NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Caesarean Section. London: RCOG Press; 2011 Nov. (NICE Clinical Guidelines, No. 132.)

Cover of Caesarean Section

Caesarean Section.

Show details

5Planned caesarean section

This chapter considers the evidence related to decisions about planned mode of birth. Other aspects of management of specific conditions or complications of pregnancy are not included because they are outside the scope of the guideline.

5.1. Breech presentation

About 4% of all singleton pregnancies are breech presentation. The proportion of breech presentation fetuses decreases with increasing gestation: 3% of term infants, 9% for those born at 33–36 weeks of gestation, 18% of those born at 28–32 weeks and 30% of those born at less than 28 weeks4. Breech presentation, is associated with cerebral palsy and handicap, due principally to the association with preterm birth and congenital malformations.57,58

Breech presentation is the primary indication for 10% of all caesarean sections (CSs). Overall 88% of pregnancies with breech presentation in England and Wales are delivered by CS (56% planned and 44% unplanned CS). However CS rates vary with gestational age, at term 91% women with a breech presentation had a CS, while at less than 28 weeks the CS rate was less than 40%.4 [evidence level 3]

External cephalic version

Interventions to promote cephalic version of babies in the breech position include external cephalic version (ECV), moxibustion and postural management. The research basis for these interventions is included in the guideline on antenatal care of healthy pregnant women.1

External cephalic version involves applying pressure to the mother's abdomen to turn the fetus in either a forward or backward somersault to achieve a vertex presentation. Recognised complications of ECV attributable to the procedure (and incidence) include:

  • fetal heart rate abnormalities: the commonest is transient bradycardia (1.1% to 16%)5962
  • placental abruption (0.4% to 1%)59,61
  • painless vaginal bleeding (1.1%)61
  • admission for induction of labour (3%).62

Two systematic reviews examined the effect of ECV at term and before term. Performing ECV at term reduced the number of non-cephalic births by 60% when compared with no ECV (6 randomised controlled trials [RCTs], n = 612 women, risk ratio [RR] 0.42, 95% confidence interval [CI] 0.35 to 0.50).63 [evidence level 1a] A reduction in caesarean section is also observed in the ECV group when compared with no ECV (6 RCTs, n = 612, RR 0.52, 95% CI 0.39 to 0.71). ECV before 37 weeks gestation does not reduce non-vertex births at term (RR 1.02, 95% CI 0.89 to 1.17).64 [evidence level 1a]

Success rates following ECV in primiparous women range from 35% to 57% and from 52% to 84% in multiparous women5962,65 [evidence level 2b]. Interventions to improve the success rates of ECV include the routine or selective use of tocolysis, the use of regional analgesia and the use of vibroacoustic stimulation.68 None of the RCTs has used newer tocolytics and the effectiveness of these is uncertain66 [evidence level 1a]. Further guidance on ECV may be found in the RCOG greentop guideline on the management of breech presentation.67

In the National Sentinel Caesarean Section Audit (NSCSA) external cephalic version was offered to 33% of women having a CS for breech presentation at term, this was the same irrespective of the woman's parity. ECV was provided by consultants, specialist registrars or staff grade obstetricians.4 [evidence level 4] If ECV was offered to all women at term, assuming a 50% success rate then it is likely this would reduce the overall CS rate by 1%.

Cost effectiveness of ECV

Six cost-effectiveness studies were identified that considered the role of ECV in decreasing the rate of CS, two in the United Kingdom and four in the USA.

The UK studies reported the cost of ECV.53 The first was a cost study that reported an expected cost of £1,452 for ECV versus £1,828 for not having ECV, an expected saving of around £380.53. The results were insensitive (i.e. did not alter the result) to changes in the cost of an ECV. The cost of CS would need to fall by £8,576 (a fall of 56%, again, a highly unlikely scenario) for the non-ECV option to be the less costly option. However, the sensitivity analysis showed that the ECV success rate would only have to fall by around 5% for the ECV option to be the less favourable option. Therefore the cost analysis cannot categorically determine which option is least costly overall. The second UK study (much smaller) found that the cost of birth with a successful ECV was £2,230 and £2,595 for an unsuccessful ECV, a cost saving of around £360 per birth.68

Four American studies have been published. One used a decision analytic modelling technique to determine the overall costs of four management options for breeches at term: ECV with planned vaginal birth, ECV with CS, selected vaginal birth and planned CS.69 The decision model used hospital charges (not costs) for vaginal birth of US$6000 and US$10,000 for CS (a wider ratio than the reported UK cost data). The expected CS rate was 25.4% (± 5.4) for ECV plus planned vaginal birth; 31.9% (± 6.6) for ECV plus planned vaginal birth; 62.6% (± 5.9) for selected vaginal birth and 88.6% (± 3.4) for planned CS. The model estimated the expected cost for each pathway (the cost of vaginal birth and CS for each option arm) and found that ECV with planned vaginal birth was the least costly option, due to the lower proportion of CS for this group (US$8071) and planned CS to be the most costly (US$9544). Whether these reported costs were statistically different is not reported. The validity of the range of probabilities used in the decision analysis were subsequently questioned.70

A study in the same year considered the costs of failed and successful ECV separately and reported a cost of US$8042 for women with failed ECV and US$5059 for women with successful ECV.71 However, the effectiveness data on which this study was based was a cohort study and not an RCT.

An American study also presented data to show that successful ECV would yield savings over unsuccessful ECV.72 The most recent US study was a much larger study of 695 women.73 This was a decision-analytic model to calculate the potential cost savings from ECV (in terms of reduced CS rates). The authors assumed that ECV would be successful in 44% of cases, of which 67% would proceed to vaginal birth and 33% to a CS. They further assumed that ECV would be unsuccessful in 56% of cases, of which only 7% would proceed to a successful vaginal birth. Given these assumptions, the model calculated a savings (in US hospital charges) of around $650 per birth. Savings from every ECV attempted (even if not successful) versus ECV not attempted were around US$3000 per birth (these are greater due to higher reported rates of CS for women not attempting ECV).

Therefore in conclusion ECV yields cost savings in comparison with CS. There is no UK-based economic evaluation comparing ECV with vaginal breech birth.

Term breech pregnancy and CS

A systematic review identified 3 RCTs (n = 2396) that evaluated the effect of mode of birth for term breech pregnancies.36,43,44,48 [evidence level 1a] The majority of the information about the effect of planned CS in the review comes from one international multi-centre RCT which is of good methodological quality (n = 2088 women, 121 centres in 26 countries).48 [evidence level 1b]

Offering planned CS reduced perinatal or neonatal death (excluding fatal anomalies) or serious neonatal morbidity (RR 0.33, 95% CI 0.19 to 0.56).36 The risk of perinatal/neonatal mortality or serious morbidity was 1.6% in the planned CS group and 5.0% in the planned vaginal birth group. The absolute risk reduction in perinatal/neonatal mortality or serious neonatal morbidity was 3.4%, therefore for every 29 CS for term breech pregnancy one baby will avoid death or serious morbidity.36 [evidence level 1a] These findings are consistent with findings from cohort studies74,75

The findings of the RCT and the systematic review are the subject of continued debate. Therefore more details about this RCT are outlined here. The RCT included a number of maternity units in the UK. About 40% of women recruited to the trial were in labour at time of randomisation. The women in labour were not further divided into stages of labour so there is no information on how many were in the second stage of labour. However advanced labour was not listed as an exclusion criterion. The RCT protocol provided guidance on management of labour. This included intermittent fetal heart monitoring (every 15 minutes in the first stage and every 5 minutes in the second stage), adequate progress in labour was defined as 0.5 cm dilatation per hour and descent of the breech to the pelvic floor within 2 hours of full cervical dilatation. Delivery of the breech could be spontaneous or assisted; the after coming head could be controlled using the Mauriceau–Smellie–Veit manoeuvre or forceps. The position of the woman for the second stage of labour was not stipulated by the protocol nor was this information collected during the trial.48 [evidence level 1b]

Sub group analysis within this RCT has been undertaken to evaluate if the effect on perinatal mortality or morbidity could be explained by specific factors.48 These effects remain consistent and are therefore not explained by differences in:

  • operator experience
  • prolonged labour
  • induction of labour with oxytocin or prostaglandins
  • augmentation of labour
  • type of breech presentation (footling or uncertain)
  • the use of epidural analgesia.

Women who were in labour were included in the RCT (therefore the findings of the trial are generalisable to women in labour); however the effect of CS on neonatal outcomes is not reported separately for this group. It is possible that the benefits and risks of caesarean section particularly during the second stage are different. Therefore further research that specifically addressed this issue was sought; however no studies evaluating the effect of CS for undiagnosed breech compared to expectant management were identified. An RCT to address this issue would require randomisation of at least 4230 women with undiagnosed breech pregnancy to either CS or vaginal birth in order to detect at least a 40% difference in neonatal morbidity.

The effects of planned CS for term breech on maternal health are less clear. The RCTs included in the systematic review assessed the impact of CS on maternal health using a variety of measures and combining the results across studies is not always possible. Where the estimates could be combined, no difference is detected in the measures of maternal morbidity (such as blood loss, blood transfusion, infection) between planned CS and planned vaginal birth.36 Estimates of composite measures of morbidity have previously been reported36 however these pooled estimates are not included in the guideline because it is unclear whether these estimates are based on person or event data. It is possible that the same woman may have more than one morbidity (for example a woman who needs additional surgery is more likely to need a blood transfusion or admission to ITU) so that composite morbidity measures based on summation of event rates rather than number of women affected can lead to spurious results.48 [evidence level 1b] Data for individual women was reported in one RCT, it did not detect any difference in composite maternal morbidity between women in the planned CS group or women in the planned vaginal birth group (RR 1.24, 95% CI 0.79 to 1.95).48 [evidence level 1b] The specific estimates of the effect of planned CS on maternal health are outlined in Table 4.5.

