Key words
induction of labour - foetal Doppler - middle cerebral artery - umbilical artery -
umbilical cord pH
Introduction
In Germany, roughly 20 – 25% of all births are induced pharmacologically, and up to
35% in the Anglo-American countries [1], [2]. The indications include post-term pregnancy, underlying maternal disease such as
pre-eclampsia or diabetes mellitus, maternal exhaustion, suspected foetal macrosomia,
subjectively sluggish foetal movements or a suspect CTG. Pharmacological induction
of labour, unlike spontaneous delivery, harbours a roughly 2.2 times increased risk
of secondary caesarean section [2], and the duration of induction is often difficult to predict.
The cervical Bishop score is generally used to assess the success of induction [3]. However, this has only limited predictive value as regards the potential need for
secondary caesarean section: with a Bishop score of 4, 5, 6 and 9, the sensitivity
and specificity are 47%/75%, 61%/53%, 78%/44% and 95%/30% respectively [4].
While the resistance index (RI) of the foetal aorta remains relatively constant over
the course of pregnancy, the resistance of the umbilical artery (UA) and middle cerebral
artery (MCA) diminishes. Severi et al. showed in 2008 that both low pulsatility indices
(PI) and low resistance in the foetal MCA are indicators that the onset of labour
is approaching [5].
The aim of this study was to find additional predictive parameters for prognostic
assessment of the success of pharmacological induction. Prior to pharmacological induction
of labour at term, Doppler measurement of the foetal UA and MCA was performed in addition
to evaluation of the Bishop score. The results were correlated with the duration of
induction until delivery, the mode of delivery and the foetal outcome. Additional
subgroup analyses were performed according to parity (primi- vs. multiparity) and
weeks of gestation at the start of induction (≤ 41 + 0 vs. > 41 + 0), since both factors
may have an influence on the duration of induction and the association between induction
duration and Doppler results.
Patients and Method
Patient recruitment
49 patients over 18 years with an uncomplicated singleton pregnancy, who presented
in our perinatal centre between October 2012 and April 2013 for induction of labour
at the estimated date of confinement (38 + 0 – 42 + 0 weeks), were included in the
study. Patients with
-
Maternal diseases such as diabetes mellitus or pre-eclampsia,
-
Ultrasonographic morphological or genetic foetal anomalies,
-
Foetal growth restriction or suspected macrosomia and
-
Pathological foetal Doppler and/or pathological CTG
were excluded from the study. The established estimated due date was checked using
the crown-rump length (CRL) in the first-trimester ultrasound scan and corrected according
to the CRL if there was a discrepancy of more than 7 days [6].
Since the German guideline on “Management of overdue and post-term pregnancy” (DGGG
2010) and the National Institute of Clinical Excellence (NICE) in its clinical guideline
on induction of labour (2008) recommend induction after 41 + 0 weeks regardless of
other maternal and/or foetal factors, we also analysed the subgroup of patients over
41 + 0 weeks.
Data collection
Immediately before planned induction of labour, the cervical dilation in cm and the
Bishop score were determined [3]. This was followed by Doppler ultrasound scan of the foetal UA and MCA using the
high-resolution ultrasound Voluson Expert 730 System, GE Voluson e® or GE Voluson Expert 8®, General Electrics, GE Medical Systems, Solingen, Germany. The ultrasound scans were
evaluated by 2 trained ultrasound experts (at least DEGUM I level); for both the resistance
index (RI) of the foetal UA and the RI of the MCA, the mean of 3 consecutive single
measurements was recorded through our internal hospital Viewpoint System (version
5). Since the blood flow patterns of the foetal vessels and thus the measured resistance
in the vessels in pregnancy depend on the week of gestation [7], the RI of the arteries was divided by the mean corresponding to the gestational
age to allow better comparability and shown as MoM (multiple of the median) [8]. The duration of induction was measured from the first dose of the medication given
to induce labour until the delivery of the baby. The mode of delivery and arterial
cord pH as a measure of foetal outcome were documented immediately after delivery.
