Mode of delivery in the early preterm infant (<28 weeks)
Prof J Drife, Department of Obstetrics and Gynaecology, University of Leeds, Clarendon Wing, Belmont Grove, Leeds LS2 9NS, UK. Email firstname.lastname@example.org
Elective caesarean section for women in labour with an immature baby might reduce the chances of fetal or neonatal death, but might also increase the risk of maternal morbidity. A review (updated in February 2004) of randomised trials comparing a policy of elective caesarean section versus expectant management with recourse to caesarean section produced six studies involving only 122 women. Differences in fetal outcome did not reach significance, but mothers undergoing elective caesarean section were more likely to have serious morbidity. Scientifically, the evidence remains inadequate. Clinically, the recommendation is that prematurity is not, in itself, an indication for caesarean section. In a survey from Israel, published in December 2004, of 2955 very low birthweight infants born at 24–34 weeks of gestation, the overall caesarean section rate was 51.7%, and the mortality rate among babies prior to discharge was lower after caesarean section (13.2 versus 21.8%). After adjustment using multiple logistic regression, caesarean section had no effect on survival except in a subgroup with amnionitis, and it was again concluded that caesarean section cannot be routinely recommended unless there are other indications. A decision model developed in the USA has compared costs and health outcomes of two options for managing labour at 24 weeks of gestation. The probabilities of both intact survival (16.8 versus 12.9%) and survival with major morbidity (39.2 versus 19.4%) are higher with willingness to perform caesarean section, but less aggressive management is the more cost-effective strategy. Large studies are few and recruitment to such studies is perceived as a major problem. For clinicians, the decision will be influenced by local circumstances.
When a woman is in spontaneous preterm labour and it becomes clear that the labour cannot be stopped, a decision needs to be made: should the obstetrician plan a vaginal delivery or should the woman undergo emergency caesarean section? Each course of action has risks and benefits, but how much does the risk/benefit balance depend on the gestation? Years ago, 28 weeks was regarded as the lower limit of neonatal viability, but nowadays, survival of babies born at less than 28 weeks is commonplace. Nevertheless, caesarean section for fetal indications at this early gestation is still controversial.
This article will summarise the evidence on the risks and benefits of caesarean section versus vaginal delivery in spontaneous preterm labour. It will ask how helpful this evidence is to the clinician and aims to shed some light on a difficult decision which obstetricians face on a regular basis. Each clinical case requires a clear answer. For an obstetrician, to do nothing is to make a decision.
In the UK, obstetric practice is increasingly influenced by official guidelines. These always include disclaimers stating that the ultimate decision-making responsibility lies with the clinician, but nonetheless they are very influential. An obstetrician who goes against them knows that he or she has to be able to justify that decision, perhaps in a court of law or more commonly to midwives and junior doctors in training, who can readily check web-based guidelines via the computer screen in the labour ward.
The most respected UK guidelines are those issued by the Royal College of Obstetricians and Gynaecologists (RCOG). The College has not yet formulated specific advice on the mode of delivery of preterm babies, but it has helped to draw up a recent national evidence-based guideline on caesarean section, known as the ‘NICE guideline’.1
The NICE guideline
‘NICE’ stands for ‘National Institute for Clinical Excellence’ (its name changed slightly in 2005, but the acronym remains the same), a body set up in 1999 to issue evidence-based guidance on clinical practice in England and Wales. Members of the public can suggest topics, and caesarean section was seen as a priority because of concern over the rapid rise in the caesarean section rate in UK, from 11.3% in 1990 to 21.3% in 2000.2
The NICE guideline, published in 2004, included the advice, ‘Do not routinely offer CS to women with preterm birth’. This reflected the general approach of the guideline, which focused on how to reduce the caesarean section rate. Perhaps because of its political overtones, the NICE guideline is not heavily relied on by obstetricians, and it is of little practical help in individual cases.
The Scottish guideline
Scotland has its own guidelines, developed by the Scottish Inter-Collegiate Guideline project.3 Obstetricians led the way, and in 1997 the Scottish Obstetric Guideline and Audit project issued quite specific guidance on preterm birth: ‘Prematurity, per se, should not be regarded as an indication for caesarean section. Vaginal delivery should be planned unless there are other indications for operative delivery’.
