Membrane Sweeping Versus Transcervical Foley Catheter for Induction of Labour in Women With Previous Caesarean Delivery

Induction of labour (IOL) is an obstetric intervention to artificially initiate labour before its spontaneous onset. It is undertaken when maternal or fetal risks of continuing pregnancy outweigh the risks of IOL. The methods of IOL include pharmacological method (prostaglandin), mechanical methods (transcervical Foley catheter and laminaria tents) and surgical method (amniotomy). Pharmacological method and mechanical methods are preferred choice for IOL in women with unfavourable cervix. Whereas surgical method is reserved for women with favourable cervix in which membranes are accessible.

With the global rise in Caesarean section, more and more women embark on pregnancy with previous Caesarean scar. It is estimated that one in every four term pregnancies requires IOL. With IOL in previous Caesarean section, the risk of uterine scar rupture with subsequent catastrophic maternal and neonatal outcomes is the greatest concern. Previous influential study highlighted that IOL with prostaglandin conferred the greatest risk of uterine rupture (2.45%) in women with one previous Caesarean delivery. On the other hand, the risks of uterine rupture were lower and comparable among women whose labour induced without the use prostaglandin (0.77%) and women with spontaneous labour (0.52%). Among the various IOL methods, prostaglandin is more likely associated with shorter interval to achieve vaginal delivery and avoidance of oxytocin augmentation. Nevertheless, mechanical methods have lower risk of uterine hyperstimulation but have more risk of infection. As a result, obstetricians are facing dilemma on deciding the safe and effective method of IOL in women with previous Caesarean delivery.

In Sibu Hospital, membrane sweeping is routinely offered to women with previous Caesarean delivery who require IOL. This technique involves the examining finger passing through the cervix to rotate against the wall of the uterus, to separate the chorionic membrane from the decidua, or to massage around the cervix if the os is closed. However, membrane sweeping may not exert its cervical ripening effect immediately and the delivery may be delayed by up to 8 days. This may render a proportion of women to resort to repeated Caesarean section for failed induction. The vaginal birth rates following membrane sweeping in women with unscarred uterus are ranging between 78%-90%. Among women with previous Caesarean delivery undergoing IOL with membrane sweeping, the reported vaginal birth rates were conflicting. A randomised controlled trial by Hamdan et al. showed vaginal birth rate of 56.1%. Another comparative study by Ramya et al. reported much lower vaginal birth rate of 17.3% which could be explained by high rate of Caesarean Section for maternal request in this study. The average intervals from sweeping to labour onset and to delivery were about 2 days and 4 days, respectively. Membrane sweeping did not increase the risk of maternal or neonatal infection, Caesarean section, postpartum haemorrhage and neonatal poor APGAR score, though discomfort during vaginal examination, minor vaginal bleeding and irregular contraction were frequently reported by women undergoing membrane sweeping. No cases of uterine rupture was reported among women with scarred uterus undergoing membrane sweeping.

Pharmacological IOL method involving prostaglandin is avoided in Sibu Hospital in view of higher risk of scar rupture and presence of limited number of consultants to cope with this debilitating event. Non-pharmacological methods of IOL such as Foley catheter, double-balloon catheter, hygroscopic cervical dilator (laminaria tent) etc are known to have lesser risk of scar rupture. Double-balloon catheters are limited in numbers in this hospital and they are costly. Laminaria tent is not available in this hospital. Foley cathter is a consumable item in this hospital and thus is readily available all the time. Compared to double-balloon catheter, Foley catheter has equivalent cervical ripening efficacy and safety profile. In addition, Foley catheter is also cheaper and has shorter induction to delivery interval. All these advantages make Foley catheter an ideal method of IOL for women with previous Caesarean delivery.

