Introduction

Caesarean section is the most common major surgical procedure. Except for a few countries in Scandinavia, caesarean sections rates have been steadily increasing worldwide over the past two decades. In many countries, almost one in three babies are born through caesarean section. In any critical discourse of caesarean section, it is best to examine the role of caesarean delivery in each of the 10 independent groups as classified by Robson. In this review, we evaluate the role of caesarean section in three specific Robson groups: for preterm births (group 10); in the delivery of singleton breech presentation (group 6 and 7); and twin pregnancies (group 8). These three groups account for 12–18% of all births.

Preterm delivery and caesarean section (Robson group 10)

Preterm birth is associated with significant morbidity and even mortality for the newborn, and the risks increase with decreasing gestational age. Worldwide, incidence of preterm delivery varies between 6 and 11% of all births. It has been long debated whether the mode of delivery alters the risks for the singleton preterm baby, both in cephalic and in breech presentation. Caesarean delivery has a theoretical advantage over vaginal delivery in premature infants. This benefit may be the result of the avoidance of prolonged labour and allows a less traumatic birth. Any benefit, however, to the newborn needs to be carefully weighed against the facts that preterm caesarean section can be technically difficult and may require carrying out a classical caesarean section with adverse risks, such as scar dehiscence, in future pregnancy. At our own centre, 34.5% of all births in Robson Group 10 are currently delivered by caesarean section. (Chong C, Su LL, Biswas A. Changing trends of caesarean section births in a tertiary teaching hospital, unpublished data).

For medically indicated preterm deliveries, an element of urgency is often present, either for maternal or fetal reasons. Understandably, in such situations, caesarean delivery is often the method of choice. Whether caesarean delivery provides any benefit to the baby in cephalic presentation in spontaneous preterm labour is debatable. Compared with vaginal delivery, caesarean delivery certainly increases maternal risks. In addition, preterm caesarean section may be technically difficult to carry out. Lower segment is usually not well formed, and a low transverse incision may not be possible. Similarly, preterm caesarean delivery may not be as atraumatic to the fetus as is often believed, particularly when the fetus is in malpresentation. A recent systematic review of all randomised-controlled trials (RCTs) compared a policy of planned immediate caesarean delivery with planned vaginal delivery for preterm birth. Only four studies involving only 116 infants could be used for analysis. The review failed to show any significant difference in birth asphyxia or birth injuries between the two groups. The review concluded that insufficient evidence is available to evaluate the use of a policy of planned immediate caesarean delivery for preterm babies. A similar systematic review conducted 11 years earlier was exactly similar, reviewing the same studies and with similar conclusion. It concluded that insufficient RCTs have been conducted to provide adequate evidence, and recruitment for such trials is difficult. It is unlikely that new large RCTs will be conducted in the near future to answer this question. Almost, all RCTs had to be stopped early because of recruitment problems, which included refusal to give consent before randomisation and withdrawal after randomisation. As such, the current policy for preterm deliveries has to be based on the best available evidence from descriptive studies and retrospective cohort studies.

Preterm delivery and caesarean section (Robson group 10)

Preterm birth is associated with significant morbidity and even mortality for the newborn, and the risks increase with decreasing gestational age. Worldwide, incidence of preterm delivery varies between 6 and 11% of all births. It has been long debated whether the mode of delivery alters the risks for the singleton preterm baby, both in cephalic and in breech presentation. Caesarean delivery has a theoretical advantage over vaginal delivery in premature infants. This benefit may be the result of the avoidance of prolonged labour and allows a less traumatic birth. Any benefit, however, to the newborn needs to be carefully weighed against the facts that preterm caesarean section can be technically difficult and may require carrying out a classical caesarean section with adverse risks, such as scar dehiscence, in future pregnancy. At our own centre, 34.5% of all births in Robson Group 10 are currently delivered by caesarean section. (Chong C, Su LL, Biswas A. Changing trends of caesarean section births in a tertiary teaching hospital, unpublished data).

