Intrapartum monitoring with cardiotocography and ST-waveform analysis in breech presentation: an observational study

Authors

  • J Kessler,

    Corresponding author
    1. Department of Obstetrics and Gynaecology, Haukeland University Hospital, Bergen, Norway
    2. Department of Clinical Science, Research Group for Pregnancy, Fetal Development and Birth, University of Bergen, Bergen, Norway
    • Correspondence: Dr J Kessler, Department of Obstetrics and Gynaecology, Haukeland University Hospital, 5021 Bergen, Norway. Email: jorg.kessler@kk.uib.no

    Search for more papers by this author
  • D Moster,

    1. Department of Health Registries, Norwegian Institute of Public Health, Bergen, Norway
    2. Department of Paediatrics, Haukeland University Hospital, Bergen, Norway
    3. Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway
    Search for more papers by this author
  • S Albrechtsen

    1. Department of Obstetrics and Gynaecology, Haukeland University Hospital, Bergen, Norway
    2. Department of Clinical Science, Research Group for Pregnancy, Fetal Development and Birth, University of Bergen, Bergen, Norway
    Search for more papers by this author

Abstract

Objective

To determine the electrocardiographic performance and neonatal outcome of pregnancies with breech presentation and planned vaginal delivery monitored with ST-waveform analysis (STAN).

Design

Prospective observational study.

Setting

University hospital, Norway; 2004–2008.

Population

Singleton pregnancies with a gestational age above 35 + 6 weeks, breech presentation, selected for vaginal delivery and monitored with STAN.

Methods

Common clinical guidelines for STAN monitoring were used. An experienced neonatologist graded the symptoms of neonatal encephalopathy. The outcome was compared with STAN-monitored high-risk deliveries in a vertex presentation (n = 5569) using logistic regression analysis.

Main outcome measure

Frequency of ST events, indications of intervention for fetal distress, and neonatal morbidity and mortality.

Results

Breech presentation occurred in 750 of 23 219 (3.2%) deliveries, 625 (83%) of which were selected for vaginal delivery. Intrapartum monitoring by STAN was performed in 433 (69%). Compared with vertex presentations, fetuses in breech presentation had a lower risk of baseline T/QRS rise during labour [odds ratio (OR) = 0.7, 95% confidence interval (95% CI) = 0.7–0.9, P = 0.003] and a higher risk for intervention as a result of preterminal cardiotocogram (OR = 2.9, 95% CI = 1.6–5.9, P = 0.001). The risks of perinatal mortality (OR = 1.8, 95% CI = 0.2–15, P = 0.6), cord metabolic acidosis (OR = 0.8, 95% CI = 0.2–3.2, P = 0.7) and moderate or severe neonatal encephalopathy (OR = 1.8, 95% CI = 0.5–6.2, P = 0.3) did not differ significantly between breech and vertex deliveries.

Conclusion

STAN can be used for the surveillance of breech presentations selected for vaginal delivery with an acceptable neonatal outcome. The electrocardiogram (ECG) pattern during labour varies with the fetal presentation.

Introduction

The Term Breech Trial (TBT) revealed an increased mortality in participants randomised to planned vaginal delivery relative to elective caesarean section.[1] This study resulted in changes to the management of breech presentation at term, with elective caesarean delivery being recommended in most countries.[2-4] The TBT has since been criticised for methodological flaws,[5] and long-term follow-up of the infants has not confirmed that the outcome differs with the mode of delivery.[6] In contrast with the recommendations of the Cochrane review,[7] several national guidelines[8-10] still support the option of vaginal delivery for breech presentation at term.

