DOI: 10.1111/1471-0528.12846
General obstetrics
www.bjog.org
A randomised controlled trial of outpatient compared with inpatient cervical ripening with prostaglandin E2 (OPRA study) C Wilkinson,a R Bryce,b,c P Adelson,d D Turnbulld a
Maternal-Fetal Medicine, Women’s and Children’s Hospital, Adelaide, b Obstetrics and Gynecology, Flinders Medical Center, Flinders University, Bedford Park, d School of Psychology, University of Adelaide, Adelaide, SA, Australia Correspondence: C Wilkinson, Department of Obstetrics and Gynecology, Women’s and Children’s Hospital, 72 King William Road, North Adelaide, SA 5006, Australia. Email [email protected] c
Accepted 3 March 2014. Published Online 14 May 2014.
Objective To compare clinical outcomes from outpatient with
inpatient cervical prostaglandin E2 ripening for low risk labour induction. Design Randomised controlled trial. Setting Two tertiary hospitals in Adelaide, Australia. Population Women with uncomplicated term pregnancies,
scheduled for induction of labour for reasons of post-dates or for social reasons. Methods Eight hundred and twenty-seven women were
randomised at induction decision to an outpatient or inpatient group. All women had CTG monitoring before and after vaginal PgE2 administration. The inpatient women were admitted and the outpatient group was discharged home overnight with instructions to return if labour established or rupture of membranes occurred. Next morning, outpatient women were admitted for rupture of membranes or further ripening. After membrane rupture, both groups received routine care. Main outcome measures Oxytocin use, maternal and fetal
outcomes, and whether planned outpatient management was achievable. Results There were no significant differences in oxytocin use (2.5% difference, CI 4.3 to 9.4), caesarean sections ( 0.59% difference, CI 6.3 to 5.1), epidural use (1.5% difference, CI 5.1 to 8.2), vaginal delivery within 24 hours ( 8.2% difference,
CI 17.6 to 1.3) or labour complications. More than half of the randomised women did not receive the intervention as they laboured spontaneously, or did not require ripening. The post-hoc analysis of women who received ripening also indicated no statistically significant differences in the patterns or results, with the exception of outpatient women having a longer mean length of active labour (mean difference 66 minutes, CI 4–128 minutes). Outpatient women who received ripening were diagnosed more frequently with non-reassuring CTG monitoring and hyperstimulation, with less than half of the women going home and remaining home overnight. Conclusions This study demonstrated no clinical advantage or
disadvantage in prostaglandin E2 outpatient cervical ripening. Uterine stimulation following prostaglandins may preclude a woman from going home or remaining at home overnight and may not be the best agent for outpatient ripening. Keywords Cervical ripening, induction of labour, outpatient ripening, prostaglandin E2. Linked article This article is commented on by WA Grobman, p. 105 in this issue. To view this mini commentary visit http://dx.doi.org/10.1111/ 1471-0528.12849. The article has journal club questions by EYL Leung, p. 106 in this issue. To view these visit http://dx.doi.org/10.1111/ 1471-0528.12847.
Please cite this paper as: Wilkinson C, Bryce R, Adelson P, Turnbull D. A randomised controlled trial of outpatient compared with inpatient cervical ripening with prostaglandin E2 (OPRA study). BJOG 2015;122:94–106.
Introduction Cervical ripening for induction of labour is one of the commonest interventions in obstetrics.1–3 In 2010, 30% of South Australian pregnant women had labour induced. Of those inductions, 58% required cervical ripening with prostaglandins.4
94
Ripening a cervix which is “unfavourable” for induction is done with prostaglandin analogues as the most common pharmaceutical method. The South Australian guidelines recommend vaginal prostaglandin E2 (PGE2) gel with hospital admission for pharmaceutical cervical ripening,5 but the necessity of hospital admission for cervical ripening, although demonstrating clinical caution, is unproven.
ª 2014 Royal College of Obstetricians and Gynaecologists
RCT of PGE2 outpatient ripening for labour induction
Outpatient cervical priming offers possible psychosocial benefits for a mother in a more familiar home environment, perhaps facilitating a more natural establishment of labour.6,7 Cost savings may also be possible.8,9 Based on this reasoning, the 2008 NICE clinical guideline recommends, “Studies are needed to assess the safety, efficiency and clinical and cost effectiveness of outpatient inductions. . ..taking into account women’s views”.10 Also supporting further studies, two recent Cochrane reviews of outpatient induction conclude that there is insufficient evidence to establish which induction techniques are most effective and safe to use in outpatient settings.3,11 Of the studies in these reviews, few examine different methods of induction directly in home and hospital settings. The primary study objective of this randomised controlled trial (acronym OPRA) was to ascertain differences in clinical outcomes associated with allowing women to go home after priming, rather than staying in hospital. The biological plausibility of clinical differences was that women who were better rested after being in their own environments, rather than having experienced a sleepless night in hospital, would be better able to cope with the stresses of labour, have different thresholds of intervention, and might labour better if they were less physically stressed. Consistent with this hypothesis, a number of previously reported smaller controlled trials using various forms of prostaglandins suggested that there was less need for oxytocin use in women allowed to go home after ripening, compared with inpatient use.6,24,26 Other clinical outcomes that we hypothesised may have been associated with this model of care and less oxytocin use, were a reduced length of labour and analgesic use, an increase in normal births, comparable perinatal and neonatal outcomes, and improved psychosocial outcomes for the mother with regard to satisfaction with care, postnatal depression and mode of infant feeding at 7 weeks. In addition, the study assessed the impact of outpatient priming on midwives and examined the cost of outpatient priming, and we performed a discrete choice experiment. The psychosocial,12 costing,9 midwives study13 and discrete choice14 results are presented elsewhere.
