bmj€¦ · confidential: for review only 2 21 abstract 22 objectives postterm pregnancy (≥42...
TRANSCRIPT
Confidential: For Review OnlyInduction of labour at 41 weeks versus expectant
management until 42 weeks (INDEX): a multicentre, open-label, randomised non-inferiority trial
Journal: BMJ
Manuscript ID BMJ-2018-047580
Article Type: Research
BMJ Journal: BMJ
Date Submitted by the Author: 19-Oct-2018
Complete List of Authors: Keulen, Judit; Amsterdam UMC, University of Amsterdam, Obstetrics ang Gynaecology Meibergdreef 9 1105 AZBruinsma, Aafke; Amsterdam UMC, University of Amsterdam, Obstetrics and GynaecologyKortekaas, Joep; Radboud University Medical Center, Obstetrics & GynaecologyVan Dillen, Jeroen; Radboud University Medical Center, ObstetricsBossuyt, Patrick; Academic Medical Center; University of Amsterdam, Dept. Clinical Epidemiology and BiostatisticsOudijk, Martijn; Amsterdam UMC, University of Amsterdam, Obstetrics ang Gynaecology Meibergdreef 9 1105 AZDuijnhoven, Ruben; Amsterdam UMC, University of Amsterdam, Meibergdreef 9 1105 AZ, Medical InformaticsKaam, Anton; Emma Childeren's Hospital, Amsterdam UMC, University of Amsterdam, NeonatologyVandenbussche, Frank; Radboud University Medical Center, Obstetricsvan der Post, Joris; Amsterdam UMC, University of Amsterdam, Obstetrics ang Gynaecology Meibergdreef 9 1105 AZMol, Ben; Monash University; Claytonde Miranda, Esteriek; Amsterdam UMC, University of Amsterdam, Obstetrics ang Gynaecology Meibergdreef 9 1105 AZ
Keywords:postterm pregnancy, induction of labour, expectant management, perinatal mortality, meconium aspiration syndrome, caesarean section, prolonged pregnancy
https://mc.manuscriptcentral.com/bmj
BMJ
Confidential: For Review Only
1
1 Title page2 Induction of labour at 41 weeks versus expectant management until 42 weeks (INDEX): a
3 multicentre, open-label, randomised non-inferiority trial
4
5 Keulen JKJa, Bruinsma Aa, JC Kortekaasb (3 first authors), van Dillen Jb, Bossuyt PMMc, Oudijk MAa,d, Duijnhoven
6 RGc, van Kaam AHe, Vandenbussche FPHAb, vd Post JAMa, Mol BWf, de Miranda Ea.
7
8 a Amsterdam UMC, University of Amsterdam, Department of Obstetrics and Gynaecology, The Netherlands, b
9 Radboud University Medical Center, Nijmegen, Department of Obstetrics & Gynaecology, Nijmegen, The
10 Netherlands, c Amsterdam UMC, University of Amsterdam, Department of Medical Informatics, The Netherlands, d
11 Department of Obstetrics and Gynaecology, University Medical Center, Utrecht, the Netherlands e Emma Children’s
12 Hospital, Amsterdam UMC, University of Amsterdam, Department of Neonatology, The Netherlands, f Department
13 of Obstetrics and Gynaecology; Monash University; Clayton, Victoria; Australia.
14 Corresponding author:
15 Esteriek de Miranda, [email protected]
16 Amsterdam UMC, University of Amsterdam, Department of Obstetrics and Gynaecology, Meibergdreef 9, 1105 AZ
17 Amsterdam
18 Telephone:+31-6-55532455
19
20 Word count: 5.430
Page 1 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
2
21 ABSTRACT 22 Objectives Postterm pregnancy (≥42 weeks) is associated with an increased risk of adverse perinatal outcomes.
23 While induction of labour (IOL) at 41 weeks is standard of care in some parts of the world, in other parts this is still
24 42 weeks. In view of the limited comparisons of these policies, we directly compared IOL at 41 weeks to IOL at 42
25 weeks in low-risk pregnant women.
26 Design Open-label, randomised controlled non-inferiority trial. The null hypothesis (EM is inferior to IOL) was
27 tested with a non-inferiority margin of 2%.
28 Setting 123 primary care midwifery practices and 45 hospitals (secondary care)
29 Participants Low risk women with an uncomplicated pregnancy
30 Interventions IOL at 41 weeks or expectant management (EM) until 42 weeks.
31 Primary and secondary outcome measures Primary outcome was a composite of perinatal mortality and neonatal
32 morbidity (Apgar 5’<7, arterial pH<7.05, meconium aspiration syndrome (MAS), plexus brachialis injury and NICU
33 admission). Secondary outcomes included maternal outcomes and mode of delivery.
34 Results Between 2012 and 2016 we randomised 900 women to IOL and 901 women to EM till 42 weeks. Median
35 gestational age at delivery was 41+0 weeks (IQR: 41+0 – 41+1) for IOL versus 41+2 weeks (IQR: 41+0 – 41+5) for
36 EM. In the IOL group, 15/900 (1.7%) participants had an adverse perinatal outcome versus 28/901 (3.1%) in the EM
37 (absolute risk difference -1.4%; 95% CI: -2.9% to 0.0%, non-inferiority p-value = 0.22). Contribution of Apgar 5’<7
38 was 11/900 (1.2%) for IOL and 23/901 (2.6%) for EM (RR: 0.48; 95% CI: 0.23 to 0.98). Fetal death occurred once
39 with IOL (0.1%) and twice with EM (0.2%), NICU admission occurred in 3/900 (0.3%) neonates after IOL versus
40 8/901 (0.9%) after EM (RR: 0.38; 95% CI: 0.10 to 1.41). There was no significant difference in composite adverse
41 maternal outcome (IOL 122/900 (13.6%) versus EM 102/901 (11.3%)) or in Caesarean section rate (both 10.8%:
42 IOL 97/900 EM 97/901).
43 Conclusions In women with uncomplicated pregnancies at 41 weeks, we could not demonstrate non-inferiority of
44 EM; instead we found a small significant difference in the risk of adverse perinatal outcome in favour of induction of
45 labour, though the chances of a good perinatal outcome were high with both strategies and the incidence of perinatal
46 mortality and NICU admission low. These findings can guide pregnant women at 41 weeks in their decision whether
47 to continue pregnancy until 42 weeks or not.
48 Trial registration number NTR3431
Page 2 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
3
49
Page 3 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
4
50 Strengths and limitations of this study
51 This concerns the largest trial until now on the 41-42 weeks dilemma
52 This research is relevant to patients and clinicians because of its potential impact on daily practice and
53 counselling of low risk women reaching 41 weeks of pregnancy.
54 This study had been performed in both midwifery and obstetrician led care
55 A limitation of this trial is the large number of women who refused to participate due to their preference for
56 a policy of expectant management until 42 weeks or induction of labour,
57
58 BACKGROUND
59 Rationale
60 Postterm pregnancy, defined as a pregnancy extended to or beyond 42+0 weeks or ≥ 294 days, is associated with
61 increased perinatal morbidity and mortality 1-10. Induction of labour (IOL) is therefore recommended after 42+0
62 weeks by the World Health Organisation and various guidelines throughout the world 10-15. Though in high resource
63 settings the overall probability of good perinatal outcome between 40 and 42 weeks is high, the risk of adverse
64 perinatal outcome increases gradually after 40 weeks 16-19.
65 Several studies have concluded that IOL from 41 weeks onwards improves perinatal outcomes, and this has
66 been confirmed in a meta-analysis 16 17 19 20. However, these results need to be interpreted with caution because of
67 trials heterogeneity caused by different outcome measures, protocols and time frames of comparison and because a
68 number of trials have compared IOL beyond 41 weeks or starting at 42 weeks, with EM far beyond 42 weeks21
69 The obstetric management of women with a pregnancy exceeding 41 weeks varies considerably between
70 and within countries. While IOL at 41 weeks has now become an accepted policy in many countries, there is no
71 consensus regarding the timing of induction in late term pregnancy in some other countries. In Sweden and the
72 Netherlands for example, EM until 42 weeks is considered standard of care in women with an uncomplicated
73 pregnancy15,22. In Norway induction should be started no later than 42+0 weeks and in Denmark delivery should take
74 place before 42+0. The RCOG/NICE guidelines recommend that women should be offered induction between 41+0
75 and 42+0 weeks 23.
76 Objective and hypothesis
Page 4 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
5
77 The aim of this study was to compare the two strategies. We anticipated that a policy of EM until 42(+0) weeks with
78 subsequent induction if necessary, being the simpler strategy, would be acceptable for a low risk population if it did
79 not lead to a substantially higher proportion of women with adverse perinatal outcomes, compared to IOL at
80 41(+0/+1) weeks.
81
82 METHODS
83 Study design
84 Since both induction of labour at 41 weeks, as well as expectant management until 42 weeks are practised in the
85 Netherlands the study was designed to investigate non-inferiority of expectant management. We conducted a
86 multicentre, open label, randomised controlled non-inferiority trial to investigate the effect of INDuction of labour at
87 41 weeks gestation to a policy of EXpectant management until 42 weeks on adverse perinatal outcomes (INDEX
88 study). Participants were recruited at 123 primary care midwifery practices and 45 hospitals (secondary care) equally
89 distributed across The Netherlands. From these 45 hospitals, 26 actively recruited participants and 19 supported the
90 study by inducing labour in women who had been recruited in a primary care setting and were allocated to induction.
91 This trial was approved by the ethics committee of the Academic Medical Centre, Amsterdam (no.
92 NL38455.018.11). The board of directors of each of the participating centres approved local execution of the study.
93 The trial has been registered in the Netherlands Trial register (no. NTR3431). We have published our study protocol
94 earlier 24. The study was performed within the Dutch Consortium for Healthcare Evaluation and Research in
95 Obstetrics and Gynaecology in cooperation with the Midwifery Research Network Netherlands.
96
97 Participants
98 Low risk women with an uncomplicated singleton pregnancy with a child in stable cephalic position at a certain
99 gestational age of 40+5 to 41+0 weeks without contra-indications for EM until 42 weeks were eligible. Gestational
100 age had to be determined by ultrasound before reaching a gestational age of 16 weeks. Women younger than 18
101 years, with ruptured membranes and/or in labour, non-reassuring fetal status (e.g. no fetal movements, abnormal fetal
102 heart rate and/or expected intra-uterine growth restriction), known fetal abnormalities (including abnormal
103 karyotype) which could influence perinatal outcome, contra indications for induction (including previous Caesarean
104 section) or contra indications for expectant management (e.g. pregnancy induced hypertension) were not eligible.
Page 5 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
6
105
106 Randomisation and masking
107 Eligible women were informed about the study at their 40 weeks antenatal check. At the next visit (40+6/41+0)
108 women were counselled by their community midwife, secondary obstetric care giver or research nurse/midwife of
109 participating centres collaborating in the Dutch Obstetric Research Consortium. After written informed consent was
110 obtained, digital vaginal examination was performed in study participants to determine the Bishop score. Sweeping
111 of the membranes was optional. Participating women were randomly allocated by a web-based program (ALEA)
112 using randomly permuted block sizes of 4 and 2, stratified by centre to IOL at 41+0-1 weeks or to EM with IOL at
113 42+0 weeks. Due to the nature of the intervention it was not possible to blind the participating women or their
114 caregivers for treatment allocation.
115
116 Procedures
117 Women allocated to IOL were scheduled for in hospital induction at 41+0-1 weeks. All women were primed and/or
118 induced according to local protocols. Women with a Bishop score <6 received cervical priming with PGE1
119 (Misoprostol, oral or vaginal), PGE2 (Dinoprostone), Foley catheter or double balloon catheter, or a combination
120 until amniotomy could be performed. Amniotomy was followed by intravenous oxytocin if required.
121 Women allocated to EM awaited spontaneous onset of labour until 42+0 weeks in midwife-led care or in a
122 hospital setting depending on the setting in which they were randomised, with monitoring according to local
123 protocol. Monitoring typically involved a combination of CTG and sonographic assessment of amniotic fluid in
124 secondary care at or between 41 and 42 weeks. Women in the EM group with ongoing pregnancies were scheduled
125 for IOL at 42+0 weeks in secondary care, following a similar IOL protocol as the intervention group.
126 In both groups labour was induced in case the maternal or fetal condition was no longer reassuring as
127 apparent from e.g. reduced fetal movements, non-optimal CTG or oligohydramnios. If prelabour rupture of
128 membranes for >24 hours or meconium stained amniotic fluid occurred, labour was also induced.
129 Relevant information regarding perinatal and maternal condition, as well as protocol deviations and their
130 reasons, were systematically collected by women’s caregivers. Every CRF was checked on completion and
131 inconsistency by the INDEX-team. Data were entered by trained data entry-staff in an online digital CRF (Oracle
132 Clinical, version 4.6.6.4.1). All adverse perinatal and maternal outcomes were checked by collecting anonymised
Page 6 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
7
133 source documents at the midwifery practice and/or hospital. All serious adverse events (SAE) were reported on a
134 case by case basis to an independent Data Safety and Monitoring Board (DSMB) and to the Dutch national internet
135 portal for the submission, review and disclosure of medical-scientific research with subjects (ToetsingOnline) 25.
136
137 Outcomes
138 The primary outcome was a composite of perinatal mortality and neonatal morbidity. Perinatal mortality was defined
139 as fetal death, intrapartum death, and neonatal death until 28 days. Neonatal morbidity was defined as having a 5
140 minutes Apgar score <7 (Apgar 5’<7), an arterial umbilical cord pH<7.05, meconium aspiration syndrome (MAS),
141 plexus brachialis injury or admission to a Neonatal Intensive Care Unit (NICU). MAS was defined as respiratory
142 distress after birth in presence of meconium stained amniotic fluid. All NICU admissions were reviewed on
143 diagnosis and presence of congenital anomaly.
144 Secondary perinatal outcomes consisted of admission to medium care, diagnosed congenital abnormality,
145 hypoglycaemia, neonatal infection or sepsis and small for gestational age (SGA, <p10) or large for gestational age
146 (LGA, >p90). Secondary maternal outcomes consisted of a composite of adverse maternal outcome and other
147 delivery outcomes. The composite adverse maternal outcome included post-partum haemorrhage (≥1000 ml (PPH),
148 manual removal of the placenta, third or fourth degree perineal tear (Obstetrical Anal Sphincter Injuries, OASIS) and
149 intensive care (ICU) admission. Other delivery outcomes concerned onset of labour, pain treatment during labour,
150 use of tocolytics, maternal intrapartum infection, meconium-stained amniotic fluid, gestational age at delivery, mode
151 of delivery, episiotomy, total postpartum blood loss, and blood transfusion.
