BMJ · 2019-11-25 · Confidential: For Review Only Induction of labour at 41 weeks versus...

Confidential: For Review OnlyInduction of labour at 41 weeks versus expectant

management until 42 weeks (the Swedish postterm induction study - SWEPIS), a multicentre, open label,

randomised, superiority trial

Journal: BMJ

Manuscript ID BMJ-2019-051538

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the Author: 11-Jul-2019

Complete List of Authors: Wennerholm, Ulla-Britt; Sahlgrenska University Hospital, Department of Obstetrics and gynecology; Sahlgrenska Academy, Institute of Clinical SciencesSaltvedt, Sissel; Karolinska Institutet, Women's and Children's Health; Karolinska University Hospital, Obstetrics and Gynaecology UnitWessberg, Anna; Sahlgrenska Academy, Instittute of Health and Care SciencesAlkmark, Mårten; Sahlgrenska University Hospital, Department of Obstetrics and gynecology; Sahlgrenska Academy, Institute of Clinical SciencesBergh, Christina; Sahlgrenska University Hospital, Department of Obstetrics and gynecology; Sahlgrenska Academy, Institute of Clinical SciencesBrismar-Wendel, Sophia; Karolinska Institute, Department of Clinical SciencesFadl, Helena; University Örebro, School of medical Health and sciences, Obstetrics and GynecologyJonsson, Maria; Uppsala University, Department of Women's and Children's HealthLadfors, Lars; Sahlgrenska University Hospital, Department of Obstetrics and gynecology; Sahlgrenska Academy, Instiutet of Clinical SciencesSengpiel, Verena; Clinical Science, Obstetrics and Gynecology, The Sahlgrenska Academy, Sahlgrenska University Hospital, GothenburgWesstrom, Jan; Uppsala University, Center for Clinical Research DalarnaWennergren, Goran; Sahlgrenska University Hospital, Department of Paediatrics; Sahlgrenska Academy, Institute of Clinical SciencesWikström, Anna-Karin; Uppsala University, Department of Women's and Children's HealthElden, Helen; Sahlgrenska Academy, Instittute of Health and Care SciencesStephansson, Olof; Clinical Epidemiology Unit, Department of Medicine, Karolinska Institutet, Medicine, SolnaHagberg, Henrik; Sahlgrenska University Hospital, Department of Obstetrics and Gynecology; Sahlgrenska Academy, Institute of Clinical Sciences

https://mc.manuscriptcentral.com/bmj

BMJ

Confidential: For Review OnlyKeywords: induction, late-term pregnancy, perinatal mortality, perinatal morbidity,

post-term pregnancy, term-pregnancy

Page 1 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Only

1

Induction of labour at 41 weeks versus expectant management until 42 weeks (the Swedish postterm induction study - SWEPIS), a multicentre, open label, randomised, superiority trial

Ulla-Britt Wennerholm1*#, Sissel Saltvedt2*, Anna Wessberg3, Mårten Alkmark1, Christina Bergh1, Sophia Brismar Wendel4, Helena Fadl5, Maria Jonsson6, Lars Ladfors1, Verena Sengpiel1, Jan Wesström7, Göran Wennergren8, Anna-Karin Wikström6, Helen Elden3**, Olof Stephansson9**, Henrik Hagberg1**

*shared first authorship; **shared senior authorship; #corresponding author1Department of Obstetrics and Gynecology, Institute of Clinical Sciences, Sahlgrenska

Academy, Gothenburg University, Sahlgrenska University Hospital, Gothenburg, Sweden

2Department of Women’s and Children’s health, Karolinska Institutet, Karolinska University

Hospital, Stockholm, Sweden3Institute of Health and Care Sciences, Sahlgrenska Academy, Gothenburg University,

Sweden

4Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm,

Sweden5Department of Obstetrics and Gynecology, Faculty of Medicine and Health, Örebro

University, Örebro, Sweden

6Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden7Center for Clinical Research Dalarna, Uppsala University, Sweden

8Department of Paediatrics, Institute of Clinical Sciences, Sahlgrenska Academy, Gothenburg

University, Sahlgrenska University Hospital, Gothenburg, Sweden

9Department of Medicine, Solna, Clinical Epidemiology Division, Karolinska Institutet,

Stockholm, Sweden

Correspondence to:

Ulla-Britt Wennerholm

Dep Obstetrics, Institute of Clinical Sciences, Sahlgrenska Hospital East

416 85 Göteborg, Sweden

[email protected]

Word count 4466

Page 2 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


2

Abstract 302 words

Objective

Neonatal mortality and morbidity, as well as maternal morbidity increase when the pregnancy

lasts more than 40 weeks, and the risks increase further as the pregnancy advances. There is

uncertainty about the obstetric management of late term pregnancies. The objective was to

evaluate if induction of labour at 41 weeks compared with expectant management and

induction of labour at 42 weeks may improve neonatal and maternal outcomes.

Design

Multicentre, open label, randomised controlled superiority trial.

Setting

Fourteen hospitals in Sweden, 2016-2018.

Participants

2760 low risk women with an uncomplicated singleton pregnancy were randomised (1:1) by

the Swedish Pregnancy Register.

Interventions

Induction of labour at 41 weeks (early induction group, n=1381), or expectant management

until induction of labour at 42 weeks (expectant management group, n=1379).

Main outcome measures

The primary outcome was a composite neonatal outcome including one or more of the

following: stillbirth, neonatal mortality, Apgar score <7 at 5 minutes, pH<7.00 or metabolic

acidosis, hypoxic ischaemic encephalopathy, intracranial haemorrhage, convulsions,

meconium aspiration syndrome, mechanical ventilation within 72 hours, and/or obstetric

brachial plexus injury. Primary analysis was by intention to treat.

Results

There was no difference in the composite primary neonatal outcome between the early

induction group 2.4% (33/1381) and expectant management group 2.3% (31/1379) (Relative

Risk [RR] 1.06; 95% confidence interval [CI] 0.65 to1.73), p=0.90). There were no perinatal

deaths in the early induction group and six perinatal deaths (five stillbirths and one early

neonatal death) in the expectant management group (p=0.032). The proportion of caesarean

delivery, instrumental vaginal delivery or any major maternal morbidity did not differ

between the groups.

Page 3 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


3

Conclusions

Induction at 41 weeks versus expectant management until 42 weeks reduces perinatal

mortality without increasing adverse maternal outcome. Labour induction should be offered

women at 41+0 weeks and could be one (of few) interventions that reduces stillbirth.

Trial registration

This trial is registered in Current Controlled Trials, ISRCTN261136529.

Keywords Induction, late-term pregnancy, perinatal mortality, perinatal morbidity, post-term

pregnancy, term-pregnancy

What is already known on this topic

Meta-analyses comparing induction of labour at or beyond term with expectant management

have shown a generally improved perinatal outcome with induction. Yet, the optimal time for

offering induction of labour is not known, particularly if induction at 41 gestational weeks

results in a better outcome compared to expectancy and induction at 42 weeks.

What this study adds

We showed that induction of labour at 41 full weeks in low risk pregnancies is associated

with a decreased risk of perinatal mortality compared with expectant management with

induction of labour at 42 full weeks. Other neonatal outcomes or caesarean delivery did not

differ between groups. Women with low risk pregnancies should be informed of the risk

profile of induction of labour versus expectant management and offered induction of labour at

41 full weeks.

Page 4 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


4

Introduction

It is well known that adverse perinatal outcomes gradually increase after 40 weeks and are

significantly elevated postterm (≥42 weeks [≥294 days]).1 The risk of stillbirth has been

shown to increase after term, and worldwide as much as 14% of stillbirth is associated with

prolonged pregnancy.2 Furthermore, maternal complications also increase with duration of

pregnancy after 40 weeks.1 To date, there is no agreement on how to manage late term (41+0–

42+0 weeks) pregnancies. The World Health Organisation (WHO) recommends induction of

labour at 41+0 weeks3 and many countries offer induction of labour between 41+0 and 42+0

weeks to avoid prolonged pregnancy.4,5

Many randomised controlled trials (RCT) have compared induction of labour with expectant

management in prolonged pregnancies, most of them with inconclusive results with respect to

perinatal mortality and major morbidity.6 The results from the latest Cochrane review from

2018 showed lower rates of caesarean delivery and perinatal death but a higher rate of

operative vaginal delivery in the induction group compared with the expectant management

group.6 Only two out of 30 included studies compared induction of labour at 41 weeks with

expectant management until 42 weeks.7,8 After the latest Cochrane review, two large RCTs

examining low risk pregnancies have been published. A large trial from the US, the ARRIVE

trial, compared induction of labour in nulliparous women at 39+0 to 39+4 weeks with

expectant management until 41+0 weeks.9 No significant difference was found in perinatal

outcome between groups while the frequency of caesarean delivery was significantly lower in

the early induction group. Another large recent trial from the Netherlands, the INDEX trial,

compared induction of labour at 41+0 to 41+1 weeks to expectant management until 42+0

weeks.10 The results could not confirm non-inferiority for adverse perinatal outcome of

expectant management, instead a significantly higher risk of adverse perinatal outcome was

found in the expectant management group. No significant difference in the rate of caesarean

delivery was found.

The objective of this study was to evaluate if induction of labour at 41+0-2 weeks compared

with expectant management until induction of labour at 42+0-1 weeks was superior in terms

of perinatal outcome in healthy women with a low risk pregnancy.

Page 5 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


5

Methods

Study design

SWEPIS (SWEdish Post term Induction Study) was a multicentre, open label, randomised

controlled superiority trial conducted in Sweden from May 2016 to October, 2018. The trial

was register-based, meaning that randomisation and most data collection was done by using

the Swedish Pregnancy Register (SPR). In all, 14 hospitals with linked antenatal clinics were

involved in the trial. Of the hospitals, four were university clinics and nine were county

hospitals comprising about 60 000 deliveries/year of the around 115 000 to 120 000

deliveries/year in Sweden. The trial was conducted according to the CONSORT guidelines

and was registered in Current Controlled Trials (ISRCTN261136529, the protocol being

available online (https://www-ncbi-nlm-nih-gov.proxy.kib.ki.se/pubmed/26951777) and as a

publication.11 The trial was undertaken within the Swedish Network for National Clinical

Studies within Obstetrics and Gynaecology (SNAKS).

Participants

Pregnant women were eligible for participation if: age ≥ 18 years; capable of understanding

oral and written information; singleton pregnancy with a fetus in cephalic presentation at

40+6 to 41+1 weeks according to ultrasound-based dating in first or early second trimester.

Exclusion criteria were: previous caesarean delivery or other uterine surgery, pre-gestational

and insulin dependent gestational diabetes, hypertensive disorder of pregnancy, known

oligohydramnios (amniotic fluid index <50 mm or deepest vertical pocket <20 mm) or small

for gestational age fetus (<-2 standard deviations [SD] according to the sex specific Swedish

reference),12 diagnosed fetal malformation, contraindication to vaginal delivery, any other

maternal condition hindering the pregnancy to proceed to 42 weeks.

General information about the study was given by posters or videos in the waiting rooms at

the antenatal clinics and by advertising in local daily newspapers. More detailed information

was provided on the study web site. At around 40 weeks, oral information about the study was

given by the midwife in Swedish or written information in any of 17 immigrant languages. In

the Stockholm region (five clinics), women were enrolled in association with a 41-week

ultrasound scan which is offered to all pregnant women in the region. This is a voluntary scan

with almost 100% coverage aiming to confirm a normal pregnancy before proceeding to 42

weeks. If the fetus had a mean abdominal diameter >110 mm and normal amniotic fluid

Page 6 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

https://www-ncbi-nlm-nih-gov.proxy.kib.ki.se/pubmed/26951777


6

volume the woman was eligible for the study. The ultrasound midwife answered questions

about the study and handled the randomisation after written informed consent was obtained.

In all other centres, women interested in taking part were invited to a visit with a research

midwife who managed patient consent and randomisation. Outside the Stockholm region, 41-

week scans were not routinely offered.

Randomisation and masking

Randomisation was done between 40+6 and 41+1 weeks. Enrolled women were allocated to

the intervention group (early induction) or the control group (expectant management group).

In the early induction group, labour was induced within 24 hours of randomisation (i. e. the

same or the next day) but not earlier than 41+0 weeks. In the expectant management group,

labour was induced in undelivered women at 42+0 to 42+1 weeks. After randomisation, no

monitoring was offered except routine follow up at the antenatal clinic consisting of blood

pressure, fundal height, and fetal heart rate monitored by dopton. Induction of labour or

caesarean delivery was performed when indicated according to local obstetric guidelines.

Allocation to either trial group, 1:1, was done online by block randomisation with a fixed

block size in the SPR (in a module specifically developed for SWEPIS by MedSciNet Inc.,

Sweden) and stratified for centre and parity (primiparity versus multiparity). Due to the nature

of the intervention, blinding was not possible for participants or caregivers, however the

statistician was blinded to group and intervention.

Procedures

Induction of labour was carried out in the same way in both groups. At admission, all women

were examined regarding blood pressure, proteinuria, fetal presentation by abdominal

palpation, cervical status, and fetal wellbeing by cardiotocography (CTG). In case of a well

engaged fetal head and a ripe cervix (Bishop score ≥ 6 for primiparas and ≥ 5 for multiparas)

amniotomy was performed, followed by oxytocin infusion after 1-2 hours without

spontaneous regular contractions. In case of an unengaged fetal head or a less ripe cervix, any

of the following methods was used, according to local routines: mechanical dilation with a

Foley-like catheter; oral misoprostol (Cytotec® or Angusta®); controlled released vaginal

Page 7 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


7

misoprostol insert (Misodel®); prostaglandin E2 vaginally (Minprostin® gel or Propess®

vaginal insert).

Outcomes

Primary outcome was a composite perinatal outcome of perinatal mortality and morbidity.

Perinatal mortality was defined as stillbirth and neonatal death (day 0-27). Neonatal

morbidity was defined as one or more of the following outcomes: Apgar score at 5 minutes

<7, pH <7.00 or metabolic acidosis (pH <7.05 and base deficit <12 mmol/l in the umbilical

artery), hypoxic ischaemic encephalopathy (HIE) I-III, intracranial haemorrhage, convulsions,

meconium aspiration syndrome, mechanical ventilation within 72 hours, obstetric brachial

plexus injury.

Secondary neonatal outcomes were: the individual components of the primary perinatal

outcome, admission to neonatal intensive care unit (NICU), macrosomia (≥4.5 kg), Apgar

score <4 at 5 minutes, therapeutic hypothermia, neonatal jaundice, neonatal hypoglycaemia,

pneumonia, sepsis, birth trauma (fracture of long bone, clavicle or skull, other neurologic

injury, retinal haemorrhage, or facial nerve palsy).

Secondary maternal outcomes were: use of epidural anaesthesia, caesarean delivery, operative

vaginal delivery, duration of active labour, chorioamnionitis, shoulder dystocia, perineal tear

III-IV, cervical laceration, uterine rupture, postpartum haemorrhage (>1000 ml), retained

placenta, postpartum infection (wound infection, endometritis, urinary tract infection, sepsis,

unspecified), preeclampsia/ hypertension, venous thromboembolism, duration of stay in

hospital, admission to intensive care unit, mortality within 42 days, breastfeeding at discharge

from hospital, and at four weeks postpartum.

Data collection

Data on maternal background, pregnancy and delivery characteristics, and neonatal outcomes

was retrieved from the SPR13 and the Swedish Neonatal Quality Register (SNQ).14 Both

registers are certified National Quality Registers initiated by Swedish healthcare

professionals. Data prospectively entered in standardized electronic medical records by

midwifes and clinicians during pregnancy, delivery and postpartum, is forwarded to the SPR

from all antenatal clinics and most delivery clinics. In the same way, the SNQ collects data on

Page 8 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


8

all newborns admitted to NICU at birth or within 28 days of life. Vital statistics on maternal

and neonatal mortality was obtained from Statistics Sweden. In all newborns with a primary

outcome the medical records were collected and scrutinised, and the same procedure was

undertaken in all women with a diagnosis of endometritis to rule out misclassification of

sepsis.

To estimate selection bias, background data and pregnancy outcome of our study population

was compared with the Swedish background population.

Safety monitoring

Before start, a Data and Safety Monitoring Board (DSMB) was formed, comprising one

statistician, one senior obstetrician and one senior midwife, with the role to supervise the trial

through periodical reviews. Serious adverse events were continuously reported by the

principle investigators to DSMB and defined as any of: perinatal or maternal death, neonate in

need of neonatal intensive care because of meconium aspiration syndrome, asphyxia,

intracranial haemorrhage or other severe condition; severe maternal morbidity with admission

to intensive care unit; complication associated with induction of labour, e.g. placental

abruption at insertion of Foley catheter or uterine rupture.

Sample size and statistical analyses

To reduce the primary outcome by one third, from 2.74% to 1.84% (superiority testing, level

of significance 0.05, power 80%, drop-out rate 10%), by induction of labour at 41 weeks as

compared with expectant management until induction at 42 weeks a sample size of 10 038

women was needed, 5019 in each randomisation group. The composite primary outcome of

2.74% was based on data on neonatal outcomes included in our primary outcome in one

Swedish region (Region Skåne) between 2000 and 2008.

The statistical analyses were carried out according to a pre-specified analysis plan by

statisticians blinded to allocation group and treatment. Main analyses were performed on the

intention to treat (ITT) population. The primary statistical analysis was the comparison

between the early induction group and the expectant management group regarding the

primary perinatal composite outcome with two-sided Fisher’s exact test at significance level

0.05. To compare secondary outcomes, Fisher’s exact test was used for dichotomous

Page 9 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


9

variables, Fisher’s non-parametric permutation test for continuous variables, Mantel Haenszel

chi-square test for ordered categorical variables and Pearsons’s chi-square test for non-

ordered categorical variables. Relative risk (RR) with 95% confidence interval (CI) between

the groups was calculated for the primary efficacy variable (the perinatal composite outcome)

and dichotomous secondary variables. Accordingly, mean differences with 95% CI between

the groups were calculated for continuous secondary variables. Data is presented by mean

with standard deviation (SD), median with 1:st and 3:rd interquartile range (IQR), number and

percentage, as appropriate.

The ITT population included all randomised women except those withdrawing consent or lost

to follow up. Women with spontaneous labour or prelabour rupture of membranes after

randomisation but before induction were included in the ITT group, as were women with

pregnancy complications necessitating interventions for medical reasons.

Complementary analyses were performed for comparison of the primary perinatal composite

outcome and secondary efficacy outcomes on the per protocol population. The per protocol

population comprised all randomised women who completed the study without significant

protocol deviation. Protocol deviation criteria were defined before start of analysing the data

and included for the early induction group: induction <41+0 weeks; labour induction,

spontaneous labour or caesarean delivery >41+2 weeks due to scheduling error or delivery

room unavailability; patient or provider preference; non-medically indicated elective

caesarean delivery. For the expectant management group protocol deviation was defined as:

induction >42+1 weeks; induction of labour <42 weeks due to scheduling error or patient or

provider preference; non-medically indicated elective caesarean delivery.

Logistic regression with treatment subgroup variable and the interaction term treatment x

subgroup variable was used to test whether the effect of treatment differed between

subgroups.

All significant tests were two-sided at the 0.05 significance level. All statistical analyses were

performed with SAS System Version 9 for Windows, SAS Inc., Cary, NC, USA.

Patient and public involvement

No patients were involved in the design, outcome measures, recruiting plans of the study, or

were asked to give advice on interpretation of results. The results of the research will be

Page 10 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


10

disseminated to the participants and general public through broadcasting, popular science

articles and newspapers. Furthermore, a survey of participant experience in the two arms of

the trial was performed which will be reported as a separate substudy.

Page 11 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


11

Results

On September 25, 2018 the DSMB recommended the steering committee of SWEPIS to stop

the study because the pre-planned interim analysis showed a statistically significant difference

in perinatal mortality between the groups. There were no perinatal deaths in the early

induction group, but six perinatal deaths in the expectant management group (five stillbirths

and one early neonatal death) (p=0.032).

Recruitment took place from May 20, 2016 through October 13, 2018. Oral and written

informed consent was obtained from 2762 women who underwent randomisation, 1383

women were assigned to induction at 41 weeks and 1379 were assigned to expectancy until

induction at 42 weeks if needed (Figure 1). Recruitment according to trial centre is shown in

in the Appendix, Supplementary Table A. After randomisation but before intervention, two

women (early induction group) withdrew their consent to participation and for their data to be

used, thus 1381 women in the early induction group and 1379 women in the expectant

management group were included in the ITT analysis.

