Summary of findings
The baby's heart beat was first thought to be heard in utero in the middle of the seventeenth or eighteenth century (Grant 1989a; Gibb 1992), but it was not until the early nineteenth century that de Kergeradee suggested that listening to the baby's heartbeat might be clinically useful (Grant 1989a). He proposed that it could be used to diagnose fetal life and multiple pregnancies, and wondered whether it would be possible to assess fetal compromise from variations in the fetal heart rate (FHR). Since then, various methods of listening to the fetal heart have been developed and introduced into maternity care (see Table 1), each with the aim of improving outcomes for babies and reducing the heartache for mothers and families when a baby dies or suffers long-term disability. Today, monitoring the fetal heart during labour, by one method or another, appears to have become a routine part of care during labour, although access to such care varies across the world.
Description of the condition
The incidence of neonatal morbidity and mortality varies around the world, although direct comparisons may be difficult because of varying definitions and classifications. Nevertheless, large differences are reported between high-income countries with average neonatal mortality rates (NMR) of four per 1000 live births) and low/middle-income countries with average NMRs of 33 per 1000 births) (Lawn 2005). Although the majority of perinatal morbidity and mortality may not be prevented by improved fetal monitoring in labour (Nelson 1996), failure in identifying abnormal FHR patterns and lack of appropriate actions are considered to be significant contributing factors (MCHRC 1997; MCHRC 1998; MCHRC 1999).
Description of the intervention
The baby's heart rate can be monitored either intermittently (at regular intervals during labour) or continuously (recording the baby's heart rate throughout labour, stopping only briefly, e.g. for visits to the toilet) as follows.
(1) Fetal stethoscope (Pinard) and hand-held Doppler
Intermittent monitoring can be undertaken either by listening to the baby's heart rate using a fetal stethoscope (Pinard), or with a hand-held Doppler ultrasound device, and by palpating the mother's uterine contractions by hand. This is known as 'intermittent auscultation'.
(2) Cardiotocograph (CTG)
The baby's heart rate and the mother's uterine contractions can be recorded electronically on a paper trace known as a cardiotocograph. This is done by using a Doppler ultrasound transducer to monitor the baby's heart rate and a pressure transducer to monitor uterine contractions, both of which are linked to a recording machine. This is known as external cardiotocography (external CTG). This is usually undertaken continuously in labour, although occasionally it is used intermittently (intermittent CTG). In most units, external CTG requires the mother to wear a belt across her abdomen while monitoring is being conducted, which restricts her mobility. An alternative means of monitoring the baby's heart rate with the CTG machine is to attach an electrode directly to the baby's presenting part, usually its head. This form of continuous monitoring is known as 'internal CTG' and requires a ruptured amniotic sac (either spontaneously or artificially) and a scalp electrode (clip) attached to the baby's head. This also restricts the woman's mobility.
The term 'electronic fetal monitoring' (EFM) is sometimes used synonymously with CTG monitoring, but is considered to be a less precise term because (1) CTG monitoring also includes monitoring the mother's contractions, and (2) other forms of fetal monitoring might also be classed as 'electronic', e.g. fetal ECG, fetal pulse oximetry.
Intermittent auscultation was the predominant method of monitoring during labour until CTGs became widely used in the latter part of the twentieth century (Enkin 2000). Although there is a lack of empirical evidence on the optimal frequency of intermittent auscultation, there is a consensus in the guidelines from professional bodies that the fetal heart should be auscultated at least every 15 minutes in the first stage of labour and at least every five minutes in the second stage of labour (ACOG 1995; Liston 2002; NCCWCH 2008; RANZCOG 2002) with each auscultation lasting at least 60 seconds (Liston 2002; NCCWCH 2008). It appears that these auscultation protocols were developed initially in the context of clinical trials and were based on 'common sense' rather than research evidence. Compliance with these guidelines, whilst maintaining contemporaneous records, poses a significant challenge for caregivers during labour who usually have multiple tasks to fulfil simultaneously.
Information and interpretation
Both intermittent auscultation and CTG provide information on the baseline heart rate (usually between 110 and 160 beats per minute in the term fetus), accelerations (transient increases in the FHR) and decelerations (transient decreases in the FHR). It is known that some aspects of labour will cause natural alterations in FHR patterns. For example, the baby's sleep FHR pattern is different from the FHR pattern when the baby is awake. External stimuli, like uterine contractions and the mother moving, can cause FHR changes, as can the administration of opiates to the mother. Some of these changes are subtle and can only be detected by continuous CTG, e.g. baseline variability, temporal shape of decelerations. Consideration needs to be given to whether such information improves detection and outcome of those babies who are truly compromised and if there are disadvantages with the technology for those who are not compromised.