152 For both the perinatal and the maternal composite outcomes we also compared the individual components.
153
154 Statistical analysis
155 Using data on adverse perinatal outcomes in the Netherlands from the Perined registry (https://www.perined.nl/), an
156 incidence of 3% of the primary composite outcome was expected with both strategies. The non-inferiority margin (∆)
157 was defined as a 2% risk difference in incidence, favouring induction. With a one-sided α of 0.05, a power (β) of
158 over 0.80 could be achieved if 900 patients were recruited in each of the arms (1800 patients in total). Non-inferiority
159 would be concluded if the lower 95% CI of the risk difference excluded a 2% higher proportion of women with an
160 adverse perinatal outcome in the group allocated to expectant management. A DSMB was established to review the
Page 7 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
8
161 accumulating data of the trial. Interim analyses were conducted on safety after 517, and 1088 patients had been
162 recruited.
163 The statistician who performed the analyses was blinded for the allocation of the participants and performed
164 the analysis according to a predefined analysis plan. The analysis of the primary outcome was done both for the
165 intention-to-treat and per-protocol data. For the per-protocol analysis, all randomised women with start of cervical
166 ripening or spontaneous onset of labour ≥41+0 weeks were selected. Subsequently, the per-protocol IOL group
167 (ppIOL) was defined as all women allocated to induction who received induction of labour before 41+2 or who had a
168 spontaneous onset of labour before induction could be started (<41+2). The per-protocol EM group (ppEM) included
169 all women allocated to EM with spontaneous onset of labour until 42+0 weeks, all women with a medical reason for
170 IOL before 42+0 weeks during EM and women with IOL ≥42+0.
171 Relative risks or median/mean differences with 95% confidence intervals were estimated for all outcomes.
172 Chi-squared, Fisher’s exact, t-test or Mann-Whitney U test statistics were used to investigate significance as
173 appropriate. We plotted Kaplan-Meier curves for the time between randomisation and birth. A difference in time to
174 birth was evaluated with the log-rank test statistic. Birth percentiles were determined using national reference data
175 for the Netherlands on birthweight, ethnicity, parity, and gestational age by week and day. All analyses were
176 performed using SAS software for Windows, version 9.4 (SAS Institute Inc., Cary, NC, USA).
177 Role of the funding source
178 The study was supported by a grant from the Netherlands Organisation for Health Research and Development
179 ZonMw (grant 171202008). This funding source had no role in study design, data collection, data analysis, data
180 interpretation, writing of the report, or the decision to submit the paper for publication. The authors had full access to
181 all the data in the study and had final responsibility for the decision to submit for publication.
182 RESULTS
183 Between 14th May 2012 and 17th March 2016, 6,088 eligible women were invited to participate in the INDEX trial,
184 of whom 4,273 declined. After randomisation but before analysis, one woman (IOL group) withdrew her consent,
185 and 13 women were found not to meet the eligibility criteria (6 IOL and 7 EM). Of the remaining 1,801 participants,
186 900 were randomly allocated to the IOL group and 901 to the EM group (Figure 1: “Flowchart”). Baseline
Page 8 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
9
187 characteristics were comparable in both groups except for nulliparity: IOL 50.8% (457/900) / EM 56.7% (511/901).
188 (Table 1: “Baseline”).
189 In the IOL group 28.9% (260/900) of the women had a spontaneous onset of labour before the planned IOL,
190 while 71.1% (640/900) had IOL, from which 59.7% (382/640) received cervical ripening (Figure 3: “Induction of
191 labour“). In the IOL group, 4.8% (43/900) of the women were not induced at 41+0 – 41+1, but at 41+2 or later.
192 In the EM group 73.7% (664/901) of the women had a spontaneous onset of labour, while 26.3% (237/901)
193 were induced (55.7% (132/237) received cervical ripening. In the EM group 9.4% (85/901) had labour induction at
194 42 weeks for postterm pregnancy, and 7.2% (65/901) were induced before 42 weeks due to medical reasons (e.g.
195 fetal condition in 15.6% (37/237), maternal condition in 9.7% (23/237), while 9.7% (87/901) in the EM group had
196 IOL on maternal request. The median gestational age at time of delivery was 41+0 weeks (IQR: 41+0 – 41+1) for the
197 IOL group and 41+2 weeks (IQR: 41+0 – 41+5) in the EM group (Table 4: “Delivery outcome”) In both groups
198 three-quarter of the women had a Bishop score of <6 at study entry. Time to delivery for both groups is shown in
199 Figure 2.
200 In the ppIOL group 15% (92/611) of the women had spontaneous onset of labour before the planned
201 induction. Of these women 11.1% (67/611) had a spontaneous onset of labour at 41+0 weeks and 4.1% (25/611)
202 had spontaneous onset at 41+1 weeks. In the ppIOL group 85% (519/611) of the women were actually induced;
203 62.5% (382/611) were induced at 41+0 and 22.4% (137/611) of the women were induced at 41+1 weeks. In the
204 ppEM group, 81% (524/647) of the women had a spontaneous onset of labour at ≥41+0 weeks and 19% (123/647)
205 were induced, 5.4% (35/647) were induced due to fetal condition ≥41+0 weeks, 3.4% (22/647) women were
206 induced due to maternal condition ≥41+0, 0.6% (4/647) women were induced due to >24h ruptured of the
207 membranes ≥41+0 weeks and 9.6% (62/647) women were induced because of post-term pregnancy (≥42+0).
208 Primary outcome
209 Perinatal outcomes are presented in Table 2 “Perinatal outcomes”. There were 15 participants with a composite
210 adverse perinatal outcome in the IOL group (1.7%) and 28 in the EM group (3.1%) (absolute risk difference -1.44%;
211 95% CI -2.88% to 0.00%; number needed to treat (NNT) 69; 95% CI: 35 to 3059). The p-value for non-inferiority
212 was 0.22 indicating that we could not exclude that EM leads to 2% or more adverse perinatal outcomes than IOL.
Page 9 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
10
213 The per-protocol analysis showed a 1.6% risk of an adverse perinatal outcome (10/611) in the ppIOL group
214 versus 2.9% (19/647) in the ppEM group (risk difference -1.30%; 95% CI: -2.99% to -0.39%, p-value for non-
215 inferiority = 0.21; see Appendix).
216 Three perinatal deaths occurred: one stillbirth in the IOL group and two stillbirths in the EM group. There
217 were no neonatal deaths. The stillbirth in the IOL group occurred in a 30 year old multiparous woman who was
218 randomised at 40+5 and scheduled for IOL at 41+1. She had reduced fetal movements at 40+6. At consultation fetal
219 death was diagnosed. She delivered a neonate of 3595 grams (p20-50). Investigations including autopsy showed no
220 explanation for the stillbirth. In the EM group a 36 year old nulliparous woman was diagnosed with a stillbirth at
221 41+3 when she was admitted to the hospital in labour. She delivered a neonate of 2945 grams (p5-10). Investigations,
222 including placental examination, showed no explanation for stillbirth, but autopsy was not performed on parents’
223 wish. The second stillbirth in the EM group occurred in a 32 year old multiparous woman and was diagnosed at 41+4
224 at a regular ‘impending postterm pregnancy’ consultation in secondary care. She delivered a neonate of 3715 grams
225 (p20-50). No autopsy was performed, but the placenta showed signs of chorioamnionitis.
226 The main contributor to the composite adverse outcome was Apgar 5’ <7: 1.2% (11/900) in the IOL group
227 and 2.6% (23/901) in the EM group (RR: 0.48; 95% CI: 0.23 to 0.98). Three of these neonates had an Apgar score 5’
228 <4, all in the EM group. The first neonate was born at 41+6 after spontaneous onset of labour and an operative
229 vaginal delivery (vacuum), because of fetal distress and failure to progress in second stage. This neonate was
230 diagnosed with MAS and was admitted to the NICU (see Table 3). The second neonate was diagnosed with sepsis
231 after spontaneous onset of labour at 40+6 and >24 hours rupture of membranes. The third concerned a neonate of
232 4320 grams, born after cervical ripening starting at 41+6 and failure to progress of second stage followed by a
233 Caesarean section at 42+2. The neonate was diagnosed with airway obstruction caused by vernix caseosa. Both were
234 admitted to a medium care unit for observation. All three recovered without complications. Admission to a NICU
235 was reported in 0.3% (3/899) in the IOL group versus 0.9% (8/899) in the EM group (RR: 0.38; 95% CI 0.10 to
236 1.41). Of the eleven children admitted to the NICU, six were diagnosed with a severe congenital disorder, three in
237 each group. MAS was diagnosed twice; only in the EM group, both neonates fully recovered. There were no plexus
238 brachialis lesions diagnosed in the study population. The two admissions because of a (suspected) infection showed
239 group B streptococcus in one neonate and sterile cultures in the other infant. One neonate was admitted because of a
240 pneumothorax. Further details of NICU admission are shown in Table 3.
Page 10 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
11
241 Arterial pH measurement was not recorded systematically and could therefore not be included in the
242 analysis. Imputing was not possible due to the many missing data (62% in IOL versus 70% in EM).
243 When stratifying by parity, we observed 2.4% nulliparous women with a composite adverse perinatal
244 outcome in the IOL group (11/457) and 4.1% in the EM group (21/511)(RR: 0.59; 95% CI: 0.29 to 1.20). In
245 multiparous women the incidence of adverse perinatal outcome was lower in both groups compared to nulliparous
246 women: 0.9% in the IOL group (4/443) and 1.8% in the EM group (7/390) (RR: 0.50; 95% CI: 0.15 to 1.71).
247
248 Secondary outcomes
249 Secondary perinatal outcomes are shown in Table 2. We did not find a difference in medium care admissions, 6.6%
250 in both groups (IOL 59/900 versus EM 60/901), . Small for gestational age (SGA: <p10), according to Dutch
251 birthweight percentiles, was equally observed in the IOL group in 6.8% (61/900) versus 6.9% (62/901) in the EM
252 group. Large for gestation age (LGA: >p90) was observed in the IOL group in 9.6% (86/900) versus 11% (99/901) in
253 the EM group. The incidence of congenital abnormalities was similar in both groups: 1.8% in the IOL group
254 (16/900) versus 2.1% in the EM group (19/901).
255 Table 4 summarizes the labour characteristics and mode of delivery. Oxytocin was significantly more often
256 given in the IOL group (59.2% (533/900) versus 39.4% (355/901) (RR: 1.50; 95% CI: 1.36 to 1.66). Meconium
257 stained amniotic fluid occurred significantly less in the IOL group (16.3% (147/900) compared to the EM group
258 (22.8% (205/901); RR: 0.72; 95% CI: 0.59 to 0.87). Ninety-seven women in each group (10.8%) had a CS (RR:
259 1.00; 95% CI 0.77 to 1.31), mainly for non-progressive labour at the first stage of labour (see table 4).
260 Maternal outcomes are shown in Table 5. The composite adverse maternal outcome occurred in 13.6%
261 (122/900) in the IOL group versus 11.3% (102/901) in the EM group (RR: 1.20; 95% CI: 0.94 to 1.53). Postpartum
262 haemorrhage (PPH) ≥ 1000 ml was the main contributor to the composite adverse maternal outcome and occurred in
263 9.1% (82/900) in the IOL group versus 8.0% (72/901) in the EM group (RR: 1.14; 95% CI: 0.84 to 1.54). Manual
264 removal of the placenta occurred in 5.1% (41/803) in the IOL group versus 4.1% (33/804) in the EM group (RR:
265 1.24; 95% CI 0.79 to 1.95). OASIS was diagnosed in 3.5% (28/803) in the IOL group versus 3.9% (31/804) in the
266 EM group (RR: 0.90; 95% CI: 0.55 to 1.49). Three mothers (0.3%) in the IOL group and two mothers (0.2%) in the
267 EM group were postpartum admitted to an ICU (RR: 1.50; 95% CI: 0.25 to 8.97), all after PPH. Blood loss in these
268 women was 3000 ml, 5100 ml and 7000 ml in the IOL group and 3390 ml and 5000 ml in the EM group. Maternal
Page 11 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
12
269 mortality did not occur. In the IOL group 265 (29.4%) of the women received epidural anaesthesia during labour
270 compared to 231 (25.6%) in the EM group (RR: 1.15; 95% CI: 0.99 to 1.33).
271
272 DISCUSSION
273 This randomised controlled trial compared the effect of induction of labour at 41 weeks versus expectant
274 management until 42 weeks on perinatal and maternal outcomes in women with an uncomplicated pregnancy. A
275 policy of IOL resulted in a median reduction of 2 days in gestational age at delivery. We found a small, but
276 significant 1.4% difference in composite adverse perinatal outcome favouring IOL, though the absolute risk of severe
277 adverse perinatal outcome (perinatal mortality, Apgar 5’ <4, NICU admission without severe congenital anomalies)
278 was very low in both groups. Our risk ratio for perinatal death is low and comparable with the risk ratio of the 4
279 studies (n=998) starting IOL (n=501) at 41 weeks versus EM (n=497) included in the Cochrane systematic review
280 ‘Induction of labour for improving birth outcomes for women at or beyond term’ (RR 0.33; 95% CI: 0.03 to 3.17)19 21
281 26.
282 Most of our primary composite outcomes can be attributed solely to the component Apgar 5’ <7 (11/15)
283 73% IOL versus (18/28) 64% EM). The cut-off of Apgar 5’ <7 was based on the earlier ACOG Committee Opinion
284 of 2006 27. In October 2015, and thus after the start of this study, the ACOG Committee stated that though it is
285 inappropriate to use an Apgar score alone to diagnose birth asphyxia, an Apgar score <4 can be considered as a
286 ‘nonspecific sign of illness’. The composite outcome including Apgar score 5’<4 instead of <7 is 0.4% (4/900) for
287 IOL and 1.3% (12/901) for EM (absolute risk difference -0.9%; 95% CI: -1.7% to -0.02%; p-value 0.06; NNT 113;
288 95%CI 57 to 4624). Congenital anomalies accounted for a substantial part of the NICU admissions in our trial,
289 although it was an exclusion criterion at study entry. It is unknown if the outcome for these children would have been
290 better when they were born earlier, though that is considered unlikely (Table 3). For these reasons, we also analysed
291 the primary composite outcome using an Apgar score 5’ <4 and NICU admission without severe congenital
292 anomalies. With these adapted adverse outcomes (perinatal mortality, Apgar 5’<4 instead of <7 and NICU admission
293 without severe congenital anomalies), the absolute risk on the composite adverse perinatal outcome is substantially
294 lower in both groups, though with a still significant risk difference in favour of IOL (0.11% (1/897) versus EM
295 (1.00% (9/898); absolute risk difference: -0.9%; 95% CI: -0.20% to -1.58%; p-value 0.01; NNT 112; 95%CI 63 to
296 491).
