Keywords
induction of labor - cervical ripening - misoprostol - Foley - quality improvement
Induction of labor is performed in more than 20% of all pregnancies in the United
States.[1] In the setting of an unfavorable cervix, labor induction is more successful when
the cervix is ripened with pharmacological methods (e.g., administration of misoprostol,
a synthetic prostaglandin) and/or through mechanical methods (e.g., placement of the
cervical Foley).[2] Misoprostol alters collagen breakdown, rearranges collagen fibers, and increases
intracellular calcium to promote cervical softening and myometrial contractions.[3]
[4] Mechanical methods appear to have two mechanisms of action: exerting pressure on
the internal cervical os promotes both mechanical dilation and release of endogenous
prostaglandins from the adjacent decidua (Ferguson's reflex).[3]
[4]
[5]
[6]
[7]
Prior studies have shown that as single agents, the cervical Foley and misoprostol
are both effective methods of cervical ripening; however, it is not clear whether
one is superior to the other.[1]
[8]
[9] Sequential use of oral misoprostol and the cervical Foley does not conclusively
appear to improve time to delivery or vaginal delivery rates in comparison to simply
using oral misoprostol alone.[7]
[10]
[11] However, the simultaneous use of oral misoprostol and the cervical Foley has consistently
been demonstrated to shorten time to delivery and decrease the risk of failed induction
of labor (IOL) in comparison to misoprostol alone. The best example of this is the
FOR MOMI trial, a large randomized controlled trial, that investigated these methods
as single agents and also in dual use in a four-arm study design.[8] They compared cervical ripening prior to induction with the cervical Foley alone,
misoprostol alone, cervical Foley and vaginal misoprostol, or cervical Foley and oxytocin.
The FOR MOMI trial demonstrated that simultaneous use of the cervical Foley and low-dose
vaginal misoprostol resulted in a shorter median time to delivery (∼4 hours) when
compared with either method alone, with no change in mode of delivery or harms to
mother or neonate.[8] These encouraging results prompted many providers to adopt dual agent ripening for
IOL. However, most of the data on dual-agent cervical ripening are randomized controlled
trials that reflect study efficacy rather than effectiveness. This study aims to address
effectiveness in the real world, namely, how implementation of an evidence-based protocol
impacted the time to delivery in two historical cohorts pre- and postimplementation
of low-dose misoprostol with the concurrent cervical Foley.
Methods
This was a retrospective before-and-after study of all nulliparous women with term
(≥37 weeks) singleton pregnancies who underwent a scheduled IOL at the University
of California San Francisco (UCSF) from May 2015 to April 2017 and from June 2017
to May 2019. Those with preterm delivery of <37 weeks, multifetal gestation, nonvertex
fetal presentation, multiparity, prior cesarean delivery, rupture of membranes prior
to initiation of IOL, and major fetal anomalies were excluded. In May 2017, the term
IOL cervical ripening protocol was changed from misoprostol 50 to 100 µg PO every
4 hours with subsequent placement of the cervical Foley once cervical dilation was
1 to 2 cm (the preimplementation protocol, based on a 2005 prospective IOL trial[12]) to misoprostol 25 µg PO every 2 hours with simultaneous placement of the cervical
Foley filled to 60 mL (the postimplementation protocol, inspired in large part by
the results of the FOR MOMI trial).[2] The decision to use 25-µg oral misoprostol administered every 2 hours (rather than
25-µg vaginal misoprostol administered every 3 hours, as per the original FOR MOMI
protocol) was based on several lines of evidence. Pharmacokinetic data indicate that
oral absorption of misoprostol is highly reliable, with minimal variability between
patients, compared with vaginal absorption.[13]
[14] However, the half-life is much shorter, thus requiring more frequent administration
to maintain serum levels.[13]
[14] The American College of Obstetricians and Gynecologists (ACOG) provides minimal
guidance surrounding the best dosing schedule for misoprostol administration.[1] However, the World Health Organization (WHO),[15] International Federation of Gynecology and Obstetrics (FIGO),[16] and the Royal College of Obstetricians and Gynaecologists (RCOG),[17] all recommend low dose but frequently administrated misoprostol, based on systematic
reviews showing this strategy to be both effective and superior in terms of safety.