Obstetricians assume, usually correctly, that guidelines represent a distillation of the most up-to-date evidence. The evidence that underpins guidelines, however, is of variable quality. The clear Scottish advice quoted above, for example, was classed as ‘grade C’, which means: ‘Requires evidence from expert committee reports or opinions and/or clinical experience of respected authorities … Indicates absence of directly applicable studies of good quality’. It seems that the UK advice, albeit unanimous, may be based on inadequate evidence.
What is the evidence?
The best source of evidence from randomised controlled trials (RCTs) is the Cochrane database of systematic reviews. Grant and Glazener reviewed the subject of ‘Elective section versus expectant management for delivery of the small baby’ in 2001 and updated their review in 2006.4 Lumley, who organised one of the RCTs, also summed up the evidence in a previous contribution to this series.5 All reviewers agree that trials on this question have been difficult to organise and have had difficulties with recruitment. There have been six RCTs, three of which involved fetuses with cephalic presentation and three with breech presentation. The total number of subjects in all the six trials was 122. End points analysed by Grant and Glazener included respiratory distress syndrome, cord pH after delivery and neonatal seizures, but arguably the two most important end points were fetal or neonatal death and major maternal morbidity. Taking all six trials together, there were eight neonatal deaths—two in the group allocated to caesarean section and six in the group allocated to expectant management. The difference between the groups is not statistically significant. Looking at maternal morbidity, eight women suffered major complications after caesarean section, including two with endotoxic shock and one with acute respiratory distress syndrome. These are unusual complications but infection may underlie spontaneous preterm labour, and so these women may be at increased risk of infective complications.
Grant and Glazener concluded from their systematic review that ‘not enough studies have been done. From their limited evidence, elective caesareans may have some benefits for babies but this has to be weighed against the increased risks to the mother of complications related to the surgery’. This is a fair conclusion from the RCTs, but unfortunately it leaves the clinician managing the individual case in a familiar quandary—how to balance the benefits to this particular baby against the risks to this particular mother.
This balance will be discussed below, but first it is appropriate to ask whether further evidence from RCTs is likely to become available. Ten years ago, Penn et al. discussed the difficulties that they had encountered with a multicentre RCT comparing elective and selective caesarean section for the delivery of the preterm infant presenting by the breech.6 They commented on the difficulty of weighing the very different risks faced by the mother and by the baby, the fact that many women will accept substantial risks to themselves in the belief that their baby will benefit, the difficulty of recruiting subjects at a stressful time and the logistic difficulties of obtaining advance consent from large numbers of women, most of whom will not go into spontaneous premature labour. They also discussed problems in implementing the randomisation when women presented in labour.
Lumley described how possible solutions to some of these problems had been considered, but commented that all six trials had stopped early because of recruitment problems.5 She explained that the reason for recruitment problems in her own trial had not been women’s unwillingness to consent in advance to randomisation, but withdrawal at a later stage, against a background of rapidly rising caesarean section rates for preterm infants.5 It is worth quoting the conclusion of Penn et al., as long ago as 1996:
Despite these difficulties, we can still see no way other than through RCTs to get reliable evidence about the relative advantages and disadvantages of alternative policies for delivery in preterm labour. Nevertheless, we recognise that it is unlikely that there will ever be a successful RCT to settle the question in respect of breech presentation (although it should still be possible to mount a trial when the presentation is cephalic). Clinicians will have to decide the best mode of delivery for each woman in close consultation with her and her family, taking into account the continuing uncertainty about whether elective CS delivery really does convey benefit to the baby and if so, whether this outweighs the known risks to the mother.
Ten years on, nothing has changed. There have been no new RCTs involving either breech or cephalic presentation. In developed countries, there is increasing emphasis on women’s choice and this, coupled with unease about the quality of informed consent in trials in developing countries, means that the prospect of better evidence from RCTs is further away than ever. Rather than bemoaning this fact, we need to look harder at the evidence that is already available.
Benefits to the baby
As far as the baby is concerned, the decision on mode of delivery may be a matter of life or death. To use less dramatic medical terminology, there are indications that the route of delivery influences neonatal mortality. Survival rates for very preterm babies have improved in recent years. In 2005, a national study from Norway summarised data from 1999 to 2000.7 At less than 23 weeks of gestation, no infants survived. At 23 weeks, the survival rate was 16% and at 27 weeks it was 82%. For infants admitted to a Neonatal Intensive Care Unit the rates were higher (39% at 23 weeks of gestation and 93% at 27 weeks).