Page 12 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
13
297 In our study MAS occurred in two neonates of the EM group. Gelisen et al. reported in a RCT with
298 comparable time frame 16/600 neonates of whom 12/300 in the EM group. We found a 10 and 20 times lower rate of
299 MAS (0.0% and 0.2% versus 1.3 and 4%) in the IOL and EM groups in comparison to the study of Gelisen. Since
300 Gelisen et al. did not specify MAS, this could be attributed to a difference in definition. Despite the considerable
301 difference in MAS between the IOL and EM groups, Gelisen et al. found no difference in NICU admission (4.3%
302 IOL versus 5.0% EM) which is expected to be associated with MAS. We found a lower rate of NICU admissions
303 also compared to the Cochrane systematic review on induction of labour >41 weeks. (0.3% IOL and 0.8% EM
304 (INDEX) versus 11% IOL and 12% EM (SR)). No details were given in the SR on NICU admission, such as
305 diagnosis, potential association with gestational age, and/or presence of congenital anomalies, which hampers a clear
306 comparison16 26.
307 We did not find differences in Caesarean section or operative vaginal delivery rates, which is consistent
308 with other large studies on induction of labour 28 29. In the only study that compared the same timeframes as our
309 study, the risk ratio on Caesarean section was comparable for both groups, though the absolute risk was twice as high
310 compared to our study. This could be due to other inclusion criteria (Gelisen et al.: Bishop score of <5) or differences
311 in policy during labour, as reflected by differences in national overall caesarean rates in Turkey (53% versus 16% in
312 the Netherlands) 30 31. The Cochrane systematic review on this topic concluded that IOL at or beyond 41 weeks is
313 associated with lower Caesarean section rates. The largest contribution to this outcome came from a randomised
314 controlled trial in which women in the control group were induced only with oxytocin according to study protocol
315 while in the induction group the use of prostaglandins was allowed in women with low Bishop scores16. Two other
316 systematic reviews including the same trial conclude that the difference in Caesarean section rate is possibly due to
317 the influence of this study with incomparable study arms 17 20 32. Population based cohort studies showed conflicting
318 results on the effect of IOL on the Caesarean section rate 33-35.
319 Our trial showed some notable results besides the main outcomes. Around 85% of participating women had
320 a Caucasian ethnicity. It has been shown that women from South Asian, African and Mediterranean origin have a
321 higher risk of perinatal mortality beyond term in comparison to Caucasian women 36. In our study, we were not able
322 to assess the effect of IOL in women with non-Caucasian ethnicity due to a low participation rate of women of other
323 ethnic origin. Also, the effect of age on adverse perinatal outcome could not be assessed because of the low number
324 of participating older mothers (>35 years).
Page 13 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
14
325 As in other studies on pregnancies at or beyond 41 weeks, most women in our study had an unfavourable
326 cervix with a Bishop Score <6 at randomisation. Although induction was planned one or two days after
327 randomisation, 29.0% of the women in the IOL group had a spontaneous onset of labour before induction started, in
328 comparison to 73.7% in the EM group. Despite the fact that women with suspected or established intra uterine
329 growth restriction were not eligible for inclusion, 7% of all children were born <10th Dutch birthweight percentile
330 (61/900 IOL and 62/901) confirming the difficulty in diagnosing term growth restricted babies. In the IOL group
331 2/61 children were born with a birthweight <p10 and had an adverse perinatal outcome: one neonate, with a
332 birthweight of 3100 grams (<p10), had an Apgar score 5’of 6 after forcipal extraction because of fetal distress. The
333 other neonate, with a birthweight of 2595 grams (<p2.3), had an Apgar score 5’of 6 after caesarean section because
334 of fetal distress, with an umbilical cord pH of 6.87, possibly due to hypotension of the mother after epidural
335 analgesia for pain relief and/ or multiple entanglement of the umbilical cord. In the EM group 3/62 were born with a
336 birthweight <p10 and had an adverse perinatal outcome: one was a stillbirth of 2945 grams as described before, one
337 with a birthweight of 2980 grams was admitted to the NICU because of a pneumothorax as described before and one
338 with a birthweight of 3040 grams had an Apgar score 5’of 6 which was attributed to maternal administration of
339 pethidine.
340
341 A major strength of our study is that it concerns a nationwide multicentre randomised controlled trial of a well-
342 defined obstetrical population at low risk, the largest trial on the comparison of IOL at 41 weeks with IOL at 42
343 weeks until now 37. There were no cases lost to follow-up.
344 In the Netherlands, expectant management until 42 weeks is the standard of care in the obstetrical low risk
345 population at 41-42 weeks according to the Dutch Obstetrical Indication List, though there is wide practice variation
346 due to preferences of women and caregivers, which complicated inclusion 15. Not all eligible women were invited,
347 and not all women who were asked participated, because of a preference for IOL or EM. Despite this selective
348 participation this trial offers the best possible representation of pregnant women reaching 41+0 weeks in the
349 Netherlands.
350 We are aware of some potential limitations of our trial. We chose to use a composite adverse perinatal
351 outcome instead of a single outcome as perinatal mortality. We considered any major adverse perinatal outcome in
352 an otherwise uncomplicated pregnancy as undesirable. It is debatable if all included adverse perinatal outcomes in
Page 14 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
15
353 our composite outcome measure are relevant to identify real severe adverse perinatal outcome with impact on baby’s
354 short-term or long-term health status. However, if we limited the criteria for Apgar 5’ <4 and excluding severe
355 congenital abnormalities, IOL resulted in a small but statistically significant risk reduction of 0.9%.
356 The non-inferiority design was chosen because we did not expect the Dutch standard policy of EM in our
357 obstetrical low risk population to be inferior to a policy of IOL but acceptable or preferable if leading to comparable
358 outcomes 38. It is good practice to use a per-protocol analysis in non-inferiority trials, as an intention-to-treat analysis
359 carries a risk of falsely rejecting the null hypothesis of inferiority. Because we did not reject the null hypothesis and
360 do not conclude non-inferiority we presented the intention-to-treat analyses, since such analyses are more common in
361 reports of clinical trials. All per protocol analyses are shown in the appendix.
362 We did not stratify randomisation by parity, because we expected a balanced allocation in both groups due
363 to the large study population. However, it did result to an imbalance: 50.8% of nulliparous women in the IOL group
364 compared to 56.7% in the EM group. We did not adjust for parity in our analysis, but after stratifying by parity in an
365 additional analysis we observed similar results. A higher incidence of the composite adverse perinatal outcome was
366 seen in the nulliparity group in both IOL (nulliparous 2.4% versus multiparous 0.9%) and EM group (nulliparous
367 4.1% versus multiparous 1.8%) as is in concordance with other studies 39.
368 Arterial pH measurement is no option in primary care and no standard policy in uncomplicated birth in most
369 hospitals in the Netherlands. Because of the high number of missing pH (60-70%) and the impossibility to impute,
370 we could not include umbilical artery pH in the composite outcome, which could have led to selection bias.
371 However, including the available umbilical arterial pH in the analyses did not alter the results.
372
373 The results of our study may be interpreted in different ways, which can have implications for standard practice. If
374 the composite outcome is interpreted straight forward, there is a small benefit of induction at 41 weeks which could
375 justify standard induction at 41 weeks. On the other hand, one can argue that a change of policy to earlier induction,
376 concerning roughly one fifth of all women with a singleton pregnancy, is too rigorous considering the relatively low
377 incidence of perinatal mortality, gestational age associated NICU admission and Apgar 5’<4, as indicator for
378 encephalopathy. This could justify EM if the woman wants to avoid IOL. In both sides of the spectrum caregivers are
379 challenged to provide neutral, evidence-based counselling of low-risk women in late term pregnancy on the pros and
380 cons of IOL. In a recent report of Walsh et al. on ‘women’s experiences of being ‘overdue’ in pregnancy’ it appeared
Page 15 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
16
381 that women did not feel they are offered a real choice when it comes to management of prolonged pregnancy, which
382 is confirmed by other studies. Induction of labour is often presented as an ‘inevitable next step’ to women without
383 giving information on alternative management strategies40-42
384
385 Our trial is the largest trial so far comparing induction of labour at 41 weeks with induction at 42 weeks.
386 Substantial larger trials are needed to evaluate differences in rare outcomes, such as perinatal mortality and NICU
387 admission. A systematic review or Individual Participant Meta-Analysis on the 41-42 weeks comparison could then
388 be performed including other and our data. Future research could also focus on long-term adverse perinatal outcome
389 of both strategies, though this requires long-term follow up of children37. In addition, a more tailored approach will
390 need identification of women who could remain pregnant until 42 weeks or are at increased risk of adverse perinatal
391 outcome (e.g. relational model).
392
393 The incidence of late term pregnancy varies between countries due to different management strategies43.
394 Women worldwide have to be counselled on the desired policy in late-term pregnancy. In this trial, IOL at 41 weeks
395 resulted in better perinatal outcome than a policy of EM until 42 weeks, though the absolute risk of severe adverse
396 outcome (perinatal mortality, NICU admission, Apgar 5’ <4) was low in both groups. As every intervention in the
397 natural birth process, the decision to induce labour must be taken with caution, as the expected benefits should
398 outweigh possible adverse effects for both mother or child44. Our results should be used to inform women
399 approaching a gestational age of 41 weeks, so they can weigh the respective outcomes and make a decision for IOL
400 at 41 weeks or to continue pregnancy until 42 weeks.
401 REGISTRATION
402 The study was registered with the Netherlands Trial Register (http://www.trialregister.nl/trialreg/index.asp), which is
403 recognised by the WHO and the International Committee of Medical Journal Editors (ICJME) under registration
404 number NTR3431 (http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=3431). The protocol is available from
405 https://bmcpregnancychildbirth.biomedcentral.com/articles/10.1186/1471-2393-14-350.
406 CONTRIBUTORS
Page 16 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
17
407 JKJK, AB and JCK are joint first authors and contributed equally to the study. EM and BWM initiated this study,
408 EM and JP supervised this study. JKJK, AB and JCK wrote the first and subsequent drafts of the paper.
409 RD conducted the statistical analyses and takes responsibility of the integrity of the data and accuracy of the data
410 analyses. PB advised on statistical issues and interpretation of the results.
411 AK is the neonatologist who reviewed all anonymized NICU admissions on case level with JCK and EM. All authors
412 have approved the final version of this manuscript submitted for publication. JKJK, AB, JCK, JD, JP and EM are
413 guarantors.
414 DECLARATION OF INTERESTS
415 All authors have completed the Unified Competing Interest form (available on request from the corresponding
416 author) and declare: only financial support from funder ZonMW, no support from any other organisation for the
417 submitted work; no financial relationships with any organisations that might have an interest in the submitted work
418 in the previous three years, no other relationships or activities that could appear to have influenced the submitted
419 work.
420 BWM is supported by a NHMRC Practitioner Fellowship (GNT1082548). BWM reports consultancy for ObsEva,
421 Merck and Guerbet.
422 ACKNOWLEDGEMENTS
423 This study was funded by ZonMw, grant number 171202008. We would like to thank all women who participated in
424 this study and all midwives, gynaecologists, residents and nurses of the participating centres. We also would like to
425 thank everyone who participated in the study setup and effectuation of the trial.
426
427 NOTES
428 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, an
429 exclusive licence on a worldwide basis to the BMJ Publishing Group Ltd to permit this article (if accepted) to be
430 published in BMJ editions and any other BMJPGL products and sublicences such use and exploit all subsidiary
431 rights, as set out in the licence.
Page 17 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
18
432 EdM affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no
433 important aspects of the study have been omitted; and that any discrepancies from the study as planned and
434 registered have been explained.