The two groups were similar at baseline (Table 1).

Page 12 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


12

Table 1 Baseline characteristics of intention to treat population

VariableEarly induction group

(n=1381)

Expectant management group

(n=1379)

Age at randomisation, years 31.2 (4.7)31.1 (28.0; 34.6)

n=1381

31.1 (4.5)30.9 (27.9; 34.2)

n=1379

Age at randomisation

Age at randomisation <35 years 1078/1381 (78.1%) 1100/1379 (79.8%)

Age at randomisation >=35 years 303/1381 (21.9%) 279/1379 (20.2%)

Parity (includes stillborn and live births)

Nulliparous 762/1381 (55.2%) 753/1379 (54.6%)

Parous 619/1381 (44.8%) 626/1379 (45.4%)

Smoking at first antenatal visit

No 1242/1274 (97.5%) 1228/1267 (96.9%)

1-9 cig/day 25/1274 (2.0%) 30/1267 (2.4%)

10 or more cig/day 7/1274 (0.5%) 9/1267 (0.7%)

Alcohol screening (AUDIT) at first antenatal visit*

0-5 points (low risk) 1102/1197 (92.1%) 1109/1190 (93.2%)

> = 6 (risk behavior) 95/1197 (7.9%) 81/1190 (6.8%)

Medical history

Psychiatric disease 95/1001 (9.5%) 108/997 (10.8%)

Prepregnancy diabetes mellitus type 1 or 2 2/1365 (0.1%) 0/1355 (0.0%)

Endocrine disease 83/1362 (6.1%) 97/1355 (7.2%)

Chronic hypertension 2/1362 (0.1%) 1/1353 (0.1%)

Height (cm) at first antenatal visit 167.5 (6.2)167 (163; 172)

n=1275

167.6 (5.9)168 (163; 172)

n=1273

Weight (kg) at first antenatal visit 70.0 (14.3)67 (60; 77)

n=1247

70.7 (14.5)68 (60; 77)

n=1241

BMI at first antenatal visit 24.9 (4.7)23.9 (21.6; 27.1)

n=1275

25.1 (4.9)24.0 (21.7; 27.4)

n=1265

BMI at first antenatal visit <30 1118/1275 (87.7%) 1081/1265 (85.5%)

BMI at first antenatal visit >=30 157/1275 (12.3%) 184/1265 (14.5%)

Last recorded weight during pregnancy, kg 83.5 (14.4)82.0 (74.0; 91.0)

n=1344

84.0 (14.6)82.0 (74.0; 92.0)

n=1336

Region of birth

Sweden 1069/1289 (82.9%) 1070/1298 (82.4%)

Other Nordic countries 74/1289 (5.7%) 84/1298 (6.5%)

Europe outside Nordic countries 20/1289 (1.6%) 18/1298 (1.4%)

Outside Europe 126/1289 (9.8%) 126/1298 (9.7%)

Highest education

Shorter than 9 years 7/1221 (0.6%) 10/1242 (0.8%)

At least 9 years primary school 47/1221 (3.8%) 47/1242 (3.8%)

High school 9 to 12 years 378/1221 (31.0%) 405/1242 (32.6%)

University or corresponding 789/1221 (64.6%) 780/1242 (62.8%)

Employment status

Employed 1098/1296 (84.7%) 1105/1301 (84.9%)

Student 101/1296 (7.8%) 101/1301 (7.8%)

Maternity leave 47/1296 (3.6%) 52/1301 (4.0%)

Page 13 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


13


(n=1381)


(n=1379)

Unemployed 17/1296 (1.3%) 18/1301 (1.4%)

Sick leave 15/1296 (1.2%) 9/1301 (0.7%)

Other 18/1296 (1.4%) 16/1301 (1.2%)

Cohabitation with partner 1215/1360 (89.3%) 1211/1353 (89.5%)

Living alone/One person household 19/1360 (1.4%) 21/1352 (1.6%)

Mode of conception

Assisted conception (IVF/ICSI) 67/1381 (4.9%) 53/1379 (3.8%)

Involuntary childlessness 160/1250 (12.8%) 146/1199 (12.2%)

BMI=body mass index, IVF= in vitro fertilization, ICSI=intracytoplasmic sperm injectionFor categorical variables n/N (%) is presented.For continuous variables Mean (standard deviation [SD]) / Median (interquartile range [IQR] Q1; Q3) / n= is presented.

*alcohol screening by AUDIT tool according to antenatal care routines15

Compared with the Swedish background population, women in the study groups had a higher

level of education and were more often born in Sweden (Appendix, Supplementary Table B).

In the early induction group, 14.1% (195/1381) of the women had spontaneous onset of

labour, 85.5% (1181/1381) underwent induction, of whom 76.6% (905/1181) had cervical

ripening, and 0.4% (5/1381) had a primary caesarean delivery (Table 2).

Page 14 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


14

Table 2 Delivery and maternal adverse outcomes in intention to treat population

Variable

Early induction group

(n=1381)


(n=1379)Relative Risk

(95% CI) p-value

Difference between groups

Mean (95% CI)

Gestational age at delivery (days) 288.8 (1.3)289

(288; 289)n=1381

291.7 (2.7)292

(289; 294)n=1379

<0.0001 -2.91 (-3.07; -2.76)

Time from randomization to delivery (days) 1.76 (1.42)2

(1; 2)n=1381

4.66 (2.64)4

(2; 7)n=1379

<0.0001 -2.91 (-3.06; -2.75)

Time from admittance to labour ward to delivery (hours)

20.1 (14.8)16.2

(9.2; 27.9)n=1380

13.6 (12.2)10.4

(4.6; 19.0)n=1378

<0.0001 6.49 (5.50; 7.50)

Preeclampsia/gestational hypertension/eclampsia

19/1381 (1.4%) 42/1379 (3.0%) 0.45 (0.26; 0.77) 0.0040

Onset of delivery

Spontaneous 195/1381 (14.1%) 920/1379 (66.7%) 0.21 (0.18; 0.24) <0.0001

Primary caesarean delivery 5/1381 (0.4%) 2/1379 (0.1%) 2.50 (0.49; 12.85) 0.45

Induction 1181/1381 (85.5%) 457/1379 (33.1%) 2.58 (2.39; 2.79) <0.0001

Mode of induction

Cervical ripening 905/1181 (76.6%) 340/457 (74.4%)

Amniotomy without oxytocin 130/1181 (11.0%) 45/457 (9.8%)

Amniotomy with oxytocin 146/1181 (12.4%) 72/457 (15.8%) 0.18

Cervical ripening

First method mechanical 343/905 (37.9%) 126/340 (37.1%) 0.99 (0.90; 1.09)

First method pharmacological 562/905 (62.1%) 214/340 (62.9%) 0.99 (0.90; 1.09) 0.84

Indication for induction

Randomisation to 41 weeks and induction according to protocol

1146/1181 (97.0%) 0/457 (0.0%)


0/1181 (0.0%) 373/457 (81.6%)

Maternal condition 1/1181 (0.1%) 41/457 (9.0%)

Fetal condition 0/1181 (0.0%) 18/457 (3.9%)

Maternal request* 7/1181 (0.6%) 23/457 (5.0%)

Other† 28/1181 (2.4%) 5/457 (1.1%) <0.0001

Use of epidural anaesthesia 729/1381 (52.8%) 669/1379 (48.5%) 1.09 (1.01; 1.17) 0.027

Chorioamnionitis 2/1381 (0.1%) 6/1379 (0.4%) 0.33 (0.07; 1.65) 0.29

Fever during labour 127/1381 (9.2%) 105/1379 (7.6%) 1.21 (0.94; 1.55) 0.15

Antibiotics during labour 248/1381 (18.0%) 262/1379 (19.0%) 0.95 (0.81; 1.11) 0.51

Meconium stained amniotic fluid 233/1238 (18.8%) 320/1127 (28.4%) 0.66 (0.57; 0.77) <0.0001

Duration of active labour 7.13 (5.39)5.67

(2.85; 10.28)n=717

8.32 (5.94)6.86

(3.76; 11.45)n=880

<0.0001 -1.19 (-1.76; -0.64)

Non-instrumental vaginal delivery 1150/1381 (83.3%) 1140/1379 (82.7%) 1.01 (0.97; 1.04) 0.71

Caesarean delivery 143/1381 (10.4%) 148/1379 (10.7%) 0.96 (0.78; 1.20) 0.79

Assisted vaginal delivery 88/1381 (6.4%) 91/1379 (6.6%) 0.97 (0.73; 1.28) 0.87

Secondary caesarean delivery 138/143 (96.5%) 146/148 (98.6%) 0.98 (0.94; 1.01) 0.42

Indication secondary caesarean delivery

Failed induction 8/138 (5.8%) 7/146 (4.8%)

Failure to progress at first stage 60/138 (43.5%) 53/146 (36.3%)

Page 15 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


15

Variable


(n=1381)



(95% CI) p-value


Mean (95% CI)

Fetal distress at first stage 35/138 (25.4%) 28/146 (19.2%)

Failure to progress and fetal distress at first stage

6/138 (4.3%) 7/146 (4.8%)

Other indication at first stage 5/138 (3.6%) 5/146 (3.4%)

Failure to progress at second stage 11/138 (8.0%) 23/146 (15.8%)

Fetal distress at second stage 6/138 (4.3%) 8/146 (5.5%)

Failure to progress and fetal distress at second stage

0/138 (0.0%) 3/146 (2.1%)

Failure operative vaginal delivery 7/138 (5.1%) 12/146 (8.2%) 0.27

Indication assisted vaginal delivery

Failure to progress 37/88 (42.0%) 28/91 (30.8%)

Fetal distress 33/88 (37.5%) 34/91 (37.4%)

Failure to progress and fetal distress 5/88 (5.7%) 7/91 (7.7%)

Maternal distress 13/88 (14.8%) 21/91 (23.1%)

Other 0/88 (0.0%) 1/91 (1.1%) 0.35

Duration of hospital stay from delivery to discharge (hours)

46.3 (27.0)43.6

(25.3; 61.6)n=1333

47.1 (29.7)44.8

(25.8; 61.2)n=1333

0.46 -0.82 (-2.99; 1.32)

Delivery and postpartum complications

Uterine rupture 0/1381 (0.0%) 0/1379 (0.0%) 1.00

Shoulder dystocia 6/1381 (0.4%) 4/1379 (0.3%) 1.50 (0.42; 5.30) 0.76

Cervical laceration 5/1381 (0.4%) 8/1379 (0.6%) 0.62 (0.20; 1.90) 0.58

Perineal lacerations III-IV 40/1381 (2.9%) 50/1379 (3.6%) 0.80 (0.53; 1.20) 0.33

Postpartum haemorrhage (>1000 ml) 140/1381 (10.1%) 146/1379 (10.6%) 0.96 (0.77; 1.19) 0.75

Maternal death 0/1381 (0.0%) 0/1379 (0.0%) 1.00

Maternal admission to intensive care unit 2/1381 (0.1%) 0/1379 (0.0%) 0.50

Postpartum infection‡ 42/1381 (3.0%) 27/1379 (2.0%) 1.55 (0.96; 2.50) 0.088

Wound infection 4/1381 (0.3%) 3/1379 (0.2%) 1.33 (0.30; 5.94) 1.00

Urinary tract infection incl. pyelonephritis 5/1381 (0.4%) 7/1379 (0.5%) 0.71 (0.23; 2.24) 0.77

Endometritis 18/1381 (1.3%) 6/1379 (0.4%) 3.00 (1.19; 7.52) 0.022

Sepsis 0/1381 (0.0%) 0/1379 (0.0%) 1.00

Venous thromboembolism 0/1381 (0.0%) 1/1379 (0.1%) 1.00

Breast feeding at discharge from delivery hospital

987/1019 (96.9%) 981/1014 (96.7%) 1.00 (0.99; 1.02) 0.98

Breast feeding 4 weeks after delivery 835/923 (90.5%) 824/939 (87.8%) 1.03 (1.00; 1.06) 0.071

CI= confidence intervalFor categorical variables n/N (%) is presented.For continuous variables Mean (standard deviation [SD]) / Median interquartile range [IQR] Q1; Q3) / n= is presented.For comparison between groups Fisher´s Exact test (lowest 1-sided p-value multiplied by 2) was used for dichotomous variables Fisher´s Exact test (2-sided) was used for dichotomous variables and Fisher´s Non-Parametric Permutation test was used for continuous variables.*maternal request outside study protocol e.g. women in the early induction group were induced on request after 41+1 weeks and women in the expectant management group were induced on request before 42+0 weeks†induction outside study protocol due to administrative errors or lack of capacity on labour ward‡ includes women with any of international classification of disease (ICD) 10 codes O85.9 (endometritis/sepsis) or O86 (wound infection, urinary tract infection incl. pyelonephritis, fever of unknown etiology postpartum and other postpartum infections

In the expectant management group, 66.7% (920/1379) of the women had spontaneous onset

of labour, 33.1% (457/1379) were induced, of whom 74.4% (340/457) had cervical ripening,

and 0.1% (2/1379) had a primary caesarean delivery. In 3.5% (48/1381) of women in the

Page 16 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


16

early induction group and in 2.0% (28/1379) of women in the expectant management group

the management was not consistent with the assigned strategy (Figure 1).

Median (IQR) time from randomisation to delivery was 2 (1; 2) days in the early and 4 (2; 7)

days in the expectant management group (p<0.0001) (Table 2, Figure 2). Median (IQR)

gestational age at delivery was 289 (288; 289) days in the early and 292 (289; 294) days in the

expectant management group (p<0.0001).

Primary outcome

The primary outcome occurred in the early induction group in 2.4% (33/1381) and in 2.2%

(31/1379) in the expectant management group (RR 1.06; 95% CI 0.65 to 1.73; p=0.90) (Table

3).

Page 17 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


17

Table 3 Perinatal outcome in intention to treat groups

Variable


(n=1381)



(95% CI) p-value


Mean (95% CI)

Primary composite outcome 33/1381 (2.4%) 31/1379 (2.2%) 1.06 (0.65; 1.73) 0.90

Sub-components of the primary composite outcome

Perinatal/neonatal mortality (stillbirth + neonatal mortality)

0/1381 (0.0%) 6/1379 (0.4%) 0.031

Stillbirth 0/1381 (0.0%) 5/1379 (0.4%) 0.062

Neonatal mortality (Live births with death day 0-27)

0/1381 (0.0%) 1/1378 (0.1%) 1.00

Neonatal morbidity 33/1381 (2.4%) 26/1374 (1.9%) 1.27 (0.76; 2.11) 0.43

Neonatal morbidity defined as at least one of the following variables:

Apgar score <7 at 5 minutes* 18/1381 (1.3%) 16/1374 (1.2%) 1.12 (0.57; 2.19) 0.88

Metabolic acidosis (denominator based on validated umbilical cord blood samples at birth)†

14/661 (2.1%) 10/644 (1.6%) 1.36 (0.61; 3.05) 0.58

Hypoxic ischaemic encephalopathy (HIE) I-III

2/1381 (0.1%) 3/1374 (0.2%) 0.66 (0.11; 3.96) 1.00

Intracranial haemorrhage 1/1381 (0.1%) 2/1374 (0.1%) 0.50 (0.05; 5.48) 1.00

Neonatal convulsions 1/1381 (0.1%) 3/1374 (0.2%) 0.33 (0.03; 3.18) 0.62

Meconium aspiration syndrome (MAS) 2/1381 (0.1%) 3/1374 (0.2%) 0.66 (0.11; 3.96) 1.00

Mechanical ventilation within first 72 hours 3/1381 (0.2%) 5/1374 (0.4%) 0.60 (0.14; 2.49) 0.72

Obstetric brachial plexus injury 4/1381 (0.3%) 1/1374 (0.1%) 3.98 (0.45; 35.56) 0.38

Additional secondary neonatal outcomes variables

Admittance to neonatal intensive care units (NICU)

55/1381 (4.0%) 82/1374 (6.0%) 0.67 (0.48; 0.93) 0.021

Days at neonatal intensive care units (NICU)

3.38 (2.97)2

(1; 6)n=55

4.59 (5.64)3 (0; 31)

(1; 6)n=81‡

0.15 -1.21 (-2.78; 0.20)

Admittance to neonatal intensive care units (NICU) >4 days

34/55 (61.8%) 45/81 (55.6%) 1.11 (0.84; 1.48) 0.58


Therapeutic hypothermia 1/1381 (0.1%) 2/1374 (0.1%) 0.50 (0.05; 5.48) 1.00

Neonatal jaundice requiring phototherapy or exchange transfusion

16/1381 (1.2%) 32/1374 (2.3%) 0.50 (0.27; 0.90) 0.027

Hypoglycaemia§ 22/1381 (1.6%) 20/1374 (1.5%) 1.09 (0.60; 2.00) 0.89

Pneumonia 8/1381 (0.6%) 13/1374 (0.9%) 0.61 (0.25; 1.47) 0.38

Sepsis 9/1381 (0.7%) 20/1374 (1.5%) 0.45 (0.20; 0.98) 0.058

Birth trauma¶ 0/1381 (0.0%) 1/1374 (0.1%) 1.00

Female 600/1381 (43.4%) 623/1379 (45.2%) 0.96 (0.88; 1.05) 0.38

Birthweight (g) 3815 (409)3804

(3536; 4090)n=1381

3875 (436)3865

(3570; 4160)n=1379

0.0002 -60.1 (-91.8; -29.6)

Small for gestational age (SGA)** 9/1381 (0.7%) 22/1379 (1.6%) 0.41 (0.19; 0.88) 0.028

Large for gestational age (LGA)** 21/1381 (1.5%) 26/1379 (1.9%) 0.81 (0.46; 1.43) 0.55

Macrosomia (>= 4500 g) 68/1381 (4.9%) 114/1379 (8.3%) 0.60 (0.45; 0.80) 0.0005

Any major birth defect†† 14/1381 (1.0%) 17/1379 (1.2%) 0.82 (0.41; 1.66) 0.72

Page 18 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


18

Variable


(n=1381)



(95% CI) p-value


Mean (95% CI)

CI=intervalFor categorical variables n/N (%) is presented.For continuous variables Mean (standard deviation [SD]) / Median interquartile range [IQR] Q1; Q3) / n= is presentedFor comparison between groups Fisher´s Exact test (lowest 1-sided p-value multiplied by 2) was used for dichotomous variables Fisher´s Exact test (2-sided) was used for dichotomous variables and Fisher´s Non-Parametric Permutation test was used for continuous variables *Apgar score of live births†validated umbilical cord samples defined as values of arterial pH <venous pH and velues of arterial pCO2 >venous pCO2‡one infant with only home based neonatal care§hypoglycaemia defined as P-glucose concentration < 2,6 mmol/L after 3 hours¶defined as any of long bone fracture, clavicular fracture, skull fracture, other neurological injury, retinal haemorrhage, facial nerve palsy**SGA and LGA defined as <-2 SD and >2 SD, respectively, according to the Swedish sex specific reference12

††minor birth defects according to EUROCAT definition excluded16

There were no cases of stillbirth or neonatal mortality (0-27 days) in the early induction group

(mortality rate 0.0%), whereas there were five stillbirths and one neonatal death (mortality

rate 0.4%) in the expectant management group (p=0.032) occurring between 41+2 weeks and

41+6 weeks. One stillbirth occurred at the labour ward close after admittance. The

postmortem examination showed a cardiovascular malformation which according to

specialists in paediatric cardiology could not be considered as lethal. There were no

explanations to the stillbirths in the other four cases. One stillborn neonate was small for

gestational age, the other stillborns had birthweights within normal range. The neonatal death

occurred due to HIE in a neonate being large for gestational age.

A low Apgar score (<7 at 5 minutes) was the main contributor to the primary outcome: 1.3%

(18/1381) in the early compared with 1.2% (16/1374) in the expectant management group

(RR 1.12; 95% CI 0.57 to 2.19; p=0.88).

Secondary perinatal outcomes

Secondary perinatal outcomes are shown in Table 3. Fewer neonates in the early induction

group were admitted to NICU; 4.0% (55/1381) in the early group versus 6.0% (82/1374) in

the expectant management group (RR 0.67; 95% CI 0.48 to 0.93; p=0.021). If neonates with a

major birth defect (n=10) were excluded (antenatally detected major birth defect was an

exclusion criterion at study entry) there was no significant difference in admittance to NICU.