Methods The trial was conducted from August 2008 to May 2011 at two tertiary referral hospitals, accounting for 39% of South Australia’s approximately 20 000 annual births.4 Women with uncomplicated term pregnancies were recruited from obstetric and midwifery clinics by a research midwife at the visit when clinicians scheduled an induction. South Australian guidelines recommend induction of labour to prevent the pregnancy exceeding 41 weeks’ gestation.5 Eligibility for the trial included; ≥18 years old; communication in English; singleton, cephalic, term pregnancy
ª 2014 Royal College of Obstetricians and Gynaecologists
(37–42 weeks); no previous caesarean section; intact membranes; not 35. Also, women with any clinical suspicion of having a fetus that was small for gestational age were required to have a normal ultrasound before being eligible. At the conclusion of the study, an analysis of clinical outcomes before and after these changes showed no significant differences, but after the changes, the recruitment rate was reduced by 25%. The two groups had comparable baseline characteristics (Table 1). The majority of women were nulliparous, from an English-speaking background, and were being induced because of prolonged pregnancy or social reasons. Median gestation at enrolment was 40 weeks plus 1 day and at cervical ripening 40 weeks plus 6 days.
Clinical course Of the 215 women in the outpatient group who received ripening, 47 (22%) were admitted as inpatients, mostly due to non-reassuring EFM or because they wished to remain
96
as inpatients (Figure 1). For the remaining 168 women who went home overnight, 64 (38%) returned to hospital before 7 am and were admitted due to contractions (n = 57), rupture of membranes (n = 6) or anxiety (n = 1). There were 12 documented cases of hyperstimulation following PGE2 insertion, defined as >five contractions per 10 minutes, with all but one case having fetal heart rate changes. Seven of the cases occurred in the outpatient group and five in the inpatient group (Table 2). All events occurred in hospital during cervical ripening and post PGE2 insertion monitoring. In four of the 12 cases, an immediate caesarean was required. Non-reassuring EFM and immediate reactions to gels were more frequently reported in the outpatient group than the inpatient group (Table S1). There were 10 cases of failed ripening (two in the outpatient group and eight in the inpatient group), whereby the woman did not establish in labour after two or more consecutive doses of PGE2 gel. Nine of the ten cases required a caesarean section for failed ripening (the tenth case was rescheduled for another day).
Maternal and neonatal outcomes There were no statistically significant differences in the two groups for any maternal and neonatal outcomes (Tables 2 and 3). Almost 50% of women received oxytocin infusion. Neonatal outcomes were also similar between the two groups, with no statistically significant differences in Apgar scores, length of stay, admission to special care, neonatal intensive care or neonatal hypoxic-ischaemic encephalopathy (HIE) (Table 3). The post-hoc analysis of women who actually received the allocated intervention (Tables 2 and 3) indicated no differences in the pattern or statistical significance in the results. The only exception to this was in relation to active length of labour, with women allocated to the outpatient group who received ripening experiencing longer labour (mean difference 66 minutes, 95%CI 4–128 minutes). One outpatient woman who was discharged home delivered within 10 minutes of arrival in the emergency department, approximately 6 hours after receiving prostaglandin ripening. There were two unintended home births (one in each randomised group); however, these women had not received prostaglandin ripening.
Serious morbidities and incidents There were six reported cases of HIE (hypoxaemic ischaemic encephalopathy). This occurred in the infants of three women in the outpatient group who received the intervention (n = 215) and in two women in the inpatient group who received the intervention (n = 210) (Table S2). Although all of these were classified as grade 1 or 2 HIE,
ª 2014 Royal College of Obstetricians and Gynaecologists
RCT of PGE2 outpatient ripening for labour induction
Figure 1. Flow diagram of participants through the trial in accordance with the CONSORT requirements of reporting randomised controlled trials.
ª 2014 Royal College of Obstetricians and Gynaecologists
97
Wilkinson et al.