Preterm breech

Breech presentation, is associated with cerebral palsy and handicap, due principally to the association with preterm birth and congenital malformations.57,58 The proportion of breech presentation fetuses decreases with increasing gestation: 9% for those born at 33–36 weeks of gestation, 18% of those born at 28–32 weeks and 30% of those born at less than 28 weeks.4,76

Overall 88% of pregnancies with breech presentation were delivered by CS. However CS rates varied by gestational age, 87% for babies born at 33–36 weeks, 81% of those born at 28–32 weeks, and 39% for babies born at less than 28 weeks.4 [evidence level 3]

The results of the term breech trial RCT are relevant for term breech pregnancies, extrapolation to preterm breech babies is inappropriate. In the Confidential enquiry into stillbirths and deaths in infancy (CESDI) Project 27/28 report, survival rates were lower for babies who were breech (84.5%) when compared to babies who were cephalic presentation (89.4%). Survival for breech presentation was significantly greater in those delivered by CS (86.5%) than those delivered vaginally (77.4%).76 [evidence level 3]

Recommendations

NumberRecommendation
9Women who have an uncomplicated singleton breech pregnancy at 36 weeks' gestation should be offered external cephalic version. Exceptions include women in labour and women with a uterine scar or abnormality, fetal compromise, ruptured membranes, vaginal bleeding or medical conditions. [A] [2004]
10Pregnant women with a singleton breech presentation at term, for whom external cephalic version is contraindicated or has been unsuccessful, should be offered CS because it reduces perinatal mortality and neonatal morbidity. [A] [2004]
NumberResearch recommendation
RR 3Further research is needed to determine the effect of caesarean section compared with vaginal birth for women with:
  • preterm breech
  • a breech presentation that is diagnosed in the second stage of labour.

5.2. Multiple pregnancy

About 15 per 1000 pregnancies are multiple gestations; the majority of these are twin pregnancies (twins 14.4 per 1000, triplets 4 per 10,000).4 There have been increases in the rates of multiple pregnancy in the last ten years that are attributed to the use of ovulation induction in fertility treatments.77,78 Perinatal mortality and morbidity such as cerebral palsy are higher among multiple births than singleton births (stillbirths: multiple: 2%. singleton: 0.5%; neonatal deaths multiple: 2.5. singleton 0.3%; RR cerebral palsy twins 4.63 (3.32–6.46.)79,80 [evidence level 3] Some of the observed increase is explained by the association of multiple pregnancy with preterm birth.4,80 Other factors which have been associated with poorer outcome in twin pregnancy include low birth weight, discordant growth between twins, monochorionic twins and being a second born twin.8185 The management of complications (such as discordant growth, monochorionic twins) and other obstetric complications in pregnancy (such as pre-eclampsia) will influence the mode of delivery decisions, however these are outside the scope of this guideline and are therefore not discussed further in this section.

Multiple pregnancy is the primary indication for 1% of caesarean sections.4 Overall 59% of twin pregnancies were delivered by CS. (37% planned and 63% unplanned CS). CS for delivery of the second twin following vaginal birth of the first baby was carried out in 3.5% of twins (n = 75). CS rates vary by gestational age, at term 60% women with a twin pregnancy had a CS, while at less than 28 weeks the CS rate was less than 29%.4 [evidence level 3] Where CS was planned for multiple pregnancy, breech presentation of the first twin was the most commonly reported indication (14%), together with previous CS (7%) and maternal request (9%). Of the unplanned caesarean sections, fetal distress was the most influential factor in 29% and “failure to progress” in 12%. Almost all triplet pregnancies (92%) were delivered by CS.4 [evidence level 3]

A systematic review that included 1 RCT (n = 60) compared CS for a second twin with a non vertex presentation to vaginal birth.37 [evidence level 1b] The methodological quality of this trial is uncertain because ‘randomisation was according to a protocol that was changed randomly by a non-involved person, without prior notice, on a time basis’.45 No difference was detected in any of the baby outcome measures, however the study is too small to accurately estimate the effect on outcomes such as neonatal birth trauma and perinatal death The study reported no difference in the average length of hospital stay (8 days compared to 5 days) and no difference in need for blood transfusion (RR 1.5 95% CI 0.27 to 8.28). Women in the planned CS group had increased risk of puerperal pyrexia compared to women in the planned vaginal birth group (RR 3.67 95% CI 1.15 to 11.69).45 [evidence level 1b]

A large number of observational studies using population based registers have been published. However the majority of these studies are analysed by actual mode of delivery rather than intended mode of delivery, the reports provide insufficient data on neonatal outcome for women who had planned CS85,86 and in the analysis paired tests have not been used to take into account that the outcome within twin pairs maybe related.87,88 One systematic review included only studies where the intended mode of delivery could be identified. The review included 3 retrospective cohort studies8991 and the RCT discussed above.45 The results from these studies were consistent and did not detect differences in neonatal morbidity such as low 5-minute Apgar score, birth trauma, neurological complications, hyperbilirubinaemia, hypoglycaemia, transient tachypnoea or secondary apnoea. The studies are too small to evaluate perinatal mortality.

Triplet and higher order multiple births are rare. They most frequently are the result of ovulation induction for treatment of fertility problems.78 Triplets are almost always born preterm and some of the poorer outcomes such as cerebral palsy seen in these infants are due to preterm birth. These and other complicating factors may influence the mode of delivery decisions. Almost all triplet pregnancies (92%) were delivered by CS.4 [evidence level 3] We identified 3 small retrospective case control studies which compared baby outcomes according to mode of birth for triplet pregnancies (119 sets of triplets in total). The babies born vaginally tended to have better outcomes such as higher Apgar scores than those delivered by CS. However these studies are analysed by actual mode of delivery rather than intended mode of delivery and do not use analysis to take into account that the outcome within triplets will be related.9294 [evidence level 2b]

Women who have multiple pregnancies have an increased risk of maternal mortality and morbidity. CEMD estimates maternal mortality is increased with multiple pregnancy (20.3 per 100 000 twin pregnancies; 215 per 100 000 triplet pregnancies, compared with 11.2 per 100 000 for singleton pregnancies).95 [evidence level 3] The effect of mode of delivery on this outcome is uncertain.

Timing of planned CS for twin pregnancy

Planned CS of twins between 36–37 weeks and 6 days is associated with increased risk of respiratory disorders (transient tachypnea [TTN] or respiratory distress syndrome [RDS]) in one or both of the twins compared to CS between 38 and 40 weeks (RR 5.94, 95% CI 0.78 to 45.01).96 [evidence level 2b] Multiple pregnancy is an established risk factor for preterm birth. About 29% of twin pregnancies are likely go into spontaneous labour before 37 weeks however CS in labour is associated with a reduced risk of respiratory disorders.4 We did not identify any studies that had evaluated the optimal timing for CS in higher order multiple births.

Recommendations

NumberRecommendation
11In otherwise uncomplicated twin pregnancies at term where the presentation of the first twin is cephalic, perinatal morbidity and mortality is increased for the second twin. However, the effect of planned CS in improving outcome for the second twin remains uncertain and therefore CS should not routinely be offered outside a research context. [C] [2004]
12In twin pregnancies where the first twin is not cephalic the effect of CS in improving outcome is uncertain, but current practice is to offer a planned CS. [GPP] [2004]
NumberResearch recommendation
RR 4RCTs are needed to evaluate the benefits and risks to mothers and babies of CS for delivery of twin and triplet pregnancies.

5.3. Preterm birth and CS

Preterm birth is the most common cause of neonatal mortality (47% of neonatal deaths are due to immaturity).76 Babies born preterm are also at increased risk of morbidity (such as cerebral palsy) however the impact of mode of delivery on outcomes is uncertain.97,98 Preterm birth may result from spontaneous preterm labour or because delivery is thought to be beneficial to the mother's (such as severe pre-eclampsia or HELLP) or baby's health (for example presumed fetal compromise). Other obstetric complications (such as multiple pregnancies and breech presentation) are associated with preterm birth and will influence the mode of delivery decisions, however detailed discussion of the appropriate management of all these situations is outside the scope of this guideline. Changing the mode of birth for preterm infants to CS has been proposed as a means of reducing the morbidity and mortality40 [evidence level 3] However when the infant is very small delivery can be difficult at CS.76 In addition upper segment caesarean section (classical) may be needed in about 10% of babies born at 27–28 weeks which may have a significant impact on future pregnancies of these women.76

A systematic review of planned CS versus expectant management for birth of the small baby identified six RCTs (n = 122).35 [evidence level 1a] Three RCTs included only breech presentation and three included only cephalic presentations. All trials were discontinued before reaching their projected sample size because of difficulties in recruitment or difficulties in weight estimation where trial entry criteria were based on birthweight.41 [evidence level 1b] About 1 in 6 of the babies allocated to CS were born vaginally, and vice versa. The findings of the review are inconclusive because there were too few events to give sufficiently precise estimates of effect that would be clinically useful.

A large number of observational studies evaluating mode of birth of preterm infants on mortality and morbidity (such as cerebral palsy) have been published. However the impact of mode delivery on neonatal outcome remains uncertain.76,97,99102 [evidence level 3]

Recommendations

NumberRecommendation
13Preterm birth is associated with higher neonatal morbidity and mortality. However, the effect of planned CS in improving these outcomes remains uncertain and therefore CS should not routinely be offered outside a research context. [C] [2004]
NumberResearch recommendation
RR 5RCTs are needed to evaluate the impact of CS on the benefits and risks to mothers and babies born preterm.

5.4. Small for gestational age

Small for gestational age (SGA) refers to a fetus that has failed to achieve a specific biometric measurement (for example abdominal circumference) or estimated weight threshold by a specific gestational age. The commonly used threshold is the tenth centile. About half of these babies are constitutionally small, others are fetuses that are not achieving their growth potential (fetal growth restriction, FGR). SGA fetuses are at greater risk of stillbirth, birth hypoxia, neonatal complications and impaired neurodevelopment. However, most term SGA infants do not have significant morbidity or mortality.103 It is beyond the scope of this guideline to consider the investigation and management of small for gestational age infants other than the effect of CS on neonatal outcome, however this topic is covered by another guideline.103

No RCTs were identified that directly reported on baby outcomes for planned CS versus planned vaginal birth for SGA babies. One RCT has compared delayed versus immediate delivery after diagnosis of fetal growth restriction. This trial reported that delayed delivery resulted in fewer CS (OR 2.7, 95% CI 1.6 to 4.5).104 [evidence level 1b] Observational data has suggested that SGA babies exposed to labour are more at risk of neonatal death than those not exposed to labour (RR 1.79, 95% CI 1.54 to 1.86).105 [evidence level 3] CS may reduce the need for neonatal resuscitation (OR 0.2, 95% CI 0.08 to 0.66).106 [evidence level 3]

The effect of CS on cerebral palsy in low birth weight babies is not certain. CS is not associated with a difference in rates of cerebral palsy.107,108 [evidence level 3] Currently available guidelines do not recommend a mode of birth for SGA babies.103 [evidence level 4]

Recommendations

NumberRecommendation
14The risk of neonatal morbidity and mortality is higher with ‘small for gestational age’ babies. However, the effect of planned CS in improving these outcomes remains uncertain and therefore CS should not routinely be offered outside a research context. [C] [2004]
NumberResearch recommendation
RR 6RCT evidence is needed to determine the effect of planned CS on neonatal mortality and morbidity for ‘small for gestational age’ babies.