Cardiotocography (CTG), which is the basic method of surveillance of high-risk labour, is hampered by a high sensitivity and low specificity for the detection of intrapartum hypoxia. As a result of the risk of cord compression and cord prolapse in breech presentations, continuous electronic fetal monitoring is essential in order to reveal fetal hypoxia in a timely manner. The use of ST-waveform analysis (STAN) of the fetal electrocardiogram (ECG) as an adjunct to CTG has been found to reduce the frequency of vaginal operative deliveries, the necessity of fetal blood sampling and the frequency of neonate transferral to the neonatal intensive care unit (NICU).[11] The largest randomised controlled trials have also demonstrated a decreased frequency of cord metabolic acidosis.[12, 13]

Despite being a common indication for electronic fetal monitoring, breech presentation was an exclusion criterion in all but one[14] of the five randomised trials on CTG and STAN.[12, 13, 15, 16] A total of 55 cases with breech presentation were randomised in the Plymouth study,[14] 30 of these to CTG and STAN. Thus, the evidence from randomised trials supporting the use of the method in breech presentation is almost negligible. The aim of the present study was therefore to investigate the fetal ECG waveform pattern and to evaluate the neonatal outcome of pregnancies with breech presentation selected for vaginal delivery, monitored with STAN.

Materials and methods

This was a prospective observational study of all singleton pregnancies with a gestational age above 35 + 6 weeks with a planned vaginal delivery that were monitored by STAN between 1 January 2004 and 31 December 2008. STAN monitoring was restricted to high-risk deliveries as defined by local guidelines.[17] The study protocol was assessed by the regional ethics committee (REK Vest, 129.08) and defined as a quality improvement project which, according to Norwegian law on medical research, does not require written patient consent for their involvement. The study population consisted of pregnancies with a fetus in breech presentation; those with cephalic presentation served as a reference. The overall results on the maternal and neonatal outcome of the study population have been published elsewhere.[17]

Women with a fetus in breech presentation were selected to undergo a planned vaginal delivery according to the following local guidelines:

  1. The estimated fetal weight did not exceed 4500 g (individual assessment between 4000 and 4500 g according to pelvic size and obstetric history).
  2. Computed tomography (CT) pelvimetry was performed [conjugata vera ≥ 11.5 cm, sum of outlet ≥ 32.5 cm (sagittal outlet + intertuberous + interspinous diameters)] in nulliparous women or those with a large discordance between previous birthweight and estimated fetal weight in the current pregnancy.

Epidural anaesthesia was advised for all women in the study population. All deliveries were performed by, or under the supervision of, a senior consultant. The shoulders and arms were delivered by Løvset's manoeuvre. Piper's forceps were used liberally for the delivery of the aftercoming head. A paediatrician was present at delivery. In cases with ruptured membranes, a fetal electrode was attached at the onset of active labour. If the CTG was normal, the fetal membranes were kept intact. Amniotomy was not performed unless CTG abnormalities occurred. The fetal electrode was most often attached to the fetal buttock, in some cases to the foot. In order to enable an ECG analysis analogous to that for cephalic presentation, the recorded ECG had to be mirror imaged by activation of the ‘Breech mode’ in the monitoring device. The diagnosis and intervention in cases of suspected intrapartum hypoxia were based on the clinical guidelines for monitoring with CTG and STAN.[18] CTG classification was conducted according to the International Federation of Gynaecology and Obstetrics.[19] CTG abnormalities were then related to the occurrence of ST events (Table 1) to determine whether or not to intervene.

Table 1. Clinical guidelines for the use of cardiotocography (CTG) and ST-waveform analysis; recommendations for intervention according to CTG abnormalities and ST events
 Intermediary CTGAbnormal CTGPreterminal CTG
Episodic T/QRS rise>0.15>0.10Immediate delivery
Baseline T/QRS rise>0.10>0.05 
Biphasic STThree biphasic eventsTwo biphasic events 

If hypoxia is indicated during the first stage of labour, clinical guidelines[18] recommend intrauterine resuscitation as the first measure and, if not successful, an operative intervention should be performed and delivery accomplished within 20 minutes. Immediate delivery is warranted if hypoxia is indicated during the second stage.[18]

An adverse neonatal outcome was defined as follows: perinatal mortality, cord metabolic acidosis [arterial pH (pHartery) < 7.05 and base deficit in the extracellular fluid (BDecf) > 12 mmol/l[20, 21]], cord acidosis at delivery [either moderate (pHartery < 7.15, <10th percentile) or severe (pHartery < 7.05, <2.5th percentile)[22]], a 5-minute Apgar score of <7, transfer to the NICU, and moderate or severe neonatal encephalopathy (NE). NE was graded according to Sarnat and Sarnat[23] by an experienced neonatologist, as described elsewhere.[17]

The occurrence of ST interval changes, the risk of ST signal disturbance and the risk of adverse outcome were compared with those for a reference population using logistic regression analysis. The characteristics of the study and background were evaluated using Pearson's chi-squared test and Student's t-test. The level of statistical significance for all statistical tests was set at P < 0.05.

