Key words
preterm contractions - tocolysis - indication and objectives - tocolytics - evidence
- safety profile
Introduction
For decades, drug-based inhibition of contractions has been a part of the treatment
concept for preterm delivery whose rate in Europe, at 5 – 18%, remains high [1]; in Germany, it was 8.6% in 2017 [2].
As the perinatal statistics from 2017 show [2], approximately 18 800 cases of tocolysis are performed in Germany annually in the
case of threatened preterm delivery, with a median duration of 3 days (up to 106 days).
For demonstrable reasons (for example, the pregnant womanʼs desire for “treatment”,
concerns about medical-legal disputes in the event of damage), tocolytics are used
too frequently and too long. In approximately 30% of pregnant women with preterm contractions,
these stop spontaneously [3]; about 50% of pregnant women deliver without tocolysis near term [4] and only 12 – 17% within one week [5]. In view of this, a differentiated indication is a precondition for performing tocolysis.
Indications for Tocolysis
Indications for Tocolysis
According to the current recommendations of the European Association of Perinatal
Medicine, tocolysis is indicated at the onset of regular preterm contractions, not
fewer than 4 contractions within 20 minutes and dynamic cervical changes (shortening/expansion
of the cervix) between 22 + 0 and 36 + 6 weeks of pregnancy [1].
The ACOG Practice Bulletin No. 171 2016 [3] recommends tocolysis in the case of regular preterm contractions and cervical dilation
of ≥ 2 cm.
From earlier investigations, it is known that measuring cervical length using ultrasound
(sensitivity 78.1%, specificity 82.7%) is superior to the digital examination (sensitivity
65.6%, specificity 72.4%) with regard to predicting preterm delivery within 7 days
[6].
Pregnant women with preterm contractions and a cervical length of ≥ 30 mm have a risk
of preterm delivery of < 2% within the next 7 days and an over 95% chance of delivering
beyond the 35th week of pregnancy without treatment [7].
The indication for tocolysis increasingly includes the qualitative (quantitative)
determination of biomarkers (foetal fibronectin, insulin-like growth factor-binding
protein-1 and placental α-microglobulin-1) in the cervicovaginal secretion to predict
preterm delivery within 7 days, in addition to the cervical length measured on ultrasound
(overview in [1]).
According to a systematic overview of prospective cohort and observational studies,
the determination of the placental α-microglobulin-1 over all areas of risk has the
highest positive predictive value for a preterm delivery within 7 days in comparison
to the two other biomarkers [8]; however, it should be taken into account here that these included non-randomised,
controlled studies must be viewed critically due to possible bias.
The combination of cervical length measurement on ultrasound and biomarker testing
in the cervicovaginal secretions should enable a differentiation to be made in pregnant
women at low risk (< 2 – 5%) for a preterm delivery within 7 days who do not need
tocolysis and pregnant women at high risk for whom an inpatient admission and tocolysis
as well as induction of foetal lung maturation is recommended. Indicative for this
recommendation were a meta-analysis by De Franco et al. [9] as well as the results of a prospective cohort study (n = 665, 24th-34th weeks of
pregnancy) from the Netherlands, with disproportionate inclusion of later weeks of
the time period mentioned and thus a significant bias [10]. Pregnant women with preterm contractions and a cervical length > 30 mm or a cervical
length of 15 – 30 mm and negative fibronectin have a preterm delivery rate < 5% within
a week. As a subsequent randomised study from the same working group [11] showed, no significant decrease in the preterm delivery rate within 7 days (8.1
vs. 2.8%) and no significant differences with regard to the median gestational age
at birth were able to be achieved in this low-risk collective (cervical length 10 – 30 mm,
negative fibronectin test, intact amniotic sac) by the oral administration of nifedipine
in comparison to placebo.
Pregnant women with regular preterm contractions, a cervical length < 15 mm or a cervical
length of 15 – 30 mm and a positive fibronectin test had a risk of preterm delivery
within 7 days of 52 and 11 – 17% respectively and thus an indication for tocolysis
[10]. In a subgroup analysis of the APOSTEL-I study, the combination of cervical length
measurement on ultrasound and the determination of foetal fibronectin proved to be
cost-effective through the reduction in inpatient admissions, tocolysis and induction
of foetal lung maturation [12]. In another subanalysis of this study, the introduction of this approach resulted
in a significant shortening of the inpatient hospital stay [13]. However, according to a recent meta-analysis from 6 randomised, controlled studies
(n = 546), the fibronectin determination alone does not lead to a reduction in the
preterm delivery rate < 28, < 32 and < 37 weeks of pregnancy and not within 7 days
or to a decrease in inpatient hospitalisations and cases of tocolysis, but to higher
inpatient costs [14]. It is currently being discussed whether, in addition to cervical length measurement
on ultrasound, the additional determination of a biomarker such as fibronectin significantly
improves the prediction of preterm delivery and thus should be included in the clinical
management in the event of threatened preterm delivery (overview in [15]). In the case of a cervical length of > 30 mm or < 20 mm, the additional determination
of a biomarker such as fibronectin is not recommended since the risk of preterm delivery
within a week in these cases is < 2 – 5% and > 25% respectively [16].