The data were analysed retrospectively from the files in 23 cases and prospectively
in 26 cases in the context of the induction.
Statistical analysis
IBM SPSS Statistics for Windows, Version 22 (IBM Corporation, New York, USA) was used
for statistical analysis. Since the data for the variables analysed in this study
(age and BMI of the mother, cervical dilation, duration of induction, RI values) were
not normally distributed, non-parametric statistical methods exclusively were used
for the analysis. Group comparisons were made by means of the Mann-Whitney U test
(for two groups) and the Kruskal Wallis H test in the case of more than two groups.
Correlation analyses were performed using the Spearman rank correlation coefficient.
A general linear model was used for multivariate analysis of the influence of cervical
dilation and Bishop score and of the UA and MCA resistance indices on the duration
of induction. In the first step, a model with cervical dilation or Bishop score and
the additional maternal parameters age (years), body mass index (BMI, kg/m2) and parity (primipara, multipara) was analysed. The next step was to examine whether
addition of the RIs of the UA or MCA significantly improves the model (significance
of the change of the F statistics), that is, whether addition of the RI enables significantly
better prediction of the duration of induction. The accuracy of the statistical models
was evaluated using the adjusted determination coefficient R2. The R2 value states which proportion of the observed variance in duration of induction can
be explained by the model; the adjusted R2 value takes the complexity of the underlying statistical model into account [9], [10]. All stated p values are two-sided, and the significance level was set at α = 0.05.
The study meets the standards of the Declaration of Helsinki, and approval to conduct
the study was obtained from the ethics committee of Ulm University (no. 325/13).
Results
Patient characteristics, indication and induction procedure
In this study, 21 primiparas and 28 multiparas were induced. The average BMI of the
study patients was 25.6 kg/m2 (range 18.4 – 39.7 kg/m2) and the average maternal age was 30.2 years (range 20 – 41 years). The most frequent
indications for induction were post-term pregnancy (n = 31), subjectively sluggish
foetal movements (n = 7) or a suspect CTG (n = 7) when the ultrasound was otherwise
normal, the patientʼs wish to end the pregnancy because of maternal exhaustion in
2 cases, and ultrasound suspicion of a relative disparity if the pregnancy was continued.
Pharmacological induction was with misoprostol in 29 cases (dosage: initial dose of
50 µg orally, followed by 75 µg every 4 hours), with Minprostin gel in 15 women who
had had a previous section (dosage: initial dose of 1 mg, repeated every 6 hours,
total 3 times/day), and with a combination of the two drugs in 3 cases. Misoprostol
(see above) was given first, and if there had been no essential progress in labour
after up to 10 doses, further priming was continued with Minprostin gel. Two multiparas
with an advanced Bishop score of ≥ 5 received an oxytocin infusion (dosage: 3 IU oxytocin
in 250 NaCl at an initial dosage of 10 ml/hour, increased individually). 40 women
(81.6%) had vaginal delivery (33 spontaneous, 7 assisted) and 9 women (18.4%) were
delivered by secondary caesarean section. The median Bishop score obtained immediately
before induction of labour was 4 (range 1 – 9). The median cervical dilation was 1 cm
(range 0 – 5 cm); cervical dilation of more than 2 cm was found in 2 cases (4.1%).
The mean duration of induction was 25.2 h (median 19 h, range 4.5 – 57.3 h). The maternal
parameters at the time of inclusion in this study, the RI data from the Doppler measurements
and the induction and delivery data are summarised in [Table 1].