Sensitivity and specificity
While specific abnormalities of the FHR pattern on CTG are proposed as being associated with an increased risk of cerebral palsy (Nelson 1996), the specificity of CTG for prediction of cerebral palsy is low with a reported false positive rate as high as 99.8%, even in the presence of multiple late decelerations or decreased variability (Nelson 1996).
FHR pattern recognition, including the relationship between the uterine contractions and FHR decelerations, are fundamental to the use of continuous CTG monitoring. Algorithms have been developed to assess and record what is normal, what requires more careful attention and what is considered abnormal requiring immediate delivery of the baby (NCCWCH 2008). However, CTG traces are often interpreted differently by different caregivers (inter-observer variation) and even by the same caregiver interpreting the same record at different times (intra-observer variation) (Devane 2005). Such variation in interpretation of CTG tracings may result in inappropriate interventions, or false reassurance and lack of appropriate intervention. Although we were unable to locate studies that sought to investigate inter- and intra-observer variation in intermittent auscultation, it would seem reasonable to suggest that intermittent auscultation is not immune to similar problems caused by inter- and intra-observer variation. However, given that the FHR parameter of interest in intermittent auscultation is the baseline FHR, it is likely that inter- and intra-observer variation is less in intermittent auscultation than that found in CTG interpretation where other aspects of FHR patterns including variability and assessment and deceleration classification require interpretation.
Fetal blood sampling is a procedure whereby a small amount of blood is taken from the baby, usually from the scalp. Performing fetal blood sampling and then measuring the parameters of acid-base balance (pH, base excess/deficit, etc) seeks to identify those babies who are truly compromised and need to be born immediately from those who are not truly compromised. It is important to establish the value of this test as an adjunct to CTG. This question is addressed by a subgroup analysis in this review.
Other methods have been considered as additional tests, but there is little evidence to support their use, for example, vibroacoustic stimulation (East 2013). Several other methods of fetal monitoring have been proposed, either as an adjunct or an alternative to CTG, e.g. pulse oximetry (Carbone 1997; East 2007), near-infrared spectroscopy (Mozurkewich 2000), fetal ECG (Neilson 2012), ST segment analysis of the fetal ECG (Luttkus 2004) and fetal stimulation tests (Skupski 2002).
Possible advantages of CTG
- More measurable parameters related to FHR patterns (see above).
- The CTG trace gives a continuous recording of the FHR and uterine activity. This is a physical record, which can be examined at anytime in labour, or subsequently if required. The examples where physical records may be useful include clinical audits, counselling parents if there has been as adverse outcome, and medico-legal situations.
Possible disadvantages of CTG
- The complexity of FHR patterns makes standardisation difficult.
- CTG prevents mobility and restricts the use of massage, different positions and/or immersion in water used to improve comfort, control and coping strategies during labour.
- Shifting staff focus and resources away from the mother may encourage a belief that all perinatal mortality and neurological injury can be prevented.
Specific situations that may influence the effectiveness or otherwise of CTG
- Induction of labour is primarily performed where it is anticipated that the outcome for the mother and/or infant would be improved were labour to be induced. Given that induction of labour includes iatrogenic stimulation of uterine activity, which puts the baby at greater risk, we determined to perform a subgroup analysis by induction of labour (RCOG 2001b).
- Preterm birth is associated with an increased risk of mortality and neurological morbidity and these babies might benefit from being monitored more intensively. Further, there is debate about what is normal for the different parameters of the CTG for preterm infants at varying gestational ages. Therefore, we performed a preterm subgroup analysis.
- Twin pregnancies carry a higher perinatal mortality rate than singleton pregnancies (RCOG 2001b), thus we conducted a subgroup analysis by twin pregnancy.
Women's and professional views
Some studies looking at women's preferences found that the support that women received from staff and labour companions was more important to them than the type of monitoring used (Garcia 1985; Killien 1989). A more recent study of women's views of routine continuous CTG in labour in the UK identified a lack of discussion about the need for, and appropriateness of CTG. In addition, women felt that CTG limited their mobility and led to an acceptance of the machine's place as the focus of attention for the women and her partner (Munro 2004).
In a synthesis of 11 studies on professionals’ views of monitoring the FHR during labour, Smith 2012 identified that, despite an absence of evidence, maternity care professionals perceive the CTG as offering 'proof' of the compromised baby and that this minimises their exposure to criticism and potential litigation. Nevertheless, professionals also recognised that the CTG offered a false sense of security.