Navathe and Berghella 2016 [17] define “true preterm labour” as regular preterm contractions in connection with
a cervical length measured transvaginally of < 20 mm or 20 – 29 mm and a positive
fibronectin test in the cervicovaginal secretions; in the case of these pregnant women,
they see an indication for inpatient admission, tocolysis and induction of foetal
lung maturation.
The NICE guideline “Preterm labour and birth” 2015 [18] discusses “diagnosed preterm labour” in pregnant women with regular preterm contractions
and a cervical length on ultrasound of ≤ 15 mm. If cervical length measurement via
vaginal ultrasound is not available/feasible, the fibronectin test for risk stratification
should be used in the event of threatened preterm delivery ≥ 30 weeks of pregnancy;
a combination of both methods is not recommended.
Tocolysis should be performed in pregnant women between 24 + 0 – 25 + 6 weeks of pregnancy
with intact amniotic sac and “suspected preterm labour” (regular, painful contractions
and cervical shortening/opening verified by vaginal examination, but not ≥ 4 cm) as
well as between 26 + 0 – 33 + 6 weeks of pregnancy in the case of “suspected” or “diagnosed
preterm labour”.
A possible algorithm for actions to be taken, published by some authors in the meantime
[1], [17], [18], [19], is shown in [Fig. 1].
Fig. 1 Possible algorithm for actions to be taken in the case of regular preterm contractions:
230/7 – 336/7.
Objectives of Tocolysis
There is no evidence to date from placebo-controlled studies that tocolytics lead
to a significant decrease in perinatal and neonatal mortality as well as in neonatal
morbidity [20], [21], [22]. However, the studies in this regard should be assessed critically owning to the
considerable heterogeneity and insufficient quality in studies/meta-analyses (among
other things, inclusion of pregnant women with advanced gestational age, no antenatal
induction of lung maturation or transfer to a perinatal centre, inadequate statistical
power for rare complications, e.g. neonatal death) [23]. According to the current view, [1], [3], [18], [22], [24] the objective of tocolysis is the prolongation of the pregnancy by at least 48 hours
to ensure completion of induction of lung maturation using corticosteroids, to enable
an in-utero transfer of the pregnant woman to a perinatal centre with a neonatal intensive
care unit and to complete foetal neuroprotection using magnesium sulphate < 32 weeks
of pregnancy [3]. These measures are evidence-based methods to decrease neonatal morbidity and mortality
[24].
Choice of tocolytic
The following criteria should be taken into account:
-
Approved or off-label use
-
Tocolytic effectiveness to prolong the pregnancy by at least 48 hours
-
Rate of maternal adverse effects/complications
-
Rate of foetal adverse effects/complications
-
Early, late, long-term morbidity
-
Practicability (e.g. mode of application), need for monitoring
-
Costs, cost effectiveness
Beta Sympathomimetics
For several decades, since the 1970s, fenoterol has been the most frequently used
tocolytic in Germany. According to a 2009 meta-analysis [25], a prolongation of the pregnancy by 48 hours can be achieved in 75% and by 7 days
in 65% of cases using beta sympathomimetics.
According to a 2012 network meta-analysis [26], beta sympathomimetics are indeed effective in prolonging pregnancy for 48 hours
(OR 2.41; 95% CI 1.27 – 4.55), but significantly less effective than calcium channel
blockers and indomethacin and also have the highest rate of maternal adverse effects
of all tocolytics in comparison to placebo (OR 22.67; 95% CI 7.51 – 73.67, see below).
A Cochrane Review published in 2014 [27] of 28 randomised, controlled studies from 1974 – 2010 (all application forms) revealed
for beta sympathomimetics – in comparison to placebo (10 studies, n = 1209) – a significant
decrease in the preterm delivery rate within 48 hours (RR 0.68; 95% CI 0.53 – 0.87)
and within 7 days (RR 0.80; 95% CI 0.65 – 0.89). However, there was no significant
decrease in the frequency of preterm deliveries before the 37th week of pregnancy
(RR 0.95; 95% CI 0.88 – 1.03, 10 studies, n = 1212), perinatal mortality (RR 0.89;
95% CI 0.46 – 1.55, 6 studies, n = 1174 children) as well as the rate of RDS (RR 0.87;
95% CI 0.71 – 1.08, 8 studies, n = 1239 children), cerebral pareses, necrotising enterocolitis
and transfer to the neonatal intensive care unit. One point of criticism is that this
Cochrane Review included 24 (out of 28) studies prior to 1990 and at that time, induction
of foetal lung maturation with corticosteroids was not a standard practice. In addition,
the target parameters such as preterm deliveries before the 37th week of pregnancy
should be looked at critically, since in many countries, no tocolysis is given starting
from 34 + 0 weeks of pregnancy. Certain morbidities such as RDS are especially relevant
in the case of an early gestational age and therefore can also, depending on the study
collective (e.g. large percentage of pregnancies beyond week 30) still demonstrate
significances with difficulty.