5.5. Placenta praevia

Placenta praevia is the primary indication for about 3% of all CS (2.2% not actively bleeding and 0.9% actively bleeding).4 The majority of low lying placenta detected at 20 weeks of gestation will resolve. If the placenta extends over the os, a repeat US should offered at 32 weeks of gestation.1 (NCC-WCH, 2008) Placenta praevia may also present with painless bleeding. CS is usually necessary when the placenta covers the internal os at 36 weeks of gestation (minor or major placenta praevia).

Women having a CS for placenta praevia are at increased risk of blood loss of greater than 1000 ml compared with women having a CS for other indications (RR 3.97, 95% CI 3.24 to 4.85).4 In the last triennial report from the Confidential Enquiry into Maternal Deaths in the UK, four deaths occurred in women with placenta praevia, three as a result of haemorrhage.95 Hence, they should have the CS carried out by an experienced operator with a consultant readily available and at a maternity unit with on-site blood transfusion services.

Recommendations

NumberRecommendation
15Women with a placenta that partly or completely covers the internal cervical os (minor or major placenta praevia) should be offered CS. [D] [2004]

5.6. Morbidly adherent placenta

Introduction

Women who become pregnant after a single previous CS have a 0.6–1.3% risk of developing placenta praevia: of these women, 11–14% will have a morbidly adherent placenta (Guise et al., 2010). With two previous CSs, there is a 1.1–2.3% risk of placenta praevia in a subsequent pregnancy and of these women, 23–40% will have a morbidly adherent placenta (Guise et al., 2010). With three or more previous caesarean sections, there is a 1.8–3.7% risk of placenta praevia in a subsequent pregnancy: of these women, 35–67% will have a morbidly adherent placenta (Guise et al., 2010).

Against this backdrop of incremental risk, CS is becoming an increasingly common mode of delivery in the UK, both as a primary and a repeat procedure. Thus, clinicians can expect to see a gradual increase in the number of women presenting in pregnancy with a morbidly adherent placenta.

Morbidly adherent placenta is associated with serious maternal morbidity including major obstetric haemorrhage, transfusion of large quantities of blood products, hysterectomy and admission to an intensive care unit. However, exsanguination and maternal death from morbidly adherent placenta is now rare in the UK (Cantwell R. et al., 2011). It is hoped that improved prenatal identification of such cases has contributed to this.

This section will review the evidence for the accuracy of imaging techniques in diagnosing morbidly adherent placenta in a pregnant woman with a previous CS who present with placenta praevia. It also reviews the evidence relating to the optimum management once the diagnosis has been made.

Accuracy of diagnostic tests

Review question

What is the accuracy of imaging techniques (colour flow ultrasound [US] and magnetic resonance imaging [MRI]) for diagnosis of a morbidly adherent placenta in pregnant women who have had a previous CS and are currently diagnosed with placenta praevia?

Overview of evidence

Five studies were included in this review (Warshak et al., 2006; Twickler et al., 2000; Masselli et al., 2008; Shih et al., 2009; Comstock et al., 2009).

Three studies were conducted in the USA (Warshak et al., 2006; Twickler et al., 2000; Comstock et al., 2009), one in Italy (Masselli et al., 2008) and one in Taiwan (Shih et al., 2009). One retrospective study examined the diagnostic accuracy of transvaginal ultrasound for diagnosis of placenta accreta in pregnant women with antenatal diagnosis of placenta praevia who had prior CS (Comstock et al., 2009). One retrospective study reported on the diagnostic accuracy of ultrasound (grey scale or colour Doppler) and MRI for diagnosis of placenta accreta in pregnant women with antenatal diagnosis of low anterior placenta and placenta praevia who had had at least one prior CS (Warshak et al., 2006). One prospective study compared the value of pelvic ultrasound with colour Doppler and MRI for diagnosis of placenta accreta, increta and percreta (Masselli et al., 2008). One prospective study introduced additional criteria for diagnosis of placenta accreta using 3D power Doppler complementary to grey scale and colour Doppler technique (Shih et al., 2009). One prospective study evaluated the diagnostic accuracy of Doppler colour-flow mapping for diagnosis of placenta accreta in pregnant women with prior CS and diagnosis of anterior low lying placenta and placenta praevia (Twickler et al., 2000).

Table 5.1. GRADE summary of findings for diagnostic accuracy of tests for placenta accreta, increta and percreta.

Table 5.1

GRADE summary of findings for diagnostic accuracy of tests for placenta accreta, increta and percreta.

Evidence statements

In the following statements these definitions have been used when summarising the levels of sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV):

  • high: 90% and above
  • moderate: 75% to 89%

Evidence was identified using a variety of ultrasounds to determine diagnostic accuracy for placenta accreta in women diagnosed with placenta praevia who had at least one prior CS. The quality of the evidence ranged from moderate to low for the included studies.

Grey scale transabdominal ultrasound

One study evaluated the diagnostic accuracy of grey scale ultrasound for placenta accreta in women diagnosed with placenta praevia who had a prior CS. The study reported a high sensitivity, a moderate specificity, a moderate PPV and a high NPV. The evidence for this test was of low quality.

Grey scale transvaginal ultrasound

One study evaluated the diagnostic accuracy of grey scale transvaginal ultrasound for placenta accreta in women at 15 to 20 weeks of gestation who were diagnosed with placenta praevia and had a prior CS. The study reported a moderate sensitivity and a low PPV. Specificity and NPV were not reported. The evidence for this test was of low quality.

One study evaluated the diagnostic accuracy of grey scale transvaginal ultrasound for placenta accreta in women at 15 to 40 weeks of gestation who were diagnosed with placenta praevia and had a prior CS. The study reported a moderate sensitivity and a low PPV. Specificity and NPV were not reported. The evidence for this test was of low quality.

Grey scale or colour Doppler ultrasound

One study evaluated the diagnostic accuracy of grey scale or colour Doppler ultrasound to diagnose placenta accreta in women diagnosed with placenta praevia who had a prior CS. The study reported a moderate sensitivity, a high specificity, a low PPV and a high NPV. The evidence for this test was of low quality.

Ultrasound colour Doppler

Three studies evaluated the diagnostic accuracy of ultrasound colour Doppler to diagnose placenta accreta in women diagnosed with placenta praevia who had a prior CS. One study reported a high sensitivity and specificity with a high PPV and a high NPV. A second study reported high sensitivity, moderate specificity, a moderate PPV and a high NPV. A third study reported a high sensitivity, a low specificity, a moderate PPV and a moderate NPV. The evidence for this test was of moderate quality in the first study and low quality in the other two studies.

MRI

Two studies evaluated the diagnostic accuracy of MRI to diagnose placenta accreta in women diagnosed with placenta praevia who had a prior CS. The first moderate quality study reported a high sensitivity, specificity, PPV and NPV, while the second reported a moderate sensitivity with a high specificity, a high PPV and a moderate NPV. The evidence for this test was of moderate quality.

3D power colour sonography

One study evaluated the diagnostic accuracy of 3D power colour sonography to diagnose placenta accreta in women diagnosed with placenta praevia who had a prior CS. The study reported a high sensitivity, a moderate specificity, a moderate PPV and a high NPV. The evidence for this test was of low quality.

Evidence to recommendations

Relative value placed on the tests considered

The GDG felt that in current practice grey scale ultrasound would not generally be used to make a decision about placenta accreta. Instead, the healthcare professional would use colour-flow ultrasound in order to highlight movement.

Trade-off between clinical benefits and harms

The GDG noted evidence from one moderate quality study and one large study of low quality that colour ultrasound is moderately accurate at ruling out morbidly adherent placenta. MRI scan is better for a more complete diagnosis (that is, considering both accurately ruling in and ruling out morbidly adherent placenta).

The GDG noted that evidence from one moderate quality prospective study showed that the diagnostic accuracy of MRI was 100% without the use of contrast dye. However, the GDG also understood that women may not wish to undergo an MRI scan for a number of reasons (such as discomfort at being enclosed in a small space, the risk of supine hypotension, the noise of the machine and the length of the procedure). The GDG therefore recognised the importance of discussing the procedure with the woman beforehand, explaining both the potential benefits and risks. This discussion should include an explanation of the degree of accuracy that can be expected, and information that the use of an MRI should enable better accuracy determining the degree of adherence.

The GDG noted that MRI is more accurate at identifying women who have a morbidly adherent placenta and therefore better able to help decision making regarding the choice of hospital advised for giving birth (local hospital or tertiary centre).

The GDG agreed with the generally held opinion that both MRI and ultrasound are safe for use in pregnancy.

Trade off between net health benefits and resources

Conclusions relating to cost effectiveness of diagnosing morbidly adherent placenta are presented in Section 13.2 where evidence relating to diagnostic accuracy of ultrasound and MRI and evidence relating to the effectiveness of antenatal diagnosis are considered together to inform the health economic modelling.

Quality of evidence

The GDG noted that in the majority of the studies, the imaging techniques were carried out prior to 30 weeks of gestation, whereas in clinical practice these scans would be more likely to be carried out after 32 weeks of gestation (since the low lying placenta scan won't generally occur until after a repeat scan for low lying placenta, which is usually carried out at 32–34 weeks).

The GDG noted that in the Warshak (2006) study looking at MRI, the person interpreting the MRI scans wasn't blinded to the results of the earlier colour ultrasound, thus potentially enhancing the diagnostic accuracy findings in favour of MRI.

The GDG noted that in all of the studies, while the high level of suspicion about morbidly adherent placenta in these women might be thought to inflate the figures for diagnostic accuracy, since this is the clinically relevant population, the figures reported are credible when generalised to clinical practice.

