Using a transcervical Foley catheter to enhance normality for induction of labour

Transcervical Foley catheter vs PGE2 gel

Jozwiak et al (2011) compared the effectiveness and safety of induction of labour using a 16-F/18-F transcervical Foley catheter filled with 30ml of saline, compared with PGE2 gel. This study had a large sample size (n=824) and was conducted in 12 hospitals in the Netherlands, between 2009 and 2010. Healthcare practice in the Netherlands is similar to the UK with similar rates of induction (van den Berg, 2015). Training on the use of the transcervical Foley catheter was provided, although it is not made clear how or to what extent. Inclusion criteria were women >37 weeks' gestation, with a singleton fetus, cephalic presentation with intact membranes, an unfavourable cervix and a Bishop score <6. Women were also included if they had oligohydramnios, hypertensive disorders or diabetes, making the study representative of the wider maternity population.

The study was an open label RCT, meaning everyone involved was aware of the treatment. Bias in the trial was therefore not quantifiable, as preference of one treatment over the other is more likely to lead to better outcomes (Aveyard, 2014). Randomised allocation occurred using a computer-generated system that was inaccessible to the recruiters, with the allocation code disclosed after allocation had taken place; this reduced the risk of selection bias and strengthened the efficacy, as both groups also had similar baseline characteristics. The primary outcome was caesarean section rate; secondary outcomes were maternal and neonatal morbidity.

Jozwiak et al (2011) found that there was no difference in caesarean section rates between the two groups, although in the transcervical Foley catheter group, more caesarean sections occured because of failure to progress. The authors noted that there was a longer interval in achieving active labour status in the transcervical Foley catheter group; however, it is widely understood that the length of the latent phase of labour is difficult to establish, poorly understood and under-researched (Nirmal and Fraser, 2016). Therefore, it would have been appropriate to evaluate the clinician's interpretation of the latent phase of labour, or the parameters used to designate failure to progress.

Jozwiak et al (2011) reported that vaginal assessments occurred more frequently in the PGE2 group. According to the Royal College of Midwives (RCM) (2012a), there is no research-based information to make recommendations for the timing and frequency of vaginal examinations, which can be painful and distressing and can have a profound effect on women. Moreover, vaginal examinations affect women's psychological and physical health, further increasing their risk of infection. The trial therefore found women were treated for an intrapartum infection more often in the PGE2 group. The authors noted that PGE2 is a mediator of the febrile response, causing a rise in body temperature, and that pyrexia is also a documented side effect. This should be explored further, as pyrexia is a trigger for the sepsis protocol, which may lead to the unnecessary initiation of intrapartum antibiotics, and as a result, it would have been valuable for the authors to have documented how many women were diagnosed with an infection, as opposed to those who were treated. The study also did not provide statistical data for neonates treated for infection. The authors found that fewer neonates were admitted to the neonatal unit in the transcervical Foley catheter group, although both groups had a similar admission rate into neonatal intensive care. Apgar scores and cord gas values showed no statistically significant difference.

A significant finding of the study was that uterine perforation and uterine rupture occurred only in the PGE2 group. In this group, women who experienced uterine hyperstimulation in either did so after the initial insertion of the PGE2, or once oxytocin infusion commenced. In contrast, hyperstimulation occurred in the transcervical Foley catheter group once oxytocin had been introduced. The RCT therefore demonstrated, using a relatively large sample, that the transcervical Foley catheter was a safe method of induction compared with PGE2 gel. The intention-to-treat basis gives strength to the findings and data analysis; however, the study did not assess or evaluate the women's experiences or satisfaction with either method. This would have been valuable in identifying issues and providing a more naturalistic insight in to the use of the transcervical Foley catheter for induction of labour.

Balloon catheters vs PGE2 gel

In contrast to Jozwiak et al (2011), Pennell et al (2009) used a qualitative element to measure women's satisfaction and efficiency of the procedure when conducting a RCT into three different methods of induction. This a strength of the study, as the researchers could then ensure that the treatment was ethical and gain insight into the women's perspectives (Entwhistle, 2005). A total of 333 nulliparous women took part, comparing the double balloon catheter (n=110), the 16-F transcervical Foley catheter filled with 30 ml saline (n=110), and PGE2 (n=113). The study took place at a tertiary hospital in Perth, Western Australia between 2001 and 2003. The inclusion/exclusion criteria was similar to that of the Jozwiak et al (2011) study, although Pennell et al (2009) used the modified Bishop Score.

To minimise variation and enhance the study's internal validity (Enkin et al, 2000; Ness Evans and Rooney, 2010), the initial pre-induction vaginal examinations were performed by one person; this ensured a greater accuracy in determining Bishop score, cervical favourability and time from induction to birth. Generalisability of vaginal examinations would result in significant variation, as progress assessed by vaginal examinations can be inaccurate when performed by different examiners (RCM, 2012a). However, as Pennell et al (2009) minimised vaginal examinations initially, the validity of the induction-to-birth time remains valid. Labour progression data could be inaccurate as it is presumed that multiple people performed the vaginal examinations throughout labour, but the study does not clarify this further.