Fewer neonates in the early induction group had hyperbilirubinemia treated with phototherapy

or exchange transfusion; 1.2% (16/1381) in the early group versus 2.3% (32 (1374) in the

expectant management group (RR 0.50; 95% CI 0.27 to 0.90; p=0.027). Macrosomia were

lower in the early induction group than in the expectant management group (4.9% [68/1381

Page 19 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


19

versus 8.3% [114/1379], RR 0.60; 95% CI 0.45 to 0.80; p<0.001). Other secondary outcomes

did not differ.

Secondary maternal outcomes

Secondary maternal outcomes are presented in Table 2. Hypertensive disorder of pregnancy

after randomisation was lower in the early induction group compared with the expectant

management group (1.4% [19/1381] versus 3.0% [42/1379], RR 0.45; 95% CI 0.26 to 0.77;

p=0.0040). Use of epidural anaesthesia was higher in the early induction compared with the

expectant management group, 52.8% (729/13819) versus 48.5% (669/1379) (RR 1.09; 95%

CI 1.01 to 1.17; p=0.027). Median (IQR) duration of active labour was shorter 5.7 hours (2.9;

10.3) in the early compared with 6.9 hours (3.8; 11.5) in the expectant management group,

respectively (p<0.0001). Mode of delivery was similar in both groups with rate of caesarean

delivery of 10.4% (143/1381) in the early induction group and 10.7% (148/1379) in the

expectant management group (RR 0.96; 95% CI 0.78 to 1.20); p=0.79). Indications for

caesarean delivery did not differ between the groups.

Endometritis occurred in 1.3% (18/1381) and 0.4% (6/1379) in the early induction and

expectant management groups (RR 3.0; 95% CI 1.19 to 7.52); p=0.022), respectively. Other

adverse maternal outcomes including postpartum haemorrhage and perineal lacerations grade

III and IV were similar in both groups.

Per protocol analysis

The prespecified analysis of the per protocol population included 1333 women in the early

induction group and 1351 women in the expectant management group. Reasons for protocol

violation are shown in Figure 1. Baseline characteristics were similar between the groups

(Appendix, Supplementary Table C). The primary perinatal adverse outcome occurred in 31

cases in the early induction group and in 31 cases in the expectant management group (RR

1.01; 95% CI 0.62 to 1.66; p=1.0) (Appendix, Supplementary Table E). There were no cases

of stillbirth or neonatal mortality (0-27 days) in the early induction group (mortality rate

0.0%), whereas there were five stillbirths and one neonatal death (mortality rate 0.4%) in the

expectant management group (p=0.032).

Secondary neonatal and maternal outcomes are shown in Appendix, Supplementary Table D

and Table E.

Subgroup analyses

Page 20 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


20

Prespecified subgroup analyses on the primary outcome and selected secondary outcomes

according to parity (parity 1 versus parity>1), maternal age (<35 years versus ≥35 years) and

BMI (BMI <30 versus BMI ≥30) were performed on the ITT population. In the ITT

population, analyses of primary outcome showed no significant difference in the treatment

effect according to parity, age or BMI (p=0.29, p=0.70, p=0.51, respectively, for the

interaction). In total, there were five stillbirths and one early neonatal death, and all of them

occurred in the expectant management group; in nulliparous 0.8% (6/753) versus in parous

women 0% (0/626), in women ≥35 years 1.1% (3/279) versus in women <35 years 0.3%

(3/1100) and in women with BMI ≥30 1.1% (2/184) versus in women with BMI <30 0.4%

(4/1081). Due to the low mortality rate (n=6) no interaction analysis on mortality could be

performed. Among nulliparous women, rate of caesarean delivery was 16.7% (127/762) in the

early induction and 17.3% (130/753) in the expectant management group (p=0.81).

When testing if the effect of early induction versus expectant management was similar across

sites (Stockholm centres versus other centres i.e. offering a routine ultrasound scan at 41+0

weeks or not) no significant interaction effect was found for the primary outcome (p=0.19) in

the ITT population. Perinatal mortality in the expectant management group was 0.0% (0/557)

in Stockholm versus 0.7% (6/822) in the other centres.

Discussion

In this large randomised trial, comparing induction of labour at 41 weeks versus expectant

management until induction at 42 weeks, we found no significant difference in the primary

composite adverse perinatal outcome, 33/1381 (2.4%) and 31/1379 (2.2%) in the early

induction and expectant management groups, respectively (RR 1.06; 95% CI 0.65 to 1.73,

p=0.90). Yet, the perinatal mortality was significantly lower in the early induction group (no

deaths) compared with the expectant management group (five intrauterine deaths, one

neonatal death) (p=0.032). Furthermore, no significant difference in caesarean delivery rates

was observed between groups, 10.4 % and 10.7% in the induction and expectant groups,

respectively (p=0.79).

Comparison with previous studies

Postterm (≥42 weeks) pregnancy is associated with an increased risk of adverse perinatal

morbidity and mortality.17-19 In fact, the risk appears to increase gradually after 39 weeks.17,18

Page 21 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


21

Results from most meta-analyses indicate that a policy of induction before 42 full weeks is

associated with decreased perinatal mortality.6, 20-22

The benefit of early induction is supported by a recently published open label multicentre

randomised trial (INDEX trial) from the Netherlands including 1801 women, in which

induction at 41 weeks was associated with a lower composite adverse perinatal outcome

(1.7%) compared with expectant management until 42 weeks (3.1%) (p=0.045).10 However,

the perinatal mortality rate did not differ significantly between the groups with one death in

the 41 weeks group and two in the 42 weeks group.

It could be argued that the higher mortality in the expectant management group in

our study is partly due to that fetal surveillance with CTG/ultrasound between 41

weeks and 42 weeks was not pursued routinely unless clinical signs of complications

occurred. This argument is supported by the fact that the perinatal death rate in the

expectant management group tended to be higher in the SWEPIS trial (6/1379;

0.44%) compared with the INDEX trial (2/901; 0.22%) where pregnancies in the

expectant management group were often monitored with CTG and/or amniotic fluid

estimation with ultrasound. Yet, the adverse perinatal outcomes in general were not

higher in the expectant group in the SWEPIS compared with the INDEX trial and the

median gestational age at delivery was higher in the expectant group in our trial (292

days) than in the INDEX trial (289 days) which could augment mortality rates.

Furthermore, no perinatal deaths occurred among women recruited in the Stockholm

region, where all women are offered a routine ultrasound at week 41+0 (before

randomisation), aiming to identify women with an elevated risk for adverse outcome.

However, the rarity of perinatal death limits the power of a subanalysis by centre.

Further, two of the five HIE cases occurred in Stockholm and the composite neonatal

Page 22 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


22

morbidity was similar in Stockholm (24/1122=2.1%) as in the other centres

(34/1632=2.1%) which does not support that the 41-week ultrasound was critical. It is

also uncertain to what extent ultrasound or CTG recordings usually performed at two

or three day intervals can prevent intrauterine or neonatal deaths4,23,24 and the

evidence supporting that fetal monitoring prevents complications of post-maturity is

considered weak.4

The occurrence of endometritis was significantly higher in the 41 weeks group than in the

expectant management group which was unexpected but could relate to the much higher rate

of induction of labour in the early induction (86%) compared with the expectant (33%) arm.

Indeed one meta-analysis suggests that the risk of postpartum infections is slightly increased

after induction in particular with balloon dilation with Foley catheters.25 However, more

recent studies indicate that the infectious morbidity is not higher for mechanical than

pharmacological methods for cervical dilation26 and the occurrence of endometritis is similar

or lower in our trial than reported in most studies on labour induction.25–27 Furthermore, the

frequency of other maternal (chorioamnionitis, wound infections, urinary tract infections) and

neonatal (sepsis, pneumonia) infections was not higher in the early induction group.

Strengths and weaknesses of the trial

Strengths of the present study is that it concerns a large national multicentre randomised

controlled trial comparing induction at 41 weeks with expectant management until induction

at 42 weeks, the latter being standard of care in Sweden at present. Despite the fact that only a

minority of eligible women were informed or accepted participation (Figure 1), the study

population was representative of a Swedish low risk population according to most baseline

characteristics (Appendix, Supplementary Table B). Another strength is that the study

participants were managed at the same level of care and methods of induction were applied

irrespective of allocation arm, which was not always the case in previous randomised trials on

postterm pregnancies.10, 28

There were some limitations of the present trial. The study was closed prematurely based on

the recommendations of the DSMB because of the significantly higher perinatal mortality in

the expectant management group; i.e. at a time point when only ~1/3 of the intended number

Page 23 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


23

of women had been recruited. It is therefore difficult to state whether the primary outcome

(composite perinatal morbidity and mortality) would have differed between the groups.

However, perinatal mortality must be considered as a more serious and important outcome

compared to the most frequently occurring components in the primary composite outcome

such as Apgar <7 at 5 minutes and metabolic acidosis (Table 3)29.

It could appear contradictory that a significant difference was found between groups in

perinatal mortality but yet no difference in the composite adverse neonatal outcome.

However, 5/6 deaths were stillbirths in the SWEPIS trial, which have a quite different

aetiology and array of risk factors30 as compared with neonatal mortality/morbidity.31

Placental abnormality/dysfunction, umbilical cord complications, and growth restriction are

considered causes of stillbirth2,30 that could well be of increasing importance in late

term/postterm pregnancy. The fact that half of the women (those recruited in the Stockholm

region) underwent ultrasound measurement of amniotic fluid volume and abdominal

diameter at 41 weeks whereas such examinations were not performed systematically

at the other centres might be regarded both as a limit and a strength. It is difficult to

determine whether outcomes were affected by this difference in policy (see above)

while such a management increases generalisability and reflects current obstetric

practice in Sweden.32

Implications for clinicians and policy makers

In conclusion, the SWEPIS trial shows that the perinatal mortality was significantly lower in

the early induction group but also that the occurrence of mortality and serious morbidity was

relatively low. The number needed to treat with induction of labour at 41 weeks to prevent

one perinatal death was 250 which is lower than previous estimates.6, 20, 21 Furthermore, the

number of caesarean deliveries, operative vaginal deliveries or any major maternal morbidity

did not differ between the early induction and expectant management groups, suggesting that

induction at 41 weeks has few medical disadvantages. Based on previous reports and the

results of the present trial, we suggest that labour induction should be offered women at 41+0

weeks and could be one (of few) interventions that reduces stillbirth.

Page 24 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


24

Contributors: UBW and SS are joint first authors and contributed equally to the study.

UBW, HH, VS and HE conceived and designed the study. UBW, HH, AW, SS, AKW, MJ,

HF and JW oversaw recruitment of study participants and collection of data at the local

centres. UBW, HH, CB, HE, OS and SS wrote the Statistical Analysis Plan together with two

statisticians (Mattias Molin, Nils-Gunnar Pehrson, the Statistical Consulting Group,

Gothenburg). UBW and MA did the data cleaning together with statistician Mattias Molin and

Per Ekman. UBW, HH, CB, SS, MA, LL, VS, SBW, OS, GW, HE and AW interpreted the

data. UBW, MA, AW, SS and HH wrote the first draft of the manuscript, which was then

critically reviewed and revised by the other co-authors. All authors approved the final version

of the manuscript for submission.

The SWEPIS study group: The midwives and doctors responsible at the local centres:

Uppsala University Hospital: Irina Sylwe; South Älvsborg Hospital: Lena Loubelo, Carolina

Bergerum, Serney Bööj; Department of Gynecology Närhälsan, Mölndal: Maria Bullarbo;

Sahlgrenska University Hospital, Göteborg: PhD candidates Anna Wessberg and Helena

Nilver, Pia Hempel, Martina Söderlund, Erica Ginström Ernstad, Monica Eriksson Orrskog;

Stockholm: Karolinska University Hospital Huddinge and Solna, South Hospital, Danderyd

Hospital, South BB, Södertälje Hospital: Helen Fagraeus, Annelie Sjölund, Eva Itzel Wiberg;

Halland Hospital: Elisabeth Johansson, Sandra Holmström, Åsa Ponten, Maud Ankardal;

Örebro Hospital: Inger Nydahl, Sofia Saarväli, Camilla Hartin; Falun Hospital: Elisabeth

Nordström, Kerstin Fransson, Visby Hospital: Madelen Jacobsson; North Älvsborg Hospital:

Maria Olsson, Anna Hagman.

Acknowledgements: Jonas Eriksson Söderling provided data from the Swedish Pregnancy

Register and performed the statistical analysis for the DSMB reports, Stellan Håkansson

provided data from the Swedish Neonatal Quality Register, Jesper Brodin provided data from

Statistics Sweden, Agneta Cedefors-Blom helped with secretarial assistance.

Mattias Molin and Per Ekman, the Statistical Consulting Group, Gothenburg, performed the

statistical analyses. Therese Svanberg at the Medical Library at Sahlgrenska University

Hospital performed the literature search.

We want to acknowledge the assistance given by the DSMB with Hans Wedel as chairman,

We also would like to thank all participating women.

Page 25 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


25

Funding: The Swedish state under the agreement between the Swedish government and the

county councils, the ALF-agreement (ALFGBG-440301, ALFGBG-718721, ALFGBG-

70940, ALFGBG-426401), the Health Technology Centre (HTA) at Sahlgrenska University

Hospital, the Foundation of the Health and Medical care committee of the Region of Vastra

Gotaland, Sweden (VGFOUREG387351, VGFOUREG640891, VGFOUREG854081),

Hjalmar Svensson Foundation, the foundation Mary von Sydow, born Wijk, donation fund,

Uppsala-Örebro Regional Research Council (RFR-556711, RFR-736891), Region Örebro

County research Committee (OLL-715501), Region Stockholm ALF 561222, 562222,

563222, Center for Clinical Research Dalarna - Uppsala University, Sweden CKFUU –

417011.

The funders had no role in study design, data collection, data analysis, data interpretation, or

writing of the report.

Competing interest: All authors have completed the ICMJE uniform disclosure form

(available on request from the corresponding author) and declare: UBW, SS, AW, MA, CB,

SBW, HF, MJ, LL, VS, JW, GW, HE, OS, HH declare no competing interests. AKW has

received free reagents (PlGF) from Roche for a prediction study of preeclampsia.

Ethical approval: The Regional Ethics Board in Gothenburg approved the study in May

2014 (Dnr: 285-14) and later its complementary applications (T 905-15, T 291-16, T 1180-16,

T 330-17, T 1066-17, T 087-18, T 347-18, T 961-18, T 1110-18).

Data sharing: The full data set is available from the corresponding author on reasonable

request.

Transparency; All authors had full access to all the data in the study and take responsibility

for the integrity of the data and the accuracy of the data analysis. The corresponding author

(UBW) affirms that this manuscript is an honest, accurate, and transparent account of the

study being reported; that no important aspects of the study have been omitted; and that any

discrepancies from the study as planned have been explained. The corresponding author

(UBW) had the final responsibility for the decision to submit for publication.

Page 26 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


26

Page 27 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


27

References

1 Olesen AW, Westergaard JG, Olsen J. Perinatal and maternal complications

related to postterm delivery: a national register-based study, 1978-1993. Am J Obstet Gynecol

2003; 189 (1): 222-7.

2 Lawn JE, Blencowe H, Waiswa P, et al. Stillbirths: rates, risk factors, and

acceleration towards 2030. Lancet 2016; 387(10018): 587-603.

3 WHO recommendations: induction of labour at or beyond term. (accessed June

9 2019). https://apps.who.int/iris/bitstream/handle/10665/277233/9789241550413-eng.pdf

4 Practice bulletin no. 146: Management of late-term and postterm pregnancies.

Obstet Gynecol 2014; 124 (2 Pt 1): 390-6.

5 Induction of labour: new NICE quality standard. Midwives 2014; 17 (3): 8.

6 Middleton P, Shepherd E, Crowther CA. Induction of labour for improving birth

outcomes for women at or beyond term. Cochrane Database Syst Rev 2018; 5: Cd004945.

7 Gelisen O, Caliskan E, Dilbaz S, Ozdas E, Dilbaz B, Ozdas E, Haberal A.

Induction of labor with three different techniques at 41 weeks of gestation or

spontaneous follow-up until 42 weeks in women with definitely unfavorable

cervical scores. Eur J Obstet Gynecol Reprod Biol. 2005 Jun 1;120 (2):164-9

8 Bruinsma A, Keulen J, Bakker J, et al. Induction of labor at 41 weeks or

expectant managment until 42 weeks- preliminary results of the INDEX trial. Am J Obstet

Gynecol 2017; 216 (1 Suppl): S27-S28 (abstr no 40)

9 Grobman WA, Rice MM, Reddy UM, et al. Labor Induction versus Expectant

Management in Low-Risk Nulliparous Women. New Engl J Med 2018; 379 (6): 513-23.

10 Keulen JK, Bruinsma A, Kortekaas JC, et al. Induction of labour at 41 weeks

versus expectant management until 42 weeks (INDEX): multicentre, randomised non-

inferiority trial. BMJ 2019; 364: l344.

11 Elden H, Hagberg H, Wessberg A, et al. Study protocol of SWEPIS a Swedish

multicentre register based randomised controlled trial to compare induction of labour at 41

completed gestational weeks versus expectant management and induction at 42 completed

gestational weeks. BMC Pregnancy Childbirth 2016; 16: 49.

12 Marsal K, Persson PH, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine

growth curves based on ultrasonically estimated foetal weights. Acta Paediatr 1996; 85 (7):

843-8.

Page 28 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

https://apps.who.int/iris/bitstream/handle/10665/277233/9789241550413-eng.pdf


28

13 Stephansson O, Petersson K, Bjork C, Conner P, Wikstrom AK. The Swedish

Pregnancy Register - for quality of care improvement and research. Acta Obstet Gynecol

Scand 2018; 97(4): 466-76.

14. Norman M, Kallen K, Wahlstrom E, Hakansson S. The Swedish Neonatal

Quality Register - contents, completeness and validity. Acta Paediatr 2019 Apr 21.

15 Goransson M, Magnusson A, Heilig M. Identifying hazardous alcohol

consumption during pregnancy: implementing a research-based model in real life. Acta Obstet

Gynecol Scand 2006; 85(6): 657-62.

16 European Surveillance of Congenital Anomalies (European Concerted Action

on Congenital Anomalies and Twins (EUROCAT). Data collection. Guideline for

Registration. (accessed May 9, 2019). http://www.eurocat-network.eu/content/EUROCAT-

Guide-1.4-section-3.2.

17 Ingemarsson I, Kallen K. Stillbirths and rate of neonatal deaths in 76,761

postterm pregnancies in Sweden, 1982-1991: a register study. Acta Obstet Gynecol Scand

1997; 76 (7): 658-62.

18 Nakling J, Backe B. Pregnancy risk increases from 41 weeks of gestation. Acta

Obstet Gynecol Scand 2006; 85 (6): 663-8.

19 Smith GC. Life-table analysis of the risk of perinatal death at term and post term

in singleton pregnancies. Am J Obstet Gynecol 2001; 184 (3): 489-96.

20 Gulmezoglu AM, Crowther CA, Middleton P, Heatley E. Induction of labour for

improving birth outcomes for women at or beyond term. Cochrane Database Syst Rev 2012;

(6): Cd004945.

21 Hussain AA, Yakoob MY, Imdad A, Bhutta ZA. Elective induction for

pregnancies at or beyond 41 weeks of gestation and its impact on stillbirths: a systematic

review with meta-analysis. BMC public health 2011; 11 Suppl 3: S5.

22 Wennerholm UB, Hagberg H, Brorsson B, Bergh C. Induction of labor versus

expectant management for post-date pregnancy: is there sufficient evidence for a change in

clinical practice? Acta Obstet Gynecol Scand 2009; 88 (1): 6-17.

23 Delaney M, Roggensack A. Guidelines for the management of pregnancy at

41+0 to 42+0 weeks. J Obstet Gynaecol Can : JOGC = Journal d'obstetrique et gynecologie

du Canada : JOGC 2008; 30 (9): 800-10.

Page 29 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

http://www.eurocat-network.eu/content/EUROCAT-Guide-1.4-section-3.2

http://www.eurocat-network.eu/content/EUROCAT-Guide-1.4-section-3.2


29

24 Nabhan AF, Abdelmoula YA. Amniotic fluid index versus single deepest

vertical pocket as a screening test for preventing adverse pregnancy outcome. Cochrane

Database Syst Rev 2008; (3): Cd006593.

25 Heinemann J, Gillen G, Sanchez-Ramos L, Kaunitz AM. Do mechanical

methods of cervical ripening increase infectious morbidity? A systematic review. Am J Obstet

Gynecol 2008; 199 (2): 177-87; discussion 87-8.

26 McMaster K, Sanchez-Ramos L, Kaunitz AM. Evaluation of a Transcervical

Foley Catheter as a Source of Infection: A Systematic Review and Meta-analysis. Obstet

Gynecol 2015; 126 (3): 539-51.

