- Top of page
Objective To use information collected by the Confidential Enquiry into Stillbirths and Deaths in Infancy to help obstetric, midwifery and paediatric practice in the management of shoulder dystocia.
Design Review of casenotes by a multidisciplinary focus group.
Sample All 56 cases reported to the Confidential Enquiry into Stillbirths and Deaths in Infancy from England, Wales and Northern Ireland in 1994 and 1995, where stillbirth or neonatal death was attributed to shoulder dystocia.
Main outcome measures Case notes were reviewed with respect to a range of perinatal variables. Comparisons were made with normative data from other studies when appropriate.
Results Maternal obesity and big babies were over-represented in pregnancies complicated by fatal shoulder dystocia. Fetal compromise was recorded in 26% of labours. The median time interval between delivery of the head and the rest of the body was only five minutes. The lead professional at the time the head was delivered was a midwife in 65% of cases. Middle grade or senior obstetric staff were supervising 47% of cases by the time the body was delivered.
Conclusions Antenatal prediction of shoulder dystocia is imprecise, and the majority of deliveries are attended by midwives. A relatively brief delay in delivery of the shoulders may be associated with a fatal outcome.
- Top of page
There is no ‘universally accepted definition’ of shoulder dystocia8, which partly accounts for the different reported incidences between 0.23% and 1.1% of vaginal vertex deliveries9,10. We were deliberately overinclusive in our selection of cases, and many that were delivered by traction of the head without any other positioning or manoeuvre may well have been excluded if a more rigorous definition of ‘true’ shoulder dystocia had been applied9. Precise definition is very difficult, as most definitions tend to include subjective statements about the degree of force required to achieve delivery. None of the records that we reviewed mentioned the so-called ‘turtle sign’ or ‘turtle-necking’ in which the head, once delivered, is tightly retracted against the perineum. This sign is not always present but when it occurs may be particularly ominous. McRoberts, who later described the primary manoeuvre for expediting delivery in cases of shoulder impaction, recounted an early exposure as a junior resident. “When the head was visible with contractions I applied low forceps as we were taught. On removal of the forceps the chin disappeared into the mother's anus. I did not realise then what a problem I had. Forty minutes later with great difficulty I delivered a dead 121b baby11.
During the two year period when these 56 cases occurred, there were approximately 1,400,000 births in England, Wales and Northern Ireland. Fatal shoulder dystocia is therefore very rare, with an approximate incidence of 0.04 per 1000 deliveries using the loose definition applied in this paper. If the definition of ‘true shoulder dystocia’ as described by Gross et al.9 had been used, then those cases where delivery was achieved by episiotomy and downwards traction would have been excluded, and the incidence would have been lower at 0.025 per 1000.
The body mass index distribution of the mothers is clearly different from the normal population of pregnant women. However, the North-West Thames data suggest that a maternal body mass index > 35 is expected in 3% and a body mass index of > 40 in 0.7% of normal pregnancies. Over the time period of CESDI data collection, therefore, about 42,000 pregnancies in England, Wales and Northern Ireland would be expected to be associated with a maternal body mass index > 35. From this population there were 13 cases of fatal shoulder dystocia reported to CESDI in which maternal body mass index was > 35. Therefore, the positive predictive value of a body mass index > 35 for predicting fatal shoulder dystocia was extremely low (13/42,000), and the positive predictive value for a body mass index of > 40 would be 4/10,000.
Previous shoulder dystocia is a widely quoted risk factor for shoulder dystocia. Smith et al.12 reported a recurrence risk of about 10% for shoulder dystocia if labour was allowed. However, Baskett and Allen13 found only one case of recurrence in 93 subsequent pregnancies in 80 women. Our data are consistent with an intrinsically low recurrence risk, in that only one mother had experienced shoulder dystocia in a previous pregnancy. The low recurrence rate may also reflect a cautious (or defensive) approach by obstetricians in subsequent pregnancies following shoulder dystocia, with a low threshold for caesarean section. In addition, women themselves may not wish to contemplate further vaginal delivery.
Macrosomia is another established risk factor for shoulder dystocia, and these cases confirmed a tendency to high birthweight. Prediction of birth weight by ultrasound examination is known to be inaccurate, especially in big babies14,15. Delpapa et al.14 found that half of 242 babies assessed by antenatal ultrasound to be over 4 kg actually had a birthweight less than 4 kg. Even if birthweight could be accurately predicted antenatally, these data show that the positive predictive value to predict fatal shoulder dystocia would be extremely low. Approximately 2% and 0.3% of all deliveries are of babies with birth-weights > 4.5 kg and > 5 kg, respectively, which corresponds to about 28,000 and 4000 babies in England, Wales and Northern Ireland over a two year period.
It is clear from the above that antenatal selection is not going to detect all babies at risk for shoulder dystocia. However, in the cases reviewed here there was often an indication from fundal height measurements that a large baby was the most likely outcome and in many cases this information did not appear to modify the plans for labour and delivery. We would agree with Gibb16 who suggests that when a large baby is suspected clinically in late pregnancy “the case sheet should be clearly marked…This primary warning should appear in the part of the notes used for admission in labour”. We would also suggest that the mother is told she has a big baby, and advised to transmit this information to the professionals involved in her labour.
