SEARCH

SEARCH BY CITATION

Keywords:

  • Breech;
  • external cephalic version;
  • female;
  • fetal/adverse effects;
  • humans;
  • pregnancy

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References

Most breech babies at term are now delivered by caesarean section. Although external cephalic version reduces the number of term breech presentations, its uptake has not been as wide as is recommended by professional bodies. This may be because of fears over safety. Pooled safety data do exist, largely from case series. However, the possibility of reporting and publication biases in such studies maybe preventing both women and obstetricians from being adequately reassured. In this series of 805 consecutive version attempts, we report an extremely low complication rate and 0.5% risk of emergency caesarean section after the procedure.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References

Breech presentation is the most common malpresentation affecting at least 20 000 babies per year in the UK alone. With delivery by planned caesarean section showing a significantly lower neonatal mortality and decreased serious neonatal morbidity,1 very few women are now delivering breech babies vaginally. This is an important contributing factor in the rising caesarean section rate with the associated increase in maternal morbidity and cost.

External cephalic version (ECV) has been clearly shown to decrease the incidence of breech presentation at term, thereby reducing the elective caesarean section rate2 although the randomised controlled trials (RCTs) may have been underpowered to detect rare adverse events. Nevertheless, the Royal College of Obstetricians and Gynaecologists (RCOG) and American College of Obstetricians and Gynecologists currently recommend that all women with an uncomplicated singleton breech presentation at term should be offered ECV.

In spite of this, acceptance of ECV among both clinicians and consumers appears to be limited, apparently because of fears over safety.3 Two large meta-analyses, largely of published case series,4,5 report a low complication rate. However, the contributing case series have been relatively small, on differing populations, and, most importantly, may be subject to reporting and publication biases. Such limitations may compromise the credibility of both meta-analyses. Those concerns are reflected in both consumer views and the professional advice given to women and may negatively affect the uptake and availability of the ECV procedure.

The aim of this article is to report the safety of ECV by documenting all complications that arose in a substantial consecutive cohort of women undergoing the procedure.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References

Data were collected prospectively for all women attending the ECV clinic held at the John Radcliffe Hospital, Oxford, between June 2000 and March 2006. Outcome data were obtained from OXMAT, the hospital records system. Local ethical approval for use of OXMAT data was obtained on 15 September 2005 (ref 05/Q1605/110).

ECV was undertaken in consenting women with stable breech presentation at or beyond 36 weeks in nulliparous women and 37 weeks in multiparous women. Contraindications were where there was a pre-existing absolute indication for caesarean section, growth restriction in association with an abnormal umbilical artery Doppler index, ruptured membranes, rhesus isoimmunisation, severe proteinuric pre-eclampsia or recent vaginal bleeding. With a previous caesarean section, uterine abnormality, known fetal abnormality or where the deepest liquor pool was less than 2 cm, the decision on attempting ECV was made on an individual basis.

A detailed ultrasound scan, but not cardiotocography (CTG), was performed before the version attempt. A forward roll technique was used, either by an operator with experience of more than 20 version attempts or by an operator under direct supervision of an experienced trainer (L.W.M.I.). If the initial attempt at ECV was unsuccessful, a second attempt was made immediately either by the trainer or with the aid of tocolysis. The second attempt was not classified as a new ECV attempt. Ultrasound was also used to aid version and visualise the fetal heart. Immediately after completion of the procedure, a CTG was performed and was continued until a normal baseline rate and variability with at least three accelerations were seen. Anti-D was administered where appropriate. A Kleihauer test was not performed. Where there was any concern about the CTG, or if a compound presentation persisted for more than 30 minutes after any ECV attempt, women were admitted. After successful version, women were referred back to the community or hospital antenatal clinic; if the attempt was unsuccessful, mode of delivery was discussed.

During the time period, a RCT was undertaken within the unit. This recruited women to assess the use of ritodrine where an initial ECV attempt had failed.6 These women were all included in our data.

Data were prospectively recorded by a research midwife on a standardised data sheet when the women attended the clinic. Information on delivery, including mode of delivery, Apgar scores, blood gas results and other fetal outcomes, were recorded from hospital records (OXMAT) within a few weeks of delivery. Data were analysed using SPSS (SPSS Inc., Chicago, IL) (version 12).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References

Eight hundred and forty-two women with breech presentation at 36 weeks and above were seen in the ECV clinic between 12 June 2000 and 21 March 2006. Twelve declined to have the procedure, and 21 women had contraindications: seven had severe oligohydramnios, four had significant fetal compromise, four had an unstable or transverse lie, two had a placenta praevia, one had a congenital fetal abnormality and one had large uterine fibroids. For two women, the contraindication was not recorded. Two women were diagnosed with an intrauterine fetal death on attending for ECV (one fetus had trisomy 21).

ECV was attempted in 807 consecutive women (96%). Two ECVs were performed to allow drainage of severe hydrocephalus and facilitate vaginal delivery with pre-existing intrauterine deaths (IUDs). These exceptional cases were excluded from analysis. Therefore, data are presented on 805 women. The mean age for the women was 31 years (range 15–44 years), mean weight 66 kg (range 39–168 kg) and mean gestation at ECV 37 + 3 weeks (range 36–42 weeks). Four hundred and seventy women (58%) were nulliparous, 31 (4%) had had a previous caesarean section, and in 357 (44%) women, the version was successful.