27 Gommers JSM, Diederen M, Wilkinson C, Turnbull D, Mol BWJ. Risk of

maternal, fetal and neonatal complications associated with the use of the transcervical balloon

catheter in induction of labour: A systematic review. Eur J Obstet Gynecol Reprod Biol

2017; 218: 73-84.

28 Hannah ME, Hannah WJ, Hellmann J, Hewson S, Milner R, Willan A.

Induction of labor as compared with serial antenatal monitoring in post-term pregnancy. A

randomized controlled trial. The Canadian Multicenter Post-term Pregnancy Trial Group. N

Engl J Med 1992; 326 (24): 1587-92.

29 Committee Opinion No. 644: The Apgar Score. Obstet Gynecol 2015; 126 (4):

e52-5.

30. Flenady V, Middleton P, Smith GC, et al. Stillbirths: the way forward in high-

income countries. Lancet 2011; 377 (9778): 1703-17.

31 Blencowe H, Calvert Ph DC, Lawn JE, Cousens S, Campbell OM. Measuring

maternal, foetal and neonatal mortality: Challenges and solutions. Best pract & res Clin obstet

& gynaecol 2016; 36: 14-29.

32. Grunewald C, Hakansson S, Saltvedt S, Kallen K. Significant effects on

neonatal morbidity and mortality after regional change in management of post-term

pregnancy. Acta Obstet Gynecol Scand 2011; 90 (1): 26-32.

Page 30 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


Figure 1 Flow chart of eligibility, randomisation, delivery, and assessment

181x234mm (150 x 151 DPI)

Page 31 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


Figure 2 Time to delivery in intention to treat groups In early induction group n=1380 because one woman was incorrectly randomised before 40+6 weeks and

delivered before 40+6 weeks

170x119mm (300 x 300 DPI)

Page 32 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlySupplementary Table A. Study period, number of possible eligible women and number of included women at the participating centres

Study period* Number of possible eligible women**

during study period

Number of included women

Included women/number of

possible eligible women

Sahlgrenska University Hospital

May 20, 2016 to October 15 2018 4657 776 16.7%

UppsalaUniversity Hospital

February 2, 2017 to October 15, 2018 1531 351 22.9%

Falun Hospital January 23, 2017 to October 15, 2018 874 119 13.6%

South Älvsborg Hospital

April 6, 2017 to October 15, 2018 1038 158 15.2%

Stockholm(Södertälje, Karolinska Solna, Karolinska Huddinge, Danderyd, South BB, South Hospital)

March 28, 2017 to October 15, 2018 8934 1122 12.6%

Örebro University Hospital

September 9, 2017 to October 15, 2018 650 65 10.0%

Halland Hospital, Varberg


Halland Hospital, Halmstad


Visby Hospital March 2, 2018 to October 15, 2018 61 30 49.2%

North Älvsborg Hospital

June 13, 2018 to October 15, 2018 279 10 3.6%

Total May 20, 2016 to October 15, 2018 19 079 2760 14.5%*Most centres stopped recruitment for a period during June, July and August**Deliveries ≥41+0 gestational weeks, only one delivery per woman, first delivery during the study period, singleton, women ≥18 years, cephalic presentation. Women with previous caesarean delivery were excluded.

Page 33 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlySupplementary Table B Baseline characteristics of Swedish background population*

Variable

Swedish background population*(n=49 714)

Age at delivery, years 30.8 (4.9)

30.6 (27.3; 34.1)

n= 49 714

Age at delivery

Age at delivery <35 years 39 581/49 714

(79.6%)

Age at delivery >=35 years 10 133/ 49 714 (20.4%)


Nulliparous 25 574/49 714 (51.4%)

Parous 24 140/49 714 (48.6%)


No 42627/44 217 (96.4%)

Yes 1590/44 217 (3.6%)

BMI at first antenatal visit 25.1 (4.8)

24.1 (21.8; 27.4)

n=45 415

BMI <30 at first antenatal visit 38 869/45 415 (85.6%)

BMI>=30 at first antenatal visit 6546/45 415 (14.4%)

Region of birth

Sweden 33 321/45 596 (73.1%)

Other Nordic countries 423/45 596 (0.9%)

Europe outside Nordic countries 3532/45 596 (7.4%)

Outside Europe 8320/45 596 (18.2%)

Highest level of education

Shorter than 9 years 1273/42 307 (3.0%)

At least 9 years primary school 2344/42 307 (5.5%)

High school 9 to 12 years 15 477/42 307 (36.6%)

University or corresponding 23 213/42 307 (54.9%)

Mode of conception

Assisted conception (IVF/ICSI) 1824/ 49 714 (3.7%)

Gestational age (days) 291 (3)

290 (288; 293)

n=49 714

Gestational age >=42+0 weeks 12 113/ 49 714 (24.4%)

Birth weight, g 3814 (453)

3800 (3500; 4110)

n=49 427

SGA 899/ 49 427 (1.8%)

LGA 968/49 427 (1.9%)

Stillbirth 59/49 714 (0.1%)

Perinatal mortality (stillbirth and early neonatal death [<7 days]) 67/49 714 (0.1%)

Page 34 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlyVariable

Swedish background population*(n=49 714)

Apgar score <4 at 5 minutes 172/49 534 (0.3%)

Apgar score <7 at 5 minutes 703/49 534 (1.4%)

Caesarean delivery 4157/49 714 (8.4%)

BMI= body mass index, SGA=small for gestational age (<-2 SD according to the sex specific Swedish reference12; LGA=large for gestational age (> 2 SD according to the sex specific Swedish reference12, IVF= in vitro fertilization, ICSI= intracytoplasmic sperm injectionFor categorical variables n/N (%) is presentedFor continuous variables Mean (standard deviation [SD])/ Median, (interquartile range [IQR] Q1; Q3) is presented*From the Swedish Pregnancy Register: Deliveries ≥ 41+0 gestational weeks in Sweden during the study period between May 15, 2016 and October 15, 2018 (only one delivery per woman, first delivery during the study period), singleton, women ≥18 years, cephalic presentation. Women with previous caesarean delivery were excluded.

Page 35 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlySupplementary Table C Baseline characteristics of per protocol population

Variable

Early induction group(n=1333)


(n=1351)

Age at randomisation, years 31.3 (4.7)31.1 (28.; 34.6)

n=1333

31.0 (4.5)30.9 (27.9; 34.2)

n=1351

Age at randomisation

Age at randomisation <35 years 1039/1333 (77.9%) 1079/1351 (79.9%)

Age at randomisation >=35 years 294/1333 (22.1%) 272/1351 (20.1%)


Nulliparous 729/1333 (54.7%) 742/1351 (54.9%)

Parous 604/1333 (45.3%) 609/1351 (45.1%)


No 1199/1231 (97.4%) 1201/1239 (96.9%)

1-9 cig/day 25/1231 (2.0%) 29/1239 (2.3%)

10 or more cig/day 7/1231 (0.6%) 9/1239 (0.7%)

Alcohol screening (AUDIT) at first antenatal visit*

0-5 points (low risk) 1068/1157 (92.3%) 1084/1164 (93.1%)

> = 6 (risk behavior) 89/1157 (7.7%) 80/1164 (6.9%)

Medical history

Psychiatric disease 89/957 (9.3%) 106/974 (10.9%)

Prepregnancy diabetes mellitus type 1 or 2 1/1317 (0.1%) 0/1327 (0.0%)

Endocrine disease 77/1314 (5.9%) 94/1327 (7.1%)

Chronic hypertension 1/1314 (0.1%) 1/1325 (0.1%)

Height (cm) at first antenatal visit 167.4 (6.2)167 (163; 172)

n=1236

167.6 (5.9)168 (163; 172)

n=1246

Weight (kg) at first antenatal visit 69.9 (14.3)67 (60.; 76)

n=1211

70.7 (14.4)68 (60.; 77)

n=1216

BMI at first antenatal visit 24.9 (4.7)23.8 (21.6; 27.1)

n=1234

25.1 (4.9)24 (21.7; 27.3)

n=1239

BMI at first antenatal visit <30 1084/1234 (87.8%) 1059/1239 (85.5%)

BMI at first antenatal visit >=30 150/1234 (12.2%) 180/1239 (14.5%)

Last recorded weight during pregnancy, kg 83.5 (14.4)82 (74.; 91)

n=1298

84.0 (14.6)82 (74.; 92)

n=1308

Region of birth

Sweden 1035/1246 (83.1%) 1047/1272 (82.3%)

Other Nordic countries 71/1246 (5.7%) 82/1272 (6.4%)

Europe outside Nordic countries 20/1246 (1.6%) 18/1272 (1.4%)

Outside Europe 120/1246 (9.6%) 125/1272 (9.8%)

Highest education

Shorter than 9 years 7/1177 (0.6%) 10/1217 (0.8%)

At least 9 years primary school 46/1177 (3.9%) 45/1217 (3.7%)

High school 9 to 12 years 365/1177 (31.0%) 395/1217 (32.5%)

University or corresponding 759/1177 (64.5%) 767/1217 (63.0%)

Employment status

Employed 1063/1251 (85.0%) 1082/1276 (84.8%)

Student 97/1251 (7.8%) 99/1276 (7.8%)

Maternity leave 45/1251 (3.6%) 52/1276 (4.1%)

Page 36 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960




(n=1351)

Unemployed 16/1251 (1.3%) 18/1276 (1.4%)

Sick leave 15/1251 (1.2%) 9/1276 (0.7%)

Other 15/1251 (1.2%) 16/1276 (1.3%)

Cohabitation with partner 1174/1312 (89.5%) 1184/1325 (89.4%)

Living alone/One person household 19/1312 (1.4%) 21/1324 (1.6%)

Mode of conception

Assisted conception (IVF/ICSI) 62/1333 (4.7%) 53/1351 (3.9%)

Involuntary childlessness 151/1207 (12.5%) 143/1173 (12.2%)

BMI=body mass index, IVF= in vitro fertilization, ICSI=intracytoplasmic sperm injectionFor categorical variables n/N (%) is presented.For continuous variables Mean (standard deviation [SD]) / Median (interquartile range [IQR] Q1; Q3) / n= is presented.

*alcohol screening by AUDIT tool according to antenatal care routines15

Page 37 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlySupplementary Table D Delivery and maternal adverse outcomes in per protocol population


(n=1333)



(95% CI) p-value

Difference between groupsMean (95% CI)

Gestational age at delivery (days) 288.7 (1.1)289

(288; 289)n=1333

291.7 (2.7)292

(289; 294)n=1351

<0.0001 -3.03 (-3.19; -2.88)

Time from randomization to delivery (days) 1.63 (1.03)2

(1; 2)n=1333

4.67 (2.65)4

(2; 7)n=1351

<0.0001 -3.04 (-3.19; -2.89)

Time from admittance to labour ward to delivery (hours)

20.1 (14.9)16.1

(9.2; 27.8)n=1332

13.6 (12.2)10.4

(4.6; 18.9)n=1350

<0.0001 6.57 (5.53; 7.59)

Preeclampsia/gestational hypertension/eclampsia

17/1333 (1.3%) 41/1351 (3.0%) 0.42 (0.24; 0.74) 0.0024

Onset of delivery

Spontaneous 182/1333 (13.7%) 917/1351 (67.9%) 0.20 (0.17; 0.23) <0.0001

Primary caesarean delivery 5/1333 (0.4%) 2/1351 (0.1%) 2.53 (0.49; 13.04) 0.44

Induction 1146/1333 (86.0%) 432/1351 (32.0%) 2.69 (2.48; 2.91) <0.0001

Mode of induction

Cervical ripening 878/1146 (76.6%) 325/432 (75.2%)

Amniotomy without oxytocin 129/1146 (11.3%) 42/432 (9.7%)

Amniotomy with oxytocin 139/1146 (12.1%) 65/432 (15.0%) 0.24

Cervical ripening

First method mechanical 334/878 (38.0%) 118/325 (36.3%) 0.97 (0.88; 1.07)

First method pharmacological 544/878 (62.0%) 207/325 (63.7%) 0.97 (0.88; 1.07) 0.63

Indication for induction


1145/1146 (99.9%) 0/432 (0.0%)


0/1146 (0.0%) 373/432 (86.3%)

Maternal condition 1/1146 (0.1%) 41/432 (9.5%)

Fetal condition 0/1146 (0.0%) 18/432 (4.2%) <0.0001

Use of epidural anaesthesia 704/1333 (52.8%) 658/1351 (48.7%) 1.08 (1.01; 1.17) 0.037

Chorioamnionitis 2/1333 (0.2%) 6/1351 (0.4%) 0.34 (0.07; 1.67) 0.30

Fever during labour 125/1333 (9.4%) 101/1351 (7.5%) 1.25 (0.98; 1.61) 0.088

Antibiotics during labour 238/1333 (17.9%) 256/1351 (18.9%) 0.94 (0.80; 1.11) 0.50

Meconium stained amniotic fluid 225/1192 (18.9%) 317/1101 (28.8%) 0.66 (0.56; 0.76) <0.0001

Duration of active labour 7.07 (5.42)5.64

(2.83; 10.24)n=692

8.37 (5.96)6.9

(3.78; 11.53)n=867

<0.0001 -1.30 (-1.87; -0.74)

Non-instrumental vaginal delivery 1112/1333 (83.4%) 1118/1351 (82.8%) 1.01 (0.97; 1.04) 0.68

Caesarean delivery 134/1333 (10.1%) 144/1351 (10.7%) 0.94 (0.75; 1.18) 0.65

Assisted vaginal delivery 87/1333 (6.5%) 89/1351 (6.6%) 0.99 (0.74; 1.32) 1.00

Secondary caesarean delivery 129/134 (96.3%) 142/144 (98.6%) 0.98 (0.94; 1.01) 0.39

Indication secondary caesarean delivery

Failed induction 8/129 (6.2%) 7/142 (4.9%)

Failure to progress at first stage 55/129 (42.6%) 52/142 (36.6%)

Fetal distress at first stage 32/129 (24.8%) 27/142 (19.0%)

Page 38 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960





(95% CI) p-value

Difference between groupsMean (95% CI)

Failure to progress and fetal distress at first stage

6/129 (4.7%) 7/142 (4.9%)

Other indication at first stage 5/129 (3.9%) 3/142 (2.1%)

Failure to progress at second stage 11/129 (8.5%) 23/142 (16.2%)

Fetal distress at second stage 6/129 (4.7%) 8/142 (5.6%)

Failure to progress and fetal distress at second stage

0/129 (0.0%) 3/142 (2.1%)

Failure operative vaginal delivery 6/129 (4.7%) 12/142 (8.5%) 0.26

Indication assisted vaginal delivery

Failure to progress 36/87 (41.4%) 28/89 (31.5%)

Fetal distress 33/87 (37.9%) 34/89 (38.2%)

Failure to progress and fetal distress 5/87 (5.7%) 6/89 (6.7%)

Maternal distress 13/87 (14.9%) 20/89 (22.5%)

Other 0/87 (0.0%) 1/89 (1.1%) 0.47

Duration of hospital stay from delivery to discharge (hours)

46.3 (27.1)43.6

(25.3; 61.6)n=1286

47.3 (29.7)45

(26.5; 61.2)n=1305

0.36 -1.03 (-3.20; 1.17)

Delivery and postpartum complications

Uterine rupture 0/1333 (0.0%) 0/1351 (0.0%) 1.00

Shoulder dystocia 6/1333 (0.5%) 4/1351 (0.3%) 1.52 (0.43; 5.38) 0.74

Cervical laceration 5/1333 (0.4%) 8/1351 (0.6%) 0.63 (0.21; 1.93) 0.60

Perineal lacerations III-IV 35/1333 (2.6%) 50/1351 (3.7%) 0.71 (0.46; 1.09) 0.14

Postpartum haemorrhage (>1000 ml) 134/1333 (10.1%) 143/1351 (10.6%) 0.95 (0.76; 1.19) 0.70

Maternal death 0/1333 (0.0%) 0/1351 (0.0%) 1.00

Maternal admission to intensive care unit 2/1333 (0.2%) 0/1351 (0.0%) 0.49

Postpartum infection* 41/1333 (3.1%) 26/1351 (1.9%) 1.60 (0.98; 2.60) 0.073

Wound infection 4/1333 (0.3%) 3/1351 (0.2%) 1.35 (0.30; 6.03) 0.99

Urinary tract infection incl. pyelonephritis 5/1333 (0.4%) 7/1351 (0.5%) 0.72 (0.23; 2.28) 0.79

Endometritis 17/1333 (1.3%) 5/1351 (0.4%) 3.45 (1.27; 9.31) 0.015

Sepsis 0/1333 (0.0%) 0/1351 (0.0%) 1.00

Venous thromboembolism 0/1333 (0.0%) 1/1351 (0.1%) 1.00

Breast feeding at discharge from delivery hospital

949/979 (96.9%) 961/992 (96.9%) 1.00 (0.98; 1.02) 1.00

Breast feeding 4 weeks after delivery 806/893 (90.3%) 814/927 (87.8%) 1.03 (1.00; 1.06) 0.11

CI= confidence intervalFor categorical variables n/N (%) is presented.For continuous variables Mean (standard deviation [SD]) / Median interquartile range [IQR] Q1; Q3) / n= is presented.For comparison between groups Fisher´s Exact test (lowest 1-sided p-value multiplied by 2) was used for dichotomous variables and Fisher´sNon-Parametric Permutation test was used for continuous variables*includes women with any of international classification of disease (ICD) 10 codes O85.9 (endometritis/sepsis) or O86 (wound infection, urinary tract infection including. pyelonephritis, fever of unknown etiology postpartum and other postpartum infections

Page 39 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlySupplementary Table E Perinatal outcome in per protocol groups

Variable


(n=1333)



(95% CI) p-value


Mean (95% CI)

Primary composite outcome 31/1333 (2.3%) 31/1351 (2.3%) 1.01 (0.62; 1.66) 1.00

Sub-components of the primary composite perinatal outcome

Perinatal/neonatal mortality (stillbirth + neonatal mortality)

0/1333 (0.0%) 6/1351 (0.4%) 0.032

Stillbirth 0/1333 (0.0%) 5/1351 (0.4%) 0.064

Neonatal mortality (Live births with death day 0-27)

0/1333 (0.0%) 1/1346 (0.1%) 1.00

Neonatal morbidity 31/1333 (2.3%) 26/1346 (1.9%) 1.21 (0.72; 2.02) 0.56

Neonatal morbidity defined as at least one of the following variables:


Metabolic acidosis (denominator based on validated umbilical cord blood samples at birth)†

13/645 (2.0%) 10/633 (1.6%) 1.28 (0.56; 2.89) 0.71

Hypoxic ischaemic encephalopathy (HIE) I-III

2/1333 (0.2%) 3/1346 (0.2%) 0.67 (0.11; 4.02) 1.00

Intracranial haemorrhage 1/1333 (0.1%) 2/1346 (0.1%) 0.50 (0.05; 5.56) 1.00

Neonatal convulsions 1/1333 (0.1%) 3/1346 (0.2%) 0.34 (0.04; 3.23) 0.63

Meconium aspiration syndrome (MAS) 2/1333 (0.2%) 3/1346 (0.2%) 0.67 (0.11; 4.02) 1.00

Mechanical ventilation within first 72 hours 3/1333 (0.2%) 5/1346 (0.4%) 0.61 (0.15; 2.53) 0.74

Obstetric brachial plexus injury 4/1333 (0.3%) 1/1346 (0.1%) 4.04 (0.45; 36.09) 0.37

Additional secondary neonatal outcomes variables

Admittance to neonatal intensive care units (NICU)

55/1333 (4.1%) 82/1346 (6.1%) 0.68 (0.49; 0.94) 0.026

Days at neonatal intensive care units (NICU)

3.38 (2.97)2

(1; 6)n=55

4.59 (5.64)3

(1; 6)n=81‡

0.15 -1.21 (-2.75; 0.19)

Admittance to neonatal intensive care units (NICU) >4 days

34/55 (61.8%) 45/81 (55.6%) 1.11 (0.84; 1.48) 0.58


Therapeutic hypothermia 1/1333 (0.1%) 2/1346 (0.1%) 0.51 (0.05; 5.58) 1.00

Neonatal jaundice requiring phototherapy or exchange transfusion

16/1333 (1.2%) 32/1346 (2.4%) 0.51 (0.28; 0.92) 0.031

Hypoglycaemia§ 22/1333 (1.7%) 20/1346 (1.5%) 1.11 (0.61; 2.03) 0.84

Pneumonia 8/1333 (0.6%) 13/1346 (1.0%) 0.62 (0.26; 1.50) 0.40

Sepsis 9/1333 (0.7%) 20/1346 (1.5%) 0.46 (0.21; 1.00) 0.065

Birth trauma¶ 0/1333 (0.0%) 1/1346 (0.1%) 1.00

Female 576/1333 (43.2%) 612/1351 (45.3%) 0.95 (0.88; 1.04) 0.29

Birthweight (g) 3814 (409)3804

(3530; 4090)n=1333

3874 (435)3860

(3570; 4155)n=1351

0.0003 -60.1 (-92.0; -28.5)

Small for gestational age (SGA)** 9/1333 (0.7%) 21/1351 (1.6%) 0.43 (0.20; 0.94) 0.045

Large for gestational age (LGA)** 20/1333 (1.5%) 24/1351 (1.8%) 0.84 (0.47; 1.52) 0.68

Macrosomia (>= 4500 g) 64/1333 (4.8%) 112/1351 (8.3%) 0.58 (0.43; 0.78) 0.0003

Any major birth defect†† 14/1333 (1.1%) 17/1351 (1.3%) 0.83 (0.41; 1.69) 0.75

Page 40 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960



(n=1333)



(95% CI) p-value


Mean (95% CI)

CI=confidence intervalFor categorical variables n/N (%) is presented.For continuous variables Mean (standard deviation [SD]) / Median interquartile range [IQR] Q1; Q3) / n= is presented.For comparison between groups Fisher´s Exact test (lowest 1-sided p-value multiplied by 2) was used for dichotomous variables Fisher´s Exact test (2-sided) was used for dichotomous variables and Fisher´s Non-Parametric Permutation test was used for continuous variables. *Apgar score of live births†validated umbilical cord samples defined as values of arterial pH < venous pH and values of arterial pCO2 >venous pCO2‡one infant with only home based neonatal care§hypoglycaemia defined as P-glucose concentration < 2,6 mmol/L after 3 hours¶defined as any of long bone fracture, clavicular fracture, skull fracture, other neurological injury, retinal haemorrhage, facial nerve palsy**SGA and LGA defined as <-2 SD and >2 SD, respectively, according to the Swedish sex specific reference12

††minor birth defects according to EUROCAT definition excluded16

Page 41 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


1

Title A Swedish multicentre register based randomised controlled superiority trial to compare induction of labour at 41 completed gestational weeks (GW) versus induction at 42 completed GW (SWEPI-Study).