The positive predictive value of many of these risk factors is so low that caesarean section may well not be appropriate. However, if a decision is made to deliver vaginally, the risk of shoulder dystocia may alter practice, both in the conduct of the labour and in determining the need for experienced obstetric and paediatric attendants at delivery. There was considerable delay in arrival of experienced obstetric staff in many of these fatal cases and in 19% no paediatric staff were present until at least three minutes after delivery of the trunk. Paediatric attendance at delivery should be planned for big babies, and in unexpected cases a paediatrician, preferably of registrar status, must be called as soon as difficulty in delivery of the shoulders is detected.
We looked carefully at the induction regimens used in these cases. There was no evidence that those women in whom medical induction proved difficult had particularly long or difficult labours.
Review of the clinical notes suggested care had been suboptimal in some cases, but it was difficult to draw firm overall conclusions concerning events at delivery. A surprising number of cases were delivered by traction alone, and neither McRoberts’ maneouvre, nor suprapubic pressure, was mentioned in the clinical notes. Excessive traction without other manoeuvres is inappropriate management, which in surviving infants may be expected to lead to an unacceptable incidence of brachial plexus injury. It is possible that appropriate maternal repositioning and/or suprapubic pressure were attempted but not recorded. Fundal pressure is of no theoretical benefit, may increase shoulder impaction and is not recommended, but was recorded in six cases.‘Wood's screw’ maneouvre was used in only two cases. Delivery of the posterior shoulder by a consultant or registrar obstetrician was necessary in 20 cases, a fact which has implications concerning immediate availability of skilled middle grade or senior staff on labour wards.
A significant minority of these infants had evidence of fetal distress before delivery of the head. This may be part of the reason that many died following a remarkably short delay in delivery of the body, even allowing for the fact that head-body delivery intervals may have been underestimated by the clinical staff involved. Animal studies would suggest that a fetus in good condition should be able to withstand at least 5–10 minutes of total anoxia and still respond to resuscitation17. Intrauterine compromise and poor condition of the fetus by the time the head delivered may have contributed to the poor outcome in some cases. Another possible explanation for the very short head/body delivery interval in some cases is that the mechanism of cerebral injury and death in shoulder dystocia may be different to cerebral hypoxia-ischaemia from other causes such as cord prolapse or placental abruption. Compression of the neck resulting in cerebral venous obstruction, excessive vagal stimulation and bradycardia, or other mechanisms, may combine with reduced arterial oxygen supply to cause clinical deterioration out of proportion to the duration of hypoxia. There was evidence from some autopsies that trauma as well as hypoxia-ischaemia may have contributed to the adverse outcome. Other mechanisms such as massive vagal stimulation would not be detectable at autopsy.
The delayed arrival of paediatric staff in many cases was certainly suboptimal, even if there is no guarantee that earlier paediatric intervention would have altered the outcome. In the majority of instances record review suggested that appropriate and competent resuscitation attempts were made. Clinical evidence of brachial plexus injury was not recorded in any of these babies, but most survived for only a brief period or were too severely encephalopathic to demonstrate focal neurological signs.
The autopsy rate of 45% was well below the target of 75% recommended by the Royal Colleges of Pathology and Obstetrics and Gynaecology18, and the quality of some autopsy reports was suboptimal. We could not determine whether the low autopsy rate resulted from professional judgement that autopsy would be noncontributory in the context of shoulder dystocia, or from a high rate of parental refusal.
The focus group found several examples of poor communication between parents and professionals in these case histories. Greater involvement of mothers in antenatal and intrapartum decisions may have helped in some cases. The group suggests that even in clinical crises it should be possible to keep the mother informed of what is happening, and one of the professionals involved should take clear responsibility for this important part of management. When there is an adverse outcome, of course, the need for good communication is even greater.
The group was well aware that a review of cases of shoulder dystocia resulting in perinatal death could not lead directly to the development of guidelines for successful management. Different management of the fatal cases may not have affected outcome. There was no control group, and we had no data from cases of shoulder dystocia with a favourable neonatal outcome.
Although there is no ‘evidence’ on which to develop a rational management protocol for shoulder dystocia, there is reasonable consensus between authors such as Roberts8, Pearson19 and Gibb16 who have published recent reviews. These data from CESDI suggest that shoulder dystocia often occurs unpredictably and senior midwifery or obstetric assistance is needed to effect delivery. Given the short interval between delivery of the head and the delivery of body in some of these fatal cases, the need for rapid response is obvious. The sequence of events during delivery was often inadequately recorded, but in many cases it seemed that increased traction alone or fundal pressure were used inappropriately. If clinical guidelines were available, they were not being followed in these cases. We would recommend that a set of clear, written guidelines, adapted according to local circumstances, is developed for every labour ward and midwifery service and reiterated to staff at regular intervals. The basic sequence of clinical manoeuvres and the procedures for calling senior staff need to be explicit and the need for a clear, complete, contemporary record of the sequence of events should be stressed.
This research was initiated and supported by CESDI who are funded by a Department of Health grant. Members of the focus group are grateful to the CESDI secretariat for their support and the Professional Steering Group and National Advisory Body of CESDI for their comments.