How the intervention might work
Through monitoring FHR changes during labour, it is hoped to identify those babies who may be compromised, or potentially compromised, by a shortage of oxygen (fetal hypoxia). If the shortage of oxygen is both prolonged and severe, babies are at risk of being born with a disability (physical and/or mental), or of dying during labour or shortly thereafter. When alterations in the FHR during labour suggest that the baby is hypoxic, or at risk of hypoxia, additional methods of assessment of fetal well-being (e.g. fetal blood sampling) may be used. Sometimes FHR alterations trigger delivery by caesarean section or by an instrument such as forceps or vacuum extractor even without recourse to additional diagnostic tests.
Why it is important to do this review
Concerns have been raised about the efficacy and safety of routine use of continuous CTG in labour (Thacker 1995). The apparent contradiction between the widespread use of continuous CTG with claims of its effectiveness in lowering early neonatal mortality and morbidity (Chen 2011) and recommendations to limit its routine use on all women (NCCWCH 2008), indicates that a regular reassessment of this practice is warranted.
Several other Cochrane reviews have addressed other methods for assessing the condition of the fetus during labour including fetal electrocardiogram/ECG (Neilson 2012); fetal pulse oximetry (East 2007); near-infrared spectroscopy (Mozurkewich 2000) and vibroacoustic stimulation (East 2013). Also, the comparison of cardiotocography versus intermittent auscultation of fetal heart as an admission test on arrival to labour ward is assessed elsewhere (Devane 2012).
The objective of this review is to evaluate the effectiveness and safety of continuous cardiotocography (CTG) when used as a method to monitor fetal well-being during labour.
Criteria for considering studies for this review
Types of studies
All randomised trials and quasi-randomised studies comparing continuous CTG during labour, with and without fetal blood sampling, with (a) no fetal monitoring, (b) intermittent auscultation of the fetal heart rate with a Pinard stethoscope or hand-held Doppler ultrasound device, or (c) intermittent CTG. Sensitivity analysis was undertaken for studies graded as low risk of bias based on sequence generation and allocation concealment. .
Types of participants
Pregnant women in labour and their babies.
Types of interventions
The main intervention of interest is continuous CTG during labour.
For the purpose of this review, the intervention is defined as an attempt to produce a continuous and simultaneous hard-copy recording of the fetal heart rate and uterine contractions in real-time throughout the woman's labour. As a guide, continuous CTG should be discontinued only for short periods (for example, visits to the toilet) and the CTG should be used for clinical decision making during labour.
Control groups of interest include: (a) no fetal monitoring, (b) intermittent auscultation of the fetal heart rate with a Pinard stethoscope or hand-held Doppler ultrasound device, or (c) intermittent CTG.
Types of outcome measures
- Perinatal mortality;
- seizures in the neonatal period, either apparent clinically or detected by electro-encephalographic recordings;
- cerebral palsy;
- caesarean section;
- instrumental vaginal birth;
- cord blood acidosis (low pH/low base excess as defined by trialists; where report included a range of pH values we have used cord pH less than 7.10 as a cut off for acidosis);
- use of all forms of pharmacological analgesia during labour and birth (including epidural but excluding anaesthesia for caesarean section).
- Hypoxic ischaemic encephalopathy (as defined by trialists);
- neurodevelopmental disability assessed at 12 months of age or more. Neurodevelopmental disability, defined as any one or combination of the following: non-ambulant cerebral palsy, developmental delay, auditory and visual impairment. Development should have been assessed by means of a previously validated tool, such as Bayley Scales of Infant Development (Psychomotor Developmental Index and Mental Developmental Index (Bayley 1993);
- Apgar less than seven at five minutes;
- Apgar less than four at five minutes;
- admission to neonatal special care and/or intensive care unit;
- fetal blood sampling;
- damage/infection to baby's head from scalp electrode or fetal blood sampling;
- caesarean section for abnormal fetal heart rate pattern and/or fetal acidosis;
- instrumental vaginal birth for abnormal fetal heart rate pattern and/or fetal acidosis;
- spontaneous vaginal birth not achieved;
- epidural analgesia;
- use of non pharmacological methods of coping with labour, e.g. transcutaneous electrical nerve stimulation, hydrotherapy;
- amniotomy (artificial rupture of membranes);
- oxytocin during labour;
- perineal trauma requiring repair (including episiotomy);
- inability to adopt preferred position during labour;
- dissatisfaction with labour and/or perceived loss of control during labour;
- postpartum depression;
- exclusively breastfeeding at discharge from hospital;
- length of stay in neonatal special care and/or intensive care unit.
Search methods for identification of studies
We searched the Cochrane Pregnancy and Childbirth Group’s Trials Register by contacting the Trials Search Co-ordinator (31 December 2012).