Table 1. Baseline characteristics Characteristic
Women randomised (ITT) n = 823 Inpatient n = 415 (%)
Parity Nulliparous Parous Mean age (SD) Current smokers Marital status Married/de facto Single Language spoken at home English only Other language Education University degree Post-high school training High school only Employment Full-time Part-time/casual Not employed Median drive to hospital Reason for induction Prolonged pregnancy To avoid prolonged pregnancy Social Other Gestation at randomisation Median weeks+ days (IQR)
303 112 28.7 42
(73.0) (27.0) (5.3) (10.1)
Outpatient n = 408 (%)
295 113 28.8 43
Inpatient ripening n = 210 (%)
152 58 28.8 22
(72.3) (27.6) (5.05) (10.4)
Outpatient ripening n = 215 (%)
157 58 28.9 24
(73.0) (27.0) (4.96) (11.1)
379 (91.3) 34 (8.2)
371 (90.9) 34 (8.3)
193 (91.9) 16 (7.6)
194 (90.2) 19 (8.8)
343 (82.7) 72 (17.3)
324 (79.4) 84 (20.6)
180 (85.7) 30 (14.3)
171 (79.5) 44 (20.5)
137 (33.0) 121 (29.2) 156 (37.6)
157 (38.5) 98 (24.0) 148 (36.2)
71 (33.8) 54 (25.7) 84 (40.0)
83 (38.6) 51 (23.7) 78 (36.2)
130 102 174 9
(31.3) (24.6) (41.9) km
149 83 164 10
(36.5) (20.3) (40.2) km
73 56 76 9
(34.8) (26.7) (36.2) km
73 53 82 10
(33.9) (24.6) (38.1) km
349 46 11 9
(84.1) (11.1) (2.7) (2.1)
342 37 19 10
(83.8) (9.0) (4.7) (2.5)
174 23 6 7
(82.9) (10.9) (2.8) (3.3)
170 23 14 8
(79.1) (10.7) (6.5) (3.7)
40+1 (39+6 40+4)
40+1 (39+6 40+4)
the subsequent case HIE rate by intention-to-treat of six in 821 women (or seven per 1000) is more than the expected rate of HIE of one to three per 1000 births that should be anticipated in technologically advanced countries.17 This was of concern in that the recruitment criteria in this trial included only low risk women at term. None of these infants showed any discernible disability from the casenote review of the paediatric follow up, so it may be HIE was overdiagnosed in this study, although recovery from grade 1 or grade 2 HIE is often quite favourable. The standard of perinatal care in South Australia is comparable to that in other western communities (WHO defined perinatal mortality rate in South Australia in 2009 was 3.5 per 1000 births.4) Only one case of HIE in the outpatient group was determined to be plausibly related to the intervention by the data monitoring committee and this resulted in the trial being temporarily stopped while the case was reviewed. The
98
(72.3) (27.7) (5.1) (10.5)
Randomised and received PGE2 gel n = 425
40+2 (40+0 40+5)
40+2 (39+6 40+4)
baby had undiagnosed growth restriction, which should have excluded his mother from the study if detected antenatally. The baby was delivered in a compromised state after a precipitate delivery soon after returning to the hospital, and was diagnosed with Grade 2 HIE (although this was reported by the neonatal long-term follow-up team as having normal development at 2 years of age). Eligibility criteria were subsequently tightened further as previously described, making it less likely that mothers of undiagnosed growth-restricted fetuses would be recruited for the remainder of the trial. There were two maternal admissions to intensive care, both in the outpatient group (one postpartum haemorrhage, one postpartum eclampsia) but neither was judged by the adverse outcomes committee to be plausibly related to the intervention. Three undiagnosed breech-presenting fetuses were also inadvertently included in the trial. All had good outcomes.
ª 2014 Royal College of Obstetricians and Gynaecologists
ª 2014 Royal College of Obstetricians and Gynaecologists
Total mean length of labour (SD minutes)
3rd stage (SD minutes)
2nd stage (SD minutes)
Indications for caesarean section Fetal distress Lack of progress Other (malpresentation etc.) Mean length of active labour, vaginal births 1st stage hours & minute (SD minutes) 7 hours 5 minutes (265 minutes)
6 hours 40 minutes (287 minutes) 1 hour 19 minutes (75 minutes) 13 minutes (17) 8 hours 12 minutes (324 minutes)
37 (40.7) 37 (40.7) 17 (18.6) (n = 316)
48 (50.5) 30 (31.6) 17 (17.9) (n = 319)
8 hours 35 minutes (299 minutes)
12 minutes (14 minutes)
1 hour 18 minutes (67)
(n = 91)
(n = 95)
6 hours 47 minutes (258)
7 hours 6 minutes (266) 231 (56.8) 85 (20.9) 91 (22.3)
86 (42.4) 117 (57.6) (n = 203)
85 (43.4) 111 (56.6) (n = 196)
227 (54.8) 92 (22.2) 95 (22.9)
(n = 203)
(n = 196)
Reason for oxytocin infusion Induction of labour Augmentation Duration of oxytocin infusion Mean (SD minutes)
Method of delivery Spontaneous vaginal Instrumental Caesarean section
192 (47.2) 203 (49.9)
Outpatient n = 407 (%)
204 (49.2) 196 (47.3)
Inpatient n = 414 (%)
Women randomised (intention to treat)
Ripening not required Oxytocin infusion
Variable
Table 2. Labour and delivery outcomes.