Purpose and aims To investigate if a policy of induction of labour at 41+0 GW (early induction) is superior, in terms of neonatal and maternal outcomes, as compared to expectant management and induction at 42+0 GW (late induction) in healthy women with low risk singleton pregnancies. Justification of the trial Pregnant women, 41 gestational weeks or more, represents 15-20% of all pregnancies. This group is at increased risk of perinatal death and morbidity, but we still lack sufficient evidence how to manage these women. Nevertheless, worldwide, induction before 42 GW has been increasingly common. We could anticipate that induction already at 41+0 GW, i.e. nearly 1/5 of all pregnant women would consume a significant amount of available health care resources and maybe inflict unnecessary harm or pain to women, which is unacceptable unless it can be shown that such a policy is associated with improvements in neonatal or maternal outcomes. It is important to point out that as obstetricians we are asked daily by pregnant women at late term for advice when induction of labour is indicated. Today, we cannot offer these women guidance based on scientific evidence but we are convinced that successful completion of the proposed study will provide the critical information needed. The ongoing international studies (35/39, ARRIVE, please see below) have different study design with induction as early as 39 GW. The Dutch study (INDEX, please see below) will have too small sample size to evaluate the effect on perinatal morbidity and mortality and an eventual effect on caesarean section rate will be difficult to interpret in view of the large proportion of home deliveries being part of standard care in the Netherlands. These studies will therefore not be able to answer the question when induction of labour should be recommended in postdate uncomplicated pregnancies. The standard of care in Sweden today is still induction at 42 GW and therefore we have a unique opportunity to do this RCT. It is important to emphasize, however, that early induction (41-42 GW) has been discussed also in Scandinavia and therefore it is urgent and timely to perform the proposed study now rather than later before a shift in management also occurs in Sweden. The results of our study – whether induction at 41 or 42 GW is optimal regarding infant risk, maternal wellbeing, long term child morbidity and cost effectiveness – will have impact on postterm pregnancies not only in Sweden, but also in other countries.

Survey of the field Postterm pregnancy (≥42+0 GW) is associated with an increased risk for the mother and infant1,2. The risk of perinatal complications such as meconium aspiration syndrome (MAS), umbilical cord complications, asphyxia, pneumonia, sepsis, convulsions, shoulder dystocia, traumatic injuries and peripheral nerve damage is higher in postterm deliveries than in deliveries at term (aOR 1.1-2.0)1. Maternal complications increase from 40 GW. The risk of puerperal infections, postpartum bleeding, disproportion, labour dystocia, emergency caesarean sections, and cervical lacerations is higher for postterm than for term pregnancies (aOR 1.2-1.6)1. Standard care in Sweden, as in many other countries, is still induction of labour at 42+0 GW. With the increased risks that accompany a postterm pregnancy, the obvious solution would be to suggest induction of labour before the pregnancy becomes postterm. Many randomised controlled trials (RCTs) have compared induction at 41 or 42 GW with awaiting spontaneous onset of labour, with or without foetal surveillance, but most of the studies are small and of low methodological quality. The largest study so far is a multicentre RCT from Canada (n=3407 women)3. In this study, women were randomised at 41 GW to induction or to expectant management with monitoring until the start of spontaneous labour, complications occurred or if the gestational age reached 44 GW. Perinatal mortality (PNM) was similarly low in the two groups but the induced group had significantly fewer caesarean sections. The study has been criticised because the different induction methods used in the groups. Prostaglandins were not used in the expectant management group, and it is speculated that this may have contributed to a

Page 42 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


2

higher caesarean section frequency. In a later Norwegian study (n=500 women), women were randomised to induction at 289 days (GW 41+2) or monitoring and induction at 300 days (GW 42+6) unless spontaneous labour occurred4. The primary outcome (a composite neonatal outcome) was similar in the two groups. No case of stillbirth occurred, and there was no difference in the rate of caesarean section between the groups. However, the study was too small to assess PNM, a problem shared by all RCTs to date. Several systematic reviews and meta-analyses of RCTs have been published5-10. They are summarized in Table 1. Table 1. Systematic reviews and meta-analyses: Induction of labour at 41+0 or 42+0 GW versus expectant management.

Outcome Myers et al, AHRQ 20025

17 RCTs

Sanchez-Ramos et al, 20036 Meta-analysis 16 RCTs, OR, 95% CI

Wennerholm et al, 20097 Meta-analysis 13 RCTs, OR, 95% CI

Hussain et al, 20118 Meta-analysis 16 RCTs OR, 95% CI

Gulmezoglu et al, 20129 Meta-analysis 19 RCTs, (3 with intervention at 38-40 GW OR, 95% CI

Wennerholm et al, 201210 Updated meta-analysis, 17 RCTs, OR, 95% CI

Perinatal death Favours induction

0.41 0.14-1.18

0.33 0.10-1.09

0.31 0.11-0.88

0.30 0.09-0.99

0.32 0.11-0.91

Meconium aspiration

No difference

0.46 0.18-1.21

0.43 0.23-0.79

0.43 0.23-0.79

0.39 0.21-0.75

0.47 0.25-0.90

Caesarean section

No difference

0.88 0.78-0.99

0.87 0.80-0.96

NA 41 weeks: 0.92 0.76-1.12

42 weeks: 0.97 0.72-1.31

0.94 0.80-1.11

In a Health Technology Assessment (HTA) performed at Sahlgrenska University Hospital (SU) in Gothenburg, 13 RCTs were included (n= 6617 women) 7. The review showed no difference in PNM but found significantly fewer cases of MAS and a lower caesarean section rate in the induction group as compared to the expectant management group. An updated HTA meta-analysis of 17 RCTs (n=7223 women) was done in 2012, with a more liberal inclusion of studies10. It showed a lower PNM in the induction group, fewer cases of MAS, and no difference in caesarean section rate. A major limitation in the individual studies is that they compared induction at 41 GW to expectant management to 42 GW and beyond, making this group very heterogeneous. In fact, many cases of adverse outcome in the expectant management group occurred after 42 completed weeks. Thus, only a few small studies of low quality (n=3 trials from India, Tunisia and Turkey, respectively, comprising 981 women) have compared induction of labour at 41 GW to a policy of expectant management to exactly 42 GW (42+0 GW). One methodological problem in individual studies and systematic reviews is that important outcomes such as PNM and hypoxic ischaemic encephalopathy (HIE) occur at a very low frequency11-13 and a large number of women needs to be induced to prevent one case of perinatal death (number needed to treat NNT=410)9. Thus, very large studies are needed to address these important outcomes. Further, only one RCT has assessed the woman’s experiences of induction or expectant policy for postterm pregnancy14. At the time of randomisation, 74% would have preferred induction if they had been able to choose. The majority of women (84%) who were induced had a positive childbirth experience. Despite that induction lead to more intensive labour, 74% preferred being induced compared with awaiting spontaneous on-set of labour. The Canadian postterm pregnancy trial from 1995 indicated that induction of labour was associated with a lower cost as compared with expectant management of postterm pregnancy15. However, as the study was performed more than 20 years ago the data may not be applicable today. Also, exact dating of pregnancy has strongly improved since these trials have been performed.

Page 43 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


3

Other studies on the topic are underway. In the United Kingdom, the 35/39 trial (ISRCTN115172751)16 has recently been completed comparing induction of labour at 39 weeks (39+0 to 39+6 GW) to expectant management for nulliparous women 35 years of age or older. The primary endpoint is caesarean section rate and the secondary endpoints include perinatal mortality and morbidity. Although results have to be awaited, the sample size of the randomised patients in this study will be too small (n=630 women, 315 in each group) to come up with definite conclusions concerning perinatal mortality and morbidity. In the United States, the NICHD network is running a study in nulliparous women comparing induction of labour at 39 GW or expectant management until at least 40+5 GW (n=6000 women, 3000 in each group) under the acronym ARRIVE (NCT01990612, https://clinicaltrials.gov/ct2/show/NCT01990612). In view of perinatal care in the United States, with a high perinatal mortality rate and its high caesarean section rate, the results will offer limited guidance for clinical management in Sweden. In the Netherlands, a similar study as the one proposed by us will be completed this year (INDEX, NTR3431, http://www.studies-obsgyn.nl/index/page.asp?page_id=1095), but the sample size (n=1800 women) will be too small to draw definite conclusions17. Moreover, the Dutch study is comparing obstetrician led care in the induction at 41 GW group to mostly midwifery led care in the induction at 42 GW group. This is different from the proposed study, where both induction at 41 GW and waiting till 42 GW will be performed in the same perinatal care setting.

So far, none of the RCTs have included or planned a long-term follow up of the children. Several reports in the literature indicate that this is an important issue. A Swedish study found a delayed development in postterm children compared with full-term children in a follow-up at the age of 4 and 4.5 years18. Recently, early markers of the metabolic syndrome and increased rate of obesity have been found in children born postterm19,20. Labour induction per se may have negative consequences as indicated in a recent Norwegian register study. Labour induction was shown to be an independent risk factor for cerebral palsy (CP) in general and for unilateral and bilateral spastic CP in particular, the risk for bilateral CP being four-fold higher when labour induction was performed in children at term21. Recently, an association between autism spectrum disorders and induced or augmented childbirth has been shown22. Some evidence suggests an oxytocin receptor deficiency in autism and the authors hypothesized that an oxytocin infusion may alter the oxytocin balance in the newborn.

Project description We submitted an application to “Klinisk behandlingsforskning (KBF)” 2015, which was accepted in the first triage step but was rejected in the final round. We have taken the criticism under consideration and the application has been re-designed and improved substantially. Trial design A multicentre register based randomised controlled superiority trial

Justification of unique Trial Design The planned study is the first randomised controlled study (RCT) utilising the national Pregnancy Register (developed by co-applicant associate Professor Olof Stephansson and colleagues) for collecting data on maternal characteristics, pregnancy, delivery and postnatal care and the Swedish Neonatal Quality Register (SNQ; developed by associate Professor Stellan Håkansson) for data on neonatal outcome. Randomisation will be performed using a computerised module linked to the Pregnancy Register. Such a design has recently been applied successfully in cardiovascular research23,24 and has several important advantages including lower cost, ease of recruitment and facilitating the translation of the study results to clinical practice. We have obtained advice concerning the study design by our international collaborator Professor Ben Mol (currently University of Adelaide, Australia) who has coordinated multiple RCTs in the Netherlands25-34. P=Population Inclusion criteria: Healthy women ≥18 years old with a normal live singleton pregnancy in cephalic presentation at 41+0-41+1 GW, who is able to understand oral and written information which will be provided in several languages.

Page 44 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


4

Exclusion criteria: Women with previous caesarean section or other uterine surgery, pregestational and insulin dependent gestational diabetes, preeclampsia, multiple pregnancy, foetus in breech or transverse position, premature rupture of the membranes, antenatally detected foetal malformations and contraindications to vaginal delivery such as placenta previa.

Setting: Swedish university clinics and county hospitals belonging to Region Västra Götaland, Region Skåne, Stockholm, Uppsala, Umeå, Region Örebro, Falun and Visby representing around 51 000 deliveries per year.

I=Intervention Intervention group: Women randomised to early induction will be induced the same day or the next day after randomisation. Induction will thus take place at GW 41+0-41+2.

C=Comparison Control group: Women randomised to expectant management/late induction will be induced at GW 42+0 (and no later than GW 42+1). After randomisation, the women are followed-up at the ordinary antenatal clinic according to clinical routine for GW 41+0-42+0.

Methods of labour induction (early at GW 41+0-41+2 or late at GW 42+0-42+1): This is a pragmatic clinical trial and the methods will be according to the clinics’ routine. Please see under “Methods for labour induction”.

O=Outcomes Primary: Primary outcome will be a composite outcome including at least one of the following: stillbirth, neonatal mortality and neonatal morbidity. Stillbirth is defined as intrauterine foetal death of a foetus that was alive at randomisation. Neonatal mortality is defined as live births with death day 0-27. Neonatal morbidity is defined as at least one of the following variables: Apgar score <7 at 5 minutes, metabolic acidosis defined as pH <7.05 and base excess >12 mmol/l in umbilical artery or pH<7.00 in umbilical artery, hypoxic-ischaemic encephalopathy (HIE) I-III, intracranial haemorrhage, neonatal convulsions, meconium aspiration syndrome (MAS), mechanical ventilation and obstetric brachial plexus injury.

Secondary: Secondary infant outcomes: admittance to neonatal intensive care unit, macrosomia, shoulder dystocia, Apgar score <4 at 5 minutes, therapeutic hypothermia, neonatal jaundice, pneumonia, sepsis, cerebral palsy, neurodevelopment at childhood follow up to the age of 4 years, death during the first 4 years of life. Secondary maternal outcomes: use of epidural analgesia, caesarean section, assisted vaginal delivery, duration of labour, perineal lacerations III-IV, postpartum haemorrhage (>1000 ml), chorioamnionitis, wound infection, urinary tract infection, endometritis, sepsis, depression/anxiety, “self-efficacy”, women’s experience of management and delivery and breast feeding. We will also measure costs and perform an economic analysis.

Sample size We anticipate that induction of labour at GW 41 as compared to induction at GW 42 will reduce the primary outcome by one third from 2.74% to 1.84% (level of significance 0.05, power 80%, dropout rate 10%). The primary composite outcome of 2.74% is based on the following rates of perinatal mortality and neonatal morbidity in GW 41+3 from Region Skåne between 2000 and 2010: HIE 0.22%, MAS 0.36%, obstetric brachial plexus injury 0.05%, Apgar score < 7 at 5 min 1.58%, stillbirth and neonatal mortality 0.53%. We need a sample size of 10 038 women to be randomised, 5 019 to induction of labour at GW 41+0 and 5 019 to expectant management. Since primary outcome is a serious and rare event, we consider that a 33% reduction is clinically relevant. In 2014, the number of deliveries at Sahlgrenska University Hospital, South Älvsborg Hospital, Malmö/Lund University Hospital, Karolinska University Hospital, Danderyd Hospital, Södersjukhuset (Stockholm South General Hospital), Södertälje Hospital, Uppsala Academic Hospital, Falun Hospital, University Hospital Örebro, Umeå University Hospital, Sunderby Hospital and Visby Hospital were 10 024, 3 018, 9 056, 8 387, 6 106, 7 294, 1 565, 3 880, 3 008, 2 864, 1 735, 1881 and 532 respectively, altogether 59 350 deliveries, representing 52% of total deliveries in Sweden. During a 3 year period we estimate a total of 178 050 deliveries at the 13 participating centres. Given a rate of 18% of women being undelivered at 41+0 GW, this will sum up to 32 049 women

Page 45 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


5

during 3 years. We estimate that around 40% (n=12 800) of these women can be included. If we instead may include 50% of the women, the power will be improved (n=16 000). Patient information and recruitment All pregnant women are given oral and written information about the study at around GW 40+3 at the antenatal care units and further information will be available on a study web site. All potentially eligible women will be consecutively contacted by the study midwife for further information. Women who are interested in participating in the study are booked for a visit at the antenatal clinic at the respective hospital at GW 41+0 to 41+1 for further information. Ultrasound examination for foetal weight estimation is performed according to the clinics’ routine. Eligible women will be enrolled and randomised after written consent. Informed consent will also be asked from both parents for future follow-up studies of the children from national health and quality registers. Randomisation A doctor or midwife at the antenatal clinic will enrol eligible women. Randomisation will be performed at the antenatal clinic at GW 41+0 to 41+1. Women allocated to early induction will be induced within 24 hours after randomisation. Block randomisation with a fix block size (unknown to the investigator) will be done online,1:1, using the Pregnancy Register with a module specifically developed for the study by MedSciNet AB, Stockholm, Sweden. Stratification will be performed for centre and parity (primiparity versus multiparity). Blinding: The statistician will be blinded to group and intervention.

Methods of labour induction (early at GW 41+0-41+2 or late at GW 42+0-42+1) Before labour induction the following examinations are performed: blood pressure, urine analysis, abdominal and cervical examination and cardiotocography (CTG). If there are no contraindications for induction of labour, induction will be performed according to the following:

1. if the foetal head is well engaged and the cervix is ripe (Bishop score ≥6 for primiparous and ≥5 for multiparous) amniotomy is performed. Oxytocin infusion will be started after 1-2 hours if there are no regular contractions.

2. if the cervix is unripe (Bishop score <6) or the foetal head is not engaged any of the following methods are used (according to the routine at the clinic):

a. mechanical dilation using a Foley like catheter (single or double balloon) b. misoprostol (Cytotec®) orally c. misoprostol (Misodel®) controlled-released vaginal insert d. prostaglandin E2 (Minprostin®) gel vaginally e. prostaglandin E2 (Propess®) vaginal insert

Mechanical dilation may be followed by prostaglandins and vice versa. If the woman is not in active labour or cervix is still unripe after 48 hours, the clinician together with the woman will decide whether to perform a caesarean section or to continue expectant management until GW 42+0. All inductions will be performed with the patient staying at the hospital. The clinic’s protocol and guidelines for induction and surveillance during induction will be followed. During the active phase of labour foetal surveillance with continuous CTG with or without ST segment analysis (STAN) in combination with foetal scalp blood sampling is recommended.

Time period The trial is planned to start in May 1, 2016 and end in April 30, 2019.

Time frame (January 2016-April 2016): Development of the randomisation module and electronic case record form (eCRF) for the Pregnancy Register by MedSciNet AB, Stockholm, Sweden. Information to antenatal clinics and delivery wards at the participating centres. 0 (May 2016): Recruitment of patients started.

18 months after start of recruitment, November 2017): 5000 patients recruited and randomised.

20 months (January 2018): An interim analysis (50% of the patients are randomised and delivered).

Page 46 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


6

36 months (May 2019): End of recruitment

37-42 months (June 2019-November 2019): Analysis of data and manuscript writing.

