Commentary on ‘A multicentre randomised controlled trial comparing elective and selective caesarean section for the delivery of the preterm breech infant’

The optimal mode of delivery of the preterm infant presenting in the breech position remains an unanswered question for obstetricians providing intrapartum care, and practice varies according to clinician and patient preference. In 1989 Penn et al. designed and began a multicentre randomised controlled trial comparing vaginal delivery with caesarean section for preterm breech infants. Unfortunately recruitment fell way short of targets, and the trial was called to a halt after 13 recruits in 27 months. The low incidence of the condition, lack of clinician equipoise and ethical issues of taking truly informed consent from laboring women are all to blame. Yet, the intrapartum period is a time when good quality trials are vital, and evidence is lacking. Are we any closer to facilitating research in this environment? The obstetric labour ward setting provides a unique and challenging environment in which rapid clinical judgement and decision-making skills must be deployed to ensure a safe outcome for both baby and mother. The safety and optimal outcome for one is often linked to, but not always directly related to, the other and conflict may arise. The natural desire of mothers to put their babies health before their own can be hard for clinicians to ignore, nor should they, but the anticipated morbidity and mortality for both mother and baby must be carefully weighed and considered in the often urgent actions we take at delivery of these infants. Underpinning these decisions should be the conscientious use of current research and evidence to make decisions about care. In 1989, Penn et al.1 designed a randomised controlled trial aiming to address one such obstetric unknown; the optimal mode of delivery of the preterm infant presenting in the breech position. At the time of trial conception, common practice dictated that the premature infant should be delivered via caesarean section. The surgery is associated with risk of complications for the mother, particularly at early gestations when the lower uterine segment may not be well formed and the classical technique may be required, as well as carrying risk to future pregnancies, but it is unclear whether caesarean or vaginal delivery is safest for the preterm baby. Before study conception, existing practice was influenced by observational retrospective data subject to significant bias and particularly by clinician ‘preference’. For example, the preterm infant born vaginally may have presented in more advanced labour when the benefit of corticosteroids may be lost, or be at extremes of viability such that a clinician may be disinclined to perform a caesarean section, falsely suggesting that vaginal delivery is dangerous. Alternatively, fetal distress in labour may result in both a caesarean section, and an infant born in poor condition, therefore favouring vaginal delivery. Nevertheless, at the time this study was considered, it was common practice to perform a caesarean section for these premature infants with the aim of avoiding presumed neonatal morbidity (to avoid the ‘stress’ of labour and risk of cervical head entrapment) despite the likely increased maternal morbidity. On this background in 1975 the Royal College of Obstetricians and Gynaecologists (RCOG) recommended a randomised controlled trial to address this issue, taken up by Penn et al. in 1989. Women in labour between 26 and 32 weeks of gestation with a singleton live pregnancy in breech presentation and no clear indication for vaginal or caesarean delivery were allocated randomly to ‘intention to deliver vaginally’ or ‘intention to deliver by caesarean section’, designed specifically to avoid the selection bias inherent in analysis according to actual mode of delivery. Twenty-six centres agreed to participate and 22 obtained timely ethics approval. Women were approached antenatally to provide consent, which was confirmed during labour if they met eligibility for randomisation. Unfortunately, after 27 months of active recruitment, and only 13 women randomised (five in the caesarean group and eight in the vaginal delivery group) from six hospitals, the study was halted because of insufficient recruitment. Limited conclusions could be made from these small numbers, although the authors note that maternal morbidity was largely restricted to those women who received caesarean section (although not all caesareans were in the group allocated to caesarean), and that neonatal morbidity is common in infants born at these early gestations regardless of mode of delivery. Given the overwhelming need for evidence to support one mode of delivery over the other (particularly as the default was interventional), the initial backing by multiple institutions, and adequate funding and infrastructural support, why did this trial fail? It is difficult to design and coordinate clinical trials for a rare condition. The incidence of preterm breech labour was estimated by the authors as 0.5%, confirmed by a survey of the participating hospitals once the trial had ceased. The pilot trial suggested that 50% of women approached would consent to participate (this rate was confirmed in the main trial). Though the authors do not cite their intended recruitment target, to sufficiently power a trial to produce real and reproducible results, sustained, collaborative multi-site recruitment would be essential. Yet, despite 22 institutions agreeing to recruit and obtaining the appropriate ethical approval, recruits came from six sites only. The difficulty in recruiting to such a trial may lie in the unique competing interests between the mother and baby. Caesarean section is known to increase maternal morbidity, and, despite a lack of convincing evidence, may reduce risk to the fetus. Given that caesarean section had become a widely accepted practice in these circumstances, albeit on the basis of flawed or biased data, it is questionable whether clinicians felt in true equipoise regarding mode of delivery; without commitment to the trial from the attending clinician, recruitment was doomed to fail. Clinicians are often reluctant to admit clinical uncertainty, particularly in times of stress or great emotion, feeling that it may undermine their autonomy or patient–doctor relationship. Furthermore, a mother may accept chance of personal risk to herself in the hope that it may help her infant. Finally, unless the attending clinician feels skilled in performing a vaginal breech delivery (a skill which is difficult to maintain given that term breech babies were increasingly delivered by elective caesarean) recruitment will be low. The failure of this trial has called the methods of recruitment into