–
0.59% ( 6.3 to 5.1)
–
– 2.5% ( 4.3 to 9.4)
Risk Diff. (95%CI)
23 minutes ( 25 to 72)
0.97 (0.76 to 1.25)
( 33 to 70)
19 minutes
1.05 (0.91 to 1.20)
Relative risk (95%CI)
5 hours 18 minutes (239 minutes) 1 hours 14 minutes (75 minutes) 13 minutes (18 minutes) 6 hours 45 minutes (278 minutes)
25 (44.6%) 16 (28.6%) 15 (26.8%) (n = 154)
(n = 56)
108 (51.4%) 46 (21.9%) 56 (26.6%)
7 hours 0 minutes
59 (49.1) 61 (50.8) (n = 120)
(n = 120)
– 120 (57.1)
Inpatient ripening n = 210 (%)
6 hours20 minutes (246 minutes) 1 hour 20 minutes (69 minutes) 11 minutes (15 minutes) 7 hours 51 minutes (286 minutes)
19 (37.2%) 26 (51.0%) 6 (11.8%) (n = 164)
(n = 51)
115 (53.5%) 49 (22.8%) 51 (23.7%)
6 hours 57 minutes
53 (41.1) 76 (58.9) (n = 128)
(n = 129)
– 129 (60.0)*
Outpatient ripening n = 215 (%)
Risk Diff. (95%CI) Rel risk (95%CI)
66 minutes (4–128)
2.9% ( 11.2 to 5.3) 0.89 (0.64–1.24)
3 minutes ( 63 to 70)
2.9% ( 6.7 to 12.6) 1.06 (0.89 to 1.29)
Women received prostaglandin priming
RCT of PGE2 outpatient ripening for labour induction
99
100 (n = 215) 2 (0.9%) 3.3 (1.6)
90 (22.1) 53 (16.8) 153 (37.6) 62 (15.2) 7 (1.7)
85 (20.5) 49 (15.4) 146 (35.3) 57 (13.8) 6 (1.4)
(n = 210) 8 (3.8%) 3.3 (1.5)
155 252
42 (10.3) 256 (62.9)
(n = 215 ripening) 89 (41.4)
Outpatient n = 407 (%)
105 261
40 (9.7) 254 (61.4)
(n = 210 ripening) 104 (49.5)
Inpatient n = 414 (%)
Women randomised (intention to treat)
2.9% ( 6.5 to 0.16)
0.3% ( 1.4 to 1.9)
1.5% ( 5.1 to 8.2)
8.2% ( 17.6 to 1.3)
Risk Diff. (95%CI)
PPH, postpartum haemorrhage; SD, standard deviation. *Includes one case of infusion for 3rd stage PPH. **Pain relief methods other than epidural or spinal may be used more than once per woman. ***Women may have more than one labour complication. ****Fetal distress includes any non–reassuring episode of EFM over the course of labour.
Length of stay mean days (SD)
Failed Ripenings
Pethidine or Fentanyl N2O and/or other Labour complications*** Meconium–stained liquor PPH >500 ml (vaginal births) Any fetal distress**** Pyrexia during labour Hyperstimulation
Labour analgesia** None Epidural or spinal
Vaginal delivery within 24 hours of ripening
Variable
Table 2. (Continued)
–
1.19 (0.40 to 3.50) 0.91 (0.51 to 1.65) 1.64 (1.19–2.28)
1.03% (0.92–1.14)
0.84 (0.68–1.03)
Relative risk (95%CI)
(n = 210) 8 (3.8%) 3.3 (1.5)
35 (16.6) 27 (12.8) 83 (39.5) 32 (15.2) 5 (2.4)
59 132
11 (5.2) 139 (66.2)
(n = 210 ripening) 104 (49.5)
Inpatient ripening n = 210 (%)
(n = 215) 2 (0.9%) 3.3 (1.7)
41 (19.0) 29 (13.5) 88 (40.9) 39 (18.1 7 (3.2)
98 139
19 (8.8) 141 (65.6)
(n = 215 ripening) 89 (41.4)
Outpatient ripening n = 215 (%)
Risk Diff. (95%CI) Rel risk (95%CI)
0.8% ( 2.2 to 4.0) 1.37 (0.44 to 4.24) 2.9% ( 6.5 to 0.16) 1.64 (1.19–2.28)
0.99 (0.86–1.14)
0.6% ( 9.6 to 8.4)
8.2% ( 17.6 to 1.3) 0.84 (0.68–1.03)
Women received prostaglandin priming
Wilkinson et al.
ª 2014 Royal College of Obstetricians and Gynaecologists
ª 2014 Royal College of Obstetricians and Gynaecologists
Relative risk (95%CI)
38 20 16 14 58
18 11 79
1 99 (24.3)
49 12
108 (26.1)
3 (0.74)
3 (0.72) –
3602 (426) 8 (2.0) 3 (0.74)
224 (55.0) 183 (45.0) 11 (2.7)
3628 (429) 10 (2.4) 3 (0.72)
218 (52.7) 196 (47.3) 9 (2.2)
1.8% ( 7.7 to 4.2)
0.5% ( 1.6 to 2.6)
0.93 (0.74–1.18)
–
9 5 53
24 5
57 (27.1)
2 (0.95)
– –
3621 (427) 7 (3.3) 1 (0.48)
1.24 (0.52 to 2.97)
108 (51.4) 102 (48.6) 5 (2.4)
– – –
–
–
10 7 40
22 10
66 (30.7)
3 (1.4)
3597 (429) 6 (2.8) 2 (0.93)
117 (54.4) 98 (45.6) 7 (3.2)
Outpatient Ripening n = =215 (%)
Inpatient Ripening n = 210 (%)
Risk Diff. (95%CI)
Inpatient n = 414 (%)
Outpatient n = 407 (%)
Women received prostaglandin priming
Women randomised (intention to treat)
3.6% ( 5.0 to 12.1) 1.13 (0.84–1.53)
0.8% ( 2.3 to 4.0) 1.37 (0.44–4.24)
Risk Diff. (95%CI) Relative risk (95%CI)
*Case involved a woman who did not require ripening and laboured spontaneously. **Special care nursery admissions include: infants boarded for maternal care, caesarean deliveries, and temperature regulation. Infants may have more than one condition.