Results

A breech presentation occurred in 750 of 23 219 (3.4%) deliveries. Among these, 625 (83%) were selected for attempted vaginal delivery, 433 (69%) of which were monitored with STAN. Pregnancies with breech presentation were characterised by predominantly nulliparous women, who were less likely to have their labour induced (Table 2). This was partly a result of a lower frequency of pregnancy complications (Table 2). Although 78% of the women selected to a trial of labour had a vaginal delivery, the frequency of emergency caesarean section was significantly higher for breech than for cephalic presentations (Table 2).

Table 2. Basic characteristics, occurrence of pregnancy complications and mode of delivery according to fetal presentation. Data are means ± standard deviation (SD) or n (%) values
 Breech (n = 433)Vertex (n = 5577) P
Maternal age (years) 29.8 ± 4.829.6 ± 5.20.6
Nulliparous 290 (67%)3335 (60%)0.002
Pregnancy complications
Diabetes mellitus1 (0.2%)66 (1.2%)0.09
Gestational diabetes4 (1%)151 (3%)0.01
Pre-eclampsia17 (4%)510 (9%)<0.001
Induced labour66 (15%)1883 (34%)<0.001
Gestational age at delivery (days) 280 ± 10284 ± 10<0.001
Post-term delivery 44 (10%)1240 (22%)<0.001
Birthweight (g) 3385 ± 5313638 ± 504<0.001
Mode of delivery
Vaginal delivery336 (78%)4701 (84%) 
Caesarean section97 (22%)876 (16%)<0.001

The ECG signal quality in fetuses with a breech presentation differed from that of the reference population, exhibiting a higher proportion of recordings with a non-interpretable ST signal (Table 3). Fetal presentation also influenced the pattern of ST interval changes during labour. Significant ST changes were less common among the breech presentations. Furthermore, these fetuses were less likely to have a baseline T/QRS rise, but were more likely to exhibit a biphasic ST waveform compared with the reference (Table 3). However, the latter difference between cephalic and breech presentations was nullified after adjustment for birthweight, gestational age at delivery and fetal sex (Table 3).

Table 3. Electrocardiogram (ECG) signal quality, ST interval changes and indication for intervention as a result of hypoxia according to fetal presentation
 Breech (n = 433)Cephalic (n = 5577)OR (95% CI)OR (95% CI)a
n (%)n (%)
  1. a

    Adjusted for birthweight, gestational age at delivery and fetal sex.

ECG signal quality
No interpretable ST23 (5.3)181 (3.2)1.7 (1.1–2.6) 
Impaired signal in first stage42 (9.7)518 (9.3)1.0 (0.8–1.5) 
Impaired signal in second stage58 (13.4)777 (14.0)1.0 (0.7–1.3) 
ST interval changes
Any ST event200 (46.2)2897 (51.9)0.8 (0.6–0.9)0.8 (0.7–0.9)
Baseline T/QRS rise168 (38.8)2621 (47.0)0.7 (0.6–0.9)0.7 (0.6–0.9)
Episodic T/QRS rise49 (11.3)535 (9.5)1.2 (0.9–1.6)1.2 (0.9–1.6)
Biphasic ST43 (9.9)391 (7.0)1.5 (1.1–2.0)1.3 (0.9–1.8)
Indication of hypoxia n = 77n = 1052  
Baseline T/QRS rise36 (46.8)684 (65.0)0.5 (0.3–0.8)0.6 (0.4–0.8)
Episodic T/QRS rise8 (10.3)120 (11.4)0.9 (0.4–1.9)0.7 (0.4–1.5)
Biphasic ST13 (16.9)115 (10.9)1.6 (0.9–3.1)1.2 (0.6–2.1)
Preterminal cardiotocography (CTG)13 (16.9)51 (4.9)4.0 (2.1–7.7)2.9 (1.6–5.9)
Abnormal CTG in second stage7 (9.1)80 (7.6)1.2 (0.5–2.7)1.1 (0.5–2.3)

The distribution of indications to intervene in cases of suspected hypoxia, as given by the STAN clinical guidelines, also varied with the type of fetal presentation. Interventions as a result of baseline T/QRS rise occurred less often in breech presentation, whereas hypoxia was more frequently indicated by a preterminal CTG. The distribution of indications for intervention was not influenced by birthweight, gestational age at delivery or fetal sex (Table 3).