Data collection and data management Data on background variables, obstetric and neonatal outcomes will be obtained from the Pregnancy Register and SNQ. An eCRF linked to the randomisation module will be used for complimentary data. Details of methods of labour induction will be registered in the eCRF. Data on women’s experiences will be collected from the Pregnancy Register, and in a subpopulation (Gothenburg, Örebro and Falun), by additional questionnaires (web based or postal) after randomisation and 3 months after delivery including the following instruments for analysis of attitudes, experiences and health related quality of life: Big Five, Euro-Qol, The Childbirth Experience Questionnaire, and The General Self-Efficacy Scale35-38. The instruments will be eligible in multiple languages. Data will be collected on randomised women and on eligible women declining participation but accepting to participate with data. Data on long-term children follow up will be obtained from the Patient-, Cause of Death- and Prescribed Drug Registers, the CPUP, (a National Quality Register for cerebral palsy), BORIS (a National Quality Register for obesity in childhood) and the newly started Child Health Register (“Svenska Barnhälsovårdsregistret”; BHVQ). Statistical analysis The primary statistical analysis will be the comparison of primary composite outcome (stillbirth, neonatal mortality and morbidity) between the two randomised groups early induction and late induction with two-sided Fisher’s exact test on the intention to treat (ITT) populations at significance level 0.05. The 95% confidence interval for the difference in proportions will be calculated for the primary variable and for all secondary dichotomous variables. Complementary analysis with adjustment for baseline variables will be performed if needed with multivariable logistic regression analyses. Women with spontaneous labour or prelabour rupture of the membranes (PROM) after randomisation but before planned induction will be included in the statistical analysis according to ITT. Women with pregnancy complications after randomisation indicating any need of intervention are included in the statistical analysis according to ITT. Foetuses/newborns with lethal birth defects will be excluded in the analysis of stillbirth, neonatal mortality and morbidity. A hierarchical testing procedure will be applied to guarantee that the probability of type I error is <0.05 for all confirmative statements. If the primary efficacy analysis will be significant at the 5% significance level then the probability mass 5% will go to the first secondary analysis. If this analysis also is significant then the probability mass will go to the second secondary efficacy analysis and so on. When the first non-significant analysis will be found all the previous analysis will be confirmatory. The distribution of the variables will be given as mean, median and SD for continuous variables and as number and percentages for categorical variables. For comparison between the two randomised groups regarding all secondary efficacy variables two sample T-test will be used for continuous variables, Fisher’s exact test for dichotomous variables, Mantel-Haenszel Chi-square test for ordered categorical variables and Chi-square test for unordered categorical variables. 95% confidence intervals for the differences between the two groups will be given where applicable. Subgroup analyses will be performed for the following pre-defined groups: 1. Women > 35 years, and 2. Women with BMI>30 and BMI >35. Interactions between randomised group effect and the following baseline variables will be analysed: parity (primiparous and multiparous), Bishop score (< 2, 2-5, ≥6). If significant

Page 47 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


7

interactions are found, subgroup analysis also will be performed for those variables. All significance tests will be two-sided and conducted at the 5% significance level. Further details of the statistical analysis will be supplied in a Statistical Analysis Plan (SAP) before the data lock. (Study Statistician: Nils-Gunnar Pehrsson). An analysis is performed when 50% of patients have been randomised and delivered (approximately 18 months after start of recruitment). This analysis is performed by a Data and Safety Monitoring Board (DSMB). The DSMB will consist of an independent statistician, one obstetrician and one midwife. The DSMB will have unblinded access to all data and will recommend the steering committee to continue or to stop the trial. The steering committee and the study statistician will not receive any results from the DSMB analysis.

Cost-effectiveness analysis Cost-effectiveness analyses will be performed, comparing induction at GW 41+0 with induction at GW 42+039,40. Cost-effectiveness measures will be calculated for the time perspective including the first 3 month after inclusion using two different effectiveness measures reflecting quality of life and morbidity/mortality: Euro-Qol (EQ-5D) and a composite effectiveness measure, comprised of stillbirth, neonatal mortality and neonatal morbidity (i.e. primary outcome). In both cases, a broad range of sensitivity analyses will be performed, which includes calculating CIs of the cost-effectiveness measures, using bootstrapping. The cost-effectiveness measures computed for the initial 3 month period will be complemented with additional calculations with a longer time perspective, including child health outcomes up to the age of 4 years.

Significance Today there is insufficient scientific support that induction of labour at 41+0 GW (late term) as compared with expectant management and induction at 42+0 GW will reduce perinatal mortality and morbidity without an increase in operative deliveries, negative delivery experiences or higher costs. Due to different study populations and level of care the generalizability of ongoing studies (35/39, ARRIVE and INDEX, please see below) will be limited. Consequently, to change the practice in Sweden, which is induction at 42 GW, it is critical and timely to investigate if induction of labour at GW 41+0 is superior to standard care since a significant proportion (15-20%) of all pregnancies last more than 41 GW.

Preliminary results In order to obtain more detailed information regarding the views of women in 41+0-42+0 GW, we executed a preliminary study including in-depths interviews of 10 women and the same women were interviewed 1-3 months after delivery. In 7 cases we also interviewed the partners of these women. Most women reported increased stress and worry as the pregnancy proceeded beyond 41+0 GW. They also experienced a lack of support and understanding from the health care system and, in particular, the women and their partners found the information by midwives and obstetricians insufficient and varied, which created a feeling of insecurity. Because of the deficient support by health care providers, many women seek more information on the internet which is partly contradictory and increases rather than relieves the stress and feeling of not being cared for. Equipment All participating centres have a fully equipped delivery ward with equipment for foetal surveillance as well as neonatal intensive care units. Obstetricians and paediatricians are on call 24/7. Need for infrastructure The Swedish Network for National Clinical Studies in Obstetrics and Gynecology (SNAKS), started in 2014, is a network of university hospitals as well as non-university clinics/hospitals in Sweden. The SNAKS initiative is supported by the Swedish Society of Obstetrics and Gynecology (SFOG). So far all university clinics and 20 non-university clinics/hospital, comprising more than half of all OB/Gyn clinics in Sweden, have joined the network. The aim of

Page 48 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


8

the network is to conduct multicentre studies of high quality, with sufficient power, engaging a majority of the Swedish hospitals and thus enhance the implementation of findings and make the health care within Ob/Gyn more evidence based. The Swedish national quality registers will be an integral part of the conduct of studies, including both randomisation and follow-up functions. This study will be among the first studies within the SNAKS. The Swedish Pregnancy Register is a certified national quality register, which contains data from early pregnancy to a few months after birth. Chair of the register is associate Professor Olof Stephansson, Karolinska Institute, Stockholm. The register will be used for randomisation with a special module developed by MedSciNet AB, Huddinge Stockholm. Pregnancy, delivery and neonatal data will be collected from the register. A majority of the variables in the Pregnancy Register are obtained directly from the standardized medical record system. Thus, data are valid as clinicians in the routine antenatal, obstetrical and delivery care enter it. A minor part of the data are entered in a web-protocol by the midwife at registration for antenatal care (country of birth, education, self-reported health status and alcohol audit). At the follow-up visit after delivery the midwife enters data on self-reported health and breastfeeding. Data from the manually entered data by midwives in antenatal care has been validated with a high quality41. The Swedish Neonatal Quality Register (SNQ) (www.snq.se) comprises data from all neonatal intensive care units (NICUs) on all newborns admitted to NICUs at birth and within 28 days after birth. Chair of the register is associate Professor Stellan Håkansson, Umeå. The Swedish Child Health Register (Svenska Barnhälsovårdsregistret; BHVQ) (www.bhvq.se) is a newly started national quality register with health data from the Child Health Centres for children between 0 and 5 years of age. Chair of the register is Dr Thomas Wallby, Uppsala. Other National Health Registers that will be used for data collection on children outcome are The National Patient Register (NPR), the Cause of Death Register and the Prescribed Drug Register (www.socialstyrelsen.se/register/halsodataregister). The Swedish health data registers are extensively used for research purposes and the quality has been deemed to be satisfactory42. National and international collaboration The following 13 centres will participate in the study: the Departments of Obstetrics and Gynecology at South Alvsborg County Hospital (SAS), Sahlgrenska University Hospital, Skane University Hospital (SUS), Karolinska Hospital Solna/Huddinge, Stockholm, Södersjukhuset (Stockholm South General Hospital), Södertälje Hospital, Danderyds Hospital, Uppsala University Hospital, Umeå University Hospital, Sunderby Hospital, University Hospital Örebro, Falun and Visby Hospital. Our aim is to include most of the centres in Sweden through SNAKS. The project will be led by a steering committee which preliminarily has the following members: associate Professor Ulla-Britt Wennerholm, Professor Henrik Hagberg, Professor Christina Bergh, associate Professor Andreas Herbst, PhD Maria Bullarbo, PhD Helen Elden, PhD Verena Sengpiel, PhD Sissel Saltvedt, PhD Sophia Brismar Wendel, associate Professor Olof Stephansson, associate Professor Anna-Karin Wikström, associate Professor Lars Ladfors, PhD Helena Fadl, PhD Jan Wesström and Professor Marie Bixo. The steering committee is responsible for the study design, the co-ordination between centres, the progress of the study, and for the results being analysed and summarised for publication. The steering committee is also principal investigators, i.e. the members are responsible for the project at their own centre for example adherence to the study protocol, recruitment and any practical problem that may arise within the frame of the project. The steering committee also represents the Swedish perinatal community, strongly expressing the need and their support for this trial. There will be further coordinating clinical investigators at each site. One midwife will be responsible for the study implementation at each site. At SU this will be Anna Wessberg, midwife and registered PhD student. Professor in Health Economics Kristian Bolin is responsible for the health economic analysis and Nils Gunnar Pehrsson is responsible for the statistical analysis. Stellan Håkansson, neonatologist at Umeå University is chair of The Swedish Neonatal Quality Register (SNQ) (www.snq.se) which comprises data from all neonatal intensive care units (NICUs) on all newborns admitted to NICUs at birth and within 28 days after birth. Professor in paediatrics Göran Wennergren is responsible for the paediatric follow up. A Data and Safety Monitoring Board (DSMB) consisting of experts (one statistician, a senior obstetrician and a midwife) will supervise the study. The DSMB will secure the project management through

Page 49 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


9

periodical reviews and will make the recommendations concerning the continuation, modification or termination of the trial. Serious adverse events (SAE) including stillbirth, neonatal death, severe neonatal morbidity, defined as intraventricular haemorrhage, asphyxia or MAS with admission to NICU, maternal death, severe maternal morbidity defined as admission to intensive care unit and events related to induction of labour e.g. uterine rupture are reported to the principal investigators, who will report to the DSMB. Other grants The current project is supported by FOU-VG (272,000 SEK, PI: H Hagberg), Medi-SAM (535,000 SEK; PI: UB Wennerholm), ALF-GBG (525,000 SEK; PI: UB Wennerholm) and RFR (Regionala forskningsrådet) Uppsala-Dalarna (430,000 SEK, PI AK Wikström). Registration of study The study is registered in Current Controlled Trials (ISRCT26113652) and will be conducted according to CONSORT 2010.

Page 50 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


10

References

1. Olesen AW et al. Am J Obstet Gynecol. 2003;189:222-227

2. Roos N et al. Acta Obstet Gynecol Scand. 2010;1003-10

3. Hannah ME et al. NEJM. 1992;326(24):1587-92

4. Heimstad R et al. Obstetrics and gynecology. 2007;109 (3):609-17

5. Myers ER et al. Evidence Report/Technology Assessment No. 53. AHRQ Publication No. 02-E018. Rockville, May 2002

6. Sanchez-Ramos L et al. Obstet Gynecol. 2003;101(6):1312-8

7. Wennerholm UB et al Acta Obstet Gynecol Scand. 2009; 88:6-17

8. Gülmezoglu AM et al Cochrane Database Syst Rev. 2012;6:CD004945

9. Hussain AA et al. BMC Public Health. 2011;11 Suppl 3:S5

10. Wennerholm U-B et al. Update of mini-HTA rapport 2007, Göteborg, Sahlgrenska Universitetssjukhuset, HTA-centrum, 2012

11. Badawi et al. BMJ. 1998;317(7172):1554-8

12. Smith J et al. BJOG. 2000;107(4):461-6

13. Thornberg E et al. Acta paediatrica. 1995;84 (8):927-32

14. Heimstad R et al. Acta Obstet Gynecol Scand. 2007;86(8):950-6

15. Goeree R et al. CMAJ. 1995;152(9):1445-50

16. Walker KF et al. BMC Pregnancy Childbirth. 2012 Dec 11;12

17. Kortekaas JC et al. BMC Pregnancy Childbirth. 2014 Oct 23;14:350

18. Lindström K et al. Acta Paediatr 2005 Sep;94(9):1192-7

19. Ayyavoo A et al. PLoS One. 2013 1;8(7):e67966

20. Beltrand J et al. J Pediatr. 2012 ;160(5):769-73

21. Elkamil AI. et al. Acta Obstet Gynecol Scand. 2011 Jan;90(1):83-91

22. Gregory SG. et al. JAMA Pediatr. 2013;167(10):959-66

23. Fröbert O et al. N Engl J Med. 2013 Oct 24;369:1587-97

24. Lagerqvist B et al. N Engl J Med. 2014 Sep 18;371:1111-20

25. Boers KE et al. BMJ. 2010;341:c7087 26. Bensdorp AJ et al. BMJ.

2015;350:g7771

27. Freeman LM et al. BMJ.

2015;350:h846 28. Mol B et al. Lancet. 2014;383:1483-9 29. Liem S et al Lancet. 2013;382:1341-9 30. Jozwiak M et al. Lancet.

2011;378:2095-103 31. Labrie J et al. N Engl J Med.

2013;369:1124-33 32. Bakker JJ et al. N Engl J Med.

2010;362:306-13 33. Koopmans CM et al. Lancet.

2009;374:979-88 34. Steures P et al. Lancet. 2006;368:216-

21 35. John O et al. In: Handbook of

personality: theory and research. Edited by Pervin L, John O. New York: Guilford Press; 1999;102-138.

36. Rabin R et al. Ann Med. 2001; 33:337-343

37. Dencker A al. BMC Pregnancy Childbirth. 2010;10:81

38. Love J et al. Qual Life Res. 2012;21:1249-1253

39. Drummond M et al. 2015 Oxford, United Kingdom, Oxford University Press.

40. Gray, A. M. 2011 Oxford, United Kingdom, Oxford University Press.

41. Petersson et al. BMC Health Services Research. 2014;14:364

42 Ludvigsson JF et al. BMC Public Health. 2011;11:450

Page 51 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlySTUDY PROTOCOL Open Access

Study protocol of SWEPIS a Swedishmulticentre register based randomisedcontrolled trial to compare induction oflabour at 41 completed gestational weeksversus expectant management andinduction at 42 completed gestationalweeksHelen Elden1*, Henrik Hagberg2, Anna Wessberg1, Verena Sengpiel2, Andreas Herbst3, Maria Bullarbo2,Christina Bergh4, Kristian Bolin5, Snezana Malbasic6, Sissel Saltvedt7, Olof Stephansson8,9, Anna-Karin Wikström10,Lars Ladfors2 and Ulla-Britt Wennerholm2

Abstract

Background: Observational data shows that postterm pregnancy (≥42 gestational weeks, GW) and late termpregnancy (≥41 GW), as compared to term pregnancy, is associated with an increased risk for adverse outcome for themother and infant. Standard care in many countries is induction of labour at 42 GW. There is insufficient scientificsupport that induction of labour at 41 GW, as compared with expectant management and induction at 42 GW willreduce perinatal mortality and morbidity without an increase in operative deliveries, negative delivery experiences orhigher costs. Large randomised studies are needed since important outcomes; such as perinatal mortality and hypoxicischaemic encephalopathy are rare events.

Methods/Design: A total of 10 038 healthy women ≥18 years old with a normal live singleton pregnancy in cephalicpresentation at 41 GW estimated with a first or second trimester ultrasound, who is able to understand oral andwritten information will be randomised to labour induction at 41 GW (early induction) or expectant management andinduction at 42 GW (late induction). Women will be recruited at university clinics and county hospitals in Swedencomprising more than 65 000 deliveries per year. Primary outcome will be a composite of stillbirth, neonatal mortalityand severe neonatal morbidity. Secondary outcomes will be other adverse neonatal and maternal outcomes, mode ofdelivery, women’s experience, cost effectiveness and infant morbidity up to 3 months of age. Data on backgroundvariables, obstetric and neonatal outcomes will be obtained from the Swedish Pregnancy Register and the SwedishNeonatal Quality Register. Data on women’s experiences will be collected by questionnaires after randomisation and3 months after delivery. Primary analysis will be intention to treat. The statistician will be blinded to group andintervention.(Continued on next page)

* Correspondence: [email protected] University, Institute of Health and Caring Sciences, SahlgrenskaAcademy, S-405 30 Gothenburg, SwedenFull list of author information is available at the end of the article

© 2016 Elden et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Elden et al. BMC Pregnancy and Childbirth (2016) 16:49 DOI 10.1186/s12884-016-0836-9

Page 52 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

http://crossmark.crossref.org/dialog/?doi=10.1186/s12884-016-0836-9&domain=pdf

mailto:[email protected]

http://creativecommons.org/licenses/by/4.0/

http://creativecommons.org/publicdomain/zero/1.0/

Confidential: For Review Only(Continued from previous page)

Discussion: It is important to investigate if an intervention at 41 GW is superior to standard care in order to reducedeath and lifelong disability for the children. The pregnant population, >41 GW, constitutes 15–20 % of all pregnanciesand the results of the study will thus have a great impact. The use of registries for randomisation and collection ofoutcome data represents a unique and new study design.

Trial registration: The study was registered in Current Controlled Trials, ISRCTN26113652 the 30th of March 2015 (DOI10.1186/ISRCTN26113652).

Keywords: Prolonged pregnancy, Postterm, Labour, Induced, Expectant management, Perinatal outcome, Neonataloutcome, Maternal outcome, Maternal preferences

BackgroundAccording to World Health Organisation (WHO), posttermbirth is defined as pregnancy duration of 294 days or longer,i.e. gestational week (GW) 42 and 0 days (42+0) or more,measured from the first day of the last menstrual period [1].This definition is arbitrary. Many studies that have analysedrisks and management of postterm birth also include preg-nancies from 41+0 GW (late term). Globally, the prevalenceof postterm birth is about 5–10 % [2] but the rate variesconsiderably between and within countries. Factors that in-fluence the prevalence are characteristics of the populatione.g. maternal age and number of primiparous women in thepopulation, the rate of preterm birth, interventions such ascaesarean section and induction of labour, the prevalence ofroutine ultrasound dating of pregnancy and pregnancy sur-veillance routines [3]. If ultrasound is performed at 40+0 or41+0 GW, fewer women will reach 42+0 GW as some preg-nancies will be diagnosed as risk pregnancies leading to in-duction (e.g. because of oligohydramniosis or small forgestational age foetuses). The prevalence of postterm birth(42+0 GW or more) in Sweden was 8.4 % between 1982 and1991 [4] and 6.9 % in 2013 [5].The aetiology of postterm birth is largely unknown [2].

Some rare, known causes of postterm birth are foetal an-encephaly, foetal adrenal hypoplasia or insufficiency andplacental sulphates deficiency. Risk factors for posttermbirth include: primiparity, advanced maternal age, mater-nal obesity, heredity, previous postterm pregnancy, anda male foetus [6–10].

Complications in postterm pregnanciesPerinatal mortality (PNM) is defined as the prevalenceof stillbirth (after GW 22+0) and neonatal mortalitywithin 7 days after birth. The PNM in Sweden between2004 and 2013 was 0.3‰ in GW 39–41 and 0.7‰ after42 GW. A Danish national register study (n = 125 043pregnancies) showed that PNM increased in womenwith postterm pregnancies as compared to women withterm pregnancies (adjusted odds ratio (aOR) 1.36; 95 %CI 1.08–1.72) [8]. A Swedish register study of deliveriesbetween 1982 and 1991 (n = 914 702 pregnancies) foundsignificantly higher rates of stillbirth in primiparous women

in GW 41 (1.86‰) and GW 42 (2.26‰) as compared toGW 40 (1.23‰) [4]. However, in multiparous women therisk of stillbirth was not significantly different in GW 41(0.50‰) and 42 (0.86‰) as compared to GW 40 (0.53‰).These figures were calculated based on the number ofwomen delivered in each GW. A recalculation with thenumber of non-delivered women in each GW (foetuses atrisk) as denominator, showed significantly higher inci-dences of stillbirth in GW 41 and 42, as compared to GW40, in primiparous women (0.62, 1.26 and 2.27‰, respect-ively), as well as in multiparous women (0.73, 1.00 and1.51‰, respectively). Risk factors for stillbirth in posttermpregnancies are maternal age and obesity [11].The risk of perinatal complications such as meconium

aspiration syndrome (MAS), umbilical cord complications,asphyxia, pneumonia, sepsis, convulsions, shoulder dys-tocia, traumatic injuries and peripheral nerve damage ishigher in postterm deliveries than in deliveries at term(aOR 1.1–2.0) [8]. A case-control study from Australiashowed a higher risk of neonatal encephalopathy in chil-dren born postterm, (GW 41: aOR 3.3 and GW 42: aOR13.2 versus GW 39) [12]. A Swedish study (n = 354 chil-dren) found an increased risk for developmental delay atthe age of 4–4.5 years in children born postterm comparedwith children born at term [13]. Recently, also increasedrates of obesity and early markers of the metabolic syn-drome were found in children born postterm [14, 15]. Thismay not be a result of prolonged pregnancy per se but ra-ther associated with foetal and/or parental genetic factorsthat also are associated with delayed parturition.Maternal complications increase from GW 40. The risk

of puerperal infections, postpartum bleeding, disproportion,labour dystocia, emergency caesarean sections, and cervicallacerations was higher for postterm than for term pregnan-cies in a Danish register study (n = 125 043 pregnancies)(aOR 1.2–1.6) [8].