These figures will probably surprise many obstetricians. Studies have consistently shown that doctors and nurses underestimate the chance of survival of very small babies, and that obstetricians’ perception of survival rates is lower than that of paediatricians.8 Furthermore, both obstetricians and paediatricians overestimate the risk of handicap. At 25 weeks, between one-third and one-half of babies who survive will be free of handicap; however, both doctors and nurses believe that the risk of handicap at this gestation is 60–80%.8,9 These misperceptions are relevant to the decision regarding mode of delivery. If the obstetrician has the impression that the baby has little chance of survival in any case, he or she is less likely to recommend intervention in the fetal interest. It is essential for both doctors and patients to know the up-to-date rates of survival and handicap-free survival in their local area. In an effort to assist in counselling, Blanco et al. in the USA developed laminated cards summarising survival rates and the risk of long-term neurodevelopmental impairment at each gestation between 23 and 28 weeks.9
The RCTs of elective caesarean section versus expectant management for delivery of the small baby suggest that neonatal death is less likely after caesarean section than after vaginal delivery, although the difference did not reach statistical significance. A similar trend was seen in a non-randomised survey from Israel, published in 2004, of 2955 very low birthweight infants born at 24–34 weeks of gestation.10 The overall caesarean section rate in this series was 51.7%, and the mortality rate among babies prior to discharge was lower after caesarean section (13.2 versus 21.8%). After adjustment using multiple logistic regression, caesarean section had no effect on survival except in a subgroup with amnionitis, and the authors also concluded that caesarean section cannot be routinely recommended unless there are other indications.10
There is a consistent trend towards reduced mortality after caesarean section. The same does not apply to morbidity, and analysis of the RCTs in the Cochrane database by Grant and Glazener showed no difference between the groups in relation to abnormal follow up in childhood.4
The trend towards reduced neonatal mortality was seen with both breech presentation and cephalic presentation. Below 28 weeks of gestation, a high proportion of deliveries involve breech presentation. At term, breech presentation involves increased risk to the baby, and it is logical to ask whether this applies to the early preterm infant, and whether this group of babies is more likely to benefit from caesarean section.
Fetal survival in breech presentation
One of the best-known RCTs in obstetrics is the ‘Term Breech Trial’, conducted between 1997 and 2000 in 121 centres in 26 countries and involving 2083 women.11 The trial had to be stopped early because a clear difference emerged between women randomised to planned caesarean section and those randomised to planned vaginal delivery. Perinatal and neonatal mortality and serious neonatal morbidity were significantly lower in the caesarean section group. Only three babies allocated to caesarean section died, one of those after a difficult vaginal delivery, but 13 babies allocated to vaginal delivery died. The risk of death does not seem to have been related to fetal size, as the weight of the 13 babies ranged from 1150 to 3650 g, with an average of 2730 g. By contrast, babies weighing more than 4 kg were overrepresented among those who suffered serious neonatal morbidity.
The results of the Term Breech Trial have changed clinical practice in the UK, but it has been emphasised that they apply only to breech presentation at term and should not be extrapolated to breech presentation before term. Why not? The scientific answer is that only women at term were recruited to the trial and therefore a separate RCT needs to be carried out involving breech presentation before term. But such a trial will never be carried out. There is no indication that the increased survival after caesarean section in the Term Breech Trial was seen only among large babies at risk of cephalopelvic disproportion. The results of the Term Breech Trial reflect the trend seen in the RCTs of preterm birth. We are doing a disservice to our patients by ignoring this obvious fact.
Risks to the mother
In the Term Breech Trial, there were no differences in maternal mortality or serious maternal morbidity between the women randomised to planned caesarean section and those randomised to planned vaginal delivery. This result differs from the trials on spontaneous preterm labour. This may be because spontaneous preterm labour involves more risk to the mother, for example, due to infection. Alternatively, it could be because maternal risks of caesarean section have been falling in recent years, while neonatal survival rates have been increasing.