The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co-ordinator and contains trials identified from:
- monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
- weekly searches of MEDLINE;
- weekly searches of EMBASE;
- handsearches of 30 journals and the proceedings of major conferences;
- weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.
Details of the search strategies for CENTRAL, MEDLINE and EMBASE, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group.
Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co-ordinator searches the register for each review using the topic list rather than keywords.
For details of additional author searches carried out for the 2006 update of this review, please see Appendix 1. We chose not to repeat these additional searches for the update as they yielded no additional studies in the original search.
Searching other resources
We searched the reference lists of retrieved studies.
We did not apply any language restrictions.
Data collection and analysis
For all included studies in this review we have used the methodology described by Higgins 2011. In previous versions of the review, we used the Cochrane methodology current at the time, with Higgins 2008 being used in the 2008 version and Higgins 2005 being used in the 2006 version, see Appendix 2, (Alfirevic 2006).
Selection of studies
Two review authors independently assessed for inclusion all the potential studies we identified as a result of the search strategy. We resolved any disagreement through discussion or, if required, we consulted the third author.
Data extraction and management
We designed a form to extract data. At least two review authors extracted the data using the agreed form. We resolved discrepancies through discussion or, if required, we consulted the third author. Data were entered into Review Manager software (RevMan 2012) and checked for accuracy.
When information regarding any of the above was unclear, we attempted to contact the authors of the original reports to provide further details.
Assessment of risk of bias in included studies
Two review authors independently assessed risk of bias for each included study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Any disagreement was resolved by discussion or by involving the third author.
(1) Sequence generation (checking for possible selection bias)
We described, for each included study, the methods used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.
We assessed the methods as:
- low risk of bias (e.g. random number table; computer random number generator);
- high risk of bias (e.g. odd or even date of birth; hospital or clinic record number);
- unclear risk of bias.
(2) Allocation concealment (checking for possible selection bias)
We described, for each included study, the method used to conceal the allocation sequence in sufficient detail and determine whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.
We assessed the methods as:
- low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
- high risk of bias (e.g. open random allocation; unsealed or non-opaque envelopes; alternation; date of birth);
- unclear risk of bias.
(3.1) Blinding (checking for possible performance bias)
We described, for each included study, all the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We have also provided any information relating to whether the intended blinding was effective. We noted where there has been partial blinding (e.g. where it has not been possible to blind participants but where outcome assessment was carried out without knowledge of group assignment). Where blinding was not possible, we assessed whether the lack of blinding was likely to have introduced bias.
We assessed the methods as:
- low, high or unclear risk of bias for participants;
- low, high or unclear risk of bias for personnel;
- low, high or unclear risk of bias for outcome assessors.
The study was judged to be ‘low risk’ when there was blinding, or where, in our judgement, the outcome or the outcome measurement was not likely to have been influenced by lack of blinding.
(3.2) Blinding of outcome assessment (checking for possible detection bias)
We described for each included study the methods used, if any, to blind outcome assessors from knowledge of which intervention a participant received. We considered that studies were at low risk of bias if they were blinded, or if we judged that the lack of blinding could not have affected the results.
We assessed the methods as:·
- low risk (no blinding of outcome assessment but the authors judged that the outcome was not likely to be influenced by this);
- high risk (no blinding of outcome assessment and the outcome measurement was likely to have been influenced by this);
- unclear risk (insufficient information to permit judgment; the study did not address this).
(4) Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations)
We described for each included study the completeness of outcome data for each main outcome, including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported, the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and any re-inclusions in analyses which we undertook.
We assessed the methods as:
- low risk (e.g. where there were no missing data or where reasons for missing data are balanced across groups);
- high risk (e.g. where missing data are likely to be related to outcomes or are not balanced across groups, or where high levels of missing data are likely to introduce serious bias or make the interpretation of results difficult);
- unclear risk of bias (e.g. where there is insufficient reporting of attrition or exclusions to permit a judgement to be made).
(5) Selective reporting bias
We described for each included study how the possibility of selective outcome reporting bias was examined by us and what we found.
We assessed the methods as:
- low risk of bias (where it is clear that all of the study’s pre-specified outcomes and all expected outcomes of interest to the review have been reported);
- high risk of bias (where not all the study’s pre-specified outcomes have been reported; one or more reported primary outcomes were not pre-specified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);
- unclear risk of bias.
(6) Other sources of bias
We described for each included study any important concerns we have about other possible sources of bias. For example, was there a potential source of bias related to the specific study design? Was the trial stopped early due to some data-dependent process? Was there extreme baseline imbalance? Has the study been claimed to be fraudulent?
