Randomised controlled trial of intrapartum fetal heart rate monitoringBMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6927.497 (Published 19 February 1994) Cite this as: BMJ 1994;308:497
- K Mahomed,
- R Nyoni,
- T Mulambo,
- J Kasule,
- E Jacobus
- Accepted 1 November 1993
Objective : To compare effectiveness of different methods of monitoring intrapartum fetal heart rate.
Design : Prospective randomised controlled trial.
Setting : Referral maternity hospital, Harare,Zimbabwe.
Subjects : 1255 women who were 37 weeks or more pregnant with singleton cephalic presentation and normal fetal heart rate before entry into study.
Interventions : Intermittent monitoring of fetal heart rate by electronic monitoring, Doppler ultrasound, use of Pinard stethoscope by a research midwife, or routine use of Pinard stethoscope by attending midwife. Main outcome measures - Abnormal fetal heart rate patterns, need for operative delivery for fetal distress, neonatal mortality, Apgar scores, admission to neonatal unit, neonatal seizures, and hypoxic ischaemic encephalopathy.
Results : Abnormalities in fetal heart rate were detected in 54% (172/318) of the electronic monitoring group, 32% (100/312) of the ultrasonography group, 15% (47/310) of the Pinard stethoscope group, and 9% (28/315) of the routine monitoring group. Caesarean sections were performed for 28% (89), 24% (76), 10% (32), and 15% (46) of the four groups respectively. Neonatal outcome was best in the ultrasonography group: hypoxic ischaemic encephalopathy occurred in two, one, seven, and 10 cases in the four groups respectively; neonatal seizures occurred only in the last two groups (six and nine cases respectively); and deaths occurred in eight,two, five, and nine cases respectively.
Conclusions : Abnormalities in fetal heart rate were more reliably detected by Doppler ultrasonography than with Pinard stethoscope, and its use resulted in good perinatal outcome. The use ofrelatively cheap ultrasound monitors should be further evaluated and promoted in obstetric units caring for high risk pregnancies in developing countries with scarce resources.
Intrapartum asphyxia is an important cause of perinatal mortality and morbidity, particularly in developing countries
As screening for asphyxia is dependent on detection of abnormalities in fetal heart rate, continuous electronic monitoring of fetal heart rate is thought to be appropriate for women in labour with a high risk of fetal hypoxia
This study compared the abilities of intermittent electronic monitoring, a hand held Doppler ultrasound monitor, and the Pinard stethoscope to detect abnormalities in fetal heart rate and their contribution to mode of delivery and fetal outcome.
The ultrasound monitor was better at detecting abnormalities in fetal heart rate than the Pinard stethoscope and was associated with lower neonatal morbidity and mortality, and perinatal outcome was no worse than that achieved with electronic monitoring
Doppler ultrasound monitoring should be promoted in developing countries where electronic monitoring is not feasible, and in first world countries ultrasonography should be evaluated as a possible alternative to electronic monitoring of fetal heart rate
In most Western countries one to two infants in every 1000 without structural abnormalities die during labour, and a similar proportion sustain long term handicap.1 In developing countries intrapartum events at term account for about 10% of stillbirths and early neonatal deaths and, although there are no data, probably for about a similar number of cases of cerebral palsy and severe mental retardation. Such events therefore continue to be an important source of potentially preventable death and damage.
The goal of fetal monitoring in labour is the early detection of a hypoxic fetus. For the monitoring to be effective the test must be performed correctly, its result must be interpreted satisfactorily, and the interpretation must provoke an appropriate and timely response.2 Evaluation of the ability of intermittent auscultation to predict poor fetal outcome is fraught with methodological problems as it is impossible to exclude the confounding effects of clinical intervention on outcome. Also the diagnosis of fetal distress is often not solely based on abnormal fetal heart rate. Of the nine randomised controlled trials comparing different methods of monitoring fetal heart rate,2 only three studied only high risk women.*RF 3-5* The results consistently showed that operative delivery for fetal distress was higher in all the intensively monitored groups and that the use of fetal blood sampling tended to limit the increased use of caesarean delivery and to promote operative vaginal delivery. The Dublin trial of low risk women showed that intrapartum deaths were equally preventable by all of the methods of fetal heart rate monitoring currently employed provided that the importance of prompt recognition of abnormal fetal heart rate was recognised.6
Continuous electronic monitoring of fetal heart rate is generally preferred in high risk obstetric units. The complexity of the monitoring equipment, however, makes it susceptible to technical and mechanical failures, and the costs of the equipment and its maintenance and replacement may be prohibitive, particularly in developing countries with scarce resources. Extensive investment in this equipment by such countries is clearly impractical, inappropriate, and often impossible. Fetal heart rate monitoring with lessexpensive and sophisticated technology may well be just as effective. 7 Auscultation with the Pinard stethoscope may be uncomfortable and unsatisfactory, particularly if analgesia is infrequently used as women may not remain still during the contraction. It is, however, often the only method of fetal monitoring available in many units in the developing world. The unreliability of intermittent auscultation may be increased by use of hand held Doppler ultrasound monitors. These are simple and inexpensive, and, while maternal heart rate may occasionally be counted, simple estimation of the fetal heart rate can be reliably performed. They also cause less maternal discomfort than the Pinard stethoscope.8
The aim of this study was to compare four different methods of monitoring fetal heart rate in labour with regard to their ability to detect abnormalities in fetal heart rate and their effect on neonatal mortality and short term neonatal morbidity.