Table 1 Patient parameters, induction data and Doppler ultrasound data at the start of induction
(n = 49).
|
Age (years)
|
Mean ± SD
|
30.2 ± 5.8
|
|
Median
|
31
|
|
Range
|
20 – 41
|
|
Body mass index (kg/m2)
|
Mean ± SD
|
25.6 ± 4.9
|
|
Median
|
24.8
|
|
Range
|
18.4 – 39.7
|
|
Parity
|
Primipara
|
28 (57.1%)
|
|
Multipara
|
21 (42.9%)
|
|
Indication for induction
|
Post-term pregnancy
|
31 (63.3%)
|
|
Sluggish foetal movements
|
7 (14.3%)
|
|
Suspect CTG
|
7 (14.3%)
|
|
Other
|
4 (8.2%)
|
|
Weeks of gestation at the start of induction
|
≤ 41 + 0
|
38 (77.6%)
|
|
> 41 + 0
|
11 (22.4%)
|
|
Induction method
|
Tablets
|
29 (59.2%)
|
|
Gel
|
15 (30.6%)
|
|
Tablets plus gel
|
3 (6.1%)
|
|
Oxytocin
|
2 (4.1%)
|
|
Cervical dilation at the start of induction (cm)
|
0
|
17 (34.7%)
|
|
1
|
21 (42.9%)
|
|
2
|
9 (18.4%)
|
|
3
|
1 (2.0%)
|
|
5
|
1 (2.0%)
|
|
Bishop score at the start of induction
|
1
|
5 (10.2%)
|
|
2
|
4 (8.2%)
|
|
3
|
10 (20.4%)
|
|
4
|
8 (16.3%)
|
|
5
|
8 (16.3%)
|
|
6
|
7 (14.3%)
|
|
7
|
1 (2.0%)
|
|
8
|
4 (8.2%)
|
|
9
|
1 (2.0%)
|
|
Missing
|
1 (2.0%)
|
|
Resistance index of the umbilical artery (MoM)
|
Mean ± SD
|
0.99 ± 0.13
|
|
Median
|
0.96
|
|
Range
|
0.76 – 1.45
|
|
Resistance index of the middle cerebral artery (MoM)
|
Mean ± SD
|
0.95 ± 0.09
|
|
Median
|
0.95
|
|
Range
|
0.76 – 1.14
|
|
Induction duration (h)
|
Mean ± SD
|
25.2 ± 15.8
|
|
Median
|
19.0
|
|
Range
|
4.5 – 57.3
|
|
Mode of delivery
|
Spontaneous
|
33 (67.3%)
|
|
Vacuum extraction
|
7 (14.3%)
|
|
Secondary section
|
9 (18.4%)
|
Associations with duration of induction
The duration of induction did not differ significantly between primiparas and multiparas
(multiparas: median 27.5 h, range 4.5 – 57.3 h; multiparas: median 16.3 h, range 8.5 – 49.0 h;
p = 0.203) and between vaginal delivery and secondary section (vaginal deliveries:
median 17.4 h, range 8.0 – 54.0 h; deliveries by secondary section: median 30.5 h,
range 4.5 – 57.3 h; p = 0.214). There were also no significant differences in the
duration of induction between the four different induction methods (p = 0.132) or
between inductions started before or after weeks 41 + 0 (p = 0.204). Neither maternal
age (rs = − 0.125, p = 0.391) nor BMI (rs = 0.135, p = 0.366) correlated significantly with the duration of induction. In all
pregnant women, there was a negative correlation between the duration of induction
and the Bishop score obtained immediately before induction of labour (rs = − 0.395, p = 0.005; [Fig. 1 a]). The initial cervical dilation, which is a single parameter in the Bishop score
that is easy to measure, likewise showed a significant negative correlation with the
duration of induction (rs = − 0.398, p = 0.005) ([Fig. 1 b]).
Fig. 1 Correlation between induction duration in hours and a Bishop score at induction of labour (rs = − 0.395, p = 0.005); b cervical dilation (cm) at induction of labour (rs = − 0.398, p = 0.005).
