NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
A summary of selected new evidence relevant to NICE clinical guideline 70 ‘Induction of labour’ (2008)
Evidence Update, No. 44
London: National Institute for Health and Care Excellence (NICE) ; 2013 . Copyright © National Institute for Health and Care Excellence, 2013.All rights reserved. This material may be freely reproduced for educational and not-for-profit purposes. No reproduction by or for commercial organisations, or for commercial purposes, is allowed without the express written permission of NICE.
1 Induction of labour. NICE clinical guideline 70 (2008).
A search was conducted for new evidence from 17 June 2011 to 15 February 2013. A total of 741 pieces of evidence were initially identified. Following removal of duplicates and a series of automated and manual sifts, 10 items were selected for the Evidence Update (see Appendix A for details of the evidence search and selection process). An Evidence Update Advisory Group, comprising topic experts, reviewed the prioritised evidence and provided a commentary.
Although the process of updating NICE guidance is distinct from the process of an Evidence Update, the relevant NICE guidance development centres have been made aware of the new evidence, which will be considered when guidance is reviewed.
If you have any comments you would like to make on this Evidence Update, please email ku.shn.ecnedive@sutcatnoc
The following table summarises what the Evidence Update Advisory Group (EUAG) decided were the key points for this Evidence Update. It also indicates the EUAG’s opinion on whether the new evidence may have a potential impact on the current guidance listed in the introduction. For further details of the evidence behind these key points, please see the full commentaries.
The section headings used in the table below are taken from the guidance.
Evidence Updates do not replace current accredited guidance and do not provide formal practice recommendations.
Induction of labour to avoid prolonged pregnancy may be associated with fewer perinatal deaths than expectant management.
Evidence for differences in rates of neonatal sepsis between induction of labour or expectant management in women with preterm prelabour rupture of membranes seems to be inconclusive.
Evidence that increasing frequency of membrane sweeping reduces the rate of induction of labour at 41 weeks’ gestation seems to be inconclusive.
Nitric oxide donors do not seem to be effective for induction of labour.Evidence for benefits of mechanical methods of induction of labour over prostaglandins seems to be inconclusive.
Compared with prostaglandin E2, Foley (balloon) catheters do not seem to lower rates of caesarean section, and are associated with longer time to delivery and higher likelihood of needing to administer oxytocin.
Maternal and fetal safety outcomes do not seem to differ whether prostaglandins are administered in the morning or evening, but women appear to prefer morning administration.
Evidence for measurement of cervical length with transvaginal ultrasonography in predicting delivery outcomes in women scheduled for induction of labour seems to be inconclusive.
These commentaries analyse the key references identified specifically for the Evidence Update. The commentaries focus on the ‘key references’ (those identified through the search process and prioritised by the EUAG for inclusion in the Evidence Update), which are identified in bold text. Supporting references provide context or additional information to the commentary. Section headings are taken from the guidance.
No new key evidence was found for this section.
NICE clinical guideline 70 (NICE CG70) recommends that women with uncomplicated pregnancies should usually be offered induction of labour between 41 and 42 weeks to avoid the risks of prolonged pregnancy. The exact timing should take into account the woman’s preferences and local circumstances.
perinatal deaths (relative risk [RR]=0.31, 95% confidence interval [CI] 0.12 to 0.88, p=0.016; 17 trials, n=7407).
instances of meconium aspiration syndrome (RR=0.50, 95% CI 0.34 to 0.73, p=0.00034; 8 trials, n=2371).
caesarean sections (RR=0.89, 95% CI 0.81 to 0.97, p=0.0067; 21 trials, n=8749).Admission to the neonatal intensive care unit was not significantly different between the induction and expectant management groups (RR 0.90, 95% CI 0.78 to 1.04, p=0.16; 10 trials, n=6161).
Risk of bias from lack of blinding was judged to be high, but blinding was not possible because of the nature of the interventions. Overall, most trials were judged to be at moderate risk of bias. However, 2 trials were published only in abstract with full publications not identified, which limited the assessment of potential bias.
This review suggests that induction of labour to avoid prolonged pregnancy may be associated with fewer perinatal deaths than expectant management. This is consistent with NICE CG70, which used a previous version of this Cochrane review from 2006 in the development of recommendations (see the full version of NICE CG70 for details). The updated results are generally similar to those of the previous review, with a narrowing of the confidence intervals for the outcomes of perinatal death. The reduction in caesarean section is now statistically significant, whereas previous results showed no significant difference.
