DOI: 10.1111/j.1471-0528.2008.01890.x www.blackwellpublishing.

com/bjog

General obstetrics

Titrated low-dose vaginal and/or oral misoprostol to induce labour for prelabour membrane rupture: a randomised trial
L Bricker,a H Peden,a AJ Tomlinson,b TK Al-Hussaini,c T Idama,d C Candelier,e M Luckas,f H Furniss,g A Davies,h B Kumar,i J Roberts,j Z Alrevick
a Obstetric Directorate, Liverpool Womens NHS Foundation Trust, Liverpool, UK b Department of Obstetrics and Gynaecology, Royal Bolton Hospital, Bolton, UK c Department of Obstetrics and Gynaecology, Assiut University Hospital, Assiut, Egypt d Department of Obstetrics and Gynaecology, Whiston Hospital, St Helens, UK e Department of Obstetrics and Gynaecology, Stepping Hill Hospital, Stockport, UK f Department of Obstetrics and Gynaecology, Leighton Hospital, Crewe, UK g Department of Obstetrics and Gynaecology, Warrington Hospital, Warrington, UK h Department of Obstetrics and Gynaecology, Royal Preston Hospital, Preston, UK i Department of Obstetrics and Gynaecology, Ysbyty Gwyned Hospital, Bangor, UK j Department of Obstetrics, Queen Mothers Hospital, Glasgow, UK k Division of Perinatal and Reproductive Medicine, University of Liverpool, Liverpool, UK Correspondence: Dr L Bricker, Liverpool Womens NHS Foundation Trust, Crown Street, Liverpool, L8 7SS, UK. Email leanne.bricker@lwh.nhs.uk

Accepted 13 July 2008. Published OnlineEarly 26 August 2008.

Objective To evaluate the clinical effectiveness and safety of

titrated low-dose misoprostol for induction of labour (IOL) in the presence of prelabour rupture of membranes (PROM).
Design Randomised controlled trial. Setting Maternity units in the UK (9) and Egypt (1). Population Women >34 weeks of gestation with PROM, singleton

viable fetus and no previous caesarean section.

Methods Subjects randomised to IOL with a titrated low-dose

misoprostol regimen (oral except if unfavourable cervix, where initial dose vaginal) or a standard induction method, namely vaginal dinoprostone followed by intravenous oxytocin if the cervix was unfavourable or intravenous oxytocin alone if the cervix was favourable.
Main outcome measures Primary outcome measures were caesarean section and failure to achieve vaginal delivery within 24 hours. Analysis was by intention to treat. Results The trial did not achieve the planned sample size of 1890 due to failure in obtaining external funding. Seven hundred and

fty-eight women were randomised (375 misoprostol and 383 standard). There were less caesarean section (14 versus 18%, relative risk [RR] 0.79; 95% CI 0.571.09) and less women who failed to achieve vaginal delivery within 24 hours in the misoprostol group (24 versus 31%, RR 0.79; 95% CI 0.631.00), but the differences were not statistically significant. Subgroup analysis showed that with unfavourable cervix, misoprostol may be more effective than vaginal dinoprostone. There was no difference in hyperstimulation syndrome. There were more maternal adverse effects with misoprostol, but no significant differences in maternal and neonatal complications.
Conclusions Titrated low-dose misoprostol may be a reasonable

alternative for IOL in the presence of PROM, particularly in women with an unfavourable cervix. Safety and rare serious adverse events could not be evaluated in a trial of this size.
Keywords Induction of labour, low-dose misoprostol, ruptured

membranes.

Please cite this paper as: Bricker L, Peden H, Tomlinson A, Al-Hussaini T, Idama T, Candelier C, Luckas M, Furniss H, Davies A, Kumar B, Roberts J, Alrevic Z. Titrated low-dose vaginal and/or oral misoprostol to induce labour for prelabour membrane rupture: a randomised trial. BJOG 2008;115:15031511.