Birth weight mean g, (SD) Congenital anomalies Admission to neonatal intensive care Hypoxic ischemic encephalopathy Perinatal death* Special care nursery admissions** Feeding problems Jaundice requiring phototherapy Respiratory problems Infection Other
Male Female Apgar
General obstetrics
www.bjog.org
A randomised controlled trial of outpatient compared with inpatient cervical ripening with prostaglandin E2 (OPRA study) C Wilkinson,a R Bryce,b,c P Adelson,d D Turnbulld a
Maternal-Fetal Medicine, Women’s and Children’s Hospital, Adelaide, b Obstetrics and Gynecology, Flinders Medical Center, Flinders University, Bedford Park, d School of Psychology, University of Adelaide, Adelaide, SA, Australia Correspondence: C Wilkinson, Department of Obstetrics and Gynecology, Women’s and Children’s Hospital, 72 King William Road, North Adelaide, SA 5006, Australia. Email [email protected] c
Accepted 3 March 2014. Published Online 14 May 2014.
Objective To compare clinical outcomes from outpatient with
inpatient cervical prostaglandin E2 ripening for low risk labour induction. Design Randomised controlled trial. Setting Two tertiary hospitals in Adelaide, Australia. Population Women with uncomplicated term pregnancies,
scheduled for induction of labour for reasons of post-dates or for social reasons. Methods Eight hundred and twenty-seven women were
randomised at induction decision to an outpatient or inpatient group. All women had CTG monitoring before and after vaginal PgE2 administration. The inpatient women were admitted and the outpatient group was discharged home overnight with instructions to return if labour established or rupture of membranes occurred. Next morning, outpatient women were admitted for rupture of membranes or further ripening. After membrane rupture, both groups received routine care. Main outcome measures Oxytocin use, maternal and fetal
outcomes, and whether planned outpatient management was achievable. Results There were no significant differences in oxytocin use (2.5% difference, CI 4.3 to 9.4), caesarean sections ( 0.59% difference, CI 6.3 to 5.1), epidural use (1.5% difference, CI 5.1 to 8.2), vaginal delivery within 24 hours ( 8.2% difference,
CI 17.6 to 1.3) or labour complications. More than half of the randomised women did not receive the intervention as they laboured spontaneously, or did not require ripening. The post-hoc analysis of women who received ripening also indicated no statistically significant differences in the patterns or results, with the exception of outpatient women having a longer mean length of active labour (mean difference 66 minutes, CI 4–128 minutes). Outpatient women who received ripening were diagnosed more frequently with non-reassuring CTG monitoring and hyperstimulation, with less than half of the women going home and remaining home overnight. Conclusions This study demonstrated no clinical advantage or
disadvantage in prostaglandin E2 outpatient cervical ripening. Uterine stimulation following prostaglandins may preclude a woman from going home or remaining at home overnight and may not be the best agent for outpatient ripening. Keywords Cervical ripening, induction of labour, outpatient ripening, prostaglandin E2. Linked article This article is commented on by WA Grobman, p. 105 in this issue. To view this mini commentary visit http://dx.doi.org/10.1111/ 1471-0528.12849. The article has journal club questions by EYL Leung, p. 106 in this issue. To view these visit http://dx.doi.org/10.1111/ 1471-0528.12847.
Please cite this paper as: Wilkinson C, Bryce R, Adelson P, Turnbull D. A randomised controlled trial of outpatient compared with inpatient cervical ripening with prostaglandin E2 (OPRA study). BJOG 2015;122:94–106.
Introduction Cervical ripening for induction of labour is one of the commonest interventions in obstetrics.1–3 In 2010, 30% of South Australian pregnant women had labour induced. Of those inductions, 58% required cervical ripening with prostaglandins.4
94
Ripening a cervix which is “unfavourable” for induction is done with prostaglandin analogues as the most common pharmaceutical method. The South Australian guidelines recommend vaginal prostaglandin E2 (PGE2) gel with hospital admission for pharmaceutical cervical ripening,5 but the necessity of hospital admission for cervical ripening, although demonstrating clinical caution, is unproven.
ª 2014 Royal College of Obstetricians and Gynaecologists
RCT of PGE2 outpatient ripening for labour induction
Outpatient cervical priming offers possible psychosocial benefits for a mother in a more familiar home environment, perhaps facilitating a more natural establishment of labour.6,7 Cost savings may also be possible.8,9 Based on this reasoning, the 2008 NICE clinical guideline recommends, “Studies are needed to assess the safety, efficiency and clinical and cost effectiveness of outpatient inductions. . ..taking into account women’s views”.10 Also supporting further studies, two recent Cochrane reviews of outpatient induction conclude that there is insufficient evidence to establish which induction techniques are most effective and safe to use in outpatient settings.3,11 Of the studies in these reviews, few examine different methods of induction directly in home and hospital settings. The primary study objective of this randomised controlled trial (acronym OPRA) was to ascertain differences in clinical outcomes associated with allowing women to go home after priming, rather than staying in hospital. The biological plausibility of clinical differences was that women who were better rested after being in their own environments, rather than having experienced a sleepless night in hospital, would be better able to cope with the stresses of labour, have different thresholds of intervention, and might labour better if they were less physically stressed. Consistent with this hypothesis, a number of previously reported smaller controlled trials using various forms of prostaglandins suggested that there was less need for oxytocin use in women allowed to go home after ripening, compared with inpatient use.6,24,26 Other clinical outcomes that we hypothesised may have been associated with this model of care and less oxytocin use, were a reduced length of labour and analgesic use, an increase in normal births, comparable perinatal and neonatal outcomes, and improved psychosocial outcomes for the mother with regard to satisfaction with care, postnatal depression and mode of infant feeding at 7 weeks. In addition, the study assessed the impact of outpatient priming on midwives and examined the cost of outpatient priming, and we performed a discrete choice experiment. The psychosocial,12 costing,9 midwives study13 and discrete choice14 results are presented elsewhere.