Other considerations

The GDG felt that as there was only one study investigating the use of 3D ultrasound, and given that it is not widely available throughout the UK, it was not appropriate to recommend its use.

The GDG also considered the relevance of evidence reviewed and recommendations to woman with other uterine scars (for example myomectomy) but in the absence of evidence pertaining specifically to this group, they did not feel this was possible.

Recommendations

NumberRecommendation
16If low-lying placenta is confirmed at 32–34 weeks in women who have had a previous CS, offer colour-flow Doppler ultrasound as the first diagnostic test for morbidly adherent placenta. [new 2011]
17If a colour-flow Doppler ultrasound scan result suggests morbidly adherent placenta:
  • discuss with the woman the improved accuracy of magnetic resonance imaging (MRI) in addition to ultrasound to help diagnose morbidly adherent placenta and clarify the degree of invasion
  • explain what to expect during an MRI procedure
  • inform the woman that current experience suggests that MRI is safe, but that there is a lack of evidence about any long term risks to the baby
  • offer MRI if acceptable to the woman. [new 2011]
18Discuss the interventions available for delivery with women suspected to have morbidly adherent placenta including cross matching of blood and planned CS with a consultant obstetrician present. [new 2011]
NumberResearch recommendation
RR 7How accurate is 3D ultrasound compared with 2D ultrasound or MRI scanning for diagnosing morbidly adherent placenta?

Effect of diagnosis on outcomes

Review question

Does a diagnosis of morbidity adherent placenta using imaging techniques lead to improved outcomes in pregnant women with a previous caesarean section currently diagnosed with placenta praevia?

Overview of evidence

Two studies were included in this review (Warshak et al., 2009; Wong et al., 2008). One study was conducted in the USA (Warshak et al., 2009) and one in New Zealand (Wong et al., 2008). One observational study examined the effects of antenatal diagnosis of placenta accreta on maternal outcomes (Wong et al., 2008) and the other observational study compared maternal and neonatal outcomes in women with an antenatal diagnosis of placenta accreta (managed by planned caesarean hysterectomy) with women in whom an antenatal diagnosis was not made (Warshak et al., 2009).

In the US study all women diagnosed with placenta accreta were offered a planned CS with hysterectomy (without attempted removal of the placenta). A caesarean hysterectomy was scheduled for 34–35 weeks of gestation after a 48 hour course of betamethasone to enhance fetal lung maturity. A multidisciplinary team involving specialists from perinatology, anaesthetics, gynaecological oncology, interventional radiology and neonatology were involved in women's care. Hysterectomies were performed under general anaesthesia. Internal iliac balloon catheters were passed preoperatively and inflated during surgery only if significant bleeding was encountered. Most women spent the first day postoperatively in the intensive care unit and stayed longer if clinically indicated. The diagnosis of placenta accreta in all women was confirmed post delivery with a histological test.

In the New Zealand study women diagnosed with placenta accreta were offered a planned CS. Five women had a hysterectomy and the uterus was conserved in two. No further details are reported regarding the package of care offered.

Maternal outcomes

Table 5.2. GRADE summary of findings for antenatal diagnosis of placenta accreta compared with no antenatal diagnosis (maternal outcomes).

Table 5.2

GRADE summary of findings for antenatal diagnosis of placenta accreta compared with no antenatal diagnosis (maternal outcomes).

Neonatal outcomes

Table 5.3. GRADE summary of findings for antenatal diagnosis of placenta accreta compared with no antenatal diagnosis (neonatal outcomes).

Table 5.3

GRADE summary of findings for antenatal diagnosis of placenta accreta compared with no antenatal diagnosis (neonatal outcomes).

Evidence statements

Maternal outcomes
Estimated blood loss

One study found that the mean blood loss in women with an antenatal diagnosis of placenta accreta was lower than in women with no antenatal diagnosis of placenta accreta. This finding was statistically significant. A second study did not find a statistically significant difference for this outcome. The evidence for this outcome was of very low quality.

Number of units of blood transfused

One study found that women with an antenatal diagnosis of placenta accreta had a lower number of packed red blood cell transfusions compared with the women who had no antenatal diagnosis of placenta accreta. This finding was statistically significant. A second study did not find a statistically significant difference for this outcome. The evidence for this outcome was of low quality.

Emergency hysterectomy

One study found that the incidence of emergency hysterectomy was lower among women with an antenatal diagnosis of placenta accreta compared with those who had no antenatal diagnosis of placenta accreta. This finding was statistically significant. The evidence for this outcome was of very low quality.

Intensive care unit (ICU) admission

Two studies did not find a statistically significant difference in the rate of ICU admission for women with an antenatal diagnosis of placenta accreta compared with women who had no antenatal diagnosis of placenta accreta. The evidence for this outcome was of very low quality.

Length of hospital stay

Two studies did not find a statistically significant difference in length of hospital stay for women with an antenatal diagnosis of placenta accreta compared with women who had no antenatal diagnosis of placenta accreta. The evidence for this outcome was of very low quality.

Bladder injuries

Two studies did not find a statistically significant difference in the rate of bladder injuries for women with an antenatal diagnosis of placenta accreta compared with women who had no antenatal diagnosis of placenta accreta. The evidence for this outcome was of very low quality.

Neonatal outcomes
Neonatal intensive care unit (NICU) admission

One study found that the rate of NICU admission was higher in neonates born to women with an antenatal diagnosis of placenta accreta than in neonates born to women with no antenatal diagnosis of placenta accreta. This finding was statistically significant. The evidence for this outcome was of very low quality.

NICU length of stay

One study did not find a statistically significant difference in NICU length of stay for neonates born to women with an antenatal diagnosis of placenta accreta compared to neonates born to women with no antenatal diagnosis of placenta accreta. The evidence for this outcome was of very low quality.

Health economics

A de novo model to compare different diagnostic strategies for morbidly adherent placenta in praevia was developed for this guideline. The results of this analysis are summarised here; further details are provided in Chapter 13.

The model compared the following diagnostic strategies:

  • none
  • ultrasound
  • MRI
  • ultrasound followed by MRI in ultrasound test positives.

There is an absence of evidence about how much a diagnosis of morbidly adherent placenta leads to improved outcomes. Even if it does, the ‘downstream’ saving and quality adjusted life years (QALY) gain from averting ‘adverse outcomes’ are further unknowns. Therefore, the model took a ‘what-if’ approach to assess what would be considered cost effective in different scenarios. There is also some uncertainty about the precise diagnostic accuracy of the different diagnostic tests, although there is at least some evidence for these, which perhaps make this uncertainty of secondary importance in terms of making guideline recommendations.

The model suggested that a diagnostic strategy of ultrasound alone was dominated by other alternatives, which meant that other strategies were likely to be cheaper and more effective. Although it has the lowest diagnostic cost, the high cost of false positives in a low prevalence population makes it the most expensive strategy overall. Furthermore, the evidence suggests that such a strategy would miss more cases than a strategy of MRI alone. This finding did not depend on assumptions about improved outcomes arising from the detection of cases.

The ‘what-if’ analysis started from the premise that identifying cases would lead to improved outcomes. Although there is an absence of evidence for this and an effect size can't be estimated, the GDG was strongly of the opinion that ‘being prepared’ offered some protection from risk. Under that premise, there were scenarios where ‘do nothing’, ultrasound plus MRI and MRI alone could be considered cost effective. However, in general a much lower effect size, QALY gain and ‘downstream’ cost saving from averting ‘adverse outcomes’ was necessary for ultrasound plus MRI to be cost effective than for MRI alone.

Evidence to recommendations

Relative value placed on the outcomes considered

The GDG recognised that although blood loss is an important outcome, the way that it has been reported is not useful for its decision making. The GDG members were particularly concerned about identifying the number of women where blood loss could be potentially life-threatening but this was not reported.

The GDG recognised that although the differences in the rates of hysterectomies appeared to be a significant finding, these results were due to the local protocol at the hospital (that is, where a placenta accreta was first discovered when performing the CS, the clinicians would perform an elective hysterectomy). The group felt that this was not a common approach and so did not wish to place any value on the differences reported.

The GDG did not feel that the findings related to ICU admission and length of hospital stay, for both women and neonates, were particularly helpful as the decision about length of stay will often be determined by local protocols.

The GDG felt that the number of bladder injuries was an important outcome. However, it recognised that neither of the studies showed a statistically significant difference for this outcome.

Trade-off between clinical benefits and harms

The GDG members believed from their experience that the main benefit of diagnosing a morbidly adherent placenta is that this allows clinicians to be prepared and to ensure that appropriate measures are taken in cases of extreme blood loss. These include ensuring that there is sufficient cross-matched blood available and that experienced specialist clinicians are available to provide support when needed.

Trade-off between net health benefits and resource use

There is insufficient evidence to determine the cost effectiveness of different diagnostic strategies for morbidly adherent placenta. The modelling undertaken for this guideline suggested that ultrasound alone was likely to be dominated (that is, there was likely to be a diagnostic strategy which was cheaper and more effective) because of a higher false positive rate. There is some evidence to support a view that ultrasound has a lower specificity than MRI, although it is not conclusive. The model also suggested that a sequential strategy of ultrasound followed by a confirmatory MRI where an ultrasound test is positive is cheaper than a strategy of MRI alone. This is because the sequential MRI test removes the costs of false positives which more than offsets the costs associated with an additional test. The sequential strategy involves a much smaller number of MRI scans than a strategy based on MRI alone and because of the substantial difference in costs between an ultrasound and MRI this means that the sequential strategy has markedly lower diagnostic costs, even if the total number of tests undertaken is higher.

Economic evaluation should consider benefits as well as costs but such evidence does not exist. Therefore, the model took a ‘what-if’ approach towards the benefits of correct diagnosis. The GDG members were strongly of the opinion that identifying cases was likely to lead to better outcomes. The model suggested that much smaller gains were necessary for ultrasound plus MRI to reach a cost-effectiveness threshold relative to “do-nothing” than for MRI alone to be considered cost effective relative to ultrasound plus MRI. Therefore, it would be difficult to justify a recommendation for a diagnostic strategy of MRI alone given existing evidence. Such a strategy is not common in current UK practice and there could be capacity issues which would hinder the implementation of such a recommendation. Although current UK practice varies, ultrasound plus MRI is an approach used in some centres. Although further evidence is required, a recommendation of ultrasound plus MRI seems to make pragmatic sense given current practice, GDG opinion and the insights available from the model produced for this guideline.

