport) and contraindications for nifedipine such as maternal cardiovascular disease or hypertension were excluded. In total, 37 women were allocated to oral nifedipine (20 mg four times per day slow release for 48 hours) and 36 women were allocated to placebo. Three women in the nifedipine group and 1 woman in the placebo group did not complete the full 48 hours of study medication. Groups were comparable at baseline except for twin pregnancies, previous preterm birth and referral from peripheral hospitals. Women who received nifedipine had more twin pregnancies (22% versus 11%), more previous preterm delivery (22% versus 11%) and more referral from peripheral hospitals (30% versus 11%) as compared to the placebo group. Outcomes of interest were composite outcome of fetal deaths, neonatal sepsis, bronchopulmonary dysplasia, necrotizing enterocolitis, intraventricular haemorrhage grade ³3 and periventricular leukomalacia grade ³2, perinatal mortality, neonatal sepsis, respiratory distress syndrome, necrotizing enterocolitis, intraventricular haemorrhage, periventricular leukomalacia and gestational age at birth.

Results

Atosiban

1. Perinatal mortality

Two studies reported perinatal mortality (Romero, 2000; Van der Windt, 2025) (Figure 1). These data were not pooled as the agreement is to start pooling when including at least three studies.

Romero (2000) reported that 11 of the 288 infants (3.8%) whose mother received atosiban died before delivery or up to 28 days after birth as compared to 5 of the 295 infants (1.7%) whose mother received placebo (RR=2.25, 95%CI 0.79 to 6.40).

Van der Windt (2025) reported that 2 of the 449 infants (0.4%) whose mother received atosiban died (stillbirth and death until 28 days postpartum) as compared to 4 of the 435 infants (0.9%) whose mother received placebo (RR=0.48, 95%CI 0.09 to 2.63).

_{Figure 1. Perinatal mortality for the comparison between atosiban and placebo}

2. Composite outcome of neonatal morbidity and mortality

Van der Windt (2025) reported a composite outcome of adverse neonatal outcomes consisting of perinatal mortality (stillbirth and death until 28 days postpartum), bronchopulmonary dysplasia, periventricular leukomalacia above grade 1, intraventricular haemorrhage above grade 2, necrotizing enterocolitis above Bell’s stage 1, retinopathy of prematurity above grade 2 or needing laser therapy, and culture-proven sepsis. Thirty-seven of the 449 infants (8.2%) whose mother received atosiban was positive for this composite outcome as compared to 40 of the 435 infants (9.2%) whose mother received placebo (RR=0.90, 95%CI 0.58 to 1.37).

3. Respiratory distress syndrome (RDS)

Two studies reported respiratory distress syndrome (Romero, 2000; Van der Windt, 2025) (Figure 2). These data were not pooled as the agreement is to start pooling when including at least three studies.

Romero (2000) reported that 64 of the 283 infants (23%) whose mother received atosiban had RDS as compared to 54 of the 292 infants (18%) whose mother received placebo (RR=1.22, 95%CI 0.89 to 1.69).

Van der Windt (2025) reported that 61 of the 449 infants (13.6%) whose mother received atosiban had infant RDS as compared to 71 of the 435 infants (16.3%) whose mother received placebo (RR=0.83, 95%CI 0.61 to 1.14).

_{Figure 2. Respiratory distress syndrome for the comparison between atosiban and placebo}

4. Neonatal sepsis

Van der Windt (2025) reported that 26 of the 449 infants (5.8%) whose mother received atosiban had culture proven neonatal sepsis as compared to 25 of the 435 infants (5.7%) whose mother received placebo (RR=1.01, 95%CI 0.59 to 1.72).

5. Necrotizing enterocolitis (NEC)

Two studies reported necrotizing enterocolitis (Romero, 2000; Van der Windt, 2025) (Figure 3). These data were not pooled as the agreement is to start pooling when including at least three studies.

Romero (2000) reported that 1 of the 283 infants (0.4%) whose mother received atosiban had NEC as compared to 5 of the 292 infants (2%) whose mother received placebo (RR=0.21, 95%CI 0.02 to 1.76).