National data on maternal morbidity are lacking. The UK National Sentinel Audit of caesarean section did not examine complication rates.2 Data on mortality are available from the long-running Confidential Enquiry into Maternal Deaths, which examines all maternal deaths in the UK and produces a detailed report every 3 years. The most recent report, covering the period 2000–2002, was published in 2004 and shows that the risks of fatal complications from caesarean section have indeed fallen in recent years.12 Maternal deaths from anaesthesia steadily fell during several decades until 1994–1996, when only one occurred in the UK. The number then increased, and in 2000–2002 there were six maternal deaths from this cause, all associated with general anaesthesia. With planned caesarean section using regional anaesthesia, the anaesthetic mortality is now effectively zero, but there are still risks when general anaesthesia is given in an emergency by a relatively inexperienced anaesthetist.12 Intraoperative haemorrhage caused or contributed to only two maternal deaths over the 9-year period from 1994 to 2002, although a past history of caesarean section was a feature of deaths from placenta praevia.
The risk of infection, which may already be increased in spontaneous preterm labour,13 is said to be increased 20-fold by caesarean section. Routine antibiotic prophylaxis, however, reduces the risk of endometritis and wound infection by two-thirds to three-quarters in both elective and emergency caesarean sections. Fatal infection after caesarean section in the UK is rare: in 2000–2002 three women died of sepsis after emergency caesarean section, but two of these were already ill from sepsis before the operation.12 RCOG guidelines on thromboprophylaxis have reduced the risk of fatal pulmonary embolism to a level similar to that following vaginal delivery.
More worrying but harder to define are the long-term risks. Despite the promotion of vaginal birth after caesarean section, there is an increased risk of caesarean section in future pregnancies. A past history of caesarean section increases the risks of uterine rupture and placenta accreta in subsequent pregnancy. In 2000–2002 in the UK, there were four deaths from placenta praevia and all four women had had at least one previous caesarean section. Nevertheless, these risks need to be kept in proportion; family size has fallen and in 2000–2002 only 3% of UK births were to women having four or more children.12
Balancing risk and benefit
In balancing risk and benefit, the obstetrician must be aware of the up-to-date evidence, and also of any bias that may affect its interpretation. The current effort in the UK to reverse the rise in caesarean section rates has united professionals and lay people and has made it more difficult to argue in favour of caesarean section for the preterm baby, despite the trend in all studies showing that caesarean section improves neonatal survival. We should bear in mind that lack of conclusive evidence in favour of caesarean section does not mean we have clear evidence in favour of vaginal delivery.
Another factor in the risk/benefit analysis is cost. A decision model developed by Cazan-London et al. in 2005 in the USA compared costs and health outcomes of two options for managing labour at 24 weeks of gestation.14 In this model, willingness to perform caesarean section not only increased the probability of intact survival (16.8 versus 12.9%) but also increased the probability of survival with major morbidity (39.2 versus 19.4%). Willingness to perform caesarean section was therefore associated with considerable cost in terms of lifetime care for handicapped survivors.14 For a family with a handicapped child, the cost may not only be financial. Cazan-London et al. concluded: ‘…both absolute survival and long-term handicaps must be considered when proceeding with aggressive intervention in extremely preterm pregnancies… Physicians must strive to present an objective and balanced discussion of the risks and benefits. The long-term implications of decisions made in often urgent circumstances must still be clearly explained to and understood by the patient and her family’.14 The patient must also agree with the decision. Talking to the patient and her family often helps the decision-making process and is a job for a senior clinician.
In practice, UK obstetricians appear to favour caesarean section for preterm birth after 28 weeks of gestation. In the UK in 2000–2001, the caesarean section rate for infants weighing less than 1500 g at 28–32 weeks was 69%, but the rate at less than 28 weeks was 25%.2 In a recent book on spontaneous preterm labour, Murphy and Twaddle sum up as follows: ‘There is a consensus now that caesarean section before 25 weeks of gestation confers little benefit in terms of survival, even for breech presentation. A second opinion is advisable if the parents disagree strongly. For breech infants with an estimated fetal weight of 1.5 kg and above, delivery by caesarean section is usually recommended. These issues need to be discussed sensitively with the parents who may be disappointed that their ‘right’ to labour and deliver normally has been taken away relatively early in pregnancy’.15 The authors add that the neonatal team should be encouraged to contribute to the decision, although it remains primarily one for the obstetrician.