435 We will make the relevant data available on written reasonable request.
436
Page 18 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
19
438 Figures and tables:
439 Figure 1 "Flowchart”
440 Figure 2 “Time to delivery”
441 Figure 3 "Flowchart cervical ripening”
442 Table 1 “Baseline characteristics of study participants by intervention group”
443 Table 2 “Perinatal outcomes”
444 Table 3 “NICU admission”
445 Table 4 “Delivery outcomes”
446 Table 5 “(Adverse) maternal outcomes”
447 Appendix Per-protocol population
448
Page 19 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
20
450 Table 1 “Baseline characteristics of study participants by intervention group”Induction of labour Expectant management
(n=900) (n=901)
Maternal age (yrs.; mean, sd) 30.6 (4.8) 30.2 (4.6)
Ethnicity
Caucasian 779 (87%) 767 (85%)
Other 121 (13%) 134 (15%)
Highest level of education
primary school 7 (0.8%) 4 (0.4%)
secondary school 37 (4.1%) 15 (1.7%)
lower / medium professional education 358 (40%) 350 (39%)
higher professional education / university 286 (32%) 322 (36%)
other / unknown 212 (24%) 210 (23.3%)
Social economic status
1 219 (24%) 251 (28%)
2 401 (45%) 365 (41%)
3 225 (25%) 233 (26%)
unknown 55 (6.1%) 52 (5.8%)
Parity
nulliparous 457 (51%) 511 (57%)
multiparous 443 (49%) 390 (43%)
Previous post-term pregnancy (≥294 days)* 51 / 443 (12%) 34 / 390 (8.7%)
Level of care at time of recruitment
primary care 851 (95%) 850 (94%)
secondary care 49 (5.4%) 51 (5.7%)
Bishop score at study entry
Nulliparous
Bishop score ≥6 47 / 457 (10%) 71 / 511 (14%)
Bishop score <6 360 / 457 (79%) 365 / 511 (71%)
missing 50 / 457 (11%) 75 / 511 (15%)
Multiparous
Bishop score ≥6 71 / 443 (16%) 46 / 390 (12%)
Bishop score <6 310 / 443 (70%) 294 / 390 (75%)
missing 62 / 443 (14%) 50 / 390 (13%)
451 Table 1 – Baseline characteristics. Analyses intention-to-treat population. 452 Data are presented as n (%) or mean (sd). * Numerator: primiparous and multiparous women
Page 20 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
21
453 Table 2 - “Perinatal outcomes”Induction of labour
Expectant management
Relative risk(95% CI)
p-value
(n=900) (n=901)
Composite adverse perinatal outcome* 15 (1.7%) 28 (3.1%) 0.54 (0.29 – 1.00) 0.045†
with Apgar score 5’ <4 instead of <7 4 (0.4%) 12 (1.3%) 0.33 (0.11 – 1.03) 0.06†
Stillbirth 1 (0.1%) 2 (0.2%) 0.50 (0.05 – 5.51) 1.00†
Neonatal death, post-partum 0 (0.0%) 0 (0.0%) n/a -
Apgar score 5 min. post-partum¤
<7 11 (1.2%) 23 (2.6%) 0.48 (0.23 – 0.98) 0.038
<4 0 (0.0%) 3 (0.3%) n/a -
Admission neonate to
intensive care (NICU) 3 (0.3%) 8 (0.9%) 0.38 (0.10 – 1.41) 0.23†
medium care 59 (6.6%) 60 (6.7%) 0.98 (0.69 – 1.39) 0.90
Meconium aspiration syndrome‡ 0 (0.0%) 2 (0.2%) n/a -
Plexus brachialis injury 0 (0.0%) 0 (0.0%) n/a -
Umbilical cord pH (arterial)
pH <7.05 16 (1.8%) 12 (1.3%) 1.06 (0.51 – 2.20) 0.88
missing 557 (62%) 629 (70%) n/c -
Congenital abnormality 16 (1.8%) 19 (2.1%) 0.84 (0.44 – 1.63) 0.61
Hypoglycaemia§ 3 (0.3%) 6 (0.7%) 0.50 (0.13 – 2.00) 0.51†
Neonatal infection / sepsis¶ 37 (4.1%) 37 (4.1%) 1.00 (0.64 – 1.56) 1.00
Sex, female 453 (50%) 463 (51%) 0.98 (0.89 – 1.07) 0.65
Birth weight (g; mean, sd) 3,685 (417.4) 3,741 (430.0) -56.6 (-95.8 – -17.4)**
0.005
Small of gestational age
<2.3rd percentile 13 (1.4%) 11 (1.2%) 1.18 (0.53 – 2.62) 0.68
<10th percentile 61 (6.8%) 62 (6.9%) 0.99 (0.70 – 1.39) 0.93
Large for gestational age
>90th percentile 86 (9.6%) 99 (11%) 0.87 (0.66 – 1.14) 0.32
>97.7th percentile 15 (1.7%) 27 (3.0%) 0.56 (0.30 – 1.04) 0.07
454 Table 2 – Perinatal outcomes (intention-to-treat). 455 Data are n (%) or mean (standard deviation). Confidence intervals are 95%. NICU: Neonatal Intensive Care Unit * 456 Composite outcome defined as: perinatal mortality, and/or a 5-minute Apgar-score below 7, and/or meconium 457 aspiration syndrome, and/or plexus brachialis injury, and/or NICU admission. † Fisher’s exact test. ¤ Apgar score of 458 the live births. ‡ Defined as respiratory distress after birth in presence of meconium stained amniotic fluid). § 459 Defined as glucose concentration <1.9 mmol/L and need for intravenous glucose. ¶ Defined as clinical suspected 460 findings or proved positive blood culture. ** Mean difference between groups with 95% confidence interval.
Page 21 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
22
461 Table 3 “NICU admission”462
Allocation NICU admission: diagnosis Congenital anomaly
Gestational age at time of birth
induction Long QT syndrome yes 40+6
induction Mild mitralis insuffiency, persistent ductus arteriosus yes 41+0
induction Interstitial lung disorder yes 41+1
expectant Diaphragm herniation, atrial septal defect, ventricular septal defect yes 40+6
expectant Muscular ventricular septal defect yes 41+5
expectant Vocal cord paresis, dysmorphic features yes 41+5
expectant Infection (suspected sterile culture) no 41+2
expectant Infection, GBS positive no 41+3
expectant Pneumothorax no 41+4
expectant Meconium aspiration syndrome no 41+2
expectant Meconium aspiration syndrome no 41+6
463
Page 22 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
23
464 Table 4 “Delivery outcomes”Induction of labour
Expectant management
Relative risk(95% CI)
p-value
(n=900) (n=901)
Gestational age at delivery (days; median, IQR) 287 (287 - 288) 289 (287 - 292) -2.1 (-2.3 – -1.9)* <0.0001 †Time randomisation to delivery (days; mean, sd) 2.1 (1.6) 4.2 (3.0) -2.2 (-2.5 – -2.0)* <0.0001 †
Level of care at onset of labour
Primary care 255 (28%) 619 (69%) n/c -
Secondary care 645 (72%) 282 (31%) n/c -
Onset of labour
Spontaneous (reference) 260 (29%) 664 (74%) 1.00 -
Induction 640 (71%) 237 (26%) 2.70 (2.41 – 3.04) <0.0001
Mode of induction
cervical ripening (catheter /prostaglandins) 382 / 640 (60%) 132 / 237 (56%) 1.07 (0.94 – 1.22) 0.30
amniotomy, without oxytocin 87 / 640 (14%) 34 / 237 (14%) 0.95 (0.66 – 1.37) 0.77
amniotomy, with oxytocin 156 / 640 (24%) 59 / 237 (25%) 0.98 (0.76 – 1.27) 0.87
Indication for induction
randomisation 634 / 640 (99%) 0 / 237 (0.0%) n/c -
post-term pregnancy 0 / 640 (0.0%) 85 / 237 (36%) n/c -
foetal condition 5 / 640 (0.8%) 37 / 237 (16%) n/c -
maternal condition 0 / 640 (0.0%) 23 / 237 (9.7%) n/c -
elective / maternal wish 1 / 640 (0.2%) 87 / 237 (37%) n/c -
>24 h of ruptured membranes 0 / 640 (0.0%) 4 / 237 (1.7%) n/c -
other 0 / 640 (0.0%) 1 / 237 (0.4%) n/c -
Use of oxytocin 533 (59%) 355 (39%) 1.50 (1.36 – 1.66) <0.0001
Use of tocolytics 28 (3.1%) 16 (1.8%) 1.75 (0.95 – 3.22) 0.07
Maternal intrapartum infection
fever during labour (≥38 °C) 50 (5.6%) 46 (5.1%) 1.09 (0.74 – 1.61) 0.67
use of antibiotics 48 (5.3%) 35 (3.9%) 1.37 (0.90 – 2.10) 0.14
Meconium stained amniotic fluid 147 (16%) 205 (23%) 0.72 (0.59 – 0.87) 0.0007
Mode of delivery
spontaneous vaginal delivery 710 (79%) 696 (77%) 1.02 (0.97 – 1.07) 0.40
operative vaginal delivery 93 (10%) 108 (12%) 0.86 (0.66 – 1.12) 0.27
(secondary) caesarean section 97 (11%) 97 (11%) 1.00 (0.77 – 1.31) 0.99
Indication successful operative vaginal delivery
failure to progress at second stage 39 / 93 (42%) 49 / 108 (45%) 0.92 (0.67 – 1.27) 0.63
suspected foetal distress 43 / 93 (46%) 37 / 108 (34%) 1.35 (0.96 – 1.90) 0.08
suspected foetal distress and failure to progress 10 / 93 (11%) 22 / 108 (20%) 0.53 (0.26 – 1.06) 0.07
maternal complication or other 1 / 93 (1.1%) 0 / 108 (0.0%) n/a -
Indication secondary caesarean section
failure to progress at first stage 29 / 97 (30%) 21 / 97 (22%) 1.38 (0.85 – 2.25) 0.19
failure to progress at second stage 12 / 97 (12%) 18 / 97 (19%) 0.67 (0.34 – 1.31) 0.24
failed OVD 6 / 97 (6.2%) 12 / 97 (12%) 0.50 (0.20 – 1.28) 0.22‡
Page 23 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
24
suspected foetal distress 24 / 97 (25%) 21 / 97 (22%) 1.14 (0.68 – 1.91) 0.61
suspected foetal distress and failure to progress at first stage
7 / 97 (7.2%) 8 / 97 (8.3%) 0.75 (0.17 – 3.26) 1.00‡
suspected foetal distress and failure to progress at second stage
4 / 97 (4.1%) 3 / 97 (3.1%) 1.00 (0.26 – 3.88) 1.00‡
maternal complication or other 15 / 97 (16%) 14 / 97 (14%) 0.93 (0.48 – 1.83) 0.84
465466 Table 4 – Delivery outcomes (intention-to-treat). 467 Data are n (%), mean and standard deviation, or median (IQR). OVD: Operative Vaginal Delivery. * Mean difference 468 between groups with 95% confidence interval. †Mann-Whitney U test. ‡Fisher’s exact test.
Page 24 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
25
469 Table 5 “(Adverse) maternal outcomes”Induction of labour Expectant
managementRelative risk(95% CI)
p-value
(n=900) (n=901)
Composite adverse maternal outcome* 122 (14%) 102 (11%) 1.20 (0.94 – 1.53) 0.15
Maternal death 0 (0.0%) 0 (0.0%) n/a -
Post-partum blood loss
< 1000 mL (reference) 818 (91%) 829 (92%) 1.00 -
≥ 1000 mL 82 (9.1%) 72 (8.0%) 1.14 (0.84 – 1.54) 0.40
1000-1500 mL 34 (3.8%) 35 (3.9%) 0.97 (0.61 – 1.54) 0.91
>1500-2000 mL 21 (2.3%) 14 (1.6%) 1.50 (0.77 – 2.93) 0.23
> 2000 mL 27 (3.0%) 23 (2.6%) 1.18 (0.68 – 2.04) 0.56
Post-partum blood loss (mL; median, IQR) 300 (200 – 500) 300 (250 – 500) - 0.18†
Transfusion (packed cells or plasma) 23 (2.6%) 17 (1.9%) 1.35 (0.73 – 2.52) 0.34
Manual removal placenta 41 (5.1%) 33 (4.1 %) 1.24 (0.79 – 1.95) 0.34
Perineal tear
episiotomy (without tear) 234 (29%) / 803 246 (31%) / 804 0.95 (0.82 – 1.11) 0.52
OASIS 28 (3.5%) / 803 31 (3.9%) / 804 0.90 (0.55 – 1.49) 0.69
third degree tear 15 (1.9%) / 803 19 (2.4%) / 804 0.79 (0.40 – 1.54) 0.49
fourth degree tear 8 (1.0%) / 803 7 (0.9%) / 804 1.14 (0.41 – 3.14) 0.80‡
episiotomy and third degree tear 4 (0.5%) / 803 2 (0.3%) / 804 2.00 (0.37 – 10.90) 0.45‡
episiotomy and fourth degree tear 1 (0.1%) / 803 3 (0.4%) / 804 0.33 (0.03 – 3.20) 0.62‡
Maternal admission (highest level of care)
intensive care 3 (0.3%) 2 (0.2%) 1.50 (0.25 – 8.97) 0.66‡
medium care 5 (0.6%) 5 (0. 6%) 1.00 (0.29 3.45) 1.00‡
ward 271 (30.1%) 277 (31%) 0.98 (0.85 – 1.13) 0.77
Indications for maternal admission
Thromboembolic complications 0 (0.0%) 0 (0.0%) n/a -
Hypertensive disorders§ 5 (0.6%) 14 (1.6%) 0.36 (0.13 – 0.99) 0.06‡
Post-partum blood loss 49 (5.5%) 52 (5.8%) 1.06 (0.72 – 1.54) 0.78
Post-caesarean 97 (11%) 97 (11%) 1.00 (0.77 – 1.31) 0.99
Pain treatment during labour 420 (47%) 386 (43%) 1.09 (0.98 – 1.20) 0.10
Remifentanil 128 (14%) 129 (14%) 0.99 (0.79 – 1.25) 0.95
Pethidine / phenergan / other opiates 60 (6.7%) 51 (5.7%) 1.18 (0.82 – 1.69) 0.38
Epidural / spinal 265 (29%) 231 (26%) 1.15 (0.99 – 1.33) 0.07
Others 1 (0.1%) 5 (0.6%) 0.20 (0.02 – 1.71) 0.22‡
470 Table 5 – (Adverse) maternal outcomes (intention-to-treat). *Defined as post-partum haemorrhage ≥ 1000 mL, 471 and/or manual removal of placenta, and/or third or fourth degree tears (Obstetrical Anal Sphincter Injuries, OASIS), 472 and/or intensive care admission, and/or maternal death. Denominator for perineal tear are vaginal deliveries only. 473 †Mann Whitney U test. ‡Fisher’s exact test. §Including (pre-)eclampsia, and HELLP syndrome.
474
Page 25 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
26
475 Appendix
476 Per-protocol population
477 Perinatal outcomes
Induction of labour
Expectant management
Relative risk(95% CI)
p-value
(n=611) (n=647)
Composite adverse perinatal outcome* 10 (1.6%) 19 (2.9%) 0.56 (0.26 – 1.19) 0.13†
with Apgar score 5’ <4 instead of <7 3 (0.5%) 8 (1.2%) 0.40 (0.11-1.49) 0.23†
Stillbirth 1 (0.2%) 2 (0.3%) 0.53 (0.05 – 5.82) 1.00†
Neonatal death, post-partum 0 (0.0%) 0 (0.0%) n/a -
Apgar score 5 min. post-partum¤
<7 7 (1.1%) 16 (2.5%) 0.46 (0.19 – 1.12) 0.08
<4 1 (0.2%) 3 (0.5%) n/a -
Admission neonate to
intensive care (NICU) 2 (0.3%) 6 (0.9%) 0.35 (0.07 – 1.74) 0.51†
medium care 39 (6.4%) 48 (7.4%) 0.86 (0.57 – 1.29) 0.45
Meconium aspiration syndrome‡ 0 (0.0%) 2 (0.2%) n/a -
Plexus brachialis injury 0 (0.0%) 0 (0.0%) n/a -
Umbilical cord pH (arterial)
pH <7.05 11 (1.8%) 10 (1.6%) 1.12 (0.49 – 2.58) 0.79
missing 366 (60%) 447 (69%) n/c -
Congenital abnormality 11 (1.8%) 12 (1.9%) 0.97 (0.42 – 2.18) 0.94
Hypoglycaemia§ 1 (0.2%) 5 (0.8%) 0.21 (0.02 – 1.81) 0.22†
Neonatal infection / sepsis 23 (3.8%) 29 (4.5%) 0.84 (0.49 – 1.44) 0.52
Sex, female 305 (50%) 337 (52%) 0.96 (0.86 – 1.07) 0.44
Birth weight (g; mean, sd) 3,697 (414.4) 3,747 (433.5) -50.2 (-97.2 – -3.2)** 0.036
Small of gestational age
<2.3rd percentile 11 (1.8%) 8 (1.2%) 1.46 (0.59 – 3.60) 0.42
<10th percentile 40 (6.6%) 46 (7.1%) 0.92 (0.61 – 1.39) 0.69
Large for gestational age
>90th percentile 54 (8.8%) 74 (11%) 0.77 (0.55 – 1.08) 0.13
>97.7th percentile 12 (2.0%) 22 (3.4%) 0.58 (0.29 – 1.16) 0.12
478 Perinatal outcomes (per-protocol). Data are n (%) or mean (standard deviation). Confidence intervals are 95%. 479 NICU: Neonatal Intensive Care Unit * Composite outcome defined as: perinatal mortality, and/or a 5-minute Apgar-480 score below 7, and/or meconium aspiration syndrome, and/or plexus brachialis injury, and/or NICU admission. ¤ 481 Apgar score of the live births. † Fisher’s exact test. ‡ Defined as respiratory distress after birth in presence of 482 meconium stained amniotic fluid). § Defined as glucose concentration <1.9 mmol/L and need for intravenous 483 glucose. ¶ Defined as clinical suspected findings, or proved positive culture. ** Mean difference between groups 484 with 95% confidence interval.
