The extraordinary results published by Rietberg et al.1 in a recent edition of the journal showed the immediate impact of the Term Breech Trial2 on medical practice and perinatal outcome in the Netherlands. They looked at caesarean section rates, perinatal deaths, asphyxia and trauma in the 33 months before and the 25 months after the publication of the Hannah trial, in a population-based study of more than 35 000 women. By increasing the caesarean section rate from 49 to 80%, perinatal mortality rates and low Apgar scores were halved, while birth trauma was quartered. At this rate, it is possible that there are more than 60 Dutch children alive, who might not be, had the Term Breech Trial not taken place, and many more Dutch children have avoided serious birth injury. There is room in the Netherlands for further improvement: another population-based retrospective study from California3 of more than 100 000 term breeches, where the overall caesarean rate was 95%, showed even lower neonatal mortality (0.6 per 1000). This might have been substantially lower had not some 1200 nulliparous women been delivered vaginally. The Californian study showed a massive increase in brachial plexus injury in vaginal breech delivery in nulliparous women and major increases in neonatal death, trauma and asphyxia. In multiparous women (some 3700 women), there were significant increases in trauma and asphyxia but not in neonatal death. The rates of adverse outcome were lower than previously reported, but it may be that the 5% who delivered vaginally represented a highly selected, low-risk group of term breeches. These findings are supported by three other population-based studies from Sweden4,5 and Denmark,6 involving some 50 000 women. In the earlier Swedish study of just less than 16 000 women with term breech,4 there was a 2.5-fold increase in infant mortality, a 12-fold increase in birth injury associated with vaginal breech delivery and a four-fold increase in neonatal convulsions associated with emergency caesarean section, compared with elective caesarean section. In the more recent Swedish study of more than 22 000 breech presentations,5 the decrease in risk of perinatal or infant death was even greater (OR 3.5) with elective caesarean delivery. In a Danish study of 15 000 nulliparous women,6 intrapartum and neonatal mortality was increased three-fold in the vaginal delivery group and maternal morbidity was significantly increased in intrapartum caesarean section, compared with elective caesarean section. It is notable that in each of the five large population-based studies, the authors concluded that elective caesarean was safer for term breech babies.
Eight further retrospective hospital-based cohort studies from highly developed countries,7–14 involving between 600 and 1500 women in each, have been published since 1998, six of them since the publication of the Term Breech Trial. None of them had the statistical power of the Term Breech Trial, but in several studies, evidence of an increase in perinatal mortality and morbidity associated with attempted vaginal delivery was presented. Several of these authors7,8,10,13,14 have advocated that there is still a place for vaginal breech delivery of term babies despite evidence of increased neonatal morbidity in most of the reports and serious morbidity and deaths in some. In the study from Graz7 of 699 term breeches, serious neonatal morbidity was increased from 0.5 to 2.3% (P= 0.12) and developmental delay (mean follow up of 57 months) was increased from 0.5 to 1.9% in trial of vaginal delivery compared with elective caesarean section, although the differences did not quite reach statistical significance. In the Dublin study8 of 641 women, after careful selection of cases suitable for vaginal breech delivery and using a strict intrapartum protocol, no differences in low Apgar scores were found. There was, however, a very high rate of intrapartum caesarean section (64% in nulliparous and 37% in parous women). Also, despite careful selection aimed at avoiding vaginal delivery for babies weighing more than 3.8 kg, more than one-third of babies weighing more than 3.8 kg actually underwent trial of vaginal delivery, indicating the difficulty of ‘careful selection’. The Lund study9 of more than 1000 women, found a 12-fold increase in neonatal neurological morbidity, including two deaths and three cases of cerebral palsy with vaginal delivery, compared with elective caesarean section (one case of cerebral palsy). The Miami study10 showed that neonatal intensive care admission was significantly increased in vaginal delivery and one intrapartum stillbirth occurred as a result of entrapment of the fetal head: despite delivery of the baby to the umbilicus, caesarean was attempted but the baby could not be resuscitated. Vaginal delivery rates were low in this study: 191 of 1021 breeches were suitable for attempted vaginal delivery, and 135 of these delivered vaginally. A report from Geneva,11 involving 705 women, found corrected neonatal morbidity to be higher, but not significantly so, in trial of vaginal delivery compared with elective caesarean section (4.5 versus 2.6%, P= 0.22); maternal morbidity was significantly lower with planned vaginal delivery. In a recent study from Birmingham,12 552 women were delivered by elective caesarean section and 881 were delivered vaginally or by intrapartum caesarean: there were no perinatal deaths in the first group and three in the second. Early neonatal morbidity was also significantly increased with attempted vaginal delivery, and while long-term morbidity was not significantly increased, it must be borne in mind that three babies in the vaginal delivery group were already dead when long-term follow up took place. A second Austrian group13 have reported only two cases of serious neonatal morbidity in the vaginal delivery group (none in the caesarean delivery group) in their study of 882 breech presentations over an 11-year period, and a Finnish group14 have reported neonatal morbidity of only 1.2% with planned vaginal delivery of 590 breeches (0.5% in the elective caesarean group of 396 breeches). Given that units with disappointing vaginal breech results are unlikely to publish the results, these single-institution reports, involving less than 8000 women between them, need to be interpreted very cautiously in view of the weightier evidence of population-based studies of some 185 000 deliveries.