Subjects and methods
The study was performed at Harare Maternity Hospital, Harare, Zimbabwe. This is a referral hospital for clinics within about 20 km radius of Harare for all high risk pregnancies. Indications for booking at Harare include obstetric and medical risk factors such as hypertension, antepartum haemorrhage, post-term pregnancy, poor obstetric history, and diabetes. Women may also be referred from surrounding clinics for any problems noted during pregnancy and labour. These include delay in the progress of labour and suspected cephalopelvic disproportion, meconium staining of the liquor, fetal distress, and hypertension. Electronic monitoring is rarely available in our labour ward, and there are no facilities for fetal blood sampling.
About 18 000 deliveries occur in the hospital annually. Any of these women seen in the maternity admission room or in the labour ward were recruited into the study provided gestation was over 37 weeks, presentation was cephalic, fetal heart rate was normal on admission, and the cervix was dilated less than 7 cm. The only high risk women excluded from the study are those presenting with placental abruption or eclampsia. The study was approved by the departmental board, and informed consent was obtained from mothers before they entered the study.
Eligible women were randomly allocated to one of four methods of monitoring intrapartum fetal heart rate by means of serially numbered sealed opaque envelopes containing the allocation. The randomisation was performed with a random permuted block of 16 numbers. Sample size was calculated to show with 95% confidence whether Doppler ultrasound scanning was at least 90% as effective as electronic monitoring in detecting abnormalities in fetal heart rate. We accepted a 20% risk of failing to show ultrasonography to be at least 90% effective when it really was. Thus, with alpha=0.05, and beta=0.2, we needed a sample size of 330 women in each group.
of fetal monitoring
Fetal heart rate was monitored by intermittent electronic monitoring, hand held Doppler ultrasonography, and use of the Pinard stethoscope. Research midwives monitored or directly supervised the monitoring of the fetal heart rate. Heart rate was also monitored according to routine clinical practice: the Pinard stethoscope was again used but by the midwife on duty in the ward.
Electronic monitoring - An external transducer was applied to the abdomen of women recruited to the study, and a continuous trace was recorded for 10 minutes in every half hour if results were normal or every 20 minutes if results were abnormal. A research midwife always observed the monitoring. The research midwives and the doctors in the labour ward were trained in detection of decelerations, and before the start of the study we were satisfied with midwives' ability to detect prolonged and late decelerations as described in the protocol. Regrettably, we did not request assessment of baseline variability.
Doppler ultrasound scanning - Huntleigh pocket Doppler ultrasound monitors were used to listen to the fetal heart. A research midwife would listen to the fetal heart rate during the last 10 minutes of every half hour, particularly before and immediately after a contraction.
Pinard stethoscope - A research midwife ensured that fetal heart rate was recorded during the last 10 minutes of every half hour with the Pinard stethoscope, particularly during and immediately after a contraction.
Routine monitoring - The midwife on duty recorded fetal heart rate with the Pinard stethoscope as was normal practice in the ward. Fetal heart rate is normally supposed to be recorded during the last 10 minutes of every half hour.
Midwives were required to report any abnormalities in fetal heart rate to the doctor in charge. Before the start of the study labour ward staff were generally advised that, irrespective of the baseline variability, a caesarean section should be performed if there were any decelerations lasting for more than 30 seconds or if there were persistent late decelerations unless vaginal delivery was imminent. Only one operating theatre was available to our busy unit, and so we expected delays if more than one woman required operative delivery at about the same time.
Data were analysed with the Epi-Info statistical package and comparisons were made with P values at 5% significance level and relative risk with 95% confidence intervals.