The higher the resistance (RI) of the MCA before the start of induction, the longer
the time from the start of induction until delivery of the baby (rs = 0.547, p < 0.001). This was the case for both primiparas (rs = 0.606, p = 0.004) and multiparas (rs = 0.505, p = 0.006; [Fig. 2 a]). A subgroup analysis according to gestation week at the start of induction showed
a significant positive association between the RI of the MCA before the start of induction
and the duration of induction for inductions started ≤ week 41 + 0 (rs = 0.593, p < 0.001), but not for inductions started >week 41 + 0 (rs = 0.274, p = 0.415; [Fig. 2 b]). Unlike MCA resistance, there was no significant correlation between the resistance
of the UA measured before induction of labour and the duration of induction (rs = − 0.055, p = 0.707). The RI of the UA and MCA measured at the start of induction
did not differ significantly between inductions where delivery was vaginal and those
in which secondary section was performed (both p > 0.3).
Fig. 2 Correlation between induction duration in hours and resistance index of the middle
cerebral artery (MoM) at induction of labour depending on a parity (primiparas: rs = 0.606, p = 0.004; multiparas: rs = 0.505, p = 0.006); b weeks of gestation at the start of induction (≤ 41 + 0: rs = 0.593, p < 0.001; > 41 + 0: rs = 0.274, p = 0.415).
A multivariable model with induction duration as dependent variable and Bishop score
together with the maternal parameters age, BMI and parity as independent factors showed
that the Bishop score had a significant effect on the duration of induction (p = 0.022).
A similar result was apparent when the model included cervical dilation instead of
the Bishop score (p = 0.018). Addition of the RI of the UA did not result in a significant
improvement either for the model with the Bishop score or for the model with cervical
dilation (both p > 0.05). By contrast, a statistical improvement resulted for both
the model with the Bishop score and for the model with cervical dilation when the
RI of the MCA was added (both p = 0.002). The model with Bishop score, RI of the MCA
and the maternal parameters age, BMI and parity together explains 27.9% of the observed
variance of the induction duration (adjusted R2 = 0.279; adjusted R2 without RI of the MCA = 0.105), while the model with cervical dilation, RI of the
MCA and the maternal parameters age, BMI and parity explains 28.3% of the observed
variance of the induction duration (adjusted R2 = 0.279; adjusted R2 without RI of the MCA = 0.114). The maternal parameters parity, age and BMI, on the
other hand, were not associated significantly with the duration of induction in any
of the models (all p > 0.05).
When the multivariable model is calculated with induction duration as dependent variable
and Bishop score or cervical dilation together with the maternal parameters age, BMI
and parity as independent factors excluding patients with start of induction in weeks
> 41 + 0, some markedly higher determination coefficients are obtained despite the
associated reduction in the sample size from 49 to 38 patients. The model with Bishop
score, RI of the MCA and the maternal parameters age, BMI and parity together explains
47.7% of the observed variance of the induction duration (adjusted R2 = 0.477; adjusted R2 without RI of the MCA = 0.242), while the model with cervical dilation, RI of the
MCA and the maternal parameters age, BMI and parity explains 45.2% of the observed
variance of the induction duration (adjusted R2 = 0.452; adjusted R2 without RI of the MCA = 0.255).
Foetal outcome
There was no significant association between the duration of induction and the foetal
outcome measured by the postnatal arterial cord pH (rs = − 0.131, p = 0.371). While no association between the resistance index of the MCA
and the foetal outcome was found (rs = 0.010, p = 0.945), there was a significant negative correlation between UA resistance
and foetal outcome (rs = − 0.287, p = 0.046). The higher the RI of the UA antenatally at the start of induction,
the lower the arterial cord pH post partum ([Fig. 3]). When patients with the start of induction in weeks > 41 + 0 were excluded, there
was still no significant association between the resistance index of the MCA and foetal
outcome measured by the postpartum arterial cord pH (rs = − 0.026, p = 0.876), while the correlation coefficient for the association between
the resistance index of the UA and the foetal outcome was increased despite the reduced
number of patients (rs = − 0.393, p = 0.015).
