Study design
This was a pragmatic, parallel group, open-label, superiority randomised
control trial of two protocols for labour induction among women with
hypertension of pregnancy. All women underwent initial cervical
preparation with low dose oral misoprostol; and those who required
ongoing induction after membrane rupture were randomised in a 1:1 ratio
to LDOM or intravenous oxytocin. The study protocol has been
published,10 but brief methods are outlined below.
Participants
Women planning to give birth in three government hospitals in central
India were recruited to this trial: Government Medical College, Nagpur
and Daga Memorial Women’s Hospital Nagpur, and Mahatma Gandhi Institute
of Medical Sciences, Sevagram, Wardha. Women with an indication for
induction due to hypertensive disease, irrespective of gestation, and
who required cervical preparation for an unfavourable cervix (Bishop’s
score ≤ 6) were recruited prior to the start of induction. All
consenting women then underwent cervical preparation with LDOM 25mcg 2
hourly. However, only those who subsequently required augmentation
following artificial rupture of membranes were randomised. Most did not
have prior ultrasound unless growth restriction was suspected clinically
prior to the development of hypertension or if the onset was before 34
weeks’ gestation. Those with a previous caesarean section, age
<18 years old, known intrauterine fetal death, or multiple
pregnancy were excluded.
Randomisation
and masking
Potential participants were informed of the study through posters in the
antenatal areas and labour ward.
Once
a clinical decision was made for induction, the woman was provided with
an own-language information sheet and a brief slide presentation to view
on a tablet. If she was unable to read, a member of the research team
read the forms to her in her own language in the presence of family
members and/or friends. If she wished to participate, then she signed
(or placed a thumb print) on the consent form. If she then required
ongoing induction after membrane rupture as part of the induction
process, she was randomised to receive misoprostol or oxytocin without
the need for additional written consent.
For randomisation, the next consecutive, sequentially numbered opaque
envelope containing the allocation was drawn from the trial dispenser by
the research assistant and opened. The treatment allocation was
generated independently using a computerised pseudo-random number
generator, stratified by centre with random block sizes of 4, 6 and 8.
Neither participant, researcher nor clinical team were blinded to the
allocated treatment.
Procedures
Induction for all women commenced with cervical preparation using 25 mcg
oral misoprostol tablets (Cipla Ltd, Goa, India) every 2 hours. Once
painful contractions started, labour monitoring commenced with
assessments every 30 minutes with a vaginal examination every 4 hours.
The next dose of oral misoprostol was omitted when 3 or more moderate or
strong contractions occurred every 10 minutes. Artificial rupture of
membranes (ARM) was performed as per routine clinical practice when the
cervix was 2cm dilated. If spontaneous rupture of membranes occurred
before that point, then the cervical preparation doses of misoprostol
were stopped. If contractions were inadequate, then LDOM could be
restarted as part of the randomised trial.
After membrane rupture, if contractions continued at 3 in 10 or more and
there was progressive cervical change (defined as at least 1cm every 2
hrs) then no further LDOM was used, and the participants were not
randomised. If, however, the contractions reduced to less than 3 in 10
or if there was no progressive cervical change, then the woman was
randomised to either continued LDOM or an oxytocin infusion. This was
the point of trial entry. If the cervix was still not favourable for ARM
after 24 hours of cervical preparation, then the decision regarding the
ongoing management was by the clinical team but was usually a caesarean
birth.
The misoprostol 25mcg was given orally every 2 hours in the absence of
adequate uterine activity. There was no titration of the misoprostol
dose, but the next dose was withheld in the presence of regular uterine
activity and only restarted if contractions became inadequate or if
there was inadequate cervical change (<1cm every 2 hours). For
the oxytocin infusion, 5IU oxytocin (Pfizer Limited, Nani Daman, India)
in 500 mL of Ringer’s lactate was given through an electronic infusion
pump at a rate of 2 mU/min, and increased every 30 min by 2 mU/min to a
maximum of 20mU/min until there were 3-4 contractions every 10 min. If
there was any suspicion of fetal distress due to excessive uterine
activity, then the oxytocin infusion was stopped, and the participant
was put in a left lateral position and continuous electronic fetal heart
rate monitoring started.
Maternal and fetal monitoring was conducted on a one-to-one basis by
graduate research assistants, specifically trained on fetal monitoring
and uterine contraction strength. Intermittent electronic fetal
monitoring was done every 30 minutes with continuous electronic fetal
monitoring in case of abnormality. Continuous fetal monitoring was also
used routinely in high-risk women when available. In case of
hyperstimulation, staff were instructed to commence electronic fetal
monitoring and, in the event of abnormality, reduce the dose of
oxytocin.
Outcomes
Outcomes were based on the Cochrane Collaboration induction of labour
generic protocol11 and the induction of labour core
outcome set.12 The primary outcome was caesarean
birth. Secondary outcomes addressed the success of the induction
process, maternal mortality and morbidity, and neonatal morbidity and
mortality. Measures of success included the randomisation to birth
interval, duration of hospital stay and satisfaction. Data was collected
using REDCap (Vanderbilt University, Tennessee, USA). A qualitative
study, situational analysis and health economic analysis were also
conducted and will be reported separately.
Sample
size calculation
In a previous study of labour induction conducted in this population,
157 (52%) women required uterine stimulation after membrane rupture
with intravenous oxytocin (standard practice), of whom 49 (31%) had a
caesarean birth.12 In a systematic review of LDOM,
those whose induction was continued after membrane rupture with LDOM had
42% less caesarean sections than those who changed to oxytocin (15% vs
26%).5 Using this data, it was estimated that a total
sample size of 520 women would provide (a) 90% power to detect a
reduction in the CS rate from 31% to 18.5% (RR 0.6), or (b) 80% power
to detect a reduction in the CS rate from 31% to 20% (RR 0.65) in
those women who receive LDOM (superiority; two sided α=0.05). It was
proposed to approach and gain consent from 1000 women of whom an
estimated 520 would require ongoing induction after cervical
preparation. At the point of requiring uterotonics for ongoing
induction, consented women would be randomized to either a protocol of
continued LDOM (n=260) or the standard oxytocin infusion (n=260).