Among women with previous Caesarean delivery underlying IOL with transcervical Foley catheter, the reported vaginal birth rates are ranging between 43.5%-71.4%. The retrospective cohort analysis by Bujold et al. compared the risk of uterine rupture among women with previous Caesarean delivery undergoing spontaneous labour, induction by amniotomy with or without oxytocin, or preinduction cervical ripening with transcervical Foley catheter. The rates of uterine rupture were similar among the groups ((1.1% vs 1.2% vs 1.6%, p = 0.81). However, transcervical Foley catheter appeared to be safer than low dose oxytocin infusion for cervical ripening before amniotomy as the latter tended to be associated with risk of uterine scar dehiscence. Another retrospective cohort study done by Gonsalves et al. showed no case of uterine rupture among 68 women with previous Caesarean section undergoing IOL with transcervical Foley catheter insertion.

The transcervical insertion of balloon catheter as a 'foreign material' may theoretically increase the risk of uterine infection. But, the current data available are conflicting. Meta-analysis by Heinemann et al. demonstrated that the Foley catheter was associated with a significantly higher rate of maternal infections, defined as fever, endometritis or chorioamnionitis, compared to the use of prostaglandins or oxytocin for induction of labour (7.6% vs 5%, pooled OR 1.5, 95%CI 1.07-2.09). On the other hand, Cochrane Review concluded that there is no evidence of an increased risk of infectious morbidity with balloon catheters. In the PROBAAT-trial, comparing the Foley catheter with the use of vaginal prostaglandin E2 gel, the rate of intrapartum infection was significantly lower in women with the Foley catheter (1% vs 3%, p = 0.035). Two other trials evaluating labour induction with balloon catheters after rupture of the membranes did not show an increased risk for maternal infection.

Maslovitz et al. reported the rate of vaginal bleeding of 1.8% following transcervical Foley catheter insertion. However, the bleeding was minor with unaltered haemoglobin levels and coagulation profile. There was also no case of cervical tear which might result in postpartum haemorrhage. Besides, risk of change of fetal vertex presentation to breech was reported to be 1.3%. This was probably related to concomitant presence of unengaged free-floating fetus and contraction following Foley catheter insertion which resulted in flipping movement of the fetus. Risk of poor APGAR and NICU admission was not significantly higher compared to other IOL methods eg. prostaglandin and oxytocin.

Till date, there is paucity of evidence on the superiority of membrane sweeping and transcervical Foley catheter in IOL in previous Caesarean delivery. Both methods have the same mechanism action by increasing local production of endogenous prostaglandin. The aim of this study is to evaluate the effectiveness of membrane sweeping and transcervical Foley catheter insertion for IOL in women with previous Caesarean delivery.

The investigators had searched electronic databases including PubMed, Cochrane Library, Scopus, Ovid, JSTOR and Google Scholar. The investigators found that there was no study comparing these two IOL methods, thus no prior information is available for sample size calculation. Based on Julious SA, the investigators recruited 24 subjects (a sample size of 12 per group) from 15 February 2018 to 8 May 2018 and performed an interim analysis to calculate the actual sample size required for a full randomised controlled trial (RCT). The justifications for this interim sample size were based on rationale about feasibility, precision about the mean and variance, and regulatory considerations. The justifications for this interim sample size are based on rationale about feasibility, precision about the mean and variance, and regulatory considerations.

The final sample size was calculated based on the study primary outcome of achievement of favourable cervix (Bishop score of 8 or more) within 48 hours of IOL. From the interim analysis, the investigators found that the rate of favourable Bishop score achieved within 48 hours of IOL was 58.3% for membrane sweeping arm and 91.7% for transcervical Foley catheter arm. The investigators used the following formula by Fleiss JL. 1981 to calculate the final sample size. With α at 0.05 and power at 0.8, total of 52 subjects (26 subject in each arm) need to be recruited. With the estimated dropout rate of 15%, the final sample size required is 60 subjects, which means a further 36 subjects (18 subjects on each arm) will be recruited to complete the study. In addition, data of the initial 24 subjects (12 per arm) recruited will be included in the final data analysis. This further recruitment will not affect the randomisation of the previous and future subjects because blocked randomisation, in a block size of 6 and an allocation ratio of 1:1 is used. Furthermore, there is no change in the methodology, scale and instrument of the research.

The investigators had applied ethical approval of the recruitment of a further 36 subjects from the Malaysia Research and Ethics Committee, Ministry of Heath Malaysia and the application was approved on 12 June 2018.