For all deliveries at UCSF, details regarding maternal demographics, labor characteristics,
and pregnancy outcomes are collected at the time of delivery by the managing clinicians
and stored within the UCSF Perinatal Database. Chart review is performed by trained
abstractors to ensure complete and accurate information, while monthly review of the
database is performed by physicians for quality assurance. We utilized the UCSF Perinatal
Database to obtain information regarding maternal demographics (age, race, body mass
index [with obesity defined as body mass index ≥ 30 kg/m2], and parity), details of antepartum and intrapartum course (pregnancy-related hypertension,
chronic hypertension, gestational diabetes, preexisting diabetes, maternal cardiac
disease, fetal growth restriction, placental abruption, gestational age at delivery,
and mode of delivery) and neonatal characteristics (birthweight, Apgar's scores, and
neonatal intensive care unit [NICU] admission).
The primary outcome was time from first misoprostol dose to delivery. Secondary outcomes
included overall cesarean delivery rate, cesarean delivery for fetal indications,
cumulative misoprostol dose, number of misoprostol administrations, need for oxytocin
augmentation, time from oxytocin initiation to delivery, postpartum hemorrhage, chorioamnionitis,
low Apgar's score (defined as 5-minute Apgar's score < 7), and NICU admission.
The Kruskal–Wallis test was used to compare median values for nonparametric continuous
variables, and Fisher's exact and Chi-squared tests were used to compare proportions
for categorical variables. Statistical analyses were performed using STATA v13.0 (StataCorp.,
College Station, TX). This study was approved by the UCSF Committee on Human Research
(study number 19–28095).
Results
During the study period, a total of 1,496 women met criteria for inclusion, with 698
women in the preimplementation group and 798 in the postimplementation group. There
were no statistically significant differences in maternal demographics between the
two groups ([Table 1]). There was a higher incidence of pregnancy-related hypertension in the preimplementation
group compared with the postimplementation group (36 vs. 28%, p < 0.001); there were no differences in other pregnancy complications or in gestational
age at delivery between the two groups.
Table 1
Maternal demographics and characteristics among nulliparous patients undergoing induction
of labor before (preimplementation) and after (postimplementation) the implementation
of an evidence-based protocol for cervical ripening
|
Demographics
|
Preimplementation (n = 698)
|
Postimplementation (n = 798)
|
p-Value
|
|
Maternal age (y)
|
33 (30–37)
|
34 (31–37)
|
0.210
|
|
Race
|
|
Caucasian
|
354 (51%)
|
431 (54%)
|
0.757
|
|
African American
|
33 (5%)
|
31 (4%)
|
|
|
Hispanic
|
63 (9%)
|
67 (8%)
|
|
|
Asian
|
161 (23%)
|
182 (23%)
|
|
|
Other
|
78 (11%)
|
80 (10%)
|
|
|
Unknown
|
9 (1%)
|
7 (1%)
|
|
|
Obese
|
97 (14%)
|
95 (12%)
|
0.250
|
|
Chronic hypertension
|
49 (7%)
|
49 (6%)
|
0.493
|
|
Pregnancy-induced hypertension
|
256 (37%)
|
225 (28%)
|
<0.001
|
|
Preexisting diabetes
|
11 (2%)
|
14 (2%)
|
0.788
|
|
Gestational diabetes
|
185 (27%)
|
205 (26%)
|
0.720
|
|
Fetal growth restriction
|
28 (3%)
|
27 (3%)
|
0.567
|
|
Placental abruption
|
11 (2%)
|
16 (2%)
|
0.534
|
|
Gestational age at delivery
|
39.9 (38.7–41.1)
|
39.9 (39.0–41.0)
|
0.882
|
Note: Continuous variables with nonparametric distribution reported as median (interquartile
range).