Gülmezoglu AM, Crowther CA, Middleton P et al. (2012) Induction of labour for improving birth outcomes for women at or beyond term. Cochrane Database of Systematic Reviews issue 6: CD004945 [PMC free article : PMC4065650 ] [PubMed : 22696345 ]
An open-label RCT (PPROMEXIL, n=536) by van der Ham et al. (2012a) compared induction of labour or expectant management in women at 34–37 weeks’ gestation with preterm prelabour rupture of membranes. Eligible women with singleton or twin pregnancies were enrolled within 24 hours of membrane rupture. The primary outcome was neonatal sepsis.
In the induction of labour group, women were induced, according to Dutch national guidelines, within 24 hours of randomisation (or if caesarean was planned, it was performed as soon as possible after randomisation). In the expectant management group, women were monitored according to local protocol until spontaneous delivery. Labour was induced if the woman reached 37 weeks of gestation, or if necessary because of other maternal or fetal indications. If caesarean section was planned in this group, it was performed as soon as labour started.
No significant difference in neonatal sepsis was seen between the induction of labour group (7 babies, 2.6%) and the expectant management group (11 babies, 4.1%; RR=0.64, 95% CI 0.25 to 1.6, p=0.346).
The authors then did a meta-analysis combining the results of their trial with a previously published Cochrane review comparing induction of labour with expectant management (Buchanan et al. 2010). A literature search identified no additional new studies to add to the meta-analysis. Data on neonatal sepsis were available from 8 trials (n=1230) including the present study, but no significant difference was seen between interventions (RR=1.06, 95% CI 0.64 to 1.76).
The authors noted that the main limitation of their trial was that it was underpowered. The sample size was calculated on expected rates of neonatal sepsis of 7.5% in the expectant management group and 2.5% in the induction of labour group. However, the rate in the expectant management group was lower than expected at 4.1%, so demonstrating a statistically significant difference would have needed a much larger number of participants. Additionally, the power calculation assumed that induction of labour would not increase the risk of neonatal sepsis. The results of some studies included in the meta-analysis showed a potentially increased risk of neonatal sepsis with induction of labour. Again, a larger sample size would have been needed to account for this possibility.
In a further RCT (PPROMEXIL-2, n=195), van der Ham et al (2012b) randomly assigned women to induction of labour or expectant management according to the trial protocol of PPROMEXIL discussed above. Again, no significant difference in neonatal sepsis was seen between the induction of labour group (3 babies, 3.0%) and the expectant management group (4 babies, 4.1%; RR=0.74, 95% CI 0.17 to 3.2). No power calculation was performed, but this trial is likely to be underpowered because the number of participants was smaller than in the original PPROMEXIL trial, which was underpowered. These results were added to the previous meta-analysis (9 studies, n=1428). The difference between interventions remained non-significant (RR=1.02, 95% CI 0.63 to 1.65).
The results provide no conclusive evidence about differences in rates of neonatal sepsis between induction of labour or expectant management in women with preterm prelabour rupture of membranes. This is consistent with NICE CG70, which recommends discussing the risks of induction of labour with the woman before a decision is made.
van der Ham DP, van der Heyden JL, Opmeer BC et al. (2012b) Management of late-preterm premature rupture of membranes: the PPROMEXIL-2 trial. American Journal of Obstetrics and Gynecology 207: e1–10 [PubMed : 22901981 ]
NICE CG70 recommends that prior to formal induction of labour, women should be offered a vaginal examination for membrane sweeping. When a vaginal examination is carried out to assess the cervix, the opportunity should be taken to offer the woman a membrane sweep. Additional membrane sweeping may be offered if labour does not start spontaneously.
NICE CG70 additionally contains a research recommendation on this topic:
multiple versus once-only membrane sweeping, at varying gestational ages, depending on parity membrane sweeping versus cervical massage?Putnam et al. (2011) did a single-centre RCT (n=350) comparing once-weekly membrane sweeping and twice-weekly membrane sweeping with control in women with an unfavourable cervix (Bishop score ≤4) at 39 weeks’ gestation. Membrane sweeping was done according to protocol until 41 weeks, when all women who had not given birth had labour induced. The primary end point was the proportion of women needing induction at 41 weeks.
Included women were older than 18 years with a singleton pregnancy who planned, and had no contraindication to, vaginal delivery. Additional inclusion criteria were reliable dating of pregnancy and ultrasound confirming the placenta was clear of the cervix.