Introduction
Prelabour rupture of membranes (PROM) occurs in approximately 8% of women at or near term.1 A large multicentre randomised trial undertaken to determine whether immediate induction of labour (IOL) was better than expectant man-

agement for PROM found similar rates of caesarean section and neonatal infection, but women viewed immediate induction more positively than expectant management.2 In recent years, misoprostol, a prostaglandin E1 analogue, which is relatively cheap, stable at room temperature and is rapidly absorbed both vaginally and orally, has emerged as an

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

1503

Bricker et al.

alternative induction agent to dinoprostone and oxytocin. It is not licensed for use in pregnancy in most countries, and this has been an issue of much debate.3 In the past 15 years, a large number of trials have been reported, which have assessed the efcacy and safety of misoprostol when used for IOL in the presence of a viable pregnancy. These trials have used both vaginal and oral misoprostol. The potential advantages of the oral route include easy, noninvasive administration and avoidance of unnecessary vaginal examinations, thus the oral route may be preferred by women and caregivers. Systematic reviews of trials of misoprostol have shown it to be more effective than other agents if used vaginally4 and as effective as other agents if used orally.5 However, denitive data on safety are still lacking. Overall, misoprostol seems to be associated with an increased incidence of uterine hyperstimulation, both compared with dinoprostone and oxytocin.4,5 Uterine hyperstimulation may predispose to more serious complications such as uterine rupture and associated maternal and neonatal morbidity and mortality. However, the incidence of adverse effects appears to be dose related, and it is generally accepted that low-dose regimens are safer, but their use may compromise clinical efcacy. Our previous trial comparing titrated low-dose oral misoprostol solution with vaginal prostaglandin E2 gel for IOL showed low-dose oral misoprostol solution to be effective for IOL with no obvious compromise in safety.6 Of particular interest was the fact that in the subgroup of women where the indication for IOL was PROM, the induction to delivery interval was reduced in the oral misoprostol group without obvious adverse effects. This led us to develop this study protocol, which aimed to compare a low-dose misoprostol regimen with the standard induction method undertaken in the presence of PROM, namely vaginal dinoprostone followed by intravenous oxytocin infusion if the cervix is unfavourable or intravenous oxytocin infusion alone if the cervix is favourable.

Methods
The trial took place in nine UK Maternity Hospitals and a Maternity Hospital in Egypt between February 2002 and November 2005 (45 months). The principal investigation site was Liverpool Womens Hospital, Liverpool, UK. Each site had a local investigating team (Appendix S1). Women with prelabour rupture of amniotic membranes with a clinical decision to induce labour were asked to participate. The inclusion criteria were: gestation >34 weeks, singleton fetus, cephalic presentation and normal admission cardiotocograph (CTG). Exclusion criteria were: multiple pregnancy, breech presentation and previous caesarean section. All eligible women were given a written information sheet and asked to sign a written consent form. The multicentre research ethics committee and local research ethics committees approved the study protocol.

Women were randomly allocated to receive either the misoprostol or the standard regimen. To allocate treatment, the next in a series of sealed consecutively numbered opaque envelopes was taken from one of four dispensers depending on parity (primiparous or multiparous) and cervical status (favourable or unfavourable by modied Bishops score). The envelopes were prepared remote from the study sites at the National Perinatal Epidemiology Unit (NPEU), Oxford, UK, using a computer-generated list of allocations. Allocations were balanced between groups using random block sizes (between four and ten) and stratied for parity and cervical status. Each envelope contained the appropriate management protocol (treatment regimen, fetal monitoring rules and management of potential adverse reactions). During the analysis, the allocation recorded on the data form was cross-checked with the randomisation schedule held at Oxford. The management protocol included in the randomisation envelope was followed unless clinical imperatives dictated otherwise. When labour commenced, electronic fetal heart rate monitoring was used. The study was not placebo controlled, therefore women and clinicians were aware of the treatment allocated. The treatment protocols were as follows. Misoprostol regimen: The smallest preparation of misoprostol available was 100 microgram tablet. To obtain doses of 25 micrograms and 50 micrograms, tablets were cut into quarters or halves using a tablet cutter. If the cervix was unfavourable (modied Bishops score 6), 25 micrograms of misoprostol was inserted vaginally deep into posterior vaginal fornix, and 4 hours later, the titrated oral misoprostol regimen was commenced. If the cervix was favourable (modied Bishops score > 6), the titrated oral misoprostol regimen was commenced with 2-hourly review. Vaginal examination was not mandatory and only performed if clinically indicated. The titrated oral misoprostol regimen was as follows: (i) if uterine contractions were adequate, no further action was taken; (ii) if uterine contractions were inadequate and/or progress was inadequate (cervical dilatation less than 1 cm per hour), 25 microgram misoprostol was given orally and (iii) if there were no uterine contractions or no progress (no further cervical dilatation), 50 microgram misoprostol was given orally. If the women had not delivered within 24 hours of commencement of trial, intravenous oxytocin was started. The maximum number of doses allowed if cervix unfavourable were 11 (25 micrograms vaginally initially and thereafter ten 25 or 50 micrograms subsequent doses 2 hourly) or if cervix favourable 12 doses (25 or 50 micrograms, 2 hourly). The maximum possible dose in a 24-hour period for the unfavourable cervix subgroup was 525 micrograms, and the favourable cervix subgroup was 600 micrograms. Standard regimen: For women with an unfavourable cervix (modied Bishops score 6), dinoprostone (3 mg tablet or 1 mg gel) was inserted into the posterior vaginal fornix and