obstetric trials into question. Principles of Good Clinical Practice state that potential trial candidates must have enough time to consider participation in a trial. For this reason in the Penn et al. study, many women were approached and gave consent in the antenatal period (although their chance of eventual eligibility was only one in 200), a time-consuming process for small returns. Consent was then confirmed, and randomisation took place if preterm breech presentation in labour occurred. The remainder were approached in labour. In general, recruiting women in labour, particularly in second stage when complications arise and fetal wellbeing is of a greater concern, poses ethical dilemmas as the woman is enduring the contractions of labour and may be anxious or drowsy after analgesia. It has been questioned whether women under these circumstances may feel coerced into providing consent,2 and research clinicians may have been wary of this. Furthermore it was claimed that many eligible women who consented antenatally were not re-approached when in labour and were lost to the study. The authors reflect as to whether this two-stage consent process was necessary and if so, whether post-randomisation consent would have been more appropriate. Unfortunately, 15 years on, we are still no closer to a definitive answer as to the optimal mode of delivery of these infants. A recent systematic review3 (n = 3557) concluded that neonatal morbidity associated with vaginal preterm breech delivery was greater than that of caesarean section, but these were nonrandomised studies. A Cochrane review of four randomised controlled trials involving 116 women4 concluded that there is not enough evidence to evaluate the use of a policy of planned immediate caesarean delivery for preterm babies; as clinicians we are none the wiser. The issues highlighted by this failed trial are not unique in obstetrics, and yet the intrapartum period is a time in which good quality trials are vital but notoriously lacking. Women are increasingly being asked to participate in research during their pregnancy. Distributing study information and obtaining informed consent for unpredictable conditions that develop acutely creates challenges for researchers, particularly in determining appropriate timing. Other specialties dealing with incapacitated patients where time is limited (such as head injury) often allow for consent to be gained retrospectively.5, 6 Whether it is appropriate to use such etiquette when recruiting during complicated labour is questionable. Two studies have reported that parturient women have the capacity to provide consent and are no less able than any other member of the patient population.7, 8 However, as both studies reported on consenting to analgesia, the generalisability of these findings is uncertain. The RCOG has since produced guidance (Advice No. 6a)9 that provides a framework for research that necessitates obtaining consent during labour. The guidance encourages investigators to abide by eight general principles, summarised in Box 1. They highlight that consent during labour is reasonable for certain situations, so long as appropriate time is available for discussion. They suggest that the consent timescales can be defined by the incidence of the condition. For very common conditions (10–100%), such as caesarean sections, the frequency of occurrence allows for the complete study information to be distributed before labour, potentially during antenatal care. For common conditions (1–10%), such as breech presentation (20% at 28 weeks of gestation to 3–4% at term),10 obtaining antenatal consent has been criticised for causing couples undue concern and overly medicalising the birthing process.7 Consequently, it is recommended that an outline of the study be provided during antenatal care, with full information provided in the event of eligibility (which may be intrapartum). As the number of clinicians on delivery suite is usually in excess of 100, ensuring Good Clinical Practice training for all staff attending women at all times during the day and night is not practical. The 24-hour nature of recruitment to low-prevalence conditions makes dedicated research staff prohibitively expensive and inefficient. Furthermore, limiting recruitment responsibilities to designated clinicians rather than a departmental team prevents the integration of research and clinical practice. The RCOG, in collaboration with the National Institute for Health and Research (NIHR), have developed a standardised toolkit for training clinicians on the components of Good Clinical Practice, specifically designed to promote the recruitment of pregnant women during labour.11 Although it does not replace the need for Good Clinical Practice training, the toolkit has been created with the purpose of imparting the fundamental principles underpinning clinical research to those clinicians who may encounter recruitment opportunities while providing routine care, but who are not regularly engaged in clinical trials. The toolkit includes a PowerPoint template into which study-specific information can be entered and a Word template that can be distributed to participants. If appropriately implemented, the toolkit has the potential to fulfil regulatory requirements, facilitate successful recruitment during labour, and promote the integration of research into clinical practice. Impetus to determine an acceptable method for obtaining perinatal consent is great. Of primary importance is maintaining a woman's autonomy during the birthing process, which includes providing access to full study information in an understandable format and at a time that promotes unbiased consideration. There is a need to develop a consent pathway that balances the interests of the labouring woman with the promotion of perinatal research, which is required if the quality of obstetric care is to continue to improve. Since the challenges of the preterm breech study, the research agenda has been significantly driven forward by the introduction of NIHR specialty groups to ensure efficient delivery of research to time and target, and the recognition that research is core business to the NHS. Reproductive health and childbirth form one of the leading specialties in terms of both volume and numbers of studies in the UK. However, the challenges of intrapartum research are unique, and many important acute interventions continue to be under researched. It is time to change and learn from this pioneering research. Not only do we need the facilities to conduct research but a clinical culture where non-evidence-based practice is unacceptable and the default should be to recruit to clinical trials, even in these challenging environments. Then we will improve care for mothers and babies. We have no interests to disclose.