For medically indicated preterm deliveries, an element of urgency is often present, either for maternal or fetal reasons. Understandably, in such situations, caesarean delivery is often the method of choice. Whether caesarean delivery provides any benefit to the baby in cephalic presentation in spontaneous preterm labour is debatable. Compared with vaginal delivery, caesarean delivery certainly increases maternal risks. In addition, preterm caesarean section may be technically difficult to carry out. Lower segment is usually not well formed, and a low transverse incision may not be possible. Similarly, preterm caesarean delivery may not be as atraumatic to the fetus as is often believed, particularly when the fetus is in malpresentation. A recent systematic review of all randomised-controlled trials (RCTs) compared a policy of planned immediate caesarean delivery with planned vaginal delivery for preterm birth. Only four studies involving only 116 infants could be used for analysis. The review failed to show any significant difference in birth asphyxia or birth injuries between the two groups. The review concluded that insufficient evidence is available to evaluate the use of a policy of planned immediate caesarean delivery for preterm babies. A similar systematic review conducted 11 years earlier was exactly similar, reviewing the same studies and with similar conclusion. It concluded that insufficient RCTs have been conducted to provide adequate evidence, and recruitment for such trials is difficult. It is unlikely that new large RCTs will be conducted in the near future to answer this question. Almost, all RCTs had to be stopped early because of recruitment problems, which included refusal to give consent before randomisation and withdrawal after randomisation. As such, the current policy for preterm deliveries has to be based on the best available evidence from descriptive studies and retrospective cohort studies.

Preterm caesarean delivery and neonatal survival

A number of studies comparing elective caesarean delivery with expectant management for delivery of the small baby suggest that neonatal death is less likely after caesarean delivery than after vaginal delivery. In an Israeli survey of 2955 very low birth-weight infants born at 24–34 weeks of gestation, the mortality rate among babies before discharge was lower after caesarean section (13.2 v 21.8%). The overall caesarean section rate in this study was 51.7%. After corrections using multiple regression analysis, however, the investigators could not find any significant survival benefit in the women who underwent caesarean section, except for a subgroup with chorio-amnionitis.

The most vulnerable group of preterm babies is the very early preterm babies born before 26 weeks gestation. In a US review of all very early preterm births, Malloy concluded that caesarean section does seem to provide survival advantages for the most immature infants delivered at 22–25 weeks of gestation, independent of maternal risk factors for caesarean section. Adjusted odds ratios showed significantly reduced risk of neonatal death for infants delivered through caesarean section at 22–25 weeks of gestation (adjusted odds ratios of 0.58, 0.52, 0.72, and 0.81 for 22, 23, 24, and 25 weeks, respectively). A similar conclusion has been reached in a few other retrospective reviews of very early preterm deliveries.

The survival benefit of caesarean delivery for very early preterm babies (<26 weeks) does not seem to extend to the intermediate (32–33 weeks) and late (34–36 weeks) preterm babies. Malloy analysed the data from the US Linked Birth and Infant Death Certificate files from the years 2000–2003. The results of his review suggest that, for low-risk preterm infants at 32–36 weeks’ gestation, independent of any reported risk factors, primary caesarean section may pose an increased risk of neonatal mortality and morbidity. After adjustment for confounders, the adjusted odds ratios (95% CI) for neonatal mortality at gestational ages of 32, 33, 34, 35, and 36 weeks were 1.69 (1.31 to 2.20); 1.79 (1.40 to 2.29); 1.08 (0.83 to 1.40); 2.31 (1.78 to 3.00); and 1.98 (1.50 to 2.62), respectively.

Preterm caesarean delivery and neurodevelopment

Periventricular leukomalacia (PVL) and intraventricular haemorrhage (IVH) are commonly described in the early preterm infant and are often associated with impaired neurological development in the survivors. It has been reported that caesarean delivery decreases the risk of IVH in early preterm birth. The results from recent studies, however, are conflicting. In a recent report from Italy, a retrospective cohort study was conducted to confirm the correlation between the occurrence of IVH and the mode of delivery. They included all newborns with gestational age 28 weeks or less admitted to the neonatal intensive care unit of a tertiary hospital in Florence. They found that vaginally born infants had a higher rate of each grade of IVH. The increase was statistically significant only for grade 3 IVH (18% v 2%; P < 0.0001) and all grades IVH (45% v 20%; P < 0.0001). They concluded that caesarean delivery decreases the risk of developing IVH in extremely preterm infants, including the most severe grades of IVH. Similarly, another study on 397 neonates weighing less than 1251 g found that vaginal delivery was associated with higher risk of PVL and severe grade IVH, especially for babies weighing less than 750 g. A similar previous study from the USA investigated the association between delivery mode and severe IVH (grade 3–4) in singleton, vertex presenting, very low birth weight (1500 g or less) live born infants. They found that the risks for severe IVH were not influenced by the mode of delivery in this group of neonates after controlling for gestational age.