Fig. 3 Correlation between the resistance index of the umbilical artery (MoM) at induction
of labour and the postpartum arterial cord pH (rs = − 0.287, p = 0.046).
Discussion
The duration of pregnancy, calculated from the first day of the last menstrual period,
is approximately 40 weeks. Prolongation of pregnancy beyond the estimated date of
delivery (EDD) harbours the risk of placental insufficiency with the consequent risk
of intrauterine foetal death [11], [12].
In 2013, Stuck et al. conducted a survey of 54 perinatal centres in Germany. This
showed that misoprostol is the agent of first choice for induction of labour in 37
(68.5%) clinics provided surgery on the uterus had not been performed previously [13]. There was remarkable heterogeneity with regard to administration route, dosage,
total daily dosage and interval between individual doses and in the overall duration
of use [13]. Like most centres, we mainly used misoprostol in our study, and used Minprostin
gel if induction was unsuccessful or the patient had previously had a section. Two
patients with a Bishop score ≥ 5 before the start of induction were given an oxytocin
infusion from the outset.
Many different factors influence the success of pharmacological induction of labour
at term. Our cohort demonstrated a significant positive correlation with the duration
of induction for both the Bishop score measured prior to induction and for the individual
parameter “cervical dilation” contained in the Bishop score. This result correlates
with the available literature [14], [15], [16].
Gonen et al. showed in 1998 that parity can serve as a predictive factor for the duration
from the start of induction until delivery, in addition to the Bishop score [17]. We found a nominally shorter induction duration for multiparas compared with primiparas
in our group of patients (median 16.3 vs. 27.5 h), but the difference did not reach
significance, due to the relatively small number of patients.
Since there is often great uncertainty with regard to the success of induction of
pregnant women, especially when the cervix is still unripe, additional predictive
parameters for better assessment of success are required, especially when weighing
induction against primary caesarean section as alternative modes of delivery.
The hypothesis underlying this study was the assumption that measurement of foetal
Doppler parameters might possibly provide an addition criterion for assessing the
duration and success of induction. However, changes in resistance in the middle cerebral
artery at an early gestational age must be distinguished from those at term. The changes
in Doppler parameters known to occur with intrauterine growth retardation (“brain
sparing”; defined as RI in the umbilical artery > 90 percentile and RI in the middle
cerebral artery < 10%) are pathological and a reaction of the foetal circulation to
impending foetal hypoxia. This contrasts with the so-called “term effect” of a fall
in the RI and PI of the MCA, which is a physiological process [5]. As the foetal head descends into the motherʼs pelvis before the expected date of
delivery, there is a physiological fall in the resistance in the middle cerebral artery,
to allow increasing oxygenation of brain structures [18]. Another assumption is that this also triggers oxytocin secretion from the foetal
pituitary, which stimulates labour. This might also explain the phenomenon of premature
labour in the case of foetuses with intrauterine growth retardation, in the sense
of an autoregulation mechanism, so as to avoid intrauterine death in these pregnancies.
It would be interesting to compare the prepartum resistance of the MCA between babies
with breech presentation and babies with cephalic presentation [19].
Assessment of resistance in the umbilical artery, the estimated foetal weight, amniotic
fluid volume and biophysical profile help to distinguish the “term effect” from pathological
centralisation of the foetal circulation in the form of “brain sparing”. For foetuses
with very low resistance in the MCA and/or other abnormalities, e.g., on cardiotocography,
pharmacological induction should be regarded critically and primary caesarean section
should possibly be considered instead.
Doppler ultrasound assessment of the foetal MCA is a test that can usually be performed
readily as part of the biometric measurements of a baby prior to planned induction
of labour. In our patients, a significant positive correlation was apparent between
the RI (MoM) of the foetal MCA measured before induction and the duration of induction;
the higher the prepartum RI of the MCA, the longer the duration of induction. The
multivariable analysis adjusted for the maternal parameters age, BMI and parity showed
that cervical dilation and resistance in the MCA exert an influence on the duration
of induction independent of one another. The two variables together explain approximately
30% of the observed variance of induction duration. A similar result is apparent when
the Bishop score is used instead of cervical dilation in the multivariable analysis.