The indicators of severe adverse neonatal outcome (cord metabolic acidosis, perinatal mortality, and moderate or severe NE) did not differ significantly between the breech and cephalic presentations (Table 4). However, there was an increased frequency of moderate or severe cord acidosis and low 5-minute Apgar score in breech presentations (Table 4). This was not reflected in an increased need for intensive care: the transfer to NICU and the proportion of neonates with a prolonged stay at NICU (>3 days) did not differ significantly between breech and cephalic presentations (Table 4).

Table 4. Neonatal outcome according to fetal presentation
 Breech (n = 433)Cephalic (n = 5577)OR (95% CI) P
n (%) Mean ± SDn (%) Mean ± SD
  1. BDecf, base deficit in the extracellular fluid; CI, confidence interval; NICU, neonatal intensive care unit; OR, odds ratio; pCO2, partial pressure of carbon dioxide; pO2, partial pressure of oxygen.

  2. a

    Arterial samples available in n = 351 breech deliveries and n = 4786 vertex deliveries; venous samples available in n = 371 breech deliveries and n = 4892 vertex deliveries.

Mortality
Intrapartum01  
Neonatal16  
Perinatal1 (0.2)7 (0.1)1.8 (0.2–15) 
5-minutes Apgar score < 7 19 (4.3)108 (1.9)2.3 (1.4–3.8) 
Transfer to NICU 25 (5.7)452 (8.1)0.7 (0.5–1.1) 
NICU stay > 3 days 11 (2.5)123 (2.2)1.2 (0.6–1.2) 
Neonatal encephalopathy
Mild4 (0.9)38 (0.7)  
Moderate2 (0.4)17 (0.3)  
Severe1 (0.2)4 (0.1)1.8 (0.5–6.2) 
Cord acid base a
Artery
pH7.19 ± 0.097.23 ± 0.08 <0.001
pCO2 (kPa)8.6 ± 1.87.5 ± 1.4 <0.001
BDecf (mmol/l)3.1 ± 3.23.3 ± 3.0 0.3
pO2 (kPa)1.6 ± 0.82.0 ± 0.9 <0.001
Vein
pH7.29 ± 0.087.31 ± 0.07 <0.001
pCO2 (kPa)6.0 ± 1.25.7 ± 1.1 <0.001
BDecf (mmol/l)4.1 ± 2.64.4 ± 2.7 0.03
pO2 (kPa)3.3 ± 0.93.4 ± 0.9 0.2
Cord metabolic acidosis 2 (0.6)35 (0.7)0.8 (0.2–3.2) 
Cord artery pH < 7.15 90 (25.6)598 (12.5)2.4 (1.9–3.1) 
Cord artery pH < 7.05 23 (6.6)123 (2.6)2.7 (1.6–4.2) 
BDecf (mmol/l) > 12 3 (0.9)44 (0.9)0.9 (0.3–3.0) 

In all cases with moderate or severe NE, the delivery of the shoulders, arms and aftercoming head was described as uncomplicated. Nevertheless, the neonates remained depressed despite immediate resuscitation (Table 5). Autopsy findings in case 1 (reactive gliosis in both hippocampi) indicated a possible prenatal brain damage. Bilateral pneumothorax was diagnosed after the initial resuscitation in case 3. None of the cases with moderate NE showed any signs of cerebral palsy when followed up by a child neurologist.

Table 5. Cases with moderate and severe neonatal encephalopathy
 Case 1Case 2Case 3
  1. BDecf, base deficit in the extracellular fluid; VD, vaginal breech delivery.

  2. a

    Born before hypothermia was established as treatment at the local neonatal intensive care unit.