1504

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

PROM IOL with low-dose misoprostol

repeated after 6 hours if labour was not established. A maximum of two doses were given. If labour was not established after 12 hours or contractions were inadequate, an oxytocin infusion was commenced as per local regimen. For women with a favourable cervix, an oxytocin infusion was commenced as per local regimen. Tachysystole (>5 contractions per 10 minutes) or hypersystole (a contraction lasting at least 2 minutes) were initially diagnosed by a combination of abdominal palpation and cardiotocography. These conditions were managed conservatively with the woman placed in left lateral position, reducing or stopping oxytocin (if it had been commenced), continuous fetal heart rate and contraction monitoring. If the forewaters were intact, articial rupture of the membranes was avoided. Hyperstimulation syndrome (tachysystole or hypersystole with fetal heart rate abnormality) was managed by changing to left lateral position and tocolysis was given if considered necessary with a ritodrine infusion (50 mg/50 ml at the rate of 20 ml/hour for 30 minutes) or terbutaline (0.25 mg subcutaneously). If there was no improvement after 30 minutes, membranes were ruptured articially (if forewaters intact) and fetal scalp blood sampling undertaken. The primary outcome measures were caesarean section as a measure of safety, and failure to achieve vaginal delivery in 24 hours as a measure of clinical effectiveness. Secondary outcome measures included intrapartum events, uterine activity, maternal adverse effects and maternal and neonatal complications. The sample size calculation was based on the two primary outcomes, namely caesarean section and vaginal delivery not achieved within 24 hours. We used the data from our previous trial6 of a low-dose oral misoprostol regimen for IOL where a subgroup of women with ruptured membranes (24 women) had a caesarean section rate of 20 and 79% had delivered within 24 hours. To show an absolute risk reduction of 5% (80% power, alpha 5%), 1890 women needed to be recruited (945 in each group).

Data were collected until mother and baby were discharged from hospital, entered on datasheets at the site of recruitment and sent to the principal site. Data were double entered by two separate individuals into an Access database at the principal site. The database was managed independently at NPEU, Oxford, and discrepancies were reported to the principal site where corrections were made after clarication sought at local site if necessary. Data were analysed using SPSS (v13.0, SPSS Inc., Chicago, IL, USA) and Revman 4.2. Analysis was by intention to treat. Differences between groups were expressed as relative risks with 95% CI. Prespecied subgroups for analysis for primary outcome measures were parity (primiparous or multiparous) and cervical status (unfavourable cervix or favourable cervix).

Results
The trial prole is shown in Figure 1 with a total of 758 women randomised. Data were not kept on women who declined, were not offered or were excluded from participation. Compliance with allocated treatment was good (98% overall). The main reason for noncompliance was that spontaneous labour ensued (eight cases); other reasons were breech presentation diagnosed after randomisation, therefore induction abandoned (two cases), women declined participation after randomisation (one case), no misoprostol in drug cupboard (two cases) and crossover (two cases). The trial did not achieve the planned sample size. Several attempts to obtain external funding were not successful. An Independent Data and Safety Monitoring Committee met once during the course of study and recommended that study should continue. At the second formal Trial Steering Group meeting in September 2005, it was concluded that the sample size was not realistically achievable particularly without funding. However, it was decided to continue for a further 3 months with a concerted recruitment drive, albeit with resource limitations. A further 44 women were recruited in

758 women recruited

375 misoprostol regimen 366 (97.8%) compliance with allocated treatment

383 standard regimen

376 (98.2%) compliance with allocated treatment

No losses to follow up
Figure 1. Trial prole.