Quality of evidence

The GDG recognised that there were only two studies that provided evidence for this question, and that the quality of the evidence for the findings from these studies was very low. GDG members noted that one of the studies only contained a small number of women in each arm and so was likely to be underpowered for rare outcomes such as bladder injury.

Given the poor quality of the evidence and lack of detail in one study about the specific management regimes used, the GDG did not feel able to make a strong recommendation for specific interventions. It was noted that in one study (Warshak, 2009) the management strategy of elective caesarean hysterectomy was used for all women, an approach which is not usual practice and thus findings from this study are not generalisable to situations where conservative management is undertaken.

In light of the large amount of blood loss associated with both arms of each study, the GDG agreed that there were steps that should be taken to minimise morbidity associated with this.

Other considerations

The group recognised that there are a number of interventions that are used in to reduce blood loss during surgery, such as balloon catheters and interventional radiology. In addition, some trusts have cell salvage equipment available that can also be used to reduce the need for cross-matched blood. There is variation in practice concerning the use of these interventions in the management of morbidly adherent placenta and a lack of evidence to support their use. Consequently, the GDG felt it important to recommend that further research is conducted in this area.

Recommendations

NumberRecommendation
19When performing a CS for women suspected to have morbidly adherent placenta, ensure that:
  • a consultant obstetrician and a consultant anaesthetist are present
  • an experienced paediatrician is present
  • a senior haematologist is available for advice
  • a critical care bed is available
  • sufficient cross-matched blood and blood products are readily available. [new 2011]
20When performing a CS for women suspected to have morbidly adherent placenta the consultant obstetrician should decide which other healthcare professionals need to be consulted or present. [new 2011]
21All hospitals should have a locally agreed protocol for managing morbidly adherent placenta that sets out how these elements of care should be provided. [new 2011]
NumberResearch recommendation
RR 8What is the effectiveness of procoagulant agents (such as recombinant factor VIIa, beriplex, tranexamic acid, fibrinogen concentrate) in reducing blood loss in women with morbidly adherent placenta?
RR 9What is the effectiveness of point of care testing for haematological indices in women with an established postpartum haemorrhage and in cases of morbidly adherent placenta in reducing maternal morbidity?
RR 10What is the effectiveness of the components of the package of care for morbidly adherent placenta such as imaging techniques (e.g. interventional radiology including balloon catheters), stenting of ureters, removal of the placenta, and cell salvage in reducing morbidity associated with maternal blood loss?
RR 11What is the appropriate gestational age of elective birth for babies of women with a morbidly adherent placenta?
RR 12What is the effectiveness of performing an elective hysterectomy to reduce morbidity associated with blood loss in women with morbidly adherent placenta?

5.7. Predicting CS for cephalopelvic disproportion in labour

Pelvimetry (clinical or X-ray) has been used to predict the need for CS in pregnant women. A systematic review of four RCTs (n = 895) assessed the effects of X-ray pelvimetry on mode of birth. Two RCTs included women with a previous CS. The women on whom pelvimetry was performed were more likely to be delivered by CS (Peto OR 2.17, 95% CI 1.63 to 2.88); There were no differences in neonatal outcomes (asphyxia, admission to neonatal unit, scar dehiscence).109 [evidence level 1a] Guidelines have recommended that pelvimetry is not used except in rare circumstances such as if the woman has had a previous fracture of the pelvis.110

Other tests to predict failure to progress (FTP) have included shoe size, maternal height and size of fetus. Observational studies have not demonstrated their value in predicting FTP in labour.111,112 [evidence level 3]

Recommendations

NumberRecommendation
22Pelvimetry is not useful in predicting ‘failure to progress’ in labour and should not be used in decision making about mode of birth. [A] [2004]
23Shoe size, maternal height and estimations of fetal size (ultrasound or clinical examination) do not accurately predict cephalopelvic disproportion and should not be used to predict ‘failure to progress’ during labour. [B] [2004]

5.8. Mother-to-child transmission of maternal infections

This section addresses CS as an intervention to reduce mother-to-child transmission (MTCT) of viral infections (such as human immunodeficiency virus [HIV]). Other interventions also impact on the risk of the mother-to-child transmission of viral infections (such as anti-retrovirals for HIV) but these topics are outside the scope of this guideline.

HIV

Introduction

Approximately 86,500 people in the UK are carriers of the human immunodeficiency virus (Health Protection Agency, 2010). About 1 in 450 pregnant women nationally are HIV positive and this rises to about 1 in 250 pregnant women in London. In general, around one quarter of all people carrying the HIV virus is unaware of their HIV positive status.

Current NICE guidelines for routine antenatal care recommend that screening for HIV should be offered to all pregnant women in the UK because various interventions can decrease the maternal to fetal (vertical transmission of the virus (Antenatal care guideline, NCC-WCH, 2008). The previous version of this guideline recommended that ‘HIV-positive women who are pregnant should be offered a planned CS because it reduces the risk of mother-to-child transmission of HIV’. However, since the publication of the original guideline there has been a growing body of evidence suggesting that for some women taking antiretroviral therapy (ART) or highly active antiretroviral therapy (HAART), the chance of vertical transmission is reduced so effectively that CS may no longer be associated with reduced vertical transmission rates compared with vaginal birth, even in the presence of a detectable viral load.

This new evidence is reviewed in this section, together with an update of the recommendations for clinical practice.

Review question

What is the effectiveness of planned caesarean section compared with planned vaginal birth at decreasing the mother-to-child transmission of the virus in pregnant women with HIV, for both low and high viral load?

Overview of evidence

Four studies were included in this review which investigated the effectiveness of planned CS compared with vaginal birth at decreasing the rates of mother-to-child transmission of the virus in pregnant women with HIV for both low and high viral load (Boer et al., 2010; Warszawski et al., 2008; Islam et al., 2010; Townsend et al., 2008).

One study was a prospective observational study (Boer et al., 2010) and three were retrospective observational studies (Townsend et al., 2008; Warszawski et al., 2008; Islam et al., 2010). Two of the studies were conducted in the UK (Islam et al., 2010; Townsend et al., 2008), one in France (Warszawski et al., 2008) and one (Boer et al., 2010) in eight Western European countries (Italy, Spain, Belgium, Netherlands, UK, Germany, Denmark and Sweden). All four studies reported mother-to-child transmission rate, viral load count and mode of birth. Two studies investigated planned CS versus planned vaginal birth (Islam et al., 2010; Townsend et al., 2008) and the other two compared planned CS with vaginal birth (planned and unplanned) (Boer et al., 2010; Warszawski et al., 2008).

Three studies reported on mother-to-child transmission rate and mode of birth in HIV infected pregnant women with low viral load defined as undetectable viral load (less than 50 copies/ml) (Boer et al., 2010; Islam et al., 2010; Townsend et al., 2008). One study reported on mother-to-child transmission rate and mode of birth in women with low viral load defined as less than 400 copies/ml (Warszawski et al., 2008).

Table 5.4 is the summary of the evidence by mother-to-child transmission rate for all the published studies identified for this review question. The results have been divided by plasma viral load count.

Table 5.4. GRADE summary of findings for mother-to-child transmission (MTCT) of HIV.

Table 5.4

GRADE summary of findings for mother-to-child transmission (MTCT) of HIV.

Evidence statements

Mother-to-child transmission in women with viral load less than 50 copies/ml on HAART

Two studies did not find a statistically significant difference in the rates of mother-to-child transmission in women on HAART having a planned CS compared with those having a vaginal birth (planned or unplanned; including intrapartum CS) with viral load less than 50 copies/ml.

One study found no reported incidences of mother-to-child transmission of HIV for women having planned vaginal birth in women with viral load less than 50 copies/ml, of whom 14 out of 23 were on HAART. The evidence for this outcome was of very low quality.

Mother-to-child transmission in women with viral load of 50 copies/ml or more up to 1000 copies/ml on HAART

One study did not find a statistically significant difference between transmission rates for women on HAART with viral load of 50 copies/ml or more up to 1000 copies/ml having a planned CS compared with those having a planned vaginal birth. The evidence for this outcome was of very low quality.

Mother-to-child transmission in women with viral load less than 400 copies/ml with and without HAART

One study found that planned CS reduces the risk of mother-to-child transmission compared with vaginal birth (planned or unplanned; including intrapartum CS) when maternal viral load is less than 400 copies/ml (some women receiving HAART, some receiving ART, a few receiving no ART; numbers not reported). This finding was statistically significant. The evidence for this outcome was of very low quality.

Mother-to-child transmission in women with viral load less than 400 copies/ml on ART

One study did not find a statistically significant difference in transmission rate for women on ART with a viral load less than 400 copies/ml having a planned CS compared with those having a planned vaginal birth. The evidence for this outcome was of very low quality.

Mother-to-child transmission in women with viral load less than 1000 copies/ml on HAART

One study reported transmission rates in women on HAART with a viral load less than 1000 copies/ml. This was slightly higher for women having a planned CS but the statistical significance was not calculable. The evidence for this outcome was of very low quality.

Mother-to-child transmission in women with viral load 1000 copies/ml or more on HAART

One study did not find a statistically significant difference between transmission rates for women on HAART with viral load of 1000 copies/ml or more having a planned CS compared with those having a vaginal birth (planned or unplanned; including intrapartum CS). The evidence for this outcome was of very low quality.

Mother-to-child transmission in women with viral load 10,000 copies/ml or more on ART (including HAART)

One study did not find a statistically significant difference between transmission rates for women on ART (including HAART) with a viral load of 10,000 copies/ml or more having a planned CS compared with those having a planned vaginal birth. The evidence for this outcome was of very low quality.

Evidence to recommendations

Relative value placed on outcomes considered

The only relevant outcome for consideration in this question was the rate of mother-to-child transmission of HIV. This outcome was then split according to the viral load (number of copies per ml) reported in each study (and in some cases further subdivided according to the treatment used).

Trade-off between clinical benefits and harms

The evidence provided a number of results, both for different viral loads and different therapies. For women with a viral load of less than 50 copies/ml on HAART, there was no significant difference in mother-to-child transmission rates between women who gave birth vaginally and those who gave birth by CS. The GDG members felt that this matched their clinical experience and so were confident in recommending that women on HAART with a viral load of less than 50 copies/ml should not be offered a CS.