Beta sympathomimetics do not influence the long-term development of the child (overview
in [28]). In twin pregnancies, the prophylactic administration of beta sympathomimetics
does not decrease the rate of preterm deliveries < 37 weeks [29]. According to a Cochrane Review from 2012 [30] which included 13 randomised, controlled studies (n = 1551), the oral use of beta
sympathomimetics following initial tocolysis in comparison to placebo/no treatment
does not lead to any significant decrease in the rate of preterm delivery < 37 weeks
(RR 1.11; 95% CI 0.91 – 1.35) or to any decrease in the transfer rate of children
to the neonatal intensive care unit (RR 1.28; 95% CI 0.68 – 2.41).
Based on the pharmacological effects with an increase in heart rate and cardiac output
as well as peripheral vasodilatation, beta sympathomimetics are above all burdened
with adverse cardiovascular effects in up to 80% of cases [27]: significantly higher incidence of palpitations (RR 10.11; 95% CI 6.56 – 15.58),
tachycardia (RR 4.08; 95% CI 1.55 – 10.73), chest pain (RR 11.3; 95% CI 3.8 – 33.46),
dyspnoea (RR 3.86; 95% CI 2.21 – 6.77) and headaches (RR 4.1; 95% CI 2.6 – 6.35) in
comparison to placebo. In addition, there are neurological symptoms, particularly
tremors (RR 10.74; 95% CI 6.2 – 18.6), inner restlessness/anxiety and insomnia. Metabolic
adverse effects involve, in particular, hyperglycaemia and hypokalaemia.
According to a prospective cohort study [31], the rate of serious, primarily maternal adverse effects is 1.7% and thus around
3.8 times higher in comparison to other tocolytics. Overall, the therapy discontinuation
rate is between 6 and 38% (RR 11.38; 95% CI 5.2 – 24.9) [21], [27].
The most serious complication of beta sympathomimetics is pulmonary oedema in approx.
0.3% of pregnant women to the point of maternal death in individual cases [21]. The decrease in diastolic cardiac filling in tachycardia promotes the development
of pulmonary oedema, particularly in unbalanced and increased volume replacement,
concomitant corticosteroid administration, preeclampsia, blood transfusions and multiple
pregnancies [32]. In comparison to the continuous application of beta sympathomimetics (fenoterol),
bolus tocolysis with fenoterol has a significantly lower rate of maternal adverse
effects [33].
No teratogenic effects are known for beta sympathomimetics; they cross the placental
barrier and can therefore lead to foetal tachycardia (in up to 28% of cases), limitations
in heart sound variability as well as neonatal hyperinsulinaemia/hypoglycaemia.
Conclusion for Clinical Practice
-
Approved for tocolysis
-
Effective for prolonging the pregnancy by 48 h (7 days)
-
Unfavourable adverse effect profile of all tocolytics and the highest therapy discontinuation
rate
-
Immobilisation of the pregnant woman in the case of parenteral use
-
Significant monitoring needed (e.g. ECG, laboratory testing), close circulatory monitoring
necessary (see [1])
In current guidelines, beta sympathomimetics are no longer recommended for tocolysis
[18], [22], [24].
NO donors
The transdermal application of nitroglycerin (patches, 10 mg/24 h) was considered
for some years to be a new, innovative method for tocolysis since it is effective,
has few side effects, is easy to apply and cost-efficient. While NO donors were not
mentioned in the 2009 meta-analysis of Haas et al. [25], the 2012 network meta-analysis [26] classified NO donors as the least effective tocolytics in comparison to placebo
for prolonging delivery by 48 hours (RR 1.19; 95% CI 0.64 – 5.33).
A meta-analysis from 2013 (13 randomised, controlled studies, n = 1302) was not able
to show any significant differences in two studies with a small number of cases (total
n = 186) between transdermal nitroglycerin and placebo with regard to prolongation
of pregnancy by 48 hours [34]. However, transdermal nitroglycerin, in comparison to beta sympathomimetics (9 studies,
n = 1024) led to a significant reduction in the rate of preterm delivery < 34 weeks
of pregnancy (20 vs. 28%, RR 0.71; 95% CI 0.51 – 0.99) and < 37 weeks of pregnancy
(44 vs. 57%, RR 0.76; 95% CI 0.60 – 0.96) without an influence on neonatal morbidity.
Moreover, transdermal nitroglycerin had a significantly lower rate of maternal adverse
effects except for headaches and hypotension. Two other randomised, controlled studies
compared transdermal nitroglycerin with nifedipine and magnesium sulphate without
significant differences with regard to prolongation of pregnancy by 48 hours. The
meta-analysis came to the conclusion that the evidence for the use of transdermal
nitroglycerin for tocolysis is insufficient [34]. A Cochrane Review from 2014, which also came to comparable results and conclusions
[35], included 12 randomised, controlled studies with 1227 pregnant women with the indication
that the studies have inadequate statistical power with regard to most variables.
In current international guidelines, NO donors are not mentioned for tocolysis or
are not recommended [1], [3], [18], [22], [24].