Methods The trial was conducted from August 2008 to May 2011 at two tertiary referral hospitals, accounting for 39% of South Australia’s approximately 20 000 annual births.4 Women with uncomplicated term pregnancies were recruited from obstetric and midwifery clinics by a research midwife at the visit when clinicians scheduled an induction. South Australian guidelines recommend induction of labour to prevent the pregnancy exceeding 41 weeks’ gestation.5 Eligibility for the trial included; ≥18 years old; communication in English; singleton, cephalic, term pregnancy
ª 2014 Royal College of Obstetricians and Gynaecologists
(37–42 weeks); no previous caesarean section; intact membranes; not 35. Also, women with any clinical suspicion of having a fetus that was small for gestational age were required to have a normal ultrasound before being eligible. At the conclusion of the study, an analysis of clinical outcomes before and after these changes showed no significant differences, but after the changes, the recruitment rate was reduced by 25%. The two groups had comparable baseline characteristics (Table 1). The majority of women were nulliparous, from an English-speaking background, and were being induced because of prolonged pregnancy or social reasons. Median gestation at enrolment was 40 weeks plus 1 day and at cervical ripening 40 weeks plus 6 days.
Clinical course Of the 215 women in the outpatient group who received ripening, 47 (22%) were admitted as inpatients, mostly due to non-reassuring EFM or because they wished to remain
96
as inpatients (Figure 1). For the remaining 168 women who went home overnight, 64 (38%) returned to hospital before 7 am and were admitted due to contractions (n = 57), rupture of membranes (n = 6) or anxiety (n = 1). There were 12 documented cases of hyperstimulation following PGE2 insertion, defined as >five contractions per 10 minutes, with all but one case having fetal heart rate changes. Seven of the cases occurred in the outpatient group and five in the inpatient group (Table 2). All events occurred in hospital during cervical ripening and post PGE2 insertion monitoring. In four of the 12 cases, an immediate caesarean was required. Non-reassuring EFM and immediate reactions to gels were more frequently reported in the outpatient group than the inpatient group (Table S1). There were 10 cases of failed ripening (two in the outpatient group and eight in the inpatient group), whereby the woman did not establish in labour after two or more consecutive doses of PGE2 gel. Nine of the ten cases required a caesarean section for failed ripening (the tenth case was rescheduled for another day).
Maternal and neonatal outcomes There were no statistically significant differences in the two groups for any maternal and neonatal outcomes (Tables 2 and 3). Almost 50% of women received oxytocin infusion. Neonatal outcomes were also similar between the two groups, with no statistically significant differences in Apgar scores, length of stay, admission to special care, neonatal intensive care or neonatal hypoxic-ischaemic encephalopathy (HIE) (Table 3). The post-hoc analysis of women who actually received the allocated intervention (Tables 2 and 3) indicated no differences in the pattern or statistical significance in the results. The only exception to this was in relation to active length of labour, with women allocated to the outpatient group who received ripening experiencing longer labour (mean difference 66 minutes, 95%CI 4–128 minutes). One outpatient woman who was discharged home delivered within 10 minutes of arrival in the emergency department, approximately 6 hours after receiving prostaglandin ripening. There were two unintended home births (one in each randomised group); however, these women had not received prostaglandin ripening.
Serious morbidities and incidents There were six reported cases of HIE (hypoxaemic ischaemic encephalopathy). This occurred in the infants of three women in the outpatient group who received the intervention (n = 215) and in two women in the inpatient group who received the intervention (n = 210) (Table S2). Although all of these were classified as grade 1 or 2 HIE,
ª 2014 Royal College of Obstetricians and Gynaecologists
RCT of PGE2 outpatient ripening for labour induction
Figure 1. Flow diagram of participants through the trial in accordance with the CONSORT requirements of reporting randomised controlled trials.
ª 2014 Royal College of Obstetricians and Gynaecologists
97
Wilkinson et al.