Van der Windt (2025) reported that 2 of the 449 infants (0.4%) whose mother received atosiban had necrotizing enterocolitis >stage 1 as compared to 6 of the 435 infants (1.4%) whose mother received placebo (RR=0.32, 95%CI 0.07 to 1.59).

_{Figure 3. Necrotizing enterocolitis for the comparison between atosiban and placebo}

6. Intraventricular haemorrhage (IVH)

Two studies reported intraventricular haemorrhage (Romero, 2000; Van der Windt, 2025) (Figure 4). These data were not pooled as the agreement is to start pooling when including at least three studies.

Romero (2000) reported that 16 of the 243 infants (7%) whose mother received atosiban had IVH as compared to 19 of the 246 infants (8%) whose mother received placebo (RR=0.88, 95%CI 0.45 to 1.62).

Van der Windt (2025) reported that 3 of the 449 infants (0.7%) whose mother received atosiban had intraventricular haemorrhage >grade 2 as compared to 6 of the 435 infants (1.4%) whose mother received placebo (RR=0.48, 95%CI 0.12 to 1.92).

_{Figure 4. Intraventricular haemorrhage for the comparison between atosiban and placebo}

7. Periventricular leukomalacia (PVL)

Van der Windt (2025) reported that 8 of the 449 infants (1.8%) whose mother received atosiban had periventricular leukomalacia >grade 1 as compared to 5 of the 435 infants (1.1%) whose mother received placebo (RR=1.55, 95%CI 0.51 to 4.70).

8. Gestational age at birth

Van der Windt (2025) reported the gestational age at birth (table 1).

Table 1. Gestational age at birth for the comparison between atosiban and placebo

9. Survival without neurodevelopmental impairment

Not reported.

Level of evidence of the literature

According to GRADE, the level of evidence of randomized controlled trials start high.

The level of evidence regarding the outcome measure perinatal mortality was downgraded to very low because of differences in the direction of the effect and differences in the gestational age of the included patient population (-1, inconsistency) and the low number of events (-2, imprecision).

The level of evidence regarding the outcome measure composite outcome of neonatal morbidity and mortality was downgraded to low because of the 95% confidence interval crossed both lines of no (clinically relevant) effect (-2, imprecision).

The level of evidence regarding the outcome measure respiratory distress syndrome was downgraded to low because of differences in the direction of the effect and differences in the gestational age of the included patient population (-2, inconsistency).

The level of evidence regarding the outcome measure neonatal sepsis was downgraded to low because of the 95% confidence interval crossed both lines of no (clinically relevant) effect (-2, imprecision).

The level of evidence regarding the outcome measure necrotizing enterocolitis was downgraded to very low because of differences in the gestational age of the included patient population (-1, inconsistency) and the low number of events (-2, imprecision).

The level of evidence regarding the outcome measure intraventricular haemorrhage was downgraded to very low because of differences in the gestational age of the included patient population (-1, inconsistency) and the low number of events (-2, imprecision).

The level of evidence regarding the outcome measure periventricular leukomalacia was downgraded to very low because of the low number of events and the 95% confidence interval crossed both lines of no (clinically relevant) effect (-3, imprecision).

The level of evidence regarding the outcome measure gestational age at birth was downgraded to moderate because of the 95% confidence interval crossed the line of no (clinically relevant) effect (-1, imprecision).

The level of evidence regarding the outcome measure survival without neurodevelopmental impairment was not reported and therefore could not be assessed with GRADE. 

Nifedipine

1. Perinatal mortality

Three studies reported perinatal mortality (Hawkins, 2021; Nijman, 2016; Vis, 2014) (Figure 5). Results were not pooled because of heterogeneity between studies.

Hawkins (2021) reported that no stillbirth or neonatal death occurred in women who received either nifedipine or placebo.