Conclusion for Clinical Practice
-
Not approved for tocolysis (off-label use)
-
More effective in comparison to beta sympathomimetics with regard to prolongation
of pregnancy by 48 hours
-
Possibly decreased acceptance due to high rate of headaches
-
Inadequate data, no recommendation in international guidelines
Cyclooxygenase Inhibitors
Cyclooxygenase Inhibitors
The nonspecific cyclooxygenase inhibitor indomethacin (initial dosage 50 – 100 mg
oral or rectal, followed by 25 – 50 mg every 4 – 6 h) is used most frequently for
tocolysis and is the only tocolytic with an anti-inflammatory effect. According to
a 2009 meta-analysis [25], delivery was postponed by 48 hours in 93% of cases and over 7 days in 76% of cases
with indomethacin < 32 weeks of pregnancy, however the low number of pregnant women
included (n = 442) limits the significance of this meta-analysis.
A subsequent network meta-analysis from 2012 [26] which included 18 studies (1980 – 2007) showed that COX inhibitors were the most
effective tocolytics for prolonging pregnancy by 48 hours in comparison to placebo
(OR 5.39; 95% CI 2.14 – 12.34) – with concomitantly the lowest rate of maternal adverse
effects and good neonatal outcome (see [Table 1]). It is surprising that a Cochrane Review from 2015 [36] which included 20 randomised, controlled studies (n = 1509) from 1980 – 2012, 15
of which were studies with indomethacin, found different results although the same
studies, for the most part, were evaluated. In this Cochrane Review, no significant
decrease in the rate of premature delivery < 48 hours was able to be demonstrated
in comparison to placebo (3 studies, n = 102) (RR 0.20; 95% CI 0.03 – 1.28). In comparison
to beta sympathomimetics, COX inhibitors led to a significant decrease in the preterm
delivery rate < 48 hours (RR 0.27; 95% CI 0.08 – 0.96) and to a greater decrease in
the preterm delivery rate < 37 weeks of pregnancy (RR 0.53; 95% CI 0.28 – 0.99; 4
studies, n = 180) without an impact on neonatal morbidity and mortality. Compared
to beta sympathomimetics, COX inhibitors had significantly lower rates of maternal
adverse effects (RR 0.19; 95% CI 0.11 – 0.31; 5 studies, n = 248) and a lower rate
of discontinuation of therapy (RR 0.39; 95% CI 0.25 – 0.62; 5 studies, n = 635).
Table 1 Network meta-analysis of 95 randomised, controlled studies (1982 – 2012). Placebo
as reference: median OR (95% CI).
|
Tocolytics
|
Prolongation of pregnancy 48 h+
|
Probability of being best
|
Neonatal mortality*
|
Probability of being best
|
RDS*
|
Probability of being best
|
Maternal adverse effects#
|
Probability of being best
|
|
+ Odds ratio > 1 favour active class
* Odds ratio < 1 favour active class
# Odds ratio > 1 favour placebo
Conclusion: Indomethacin and calcium channel blockers have the highest probability
of delaying delivery and improving neonatal and maternal outcome
Haas DM et al. BMJ 2012; 345: e6226 [26]
|
|
Betamimetics
|
2.41 (1.27 – 4.55)
|
0.01
|
0.62 (0.14 – 2.48)
|
0.12
|
0.85 (0.5 – 1.45)
|
0.14
|
22.68 (7.5 – 73.7)
|
< 0.01
|
|
PG inhibitors
|
5.39 (2.14 – 12.34)
|
0.83
|
0.62 (0.04 – 4.63)
|
0.28
|
0.87 (0.4 – 1.75)
|
0.20
|
1.63 (0.4 – 6.85)
|
0.21
|
|
Calcium channel blockers
|
2.71 (1.17 – 5.9)
|
0.06
|
0.39 (0.09 – 1.5)
|
0.41
|
0.71 (0.37 – 1.43)
|
0.47
|
3.50 (1.0 – 16.9)
|
0.01
|
|
Magnesium sulphate
|
2.76 (1.58 – 4.94)
|
0.02
|
0.97 (0.3 – 3.3)
|
0.03
|
0.99 (0.55 – 1.7)
|
0.03
|
8.15 (2.47 – 27.7)
|
< 0.01
|
|
Atosiban
|
2.02 (1.1 – 3.8)
|
0.01
|
0.62 (0.16 – 2.35)
|
0.13
|
0.89 (0.55 – 1.37)
|
0.10
|
1.99 (0.61 – 6.94)
|
0.08
|
|
NO donors
|
1.91 (0.64 – 5.33)
|
0.04
|
No information
|
–
|
No information
|
–
|
3.2 (0.4 – 20.84)
|
0.10
|
The Cochrane Review concludes that because of the low number of cases in studies,
the insufficient information on the safety profile and long-term consequences, the
evidence for the use of COX inhibitors for tocolysis is not sufficient.