No significant differences were seen in the proportion of women admitted for induction of labour at 41 weeks. The rates were 34% in the control group, 27% in the once-weekly membrane sweep group and 23% in the twice-weekly group (p=0.149). At baseline, the group who had twice-weekly membrane sweeping had a significantly higher proportion of women with Bishop score of 3–4 (83%) than the once weekly group (71%, p=0.04) or the control group (67%, p<0.001). The authors conducted analyses to adjust for differences in Bishop score, but the results remained non-significant.
The authors discussed the lack of balance of Bishop score between groups and noted that in hindsight the randomisation should have been stratified by Bishop score. The authors additionally noted a potential limitation in that they did not record data for women whose cervix was so unfavourable that it would not allow a full membrane sweep, which may have provided useful additional information.
The results of this study provide no conclusive evidence that increasing frequency of membrane sweeping reduces the rate of induction of labour at 41 weeks’ gestation. This evidence is unlikely to have an impact on NICE CG70 and further research is needed to answer the research recommendation on this topic.
NICE CG70 states that vaginal nitric oxide donors should not be used for induction of labour.
In a Cochrane review, Kelly et al. (2011) investigated the effects of nitric oxide donors for cervical ripening or induction of labour. The analysis included 10 randomised trials (n=1889) that compared nitric oxide donors with placebo, vaginal or cervical PGE2, and vaginal misoprostol. These comparisons were analysed separately.
Primary outcomes were vaginal delivery within 24 hours, uterine hyperstimulation with fetal heart rate changes, caesarean section, serious neonatal morbidity or perinatal death and serious maternal morbidity or death.
Nitric oxide donors did not show significant benefits in the primary outcomes compared with any other interventions assessed. Maternal side effects of headache and nausea were significantly higher in the nitric oxide group in several studies. All studies were judged to be of a generally high standard with low risk of bias.
Nitric oxide donors do not seem to be effective for induction of labour, which is consistent with recommendations in NICE CG70 not to use this intervention.
Kelly AJ, Munson C, Minden L (2011) Nitric oxide donors for cervical ripening and induction of labour. Cochrane Database of Systematic Reviews issue 6: CD006901 [PubMed : 21678363 ]
NICE CG70 states that mechanical procedures (balloon catheters and laminaria tents) should not be used routinely for induction of labour.
Jozwiak et al. (2012) conducted a Cochrane review of 71 RCTs (n=9722) of mechanical methods for induction of labour compared with placebo or no treatment, prostaglandins, or oxytocin. The mechanical methods studied were laminaria tents or synthetic equivalents, catheters, or extra-amniotic infusion by catheter. Comparisons were of any method against any other or placebo or no treatment.
The primary outcomes of interest were vaginal delivery within 24 hours, uterine hyperstimulation with fetal heart rate changes, caesarean section, serious neonatal morbidity or perinatal death and serious maternal morbidity or death.
Many studies did not report on the outcome of vaginal delivery within 24 hours, and most of those reporting this outcome did not find any significant difference between mechanical methods and either no treatment or vaginal PGE2. Most studies reported on caesarean section rates. No significant differences in caesarean section rates were seen for mechanical methods compared with no treatment or prostaglandins. Mechanical methods were associated with lower rates of caesarean section than oxytocin (RR=0.62, 95% CI 0.42 to 0.90, p=0.011; 5 studies, n=398). Less uterine hyperstimulation with fetal heart rate changes occurred when balloon catheters were used, compared with any prostaglandins (RR 0.19, 95% CI 0.08 to 0.43, p=0.000063; 9 studies, n=1931). About half of included studies did not report details of allocation concealment.
These results provide no conclusive evidence that mechanical methods of induction of labour have benefits over prostaglandins, which is consistent with NICE CG70.
Jozwiak M, Bloemenkamp KWM, Kelly AJ et al. (2012) Mechanical methods for induction of labour. Cochrane Database of Systematic Reviews issue 3: CD001233 [PubMed : 22419277 ]
In an open-label RCT (n=824), Jozwiak et al. (2011) compared induction of labour using a Foley (balloon) catheter with induction of labour using vaginal PGE2 in women at term with an unfavourable cervix (Bishop score <6). The primary outcome was caesarean section rate.
In the Foley catheter group, women had a catheter inserted past the internal os of the cervix, and the balloon was inflated with 30 ml of sterile saline or water. Women then had 1 hour of bed rest with fetal and uterine monitoring. Once the catheter was spontaneously expelled or the Bishop score was 6 or more, amniotomy was done. If uterine activity was insufficient an infusion of oxytocin was given. In the PGE2 group, 1 mg PGE2 gel was administered, then another 1 mg after 6 hours. When the Bishop score was 6 or more, amniotomy was done and oxytocin infusion started. In both groups, if the cervix remained unfavourable after 48 hours of treatment, the woman had 1 day of rest then induction started again.