Postterm pregnancy: induction of labour orexpectant managementRandomised controlled studies (RCTs)With the increased risks that accompany a posttermpregnancy, the obvious solution would be to suggest

Elden et al. BMC Pregnancy and Childbirth (2016) 16:49 Page 2 of 10

Page 53 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

http://www.isrctn.com/ISRCTN26113652

http://dx.doi.org/10.1186/ISRCTN26113652

Confidential: For Review Onlyinduction of labour before the pregnancy becomes post-term. Several RCTs have compared induction withawaiting spontaneous onset of labour, with or withoutfoetal surveillance, but most of the studies are small andof low methodological quality. The largest study so far isa multicentre RCT from Canada (n = 3407). The womenwere randomised at 41 GW to induction or to expectantmanagement with monitoring until the start of spontan-eous labour [16]. The PNM was similarly low in the twogroups but the induced group had significantly fewercaesarean sections. The study has been criticised becausethe induction methods were different in the groups.Prostaglandins were not used in the expectant manage-ment group, and it is speculated that this may have con-tributed to a higher caesarean section frequency in thisgroup. In a later Norwegian study (n = 500 women),women were randomised to induction at 289 days (GW41+2) or monitoring and induction at 300 days (GW 42+6) unless spontaneous labour [17]. The primary out-come (a composite neonatal outcome) was similar in thetwo groups. No case of stillbirth occurred, and there wasno difference in the rate of caesarean sections. However,the study was too small to assess PNM, a problem ac-companied by all RCTs to date. Several systematic re-views and meta-analyses of RCTs have been published.In a Health Technology Assessment (HTA) performed atSahlgrenska University Hospital (SU) in Gothenburg, 13RCTs were included (n = 6617 women) [18]. The reviewshowed no difference in PNM but found significantlyfewer cases of MAS and a lower caesarean section ratein the induction group as compared to the expectantmanagement group [18]. An updated HTA meta-analysisof 17 RCTs (n = 7223 women) was done in 2012, with amore liberal inclusion of studies (a study from 1969, ab-stracts and articles in Spanish were added) [19]. Itshowed a lower PNM in the induction group, fewercases of MAS, and no difference in caesarean sectionrate. However, a large number of women need to be in-duced to prevent one case of perinatal death (NNT =469). A methodological problem in individual studiesand systematic reviews is that important outcomes suchas PNM and hypoxic ischaemic encephalopathy (HIE)occur at a very low frequency [20–22]. Thus, very largestudies are needed to address these important outcomes.

Observational studiesA recent register study was based on Medical BirthRegistry (MBR) data on all pregnancies more than 41+2

weeks in Sweden between 2000 and 2007 (n = 119 198pregnancies) [23]. The management of postterm preg-nancy in the Stockholm region was changed in 2005from induction of labour at GW 43+0 to induction at nolater than GW 42+0. Decreases in the rates of PNM (by48 %), MAS (by 51 %) and low Apgar scores (by 31 %)

were observed. Sweden was divided into three regionsbased on the rate of deliveries at more than 42+2 weeks.It was shown that foetal surveillance differs and that re-gions with a high rate of deliveries at >GW 42+2 had ahigher rate of neonates with MAS and low Apgar scores.A recent population-based retrospective cohort study

from Scotland (n = 1 271 549 pregnancies) showed thatinduction of labour for non-medical reasons at GW 37,38, 39, 40 and 41 were associated with reduced PNM,without an increased risk of caesarean section. However,more children were admitted to the neonatal ward inthe induced group [24]. An older observational study[25] compared a group of uncomplicated pregnanciesbetween 1997 and 1999 (n = 2176) with a group of un-complicated pregnancies between 1999 and 2002 (n =3716). Management was changed in 1999 from induc-tion of labour at GW 42+0 to induction at GW 41+0.There was no difference in neonatal outcome includingstillbirth between the two time periods, but there weremore prolonged labours and caesarean sections due tofailed inductions between 1999 and 2002.

Risks of labour inductionThe RCTs referred to above [16, 17] and the systematicreviews show that induction compared with expectantpolicy in GW 41+0 or later did not increase the frequencyof caesarean sections or instrumental vaginal deliveries.Earlier observational studies, comparing induction withspontaneous start of labour at the same gestational age,showed higher rates of caesarean section after induction,particularly in women with an unripe cervix [26–28]. Aretrospective cohort study attempted to mimic the clinicalsituation by comparing women with induction in a spe-cific GW (GW 38, 39, 40, 41) with women being undeliv-ered in the same GW, who were either subsequentlyinduced or went into spontaneous labour [29]. As withearlier studies, a higher frequency of caesarean sectionswas found after induction, which was statistically signifi-cant for GW 38, 39 and 40. At GW 41+0 there was no dif-ference. A recently published meta-analysis of 31 RCTs(19 studies addressing induction for postterm birth and 12inductions for other indications) found that induction ledto a lower risk of caesarean section (OR 0.83; 95 % CI0.76–0.92) [30]. Similar findings are reported in anotherrecent meta-analysis of 157 RCTs [29, 31].A register study with data from the Norwegian MBR

between1996 and 1998 (n = 176 591 pregnancies) andthe Norwegian Cerebral Palsy Register (n = 373 children)found an independent association between induction oflabour at term and bilateral cerebral palsy (CP) (OR 3.7,95 % CI 1.8–7.5) after adjustment for maternal disease,gestational age, birth weight, prelabour rupture of themembranes [32]. However, the authors concluded thatthey were only able to speculate on possible mechanisms


Page 54 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Onlyleading to CP after labor induction due to lack of de-tailed data. Recently, an association between autismspectrum disorders and induced or augmented child-birth has been shown [33]. Some evidence suggests anoxytocin receptor deficiency in autism and the authorshypothesized that an oxytocin infusion may alter theoxytocin balance in the newborn.There is no consensus as to the optimal time of labour

induction or the optimal way of antenatal surveillance.Recently, guidelines from the Danish Society of Obstetricsand Gynaecology (DSOG 2011) and the Norwegian Dir-ectorate of Health (2012) have been published. Accordingto DSOG, all pregnant women should be delivered beforeGW 42+0. A clinical examination is recommended in GW41+0 including cardiotocography (CTG) and ultrasound.Induction will be planned for at GW 41+2–41+5, depend-ing on local tradition. In Norway, a check is also recom-mended in GW 41+0–41+2 with a clinical examination,ultrasound, CTG and amniotomy if possible. Inductionshould be started no later than GW 42+0. Induction inGW 41 is recommended for women who are older(>38 years) or obese, who have gestational diabetes or aresuspected to have a foetus being small for gestational age(SGA). The American College of Obstetricians andGynaecologists (ACOG) recommends starting monitoringbetween GW 41+0 and 42+0 with CTG and measurementof the amniotic fluid index (AFI) twice a week. In GreatBritain, The Royal College of Obstetricians and Gynaecol-ogists (RCOG) recommend induction between GW 41+0

and 42+0. The Society of Obstetricians and Gynaecologistsof Canada (SCOG) recommends antenatal examination inGW 41–42, with local guidelines for further action.

Health economic considerationsThe Canadian Multicentre Postterm pregnancy trialfrom 1995 indicated that induction of labour was associ-ated with a lower cost as compared with expectant man-agement of postterm pregnancy [34]. However, as thestudy was performed more than 20 years ago the datamay not be applicable today. Induction of labor at 41GW versus expectant management with antenatal test-ing in nulliparous women resulted in improved obstetricoutcomes and was cost-effective when analysed with adecision-analytic model [35]. The incremental cost was$10,945 per quality-adjusted life year (QUALY) gained(the range of cost-effectiveness thresholds is commonlyset at $ 50,000–100,000 per QUALY in the USA). Be-sides the analyses being based on a theoretical cohort,the authors stated some limitations in the cost-effectiveanalysis. Much of the data utilized was more than10 years old and health care costs have rapidly increased.In addition, they conducted sensitivity analysis over wideranges of cost inputs and concluded that better cost data

in this area would facilitate more accurate estimates ofthe cost-effectiveness of induction of labour [35].

Women’s experiencesA Swedish study (n = 1111 women) found that inductionof labour irrespective of cause for induction was associ-ated with a more negative experience of childbirth [36].This finding is inconsistent with a cross-sectional study(n = 252 women) from Nigeria, which showed that 71 %of the women expressed satisfaction with the inductionprocess [37]. Only one RCT has assessed the woman’sexperiences of induction or expectant policy for post-term pregnancy (n = 508 women) [38]. At the time ofrandomisation, 74 % would have preferred induction ifthey had been able to choose. The majority of women(84 %) who were induced had a positive childbirth ex-perience. Despite that induction lead to more intensivelabour, 74 % preferred being induced compared withawaiting spontaneous on-set of labour (p = 0.001).

AimThe aim with this study is to evaluate if a policy of in-duction of labour at 41 GW (early induction) is superior,in terms of neonatal and maternal outcomes, as com-pared to expectant management and induction at 42+0

(late induction) in healthy women with a low risk single-ton pregnancy.

Methods/DesignThis multicentre registry based randomised controlled trialnamed SWEPIS (SWEdish Post-term Induction Study), willbe performed at 4 university clinics and 1 county hospitalin Sweden from 01-09-2015 to 31-08-2018. The centrescomprise more than 65 000 deliveries per year. Eligible par-ticipants are healthy women ≥18 years old, who are able tounderstand oral and written information, with a normal livesingleton pregnancy in cephalic presentation at 41+0 GW,estimated with a first or second trimester ultrasound.Women will be excluded if they have had a previous caesar-ean section or other uterine surgery, pregestational and in-sulin dependent gestational diabetes, hypertensive disordersin pregnancy including preeclampsia, multiple pregnancy,foetus in breech or transverse position, oligohydramniosis(amniotic fluid index <50 mm or deepest vertical pocket<20 mm), small for gestational age (<−22 % according to aSwedish reference [39], foetal malformations and contrain-dications to vaginal delivery such as placenta previa.The Regional Ethics Board in Gothenburg approved

the study in May 2014 (Dnr: 285–14).General information of the study will be available at

each antenatal care unit where women who most likelychose to give birth at any of the including hospitals willattend. All pregnant women are given oral and writteninformation about the study at around GW 40 at the


Page 55 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Onlyantenatal care units. All potentially eligible women will beconsecutively contacted for further information. Womenwho are interested in participating in the study are bookedfor a visit at the antenatal clinic at the respective hospitalat GW 41+0–41+1. Further information is given and writ-ten informed consent is obtained. Thereafter, an ultra-sound examination is performed. Women with a foetuswith an estimated foetal weight <22 % according to theSwedish gestational and gender adjusted reference curveand/or AFI <50 mm or DVP <20 mm will not be rando-mised but will be followed-up according to clinical rou-tine. Consenting eligible women will be enrolled andrandomised. Eligible women who do not want to partici-pate in the study will be asked to fill in a web based orpostal written consent for collection of information ondemographic background variables, previous posttermpregnancy and obstetric and neonatal outcomes (this is tomake it possible to check for selection bias).

RandomisationFigure 1 shows randomisation and progress through thetrial. The doctor or midwife who performs the ultra-sound will enrol eligible women. Randomisation will becompleted at the antenatal clinic at GW 41+0 to 41+1.Block randomisation with a fix block size (unknown to

the investigator) will be done online, 1:1, using the Preg-nancy Register with a module specifically developed forthe study by MedSciNet AB, Stockholm, Sweden. Strati-fication will be performed for centre and parity (primi-parity and multiparity).

Labour inductionWomen in the intervention group (early induction) willbe induced within 24 h after randomisation. Women inthe control (expectant management/late induction) willbe induced at GW 42+0 (and no later than GW 42+1).After randomisation, women in the control group arefollowed-up at the ordinary antenatal clinic according toclinical routine for GW 41+–42+0.

Methods of labour induction (early at GW 41+0–41+2 orlate at GW 42+0–42+1)Before labour induction the following examinations areperformed: blood pressure, urine analysis, abdominaland cervical examination and CTG. If there are no con-traindications for induction of labour, induction will beperformed according to the following:

1. if the foetal head is well engaged and the cervix isripe (Bishop score ≥6 for primiparous and ≥5 for

Fig. 1 Flow-shart of the study


Page 56 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Onlymultiparous) amniotomy is performed. Oxytocininfusion will be started after 2 h if no or not regularcontractions.

2. if the cervix is unripe (Bishop score <6) or the fetalhead is not engaged any of the following methodsare used (according to the routine at the clinic):a. mechanical dilation using a Foley like catheter

(BARD or Cook)b. misoprostol (Cytotec®) orallyc. misoprostol (Misodel®) controlled-released vaginal

insertd. prostaglandin E2 (Minprostin®) gel vaginallye. prostaglandin E2 (Propess®) vaginal insert

Mechanical dilation may be followed by prostaglandinsand vice versa. If the woman is not in active labour orcervix is still unripe after 48 h, the clinician together withthe woman will decide whether to perform a caesareansection or to continue expectant management until GW42+0. All inductions will be performed with the patientstaying at the hospital until delivery. The clinic’s protocoland guidelines for induction and surveillance during in-duction will be followed. During the active phase of labourfetal surveillance with continuous CTG with or withoutST segment analysis (STAN) in combination with foetalscalp blood sampling is recommended.

Outcome measuresPrimary outcome measureA composite of stillbirth, neonatal mortality and neonatalmorbidity. Stillbirth is defined as intrauterine foetal deathof a foetus that was alive at randomisation. Neonatal mor-tality is defined as live births with death day 0–27. Neo-natal morbidity is defined as at least one of the followingvariables: Apgar score <7 at 5 min, metabolic acidosis de-fined as pH <7.05 and base deficit >12 mmol/l in umbilicalartery or pH <7.00 in umbilical artery, HIE I-III, intracra-nial haemorrhage, neonatal convulsions, meconium aspir-ation syndrome (MAS), mechanical ventilation, obstetricbrachial plexus injury.

Secondary outcome measures- neonatesAdmittance to neonatal intensive care unit, birth weight,macrosomia (>4.5 kg), Apgar score <4 at 5 min, thera-peutic hypothermia, neonatal jaundice, pneumonia, sep-sis, costs.

Secondary outcome measures-mothersUse of epidural analgesia, caesarean section, assisted vagi-nal delivery, duration of labour, episiotomy, perineal lacer-ations III-IV, shoulder dystocia, postpartum hemorrhage(>1000 ml), chorioamnionitis, wound infection, urinarytract infection, endometritis, sepsis, costs.

Data on women’s attitudes, experiences and health-related quality of life will be collected, only in the RegionVastra Götaland subpopulation, by questionnaires (webbased or postal).

After randomisation Personality (Big Five), Depression/anxiety (HADS), Health related quality of life (Euro-Qol –VAS & Euro-Qol-5D, Self-efficacy (The General Self-Efficacy Scale (GES) [40–44].Health-related quality of life (Euro-Qol –VAS & Euro-

Qol-5D, the General Self-Efficacy Scale (GES),

Three months after delivery Depression/anxiety (Hos-pital anxiety and depression scale, HADS) [44], TheChildbirth Experience Questionnaire (CEQ) [41]. Theinstruments will be eligible in English, Spanish, Arabic,Bosnian, Croatian, Serbian, Farsi and Somali language.For centres with the same medical record system

e.g.Siemens Obstetrix (Siemens Medical solutions) wewill register data on satisfaction with delivery (Visualanalog scale 0–10) [45].

Data collectionData on background variables, obstetric and neonataloutcomes will be obtained from the Swedish PregnancyRegister and Swedish Neonatal Quality Register (SNQ)after delivery. The Pregnancy Registry is a certified na-tional quality registry initiated by the Swedish Healthcare that collect and process information all the wayfrom early pregnancy to a few months after birth. SNQis a certified national quality registry initiated by theSwedish Health care that collect and process informa-tion about the neonate (www.snq.se).

Statistical analysisWe anticipate that induction of labour at GW 41 ascompared to induction at GW 42 will reduce the pri-mary outcome by one third from 2.74 to 1.84 % (level ofsignificance 0.05, power 80 %, dropout rate 10 %). Theprimary composite outcome of 2.74 % is based on thefollowing rates of perinatal mortality and neonatal mor-bidity in GW 41+3 from Region Skane between 2000 and2010: HIE 0.22 %, MAS 0.36 %, obstetric brachial plexusinjury 0.05 %, Apgar score <7 at 5 min 1.58 %, stillbirthand neonatal mortality 0.53 %. We need a sample size of10 038 women to be randomised, 5 019 to induction oflabour at GW 41+0 and 5 019 to expectant management.Since primary outcome is a serious and rare event, weconsider that a 33 % reduction is clinically relevant. In2013, Region of Vastra Gotaland, Region Skane,Stockholm and Uppsala had 18 261, 14 719, 28 335 and3 787 deliveries, respectively, altogether 65 102 deliver-ies. During a 3 year period we calculate with 195 306 de-liveries. Given a rate of 18 % of women being


Page 57 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

http://www.snq.se/

Confidential: For Review Onlyundelivered at 41+0 GW, this will give us 34 755 womenduring 3 years. We estimate that around 33 % (11 469)of these women can be included.The primary statistical analysis will be the comparison

of primary composite outcome (stillbirth, neonatal mor-tality and morbidity) between the two randomisedgroups with two-sided Fisher’s exact test on theintention to treat (ITT) [46] populations at significancelevel 0.05. Adjustments for baseline variables if necessaryand interaction analyses will be performed with multi-variable logistic regression analyses. For variables withinteractions p <0.10 with primary efficacy variable, sub-group analyses will be performed. Women with spontan-eous labour or prelabour rupture of the membranes(PROM) after randomisation but before planned induc-tion will be included in the statistical analysis accordingto ITT. Women with pregnancy complications after ran-domisation indicating any need of intervention are in-cluded in the statistical analysis according to ITT.Foetuses/newborns with lethal birth defects will be ex-cluded in the analysis of stillbirth, neonatal mortalityand morbidity. The median, 95 % CI, quartiles, meansand SD are calculated when appropriate. The Mann–Whitney U test is used to compare differences betweenthe two groups concerning continuous variables, Fisher’sexact test for dichotomous variables, Mantel-HaenszelChi-square test for ordered categorical variables andChi-square test for unordered categorical variables.The statistician will be blinded to group and interven-

tion. Otherwise blinding is not possible for practicalreasons.

Cost-effectiveness analysesCost-effectiveness analyses will be performed comparinginduction at GW 41+0 with induction at GW 42+0. Theprimary measures of effectiveness will be a composite ofstillbirth, neonatal mortality and neonatal morbidity (i.e.primary outcome) [47]. A broad range of sensitivity ana-lyses will be performed, which includes calculating CIsof the cost-effectiveness measures, using bootstrapping[47]. In order to extrapolate the outcomes observed inthis study to a longer time span we will utilize a simulationmodel. The model will be constructed and parameterizedusing estimates of relevant parameters obtained within theproject. Moreover, associations between neonatal outcomesand health outcomes later in life that was not studied inthis project will be collected from the epidemiology litera-ture. Initially, two modelling structures will be considered:a Markov structure and a discrete event structure [48]. Thetypical Markov model uses a cohort and state-transitiontechnique, while the discrete event structure means thatsimulations are event-based and performed at the individ-ual level. The choice between these two main alternatives

will be performed according to the guidelines provided inBrennan et al. [48].