COX inhibitors rapidly cross the placental barrier and lead to inhibition of foetal
prostaglandin synthesis with potential consequences for the foetus: as a result of
an increased vasopressin effect with decrease in the renal blood flow [37], there can be decreased urine production (rarely to the point of renal failure)
following indomethacin in foetuses, resulting in oligohydramnios in 5 – 15% of cases,
and even in up to 70% of cases in the case of use for more than 72 h [38]. After stopping indomethacin, the amount of amniotic fluid should normalise once
again within 24 (− 96) h [39]. According to the recommendations of the European Association of Perinatal Medicine
2017 [1] the amount of amniotic fluid should be checked prior to starting therapy and after
48 – 72 h. Indomethacin should only be given if there is a normal amount of amniotic
fluid and if oligohydramnios occurs, it should be discontinued or at least the dose
should be reduced [1].
To avoid premature closure of the ductus arteriosus, it is recommended to administer
COX inhibitors only up to the 32nd week of pregnancy for 48 hours. Under these conditions,
only one case with antenatal closure of the ductus arteriosus was identified among
403 pregnant women with indomethacin exposure [40], [41]. Prior to 32 weeks of pregnancy, an echocardiographic examination of the foetal
ductus arteriosus with assessment of the tricuspid valve (tricuspid regurgitation)
is recommended if treatment lasts > 48 hours [1]. In accordance with this in the Cochrane Review from 2015 and in a subsequent meta-analysis
[42], there was not found to be an elevated risk of a significant increase in neonatal
mortality or the rate of RDS, bronchopulmonary dysplasia, patent ductus arteriosus
and pulmonary hypertension or intraventricular bleeding of all levels of severity.
However, in the meta-analysis [42] which included 27 observational studies, (n = 1731) there was a significantly increased
risk of
-
Intraventricular bleeding, grade III/IV RR 1.29 (95% CI 1.06 – 1.50),
-
necrotising enterocolitis RR 1.36 (95% CI 1.08 – 1.71) and
-
periventricular leucomalacia RR 1.59 (95% CI 1.17 – 2.17).
However, in this meta-analysis, there was no information provided on the dosage and
duration of the indomethacin administration and also no details given on the percentage
of pregnant women who received antenatal induction of lung maturation with corticosteroids.
Likewise, it should be taken into account that indomethacin was/is used in the very
early weeks of pregnancy in which the previously described complications naturally
occur more frequently.
Conclusion for Clinical Practice
-
Off-label use
-
Easy application, cost-effective
-
Use only until 32 weeks of pregnancy for 48 hours
-
Indications of increased neonatal morbidity (severe intraventricular bleeding, necrotising
enterocolitis, periventricular leucomalacia): inform paediatrician of administration!
-
Varying recommendations (controversial data, to some extent)
-
not recommended/insufficient data: Cochrane Review 2015 [36], Hammers et al. 2015 [42], Nijman et al. 2016 [5]. WHO Recommendations 2015 [24]
-
recommended with restrictions: Navathe and Berghella 2016 [17], European recommendations 2017 [1]
-
recommended: among others: 2009 meta-analysis [25], 2012 network meta-analysis [27], ACOG Practice Bulletin 2016 [3]
Magnesium Sulphate
The data on magnesium sulphate as a tocolytic are controversial. According to the
meta-analysis from 2009 [25], a prolongation of pregnancy by 48 hours can be achieved in 89% of cases with magnesium
sulphate and by up to 7 days in 61%. A network meta-analysis from 2012 [26] which included 29 studies confirmed the high tocolytic efficacy of magnesium sulphate
for prolonging pregnancy by 48 hours (OR 2.76; 95% CI 1.58 – 4.94) without influencing
neonatal mortality (OR 0.97; 95% CI 0.29 – 3.29). However, in a probability analysis,
this work concluded that magnesium sulphate is not the most effective tocolytic with
the fewest adverse effects. Contrary results were seen in the Cochrane Review from
2014 [43] which included 37 studies with 3571 pregnant women (period from 1982 – 2012). This
did not reveal any evidence (4 studies) that magnesium sulphate, in comparison to
placebo/no treatment, is more effective for prolonging pregnancy by 48 hours (RR 0.57;
95% CI 0.28 – 1.15) and increasing the gestational age until delivery. In addition,
a borderline increase in cases of perinatal/neonatal death following magnesium sulphate
was seen (risk ratio 4.56; 95% CI 1.00 – 20.86, 2 studies, 257 children). According
to this, the Cochrane Analysis concluded that magnesium sulphate is ineffective for
prolonging pregnancy or avoiding preterm delivery. With regard to these controversial
results, it is interesting that the network meta-analysis (2012) and the Cochrane
Review (2014) evaluated nearly the same studies covering a comparable period – in
the Cochrane Review, only one study after the period covered by the network meta-analysis
was included [44].
The myometrium-relaxing effect of magnesium sulphate as well as the scope and frequency
of adverse effects are dose-dependent (overview in [45]). Thus at an initial dose of 4 g and a maintenance dose of 2 g/hour as an infusion,
the rate of success with regard to prolongation of pregnancy by 48 hours was between
60 – 77%; at an initial dose of 6 g followed by 2 (4) g/hour, it was between 84 – 94%
[46]. In a randomised comparative study between nifedipine and magnesium sulphate (4 g
magnesium sulphate initially, followed by 2 – 4 g as maintenance dose), prolongation
of pregnancy by 48 hours was able to be achieved in 87% of cases with magnesium sulphate,
however the rate of maternal adverse effects – 62% – was high (including 3 cases of
pulmonary oedema) [47]. According to a current retrospective case-control study (n = 150), magnesium sulphate
significantly increases the risk of pulmonary oedema (adjusted OR 3.51; 95% CI 1.26 – 9.76;
[48]). This is in line with a Cochrane Review from 2015 on tocolysis with magnesium sulphate
at various dosages [49].