Page 26 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
27
486 Delivery outcomes
Induction of labour
Expectant management
Relative risk(95% CI)
p-value
(n=611) (n=647)
Gestational age at delivery (days; median, IQR) 288 (287 - 288) 290 (288 - 292) -2.3 (-2.5 – -2.1)* <0.0001 †Time randomisation to delivery (days; mean, sd) 2.2 (1.5) 4.5 (2.9) -2.2 (-2.5 – -2.0)* <0.0001 †
Level of care at onset of labour
Primary care 96 (16%) 474 (73%) n/c -
Secondary care 515 (84%) 173 (27%) n/c -
Onset of labour
Spontaneous (reference) 92 (15%) 524 (81%) 1.00 -
Induction 519 (85%) 123 (19%) 4.47 (3.80 – 5.26) <0.0001
Mode of induction
cervical ripening (catheter /prostaglandins) 304 / 519 (59%) 71 / 123 (58%) 1.01 (0.86– 1.20) 0.86
amniotomy, without oxytocin 72 / 519 (14%) 12 / 123 (9.8%) 1.42 (0.80 – 2.54) 0.23
amniotomy, with oxytocin 135 / 519 (26%) 33 / 123 (27%) 0.97 (0.70 – 1.34) 0.85
Indication for induction
randomisation 519 / 519 (100%) 0 / 123 (0.0%) n/c -
post-term pregnancy 0 / 519 (0.0%) 62 / 123 (50%) n/c -
foetal condition 0 / 519 (0.0%) 35 / 123 (29%) n/c -
maternal condition 0 / 519 (0.0%) 22 / 123 (18%) n/c -
elective / maternal wish 0 / 519 (0.0%) 0 / 123 (0.0%) n/c -
>24 h of ruptured membranes 0 / 519 (0.0%) 4 / 123 (3.3%) n/c -
other 0 / 519 (0.0%) 0 / 123 (0.0%) n/c -
Use of oxytocin 402 (66%) 255 (39%) 1.69 (1.49 – 1.87) <0.0001
Use of tocolytics 20 (3.3%) 13 (2.0%) 1.63 (0.82 – 3.25) 0.17
Maternal intrapartum infection
fever during labour (≥38 °C) 33 (5.4%) 35 (5.4%) 0.99 (0.63 – 1.59) 0.99
use of antibiotics 34 (5.6%) 22 (3.4%) 1.64 (0.97 – 2.77) 0.07
Meconium stained amniotic fluid 82 (13%) 163 (25%) 0.53 (0.42 – 0.68) 0.0001
Mode of delivery
spontaneous vaginal delivery 484 (79%) 493 (76%) 1.04 (0.98 – 1.10) 0.20
operative vaginal delivery 56 (9.2%) 80 (12%) 0.74 (0.54 – 1.02) 0.07
(secondary) caesarean section 71 (12%) 74 (11%) 1.02 (0.75 – 1.38) 0.92
Indication successful operative vaginal delivery
failure to progress at second stage 27 / 56 (48%) 37 / 80 (46%) 1.04 (0.73 – 1.49) 0.82
suspected foetal distress 24 / 56 (43%) 29 / 80 (36%) 1.18 (0.78 – 1.80) 0.43
suspected foetal distress and failure to progress 5 / 56 (8.9%) 14 / 80 (18%) 0.51 (0.19 – 1.34) 0.21‡
maternal complication or other 0 / 56 (0.0%) 0 / 80 (0.0%) n/a -
Indication secondary caesarean section
failure to progress at first stage 27 / 71 (38%) 16 / 74 (22%) 1.76 (1.04 – 2.98) 0.04
failure to progress at second stage 8 / 71 (11%) 12 / 74 (16%) 0.69 (0.30 – 1.60) 0.39
failed OVD 3 / 71 (4.2%) 9 / 74 (12%) 0.35 (0.10 – 1.23) 0.13‡
suspected foetal distress 14 / 71 (20%) 18 / 74 (24%) 0.81 (0.44 – 1.50) 0.51
Page 27 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
28
suspected foetal distress and failure to progress at first stage 3 / 71 (4.2%) 7 / 74 (9.5%) 0.45 (0.12 – 1.66) 0.33‡suspected foetal distress and failure to progress at second stage
3 / 71 (4.2%) 3 / 74 (4.1%) 0.96 (0.22 – 4.99) 1.00‡
maternal complication or other 13 / 71 (18%) 9 / 74 (12%) 1.51 (0.69 – 3.30) 0.31
487 Delivery outcomes (per-protocol). Data are n (%), mean and standard deviation, or median (IQR). OVD: Operative 488 Vaginal Delivery. * Mean difference between groups with 95% confidence interval. †Mann-Whitney U test. 489 ‡Fisher’s exact test.
490
Page 28 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
29
492 (Adverse) maternal outcomes
Induction of labour Expectant management
Relative risk(95% CI)
p-value
(n=611) (n=647)
Composite adverse maternal outcome* 84 (14%) 75 (12%) 1.19 (0.89 – 1.59) 0.25Maternal death 0 (0.0%) 0 (0.0%) n/a -
Post-partum blood loss
< 1000 mL (reference) 552 (90%) 592 (92%) 1.00 -
≥ 1000 mL 59 (10%) 55 (8.5%) 1.14 (0.80 – 1.61) 0.48
1000-1500 mL 23 (3.8%) 27 (4.2%) 0.92 (0.53 – 1.58) 0.76
>1500-2000 mL 15 (2.5%) 10 (1.6%) 1.59 (0.72 – 3.52) 0.25
> 2000 mL 21 (3.4%) 18 (2.8%) 1.24 (0.67 – 2.31) 0.49
Post-partum blood loss (mL; median, IQR) 300 (200 – 500) 300 (250 – 500) - 0.18†
Transfusion (packed cells or plasma) 18 (3.0%) 13 (2.0%) 1.47 (0.72 – 2.97) 0.29
Manual removal placenta 29 (5.4%) 26 (4.5 %) 1.18 (0.71 – 1.98) 0.52
Perineal tear
episiotomy (without tear) 150 / 540 (28%) 181 / 573 (32%) 0.88 (0.73 – 1.05) 0.17
OASIS 16 / 540 (3.0%) 21 / 573 (3.7%) 0.81 (0.43 – 1.53) 0.51
third degree tear 10 / 540 (1.9%) 14 / 573 (2.4%) 0.79 (0.40 – 1.55) 0.49
fourth degree tear 5 / 540 (0.9%) 5 / 573 (0.9%) 1.06 (0.31 – 3.64) 1.00‡
episiotomy and third degree tear 0 / 540 (0.0%) 1 / 573 (0.2%) - -
episiotomy and fourth degree tear 1 / 540 (0.2%) 1 / 573 (0.2%) 1.06 (0.07 – 16.92) 1.00‡
Maternal admission (highest level of care)
intensive care 3 (0.5%) 2 (0.3%) 1.59 (0.27 – 9.47) 0.61‡
medium care 4 (0.7%) 4 (0.6%) 1.06 (0.27 4.22) 1.00‡
ward 196 (32%) 209 (32%) 0.99 (0.85 – 1.17) 0.93
Indications for maternal admission
Thromboembolic complications 0 (0.0%) 0 (0.0%) n/a -
Hypertensive disorders§ 3 (0.5%) 13 (2.0%) 0.24 (0.07 – 0.85) 0.02‡
Post-partum blood loss 34 (5.6%) 40 (6.2%) 0.90 (0.58 – 1.40) 0.64
Post-caesarean 71 (12%) 74 (11%) 1.02 (0.75 – 1.38) 0.92
Pain treatment during labour 309 (51%) 278 (43%) 1.18 (1.05 – 1.33) 0.007
Remifentanil 93 (15%) 89 (14%) 1.11 (0.85– 1.45) 0.46
Pethidine / phenergan / other opiates 47 (7.7%) 33 (5.1%) 1.51 (0.98– 2.32) 0.06
Epidural / spinal 194 (32%) 173 (27%) 1.19 (1.00 – 1.41) 0.05
Others 1 (0.2%) 4 (0.6%) 0.26 (0.03 – 2.36) 0.38‡
493 (Adverse) maternal outcomes (per-protocol). *Defined as post-partum haemorrhage ≥ 1000 mL, and/or manual 494 removal of placenta, and/or third or fourth degree tears (Obstetrical Anal Sphincter Injuries, OASIS), and/or 495 intensive care admission, and/or maternal death. Denominator for perineal tear are vaginal deliveries only. †Mann 496 Whitney U test. ‡Fisher’s exact test. §Including (pre-)eclampsia, and HELLP syndrome.
497
Page 29 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
30
498 REFERENCES 499 1. International classification of diseases. WHO Collaborating Centers. Epidemiol Bull 500 1998;19(3):7-11.501 2. Hilder L, Costeloe K, Thilaganathan B. Prolonged pregnancy: evaluating gestation-specific 502 risks of fetal and infant mortality. Br J Obstet Gynaecol 1998;105(2):169-73.503 3. Ingemarsson I, Kallen K. Stillbirths and rate of neonatal deaths in 76,761 postterm pregnancies 504 in Sweden, 1982-1991: a register study. Acta Obstet Gynecol Scand 1997;76(7):658-62.505 4. Nakling J, Backe B. Pregnancy risk increases from 41 weeks of gestation. Acta Obstet Gynecol 506 Scand 2006;85(6):663-8. doi: 10.1080/00016340500543733507 5. Chantry AA, Lopez E. [Fetal and neonatal complications related to prolonged pregnancy]. J 508 Gynecol Obstet Biol Reprod (Paris) 2011;40(8):717-25. doi: 10.1016/j.jgyn.2011.09.007509 6. Chantry AA. [Epidemiology of prolonged pregnancy: incidence and maternal morbidity]. J 510 Gynecol Obstet Biol Reprod (Paris) 2011;40(8):709-16. doi: 10.1016/j.jgyn.2011.09.006511 7. De Los Santos-Garate AM, Villa-Guillen M, Villanueva-Garcia D, et al. Perinatal morbidity 512 and mortality in late-term and post-term pregnancy. NEOSANO perinatal network's 513 experience in Mexico. J Perinatol 2011;31(12):789-93. doi: 10.1038/jp.2011.43514 8. Hermus MA, Verhoeven CJ, Mol BW, et al. Comparison of induction of labour and expectant 515 management in postterm pregnancy: a matched cohort study. J Midwifery Womens Health 516 2009;54(5):351-6. doi: 10.1016/j.jmwh.2008.12.011517 9. ACOG Committee Opinion No 579: Definition of term pregnancy. Obstet Gynecol 518 2013;122(5):1139-40. doi: 10.1097/01.AOG.0000437385.88715.4a519 10. Vayssiere C, Haumonte JB, Chantry A, et al. Prolonged and post-term pregnancies: 520 guidelines for clinical practice from the French College of Gynecologists and 521 Obstetricians (CNGOF). Eur J Obstet Gynecol Reprod Biol 2013;169(1):10-6. doi: 522 10.1016/j.ejogrb.2013.01.026523 11. American College of O, Gynecologists. Practice bulletin no. 146: Management of late-term 524 and postterm pregnancies. Obstet Gynecol 2014;124(2 Pt 1):390-6. doi: 525 10.1097/01.AOG.0000452744.06088.48526 12. Mandruzzato G, Alfirevic Z, Chervenak F, et al. Guidelines for the management of postterm 527 pregnancy. Journal of perinatal medicine 2010;38(2):111-9. doi: 10.1515/JPM.2010.057 528 [published Online First: 2010/02/17]529 13. NICE. Induction of labour: new NICE quality standard. Midwives, 2014:8.530 14. Workgroup SAPPG. Clincal Guideline Prolonged Pregnancy, 2017.531 15. Netherlands Association for Obstetrics and Gynaecology (Nederlandse Vereniging voor 532 Obstetrie en Gynaecologie). Prolonged Pregnancy (Seroniteit), 2007.533 16. Gulmezoglu AM, Crowther CA, Middleton P, et al. Induction of labour for improving birth 534 outcomes for women at or beyond term. Cochrane Database Syst Rev 535 2012(6):CD004945. doi: 10.1002/14651858.CD004945.pub3536 17. Hussain AA, Yakoob MY, Imdad A, et al. Elective induction for pregnancies at or beyond 41 537 weeks of gestation and its impact on stillbirths: a systematic review with meta-analysis. 538 BMC Public Health 2011;11 Suppl 3:S5. doi: 10.1186/1471-2458-11-S3-S5539 18. Cheng YW, Nicholson JM, Nakagawa S, et al. Perinatal outcomes in low-risk term 540 pregnancies: do they differ by week of gestation? Am J Obstet Gynecol 2008;199(4):370 541 e1-7. doi: 10.1016/j.ajog.2008.08.008
Page 30 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
31
542 19. Middleton P, Shepherd E, Crowther CA. Induction of labour for improving birth outcomes 543 for women at or beyond term. Cochrane Database Syst Rev 2018;5:CD004945. doi: 544 10.1002/14651858.CD004945.pub4 [published Online First: 2018/05/10]545 20. Wennerholm UB, Hagberg H, Brorsson B, et al. Induction of labor versus expectant 546 management for post-date pregnancy: is there sufficient evidence for a change in clinical 547 practice? Acta Obstet Gynecol Scand 2009;88(1):6-17. doi: 10.1080/00016340802555948548 21. Keulen JKJ, Bruinsma A, Kortekaas JC, et al. Timing induction of labour at 41 or 42 weeks? 549 A closer look at time frames of comparison: A review. Midwifery 2018;66:111-18. doi: 550 10.1016/j.midw.2018.07.011 [published Online First: 2018/09/01]551 22. SFOG. Induktion av förlossning. In: Riktlinje, ed., 2016.552 23. NICE Guideline. Induction of Labour. London2008.553 24. Kortekaas JC, Bruinsma A, Keulen JK, et al. Effects of induction of labour versus expectant 554 management in women with impending post-term pregnancies: the 41 week - 42 week 555 dilemma. BMC Pregnancy Childbirth 2014;14:350. doi: 10.1186/1471-2393-14-350 556 [published Online First: 2014/10/24]557 25. Available from: https://www.toetsingonline.nl/.558 26. Gulmezoglu AM, Crowther CA, Middleton P, et al. Induction of labour for improving birth 559 outcomes for women at or beyond term. The Cochrane database of systematic reviews 560 2012;6:CD004945. doi: 10.1002/14651858.CD004945.pub3 [published Online First: 561 2012/06/15]562 27. Committee on Obstetric Practice A, American Academy of P, Committee on F, et al. ACOG 563 Committee Opinion. Number 333, May 2006 (replaces No. 174, July 1996): The Apgar 564 score. Obstet Gynecol 2006;107(5):1209-12.565 28. Ten Eikelder ML, Oude Rengerink K, Jozwiak M, et al. Induction of labour at term with oral 566 misoprostol versus a Foley catheter (PROBAAT-II): a multicentre randomised controlled 567 non-inferiority trial. Lancet 2016;387(10028):1619-28. doi: 10.1016/S0140-568 6736(16)00084-2569 29. Koopmans CM, Bijlenga D, Groen H, et al. Induction of labour versus expectant monitoring 570 for gestational hypertension or mild pre-eclampsia after 36 weeks' gestation (HYPITAT): 571 a multicentre, open-label randomised controlled trial. Lancet 2009;374(9694):979-88. doi: 572 10.1016/S0140-6736(09)60736-4573 30. Gelisen O, Caliskan E, Dilbaz S, et al. Induction of labor with three different techniques at 41 574 weeks of gestation or spontaneous follow-up until 42 weeks in women with definitely 575 unfavorable cervical scores. Eur J Obstet Gynecol Reprod Biol 2005;120(2):164-9. doi: 576 10.1016/j.ejogrb.2004.08.013577 31. OECD Health Statistics 2017. 2017578 32. Hannah ME, Hannah WJ, Hellmann J, et al. Induction of labor as compared with serial 579 antenatal monitoring in post-term pregnancy. A randomized controlled trial. The 580 Canadian Multicenter Post-term Pregnancy Trial Group. N Engl J Med 581 1992;326(24):1587-92. doi: 10.1056/NEJM199206113262402582 33. Burgos J, Rodriguez L, Otero B, et al. Induction at 41 weeks increases the risk of caesarean 583 section in a hospital with a low rate of caesarean sections. J Matern Fetal Neonatal Med 584 2012;25(9):1716-8. doi: 10.3109/14767058.2012.663018585 34. Roos N, Sahlin L, Ekman-Ordeberg G, et al. Maternal risk factors for postterm pregnancy and 586 cesarean delivery following labor induction. Acta Obstet Gynecol Scand 587 2010;89(8):1003-10. doi: 10.3109/00016349.2010.500009
Page 31 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
32
588 35. Mahomed K, Pungsornruk K, Gibbons K. Induction of labour for postdates in nulliparous 589 women with uncomplicated pregnancy - is the caesarean section rate really lower? J 590 Obstet Gynaecol 2016;36(7):916-20. doi: 10.1080/01443615.2016.1174824591 36. Ravelli AC, Schaaf JM, Eskes M, et al. Ethnic disparities in perinatal mortality at 40 and 41 592 weeks of gestation. Journal of perinatal medicine 2013;41(4):381-8. doi: 10.1515/jpm-593 2012-0228594 37. Elden H, Hagberg H, Wessberg A, et al. Study protocol of SWEPIS a Swedish multicentre 595 register based randomised controlled trial to compare induction of labour at 41 completed 596 gestational weeks versus expectant management and induction at 42 completed 597 gestational weeks. BMC Pregnancy Childbirth 2016;16:49. doi: 10.1186/s12884-016-598 0836-9599 38. Perinatal Registration Netherlands Yearbook,. 2011600 39. Lindegren L, Stuart A, Herbst A, et al. Improved neonatal outcome after active management 601 of prolonged pregnancies beyond 41(+2) weeks in nulliparous, but not among multiparous 602 women. Acta obstetricia et gynecologica Scandinavica 2017;96(12):1467-74. doi: 603 10.1111/aogs.13237604 40. Walsh D RJ, Khan K et al. NHS Research Authority, University of Nottingham, . Women’s 605 experiences of being ‘overdue’in pregnancy. NHS Research Authority, University of 606 Nottingham. 2017607 41. Moore JE, Low LK, Titler MG, et al. Moving toward patient-centered care: Women's 608 decisions, perceptions, and experiences of the induction of labor process. Birth 609 2014;41(2):138-46. doi: 10.1111/birt.12080610 42. Stevens G, Miller YD. Overdue choices: how information and role in decision-making 611 influence women's preferences for induction for prolonged pregnancy. Birth 612 2012;39(3):248-57. doi: 10.1111/j.1523-536X.2012.00554.x613 43. Zeitlin J, Blondel B, Alexander S, et al. Variation in rates of postterm birth in Europe: reality 614 or artefact? BJOG 2007;114(9):1097-103. doi: 10.1111/j.1471-0528.2007.01328.x615 44. WHO recommendations: Intrapartum care for a positive childbirth experience: World Health 616 Organization 2018.