Although it is possible that the incidence of PVL and severe-grade IVH are increased in vaginally delivered extremely preterm neonates, studies on longer term neurodevelopmental outcome have failed to show any significant effect of the mode of delivery. A study from the UK included all infants weighing less than 1250 g born over a 9-year period and followed up at 2 years of age for assessment of the neurodevelopmental status by an independent paediatrician. Out of the 213 infants analysed, 103 were vaginally delivered and 110 by caesarean section. The investigators did not find any significant difference in the overall incidence of neurodisability in the infants born by caesarean section compared with those delivered vaginally. A large retrospective cohort study of extremely-low-birth-weight infants (401–1000 g) who were born by caesarean delivery and cared for in the National Institute for Child Health and Human Development Neonatal Network, during 1995 to 1997, looked at the incidence of severe IVH and PVL in the neonatal period and neurodevelopmental outcome at 18–22 months of age. Complete follow-up data set were available for 1273 infants, of which 667 were born without labour and 606 were born with labour. By univariate analysis, infants who were born by caesarean delivery with labour had a higher incidence of grade 3–4 IVH (23.3% v 12.1%; P < 0.001), PVL (8.5% v 4.7%; P < 0.02), and neurodevelopmental impairment (41.7% v 34.6%; P < 0.02). After logistic regression analysis that controlled for all maternal and neonatal demographic and clinical variables, however, the significant differences in grade 3–4 IVH, PVL, and neurodevelopmental impairment were no longer evident. The authors concluded that labour does not seem to play a significant role in adverse neonatal outcomes and neurodevelopmental impairment at 18–22 months of corrected age.

Delivery of extremely preterm infants, less than 26 weeks’ gestation, by caesarean section has increased significantly in many countries, including the USA, over the past decade. Current data do not support or justify this increase. The appropriateness of intervening with caesarean sections for these very immature infants needs further evaluation and cannot be supported at present. We should bear in mind, however, that lack of conclusive evidence in favour of caesarean section does not mean we have clear evidence in favour of vaginal delivery.

Breech presentation and caesarean delivery (Robson group 6 and 7)

Obstetricians have long debated the role of caesarean delivery for breech infants as a potentially safer mode of delivery. The results of the Term Breech Trial, published in 2000, finally gave solid ‘evidence base’ to this long-held opinion. This was a large multicentric, randomised-controlled trial of planned caesarean section versus planned vaginal breech delivery. Perinatal mortality, neonatal mortality, or serious neonatal morbidity was lower for the planned caesarean delivery group (17 out of 1039) than for the planned vaginal delivery group (52 out of 1039) (RR 0.33, 95% CI 0.19 to 0.56). Findings were also consistent for mortality alone (0.23, 0.07 to 0.81) and serious neonatal morbidity (0.36, 0.19 to 0.65). In countries with a low perinatal mortality rate, the trial showed no difference in perinatal mortality between a planned caesarean section and a trial of vaginal breech delivery, but a striking difference in ‘serious’ short-term neonatal morbidity (0.4% v 5.1%). This resulted in a major shift in guidelines and clinical practice regarding breech birth. In many countries, it has pushed assisted vaginal delivery of the breech-presenting fetus to the history books. Since the publication of the Term Breech Trial results, almost all breech presenting babies are being delivered by Term Breech Trial and, as a result, training opportunity and expertise in assisted vaginal delivery is fast disappearing. In many countries of the world, women with a breech fetus are no longer given the option of a vaginal breech delivery.