No losses to follow up

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

1505

Bricker et al.

Table 1. Characteristics of women at trial entry Misoprostol Number randomised Age (years), median (range) Gestation (weeks), mean (SD) Parity, n (%) Primiparous Cervical status, n (%) Favourable (modied BS . 6) BS, Bishops score. 375 29 (1543) 38.9 (1.7) 214 (57) 154 (41) Standard 383 28 (1642) 38.8 (1.7) 227 (59) 149 (39)

only two of the recruiting hospitals, and a decision was made to stop recruitment. At no point was the steering group aware of the interim outcome results. Only recruitment gures were considered. The characteristics of trial entry are shown in Table 1. All demographic and other characteristics at trial entry were balanced between groups. In the unfavourable cervix group, 26% of women had only one dose of misoprostol (i.e. the initial vaginal dose), 23% needed two doses and 49% needed three or more doses (Figure 2). In this group of women, the median misoprostol dose used was 75 micrograms (range 0525 micrograms), and the majority of individual doses used were 25 micrograms (n = 576, 25 micrograms doses; n = 152, 50 micrograms doses). In the standard regimen group, 39% of women needed only one dose of dinoprostone with 59% requiring a second dose 6 hours later. In the favourable cervix group, 28% of women had only one dose of misoprostol, 37% needed two doses and 32% needed three or more doses (Figure 2). In this group of

women, the median dose used was 50 micrograms (range 0275 micrograms). The majority of individual doses used were 25 micrograms (n = 152, 25 micrograms doses; n = 51, 50 micrograms doses). The primary outcome measures are shown in Table 2. There were less caesarean sections and less failures to achieve vaginal delivery within 24 hours of induction in the misoprostol group, but the differences did not reach statistical signicance. The secondary outcome measures for labour and delivery are shown in Table 3. The use of misoprostol for IOL was associated with a statistically signicant reduction in the use of oxytocin augmentation. Data for this outcome were analysed for the unfavourable cervix subgroup only as controls in the favourable subgroup were all induced with oxytocin. Women in the misoprostol group used less epidural for analgesia and had less caesarean section for failure to progress. There was no difference in the clinically important outcomes of uterine hyperstimulation and hyperstimulation syndrome. Maternal adverse effects and complications are shown in Table 4. Overall, misoprostol was associated with more maternal adverse effects, but for individual adverse effects, there was no signicant difference. Misoprostol induction was abandoned only once due to unacceptable adverse effects of nausea and vomiting. The majority of maternal complications in both groups were due to postpartum haemorrhage, most of which were blood loss between 500 and 1000 ml. There were three serious adverse events (two of which resulted in maternal intensive therapy unit [ITU] admission) reported in the misoprostol group and none in the standard group. There were no cases of uterine rupture or maternal death. The three serious adverse events were as follows: (i) 1 hour and 15 minutes after the last dose of 50 microgram

Figure 2. Number of doses of misoprostol.

1506

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

PROM IOL with low-dose misoprostol

Table 2. Primary outcomes expressed as n (%) Misoprostol Standard Number randomised Caesarean section Vaginal delivery not achieved within 24 hours 375 54 (14) 91 (24) 383 70 (18) 117 (31) RR (95% CI)

0.79 (0.571.09) 0.79 (0.631.00)

misoprostol precipitate, labour ensued and there was severe postpartum haemorrhage (PPH) following normal vaginal delivery due to uterine atony, which did not respond to medical treatment and required peripartum hysterectomy and admission to ITU; (ii) two hours and 36 minutes after the last dose of misoprostol, severe PPH due to cervical tear following normal vaginal delivery, but extent of blood loss was due to delayed diagnosis of cervical tear; (iii) cardiac arrest under general anaesthetic during evacuation of retained products of conception soon after delivery, subsequently diagnosed as peripartum cardiomyopathy and required admission to ITU. Neonatal outcomes and complications are shown in Table 5. There were no differences in neonatal outcomes, and in general, the incidence of markers of neonatal morbidity was low. There were no neonatal deaths and only one neonatal complication (a case of neonatal sepsis) in the control group.