The group noted that one study with the majority of women on ART (Warszawski et al., 2008) did not show a statistically significant difference in the mother-to-child transmission rate between women giving birth vaginally and those giving birth by CS in women with a viral load of less than 400 copies/ml. However, the group recognised that the study did include a number of women on HAART and that this might have affected the results. As a result, the group agreed that either a vaginal birth or CS could be considered for women on ART with a viral load of 50–400 copies/ml.

The evidence also suggested that for women on HAART with a viral load of 50 copies/ml or more up to 1000 copies/ml there was no significant difference in transmission rates between women who gave birth vaginally and those who gave birth by CS. Further discussion of all the evidence pertaining to viral loads of more than 400 copies/ml led the GDG members to decide that the evidence was of too low a quality to change current practice and they felt it important to remain cautious in these instances. They agreed that women on HAART with a viral load less than 400 copies/ml should not be offered a CS on the grounds of their HIV status but that all women with a viral load of more than 400 copies/ml should be offered a CS regardless of the therapy being received.

One study that included women receiving different therapies showed a significant reduction in the risk of mother-to-child transmission with planned CS (Boer et al., 2010). The GDG felt that this difference was likely to be due to the fact that the study included women receiving no therapy. As a result, the group agreed that it was appropriate to adopt a cautious approach and recommended that women on no therapy should be offered a CS regardless of their viral load.

Trade-off between health benefits and resources

The economic modelling for CS compared with vaginal birth for women with healthy, uncomplicated pregnancy suggests vaginal birth may be more cost effective, although the evidence for this is not strong and is based upon incomplete outcome data. However, where a woman is HIV positive the balance may be tipped in favour of CS in order to reduce the risk of mother-to-child transmission of HIV and the subsequent long-term health loss and treatment costs for the baby. The point at which the balance favours CS above vaginal birth depends upon the relative risk of HIV transmission. The GDG members decided, based on the evidence reviewed plus their own experience, that this tipping point comes at viral loads of more than 400 copies/ml, regardless of therapy. The use of HAART reduces the risk of transmission and so where there is low viral load (less than 400 copies/ml) it is appropriate to advise vaginal birth.

Quality of evidence

All evidence reviewed for this question was of very low quality. This was mostly due to the studies being retrospective observational studies that were underpowered and had flaws in reporting, meaning it was not always possible to determine actual mode of birth for all women within each study group. The low quality of evidence meant the GDG remained cautious with its recommendations.

Other considerations

The GDG agreed that further UK-based data about the diagnosis of HIV in pregnant women, its treatment, the mode of birth chosen in different circumstances and mother-to-child transmission rates were all required. The GDG was aware that some of this information is currently collected by the RCOG and so included a recommendation that this data continue to be collected.

Recommendations

NumberRecommendation
24As early as possible give women with HIV information about the risks and benefits for them and their child of the HIV treatment options and mode of birth so that they can make an informed decision. [new 2011]
25Do not offer a CS on the grounds of HIV status to prevent mother-to-child transmission of HIV to:
  • women on highly active anti-retroviral therapy (HAART) with a viral load of less than 400 copies per ml or
  • women on any anti-retroviral therapy with a viral load of less than 50 copies per ml.
Inform women that in these circumstances the risk of HIV transmission is the same for a CS and a vaginal birth. [new 2011]
26Consider either a vaginal birth or a CS for women on anti-retroviral therapy (ART) with a viral load of 50–400 copies per ml because there is insufficient evidence that a CS prevents mother-to-child transmission of HIV. [new 2011]
27Offer a CS to women with HIV who:
  • are not receiving any anti-retroviral therapy or
  • are receiving any anti-retroviral therapy and have a viral load of 400 copies per ml or more. [new 2011]
28Researchers and national bodies responsible for the collection of UK population data should continue to collect data about HIV diagnoses in pregnant women, including treatment, mode of birth, and mother-to-child transmission rates. [new 2011]

Hepatitis B virus

Serological screening for hepatitis B should be offered to all pregnant women.1 The prevalence of hepatitis B surface antigen (HBsAg) in pregnant women in the UK has been found to range from 0.5 to 1%.129,130 [evidence level 3]. The wide range in prevalence rates is most likely due to wide variation in prevalence among different ethnic groups.131 [evidence level 3]

Hepatitis B immunoglobulin and hepatitis B vaccine reduce mother-to-child transmission (MTCT). The vaccine and immunoglobulin are given to the infant at birth followed by either a one month and six month dose or at 5 weekly intervals.132,133 [evidence level 1b]

Most MTCT occurs at birth or postnatally. Transmission at birth may be due to microperfusion of maternal blood into the infant's circulation during placental separation or by the infant swallowing maternal blood, amniotic fluid or vaginal secretions at vaginal birth.134 It has been suggested that CS could further reduce MTCT however no RCTs have addressed this issue. One cohort study was identified (n = 447 infants). The methodology of this study is not clearly reported and the generalisability of the findings is not clear.135 [evidence level 2a]

Recommendations

NumberRecommendation
29Mother-to-child transmission of hepatitis B can be reduced if the baby receives immunoglobulin and vaccination. In these situations pregnant women with hepatitis B should not be offered a planned CS because there is insufficient evidence that this reduces mother-to-child transmission of hepatitis B virus. [B] [2004]
NumberResearch recommendation
RR 13RCTs are needed to evaluate the effect of planned CS in addition to immunoglobulin and vaccination on MTCT of hepatitis B.

Hepatitis C virus

Women are not routinely offered screening for hepatitis C infection in the UK.1 The prevalence of hepatitis C virus (HCV) in women of child-bearing age is not known as large scale serological studies have not been done. It is however estimated that 1–2% of women of child-bearing age in the US are positive for antibody to HCV.136 An estimate for EU countries is 0.9% (0.1–3%).136

Mother-to-child transmission (MTCT) of HCV was first described in the early 1990s and is now widely recognized. The risk of MTCT of HCV is usually low at 3–5% but higher rates of 10–20% are observed among HIV co-infected women.136 [evidence level 3] A cohort study involving 441 mother-child pairs from the UK and Ireland of which 5% were known to be HIV-positive, estimated overall MTCT risk at 6.7% (95% CI 4.1 to 10.2). Women co-infected with HIV and HCV had a 3.8 times higher risk of transmitting HCV to their infants than HIV-negative women.137 [evidence level 2b]

The effect of mode of birth on the risk of MTCT of HCV has not been evaluated in RCTs. We identified a pooled retrospective analysis of prospectively collected data on 1474 HCV infected women from 36 centres in eight European countries.138 [evidence level 3] For women with hepatitis C infection, there was no difference in risk of vertical transmission by mode of birth (OR 1.19, 95% CI 0.64 to 2.20). This lack of association persisted with adjustment for breastfeeding status, geographic region and maternal age at birth (OR 1.26, 95% CI 0.68 to 2.34), (OR 1.29, 95% CI 0.69 to 2.42) and (OR 1.17, 95% CI 0.59 to 2.31).138 [evidence level 3]

Within this study subgroup analysis of women co-infected with HIV (n = 503, 35.4%), reported that the risk of vertical transmission for HCV was reduced by 60% with CS (OR 0.43, 95% CI 0.23 to 0.80). Of the HIV co-infected women, 14 (7.3%) were classified as clinical stage C, the remainder of the women are described as being asymptomatic. There is no mention of whether any of the women were on anti-retroviral therapy. Thirteen (2.6%) of the HIV co-infected women breastfed their infants.138 [evidence level 3]

Recommendations

NumberRecommendation
30Women who are infected with hepatitis C should not be offered a planned CS because this does not reduce mother-to-child transmission of the virus. [C] [2004]
31Pregnant women who are co-infected with hepatitis C virus and HIV should be offered planned CS because it reduces mother-to-child transmission of both hepatitis C virus and HIV. [C] [2004]

Genital herpes simplex virus

Genital herpes simplex virus (HSV) infection is an ulcerative sexually transmitted infection which can recur and is associated with considerable physical and psychological morbidity. Genital ulcers may cause pain but can be asymptomatic (for example cervical lesions). Between 1972 and 2001, there was a 9–fold increase in the incidence of genital HSV diagnosed in women in the UK.139 [evidence level 3] Currently HSV-2 antibody prevalence in England and Wales is 3% in men and 5% in women.140 [evidence level 3]

Neonatal HSV can cause severe systemic disease and is associated with a high mortality rate. Active surveillance in the UK suggests that neonatal HSV infection occurs in 1.65 per 100,000 live births.141 [evidence level 3] Neonatal HSV may result from contact of the newborn with the birth canal of an infected mother.

Primary HSV infection and MTCT of HSV

The accepted practice of offering CS to women with HSV infection is based on three case series. The first study included 101 pregnant women with HSV (both primary and recurrent disease). This study found the risk of neonatal herpes to be highest for women who acquired primary infection during the third trimester (3 cases of neonatal infection out of 9 cases of exposure).142 [evidence level 3] Subsequently a study evaluating screening for HSV identified 94 women who acquired HSV during pregnancy but with no MTCT to the infants. There were an additional 9 women who acquired genital HSV near the onset of labour and in this group, 4 of the 9 infants developed neonatal HSV infection.143 [evidence level 3] A study of 15,923 asymptomatic women in early labour reported isolating HSV from 56 women of whom 18 (35%) had a primary infection. Neonatal HSV developed in 6 infants (33%).144 [evidence level 3] None of the studies are large enough to address the effect of mode of birth on MTCT.

Despite limited evidence the high mortality associated with neonatal herpes means there is consensus about current practice to offer CS for primary infection.145,146

Recurrent HSV infection and history of HSV infection and MTCT

Observational data suggests that the risk of neonatal infection with recurrent HSV is lower than with primary HSV infection (8% with recurrent infection and 33% with primary HSV infection).147,148 [evidence level 3] In the Netherlands there has not been a policy of CS for women with recurrent HSV since 1987, and this practice has not resulted in an associated increase in HSV neonatal infections.149 [evidence level 3]

Recurrent HSV may not cause symptomatic lesions, for example with cervical ulceration. A study of 15923 asymptomatic women in early labour reported isolating HSV from 34 women, neonatal HSV developed in 1of the infants (3%).144 [evidence level 3] To prevent MTCT of HSV in asymptomatic women antenatal screening using HSV cultures was proposed, but this test also did not predict infants risk at birth.150 [evidence level 3]

Three RCTs evaluate using oral acyclovir from 36 weeks to prevent recurrence of HSV at the time of birth. These found a reduction in CS for HSV; however do not report the effect of acyclovir on MTCT.151153 [evidence level 1b]

A survey of obstetricians in the UK found there was no consensus of opinion or practice for recurrent disease or a history of disease.146 [evidence level 3]

Cost effectiveness of CS to prevent MTCT of HSV

Three American studies have considered the factors that promote or inhibit the cost-effectiveness of various strategies to prevent MTCT of HSV.154156 Two studies by the same author have examined the additional efficacy, risks, and costs of CS for three groups of women: those presenting with primary HSV; women with a history of HSV; and women with no clinical HSV or history of HSV. The first study was a decision analytic model using data from a review of 19 studies.154 Marginal (additional) costs and benefits over and above standard delivery were calculated.