One hundred and twenty-four women were recruited into our RCT examining the use of ritodrine tocolysis for repeat ECV attempts.6 Therefore, 62 women received intravenous ritodrine infusion before a second attempt at ECV as part of the study. This trial showed significant improvement in success rates for repeat ECV with tocolysis and so a further 44 women received intravenous tocolysis for a second attempt after completion of the trial.

Table 1 shows the complications of ECV. There was one unexplained antepartum stillbirth (0.1%) after a successful and uncomplicated ECV performed at 37 weeks. Two subsequent antenatal visits were recorded and normal fetal movements were experienced, but at 40 + 5 weeks, an unexplained fresh stillbirth was diagnosed. There were no cases of uterine rupture, scar dehiscence, neonatal encephalopathy or evidence of fetal trauma (fractures or bruising).

Table 1.  Adverse events after ECV
 n (%)
  • *

    Blood loss of approximately 150 ml after ECV necessitating immediate delivery.

  • **

    Data on Apgar scores and blood gas results were not available for all deliveries; for example, from home births.

Women in whom ECV was attempted805 (100)
Perinatal mortality1 (0.1)
ECV-attributable mortality0 (0)
Abnormal CTG immediately after ECV13 (2)
Emergency caesarean immediately after ECV4 (0.5)
Admission to hospital for observation after ECV39 (5)
Uterine rupture or scar dehiscence0 (0)
Suspected placental abruption*1 (0.1)
Neonatal encephalopathy0 (0)
Fetal trauma e.g. fractured bones0 (0)
Apgar score of less than 7 at 5 minutes**3/800 (0.3)
Arterial pH at delivery <7.05**9/489 (1.8)
Admission to neonatal unit36 (4)

Four women were delivered by emergency caesarean section at the time of the attempted version (0.5%). Two had abnormal CTG recordings for more than 20 minutes after the procedure. One of these neonates was subsequently diagnosed with trisomy 21. A third woman experienced vaginal bleeding with a normal CTG after a failed ECV. No definitive evidence of placental abruption was found at caesarean section. Another woman’s membranes ruptured during the failed ECV attempt. All three congenitally normal babies were born with normal Apgar scores and cord pH levels.

Thirteen women (2%) had transient CTG abnormalities immediately after the procedure: in all but the two women (above), these settled spontaneously. Thirty-nine women (5%) were admitted for observation after attempted ECV.

Of the 357 women who had a successful version, 11 were not cephalic at delivery (spontaneous reversion rate 3%). Of those 448 with an unsuccessful version, 14 were a cephalic presentation at delivery (spontaneous version rate after failed ECV 3%). Of the 345 women who had a successful version and were cephalic at the onset of labour and attempted a vaginal delivery, 33 (9.5%) were delivered by emergency caesarean section, 57 (16.5%) by instrumental delivery and 255 (74%) delivered normally.

Thirty-six (4%) babies were admitted to the neonatal unit; 18 (4%) after unsuccessful and 18 (5%) after successful versions (χ2= 1.296, P= 0.523). Five (0.6%) babies had Apgar scores of 7 or less at 5 minutes.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References

This is, to the best of our knowledge, the largest single series of ECV attempts documenting fetal safety published in the English language. Although our numbers are less than that in the systematic reviews,4,5 our results are similar, and the consecutive nature of our cases should add weight to their reliability by allaying fears regarding reporting or publication bias. The overall perinatal mortality was 0.1%, an unsurprising figure in a large group of term pregnancies, with no fetal deaths attributable to ECV. Uterine rupture and fetal trauma were not experienced, and suspected placental abruption occurred in 0.1%. We also report a 3% rate of spontaneous reversion after a successful version and 3% rate of spontaneous version after a failed ECV. These figures should be useful for counselling women.

The limitations relate to the nature of all observational studies. However, case series are able to show the incidence of a particular adverse event, even if not prove it to be directly attributable to the intervention being carried out. It would not have been appropriate to perform a case–control study, as the control group would have had to consist of women with breech presentation not having ECV. This would have been unethical given the current RCOG guidelines and known reduction in caesarean section rate with successful ECV. However, the lack of a control group means that we cannot add to existing literature regarding intrapartum intervention figures after successful ECV, which suggests a slightly higher rate of intrapartum caesarean section when compared with babies that were always cephalic. Similarly, although our neonatal admission rate is similar to persistently cephalic neonates at term in this unit, we cannot conclude that ECV does not increase neonatal admission rates. Finally, our small numbers of women with a previous caesarean section prevent confident conclusions about safety but are in accordance with the reassuring literature currently available.

Where the operator or supervisor is experienced in ECV, serious complications are, at worst, very rare. While this could be attributed to a relatively low success rate, our women were predominantly Caucasian and 58% were nulliparous: groups associated with low ECV success rates. Furthermore, our exclusion criteria appear to be less strict than in most other series.

Many obstetricians and women still need to be convinced of the safety of ECV. However, local experience is encouraging as efforts to disseminate our results to midwives, GPs and surrounding local hospitals have resulted in a dramatic increase in the number of women attending ECV clinic and accepting the offered procedure. Our data support the view that the systematic reviews of hitherto published case series do reflect the true incidence of complications and do not under-represent the adverse events. While case reports of complications exist for most procedures, particularly when performed by inexperienced operators without safety precautions, this does not mean that the procedure itself is inherently dangerous. We do not advocate complacency but stress the need for stringent local audit of safety. Our data suggest that women should be counselled that ECV is extremely safe but has a 0.5% risk of emergency caesarean section at the time of the procedure.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References

No external funding was received; all the authors are currently employed on NHS contracts.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. References