Prespecied subgroup analyses for primary outcome measures are shown in Table 6. There was no statistically signicant difference in caesarean section rates in the subgroup analysis; however, women with unfavourable cervix allocated to misoprostol were more likely to achieve vaginal delivery within 24 hours. Exploratory analysis was undertaken using KaplanMeier survival plots showing time from trial entry to delivery. Overall, there was no significant difference between the misoprostol and the standard regimen (Figure 3). However, in women with a favourable cervix, the standard regimen (oxytocin) was more effective than the misoprostol regimen (oral misoprostol) (Figure 4), while for women with an unfavourable cervix, the misoprostol regimen was more effective than a combination of dinoprostone and oxytocin (Figure 5).

Discussion
Although we found no signicant difference in primary outcomes (caesarean section and vaginal delivery not achieved in 24 hours), there was a trend towards better outcomes with titrated oral misoprostol. In particular, misoprostol was associated with less need for oxytocin augmentation, less epidural analgesia and less caesarean section for failure to progress. We believe that these results reect our targeted rather than one size ts all approach to IOL with PROM. For women with unfavourable cervices, a combination of a small dose of vaginal misoprostol (25 micrograms) followed by a titrated oral

Table 3. Secondary outcomes of labour and delivery expressed as n (%) or mean (SD) Misoprostol Number randomised Labour Oxytocin augmentation* Number of vaginal examinations, mean (SD) Epidural analgesia Opioid analgesia No analgesia Meconium Intrapartum vaginal bleeding Intrapartum fetal blood sampling Uterine hyperstimulation without FHR changes (tachysystole/hypersystole) Hyperstimulation syndrome Tocolysis for excessive uterine activity Delivery Instrumental vaginal delivery Caesarean section for fetal distress Caesarean section for failure to progress Randomisation to delivery interval (hours), mean (SD) Estimated blood loss at delivery (ml) FHR, fetal heart rate; WMD, weighted mean difference. *Data for unfavourable cervix subgroup only. 375 52/220 (24) 3.8 (2.3) 89 (24) 222 (59) 39 (10) 50 (13) 22 (6) 33 (9) 18 (5) 7 (2) 6 (2) 59 (16) 24 (6) 19 (5) 13.6 (8.2) 301 (220) Standard 383 155/235 (66) 3.9 (2.1) 118 (31) 232 (61) 30 (8) 36 (9) 20 (5) 41 (11) 13 (3) 6 (1) 4 (1) 59 (15) 21 (6) 36 (9) 14.7 (8.8) 306 (214) RR 0.36 (0.280.46) WMD 20.10 (20.41 to 0.21) RR 0.77 (0.610.97) RR 0.98 (0.871.10) RR 1.33 (0.842.09) RR 1.42 (0.952.12) RR 1.12 (0.622.02) RR 0.82 (0.531.27) RR 1.41 (0.702.84) RR 1.43 (0.464.47) RR 1.53 (0.445.39) RR 1.02 (0.731.42) RR 1.17 (0.662.06) RR 0.54 (0.310.92) WMD 21.1 (22.31 to 0.11) WMD 25 (236 to 26) RR or WMD (95% CI)

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

1507

Bricker et al.

Table 4. Maternal adverse effects and complications expressed as n (%) Misoprostol Number randomised Maternal adverse effects Maternal adverse effects (all)* Nausea Vomiting Diarrhoea Shivering Pyrexia . 38C during labour Maternal complications Maternal complications (all)** PPH PPH mild (.500 ml, ,1000 ml) PPH moderate (.1000 ml, ,1500 ml) PPH severe (. 1500 ml) Ruptured uterus Third-degree tear Retained placenta Postpartum pyrexia .38C Sepsis conrmed with blood culture Admission to ITU Maternal death *Some women had more than one adverse effect. **Some women had more than one complication. 375 71 (19) 35 (9) 41 (11) 4 (1) 19 (5) 8 (2) 71 (19) 57 (15) 49 (13) 5 (1) 3 (1) 0 5 (1) 10 (3) 4 (1) 0 2 (0.5) 0 Standard 383 51 (13) 30 (8) 30 (8) 0 15 (4) 6 (2) 83 (22) 67 (18) 62 (16) 3 (1) 2 (0.5) 0 3 (1) 12 (3) 4 (1) 0 0 0 1.42 (1.021.98) 1.19 (0.751.90) 1.37 (0.872.14) 9.19 (0.50170.13) 1.29 (0.672.50) 1.36 (0.483.89) 0.87 (0.631.20) 0.87 (0.661.16) 0.81 (0.571.14) 1.70 (0.417.07) 1.53 (0.269.12) 1.70 (0.417.07) 0.85 (0.371.95) 1.02 (0.264.05) 5.11 (0.25106.01) RR (95% CI)