Since the German guideline on “Management of overdue and post-term pregnancy” (DGGG
2010) and the National Institute of Clinical Excellence (NICE) in its clinical guideline
on induction of labour (2008) recommend induction after 41 + 0 weeks regardless of
other maternal and/or foetal factors, we also analysed the subgroup of patients over
41 + 0 weeks. This showed a significant positive association between the RI of the
MCA before the start of induction and induction duration for inductions started ≤ 41 + 0
weeks but not for inductions started > 41 + 0 weeks. When the multivariable model
is calculated with induction duration as dependent variable and Bishop score or cervical
dilation together with the maternal parameters age, BMI and parity as independent
factors excluding patients with start of induction in weeks > 41 + 0, markedly higher
determination coefficients are obtained despite the associated reduction in the sample
size from 49 to 38 patients. The model with Bishop score, RI of the MCA and the maternal
parameters age, BMI and parity together then explains 47.7% of the observed variance
of the induction duration, while the model with cervical dilation, RI of the MCA and
the maternal parameters age, BMI and parity explains 45.2% of the observed variance
of the induction duration.
In our study, no correlation was shown between the RI of the UA and the time from
the start of induction of labour until delivery, and there were no significant differences
in the RI of the UA measured at the start of induction between inductions where delivery
was vaginal and inductions in which delivery was by secondary section. However, there
was a significant association between the RI of the UA and the arterial cord pH of
the neonate. Babies with a prepartum umbilical RI in the upper normal range showed
significantly worse arterial cord pH post partum. Since the resistance of the UA provides
information about placental function and thus about the intrauterine supply of the
foetus, increased resistance in the UA explains a relative impairment of placental
function and therefore supply to the foetus [20], [21]. Unlike the UA, this study did not find any association between the RI of the MCA
measured at the start of induction and the babyʼs postpartum arterial cord pH. This
result is consistent with studies in which an association between RI in the MCA and
the foetal outcome was found only in risk populations [22], [23].
In summary, this study shows that Doppler examination of the MCA at the start of induction
can provide additional information about the probable duration of induction in addition
to the parameters used clinically to date such as Bishop score or cervical dilation
and parity to estimate the duration and success of induction of labour at term. Doppler
measurement of the UA at the start of induction allows prognostic assessment of the
postnatal foetal outcome. The MCA and Bishop score have an independent influence on
the duration of induction, but in our study they can explain about 30% of the variation
in induction duration, but about 45% when patients whose induction is started in weeks
> 41 + 0 are excluded. However, our study is based on a small number of patients and
therefore has low statistical power.
Other limitations of the study consist of the heterogeneity of the patients with regard
to parity, indication for induction, drug employed and the fact that some of the data
were obtained retrospectively and some prospectively. However, this better reflects
the real situation in routine clinical practice than would be the case in a prospective
study with uniform induction indications and methods. Further and larger studies are
necessary, however, to further optimise the prognostic conclusion regarding induction
duration and success at term and develop an algorithm consisting of
-
Parity
-
Bishop score or cervical dilation and
-
MCA results.
Another possible study would be to compare the prepartum RI of the MCA between babies
with breech presentation and those with cephalic presentation to investigate the underlying
regulatory mechanisms. Ideally, this would deliver evidence about the course and success
of induction in pregnant women at term, which might enable the mode of delivery likely
to be successful to be planned prospectively.
Conclusion for Practice
When estimating the duration of induction at term, Doppler measurement of the middle
cerebral artery appears to be an additional prognostic factor in addition to the established
Bishop score. The resistance in the umbilical artery correlated with the foetal arterial
cord pH.