Of the 1255 women recruited to the study, 318 were allocated to intermittent electronic fetal monitoring, 312 to intermittent Doppler ultrasound scanning, 310 to monitoring with the Pinard stethoscope, and 315 to routine monitoring. Table I shows the indications for booking at Harare Maternity Hospital and for referral from surrounding clinics. Table II shows that baseline maternal and fetal characteristics were similar in the four groups (differences not significant at the 5% level). Antenatal problems included any condition for which a woman sought medical treatment, but these have not been listed as there were over 40 different conditions: the main problems were hypertension (67, 63, 58, and 65 of the women monitored electronically, by ultrasound scanning, by the Pinard stethoscope, and by routine methods respectively), urinary tract infection (13, 13, 14, and 4 women respectively), and post-term pregnancy (13, 16, 10, and 6 women respectively).
For 18 of the women allocated to electronic monitoring, delivery was either too rapid for monitoring or it was prevented by technical problems with the machine, but these women have been included in the group for statistical analysis. In 24 cases traces were of poor quality, usually because of frequent loss of contact, machine failure, or poor marking on the paper. However, as guidelines for intervention were quite crude - relying on the presence of late decelerations rather than baseline variability - this did not affect the clinical management in labour in any of the cases. All the women in the other groups were monitored as allocated.
Despite the provision of guidelines for appropriate actions to be taken when an abnormal fetal heart rate was detected, we noted three problems. Failure to perform caesarean section despite indications for the operation occurred in similar proportions of the four groups (8 (3%), 10 (3%), 15 (5%), and 12 (4%) of women monitored electronically, by ultrasonography, by the Pinard stethoscope, and by routine methods respectively). In a few cases in each group (7 (2%), 9 (3%), 5 (2%), and 5 (2%)) decisions were delayed, all apparently because no doctor was available in the labour ward at the relevant time. There were also delays in operations in each group (10 (8%), 4 (1%), 5 (2%), and 5 (2%)).
Detection of abnormal fetal heart rate
Some abnormality in fetal heart rate was detected in 172 (54%) of the electronically monitored group, 100 (32%) of the Doppler ultrasonography group, 47 (15%) of the Pinard stethoscope group, and 28 (9%) of the routine monitoring group. Compared with routine monitoring, the relative risk (95% confidence interval) of detecting abnormal fetal heart rate was 6.1 (4.2 to 8.8) with electronic monitoring, 3.6 (2.4 to 5.3) with Doppler ultrasonography, and 1.7 (1.1 to 2.7) with the Pinard stethoscope. All these differences were significant. Prolonged early and late decelerations were detected in 66 (21%), 62 (20%), 30 (10%), and 23 (7%) of the electronic monitoring, ultrasonography, Pinard stethoscope, and routine monitoring groups respectively. However, the time between detection of the abnormality and delivery was similar in the four groups, being 60-90 minutes in 30% of such cases in each group.
Labour and delivery
There was no significant difference in the mean duration of labour in the four groups and in 17-19% of cases in each group labour lasted more than 18 hours (table III). Similarly, among the women who delivered vaginally, there was a high proportion (18%, 16%, 18%, and 15% in each group) in whom the second stage of labour lasted for more than 30 minutes from when they started to push.
Delivery by caesarean section was more common in the electronic monitoring group (89 (28%) of the women) and the ultrasonography group (76 (24%)) than in the Pinard stethoscope group (32 (10%)) and the routine monitoring group (46 (15%)). Compared with the routine monitoring group, the relative risk (95% confidence interval) of delivery by caesarean section was 1.8 (1.4 to 2.3) with electronic monitoring, 1.6 (1.2 to 2.0) with ultrasonography, and 0.9 (0.6 to 1.2) with the Pinard stethoscope. Fetal distress was the indication for caesarean section for 56 (63%) of those who underwent the operation in the electronic monitoring group and 51 (67%) of those in the ultrasonography group, significantly higher proportions than in the Pinard stethoscope group (13 (41%)) and the routine monitoring group (19 (41%)). Caesarean sections for other intrapartum emergencies were equally common in the groups. There was no difference between the groups in terms of vacuum delivery.
Table IV shows the neonatal morbidity and mortality in the groups. The proportion of infants with low Apgar scores at five minutes after birth was slightly higher in the Pinard stethoscope group and the routine monitoring group. Compared with the routine monitoring group, the relative risk of an Apgar score below 6 was 0.7 (0.2 to 1.8) for the electronic monitoring group, 0.3 (0.1 to 1.2) for the ultrasonography group, and 0.9 (0.4 to 2.3) for the Pinard stethoscope group; differences were not significant. Admission to the neonatal unit was lowest in the ultrasonography group, with a risk of 0.6 (0.4 to 0.9) relative to the routine monitoring group. Neonatal seizures were recorded in the Pinard stethoscope group (six) and routine monitoring group (nine) only. There were two cases of hypoxic ischaemic encephalopathy in the electronic monitoring group and one in the ultrasonography group but seven and 10 in the Pinard stethoscope and routine monitoring groups.