dose stems from pharmacokinetic7,8 and effectiveness data.4,5 A signicant reduction in randomisation to delivery with this misoprostol combination suggests that vaginal misoprostol may be a better cervical ripening agent than vaginal dinoprostone. In the subgroup of women with a favourable cervix, intravenous oxytocin seems to be more effective than oral misoprostol in terms of time from randomisation to delivery. Although the regimen we evaluated seemed to be clinically effective, the limited power of our study has resulted in inability to draw rm conclusions in this regard.

The compliance with the titrated oral misoprostol regimen was excellent. Although unique and seemingly complicated, after only few cases, the staff to became condent with the protocol. This is probably due to the fact that the principles of the protocol are similar to those of the titrated intravenous oxytocin. Although there were more adverse effects overall in the misoprostol group, they were self-limiting and only one woman abandoned misoprostol due to adverse effects. For adverse effects, which are of concern and attributed to misoprostol,

Table 5. Neonatal outcomes and complications expressed as n (%) or mean (SD) Misoprostol Number randomised Neonatal outcomes Birthweight, mean (SD) Apgar ,7 at 5 minutes Cord venous pH ,7.0 Cord venous base excess .12 Admission to neonatal unit Admission to neonatal unit . 24 hours Neonatal complications Sepsis conrmed with blood culture Seizures Neonatal death 375 3301 (491) 11 (3) 0 8 (2) 6 (2) 5 (1) 0 0 0 Standard 383 3335 (668) 8 (2) 0 3 (1) 10 (3) 7 (2) 1 (0.3) 0 0 WMD 234 (2107 to 38) RR 1.40 (0.573.45) RR 2.72 (0.7310.19) RR 0.61 (0.221.67) RR 0.73 (0.232.28) RR 0.34 (0.018.83) RR or WMD (95% CI)

1508

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

PROM IOL with low-dose misoprostol

Table 6. Subgroup analyses Misoprostol Standard Primiparous Caesarean section Vaginal delivery not achieved within 24 hours Multiparous Caesarean section Vaginal delivery not achieved within 24 hours Favourable cervix Caesarean section Vaginal delivery not achieved within 24 hours Unfavourable cervix Caesarean section Vaginal delivery not achieved within 24 hours 214 41 (19) 71 (33) 227 57 (25) 86 (38) RR (95% CI)

0.76 (0.531.09) 0.88 (0.681.13)

161 13 (8) 20 (12)

156 13 (8) 31 (20)

0.97 (0.462.02) 0.63 (0.371.05)

155 17 (11) 24 (16)

148 20 (14) 24 (16)

0.81 (0.441.49) 0.95 (0.571.60) Figure 4. KaplanMeier plot of time to delivery: favourable cervix (log rank test, P = 0.02).

220 37 (17) 67 (31)

235 50 (21) 93 (40)

0.79 (0.541.16) 0.77 (0.600.99)

namely shivering, diarrhoea and pyrexia, there was no difference. This is likely to be due to the fact that we used a low-dose regimen, and adverse effects are known to be dose related. Uterine hyperstimulation and uterine hyperstimulation syndrome were not increased in the misoprostol group. Again, it is well known that these events are dose related. There were no cases of uterine rupture and serious neonatal complications. However, the three cases of serious adverse events in the misoprostol group were of concern. While all

three cases were not thought to directly associate with misoprostol, we should remain vigilant regarding the safety of misoprostol and indeed of all prostaglandin induction agents. It is regrettable that we had to stop this trial before achieving the prespecied sample size. This has resulted in limited power of our trial to provide clinically meaningful recommendations. However, our results need to be considered in the context of available evidence about low-dose misoprostol regimens (50 micrograms) for IOL following PROM. In the published literature, there are four trials that specically set out to evaluate low-dose misoprostol compared with other regimens for IOL following PROM.912 Three of these trials evaluated low-dose vaginal misoprostol regimens9,10,12 and only one evaluated an oral misoprostol regimen.11 All four trials had small numbers (109197 women), and therefore,