According to Elliott et al. [50], magnesium sulphate is equieffective to other tocolytics at a sufficiently high
dose reaching magnesium levels between 5.0 – 8.0 mg/100 ml (Note: The patellar tendon
reflex disappears between 9 – 13 mg/100 ml).
Except in the U. S., magnesium sulphate is no longer recommended for tocolysis [51] in current reviews and guidelines [1], [4], [17], [22], [43], [52].
Oxytocin Receptor Antagonists (Atosiban)
Oxytocin Receptor Antagonists (Atosiban)
The advantage of the selective oxytocin receptor antagonist atosiban approved for
tocolysis, in comparison to other “ubiquitously” effective tocolytics is the primarily
uterus-specific effect without a significant influence on the cardiovascular system,
the CNS, the kidney and the lung as well as without negative metabolic effects (overview
in [53]). With low placental passage of 10 – 12%, no relevant adverse effects on the foetus/child
are known, even after a 2-year observation period. Except for “hypersensitivity to
the drug”, there are no substance-specific contraindications. Disadvantages are the
i. v. application and associated immobilisation of the pregnant woman as well as the
drug costs which are significantly higher as compared to other tocolytics. A Cochrane
Review from 2014 [54] included 14 studies with 2485 pregnant women with the following results: oxytocin
receptor antagonists (atosiban and barusiban) did not prove to be significantly more
effective, compared to placebo, in prolonging pregnancy by 48 hours (4 studies, n = 854;
RR 1.05; 95% CI 0.15 – 7.43) and no significant differences were seen with regard
to the rate of preterm delivery < 37 weeks of pregnancy, perinatal mortality and neonatal
morbidity.
In 8 studies (n = 1402) atosiban was compared to beta sympathomimetics; this revealed
no significant differences between both tocolytics with regard to prolongation of
pregnancy by 48 hours (RR 0.89; 95% CI 0.66 – 1.22) as well as perinatal mortality
(RR 0.55; 95% CI 0.21 – 1.48; 3 studies with 812 children). The difference in the
rate of maternal adverse effects (RR 0.38; 95% CI 0.21 – 0.68) and treatment discontinuations
(RR 0.05; 95% CI 0.02 – 0.11; 5 studies, n = 1161) was significant. A large prospective
cohort study revealed mild adverse effects in 0.2% of the pregnant women, but no severe
adverse effects (RR 0.07; 95% CI 0.01 – 0.4; [31]).
In the meta-analysis of Haas et al. from 2009 [25], which included 1249 pregnant women, prolongation of pregnancy by 48 hours was achieved
in 86% of cases and by 7 days in 78% of cases. The network meta-analysis from 2012
[26] revealed a significant tocolytic effect for atosiban in the prolongation of pregnancy
by 48 hours (OR 2.02; 95% CI 1.10 – 3.80) without an influence on the rate of RDS
and perinatal mortality.
Of interest is a prospective, randomised study from Germany [55] on 105 pregnant women between 24 + 0 to 33 + 6 weeks of pregnancy in which the pulsatile
application of fenoterol was compared to atosiban at a standard dosage. This revealed
no significant differences with regard to prolongation of pregnancy by 48 hours (79.8
vs. 86.3%) and by 7 days (66.7 vs. 78.4%). However, on fenoterol, adverse cardiovascular
effects were observed in 78% of cases and on atosiban in only 4% of cases (treatment
discontinuation rate 9% versus 0). The data on the comparison of nifedipine and atosiban
for acute tocolysis were unclear since 3 randomised studies with small numbers of
cases had controversial results (overview in [56]). This was clarified by the APOSTEL-III study, a randomised, controlled study (n = 510)
on pregnant women with regular preterm contractions, a cervical length of ≤ 10 mm
or a cervical length of 11 – 30 mm and a positive fibronectin test [57]. In this study, 20 mg nifedipine were given in the first hour, followed by 20 mg
nifedipine sustained-release every 6 h (up to 48 h) and compared to a standard dosage
of atosiban. No significant differences were seen between nifedipine and atosiban
with regard to prolongation of pregnancy by ≥ 48 hours (68 vs. 66%) and by 7 days
(51 vs. 45%), no significant differences with regard to perinatal mortality (5 vs.
2%) and neonatal morbidity (14 vs. 15%); given a higher rate of maternal adverse effects,
the treatment discontinuation rate following nifedipine was not significantly higher
(6 vs. 3%).