In the electronic monitoring group there were three stillbirths, all because of poor adherence to the study protocol, and five neonatal deaths from hypoxia caused by delays in delivery, either because of a delayed decision to operate or because the operating theatre was not available. In the ultrasonography group there were two deaths: one stillbirth, due to non-adherence to the protocol and an inappropriate attempt at operative vaginal delivery, and one infant born severely hypoxic because the operating theatre was initially unavailable. The five deaths in the Pinard stethoscope group were all because of asphyxia resulting from delays in delivery. Of the nine deaths in the routine monitoring group, only tow were due to inappropriate management or a delay in delivery: in the other seven cases the fetal heart rate had been recorded as normal before delivery of a fresh stillbirth or a severely asphyxiated baby.
Harare Maternity Hospital is a teaching hospital. The standard management of women in labour includes use of a partogram, and oxytocin is frequently used if the rate of cervical dilatation is too slow and there is no obvious evidence of cephalopelvic disproportion.9,10 Epidural analgesia is not available in the labour ward. Since electronic fetal monitoring is not generally available, in this study we had to give guidelines on the appropriate action to be taken on the basis of fetal heart rate abnormality. The final management decision was, however, left to the team on call in the labour ward that day. There were no facilities to perform fetal blood sampling.
Electronic monitoring of fetal heart rate identified nearly twice as many fetuses with an abnormal pattern as did the other methods, but many of these were mild or variable decelerations or tachycardias alone. The number of fetuses with detected prolonged or late decelerations (the indications for intervention) was only slightly greater in the electronic monitoring group than in the ultrasonography group but was substantially greater than in the two groups being monitored with the Pinard stethoscope. This is reflected in the greater number of caesarean deliveries in the first two groups. The rate of caesarean section in the electronic monitoring and ultrasonography groups was higher than the overall rate of 18% for the hospital and could have been due to the large number of primigravid women in the study (randomisation ensured equal numbers of primigravidas in all four groups). It could also be due to the absence of fetal blood sampling to make a definitive diagnosis of fetal hypoxia. Investigators have previously noted the high rate of false positive diagnosis of fetal distress with abnormal patterns. Only 20% to 40% of ominous patterns are thought to be associated with fetal pH values that are considered acidotic.11 This increased rate of operative delivery in the electronically monitored group has also been noted in other studies.4,5
Although there was no significant difference between the groups in neonatal Agar scores, there were more babies with low Apgar score in the two groups monitored with the Pinard stethoscopes. Significantly fewer babies in the ultrasonography group needed admission to the neonatal unit compared with the other groups. Neonatal seizures occurred only in the groups monitored with the Pinard stethoscope. This finding concurs with the results of the pooled analysis conducted by Chalmers12 and the results from the Dublin trial.6
There was, however, no significant difference between the electronic monitoring and ultrasonography groups in terms of neonatal seizures and hypoxic ischaemic encephalopathy. This may be due to the small sample size of our study, but it does indicate that the two methods were similar in their capacity to detect important abnormalities in fetal heart rate. Exclusion of preterm infants, with the resultant reduction in corrected neonatal morbidity and mortality to low levels, would increase the need for a larger sample size. Although electronic monitoring was more sensitive in picking up any fetal heart rate abnormality, there were still two infants born with hypoxic ischaemic encephalopathy. These were due to delays in performing operative deliveries and emphasise the need for appropriate and timely action irrespective of what method is used for monitoring fetal heart rate.Several of the neonatal deaths that occurred were due to delays in action being taken, and this was certainly the explanation for the apparently better outcome for the ultrasonography group compared with the electronic group in terms of perinatal deaths. Regular training of midwives can help to ensure that decisions and appropriate action are made promptly.
Two obvious weaknesses of this study are that we did not request detection of reduced variability of fetal heart rate, a potential advantage of electronic monitoring over the other methods, and our inability to confirm fetal distress by fetal blood sampling. Nevertheless, our results show that Doppler ultrasound scanning was better than the Pinard stethoscope in detecting abnormalities in fetal heart rate and was associated with fewer babies with seizures and hypoxic encephalopathy and with lower neonatal mortality. Perinatal outcome was no worse than that achieved with electronic monitoring, and the rates of caesarean section with the two methods were similar. We therefore conclude that, when fetal acidosis cannot be confirmed by fetal blood sampling and reliance must be placed on abnormalities in fetal heart rate, asphyxia can be detected with a hand held Doppler ultrasound scanner just as reliably as with intermittent electronic monitoring. The use of such ultrasound scanners should be promoted in obstetric units in the developing world that care for high risk pregnancies.
We thank all the labour ward staff for their cooperation during the period of the study. This study was funded by World Health Organisation Special Programme for Research, Development, and Research Training in Human Reproduction.