The alternative to vaginal breech delivery, elective caesarean section, has become a very safe procedure. Recent British data15 identified one maternal death in 78 000 elective caesareans, even lower than the rate of 5.6 per 100 000 for repeat elective caesarean section reported by Wen et al.16 in their study of more than 300 000 Canadian women having elective repeat caesarean section. Similarly, a study from Washington State of more than 265 000 nulliparous women17 showed, with logistic regression analysis, that women who underwent caesarean section were not at increased risk of death compared with those who had vaginal delivery. In a population-based study from Nova Scotia, involving more than 18 000 women,18 maternal morbidity in elective caesarean was confined to postpartum febrile morbidity, while the more serious morbidity, postpartum haemorrhage, was reduced by elective caesarean section; the highest morbidity was in assisted vaginal delivery and intrapartum caesarean section. In a recent prospective study of more than 1600 women, comparing planned caesarean section with planned vaginal delivery, McAuliffe et al.19 found no difference between the groups at 12-month follow up in perception of pain, pelvic floor function, sexual function, coping, postpartum stress, anxiety and depression. The recent article from Patel et al.20 on a cohort of more than 10 000 women showed that elective caesarean section does not cause (nor does it prevent) postnatal depression. Elective caesarean section is not risk free: the most serious long-term risk is placenta praevia percreta, which carries a maternal mortality rate of 7%21 and is one of the most challenging conditions an obstetrician can encounter, as demonstrated by the dramatic case reported by Leaphart et al.22 However, an elective caesarean section for breech presentation can be followed by successful vaginal delivery in the next pregnancy: in a Dublin study,23 85% of those women who had trial of vaginal delivery in their second pregnancy, having had an elective caesarean section for breech presentation in their first pregnancy, delivered vaginally.
Importantly, the Dutch report1 seems to answer emphatically criticisms24,25 of the Term Breech Trial. In this regard, it has brought the term breech debate towards a final conclusion. It replaces the Term Breech Trial as the most important reference on the subject, and it might well be the final word on the matter. It only remains to be seen if the Dutch results can be replicated in other countries: the Netherlands had an unusually low caesarean rate to begin with, so one might expect less dramatic improvements in other countries. It is unlikely that the Term Breech Trial will be repeated in fully developed countries: in a systematic review of the literature by a group from Norway,26 it was concluded that with Scandinavian neonatal mortality and morbidity rates, a replication of the Term Breech Trial in a Scandinavian setting would be impractical because it would require some 10 000 women in each arm.
So what now for advocates of vaginal breech delivery? Certainly, they have their work cut out for them. They will need to develop very robust consent procedures, and they might be well advised to abandon vaginal breech delivery in nulliparous women altogether. Even in multiparous women, where there is not so great a risk of perinatal death from vaginal breech delivery, the Californian data3 show a substantial risk of morbidity. It was, perhaps, a surprising finding of the Term Breech Trial that multiparity offered the baby relatively little protection from serious morbidity. Women wishing to have a trial of vaginal breech delivery need to be counselled about this. Disturbingly, but hardly surprisingly, it was noted by the authors of the Californian study3 that women who were of ethnic minority, less educated or uninsured were more likely to have a vaginal breech delivery than were women who were white, college educated and privately insured. Vaginal breech delivery is becoming uncommon. Even where vaginal breech delivery is still advocated,8 the number of vaginal breech deliveries in nulliparous women per ‘experienced obstetrician’ was less than one a year. This is hardly enough to maintain one’s own skills, let alone to teach them to the next generation of young obstetricians. This is worrying because there is some evidence that young obstetricians approach breech delivery with a certain amount of naivety: in a UK survey,27 92% of registrars, who had carried out a median of 16 vaginal breech deliveries, described themselves as confident and none described themselves as unconfident; whereas only 55% of consultants, who had carried out a median of 162 vaginal breech deliveries, described themselves as confident and 18% described themselves as unconfident.
Careful advocates of vaginal breech delivery will certainly not cause many perinatal deaths. For example, a large maternity unit, delivering 7000 babies per annum, would manage about 2000 term breeches in a 10-year period. With a fairly liberal elective caesarean rate of 50% and an intrapartum caesarean rate of a further 25%, extrapolating from the population-based studies mentioned above, it might have fewer than eight perinatal deaths caused by vaginal breech delivery in a decade, and it would avoid about 500 caesareans. Thus, they would need to perform at least 60 extra caesareans for each perinatal death avoided. In the Term Breech Trial it was suggested that in countries reported to have a low perinatal mortality rate, one baby might avoid death or serious morbidity for every seven extra caesareans performed, but in the study of Rietberg et al., Dutch obstetricians performed 192 extra caesareans to avoid one perinatal death, 82 caesareans to avoid a death or birth trauma but only 18 to avoid one baby dying or having birth trauma or having a 5-minute Apgar score of less than 7. By increasing the caesarean rate to 95% to avoid 50% of the remaining perinatal deaths (were that possible), 164 further caesareans would be required for each death avoided.
The question is: is it worth doing 60 or 160 more caesareans, which nowadays cost nothing in terms of maternal mortality, morbidity or economically,28 to avoid one baby’s death? Is that too extravagant? Exactly how many caesareans is one baby’s life worth? And is it reasonable that this arbitrary decision be made on the basis of an individual obstetrician’s personal enthusiasm for vaginal breech delivery?