Table 1. Baseline characteristics Characteristic
Women randomised (ITT) n = 823 Inpatient n = 415 (%)
Parity Nulliparous Parous Mean age (SD) Current smokers Marital status Married/de facto Single Language spoken at home English only Other language Education University degree Post-high school training High school only Employment Full-time Part-time/casual Not employed Median drive to hospital Reason for induction Prolonged pregnancy To avoid prolonged pregnancy Social Other Gestation at randomisation Median weeks+ days (IQR)
303 112 28.7 42
(73.0) (27.0) (5.3) (10.1)
Outpatient n = 408 (%)
295 113 28.8 43
Inpatient ripening n = 210 (%)
152 58 28.8 22
(72.3) (27.6) (5.05) (10.4)
Outpatient ripening n = 215 (%)
157 58 28.9 24
(73.0) (27.0) (4.96) (11.1)
379 (91.3) 34 (8.2)
371 (90.9) 34 (8.3)
193 (91.9) 16 (7.6)
194 (90.2) 19 (8.8)
343 (82.7) 72 (17.3)
324 (79.4) 84 (20.6)
180 (85.7) 30 (14.3)
171 (79.5) 44 (20.5)
137 (33.0) 121 (29.2) 156 (37.6)
157 (38.5) 98 (24.0) 148 (36.2)
71 (33.8) 54 (25.7) 84 (40.0)
83 (38.6) 51 (23.7) 78 (36.2)
130 102 174 9
(31.3) (24.6) (41.9) km
149 83 164 10
(36.5) (20.3) (40.2) km
73 56 76 9
(34.8) (26.7) (36.2) km
73 53 82 10
(33.9) (24.6) (38.1) km
349 46 11 9
(84.1) (11.1) (2.7) (2.1)
342 37 19 10
(83.8) (9.0) (4.7) (2.5)
174 23 6 7
(82.9) (10.9) (2.8) (3.3)
170 23 14 8
(79.1) (10.7) (6.5) (3.7)
40+1 (39+6 40+4)
40+1 (39+6 40+4)
the subsequent case HIE rate by intention-to-treat of six in 821 women (or seven per 1000) is more than the expected rate of HIE of one to three per 1000 births that should be anticipated in technologically advanced countries.17 This was of concern in that the recruitment criteria in this trial included only low risk women at term. None of these infants showed any discernible disability from the casenote review of the paediatric follow up, so it may be HIE was overdiagnosed in this study, although recovery from grade 1 or grade 2 HIE is often quite favourable. The standard of perinatal care in South Australia is comparable to that in other western communities (WHO defined perinatal mortality rate in South Australia in 2009 was 3.5 per 1000 births.4) Only one case of HIE in the outpatient group was determined to be plausibly related to the intervention by the data monitoring committee and this resulted in the trial being temporarily stopped while the case was reviewed. The
98
(72.3) (27.7) (5.1) (10.5)
Randomised and received PGE2 gel n = 425
40+2 (40+0 40+5)
40+2 (39+6 40+4)
baby had undiagnosed growth restriction, which should have excluded his mother from the study if detected antenatally. The baby was delivered in a compromised state after a precipitate delivery soon after returning to the hospital, and was diagnosed with Grade 2 HIE (although this was reported by the neonatal long-term follow-up team as having normal development at 2 years of age). Eligibility criteria were subsequently tightened further as previously described, making it less likely that mothers of undiagnosed growth-restricted fetuses would be recruited for the remainder of the trial. There were two maternal admissions to intensive care, both in the outpatient group (one postpartum haemorrhage, one postpartum eclampsia) but neither was judged by the adverse outcomes committee to be plausibly related to the intervention. Three undiagnosed breech-presenting fetuses were also inadvertently included in the trial. All had good outcomes.
ª 2014 Royal College of Obstetricians and Gynaecologists
ª 2014 Royal College of Obstetricians and Gynaecologists
Total mean length of labour (SD minutes)
3rd stage (SD minutes)
2nd stage (SD minutes)
Indications for caesarean section Fetal distress Lack of progress Other (malpresentation etc.) Mean length of active labour, vaginal births 1st stage hours & minute (SD minutes) 7 hours 5 minutes (265 minutes)
6 hours 40 minutes (287 minutes) 1 hour 19 minutes (75 minutes) 13 minutes (17) 8 hours 12 minutes (324 minutes)
37 (40.7) 37 (40.7) 17 (18.6) (n = 316)
48 (50.5) 30 (31.6) 17 (17.9) (n = 319)
8 hours 35 minutes (299 minutes)
12 minutes (14 minutes)
1 hour 18 minutes (67)
(n = 91)
(n = 95)
6 hours 47 minutes (258)
7 hours 6 minutes (266) 231 (56.8) 85 (20.9) 91 (22.3)
86 (42.4) 117 (57.6) (n = 203)
85 (43.4) 111 (56.6) (n = 196)
227 (54.8) 92 (22.2) 95 (22.9)
(n = 203)
(n = 196)
Reason for oxytocin infusion Induction of labour Augmentation Duration of oxytocin infusion Mean (SD minutes)
Method of delivery Spontaneous vaginal Instrumental Caesarean section
192 (47.2) 203 (49.9)
Outpatient n = 407 (%)
204 (49.2) 196 (47.3)
Inpatient n = 414 (%)
Women randomised (intention to treat)
Ripening not required Oxytocin infusion
Variable
Table 2. Labour and delivery outcomes.