  3. b

    During a 4-year follow-up.

  4. c

    Relates to the time from indication of hypoxia until delivery.

Gestational age (weeks) 40 + 342 + 140 + 6
Mode of delivery VDVDVD
Birthweight (g) 324036403160
Apgar score at 1/5 minutes 3/21/51/1
Cord acid base
Artery
pH/BDecf (mmol/l)7.04/9.27.12/4.1
Vein
pH/BDecf (mmol/l)7.24/5.27.16/10.9
Indication of hypoxia Biphasic ST
Time indication–deliveryc(min)167
Neonatal encephalopathy SevereModerateModerate
Hypothermia YesNoaYes
Survival NoYesYes
Cerebral palsy b NoNo

Discussion

Main findings

The present prospective observational study demonstrated the feasibility of using STAN technology to monitor vaginal breech deliveries. The pattern of ST interval changes during labour was different between breech and cephalic presentations. The frequency of severe adverse neonatal outcome did not differ significantly between breech presentation and high-risk vertex deliveries monitored by STAN; however, this finding must be interpreted with caution because of the small sample of the study.

Strengths

The delivery of a significant number of babies in breech presentation allowed the medical staff to become experienced at such deliveries, and encouraged the development of effective teamwork. The uniform management, thorough evaluation of intrapartum monitoring and neonatal outcome represent strengths of the present study. Furthermore, the sample in this study was considerably larger than those in previous investigations on fetal monitoring in breech presentation.

Limitations

The observational design of the present study prevents any direct comparison with other monitoring techniques (e.g. fetal blood sampling). A major limitation of this study is the size of the study population, which limits the ability to draw conclusions on rare, serious, adverse events. Furthermore, cases with cord acidaemia could have been missed as cord acid–base samples were not available in all cases. The selection of the reference population can also be questioned. Vertex deliveries monitored by STAN were chosen in order to elaborate upon the potential differences in ECG waveform analysis related to fetal presentation.

Interpretation

The proportion of pregnancies with breech presentation selected for vaginal delivery was higher than in previously published data from our own institution[24] and from other recent observational studies from Norway,[25] Finland[26] and Belgium/France (the PREMODA study).[10] This can be partly explained by a reduction in the use of pelvimetry during the study period. The proportion of women who had a vaginal delivery in the present study did not differ from those in the aforementioned studies.

Most guidelines on the intrapartum management of a vaginal breech delivery[9, 10, 27] include electronic fetal monitoring as a prerequisite because of the increased risk of cord prolapse[28] and cord compression.[29] However, there has been little research on methods of fetal monitoring other than CTG during vaginal breech deliveries. Some authors have advocated fetal blood sampling during the late first and second stages of labour. A threshold for intervention of pH < 7.15 has been suggested based on 30 cases, as this cut-off is predictive of a 1-minute Apgar score of <4.[30] Another study that included 10 cases revealed strong correlations between pH values obtained from the fetal buttock within 15 minutes before delivery and umbilical arterial and venous pH at delivery.[31] However, the usefulness of fetal blood sampling as an additional method in fetal monitoring has only been tested in one randomised study.[32] Breech presentation was an exclusion criterion in that study, and the results did not show any improvements in neonatal and maternal outcomes. The introduction of lactate measurement in fetal blood sampling resulted in a higher success rate in sampling and assessment than pH measurement, without any improvement in neonatal outcome.[33] No reference ranges for lactate samples taken from the fetal buttock have been published.

Fetal blood sampling provides only isolated temporal measurements, whereas analysis of the fetal ECG gives continuous information on the fetal cardiac metabolism.[34] An optimal signal quality is necessary for the timed detection of significant ST changes.[18] The ECG signal quality has not been assessed previously in relation to fetal presentation. The higher proportion of recordings with no interpretable ST signal in breech presentations in the present study could be a result of the heart being further from the buttock than from the head, spatial differences in the conductive properties of the body[35] or the greater difficulty of applying the fetal electrode to the fetal buttock, resulting in less stable signal conduction.