Nijman (2016) reported perinatal mortality. Two of the 27 infants (7.4%) whose mother received nifedipine died, while no deaths occurred in infants whose mother received placebo (RR=5.18, 95%CI 0.26 to 103.15).

Vis (2014) reported fetal deaths. Two of the 37 women (5.4%) who received nifedipine had a fetal death, while no fetal deaths occurred in women who received placebo (RR=4.18, 95%CI 0.24 to 98.02).

_{Figure 5. Perinatal mortality for the comparison between nifedipine and placebo}

2. Composite outcome of neonatal morbidity and mortality

Two studies reported a composite outcome of neonatal morbidity and mortality (Nijman, 2016; Vis, 2014) (Figure 6). These data were not pooled as the agreement is to start pooling when including at least three studies and because of heterogeneity between the studies.

Nijman (2016) reported a composite of adverse perinatal outcome consisting of perinatal death, bronchopulmonary dysplasia, periventricular leukomalacia > grade 1, intraventricular haemorrhage > grade 2, necrotizing enterocolitis > stage 1 and culture proven sepsis. Nine of the 27 infants (33.3%) whose mother received nifedipine was positive for this composite outcome as compared to 9 of the 28 infants (32.1%) whose mother received placebo (RR=1.04, 95%CI 0.49 to 2.21).

Vis (2014) reported a composite of adverse neonatal outcome consisting of fetal deaths, neonatal sepsis, bronchopulmonary dysplasia, necrotizing enterocolitis, intraventricular haemorrhage grade ≥3, periventricular leukomalacia grade ≥2. Three of the 37 women (8.1%) who received nifedipine had an infant who was positive for this composite outcome, while no infants whose mother received placebo were positive (RR=6.82, 95%CI 0.36 to 8.29). 

_{Figure 6. Composite outcome of neonatal morbidity and mortality for the comparison between nifedipine and placebo}

3. Respiratory distress syndrome (RDS)

Vis (2014) reported that 2 of the 36 women (5.6%) who received nifedipine had an infant with RDS, while this did not occur in infants whose mother received placebo (RR=5.00, 95%CI 0.25 to 100.63).

4. Neonatal sepsis

Two studies reported neonatal sepsis (Nijman, 2016; Vis, 2014) (Figure 7). These data were not pooled as the agreement is to start pooling when including at least three studies.

Nijman (2016) reported that 6 of the 27 infants (22.2%) whose mother received nifedipine had neonatal sepsis as compared to 7 of the 28 infants (25%) whose mother received placebo (RR=0.89, 95%CI 0.34 to 2.31).

Vis (2014) reported that 1 of the 37 women (2.7%) who received nifedipine had an infant with neonatal sepsis, while this did not occur in infants whose mother received placebo (RR=2.92, 95%CI 0.12 to 69.43).

_{Figure 7. Neonatal sepsis for the comparison between nifedipine and placebo}

5. Necrotizing enterocolitis (NEC)

Two studies reported necrotizing enterocolitis (Nijman, 2016; Vis, 2014) (Figure 8). These data were not pooled as the agreement is to start pooling when including at least three studies.

Nijman (2016) reported that 3 of the 27 infants (11.1%) whose mother received nifedipine had NEC, while this did not occur in infants whose mother received placebo (RR=7.25, 95%CI 0.39 to 134.07).

Vis (2014) reported that 2 of the 37 women (5.4%) who received nifedipine had an infant with NEC, while this did not occur in infants whose mother received placebo (RR=4.87, 95%CI 0.24 to 98.02).

_{Figure 8. Necrotizing enterocolitis for the comparison between nifedipine and placebo}

6. Intraventricular haemorrhage (IVH)

Nijman (2016) reported that 1 of the 27 infants (3.7%) whose mother received nifedipine had IVH > grade 2, while this did not occur in infants whose mother received placebo (RR=3.11, 95%CI 0.13 to 73.11).

Vis (2014) reported that women who either received nifedipine or placebo did not have an infant with IVH grade ≥ 3.