The study, which is also supported by the findings of Jozwiak et al (2011), found no difference in caesarean section rates between the groups; however, a non-reassuring fetal heart rate trace was the main indicator for medical intervention in the PGE2 group. Both studies identified that hyperstimulation occured primarily in the PGE2 group, and that uterine hyperstimulation only occurred in the transcervical Foley catheter group once an oxytocin infusion had been commenced. In view of this evidence, uterine hyperstimulation is a common reccurring adverse risk associated with PGE2.

However, the similar caesarean section rate in both studies could be associated with the regular intermittent use of a cardiotocography (CTG). The evidence suggests that the use of a CTG is associated with an increased risk of caesarean section and instrumental births, as clinicians are more likely to intervene (Scamell and Alaszewski, 2016). The transcervical Foley catheter means that CTG monitoring is only required once oxytocin infusion is commenced. In the UK, intermittent fetal monitoring is required via CTG during the induction of labour with PGE2, as there are adverse associated risks with the drug (NICE, 2008; RCOG, 2008). It therefore would have been beneficial for the women in the transcervical Foley catheter group to be monitored similarly to those in the PGE2 group, as these results might imply that women using the transcervical Foley catheter are at more of a risk of caesarean section.

Pennell et al (2009) also found that fetal blood sampling occurred mostly in the PGE2 group, in which five neonates had fetal blood samples taken and severe hypoxia was evident in four of the five. There were no reported episodes of hypoxia in the transcervical Foley catheter group. This contradicts the study by Jozwiak et al (2011), who found no variance in cord gas values, and led to Pennell et al (2009) reporting worse neonatal outcomes in the PGE2 group.

Contrary to Jozwiak et al (2011), Pennell et al (2009) found no increase in maternal temperature during labour. This is an unexpected finding, as Pennell et al (2009) used up to three doses of 2 mg of PGE2, which is larger than recommended by the drug manufacturer or RCOG guidelines (National Collaborating Centre for Women's and Children's Health, 2008).

Pennell et al (2009) used a visual analogue scale (VAS) to assess satisfaction, pain on insertion and pain on ripening. This was the only study that attempted to assess women's experiences, although it was not in-depth research. Although there are limitations to VAS systems, such as the patient's ability to interpret their pain score, and it is reportedly weaker than other visual systems (Pasero and McCaffery, 2011), it positively identified that women preferred the transcervical Foley catheter to the PGE2. The transcervical Foley catheter was reported to be the least painful; as pain is a contributing factor of woman's holistic health that negatively affects her experience (NICE, 2008), this is a fundamental finding of the study. Women in the double balloon catheter group had the longest time from induction to delivery, while women in the transcervical Foley catheter group reported the shortest time. This is an important variant in the trial, because a longer induction-to-labour period negatively affects the woman's wellbeing and increases the likeliness of further medical intervention (RCM, 2012b; WHO, 2010).

Optimal transcervical Foley catheter volume

Gu et al (2015) conducted an investigation on the optimal filling volume for the transcervical Foley catheter during induction. The study also aimed to improve vaginal delivery time by assessing the effectiveness of the transcervical Foley catheter at 12 hours and 24 hours. Groups 1 and 2 had a transcervical Foley catheter filled with 30 ml normal saline for 12 or 24 hours, while groups 3 and 4 had a transcervical Foley catheter filled with 80 ml normal saline for 12 or 24 hours. However, of the included studies, this was the only study that did not mention the size of the transcervical Foley catheter used, and as a result, it is difficult to make recommendations on the type of transcervical Foley catheter to use in future practice.

The study was conducted in China, where 504 women took part after selection and computer-generated randomisation. The study had similar inclusion and exclusion criteria to that of Jozwiak et al (2011) and Pennell et al (2009). All medical staff involved in the trial were given advanced training on assessing Bishop score, use of a transcervical Foley catheter for induction of labour. They were also given clear directives given for the interpretation of active labour and hyperstimulation status, and their management, which enabled the researchers to minimise premature intervention or misinterpretation of the findings during the trial (Aveyard, 2014).