617
Page 32 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
1
1 Figure 1 "Flowchart”
2 Randomisation, treatment and follow-up of participants
3 CONSORT 2010 Flow Diagram
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Assessed for eligibility (n=6088)
Excluded (n=4273) refused policy by randomisation (n=772) wanted induction at 41w0d (n=616) wanted expectant management until 42w0d (n=2603) other reasons (eg. planned home delivery) (n=282)
Analysed (n=900) Excluded from analysis (n=0)
Per-protocol immediate induction (ppIOL): n=611
Lost to follow-up (n=0)
Allocated to intervention: induction of labour 41w0d (n=907) Received allocated intervention (n=900) Did not receive allocated intervention: (n=7)Retreat informed consent (n=1) (before intervention started)Non eligible (n=6)
Too early/late for randomisation (n=3) (40w0d; 40w4d; 41w2d)
Incorrect gestational age and pregnancy induced hypertension (n=1) (38w0d)
Broken membranes (n=1) Maternal disease (n=1)
Lost to follow-up (n=0)
Allocated to control: policy of expectant management until 42w0d with subsequent induction (n=908) Received allocated intervention (n=901) Did not receive allocated intervention: (n=7)Non eligible (n=7)
Too early for randomisation (n=2) (39w0d; 39w4d)
Non stable-cephalic position (n=1) High blood pressure (n=1) Non- reassuring fetal status (n=2) Previous Caesarean section (n=1)
Analysed (n=901) Excluded from analysis (n=0)
Per-protocol expectant management (ppEM): n=647
Allocation
Analysis
Follow-Up
Randomised (n=1815)
Enrollment
Page 33 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only1 Figure 2 “Time to delivery (days)
2
34
Page 34 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only1 Figure 3 "Flowchart cervical ripening”
2
Page 35 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX TRIAL
INDUCTION OF LABOUR AT 41 WEEKS VERSUS EXPECTANT MANAGEMENT UNTIL
42 WEEKS
STATISTICAL ANALYSIS PLAN
EUDRA CT no. n/a
Dutch Clinical trial registry no. NTR3431
Principal investigator, centre Dr E. de Miranda
Coordinating investigator Prof. J.A.M. van der Post
Sponsor Academisch Medisch Centrum, Amsterdam
SAP version, date version 1, 24 Apr 2017
Trial methodologist Dr R.G. Duijnhoven
SAP author Dr R.G. Duijnhoven
Page 36 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 2 of 16
Page 37 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 3 of 16
Table of contents
1. Index of tables ................................................................................................................................. 5
2. Index of figures ................................................................................................................................ 5
3. SAP signatures ................................................................................................................................. 5
4. Introduction ..................................................................................................................................... 6
5. Study protocol ................................................................................................................................. 6
5.1 Background .............................................................................................................................. 6
5.2 Rationale .................................................................................................................................. 6
5.3 Objective.................................................................................................................................. 6
6. Endpoints ......................................................................................................................................... 6
6.1 Primary endpoint ..................................................................................................................... 6
6.2 Secondary endpoints ............................................................................................................... 6
7. Study methods................................................................................................................................. 7
7.1 Study design ............................................................................................................................ 7
7.2 Treatment of subjects ............................................................................................................. 7
7.3 Blinding .................................................................................................................................... 8
7.4 Randomisation procedure ....................................................................................................... 8
8. Sample-size ...................................................................................................................................... 8
9. Analysis considerations ................................................................................................................... 8
9.1 Analysis populations ................................................................................................................ 8
9.2 Covariates and Subgroups ....................................................................................................... 9
9.3 Missing Data ............................................................................................................................ 9
9.4 Interim Analyses and Data Monitoring ................................................................................... 9
9.5 Multi-centre Studies .............................................................................................................. 10
9.6 Multiple Testing ..................................................................................................................... 10
10. Presentation of study results .................................................................................................... 10
10.1 Recruitment ........................................................................................................................... 10
10.2 Protocol violations ................................................................................................................. 10
10.3 Baseline characterisations ..................................................................................................... 10
10.4 Comorbidities and conditions ............................................................................................... 10
10.5 Prior and concurrent medication .......................................................................................... 10
10.6 Treatment compliance .......................................................................................................... 10
Page 38 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 4 of 16
11. Efficacy analyses ........................................................................................................................ 10
11.1 Primary endpoint analysis ..................................................................................................... 10
11.2 Secondary efficacy analyses .................................................................................................. 10
12. Safety analyses .......................................................................................................................... 10
12.1 Adverse events ...................................................................................................................... 10
12.2 Deaths, Serious Adverse Events and other Significant Adverse Events ................................ 11
12.3 Other safety outcomes .......................................................................................................... 11
13. Other analyses ........................................................................................................................... 11
14. Tables......................................................................................................................................... 12
15. Figures ....................................................................................................................................... 16
1. Index of tables
Table 1 – Baseline characteristics. ......................................................................................................... 12
Table 2 – Perinatal outcomes. ............................................................................................................... 13
Table 3 – Delivery outcomes ................................................................................................................. 14
Table 4 – (Adverse) maternal outcomes.. ............................................................................................. 15
Page 39 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 5 of 16
SAP signatures
I give my approval for the attached SAP entitled INDEX trial dated 20 April 2017.
Primary investigator
Name: E. de Miranda
Signature: _______________________
Date: _______________________
Methodologist (statistician)
Name: R.G. Duijnhoven
Signature: _______________________
Date: _______________________
Page 40 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 6 of 16
2. Introduction
This Statistical Analysis Plan (SAP) covers the main analyses for the INDEX trial. Parallel to the trial an economic evaluation (cost-effectiveness and budget impact analysis) is conducted. The analyses for the CEA and BIA are outside the scope of this SAP.
3. Study protocol
3.1 Background
Post-term pregnancy, a pregnancy exceeding 294 days or 42 completed weeks is associated with increased perinatal morbidity and mortality. It is therefore considered a high-risk condition which requires specialist surveillance and induction of labour. There is however, uncertainty on the policy concerning the timing of induction for post-term pregnancy or impending post-term pregnancy, causing variation in practice between caregivers. Previous studies on induction at or beyond 41 weeks versus expectant management differed in results on perinatal outcome. Meta-analyses have shown better outcomes with induction at 41 weeks. Sample sizes however were limited and the (methodological) quality of the available studies was suboptimal.
3.2 Hypothesis
Expectant management until 42 weeks gestation is expected to be non-inferior regarding safety and effectiveness as compared to induction of labour at 41 weeks.
3.3 Objective
To investigate the non-inferiority of a policy of expectant management until 42 weeks gestation, as compared to induction of labour at 41 weeks in women without contra-indications for expectant management.
4. Endpoints
4.1 Primary endpoint
Primary outcome is a composite of perinatal mortality and neonatal morbidity. The composite outcome is defined as either of the following: perinatal mortality, 5-minute Apgar-score below 7, umbilical artery pH below 7.05, meconium aspiration syndrome, plexus brachialis injury (with and without association with shoulder dystocia), or admission to a Neonatal Intensive Care Unit.
4.2 Secondary endpoints
Secondary outcomes are maternal outcomes and other perinatal outcomes. Maternal outcomes include operative delivery and Caesarean section, severe perineal injury (third- or fourth-degree perineal tear), post-partum haemorrhage (≥ 1000 mL), and the use of analgesia (epidural, remifentanil, pethidin).
Page 41 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 7 of 16
5. Study methods
5.1 Study design and setting
The INDEX study is a randomised controlled non-inferiority study, investigating non-inferiority of expectant management as compared to induction of labour.
The study is set in the Dutch Obstetric Consortium: a collaboration of obstetric centres in the Netherlands in cooperation with the Midwifery Research Network of the Netherlands (MRNN). Approximately 200 centres, including university hospitals, teaching hospitals, nonteaching hospitals and midwifery practices participate in the trial.
5.2 Inclusion criteria
Women at low obstetrical risk, aged ≥ 18 years with a singleton pregnancy in stable cephalic position and an ultrasound confirmed gestational age of 40 + 5 to 41 + 0. Patients should have no contra-indications for expectant management until 42 weeks.
5.3 Exclusion criteria
Obstetrical indications for secondary care, including:
- hypertension (systolic 140 mmHg and/or diastolic 90 mmHg or more); - proteinuria (≥3 g/L); - pre-existent maternal heart or kidney diseases; - gestational diabetes; - previous Caesarean section; - multiple pregnancy; - intra-uterine growth retardation; - non-reassuring foetal status, including no foetal movements, abnormal foetal heart rate,
known foetal abnormalities which could influence perinatal outcome; - abnormal karyotype; - ruptured membranes at time of randomisation and a non-reassuring foetal status at time of
randomisation).