Figure 3. KaplanMeier plot of time to delivery: overall (log rank test, P = 0.057).

Figure 5. KaplanMeier plot of time to delivery: unfavourable cervix (log rank test, P = 0.006).

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

1509

Bricker et al.

our study provides the largest data set evaluating a unique titrated low-dose misoprostol regimen for PROM. In this trial, the women and caregivers were not blind to allocation; therefore, decisions to intervene may have been susceptible to bias such as earlier recourse to caesarean section due to anxiety about experimental therapy. However, enthusiasm for the regimen could have led to bias in the opposite direction. We did not evaluate womens and caregivers views and satisfaction with the different methods of induction. In general, there is a lack of data in this regard. Our experience suggests that oral misoprostol is acceptable as it is easy to administer, and uncomfortable vaginal examination can be avoided. In a similar trial of low-dose oral misoprostol versus vaginal dinoprostone for IOL in women with intact membranes, women preferred the oral route.13 It is difcult to compare the clinical ndings of this trial with other trials as it was unique for three reasons. First, as above mentioned, it was a titrated regimen used throughout labour. Second, a low-dose of misoprostol was given vaginally at the start of induction for women with unfavourable cervix prior to commencing the oral misoprostol regimen. Third, the intervention in the control group was pragmatic reecting usual practice depending on cervical status (dinoprostone with or without oxytocin for unfavourable cervix, and oxytocin only for favourable cervix). The other trials evaluating low-dose misoprostol for PROM used either vaginal misoprostol or oral misoprostol regimens, not a combination, and compared misoprostol with either dinoprostone or oxytocin. The ndings, however, are in keeping with emerging ndings from systematic reviews, namely that low-dose misoprostol is effective at achieving vaginal delivery, particularly for women with unfavourable cervix.

Contribution to authorship
L.B. and Z.A. conceived and designed the trial and with H.P. formed the trial management group (TMG). Some local coordinators contributed to the nal protocol development prior to the trial commencing. All other named authors were instrumental in running the trial in their local units and acquiring data. The TMG coordinated the trial in all centres with H.P. acting as the daily contact person for the local coordinating teams, liaising with the principal investigator (L.B.) as required. L.B. and Z.A. analysed and interpreted the data initially. L.B. drafted the manuscript and all other named authors contributed to critically reviewing and revising the manuscript for nal submission. All authors approved the version to be submitted.

Details of ethics approval

Multi research ethics approval: South East MREC; Reference MREC 01/1/13; 18 April 2001. Each centre obtained local research ethics approval as follows: Liverpool Womens Hospital NHS Foundation Trust, UK Liverpool REC committee, Ref: 2K/207, April 2001. Royal Bolton Hospital, UK Bolton REC committee, Ref: 384/10/01, 10 October 2001. Assiut University Hospital, Egypt Committee of Medical Ethics, Faculty of Medicine, Assiut University, Egypt, Ref: 15/12/2003, 10 December 2003. Whiston Hospital, St Helens, UK St Helens and Knowsley LREC, Ref: 2001/2002-54, 21 November 2001. Stepping Hill Hospital, Stockport, UK Stockport LREC, Ref: EE/SPR/MREC, 23 November 2001. Leighton Hospital, Crewe, UK South Cheshire LREC, Ref: C240/01, 15 January 2002. Warrington Hospital, UK North Cheshire LREC, Ref: NC01-38, early 2002. Royal Preston Hospital, UK Preston, Chorley and South Ribble LREC, Ref: 2001.11.11, 2 July 2002. Ysbyty Gwyned Hospital, Bangor, UK North Wales Health Authority REC (West), Ref: (no ref on letter of approval), 28 March 2002. Queen Mothers Hospital, Glasgow, UK Yorkhill REC, Ref: Mc56/03, 13 December 2003.