The different pattern of ST events in breech presentations could be caused by differences in birthweight and gestational age at delivery (Table 2). Prematurity and low birthweight have previously been shown to be associated with a higher density of ST events in general, and biphasic ST waveform in particular.[36] The latter is also more prevalent among male fetuses.[36] In the present study, adjustment for gestational age at delivery removed differences in the occurrence of biphasic ST waveforms. However, the association between breech presentation and a decreased proportion of cases with baseline T/QRS rise remained significant after this adjustment. ST interval changes not only evolve as a result of hypoxia; a T/QRS rise is observed in one-half of cases[36] as a result of fetal arousal caused by physiological sympathetic stimulation during labour.[37] The effect of a given external stimulus may depend on fetal presentation: it has been found that the response to vibroacoustic stimulation close to term is lower in breech than in cephalic presentations.[38] Furthermore, the sensitivity of somatosensory perceptions in humans is higher in the face than in the lower extremities and buttocks.[39] As a consequence, the mechanical forces affecting the presenting part may result in higher levels of fetal arousal when applied to the fetal skull, eyes and face relative to the feet and buttocks, which could explain the different patterns of ST interval changes between breech and cephalic presentations.

CTG remains the basic method in STAN monitoring. Early research on the changes in fetal heart rate that occur during breech delivery has highlighted the frequent occurrence of variable and/or late decelerations being the result of repetitive cord compression, and their significance for neonatal outcome.[29] Decreased variability was present in eight of 42 cases (19%), and all but two had a 1-minute Apgar score of <5 or the need for respiratory support.[40] In contrast, there are no published data on the prevalence of preterminal CTG relative to fetal presentation. The present study found that a preterminal CTG as a primary indicator of hypoxia was more common in breech than in cephalic presentations. The prevalence of preterminal CTG in breech presentations (3%; 95% CI = 1.8–5.1) did not differ significantly from those reported for fetuses of mothers with diabetes mellitus (8.6%; 95% CI = 4.0–17.5) or gestational diabetes (1.0%; 95% CI = 0.3–2.8).[41] The progression of CTG abnormalities to a preterminal pattern without concomitant ST changes has been found in pregnancies complicated by diabetes mellitus,[41] intrapartum sepsis,[42] and in breech presentation (in the present study). This deserves further investigation.

In line with previous research,[43, 44] we found that breech presentation was associated with acidaemia, hypercapnia and hypoxaemia, with the differences being more prominent in the umbilical artery than in the vein. Another study found that cord acid–base levels at delivery did not vary with the type of fetal presentation.[45]

Although respiratory acidosis is usually well tolerated by the fetus, cord metabolic acidosis is associated with significant neonatal short-term morbidity.[46] Thus, it is reassuring that the frequency of neonates with BDecf > 12 mmol/l was more than 60% lower in the present study than in the TBT.[1] Furthermore, the increased frequency of cord acidosis in breech presentation was not paralleled by an increase in NICU transfer or NE.

Despite no indication of hypoxia during labour and the absence of acidosis at delivery, one neonate died after a vaginal breech delivery. Prenatal brain damage could have influenced the ability of that fetus to tolerate peripartum asphyxia. Delayed intervention was present in one-third of cases with moderate or severe NE, which is in line with previous findings in a high-risk population.[17]

Conclusions

STAN monitoring is technically feasible during vaginal breech delivery, with an acceptable neonatal outcome. The differential pattern of ST interval changes in breech presentation may prompt a further refinement of clinical guidelines on STAN monitoring.

Disclosure of interests

JK received a lecture fee from Neoventa Medical, Mølndal, Sweden on one occasion. SA and DM do not have any conflicts of interest.

Contribution to authorship

JK collected and analysed the data, and wrote the manuscript. SA collected data and revised the manuscript. DM evaluated the medical records of neonates transferred to the NICU and revised the manuscript.

Details of ethics approval

The study protocol was assessed by the regional ethics committee (REK Vest, 129.08) and defined as a project of quality improvement which, according to Norwegian law on medical research, does not require written patient consent for their involvement.

Funding

JK was supported by the Western Norway Regional Health Authority.

Acknowledgements

We thank Berit Aldahl, Ingrid Borthen, Eirin Butt, Thomas Hahn, Ferenc Macsali, Grete Midbøe and Elisabeth Wik for their support in data collection.

Ancillary