5.4 Treatment of subjects
5.4.1 Control group: induction of labour at 41 + 0 or 41 + 1 weeks.
Women randomised to induction of labour who are under the care of a primary care midwife, will be referred to a participating hospital for induction of labour. Induction of labour will be started at 41 + 0 to 41 + 1 weeks. Women with a cervix that is judged to be ripe at vaginal examination (Bishop score of 6 or more), will have labour induced with amniotomy followed by intravenous oxytocin according to local protocol. In case rupturing of membranes is not possible, cervical ripening will be accomplished in accordance with the national guidelines. In case the cervix is still judged to be unripe the day after priming, cervical ripening will be repeated.
Page 42 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 8 of 16
5.4.2 Intervention group: expectant management until 42 weeks.
Women allocated to expectant management await spontaneous onset of labour until week 42. If labour has not started, monitoring is according to local protocol. This reflects current care in the Netherlands. Monitoring can consist of consultations, electronic foetal heart rate monitoring and ultrasound assessment of amniotic fluid. An increase of the frequency of these checks as well as admission to the hospital is based on the judgment of the midwife or clinician in charge.
In the expectant management group, intervention will occur in case the foetal condition does not justify further expectant management, such as reduced foetal movements reported by the mother, non-optimal foetal heart rate on CTG, or oligohydramnios. If an indication for induction of labour occurs, such as pre-labour rupture of membranes for >24 hours or meconium stained amniotic fluid, referral to secondary care for labour induction is indicated according to the national Obstetrical Indication List.
Women with uncomplicated pregnancies who are still in primary care will be referred to secondary care at 42 + 0 weeks for induction of labour following the procedure as stated for the intervention group.
5.5 Blinding
The study has an open label, as it is impossible to blind either health care workers or patients for the intervention.
5.6 Randomisation procedure
Randomisation will be done using permuted blocks of sizes two and four, and stratified by recruiting site. Actual randomisation was done online using TEN-ALEA.
6. Sample-size
The sample size was calculated assuming an incidence of the primary outcome (composite adverse perinatal outcome) of 3% in the induction of labour group, and a non-inferiority margin of 2% (margin at the absolute risk at 5%). Using an one-sided alpha of 5%, a total of 1800 patients (900 per treatment group) achieves a power of 80%.
7. Analysis considerations
7.1 Analysis populations
7.1.1 Intention to Treat Population
The Intention-to-Treat (ITT) populations will consist of all patients who have been allocated one of the two treatments, irrespective of treatment received.
7.1.2 Per Protocol Population
The per protocol population is defined as:
Page 43 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 9 of 16
Expectant management
All women randomised to expectant management who: - came into labour spontaneously on or after day 41-0; - had labour induced on or after day 42-0; - had labour induced on or after day 41-0 for:
o foetal condition; o maternal condition; o ruptured membranes for 24 hours of longer.
Induction of labour
All women randomised to induction of labour who: - came in to labour spontaneously on or after day 41-0 and before day 41-3; - had labour induced on or after day 41-0 and before day 41-3.
7.2 Covariates and Subgroups
Outcomes will not be estimated with adjustment for prognostic factors, and no subgroup analyses will be conducted.
7.3 Missing Data
Missing data will not be imputed.
7.4 Interim Analyses and Data Monitoring
7.4.1 Purpose of Interim Analyses
A Data and Safety Monitoring Board will be installed, and interim analyses will be conducted for efficacy as well as to monitor sample size assumptions.
7.4.2 Planned Schedule of Interim Analyses
One interim analysis will be done after the outcomes of the first 900 patients have become available.
7.4.3 Scope of Adaptations
Extension of recruitment may be indicated if sample size the initial sample size assumptions do not hold, and a larger sample size is required.
7.4.4 Stopping Rules
The stopping rule applied is according to Haybittle-Peto, with an alpha of 0.001 for interim analyses, and 0.05 for the final analyses at the end of the trial.
7.4.5 Adjustment of Confidence Intervals and p-values
Page 44 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 10 of 16
Inferential testing as part of the interim analysis will only be done for the primary outcome (composite adverse perinatal outcome), applying an alpha of 0.001 for statistical significance.
7.5 Multicentre studies
The INDEX study is a multicentre study, recruiting patients in both primary care maternity clinics, as well as in hospitals.
The randomisation is stratified by centre. Data from all sites will be pooled for analyses.
7.6 Multiple Testing
There will be no adjustment for multiple testing.
There will be no subgroup analyses.
8. Presentation of study results
8.1 Recruitment
The recruitment of study participants will be presented using the CONSORT flow diagram.
8.2 Protocol violations
Protocol violations, especially for induction of labour in the expectant management group, will be analysed using the per-protocol population (see Per Protocol Population, 7.1.2).
8.3 Baseline characterisations
The baseline characteristics will be presented for the total population as randomised (intention-to-treat), using the format of the mock table included in paragraph 11.1.
Data will be presented using absolute numbers with percentages for discrete outcomes. Continuous outcomes will be presented as means with standard deviation, or medians with interquartile ranges.
9. Primary endpoint analysis
The main analysis for the primary outcome will be done using the intention-to-treat and per-protocol populations. Using the Farrington-Manning test, the non-inferiority hypothesis will be tested using the predetermined non-inferiority margin of 2% (risk difference).
10. Secondary endpoint analyses
10.1 Adverse perinatal outcomes
Adverse perinatal outcomes are part of the primary outcome. The factors of the composite outcome will be presented separately, together with additional secondary outcomes. These include amongst other things congenital anomaly, hypoglycaemia, neonatal infection or sepsis.
Page 45 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 11 of 16
Absolute numbers, together with percentages will be given. Relative risks with 95% confidence intervals will be calculated for dichotomous outcomes with p-values using Chi-squared test, or Fisher’s exact test. Differences in means will be given with 95% confidence intervals and p-values using the t-test (Satthertwaite as appropriate). In case off non-symmetrically distributed data, medians will be given with accompanying 95% confidence intervals and p-value using the Mann-Whitney U test or Hodges-Lehmann test.
The data will be presented using the format of the mock table included in paragraph 11.2.
10.2 Adverse maternal outcomes
Adverse maternal outcomes include amongst other things caesarean section, operative vaginal delivery, post-partum haemorrhage, intrapartum infection, use of medication, and obstetric anal sphincter injuries (OASIS).
Absolute numbers, together with percentages will be given. Relative risks with 95% confidence intervals will be calculated for dichotomous outcomes with p-values using Chi-squared test, or Fisher’s exact test. Differences in means will be given with 95% confidence intervals and p-values using the t-test (Satthertwaite as appropriate). In case off non-symmetrically distributed data, medians will be given with accompanying 95% confidence intervals and p-value using the Mann-Whitney U test or Hodges-Lehmann test.
The data will be presented using the format of the mock tables included in paragraphs 11.3 and 11.4.
Page 46 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 12 of 16
11. Tables
11.1 Baseline characteristics
Immediate induc. labour Expectant management (n=NNN) (n=NNN)
Maternal age (yrs.; mean, sd) xx.x (x.x) xx.x (x.x) Ethnicity
Caucasian NNN (xx.x%) NNN (xx.x%) Other NNN (xx.x%) NNN (xx.x%)
Highest level of education primary school NNN (xx.x%) NNN (xx.x%) secondary school NNN (xx.x%) NNN (xx.x%) lower / medium professional education NNN (xx.x%) NNN (xx.x%) higher professional education / university NNN (xx.x%) NNN (xx.x%) other / unknown NNN (xx.x%) NNN (xx.x%)
Social economic status 1 NNN (xx.x%) NNN (xx.x%) 2 NNN (xx.x%) NNN (xx.x%) 3 NNN (xx.x%) NNN (xx.x%) unknown NNN (xx.x%) NNN (xx.x%)
Parity nulliparous NNN (xx.x%) NNN (xx.x%) multiparous NNN (xx.x%) NNN (xx.x%)
Previous post-term pregnancy (≥294 days)* NNN /NNN (xx.x%) NNN /NNN (xx.x%) Level of care at time of recruitment
primary care NNN (xx.x%) NNN (xx.x%) secondary care NNN (xx.x%) NNN (xx.x%)
Bishop score at study entry Nulliparous
Bishop score ≥6 NNN / NNN (xx.x%) NNN / NNN (xx.x%) Bishop score <6 NNN / NNN (xx.x%) NNN / NNN (xx.x%) missing NNN / NNN (xx.x%) NNN / NNN (xx.x%)
Multiparous Bishop score ≥6 NNN / NNN (xx.x%) NNN / NNN (xx.x%) Bishop score <6 NNN / NNN (xx.x%) NNN / NNN (xx.x%) missing NNN / NNN (xx.x%) NNN / NNN (xx.x%)
Table 1 – Baseline characteristics. Data are presented as n (%) or mean (sd). * Numerator: primiparous and multiparous women.
Page 47 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 13 of 16
11.2 Perinatal outcomes
Immediate induc. labour
Expectant management
Relative risk (95% CI)
p-value
(n=NNN) (n=NNN)
Composite adverse perinatal outcome* NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx with 5 min. Apgar score <4 NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx
Stillbirth NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx Neonatal death, post-partum NN (x.x%) NN (x.x%) n/a - Apgar score 5 min. post-partum
<7 NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx <4 NN (x.x%) NN (x.x%) n/a -
Admission neonate to
intensive care (NICU) NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx medium care NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx
Meconium aspiration syndrome NN (x.x%) NN (x.x%) n/a -
Plexus brachialis injury NN (x.x%) NN (x.x%) n/a - Umbilical cord pH (arterial)
pH <7.05 NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx missing NN (x.x%) NN (x.x%) n/c -
Congenital anomaly NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx Hypoglycaemia NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx Neonatal infection / sepsis NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx Sex, female NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx Birth weight (g; mean, sd) x,xxx (xxx.x) x,xxx (xxx.x) x.xx (x.xx –
x.xxx
Small of gestational age <2.3rd percentile NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx <10th percentile NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx
Large for gestational age >90th percentile NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx >97.7th percentile NN (x.x%) NN (x.x%) x.xx (x.xx – x.xx) x.xxx
Table 2 – Perinatal outcomes. Data are n (%) or mean (standard deviation). Confidence intervals are 95%. NICU: Neonatal Intensive Care Unit * Composite outcome defined as: perinatal mortality, and/or a 5-minute Apgar-score below 7, and/or meconium aspiration syndrome, and/or plexus brachialis injury, and/or NICU admission. † Mean difference between groups with 95% confidence interval.
Page 48 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 14 of 16
11.3 Delivery outcomes
Immediate induc. labour
Expectant management
Relative risk (95% CI)
p-value
(n=NNN) (n=NNN)
Gestational age at delivery (days; median, IQR) xxx (xxx - xxx) xxx (xxx - xxx) xxx (xxx - xxx)* x.xxx † Time randomisation to delivery (days; mean, sd) x.x (x.x) x.x (x.x) xxx (xxx - xxx)* x.xxx †
Level of care at onset of labour Primary care NNN (x.x%) NNN (x.x%) n/c - Secondary care NNN (x.x%) NNN (x.x%) n/c -
Onset of labour Spontaneous (reference) NNN (xx.x%) NNN (xx.x%) 1.00 - Induction NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Mode of induction cervical ripening (catheter /prostaglandins) NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx amniotomy, without oxytocin NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
amniotomy, with oxytocin NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Indication for induction
randomisation NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c -
post-term pregnancy (≥42 weeks) NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c - foetal condition NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c - maternal condition NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c -
elective / maternal wish NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c - >24 h of ruptured membranes NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c - other NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/c -
Use of oxytocin NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Use of tocolytics NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Maternal intrapartum infection
fever during labour (≥38 °C) NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx use of antibiotics NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Meconium stained amniotic fluid NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Mode of delivery vaginal delivery (non OVD) NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx operative vaginal delivery NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
(secondary) caesarean section NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Indication successful operative vaginal delivery
failure to progress at second stage NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
suspected foetal distress NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx suspected foetal distress and failure to progress NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx maternal complication or other NNN / NNN (xx.x%) NNN / NNN (xx.x%) n/a -
Indication secondary caesarean section failure to progress at first stage NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx failure to progress at second stage NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
failed OVD NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx suspected foetal distress NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx susp. foetal distr. and failure to progr. at first stage NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx susp. foetal distr. and failure to progr. at sec. stage NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx maternal complication or other NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Table 3 – Delivery outcomes. Data are n (%) or mean (standard deviation). Confidence intervals are 95%.
Page 49 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 15 of 16
11.4 (Adverse) maternal outcomes
Immediate induc. labour
Expectant management
Relative risk (95% CI)
p-value
(n=NNN) (n=NNN)
Composite adverse maternal outcome* NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Maternal death NNN (xx.x%) NNN (xx.x%) n/a - Post-partum blood loss
< 1000 mL (reference) NNN (xx.x%) NNN (xx.x%) 1.00 - ≥ 1000 mL NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
1000-1500 mL NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx >1500-2000 mL NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx > 2000 mL NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Post-partum blood loss (mL; median, IQR) xxx (xxx - xxx) xxx (xxx - xxx) xxx (xxx – xxx) x.xxx† Transfusion (packed cells or plasma) NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Manual removal placenta NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx Perineal tear
episiotomy (without tear) NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx OASIS NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
third degree tear NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx fourth degree tear NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx episiotomy and third degree tear NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx episiotomy and fourth degree tear NNN / NNN (xx.x%) NNN / NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Maternal admission (highest level of care) intensive care NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx medium care NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx ward NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Indications for maternal admission thromboembolic complications NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx hypertensive disorders‡ NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx post-partum blood loss NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx post-caesarean NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Pain treatment during labour NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx remifentanil NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx pethidine / phenergan / other opiates NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx epidural / spinal NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx others NNN (xx.x%) NNN (xx.x%) x.xx (x.xx – x.xx) x.xxx
Table 4 – (Adverse) maternal outcomes. *Defined as post-partum haemorrhage ≥ 1000 mL, and/or manual removal of placenta, and/or third or fourth degree tears (Obstetrical Anal Sphincter Injuries, OASIS), and/or intensive care admission, and/or maternal death. Denominator for perineal tear are vaginal deliveries only. †Mann Whitney U test. ‡Including (pre-) eclampsia, and HELLP syndrome.
Page 50 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
INDEX trial Statistical Analysis Plan 24 April 2017 Page 16 of 16
12. Figures
12.1 Flow chart of participants
The flow of study participants will be presented using the CONSORT flow-chart for clinical trial participants.
12.2 Kaplan-Meier plots
Kaplan-Meier plot for time to delivery. Plots will be made using the intention-to-treat and per-protocol populations.