Conclusion for Clinical Practice
-
Approved (not in the USA)
-
Tocolytic efficacy comparable to beta sympathomimetics and nifedipine
-
Associated with few adverse effects for mother and child
-
Intravenous application and immobilisation necessary
-
High drug costs
-
Recommendations: particularly in the case of contraindications to nifedipine and indomethacin
[18] or as a first-line tocolytic [1], suitable tocolytic in the case of pre-existing metabolic and cardiovascular diseases
in the mother [23]
Calcium Channel Blockers (e.g. Nifedipine)
Calcium Channel Blockers (e.g. Nifedipine)
Nifedipine is used most frequently (half-life 2 – 3 hours, maximum plasma concentration
after a 10-mg tablet 30 – 60 min); it is not approved for tocolysis. The dosage recommendations
differ, e.g.
-
NICE 2015 [18]: initially 20 mg oral, followed by 10 – 20 mg three to four times daily, depending
on uterine activity (see also [24]).
-
ACOG 2016 [3]: 30 mg nifedipine initially, then 10 – 20 mg every 4 – 6 h.
Dose-dependent adverse effects such as flushing, headaches, dizziness, palpitations,
tachycardia and hypotension may occur. The rate of serious maternal adverse effects
is 0.9% and is significantly higher than after atosiban; the rate of treatment discontinuation
as a result of marked hypotension is < 0.5% [31] and isolated cases with intrauterine death have been described in this connection
[58], [59] whereby a causal connection must be critically scrutinised.
Caution is needed in the case of pre-existing cardiac diseases and pre-existing hypotension:
The use of slow-release preparations can lead to maternal hypotension which persists
for hours [60] and the sublingual application of nifedipine can lead to sudden and rapid decreases
in blood pressure with consecutive foetal hypoxia [61].
According to a meta-analysis from 2010 [62], an approximately 4-fold increase in maternal adverse effects should be expected
if a daily dose of > 60 mg is exceeded (OR 3.78; 95% CI 1.27 – 11.2); in particular,
there is a three-fold increase in tachycardia and an 8- to 9-fold increase in hypotension
which can have serious consequences for the patient with cardiac problems. Isolated
cases of myocardial infarcts have been published [63], however also of pulmonary oedema, above all, following nicardipine and also following
nifedipine at daily dosages ≥ 150 mg in connection with multiple pregnancies, intramuscular
corticosteroid administration and high volume replacement [63].
Calcium channel blockers do not have any significant negative effects on the child,
also not on utero-foetoplacental circulation, in particular [64].
Long-term investigations on children following in-utero exposure with nifedipine did
not reveal any negative effects in children aged 9 – 12 years [65].
According to a meta-analysis of Conde-Agudelo et al. in 2011 (26 studies, n = 2124;
[66]) in 2014, a Cochrane Review [67] was published which took 38 studies (n = 3550) into account, 35 of which were with
nifedipine. The significant heterogeneity and the quality of the studies, which is
to some extent low, should be pointed out. In comparison to placebo/no therapy (2
studies, n = 173) there was a significant decrease in the rate of preterm delivery
within 48 hours (RR 0.30; 95% CI 0.21 – 0.43) with an increased rate of maternal adverse
effects. Calcium channel blockers demonstrated significant advantages over beta sympathomimetics:
a prolongation of pregnancy by a median of 4.4 days, a decrease in the rate of preterm
delivery < 32 weeks of pregnancy (RR 0.89; 95% CI 0.80 – 0.89), the rate of RDS (RR
0.64; 95% CI 0.48 – 0.86), of necrotising enterocolitis (RR 0.25; 95% CI 0.05 – 0.96),
of intraventricular haemorrhage (RR 0.53; 95% CI 0.34 – 0.84) as well as admissions
to the neonatal intensive care unit (RR 0.74; 95% CI 0.63 – 0.87). Moreover, there
was a significantly lower rate of maternal adverse effects (RR 0.36; 95% CI 0.24 – 0.53).
The studies comparing calcium channel blockers to magnesium sulphate, NO donors and
indomethacin are inadequate.
In the meta-analysis from 2009 [25] delivery was able to be postponed by 48 hours with calcium channel blockers in 66%
of cases and by 7 days in 62% of cases. According to the network meta-analysis from
2012 [26], calcium channel blockers are the most effective tocolytic, after indomethacin,
for postponing delivery by 48 hours, with a low rate of maternal adverse effects and
good neonatal outcome.
Conclusion for Clinical Practice
-
Not approved
-
Simple application mode
-
High tocolytic efficacy, equieffective to beta sympathomimetics and atosiban
-
Low rate of maternal adverse effects
-
No negative effects on the foetus/newborn, no negative long-term effects on the child
-
Recommended as first-line tocolytic: NICE 2015 [18], WHO 2015 [24], French guideline 2017 [22]
An overview of the tocolytics currently used can be found in [Table 2].