In September 2004, the investigators of the Term Breech Trial published a 2-year follow-up of children from the study and reported no difference between the two groups in morbidity or mortality. In another article, they reported that maternal outcomes at 2 years postpartum are similar after planned caesarean section and planned vaginal birth for the singleton breech fetus at term. These reports have shaken the very premise of the earlier conclusions drawn from the Term Breech Trial. Since then, the methodology and implications of the Term Breech Trial have been questioned. It is being asked whether the findings of Term Breech Trial have been misinterpreted, and whether this has led to a premature change in clinical practice. The Term Breech Trial has been mainly criticised in the following three areas: (1) inadequate case selection and intrapartum management; (2) maternity units with markedly different skill levels grouped together; and (3) short-term morbidity used as a surrogate marker for long-term neurological impairment. It is becoming obvious now that the Term Breech Trial had a number of serious flaws, and the results of this trial should no longer be considered in formulation of any guideline on breech birth. The trial even included fetuses that were already dead in-utero before labour started. A systematic review of randomised trials comparing a policy of intended caesarean section with a policy of intended vaginal breech delivery included three trials with 2396 participants. A total of 2088 participants of this review were contributed by the Term Breech Trial and, as expected, the results of this review were similar to the results of the Term Breech Trial.

In 2006, the results of a large prospective descriptive study of more than 8000 breech births in 174 centres in France and Belgium were published (PREMODA study). Trial of labour was undertaken in 31% and 71% of these women ( n = 1796) delivered vaginally. Unlike the practice in many countries, radiologic pelvimetry was used to ascertain pelvic adequacy in 82% of cases. No difference was reported in perinatal mortality (0.08% v 0.15%) or serious neonatal morbidity (1.6% v 1.45%) between a trial of labour and planned caesarean section. The only difference in outcome was a 0.16% incidence of 5-min Apgar score less than 4 in the trial of labour group compared with 0.02% in the planned caesarean section group. Although this was not an RCT, the results of the PREMODA trial gives a fair estimate of the outcome of well-conducted and well-monitored vaginal breech delivery in well-selected cases. The results of this large prospective descriptive study should also compel us to re-evaluate the role of radiologic pelvimetry in appropriate case selection.

In counselling women with breech presentation at term, the following two important facts should be taken into account. Firstly, the 2-year follow-up data of the Term Breech Tria1 shows that women had a 97% chance of having a neurologically normal 2-year-old, regardless of planned mode of birth. Those randomised to a trial of labour had a 6% absolute lower chance (or 30% relative risk reduction) of having a 2-year-old child with unspecified medical problems, suggesting some lasting benefit of labour to the newborn immune system. Secondly, the benefits, if any, of elective caesarean delivery for the child should be carefully weighed against the various short-term and long-term consequences of caesarean delivery for the mother. It has been estimated that, for every infant potentially saved by a caesarean section, one woman will experience a uterine rupture during a subsequent pregnancy if vaginal delivery is attempted. A study from the Netherlands estimated that, in the 4 years after publication of the Term Breech Trial, the increase of about 8500 elective caesarean sections probably prevented 19 perinatal deaths. It also resulted in four maternal deaths, however, that may have been avoidable. It is also estimated that, in future pregnancies, nine perinatal deaths can be expected as a result of the uterine scar, and 140 women will have potentially life-threatening complications from the uterine scar.

Caesarean section should no longer be an automatic management option for all breech presentation at term. Vaginal breech delivery should again be offered in well-selected cases. In fact, in 2006, the Royal College of Obstetricians and Gynaecologists and the American College of Obstetrics and Gynecology replaced their restrictive 2001 breech guidelines with new versions supportive of selected vaginal breech birth. This should, however, only be made available where an obstetrician experienced in vaginal breech delivery is available. This is unlikely to happen in the near future, unless corrective affirmative actions are taken soon.

Preterm breech and caesarean delivery

Even before the publication of the results of the Term Breech Trial in 2000, many obstetricians offered planned caesarean delivery for the preterm fetus in breech presentation. Such a policy was often based on the perceived additional risk of vaginal breech delivery of the preterm fetus, namely, the risk of entrapment of the head by an incompletely dilated cervix, increased vulnerability of the preterm ‘after coming head’ to distortional trauma and the increased risk of cord prolapse in labour. Since the publication of the Term Breech Trial, in many countries, it has become routine to advocate caesarean delivery for breech birth for both term and the preterm fetus. Although this is questionable, extending the verdict to the preterm breech is definitely unjust.