A primary outcome of the study was achieving vaginal delivery within 24 hours, similar to Pennell et al (2009), who also considered time of induction to delivery as an important factor. It is reported (NICE, 2014a) that women consider the time from induction to delivery as an important factor when undergoing induction of labour, whereby the longer the induction process, the greater the level of maternal distress (Coates et al, 2019). Prolonged labour is also associated with higher infection rates, increased use of antibiotics and oxytocin infusion (NICE, 2014b). Gu et al (2015) found that there was a shorter induction to delivery time in the 30 ml/12 hour transcervical Foley catheter group. Despite this, Gu et al (2015) used logistic regression analysis and found that independent variables such as gestational age, parity and birth weight were contributory factors in achieving a vaginal delivery within 24 hours. Their original finding was therefore contradictory, as after reviewing the data, no statistical significance was apparent in all groups from time of induction to delivery. However, the study showed that, once in active labour, women who were induced displayed similar progress to that of women in spontaneous labour, which raises questions regarding the parameters for timing and progression when being induced. The findings from Gu et al (2015) suggested that these time limitations are unnecessary could lead to further intervention.

Caesarean section rates were reported to be the lowest in the 30 ml/12 hour group, although not significantly different to the other groups. However, the study found that failure to progress was the most common cause for caesarean section, which further supports the findings of Jozwiak et al (2011). Only one episode of hyperstimulation occurred, which is statistically lower than the hyperstimulation rates reported by Jozwiak et al (2011) and Pennell et al (2009), as PGE2 was not used as a variable. Crane et al (2001) found that 48.6% of women who underwent induction of labour by PGE2 experienced uterine hyperstimulation, resulting in a higher incidence of cord compression and fetal compromise (Strachan, 2009). This corroborates with the findings of Gu et al (2015), who reported low rates of hyperstimulation across all types of transcervical Foley catheter used in the study. Although PGE2 was not compared, this suggests that hyperstimulation noted by Cromi et al (2011) and Pennell et al (2009) is related to PGE2 and explains why there were such low rates of uterine hyprerstimulation reported in the study by Gu et al (2015).

The incidence of postpartum haemorrhage was relatively high: of the 376 women; 19.4% experienced postpartum haemorrhage, the highest incidence of all studies reviewed. This may be due to the women's pre-existing risk factors, although Pennell et al (2009) also included women with similar characteristics to Gu et al (2015) and found no statistically significant incidence of postpartum haemorrhage, despite reporting that more women suffered a blood loss of >1000 ml in the PGE2 group. Jozwiak et al (2011) also reported similar rates of postpartum haemorrhage between groups; however, more women required a blood transfusion in the PGE2 group. This variation between Gu et al (2015) and the other two studies could be due to the wider healthcare, population and demographic differences.

Optimal volume and time for induction

Cromi et al (2011) conducted an RCT of 397 women in Italy during 2008-2010, which compared three methods of induction of labour. An 18-F transcervical Foley catheters filled with 50 ml saline were allocated to two of the three groups, to test the vaginal delivery outcome at 12 and 24 hours, similar to Gu et al (2015). The third group was allocated to PGE2 controlled-release pessary. The primary aim of the study was to determine if there was an optimal time of efficacy of the transcervical Foley catheter. The women had similar baseline characteristics, and inclusion and exclusion criteria were similar to that of Jozwiak et al (2011), Pennell et al (2009) and Gu et al (2015). Cromi et al (2011) included women who were >34 weeks', but did not explain why.

Caesarean section rates were similar between the three groups and comparable with the findings from the studies reviewed thus far. However, expedited caesarean section deliveries occurred more frequently in the PGE2 groups, on the grounds of non-reassuring fetal heart rate, suggesting that PGE2 has an effect on the fetal heart. During the active phase of labour, caesarean section rates remained comparable between the three groups, as with Pennell et al (2009). Cromi et al (2011) reported no episodes of hyperstimulation in either of the transcervical Foley catheter groups, with hyperstimulation occurring in 6.1% of the PGE2 group. Neonatal admissions to intensive care were higher in the PGE2 group, in contrast to Jozwiak et al (2011) who reported that admission rates were similar between Foley catheter and PGE2 groups.

Cromi et al (2011) reported that oxytocin infusion was used more frequently in both transcervical Foley catheter groups. As oxytocin use is associated with an increased risk of postpartum haemhorrhage (Subramanian and Penna, 2009; Gilbert, 2010), increased adverse risks, particularly in women with pre-existing maternal conditions, would be expected, as shown by Gu et al (2015). However, Cromi et al (2011) did not report on the outcomes of these adverse risks, and so if the study were to be repeated, reporting rates of postpartum haemhorrhage would be beneficial.

Kaplan-Meier survival curves showed that the 12-hour transcervical Foley catheter group took longer than the PGE2 group to progress; however, once in active labour, the transcervical Foley catheter group made significant progress to continue and give birth in a much shorter period, similar to studies by Pennell et al (2009) and Gu et al (2015). Cromi et al (2011) therefore recommend the use of a transcervical Foley catheter for 12 hours, which is enough time bring about cervical changes and allow women to progress. This is a key recommendation, also discussed by Jozwiak et al (2011); however, Cromi et al (2011) also reiterate that time from induction to delivery is an important factor when caring for women undergoing induction of labour as it improves their overall experience and holistic health (WHO, 2010; RCM, 2012b).