There was no difference in the primary outcome of time to delivery between the two
groups (median of 29 hours for the preimplementation group vs. 30 hours in the postimplementation
group, p = 0.68; [Table 2]). The preimplementation group had a significantly higher cumulative misoprostol
dose than the postimplementation group (150 vs. 100 µg, p < 0.001), while the postimplementation group had significantly more misoprostol administrations
(4 vs. 2 doses, p < 0.001). A smaller proportion of the preimplementation group required oxytocin augmentation
when compared with the postimplementation group (81 vs. 88%, p = 0.001; [Table 2]).
Table 2
Pregnancy outcomes among nulliparous patients undergoing induction of labor before
(preimplementation) and after (postimplementation) the implementation of an evidence-based
protocol for cervical ripening
|
Demographics
|
Preimplementation (n = 698)
|
Postimplementation (n = 798)
|
p-Value
|
|
Time from misoprostol start to delivery (h)
|
29.1 (20.5–41.7)
|
29.9 (20.1–40.5)
|
0.689
|
|
Cumulative misoprostol dose (µg)
|
150 (50–250)
|
100 (50–150)
|
0.0001
|
|
Total number of misoprostol administrations
|
2 (1–3)
|
4 (2–6)
|
0.0001
|
|
Required oxytocin for augmentation
|
568 (81%)
|
700 (88%)
|
0.001
|
|
Time from Pitocin start to delivery (h)[a]
|
16 (9.27–22.3)
|
15 (9.70–21.78)
|
0.817
|
|
Cesarean delivery for any indication
|
208 (30%)
|
206 (26%)
|
0.086
|
|
Cesarean delivery for fetal indication
|
74 (11%)
|
67 (8%)
|
0.145
|
|
Postpartum hemorrhage
|
74 (11%)
|
119 (15%)
|
0.013
|
|
Blood transfusion
|
41 (6%)
|
49 (6%)
|
0.829
|
|
Chorioamnionitis
|
70 (10%)
|
94 (12%)
|
0.280
|
|
Low Apgar's score
|
29 (4%)
|
50 (6%)
|
0.067
|
|
NICU admission
|
117 (17%)
|
76 (10%)
|
<0.001
|
Abbreviation: NICU, neonatal intensive care unit.
Note: Continuous variables with nonparametric distribution reported as median (interquartile
range).
a
n = 568 for preimplementation group and n = 700 for postimplementation group.
There was no statistical difference in cesarean deliveries overall or cesarean deliveries
for fetal indication (preimplementation of 30% and postimplementation of 26% [p = 0.086] for any indication; 8 vs. 11% [p = 0.145] for fetal indication). Multivariate logistic regression to control for potential
confounders (maternal age, preexisting diabetes mellitus, gestational diabetes mellitus,
chronic hypertension, and pregnancy-related hypertensive disease) failed to find a
statistically significant difference in cesarean delivery (adjusted odds ratio [aOR] = 0.80,
95% confidence interval [CI]: 0.63–1.00) or cesarean delivery for fetal indication
(aOR = 0.74, 95% CI: 0.52–1.05) between the two groups.
The postimplementation group had a significantly higher postpartum hemorrhage rate
than the preimplementation group (15 vs. 11%, p = 0.013); however, there was no statistically significant difference in the rate
of postpartum blood transfusion between the two groups (6 vs. 6%, p = 0.83). There was no difference in the rate of low Apgar's scores between the two
groups, but there was as a significantly higher rate of NICU admission in the preimplementation
group compared with the postimplementation group (17 vs. 10%, p < 0.001; [Table 2]).
Discussion
We found that implementation of an evidence-based concurrent dual agent cervical ripening
protocol did not change the median time to delivery or likelihood of vaginal delivery.