Table 2 Overview: common tocolytics.
|
Tocolytics
|
Approved
|
Prolongation of pregnancy* by 48 h
|
Adverse effects – mother
|
Adverse effects – child
|
Mode of application
|
Amount of monitoring needed
|
Drug costs
|
|
*evaluated according to meta-analyses of Haas et al. 2009 and 2012 [25], [26]
To some extent no direct comparative studies
Preference: Nifedipine, indomethacin < 32 weeks of pregnancy (48 h) (off-label use);
Atosiban: approved
|
|
Beta sympathomimetics (continuous i. v.)
|
+
|
++
|
+++
|
++
|
i. v.
|
high
|
moderate
|
|
Indomethacin
|
–
|
++(+)
|
(+)
|
+(+)
|
rectal/oral
|
low
|
very low
|
|
Atosiban
|
+
|
++
|
(+)
|
–
|
i. v.
|
low
|
high
|
|
Magnesium sulphate
|
–
|
Controversially dose-dependent
|
Dose-dependent ++
|
++
|
i. v.
|
moderate
|
moderate
|
|
Calcium channel blockers (nifedipine)
|
–
|
++(+)
|
+
|
(+)
|
oral
|
low
|
very low
|
|
NO donors
|
–
|
++
|
++
|
–
|
transdermal
|
low
|
low
|
Discussion
Next to induction of labour, tocolysis is among the most common obstetric interventions.
The lack of current, randomised, controlled and, above all, placebo-controlled studies
is noteworthy, and despite efforts regarding innovative new developments, the stagnation
in the clinical approval and availability of new tocolytics. Taking the current indications
for tocolysis into account – not only to treat the symptom of “preterm contractions”
but only those which have an effect on the cervix – placebo-controlled studies should,
for understandable reasons, be a thing of the past. The implementation of randomised,
controlled studies has become increasingly more difficult not only in Germany in recent
years. The recruitment of a sufficiently high number of “study participants” has,
among other things, also become more difficult, particularly in the case of off-label
use of the tocolytic (only fenoterol and atosiban are approved) in light of the pregnant
womenʼs increased need/desire for information, given a rapid increase in easily accessible
sources of information, resulting in uncertainty or even a refusal of the pregnant
women to take part in studies. The legal and bureaucratic hurdles for conducting randomised-controlled
studies have grown, while at the same time, there has been a significant decrease
in third-party funding for obstetrical clinical studies and the financial support
of the industry for studies in which affordable drugs which have already been introduced
and are established in other areas of indication are used. As our own research revealed,
the manufacturer of nifedipine, for example, is not interested in conducting costly
placebo-controlled studies as a precondition for approval during pregnancy in the
case of such a “low-price product”. It is therefore hardly surprising that, since
2016, no randomised, controlled studies on acute tocolysis with indomethacin and only
one randomised, controlled study on acute tocolysis with nifedipine has been published;
according to PubMed research, there have also been no new studies on atosiban since
2014.
Another problem is the inconsistent and, to some extent, controversial data on the
use of tocolytics. The small amount of evidence, particularly of the serious maternal
adverse effects of tocolytics and the neonatal morbidity and mortality from Cochrane
and meta-analyses, primarily stems from the considerable heterogeneity (e.g. different
indications, inclusion of pregnant women with advanced gestational age, no antenatal
induction of lung maturation and no in-utero transfer to a perinatal centre, cumulative
data for different forms of application and dosages) and the mostly moderate to low
quality of the studies included as well as the inadequate statistical power for rare,
serious complications or perinatal/neonatal death. For these reasons, meta-analyses
have been unable to prove to date that tocolytics are able to reduce the serious neonatal
morbidity and mortality. On the other hand, it should be understandable that, through
prolongation of pregnancy by 48 hours which can be achieved with tocolytics in approximately
80% of cases which enables completion of induction of foetal lung maturation and the
in-utero transfer to a perinatal centre as well as the implementation of foetal neuroprotection
with magnesium sulphate, especially before the 30th week of pregnancy, a decrease
in the serious neonatal morbidity and mortality can be reached. Upon critical review,
it is also striking that Cochrane and meta-analyses, despite including nearly identical
studies, arrive at contrary statements (e.g. magnesium sulphate: Cochrane Review 2014
versus network meta-analysis 2012) and in addition, predominantly “older” studies
published before the turn of the millennium influence the results in which the above-mentioned
evidence-based measures were not yet standard. For example, in the Cochrane Review
from 2015 [36] on the COX inhibitors which covered 20 studies, 12 studies were published prior
to 2000; in the Cochrane Review by Crowther et al. [43] on magnesium sulphate as a tocolytic, 25 of 37 studies were published prior to 2000.
The uncertainty of the data is also reflected by the fact that nearly unanimously
at the end of current reviews and meta-analyses, there is a call for additional placebo-controlled
or randomised, controlled studies which, however, are clearly not being performed.
Independent of the insufficient data and lack of evidence, particularly with regard
to the improvement in neonatal outcome, and this is a priority, drug tocolysis is
and remains an indispensable measure in everyday obstetrical practice. To date, there
is no “ideal” tocolytic. In the authorsʼ opinion, nifedipine and atosiban are suitable
tocolytics with regard to efficacy, adverse effect profile and effects on the child.
Indomethacin is a potent tocolytic with anti-inflammatory effects and a low rate of
maternal adverse effects and should be justified, especially in early preterm delivery,
however it is not clear whether or not indomethacin leads to an increase in foetal
complications.
Beta sympathomimetics should no longer be used for tocolysis due to the high rate
of maternal and foetal adverse effects and magnesium sulphate should not be used due
to the controversial study results and the adverse effect profile at a high dose.