Caesarean section rate did not differ significantly between the Foley catheter group (23%) and the PGE2 group (20%, p=0.38). Oxytocin augmentation was significantly lower in the PGE2 group than in the Foley catheter group (59% versus 86% respectively, p<0.0001). Time from start of induction to birth was also significantly shorter in the PGE2 group (median=18 hours, interquartile range 12–33 hours) than in the Foley catheter group (median=29 hours, interquartile range 15–35 hours, p<0.0001). Admissions to the neonatal ward were significantly higher in the PGE2 group (20%) compared with the Foley catheter group (12%, p=0.0019). Two serious maternal adverse events were recorded, both in the PGE2 group: a uterine rupture and a uterine perforation.
The authors then performed a systematic review of studies comparing induction using a Foley catheter with induction using PGE2, and found 2 studies in addition to their own. A meta-analysis of these 2 trials plus results from the present study (n=1431) suggested that there was no significant difference in the rate of caesarean section between Foley catheters and PGE2 (OR=1.02, 95% CI 0.80 to 1.30, p=0.86).
The authors noted that the estimated 8% reduction in the rate of caesarean sections used in the power calculation was optimistic, and that in view of the results a non-inferiority design would have been more appropriate. Additionally, the authors did not assess the acceptability of the intervention to women. However, in the Foley catheter group, 2 women requested not to receive the intervention at randomisation and 5 women discontinued the intervention on request. This may indicate lower acceptability of Foley catheters than PGE2 gel, because no women requested not to use, or to discontinue, PGE2 gel.
Henry et al. (2013) reported an RCT (n=101) comparing a Foley catheter as outpatient treatment with inpatient PGE2 gel for induction of labour. Women with pregnancies of gestational age of 37 weeks or more, a Bishop score of less than 7 and cervical dilation of less than 2 cm were eligible for inclusion. The primary outcomes were proportion of vaginal births within 12 hours of admission to the birthing unit and total number of inpatient hours from randomisation to delivery.
The Foley catheter was inserted and inflated with 30 ml of sterile water. Fetal monitoring was then performed for at least 30 minutes before the woman was discharged. Women were given pain relief and a sedative with instructions to use if needed, and asked to return to the birthing unit at 7 am the next morning unless they had any concerns or if labour occurred. In the PGE2 group, PGE2 was administered and then fetal monitoring was performed for at least 30 minutes. Pain relief and sedation were given if needed, and the woman was transferred to the birthing unit at 7 am the next morning unless labour started or there were concerns about the mother or baby. In the birthing unit, women had amniotomy plus oxytocin infusion. If this was not possible or the cervix remained unfavourable, further clinical management was undertaken.
Vaginal delivery within 12 hours occurred significantly more often in the PGE2 gel group (53%) compared with the Foley catheter group (28%, p=0.011). Total inpatient stay from randomisation to birth was significantly longer in the PGE2 gel group (mean=32.4 hours, SD16.9 hours) compared with the Foley catheter group (mean=21.3 hours, SD 10.1 hours, p<0.001). However, women in the Foley catheter group had a longer stay in hospital after giving birth, so the difference in total inpatient stay (mean=96 hours, SD 38 hours) was not significantly lower than the PGE2 group (mean=105 hours, SD 38 hours, p=0.267). Significantly more women in the Foley catheter group needed oxytocin (88%) than in the PGE2 group (59%, p=0.01). A survey of maternal satisfaction suggested that this was higher in the Foley catheter group.
The authors noted that recruitment was lower than the sample size estimated in the power calculation (200 women) for the first primary outcome (proportion of vaginal births in 12 hours). The study met the sample size for the second primary outcome (time from randomisation to birth). Additionally, the choice of a mechanical outpatient method compared with a chemical inpatient method may have limited the assessment of inpatient versus outpatient and mechanical versus chemical means. Finally they acknowledged that using total inpatient hours would have been a better primary outcome than inpatient hours before delivery.
The results of these studies suggest that compared with PGE2, Foley catheters do not seem to lower rates of caesarean section, and are associated with longer time to delivery and higher likelihood of needing to administer oxytocin. Therefore, this evidence is unlikely to have an impact on NICE CG70.