Adopting a programme of offering routine CS for women with a history of HSV, 9 neonatal cases would be averted per million births at an estimated cost of US$2.5 million per case of neonatal HSV averted. For women with primary HSV, 18 neonatal cases prevented per million with estimated cost saving of US$38,000 per case of neonatal HSV averted.154 However more data on transmission rates and the efficacy of CS are required to make these estimates robust.154

A later study155 modelled the cost-effectiveness of four strategies to prevent MTCT of HSV in women with at least one previous episode of HSV. CS only, acyclovir prophylaxis in late pregnancy with vaginal birth, acyclovir prophylaxis in late pregnancy with screening and follow-up, and a ‘do nothing’ option. The incremental cost per case prevented compared with ‘do nothing’ was highest for CS with 2.8 cases prevented at an additional cost of US$1.3 million, and lowest for acyclovir prophylaxis with screening and follow-up of neonates (an additional cost of US$400,300). This suggests that acyclovir therapy with follow-up was a more cost-effective strategy than CS alone.

The third paper examined whether acyclovir suppression was a more cost-effective option compared to offering CS only to women with a history of HSV.156 The analysis showed that CS rate was the most sensitive variable (since it represents a high proportion of the total costs). The authors concluded that acyclovir suppression was a cost-effective alternative to CS for women with a history of genital herpes in agreement with analysis of the authors of the previous two papers. However, given the lack of data around the estimates of costs, the small sample size (46 women presenting with HSV or with a history of HSV) and the setting of the study, the findings are of limited value to this guideline.

In conclusion CS is the preferred (the most cost-effective and cost-saving) option in women presenting with primary HSV late in pregnancy. Acyclovir prophylaxis may be a more cost-effective option for women with recurrent HSV.

Recommendations

NumberRecommendation
32Women with primary genital herpes simplex virus (HSV) infection occurring in the third trimester of pregnancy should be offered planned CS because it decreases the risk of neonatal HSV infection. [C] [2004]
33Pregnant women with a recurrence of HSV at birth should be informed that there is uncertainty about the effect of planned CS in reducing the risk of neonatal HSV infection. Therefore, CS should not routinely be offered outside a research context. [C] [2004]
NumberResearch recommendation
RR 14RCTs are needed to determine whether planned CS should be offered to prevent MTCT of HSV to women with recurrence of HSV at birth and in women in whom the primary HSV infection occurs in the first trimester of pregnancy.

5.9. Maternal request for CS

Introduction

In general, CS is a safe operation, especially when performed as a planned procedure. CS rates are rising worldwide, with an increasing proportion being undertaken in response to maternal request, in contrast to those that are performed for obstetric indications. There are many reasons for such requests but these are not always revealed by the women or adequately explored and clearly documented by their carers. This chapter addresses the issue of CS requested by women who have no apparent clinical reason for requesting a CS or who report fear of giving birth vaginally.

Rates of maternal request for CS

We identified 19 observational studies that report rates of maternal request for CS. Twelve of these are included in a systematic review (n = 13285)157 and seven studies have been published since the review.4,157162 The largest of these studies were a survey of women attending antenatal clinics in Sweden (n = 3061)160 and a survey of women's views of childbirth carried out within the National Sentinel Caesarean Section Audit [NSCSA] (n = 2475).4

The rates of preference for CS expressed by the women that were surveyed during pregnancy in UK, Australia and Sweden range from 6% - 8%.4,157,158,160 [evidence level 3]

Within these studies there was a consistent relationship between women's preference for CS and either previous CS, previous negative birth experience, a complication in the current pregnancy or a fear of giving birth.4,157,160 The main reason given for preference for CS was that it was perceived to be safest for the baby. The main reason given by those who expressed a preference for vaginal birth was the experience of a natural event. One study157 concluded that maternal request for CS seems to be a marker for previously negative birth experiences and should prompt enquiries to address any issues or concerns.157 [evidence level 3]

Fear of childbirth

It is estimated that about 6%–10% of pregnant women experience fear of childbirth.163,164 [evidence level 3] Fears concerning childbirth such as pain, obstetric injury, unplanned CS, health care staff and the effects on family life have been reported to be more common among primiparous compared to multiparous women, and among those who had not attended antenatal classes.165 [evidence level 3] Fear of health care workers was reported to be more common among women who either had problems in the current pregnancy or those who were intending a planned CS.165 [evidence level 3] Manifestations of this fear included stress symptoms influencing everyday life, nightmares, a wish to have CS and a wish to avoid the current pregnancy and childbirth.165 [evidence level 3]

Fear of childbirth has been measured using different scoring systems.167 One case–control study found that women who requested planned CS due to fear of child birth were more likely to have also experienced a spontaneous miscarriage (OR 1.73, 95% CI 1.05 to 2.85), a longer time between pregnancies (OR 1.44, 95% CI 1.19 to 1.75), a longer duration of second stage of labour and a previous assisted vaginal birth (OR 4.50, 95% CI 2.18 to 9.31) or “emergency” CS (OR 26.91, 95% CI 11.86 to 61.07).166 [evidence level 3] Previous infertility, induction of labour, epidural analgesia, duration or intervention in the third stage of labour in a previous pregnancy were not found to be associated with fear of childbirth in this study.166

Another study reported that women who had “emergency” CS had higher scores for fear of childbirth during pregnancy compared to those who had vaginal births.167 However a prospective study carried out in the U.K. did not find an association between fear of child birth and ‘emergency’ CS (OR 1.00, 95% CI 0.98 to 1.01).26 [evidence level 3]

One RCT randomised women referred to an antenatal clinic for fear of child birth to receive either cognitive behavioural therapy or usual care. No difference was detected in the proportion of women who chose to deliver by CS (OR 0.82, 95% CI 0.50 to 1.36), however fewer women in the intervention group who had vaginal births reported fear of pain in labour and had shorter labours.168 [evidence level 1b]

Responding to requests for CS

Obstetricians estimate that they agree to perform a CS for about half of the requests they receive.4 [evidence level 3] A woman's request for CS is the ‘start of a continuing dialogue and process’ during which a negotiated plan of care can be developed which enables women to continue to feel in control with the support of her healthcare providers.169 [evidence level 4]

When a woman requests a CS the first response should be to determine the reason for the request and the factors that are contributing to the request. This can then be followed by the provision of information that compares the risks and benefits of planned CS and vaginal birth (see Tables 4.5 and 4.6).

Review question

What is the appropriate care pathway for women who request a primary caesarean section where there is no obstetric or medical indication?

Overview of evidence

When addressing this question for the guideline, the GDG hoped to find evidence relating to the effectiveness of interventions for providing care for women requesting a CS. Unfortunately, no such evidence was identified.

One prospective cohort study conducted in Sweden investigated postpartum outcomes in women having their first baby planning CS in the absence of medical indication compared to those planning a vaginal birth (Wiklund et al., 2007). Questionnaires were completed by the women on recruitment prior to giving birth, and at two days and three months postpartum. The outcomes recorded included their reason for the request, self-estimated health, expectations of birth and experience of delivery as well as duration of breastfeeding and time to re-establishment of sexual life. The study also had a secondary aim which was to study whether postpartum depression was more common in the group planning CS.

Evidence profile

Tables 5.5 and 5.6 summarise the evidence from this study which included 91 women who planned to have a CS and 266 consecutively recruited controls who were planning a vaginal birth.

Table 5.5. GRADE summary of findings for comparison of planned CS versus vaginal birth (maternal outcomes).

Table 5.5

GRADE summary of findings for comparison of planned CS versus vaginal birth (maternal outcomes).

Table 5.6. GRADE summary of findings for comparison of planned CS versus vaginal birth (neonatal outcomes).

Table 5.6

GRADE summary of findings for comparison of planned CS versus vaginal birth (neonatal outcomes).

Evidence statements

Maternal outcomes

The evidence for all of the following maternal outcomes was of very low quality.

Maternal hospital stay

One study found that women who had a planned CS remained in hospital for longer than women having a planned vaginal birth. This finding was statistically significant.

Birth experience

One study found that women with a planned CS reported a higher satisfaction score regarding their birth experience 2 days after birth compared with women having a planned vaginal birth and this effect remained at 3 months. These findings were both statistically significant.

Uncomplicated breastfeeding

One study did not find a statistically significant difference in breastfeeding rates at 2 days postpartum between women who had a planned vaginal birth compared with women who had a planned CS. However, the same study found that more women who had a planned vaginal birth were breastfeeding at 3 months postpartum compared with women who had a planned CS. This finding was statistically significant.

Coitus

One study did not find a statistically significant difference in the numbers of women resuming coitus at 3 months following a planned CS compared with those who had a planned vaginal birth.

Family planning

One study found that more women who had a planned vaginal birth had plans for a second child at 3 months postpartum compared to women having a planned CS. This finding was statistically significant.

Depression

One study did not find a statistically significant difference in signs of postnatal depression comparing women who had given birth by planned CS compared with those who had had a planned vaginal birth.

Neonatal outcomes

Neonatal intensive care unit care

One study found did not find a statistically significant difference in the number of neonates who received neonatal intensive care following planned CS compared with planned vaginal birth. The evidence for this outcome was of very low quality.

Health economics

A model to compare the cost effectiveness of maternal request CS versus planned vaginal birth in primiparous women without any medical or obstetric indication for CS was developed. Full details of this model are presented in Chapter 13 but a summary is provided here.