Serious adverse eventsThe following serious adverse events (SAE) will be iden-tified: stillbirth, neonatal death, severe neonatal morbid-ity, defined as intraventricular haemorrhage, asphyxia orMAS with admission to NICU, maternal death, severematernal morbidity defined as admission to intensivecare unit and events related to induction of labour e.g.uterine rupture. All SAEs are reported to the main in-vestigators, who will report to the Data and Safety Moni-toring Board (DSMB).

Data safetyThe steering committee’ is responsible for the study de-sign, the co-ordination between centres, the progress ofthe study, and for the results being statistically analysedand summarised for publication. The steering committeeare also principal investigators, i.e. they are responsiblefor the project at their own centre for example adher-ence to the study protocol, recruitment and any practicalproblem that may arise within the frame of the project.The DSMB consist of experts (one statistician, a seniorobstetrician and a midwife) who will supervise the study.The DSMB will secure the project management throughperiodical reviews and will make the recommendationsconcerning the continuation, modification or termin-ation of the trial.

Trial statusBesides the literature search resulting in the systematicreview and meta-analysis and the HTA reports [18, 19]there are no clinical results. Overall recruitment startdate: 01/09/2015 and overall trial end date: 31/12/2018.

DiscussionThe planned study is the first RCT utilising the nationalPregnancy Register (developed by co-applicant OS) forcollecting data on maternal characteristics, pregnancy,delivery and postnatal care and SNQ for data on neo-natal outcome. Randomisation will be performed using acomputerised module linked to the Pregnancy Register.Such a design has recently been applied successfully incardiovascular research [49, 50] and has several import-ant advantages including costs, ease of recruitment andfacilitating the translation of the study results to clinicalpractice.Our primary and secondary outcome measures are in

the interest of the woman, the expectant partner and thechild. There are no high quality studies with adequatesample size investigating if early induction is superior tostandard care (expectant management/late induction) interms of maternal and neonatal outcomes. The optimal


Page 58 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Onlytiming of offering induction of labour to women at lateterm or postterm therefore needs further investigation.Women’s experiences and opinions about the choiceshave not been adequately evaluated. In this study, wealso aim to evaluate women’s attitudes and experienceswith questionnaires. In the early induction group we willhave a higher proportion of induction of labour, withpossible negative delivery experiences. In the late induc-tion group, the proportion of induction will be lower,but the experience of another week of waiting “post duedate” is not evaluated. Further, the foetal growth con-tinues between 41 and 42 GW, leading to an expectedhigher infant birth weight and more “macrosomic in-fants” (>4.5 kg) in the late compared to the early induc-tion group. Giving birth to macrosomic infants isassociated with more complications, leading to a pos-sible more negative experience in the late inductiongroup. The outcome “women’s experiences and opin-ions” of each intervention (early or late induction) isthus complex and therefore an important outcome ofour study. One of the advantages of the register-baseddesign is that the primary and secondary outcomes willbe collected continuously through national quality andhealth registers. These data can be used for implementa-tion after the study has been finalized. The conse-quences of a new clinical management can be evaluated(reduction in e.g. perinatal mortality, HIE, caesarean sec-tion etc.) using national data from the Pregnancy Regis-ter and SNQ. The study results will fill several gaps inhealth care knowledge and contribute to establish goodevidence-based routines regarding management of post-term pregnancy. Support for if an early intervention inthe natural course of pregnancy is superior for thehealth of children and mothers is important, to justifyan intervention (labour induction) in 15–20 % of allwomen whose pregnancy has lasted ≥41+0 GW in the fu-ture. Sweden is suited to run a high quality RCT: publichealth care system, almost 100 % of pregnant women at-tend antenatal care, nearly everybody gives birth at a de-livery hospital and there is usually (except theStockholm region) only one hospital with a deliveryward for a certain geographic region (reduces selectionbias), low rates of maternal and perinatal mortality, com-parably low caesarean section rates and opposed tomany other countries we have not yet started to induceearlier so we still have the possibility to run this RCT.Also, this project is the first study supported by the re-cently started Swedish National Network on ClinicalStudies in obstetrics and gynaecology (SNAKS). The net-work comprises one member from each university re-gion in Sweden and is supported by the Swedish Societyof Obstetrics and Gynaecology (SFOG). If the interven-tion, early induction, is shown to be superior to expect-ant management/late induction, new local, regional and

national guidelines may therefore be developed within ashort time frame (approximately 3 years).

Project organization and co-workersThe following five centres will participate in the study:the Departments of Obstetrics and Gynecology at SouthAlvsborg County Hospital (SAS), Sahlgrenska UniversityHospital, Skane University Hospital (SUS), KarolinskaHospital Solna/Huddinge, Stockholm, Uppsala UniversityHospital. Our aim is to include most of the centres inSweden through SNAKS. The project will be led by asteering committee which preliminarily has the followingmembers: Associate professor Ulla-Britt Wennerholm,Professor Henrik Hagberg, Professor Christina Bergh, As-sociate professor Andreas Herbst, PhD Maria Bullarbo,PhD Helen Elden, PhD Verena Sengpiel, PhD Sissel Salt-vedt, Associate professor Olof Stephansson, Associateprofessor Anna-Karin Wikström and Associate professorLars Ladfors. The steering committee is responsible forthe study design, the co-ordination between centres, theprogress of the study, and for the results being analysedand summarised for publication. The steering committeeare also principal investigators, i.e. they are responsible forthe project at their own centre for example adherence tothe study protocol, recruitment and any practical problemthat may arise within the frame of the project.There will be further coordinating clinical investigators

at each site. One midwife will be responsible for thestudy implementation at each site. At SU this will beAnna Wessberg, midwife and registered PhD student.Professor in Health Economics Kristian Bolin is respon-sible for the health economic analysis and Nils GunnarPehrsson is responsible for the statistical analysis.

AbbreviationsACOG: American College of Obstetricians and Gynaecologists; AFI: amnioticfluid index; aOR: adjusted odds ratio; BMI: body mass Index; CI: confidenceinterval; CTG: cardiotocography; DSMB: data and safety monitoring board;DSOG: Danish Society of Obstetrics and Gynaecology; eCRF: electronic caserecord form; GW: gestational week; HADS: hospital anxiety and depression scale;HIE: ischaemic encephalopathy; HTA: health technology assessment;ITT: intention to treat; MAS: meconium aspiration syndrome; MBR: medical birthregistry; NNT: numbers needed to treat; NM: neonatal mortality; PNM: perinatalmortality; QUALYS: quality-adjusted life year; RCT: randomised controlled trial;RCOG: Royal College of Obstetricians and Gynaecologists; SAE: seriousadverse events; SAS: South Alvsborg County Hospital; SCOG: Society ofObstetricians and Gynaecologists of Canada; SD: standard deviation;SFOG: Swedish Society of Obstetrics and Gynaecology; SNAKS: Swedish Networkon Clinical Studies; SNQ: Swedish Neonatal Quality Register; SU: SahlgrenskaUniversity Hospital; SUS: Skane University Hospital; SWEPIS: SWEdish Post-termInduction Study; VAS: visual analog scale; WHO: World Health Organisation.

Competing interestsThe authors declare that they have no competing interests.

Authors’ contributionsHE, HH, AW, VS, AH, MB, CB, KB, SM, SS, OS, AKW, LL, UBW participated in thedesign of the study and helped to draft the study protocol. UBW and HH aregrant holders. All authors read and approved the final manuscript. Allauthors contributed to refinement of the study protocol and approvedthe final manuscript.


Page 59 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review OnlyAcknowledgementsThe Foundation of the Health and Medical care committee of the Region ofVastra Götaland, Sweden (Grant number: VGFOUREG-387351) and The Swed-ish government and the county councils concerning economic support ofresearch and education of doctors (ALF-agreement) supported this work.

FundingThis study is supported by grants from the Foundation of the Health andMedical care committee of the Region of Vastra Gotaland, Sweden (VGFOUREG-387351) (H.H.), grants from the Health Technology Centre (HTA) (UB.W.) andgrants from Swedish governmental grants to researchers in the public healthservice (ALFGBG-440301) (UB.W.)

Author details1Gothenburg University, Institute of Health and Caring Sciences, SahlgrenskaAcademy, S-405 30 Gothenburg, Sweden. 2Gothenburg University, Perinatalcentre, Department of Obstetrics and Gynecology, Institute of ClinicalSciences, Sahlgrenska Academy, Sahlgrenska University Hospital, East, S-41685 Gothenburg, Sweden. 3Skane University Hospital, S-214 28 Malmo,Sweden. 4Gothenburg University, Reproductive Medicine, Department ofObstetrics and Gynecology, Institute of Clinical Sciences, SahlgrenskaAcademy, Sahlgrenska University Hospital, S-416 85 Gothenburg, Sweden.5Gothenburg University, Department of Economics and Statistics, SchoolBusiness, Economics and Law, P.O. Box 640S-405 30 Gothenburg, Sweden.6South Alvsborg County Hospital, Department of Obstetrics and Gynecology,S- 501 82 Boras, Sweden. 7Department of Obstetrics and Gynecology,Karolinska University Hospital, 171 76 Stockholm, Sweden. 8ClinicalEpidemiology Unit, Department of Medicine, Solna, Karolinska Institute, S-17176 Stockholm, Sweden. 9Department of Women’s and Children’s Health,Division of Obstetrics and Gynecology, Solna, Karolinska University Hospital,S-171 76 Stockholm, Sweden. 10Uppsala University, Women’s and Children’sHealth, Akademiska Hospital, SE-751 85 Uppsala, Sweden.

Received: 2 April 2015 Accepted: 3 March 2016

References1. WHO. Recommended definitions, terminology and format for statistical

tables related to the perinatal period and use of a new certificate for causeof perinatal deaths. Modifications recommended by FIGO as amendedOctober 14, 1976. Acta Obstet Gynecol Scand. 1977;56:247–53.

2. Shea KM, Wilcox AJ, Little RE. Postterm delivery: a challenge forepidemiologic research. Epidemiology (Cambridge, Mass). 1998;9:199–204.

3. Norwitz ER, Snegovskikh VV, Caughey AB. Prolonged pregnancy: whenshould we intervene? Clin Obstet Gynecol. 2007;50:547–57.

4. Ingemarsson I, Kallen K. Stillbirths and rate of neonatal deaths in 76,761postterm pregnancies in Sweden, 1982–991: a register study. Acta ObstetGynecol Scand. 1997;76:658–62.

5. Pregnancies, Deliveries and Newborn Infants. The Swedish Medical BirthRegister 1973–2013. Stockholm: Socialstyrelsen; 2014.

6. Mittendorf R, Williams MA, Berkey CS, Lieberman E, Monson RR. Predictorsof human gestational length. Am J Obstet Gynecol. 1993;68:480–4.

7. Bakketeig LS, Hoffman HJ. The tendency to repeat gestational age and birthweight in successive births, related to perinatal survival. Acta ObstetGynecol Scand. 1983;62:385–92.

8. Olesen AW, Westergaard JG, Olsen J. Perinatal and maternal complicationsrelated to postterm delivery: a national register-based study, 1978–1993. AmJ Obstet Gynecol. 2003;189:222–7.

9. Olesen AW, Westergaard JG, Olsen J. Prenatal risk indicators of a prolongedpregnancy. The Danish Birth Cohort 1998–2001. Acta Obstet Gynecol Scand.2006;85:1338–41.

10. Oberg AS, Frisell T, Svensson AC, Iliadou AN. Maternal and fetal geneticcontributions to postterm birth: familial clustering in a population-basedsample of 475,429 Swedish births. Am J Epidemiol. 2013;177:531–7.

11. Aune D, Saugstad OD, Henriksen T, Tonstad S. Maternal body mass indexand the risk of fetal death, stillbirth, and infant death: a systematic reviewand meta-analysis. JAMA. 2014;311:1536–46.

12. Badawi N, Kurinczuk JJ, Keogh JM, Alessandri LM, O’Sullivan F, Burton PR, et al.Antepartum risk factors for newborn encephalopathy: the Western Australiancase-control study. BMJ. 1998;317:1549–53.

13. Lindstrom K, Fernell E, Westgren M. Developmental data in preschoolchildren born after prolonged pregnancy. Acta Paediatrica (Oslo, Norway :1992). 2005;94:1192–7.

14. Ayers S, Collenette A, Hollis B, Manyonda I. Feasibility study of a Latest Dateof Delivery (LDD) system of managing pregnancy. J Psychosom ObstetGynaecol. 2005;26:167–71.

15. Beltrand J, Soboleva TK, Shorten PR, Derraik JG, Hofman P, Albertsson-Wikland K, et al. Post-term birth is associated with greater risk ofobesity in adolescent males. J Pediatr. 2012;160:769–73.

16. Hannah ME, Hannah WJ, Hellmann J, Hewson S, Milner R, Willan A.Induction of labor as compared with serial antenatal monitoring inpost-term pregnancy. A randomized controlled trial. The CanadianMulticenter Post-term Pregnancy Trial Group. N Engl J Med.1992;326:1587–92.

17. Heimstad R, Skogvoll E, Mattsson LA, Johansen OJ, Eik-Nes SH, Salvesen KA.Induction of labor or serial antenatal fetal monitoring in posttermpregnancy: a randomized controlled trial. Obstet Gynecol. 2007;109:609–17.

18. Wennerholm UB, Hagberg H, Brorsson B, Bergh C. Induction of labor versusexpectant management for post-date pregnancy: is there sufficient evidencefor a change in clinical practice? Acta Obstet Gynecol Scand. 2009;88:6–17.

19. Wennerholm U-B, Flisberg A, Hagberg H, Ladfors L, Jivegård L, Svanberg T,et al. Induktion av förlossning vid 41 fullgångna till och med 42 fullgångnagraviditetsveckor, en uppdatering av mini-HTA VGR 2007 [Induction oflabour at 41 completed gestational weeks until 42 completed completedgestational weeks, update of mini-HTA VGR 2007]. In. Göteborg: VästraGötalandsregionen: Göteborg: Västra Götalandsregionen, SahlgrenskaUniversitetssjukhuset, HTA-centrum; 2012.

20. Badawi N, Kurinczuk JJ, Keogh JM, Alessandri LM, O’Sullivan F, Burton PR, et al.Intrapartum risk factors for newborn encephalopathy: the Western Australiancase-control study. BMJ. 1998;317:1554–8.

21. Smith J, Wells L, Dodd K. The continuing fall in incidence of hypoxic-ischaemic encephalopathy in term infants. BJOG. 2000;107:461–6.

22. Thornberg E, Thiringer K, Odeback A, Milsom I. Birth asphyxia: incidence,clinical course and outcome in a Swedish population. Acta Paediatrica(Oslo, Norway : 1992). 1995;84:927–32.

23. Grunewald C, Hakansson S, Saltvedt S, Kallen K. Significant effects on neonatalmorbidity and mortality after regional change in management of post-termpregnancy. Acta Obstet Gynecol Scand. 2011;90:26–32.

24. Stock SJ, Ferguson E, Duffy A, Ford I, Chalmers J, Norman JE. Outcomes ofelective induction of labour compared with expectant management:population based study. BMJ. 2012;344:e2838.

25. Fok WY, Chan LY, Tsui MH, Leung TN, Lau TK, Chung TK. When to inducelabor for post-term? A study of induction at 41 weeks versus 42 weeks. EurJ Obstet Gynecol Reprod Biol. 2006;125:206–10.

26. Luthy DA, Malmgren JA, Zingheim RW. Cesarean delivery after elective inductionin nulliparous women: the physician effect. Am J Obstet Gynecol. 2004;191:1511–5.

27. Vahratian A, Zhang J, Troendle JF, Sciscione AC, Hoffman MK. Laborprogression and risk of cesarean delivery in electively induced nulliparas.Obstet Gynecol. 2005;105:698–704.

28. Roos N, Sahlin L, Ekman-Ordeberg G, Kieler H, Stephansson O. Maternal riskfactors for postterm pregnancy and cesarean delivery following laborinduction. Acta Obstet Gynecol Scand. 2010;89:1003–10.

29. Caughey AB, Nicholson JM, Cheng YW, Lyell DJ, Washington AE. Inductionof labor and cesarean delivery by gestational age. Am J Obstet Gynecol.2006;195:700–5.

30. Wood S, Cooper S, Ross S. Does induction of labour increase the risk ofcaesarean section? A systematic review and meta-analysis of trials inwomen with intact membranes. BJOG. 2013;121:6.

31. Mishanina K, Rogozinska E, Thatthi T, Uddin-Khan R, Khan KS, Meads C. Useof labour induction and risk of cesarean delivery: a systematic review andmeta-analysis. CMAJ. 2014;186:9.

32. Elkamil AI, Andersen GL, Salvesen KA, Skranes J, Irgens LM, Vik T. Inductionof labor and cerebral palsy: a population-based study in Norway. ActaObstet Gynecol Scand. 2011;90:83–91.

33. Gregory SG, Anthopolos R, Osgood CE, Grotegut CA, Miranda ML. Associationof autism with induced or augmented childbirth in North Carolina BirthRecord (1990–1998) and Education Research (1997–2007) databases.JAMA Pediatr. 2013;167:959–66.

34. Goeree R, Hannah M, Hewson S. Cost-effectiveness of induction of labourversus serial antenatal monitoring in the Canadian Multicentre PosttermPregnancy Trial. CMAJ. 1995;152:1445–50.


Page 60 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Only35. Kaimal AJ, Little SE, Odibo AO, Stamilio DM, Grobman WA, Long EF, et al.

Cost-effectiveness of elective induction of labor at 41 weeks in nulliparouswomen. Am J Obstet Gynecol. 2011;204:2.

36. Waldenstrom U. Experience of labor and birth in 1111 women. J PsychosomRes. 1999;47:471–82.

37. Ezeanochie M, Olagbuji B, Ande A. Women’s concerns and satisfaction withinduced labour at term in a Nigerian population. Niger Postgrad Med J.2013;20:1–4.

38. Heimstad R, Romundstad PR, Hyett J, Mattsson LA, Salvesen KA. Women’sexperiences and attitudes towards expectant management and inductionof labor for post-term pregnancy. Acta Obstet Gynecol Scand. 2007;86:950–6.

39. Marsal K, Persson PH, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterinegrowth curves based on ultrasonically estimated foetal weights. ActaPaediatrica (Oslo, Norway : 1992). 1996;85:843–8.

40. Love J, Moore CD, Hensing G. Validation of the Swedish translation of theGeneral Self-Efficacy scale. Qual Life Res. 2012;21:1249–53.

41. Dencker A, Taft C, Bergqvist L, Lilja H, Berg M. Childbirth experience questionnaire(CEQ): development and evaluation of a multidimensional instrument. BMCPregnancy Childbirth. 2010;10:81. doi:10.1186/1471-2393-10-81.

42. Rabin R, De Charro F. EQ-5D: a measure of health status from the EuroQolGroup. Ann Med. 2001;33:337–43.

43. John O, Srivastava S. The big five trait taxonomy: history, measurement andtheoretical perspectives. In: Pervin L, John O, editors. Handbook of personality:theory and research..dn. New York: Guilford Press; 1999. p. 102–38.

44. Matthey S, Ross-Hamid C. Repeat testing on the Edinburgh DepressionScale and the HADS-A in pregnancy: differentiating between transient andenduring distress. J Affect Disord. 2012;141:213–21.

45. Huskinsson E. Measurement of pain. Lancet. 1974;9:1127–31.46. Gould A. A new approach to the analysis of clinical drug trials with

withdrawals. Biometrics. 1980;36:721–7.47. Drummond M. Methods for the Economic Evaluation of Health Care

Programmes. Oxford: Oxford Medical Publications; 2005.48. Brennan A, Chick SE, Davies R. A taxonomy of model structures for

economic evaluation of health technologies. Health Econ. 2006;5:1295–310.49. Frobert O, Lagerqvist B, Olivecrona GK, Omerovic E, Gudnason T, Maeng M, et al.

Thrombus aspiration during ST-segment elevation myocardial infarction. N Engl JMed. 2013;369:1587–97.

50. Lagerqvist B, Frobert O, Olivecrona GK, Gudnason T, Maeng M, Alstrom P, et al.Outcomes 1 year after thrombus aspiration for myocardial infarction. N Engl JMed. 2014;371:1111–20.

• We accept pre-submission inquiries

• Our selector tool helps you to find the most relevant journal

• We provide round the clock customer support

• Convenient online submission

• Thorough peer review

• Inclusion in PubMed and all major indexing services

• Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step:


Page 61 of 60


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

http://dx.doi.org/10.1186/1471-2393-10-81

BMJ · 2019-11-25 · Confidential: For Review Only Induction of labour at 41 weeks versus...

Documents

Transcript of BMJ · 2019-11-25 · Confidential: For Review Only Induction of labour at 41 weeks versus...