Page 51 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only
CONSORT 2010 Flow Diagram
Assessed for eligibility (n= )
Excluded (n= )
Not meeting inclusion criteria (n= )
Declined to participate (n= )
Other reasons (n= )
Analysed (n= )
Excluded from analysis (give reasons) (n= )
Lost to follow-up (give reasons) (n= )
Discontinued intervention (give reasons) (n= )
Allocated to intervention (n= )
Received allocated intervention (n= )
Did not receive allocated intervention (give
reasons) (n= )
Lost to follow-up (give reasons) (n= )
Discontinued intervention (give reasons) (n= )
Allocated to intervention (n= )
Received allocated intervention (n= )
Did not receive allocated intervention (give
reasons) (n= )
Analysed (n= )
Excluded from analysis (give reasons) (n= )
Allocation
Analysis
Follow-Up
Randomized (n= )
Enrollment
Page 52 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review OnlyManuscript number: THELANCET-D-18-05310
Article title: Induction of labour at 41 weeks versus expectant management until 42 weeks (INDEX): a multicentre, open-label, randomised non-inferiority trial
Dear Dr. de Miranda,
Thank you for submitting your manuscript to The Lancet for consideration. This is an interesting manuscript in an important area, so the editors sought external views to add to our own.
After discussing the paper further, we were unable make this submission a priority for our general readers when compared with other papers also under consideration. As a result, I regret to inform you that we have decided to decline this submission.
We appreciate the decision not to pursue your article further at The Lancet is disappointing, but we would like to suggest EClinicalMedicine as an alternative home. Having discussed your work with Dr Marta Koch ([email protected]), the Editor-in-Chief of EClinicalMedicine, she has expressed an interest in your manuscript for her journal. EClinicalMedicine is The Lancet's new gold open-access journal publishing quality clinical research. I have copied Marta in to this message to facilitate communication about the possibility of transferring your manuscript to EClinicalMedicine, if you and your co-authors agree.
We are sorry that we cannot be more positive on this occasion, but we hope this decision does not deter you from considering us again in the future.
Yours sincerely,
Dr Astrid James
Deputy Editor
The Lancet
Page 53 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review OnlyReviewers' comments:
Reviewer #1:
The present multicentre, open-label, non-inferiority RCT was performed to examine the efficacy and safety of induction of labour at 41 weeks compared to expectant management until 42 weeks. The trial participants were recruited at 123 primary care and 45 secondary care centres in the Netherlands between May 2012 and March 2016. The primary endpoint was a composite of perinatal mortality and neonatal morbidity. The manuscript is very nicely written and the interpretation of the findings is well balanced taking both the clinical and statistical significance into account. The fact the risk of perinatal adverse outcomes is low in both arms and mostly related to Apgar score is discussed in detail in terms of clinical implications. The authors may wish to consider the following comments:
1. The statistical analysis section should provide guidance on how the results should be interpreted in relation the non-inferiority margin and 95% CI i.e. that non-inferiority would be concluded if the lower 95% CI is not lower than 2.
We adapted this in the manuscript.
2. The per-protocol result of the primary endpoint should be reported in the same section with the intention to treat analysis.
We have included the per protocol primary endpoint now in the same section of the ITT analysis.
3. In non-inferiority trials, the per-protocol analysis is very important as the intention to treat analysis is potentially biased towards the null which helps concluding non-inferiority while the per-protocol analysis is potentially biased away from the null.
We added an explanation in the text of the reason why we set the focus on the intention to treat analysis: It is good practice to use a per-protocol analysis in non-inferiority trials, as an intention-to-treat analysis carries a risk of falsely rejecting the null hypothesis of inferiority. Because we did not reject the null hypothesis and do not conclude non-inferiority we presented the intention-to-treat analyses, since such analyses are more common in reports of clinical trials. All per protocol analyses are shown in the appendix.
4. Table 2: I am not sure where SGA<2.3rd percentile and LGA>97.7th percentile come from?
We add an explanation that this concerns Dutch population based birthweight percentiles.
5. The statistical analysis section in the protocol is very limited suggesting a final statistical analysis plan was developed before the end of the study. This should be submitted with the manuscript to check whether all analyses were according to the final analysis plan. Any deviations from the statistical analysis plan should be reported.
There is indeed a statistical analysis plan, we have added this to the submitted manuscript.
Page 54 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review OnlyI would have expected subgroup analyses according to maternal age considering the potential increased risk of stillbirth among advanced maternal age women. Does the data allow such a post-hoc analysis or the number of events is small in older women?
According to the analysis plan we report only outcomes of the intention to treat population and the per protocol population. However, to provide more insight we added the maternal age in the cases of foetal death.
6. The numbers needed to treat should be reported which would help the discussion and debate about the results and their potential impact on clinical practice.
We added the numbers needed to treat in the result section.
7. The Abstract should state the non-inferiority margin otherwise it is hard to interpret the results in terms of non-inferiority.
We added the non-inferiority margin in the abstract.
8. The Methods section should state whether the analysis was based on intention to treat or per-protocol. The non-inferiority analysis should be based on the per-protocol population with the intention to treat analysis as a supporting analysis.
This is explained in the methods section. See further our explanation at point 2.
9. The Findings section report that 901 women were allocated to the IOL arm while 900 were allocated to EM until 42 weeks, however, the result suggest there were 900 in the IOL arm and 901 in the EM arm.
We corrected this.
10. Looking at Figure 1 it seems there were 900 women in the IOL arm and 901 in the EM arm.
This is correct.
Reviewer #2: To authors,
This randomized study compared the perinatal outcome between induction at 41 weeks vs. expectant till 42 weeks. Although the former yielded significantly better perinatal outcome (mainly due to decreased incidence of low Apgar score) than the latter, observed poor outcome' incidence was very low in either strategy group. The authors did not conclude which, induction vs. expectant, is better but concluded that this should be informed to women to better decide which strategy they would like to choose/employ. The study is well designed and the theme is worth studying. In many developed countries, expectant till 42+0 is actually difficult. Netherland may have provided a good chance/place to study this issue. There are some limitations in this study, but the authors definitely state these limitations and, in my opinion, these limitations do not distort the present result.
Major comments
Page 55 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review OnlyI have no major advise which may improve this study/paper.
Minor comments
1. The prerequisite of this study is: 1) induction can be done safely, 2) in expectant group, the caregivers have enough ability to "picking up" the women requiring induction/cesarean. These two are mandatory factors. I believe that this should be very shortly mentioned in Discussion. For example, around line 268-280 paragraph.
Induction of labour is standard practise in obstetrical care and the safety of the procedure has been studied in many trials. We don’t think that this needs further clarification. The reviewer suggests to mention that caregivers have enough ability to ‘picking up’ the women requiring induction/caesarean. We feel that it is self-evident that care is focused on risk selection and that it is not necessary to explain this further.
2. Figure 3 indicates that 29% of women in induction group actually had spontaneous onset of labor without induction. Then, if you may have started induction at 41+2, many fractions of induction group (possibly a half or 1/3) had not received induction: induction, and thus the decision of induction vs expectant itself, was not necessary. You started induction at 41+0 or 41+1 in somewhere, and also stated 41+0 in other places. Since the actual delivery date between the two groups were only slightly differ (by 2 days), these two are important. In short, first, did you do induction at 41+0 OR (41+0 or 41+1), which? If the latter is the case, why did you not set it just 41+0? Please state the reason.
That the two groups would differ by two days became clear after analysis. We wrote in our procedure section that induction could take place at 41+0 -1. This a margin of just one day. As is known, obstetrical care is unplannable care in many cases. For logistic reasons it is not always possible to plan an elective induction on the preferable day (41+0), especially if labour rooms are required for indicated inductions or emergency care.
Second, if you may have started induction at 41+2, then, there may have been significant fraction of women who need not induction but may have had spontaneous labor onset. This scenario, although the study population may become decreased, may be worthy to discuss because actually many women will spontaneously deliver (without induction) between 41+0 and 41+2. I understand that to make the study design simple and also to obtain sufficient study population, you may have chosen this 41+0 approach. But, please shortly touch this issue for readers' better understanding. This is because the actual delivery date differs only by 2 days between the two groups.
We compared two well defined management strategies because of reasons we explained in the introduction. We feel that we have adequately reported and discussed the outcomes of this comparison, including the fact that a significant fraction of women went into labour before induction could take place (and that another fraction of women allocated to expectant management chose for induction before 42 weeks). That a part of the induction group had spontaneous onset of labour before induction could take place and a part of the expectant management group was electively induced is a result of the followed strategy. We also performed a per protocol analysis with comparable results as the intention to treat analysis as we have described in the manuscript.
3. References 14, 15, 29, 36, and 42: If you have internet address of something else, please specify these. Otherwise, readers cannot access or confirm the ones.
Page 56 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only We have adapted this.
Reviewer #3:
Thank you for the opportunity to review this manuscript. The dilemma of induction of labour at 41 weeks or expectant management until 42 weeks is a highly relevant and interesting topic. To conduct a randomised controlled trial on the topic is challenging, still a good choice of design. The workload put in this trial deserves respect. Nevertheless, I have some major concerns.
Major comments
1. The complexity of the material represents a basis for bias and insecurity:
We don’t agree with the reviewer comments regarding ‘a basis for bias and insecurity’. We estimated an effect ( a difference in outcomes between the two strategies). What the reviewer mentioned concerns the outcome but does not influence the magnitude of the effect, due to the large sample size of the trial.
2. There were 123 primary care units and 45 hospitals included for recruitment and/or treatment. Given the importance of settings for birth on birth outcome, the impact of place of birth could be considered a bias. The centers are not presented unless for the fact that it was used for stratification.
There were 123 primary care practices in which women were randomized. This concerns the level of care at time of randomization (during pregnancy) not place of birth. See our table 4: Delivery outcomes for level of care at start labour. Childbirth does not take place in midwifery practices in The Netherlands. The vast majority of births are in the hospital (87%) under responsibility of a midwife (17%, low risk birth) or a gynaecologist (70%, indication for secondary care) (Perined, 2016) .Our study can also be viewed as a comparison of outcomes of secondary maternity care versus primary maternity care in late term pregnancy for women with a low risk pregnancy. Therefore level of care and place of birth should not be considered as basis of bias, it is an inherent part of the comparison of two strategies.
3. Both primiparous and multiparous women are included without stratifying for this. Even though one could expect the numbers to be equal in the groups (which was not the case in this study), the impact of parity on birth outcome should be considered prior to commencement of the study.
We did consider this when designing the study and we assumed that randomisation in this large trial would result in an equal distribution. There are slightly more nulliparous in the EM group in our study but the distribution is by chance. As we have described in our manuscript we did not adjust for parity in our analysis, but after stratifying by parity in an additional analysis we observed similar results.
4. Various methods for induction and cervical ripening were used and differences between the groups were shown. The impact of the different methods might bias the results.
There are no significant differences between the groups regarding the used methods for cervical ripening.
Page 57 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review Only5. It is quite challenging to use "transfers of neonates to NICU" as an indication for adverse
perinatal outcome in trials with several centres without knowing if NICU was equal and available at all included centers.
The Netherlands is divided in 10 NICU regions. Neonates with a NICU indication after birth are transferred to the regional NICU by babylance (ambulance specially designed for transport of sick neonates). So NICU care is available for all neonates who need specialised intensive care.
6. It is challenging to include participants in a RCT on the topic, as shown by the large number (4273) of women who refused to participate. Nevertheless, it is important to know basic characteristics of the included and excluded women in order to strengthen the generalizability.
We agree that it is interesting to know the basic characteristics of women who refused randomisation. This will be analysed later in a separate cohort study.
7. Given the different analyses, many tables and challenging variables, it would be interesting to now if there is an statistical analysis plan presenting the prespecified outcomes and analyses, and if all analyses were conducted according to the plan.
Yes, we will include the analysis plan.
8. The per-protocol EM group (ppEM) included all women allocated to EM with spontaneous onset of labour until 42+0 weeks, all women with medical reason for IOL before 42+0 weeks during EM and women with IOL≥42+0 (page 6) The number of ppEM in the analyses is set to 647 (Appendix table), while 664 had a spontaneous onset of labour in EM group, I do not quite get the math here.
We checked the numbers and clarified this in the text.
9. Page 3, line 58-60; The standard of care in the Netherlands and in Scandinavia is presented using a guideline from the Netherlands in 2007 as reference
We added the relevant information for Sweden, Norway and Denmark.
Minor comments
10. Page 2, line 28; it seems that the numbers in each group have been switched (901 women to IOL)
We corrected this.
11. Page 5, line 123; the abbreviation DSMB is used and not explained until page 6
We corrected this.
12. Percentages are not presented with consistently number of decimals in the manuscript (eg. Page 7, lines177 and 183), nor in the tables (eg. Table 1)
Page 58 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
Confidential: For Review OnlyWe corrected this in the text.
13. The reference list needs revision. Eg. I do think that the reference nr 29 is incomplete and a web page is presented with a www. address only (23)
We corrected this.
Reviewer #4:
1. The paper is well written and very interesting when comparing induction of labour vs expectant management of pregnancies beyond 41 weeks. The question about using a non-inferiority hypothesis is original. However, I was a bit confused in the results section because you refer to the non-inferiority p and the table shows the p-value corresponding to Fisher's exact test. Could you clarify this point? In the section of statistical analysis this question could also have been better explained when using non-inferiority p.
The non-inferiority p-value is only used for the primary outcome (risk difference) and the Fisher’s exact for the other outcomes (RR).
2. In the discussion, which is very long, this whole paragraph "Our trial showed some remarkable results besides the main outcomes. Around 85% of participating women had a Caucasian ethnicity. It has been shown that women from South Asian, African and Mediterranean origin have a higher risk of perinatal mortality beyond term in comparison to Caucasian women [34]. In our study, we were not able to assess the effect of IOL in women with non-Caucasian ethnicity due to low participation rate of women of other ethnic origin " does not correspond to any objective nor can it be supported by the data and could be omitted.
It is correct that it was not an objective of our study but we mentioned this in the discussion because many readers will be interested if we have looked at ethnic disparities regarding the outcomes in this time frame of comparison.
3. There was no significant difference in perinatal mortality. While the discussion evaluates the issue of Apgar and its limitations, it could give more emphasis to the question of perinatal mortality, which is different from other studies, such as the Cochrane systematic review.
In the discussion section we have included now a reference to a recently published review of our INDEX-group in which we analysed the cases of perinatal mortality in the Cochrane systematic review according to time and cause of death.
Overall a great job and I highly recommend its acceptance.
Page 59 of 59
https://mc.manuscriptcentral.com/bmj
BMJ
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960