7. Periventricular leukomalacia (PVL)

Nijman (2016) reported that no PVL >grade 1 occurred in infants whose mother received either nifedipine or placebo.

Vis (2014) reported that women who either received nifedipine or placebo did not have an infant with IVH grade ≥ 2.

8. Gestational age at birth

Nijman (2016) reported that infants whose mother received nifedipine were born at a median gestational age of 32.0 weeks (IQR=29.1 to 33.3 weeks) as compared to 30.0 weeks (IQR=26.3 to 32.1 weeks).

Vis (2014) reported that infants whose mother received nifedipine were born at a median gestational age of 37+0 weeks (IQR=34+6 to 38+5 weeks) as compared to 38+2 weeks (IQR=37+0 to 39+6 weeks).

9. Survival without neurodevelopmental impairment

Not reported.

Level of evidence of the literature

According to GRADE, the level of evidence of randomized controlled trials start high.

The level of evidence regarding the outcome measure neonatal mortality and gestational age were downgraded to very low because of study limitations (-1, risk of bias), limitations in the study population (-1, indirectness) and the optimal information size was not achieved
(-2, imprecision).

The level of evidence regarding the outcome measures composite outcome of neonatal morbidity and mortality, respiratory distress syndrome, neonatal sepsis and necrotizing enterocolitis were downgraded to very low because of study limitations (-1, risk of bias), limitations in the study population (-1 indirectness), and the broad confidence intervals crossing both lines of no (clinically relevant) effect (-2, imprecision).

The level of evidence regarding the outcome measures intraventricular haemorrhage and periventricular leukomalacia were downgraded to very low because of study limitations (-1, risk of bias), differences in the definition of the outcome (-1, inconsistency), limitations in the study population (-1 indirectness), and the optimal information size was not achieved
(-2, imprecision).

The level of evidence regarding the outcome measure survival without neurodevelopmental impairment was not reported and therefore could not be assessed with GRADE. 

A systematic review of the literature was performed to answer the following question:

What are the (un)favorable effects of treating pregnant women with a threatened preterm birth between 24 and 34 weeks with tocolytics for a maximum of 48 hours compared to no tocolysis on the morbidity and mortality of the child?

Relevant outcome measures

The guideline development group considered perinatal mortality and survival without neurodevelopmental impairment as critical outcome measures for decision making; and neonatal sepsis, composite outcome of neonatal morbidity and mortality, respiratory distress syndrome, intraventricular haemorrhage, periventricular leukomalacia, and gestational age at birth as important outcome measures for decision making.

A priori, the working group did not define the outcome measures listed above but used the definitions used in the studies.

The working group defined a 1% difference for perinatal death (RR < 0.99 or > 1.01) and a 10% difference for survival without neurodevelopmental impairment (RR < 0.90 or > 1.10) as a minimal clinically (patient) important difference. For the other outcomes, a 25% difference for dichotomous outcomes (RR < 0.8 or > 1.25) and 0.5 SD for continuous outcomes was taken as minimal clinically (patient) important difference.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms from 2010 until the 13^th of December, 2023. The search was updated on the 3^rd of April, 2025 because of the publication of the APOSTEL-8 trial. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 670 hits. Studies were selected based on the following criteria:

• Systematic review (searched in at least two databases, and detailed search strategy, risk of bias assessment and results of individual studies available), randomized controlled trial, or observational studies comparing inhibition of contractions for 48 hours with no inhibition of contractions or placebo;
• The study population had to meet the criteria as defined in the PICO; and
• Full-text English language publication.

Thirty-eight studies were initially selected based on title and abstract screening. After reading the full text, 36 studies were excluded (see the table with reasons for exclusion under the tab Methods), and two studies were included. Wilson 2022 defined a broader PICO than the PICO defined for this module (it also included studies about other tocolytics and comparisons other than placebo or no tocolytics). Therefore, five randomized controlled trials included in the review were selected for the literature analysis (Hawkins 2021; Nijman 2016; Romero 2000; Vis 2014; Van der Windt 2025).

Results

Five studies were included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.