Our study expands on the FOR MOMI trial by examining the real-world effects of clinically
implementing a protocol of low-dose misoprostol with the concurrent cervical Foley
placement.[8] Unlike the FOR MOMI trial, we did not find a decrease in the median time to delivery
with the use of this protocol. Importantly, the median time to delivery in the FOR
MOMI trial for the misoprostol with the concurrent cervical Foley group was 13 hours,
whereas the median time to delivery in the post implementation group in our study
was 30 hours. This discrepancy may be due numerous factors. First, the FOR MOMI trial
included both multiparous and nulliparous patients, whereas our study included only
nulliparous patients, who have longer induction to delivery times.[3]
[11] Second, the cesarean rate for nulliparous woman in the FOR MOMI trial receiving
dual- agent was 44%, compared with 26% in our population. Our lower cesarean rate
likely reflects differences in the institutional management of labor and may be due
to fewer cesareans for failed induction. Lastly, the variation in route of misoprostol
administration (vaginal in FOR MOMI, oral in the present study) may have also contributed
to the differences in time to delivery. A recent meta-analysis demonstrated that among
all prostaglandins, low-dose (<50 µg) oral misoprostol had the lowest probability
of cesarean delivery over time, whereas vaginal misoprostol had the highest probability
of achieving a vaginal delivery within 24 hours.[18]
[19] This finding aligns with our report of low cesarean delivery rates in the postimplementation
(low-dose oral misoprostol) group but longer time to delivery than the FOR MOMI trial,
which used the vaginal route.
This study also demonstrated that the postimplementation group had a significantly
higher rate of postpartum hemorrhage, which is likely due to the implementation of
quantitative blood loss (QBL) measurement during the postimplementation period, rather
than a true increase in blood loss at delivery. QBL measurements are more accurate
than estimated blood loss (EBL) measurements, which tend to underestimate hemorrhage.[20] This is supported by the stable rate of transfusions across the time periods.
We also found that the postimplementation group had significantly fewer NICU admissions,
a result that is congruent with multiple studies that used similar doses of misoprostol
in combination with the cervical Foley.[21]
[22] A recent meta-analysis of eleven randomized controlled trials found that the combination
of misoprostol and the cervical Foley was associated with a reduced risk of NICU admission
(relative risk [RR] = 0.75, 95% CI: 0.58–0.97, p = 0.03) compared with either method alone.[21] To our knowledge, there were no other significant changes in NICU admission criteria
to explain this finding.
Strengths and Limitations
Strengths and Limitations
The major strength of this study is the evaluation of the implementation of an evidence-based
cervical ripening method into real-world practice, with a sample size six times larger
than the original FOR MOMI trial, which included 123 women in the misoprostol plus
the cervical Foley group. Furthermore, our institutional cesarean delivery rate is
much lower, likely reflecting different practice patterns and strategies of labor
management. Institutions may need to compare their cesarean delivery rates to those
reported in randomized controlled trials of induction methods to determine whether
those data will be generalizable to their own populations.
The major limitation of this study is its retrospective nature, which relies on historical
cohorts and is limited to what has previously been documented in the electronic medical
record. Of note, the indication for NICU admission was not included in the data analysis.
Furthermore, the comparison included changes made in dosing of oral misoprostol and
in timing of the cervical Foley placement, which limits our ability to determine which
changes contributed to the differences we found. Additionally, while the low-dose
misoprostol with the concurrent cervical Foley protocol aimed to have the simultaneous
cervical Foley and misoprostol, many women still received sequential methods if providers
felt that the cervical Foley placement would be difficult; we do not have clear data
on the frequency of this occurrence.
Conclusion
Based on this study and the existing literature, no method of cervical ripening is
clearly superior to the others outside of the context of a randomized controlled trial;
even then, the differences between the methods are the modest at the best (∼2–4 hours).
As such, we recommend that undergoing cervical ripening for IOL be counseled in a
shared decision-making model. This counseling should include the beneficial evidence
of dual agent cervical ripening over individual methods, as well as the differences
in timing and frequency of misoprostol administration, timing of the cervical Foley
placement, and the impacts on time to delivery and cesarean delivery. Shared decision
making in maternity care can decrease decisional conflict, increase satisfaction with
the delivery experience, generate more positive feelings toward the newborn, and decrease
depressive symptoms.[23]
[24]
In conclusion, implementing an evidence-based protocol of low-dose misoprostol with
the concurrent cervical Foley was not associated with a change in median time to delivery
or vaginal delivery rates. While randomized controlled trial data are useful for calculating
efficacy of induction methods, real-world effectiveness of these interventions will
vary based on local and institutional practice and should be considered when choosing
an induction method.