Risks for the two modes of birth were taken from a clinical review undertaken for this guideline update comparing outcomes by planned mode of birth rather than actual mode of birth (see Section 4.2). The analysis considered the costs of birth and ‘downstream’ costs associated with the outcomes reported in the clinical review and found that a planned vaginal birth was approximately £700 cheaper than a maternal request CS. This implies that the NHS could save £4.9 million for every percentage point reduction in CSs if the characteristics of the population were similar to those of women included within the guideline model. A cost utility analysis found that planned vaginal birth dominated maternal request CS.

However, there may be other outcomes, such as urinary incontinence, which were not reported in the studies that were included in the clinical review which make the findings reported above more uncertain. Sensitivity analysis suggested that this could, under certain assumptions, produce a different cost effectiveness result.

Evidence to recommendations

Relative value placed on outcomes considered

The GDG agreed that the most important outcomes to consider were women's birth experience along with women's satisfaction and experiences of care. The GDG members noted that these are difficult outcomes to measure, given the disparate reasons that women request a CS.

The GDG also felt that women's mental health was an important outcome to consider. It was acknowledged that not agreeing to a request for a CS could have a negative impact on a woman's mental health and potentially lead to a long-term need for psychological support postnatally.

The GDG noted that the length of hospital stay will not always be an important consideration for women (as they felt that women would be aware and accept that a surgical procedure will be associated with inpatient stay). However, GDG members agreed that it was important to recognise the increased cost and resources required associated with CS.

Trade-off between clinical benefits and harms

From the evidence reviewed for maternal request, the GDG noted that CS is associated with a longer hospital stay and a higher rate of women not breastfeeding at 3 months. However, the GDG weighed this against the finding that women who had a CS described a significantly better birthing experience, both immediately postnatally and 3 months after birth.

The GDG noted that the findings for breastfeeding might have been influenced by the different demographic profile of the two groups of women. Women in the planned CS group were significantly older than those in the planned vaginal birth group, more likely to have come from abroad, less likely to have received parenthood education and less likely to report their perceived health as good compared with women in the planned vaginal birth group.

The GDG members noted that the findings for depression were poorly reported and they did not feel that they were helpful. They were aware from their own experience that if some women did not receive a requested CS it could lead to poorer mental health outcomes, such as anxiety, both during and after the pregnancy, and difficulty bonding with their baby.

Trade-off between net health benefits and resource use

The GDG noted that there was likely to be an increased resource use with CS due to the increased length of hospital stay.

An economic model developed for this guideline suggested that planned vaginal birth was cost effective compared to a maternal request CS. However, this finding was limited to outcomes that were reported in the included studies for the clinical review undertaken for this guideline (see Section 4.2). A sensitivity analysis suggested that the inclusion of adverse outcomes not reported, such as urinary incontinence, could make the conclusion regarding cost effectiveness less certain. On balance, this model does not provide strong evidence to refuse a woman's request for CS on cost effectiveness grounds.

The GDG agreed that there was likely to be a cost associated with providing psychological support to those women who experience mental health problems as a result of not receiving a CS on request. However, it noted that this was only likely to be the case for a small proportion of women. The GDG's experience of caring for women requesting a CS was that anxiety about giving birth vaginally was often at the root of the request; for example as a result of a previous poor birth experience. The GDG believed that when women are given the opportunity to discuss these anxieties in a supportive environment, the anxieties can often be reduced to the point where the woman is able to choose a planned vaginal birth. The GDG agreed this was the preferred approach. It was not felt to be necessary for the person providing this psychological support to be a mental health expert unless clinically indicated, but rather that it could be provided by a member of the maternity team, such as a midwife or obstetrician. It was felt that the extra resource required to provide this support would be offset by resources saved where a request for planned CS was appropriately changed to a planned vaginal birth as a result of addressing a woman's anxieties or concerns antenatally. However, in situations where a woman persists in her request for a CS following provision of the opportunity to discuss and explore her reasons for the request, the GDG believed that the potential for psychological harm caused by denying this request was sufficient to warrant this unacceptable in terms of the woman's health: it also has the potential to be costly in terms of long-term need for psychological support. It was concluded, therefore, that if a vaginal birth is not an acceptable option to the woman after discussion and the offer of support, she should be supported in her choice of a planned CS.

The GDG was aware of instances where women had been offered referral to a perinatal mental health expert and that this expert had not been granted access to the planned place of birth. The GDG recognised that having this access is important in order to provide appropriate support and adequately address any anxieties regarding the birth setting. As a result, the group agreed that it was appropriate to recommend that the healthcare professional providing this care be given access to the planned place of birth.

Quality of the evidence

The GDG was hoping to find evidence of the effectiveness of antenatal interventions aimed at supporting women who request a CS in the absence of medical indications. Unfortunately, no such evidence was identified. The one included study compared outcomes for women requesting and receiving a CS with those who had a planned vaginal birth. This information was only marginally useful in helping the GDG to decide its recommendations.

The women in the two groups were significantly different in a number of characteristics at baseline: compared to the planned vaginal birth group, the women who had a planned CS were older, more had come from abroad and more had had IVF, although fewer reported their pregnancy was planned or that they had received parenthood education or perceived their health as good. The groups were only similar in terms of having a university education and in the number of smokers.

Analysis was not performed to assess the effects of these differences on the results obtained from the questionnaires. In addition, a sub-group/per protocol analysis was not performed to estimate outcomes separately for women who planned a vaginal birth but subsequently had an unplanned CS (n = 29, 11%) or an instrumental delivery (n = 36, 13%).

The study was conducted in a middle-to-high income urban area in Sweden and the women were highly educated. The GDG considered the results to be relevant to a UK population but noted that the study was not representative of women from a low socio-economic background. The GDG agreed that the quality of the study means that it is of limited relevance.

The GDG noted that there was no evidence comparing women who requested a CS and received one with those who wanted a CS and didn't receive one. This would have been a useful comparison.

It was also noted that there appeared to be incomplete reporting of some of the findings, such as postnatal depression, which undermined the validity of those findings.

Other considerations

The GDG was aware that some groups of women, such as women who don't speak English as a first language, can find it more difficult to express their concerns. They recognised the importance of ensuring that all women are encouraged to discuss their concerns with a healthcare professional at an early stage in pregnancy. Discussions with women requesting a CS need to sensitively explore reasons behind the request, including: previous birth trauma, women's perceptions of the risks of both vaginal birth and CS; women's perceptions of vaginal birth, including misconceptions and lack of knowledge about birth; and planning a date for giving birth and convenience.

The GDG also believed it was important for an individual obstetrician to be able to exercise their own beliefs about what is the best course of action in any given situation. Thus, if an obstetrician feels a woman's request for CS is not appropriate after the woman has received appropriate counselling and support, then the obstetrician should be able to decline to support the women's request. However, this does not overrule the woman's rights to express a preference for a CS, and in this instance the obstetrician should transfer care of the woman to an obstetrician who is happy to support the woman's choice.

Recommendations

NumberRecommendation
34When a woman requests a CS explore, discuss and record the specific reasons for the request. [new 2011]
35If a woman requests a CS when there is no other indication, discuss the overall risks and benefits of CS compared with vaginal birth (see tables 4.5 and 4.6) and record that this discussion has taken place. Include a discussion with other members of the obstetric team (including the obstetrician, midwife and anaesthetist) if necessary to explore the reasons for the request, and to ensure the woman has accurate information. [new 2011]
36When a woman requests a CS because she has anxiety about childbirth, offer referral to a healthcare professional with expertise in providing perinatal mental health support to help her address her anxiety in a supportive manner. [new 2011]
37Ensure the healthcare professional providing perinatal mental health support has access to the planned place of birth during the antenatal period in order to provide care. [new 2011]
38For women requesting a CS, if after discussion and offer of support (including perinatal mental health support for women with anxiety about childbirth), a vaginal birth is still not an acceptable option, offer a planned CS. [new 2011]
39An obstetrician unwilling to perform a CS should refer the woman to an obstetrician who will carry out the CS. [new 2011]
NumberResearch recommendation
RR 15What support or psychological interventions would be appropriate for women who have a fear of vaginal childbirth and request a CS?

Interventions for evaluation could include:
  • support from a named member of the maternity team
  • continuity of carer
  • formal counselling
  • cognitive behavioural therapy.
Outcomes could include:
  • mode of birth planned at term
  • psychological outcomes (postnatal depression, post-traumatic stress disorder, self-esteem, mother-infant bonding)
  • breastfeeding.
Why this is important

Fear of vaginal childbirth may stem from:
  • fear of damage to the maternal pelvic floor
  • damage to the baby during childbirth
  • self-doubt on the ability to physically achieve vaginal birth
  • previous childbirth experience
  • unresolved issues related to the genital area.
Currently there is a wide variation in practice and limited resources lead to limited availability of effective interventions. Interventions that may be appropriate include:
  • antenatal clinics dedicated to providing care for women with no obstetric indications who request a CS
  • referral to a psychologist or a mental health professional
  • referral to an obstetric anaesthetist
  • intensive midwifery support.
Continuity of healthcare professional support from the antenatal to the intrapartum periods and ‘one to one’ midwifery care during labour are also often lacking and may make a difference to women who are anxious or afraid.

All of these interventions have different resource implications and there is no clear evidence to suggest that any are of benefit. The proposed research would compare in a randomised controlled trial two or more of these interventions in women requesting a CS. In the absence of any evidence, there is a case for comparing these interventions with routine antenatal care (that is, no special intervention).

This research is relevant because it would help to guide the optimal use of these limited resources and future guideline recommendations.
RR 16Medium to long term quality of life study comparing psychological and physical outcomes in women who have had a requested and given birth by CS compared with women who plan a vaginal birth.
RR 17Qualitative and quantitative research should be carried out to look at the reasons that lead to pregnant women's request for CS
RR 18The effect of counselling and other interventions such as second opinion and provision of support on the likelihood of CS for women who express a preference for CS need further evaluation.

The original guideline indicated that the repeat scan should be offered at 36 weeks. This sentence has now been updated in line with the updated Antenatal Care guideline (RCOG, 2008).

Footnotes

1

The original guideline indicated that the repeat scan should be offered at 36 weeks. This sentence has now been updated in line with the updated Antenatal Care guideline (RCOG, 2008).

Copyright © 2011, National Collaborating Centre for Women's and Children's Health.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

Bookshelf ID: NBK115297
PubReader format: click here to try

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...