Only one prospective randomised study has been conducted on the preterm breech, an RCT of only 38 women in preterm labour (28–36 weeks) with breech presentations. Of these, 20 were randomised to attempt vaginal delivery and 18 were randomised to caesarean delivery. Of the vaginal delivery group, 25% of women underwent caesarean delivery for non-reassuring fetal heart rate tracing. Five neonatal deaths occurred in the vaginal group, and one neonatal death occurred in the caesarean delivery group. Because of small sample size, however, it is difficult to reach definitive conclusions about the safety of vaginal breech delivery for the preterm breech.

A large number of retrospective studies have been conducted. Probably the largest retrospective cohort study comes from California, in which all preterm breech births ( n = 14417) over a 10-year period from 1991 to 1999 were analysed. Vaginal breech delivery of low birth weight newborns in nulliparous women was associated with significantly increased neonatal mortality in newborns weighing 500–1000 g (OR 11.7; 95% CI 7.9 to 17.2); 1001–1500 g (OR 17.0; 95% CI 6.8 to 42.7); 1501–2000 g (OR 7.2; 95% CI 2.4 to 21.4); and 2001–2500 g (OR 6.6; 95% CI 2.1 to 21.2) compared with breech delivery by caesarean in nulliparous women. Birth trauma was also increased in vaginal breech birth of newborns weighing 1500–2000 g (OR 3.8; 95% CI 1.4 to 10.1) and 2001–2500 g (OR 4.5; 95% CI 2.6, 7.9) compared with caesarean breech delivery in nulliparous women. The investigators concluded that caesarean delivery may be a safer route of delivery for preterm breech fetuses. It is difficult, however, to draw meaningful conclusions from such population-based data in which only 14% of preterm breech births were vaginal delivery. It is likely that these preterm vaginal deliveries would be subject to selection bias when the general trend is to carry out caesarean section for preterm breech birth.

In a Canadian retrospective study, neonates delivered as breech and admitted to participating neonatal intensive care units in the Canadian Neonatal Network between 2003 and 2007 were included. Similar to the Californian retrospective study, of 3552 preterm breech infants, 2937 (83%) were delivered by caesarean section and 615 (17%) by vaginal delivery. Multivariate regression analysis with adjustment for perinatal risk factors showed that vaginal delivery was associated with an increased risk of death (OR 1.7; 95% CI 1.3 to 2.3). Similar conclusions were also reached in another large retrospective analysis from the USA.

A recent Turkish retrospective study compared neonatal outcome in 1537 women with breech presentation. Of these, 478 women had vaginal breech birth. The investigators found that neonatal mortality was significantly increased only in the vaginally delivered breech neonates weighing between 1000 and 1500 g compared with those delivered by caesarean section. A French retrospective study comparing planned vaginal delivery with planned caesarean delivery for breech presentation between 26 and 29 weeks, found that entrapment of the aftercoming head occurred twice as commonly in the vaginal delivery group.

It seems from these retrospective studies that vaginal breech delivery of preterm infants, at least for those weighing between 750 and 1500 g (26–32 weeks) is associated with increased neonatal mortality. Less certain benefit was shown with caesarean delivery if the fetal weight was heavier than 1500 g (roughly over 32 weeks). This conclusion, however, is based on retrospective data that are likely flawed by selection bias. Appropriate answers can only be obtained from well-conducted RCTs. Unfortunately, such an RCT is unlikely to be undertaken in the near future. A UK multicentre RCT to determine the optimum mode of delivery for women in preterm breech labour between 26 and 32 weeks gestation, had to be terminated after 17 months because of low recruitment. The Maternal-Fetal Medicine Units Network of the National Institute of Child Health and Human Development also considered a multicentre, randomised clinical trial of attempted vaginal delivery compared with elective caesarean delivery for 24–28-week breech fetuses. The study, however, was never started because of anticipated difficulty with recruitment, inadequate numbers to show statistically significant differences, and medical–legal concerns.