–
0.59% ( 6.3 to 5.1)
–
– 2.5% ( 4.3 to 9.4)
Risk Diff. (95%CI)
23 minutes ( 25 to 72)
0.97 (0.76 to 1.25)
( 33 to 70)
19 minutes
1.05 (0.91 to 1.20)
Relative risk (95%CI)
5 hours 18 minutes (239 minutes) 1 hours 14 minutes (75 minutes) 13 minutes (18 minutes) 6 hours 45 minutes (278 minutes)
25 (44.6%) 16 (28.6%) 15 (26.8%) (n = 154)
(n = 56)
108 (51.4%) 46 (21.9%) 56 (26.6%)
7 hours 0 minutes
59 (49.1) 61 (50.8) (n = 120)
(n = 120)
– 120 (57.1)
Inpatient ripening n = 210 (%)
6 hours20 minutes (246 minutes) 1 hour 20 minutes (69 minutes) 11 minutes (15 minutes) 7 hours 51 minutes (286 minutes)
19 (37.2%) 26 (51.0%) 6 (11.8%) (n = 164)
(n = 51)
115 (53.5%) 49 (22.8%) 51 (23.7%)
6 hours 57 minutes
53 (41.1) 76 (58.9) (n = 128)
(n = 129)
– 129 (60.0)*
Outpatient ripening n = 215 (%)
Risk Diff. (95%CI) Rel risk (95%CI)
66 minutes (4–128)
2.9% ( 11.2 to 5.3) 0.89 (0.64–1.24)
3 minutes ( 63 to 70)
2.9% ( 6.7 to 12.6) 1.06 (0.89 to 1.29)
Women received prostaglandin priming
RCT of PGE2 outpatient ripening for labour induction
99
100 (n = 215) 2 (0.9%) 3.3 (1.6)
90 (22.1) 53 (16.8) 153 (37.6) 62 (15.2) 7 (1.7)
85 (20.5) 49 (15.4) 146 (35.3) 57 (13.8) 6 (1.4)
(n = 210) 8 (3.8%) 3.3 (1.5)
155 252
42 (10.3) 256 (62.9)
(n = 215 ripening) 89 (41.4)
Outpatient n = 407 (%)
105 261
40 (9.7) 254 (61.4)
(n = 210 ripening) 104 (49.5)
Inpatient n = 414 (%)
Women randomised (intention to treat)
2.9% ( 6.5 to 0.16)
0.3% ( 1.4 to 1.9)
1.5% ( 5.1 to 8.2)
8.2% ( 17.6 to 1.3)
Risk Diff. (95%CI)
PPH, postpartum haemorrhage; SD, standard deviation. *Includes one case of infusion for 3rd stage PPH. **Pain relief methods other than epidural or spinal may be used more than once per woman. ***Women may have more than one labour complication. ****Fetal distress includes any non–reassuring episode of EFM over the course of labour.
Length of stay mean days (SD)
Failed Ripenings
Pethidine or Fentanyl N2O and/or other Labour complications*** Meconium–stained liquor PPH >500 ml (vaginal births) Any fetal distress**** Pyrexia during labour Hyperstimulation
Labour analgesia** None Epidural or spinal
Vaginal delivery within 24 hours of ripening
Variable
Table 2. (Continued)
–
1.19 (0.40 to 3.50) 0.91 (0.51 to 1.65) 1.64 (1.19–2.28)
1.03% (0.92–1.14)
0.84 (0.68–1.03)
Relative risk (95%CI)
(n = 210) 8 (3.8%) 3.3 (1.5)
35 (16.6) 27 (12.8) 83 (39.5) 32 (15.2) 5 (2.4)
59 132
11 (5.2) 139 (66.2)
(n = 210 ripening) 104 (49.5)
Inpatient ripening n = 210 (%)
(n = 215) 2 (0.9%) 3.3 (1.7)
41 (19.0) 29 (13.5) 88 (40.9) 39 (18.1 7 (3.2)
98 139
19 (8.8) 141 (65.6)
(n = 215 ripening) 89 (41.4)
Outpatient ripening n = 215 (%)
Risk Diff. (95%CI) Rel risk (95%CI)
0.8% ( 2.2 to 4.0) 1.37 (0.44 to 4.24) 2.9% ( 6.5 to 0.16) 1.64 (1.19–2.28)
0.99 (0.86–1.14)
0.6% ( 9.6 to 8.4)
8.2% ( 17.6 to 1.3) 0.84 (0.68–1.03)
Women received prostaglandin priming
Wilkinson et al.
ª 2014 Royal College of Obstetricians and Gynaecologists
ª 2014 Royal College of Obstetricians and Gynaecologists
Relative risk (95%CI)
38 20 16 14 58
18 11 79
1 99 (24.3)
49 12
108 (26.1)
3 (0.74)
3 (0.72) –
3602 (426) 8 (2.0) 3 (0.74)
224 (55.0) 183 (45.0) 11 (2.7)
3628 (429) 10 (2.4) 3 (0.72)
218 (52.7) 196 (47.3) 9 (2.2)
1.8% ( 7.7 to 4.2)
0.5% ( 1.6 to 2.6)
0.93 (0.74–1.18)
–
9 5 53
24 5
57 (27.1)
2 (0.95)
– –
3621 (427) 7 (3.3) 1 (0.48)
1.24 (0.52 to 2.97)
108 (51.4) 102 (48.6) 5 (2.4)
– – –
–
–
10 7 40
22 10
66 (30.7)
3 (1.4)
3597 (429) 6 (2.8) 2 (0.93)
117 (54.4) 98 (45.6) 7 (3.2)
Outpatient Ripening n = =215 (%)
Inpatient Ripening n = 210 (%)
Risk Diff. (95%CI)
Inpatient n = 414 (%)
Outpatient n = 407 (%)
Women received prostaglandin priming
Women randomised (intention to treat)
3.6% ( 5.0 to 12.1) 1.13 (0.84–1.53)
0.8% ( 2.3 to 4.0) 1.37 (0.44–4.24)
Risk Diff. (95%CI) Relative risk (95%CI)
*Case involved a woman who did not require ripening and laboured spontaneously. **Special care nursery admissions include: infants boarded for maternal care, caesarean deliveries, and temperature regulation. Infants may have more than one condition.
Birth weight mean g, (SD) Congenital anomalies Admission to neonatal intensive care Hypoxic ischemic encephalopathy Perinatal death* Special care nursery admissions** Feeding problems Jaundice requiring phototherapy Respiratory problems Infection Other
Male Female Apgar