Conclusions
Since the patent for misoprostol expired, licensed products for reproductive health indications have become available in some countries (Brazil, France and Egypt) and prelicensing trials for low-dose misoprostol are underway in the UK (www.misoprostol.org/File/availability.php). In resource poor settings where licensed prostaglandins are not affordable, misoprostol is being used frequently and misoprostol 25 micrograms is included in the World Health Organizations Interagency List of Essential Medicines for Reproductive Health (2006). In the interest of safety, we advocate the use of low-dose regimens including the one evaluated in this study.

Funding
H.P. was funded by North West NHS Research and Development for 1 year to set up and coordinate the trial regionally. The trial had no other sources of funding. Some of the local coordinating centres funded midwifery time for the trial.

Disclosure of interests
Z.A. and H.P. worked with Controlled Therapeutics and Alliance on phase III trials of misoprostol for cervical ripening and induction of labour.

Acknowledgements
The authors would like to thank Eileen Reynolds (Principal and Clinical Trials pharmacist, Liverpool Womens Hospital); Sarah Ayers (NPEU, Oxford) who prepared randomisation

1510

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

PROM IOL with low-dose misoprostol

schedule and envelopes and designed and managed the database; and the midwives, pharmacists and women of the collaborating hospitals.

Supporting information
Additional Supporting Information may be found in the online version of this article: Appendix S1. The PROMMIS Trial collaborative group. Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. j

References
1 Gunn GC, Mishell DR Jr, Morton DG. Premature rupture of the fetal membranes: a review. Am J Obstet Gynecol 1970;106:46983. 2 Hannah ME, Olhsson A, Farine D, Hewson BA, Hodnett ED, Myhr TL, et al. Induction of labour compared with expectant management for prelabour rupture of the membranes at term. N Eng J Med 1996;334: 100510. 3 Goldberg AB, Greenberg MB, Darney PD. Misoprostol and pregnancy. N Engl J Med 2001;344:3847.

4 Hofmeyr GJ, Gulmezoglu AM. Vaginal misoprostol for cervical ripening and induction of labour. Cochrane Database Syst Rev 2003;CD000941. DOI: 10.1002/14651858.CD000941. 5 Alrevic Z, Weeks A. Oral misoprostol for induction of labour. Cochrane Database Syst Rev 2006;CD001338. DOI: 10.1002/14651858. CD001338.pub2. 6 Hofmeyr GJ, Alrevic Z, Matonhodze B, Brocklehurst P, Campbell E, Nikodem VC. Titrated oral misoprostol solution for induction of labour: a multi-centre, randomised trial. BJOG 2001;108:9529. 7 Tang OS, Schweer H, Seyberth HW, Lee SWH, Ho PC. Pharmacokinetics of different routes of administration of misoprostol. Hum Reprod 2002;17:3326. 8 Khan R, El-Refaey H, Sharma S, Sooranna D, Stafford M. Oral, rectal and vaginal pharmacokinetics of misoprostol. Obstet Gynecol 2004; 103:86670. 9 Wing DA, Paul RH. Induction of labour with misoprostol for premature rupture of membranes beyond 36 weeks gestation. Am J Obstet Gynecol 1998;178:S93. 10 Sanchez-Ramos L, Chen AH, Kaunitz AM, Gaudier FL, Delke I. Labor induction with intravaginal misoprostol in term premature rupture of membranes: a randomized study. Obstet Gynecol 1997;89:90912. 11 Butt KD, Bennett KA, Crane JM, Hutchens D, Young DC. Randomised comparison of oral misoprostol and oxytocin for labour induction in term prelabour membrane rupture. Obstet Gynecol 1999;94:9949. 12 Frohn WE, Simmons S, Carlan SJ. Prostaglandin E2 gel versus misoprostol for cervical ripening in patients with premature rupture of membranes after 34 weeks. Obstet Gynecol 2002;99:20610. 13 Dodd JM, Crowther CA, Robinson JS. Oral misoprostol for induction of labour at term: randomised controlled trial. BMJ 2006;332:50911.

2008 The Authors Journal compilation RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

1511