We assessed whether each study was free of other problems that could put it at risk of bias:
- low risk of bias;
- high risk of bias;
- unclear risk of bias.
(7) Overall risk of bias
We made explicit judgements about risk of bias for important outcomes both within and across studies. With reference to (1) to (6) above, we assessed the likely magnitude and direction of the bias and whether we considered it was likely to impact on the findings. We explored the impact of the level of bias through undertaking sensitivity analyses - see Sensitivity analysis.
Measures of treatment effect
For dichotomous data, we presented the results as risk ratio (RR) with 95% confidence intervals (95% CI).
For continuous data, we presented the results as mean difference (MD) with 95% CI. When pooling data across studies we estimated the MD if the outcomes were measured in the same way between trials. We planned to use standardised mean difference to combine trials that measured the same outcome, but used different methods.
Unit of analysis issues
We have not identified any cluster trials as yet, and do not anticipate any cluster-randomised trials on this topic. However, should we find any, we will include them using the guidance in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Dealing with missing data
We noted the levels of attrition for included studies. We explored, using sensitivity analysis, the impact of including studies with high levels of missing data in the overall assessment of treatment effect.
Intention-to-treat analysis (ITT)
For all outcomes, we analysed the data, as far as possible, on an intention-to-treat basis, i.e. we attempted to include all participants randomised to each group in the analyses. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes are known to be missing ('available case' analysis).
We analysed data on all participants with available data in the group to which they are allocated, regardless of whether or not they received the allocated intervention. If in the original reports participants were not analysed in the group to which they were randomised, and there was sufficient information in the trial report, we attempted to restore them to the correct group.
Assessment of heterogeneity
We assessed statistical heterogeneity in each meta-analysis using the T², I² and Chi² statistics. We regarded heterogeneity as substantial if I² is greater than 50% and either T² is greater than zero, or there is a low P value (less than 0.10) in the Chi² test for heterogeneity.
Assessment of reporting biases
We investigated reporting biases (such as publication bias) using funnel plots for outcomes with 10 or more studies in the meta-analysis. We assessed funnel plot asymmetry visually. If asymmetry was suggested by a visual assessment, we planned to perform exploratory analyses to investigate it.
We only synthesised data where we thought it was appropriate, i.e. where trials were examining similar interventions, with similar populations and methods. We used a fixed-effect meta-analysis for combining data. However, where there was clinical or methodological heterogeneity between studies sufficient to suggest that treatment effects might differ between trials, we used random-effects meta-analysis.
Subgroup analysis and investigation of heterogeneity
We carried out the following subgroup analyses:
- high risk for perinatal mortality and morbidity (as defined by trialists) versus low risk (absence of identifiable risk factors associated with increased in perinatal mortality and morbidity as defined by trialists);
- spontaneous onset of labour versus induction of labour;
- preterm (less than 37 + 0 weeks) versus term (greater than 37 + 0 weeks);
- singleton pregnancy versus twin pregnancy;
- access to fetal blood sampling (FBS) during labour versus no access to FBS during labour;
- primiparous versus multiparous.
The review’s primary outcomes were used in subgroup analyses.
We assessed subgroup differences by interaction tests available within RevMan (RevMan 2012). We reported the results of subgroup analyses quoting the χ2 statistic and P values, and the interaction test I² value. We assumed a P value of more than 0.05 indicative that subgroup results are no different from the overall effect. If a P value was less than 0.05 we did not automatically assume that the intervention has a different effect, but planned to interpret the findings in the clinical context and in the context of plausibility.
We carried out sensitivity analyses to explore the effect of trial quality for the primary outcomes in the review. Where there was risk of bias associated with a particular aspect of study quality (e.g. inadequate allocation concealment), this was explored by sensitivity analyses.
Description of studies
Results of the search
Our search strategy identified 383 citations corresponding to 17 studies for potential inclusion. Of those, 13 studies with 37,715 women participating were included (Athens 1993; Copenhagen 1985; Dallas 1986; Denver 1976; Denver 1979; Dublin 1985; Lund 1994; Melbourne 1976; Melbourne 1981; New Delhi 2006; Pakistan 1989; Seattle 1987; Sheffield 1978) and four were excluded (Harare 1994; Ioannina 2001; Manchester 1982; North America 2000).
Twelve studies, with 33,681 women participating, compared continuous CTG with intermittent auscultation (Athens 1993; Copenhagen 1985; Dallas 1986; Denver 1976; Denver 1979; Dublin 1985; Melbourne 1976; Melbourne 1981; New Delhi 2006; Pakistan 1989; Seattle 1987; Sheffield 1978). Five studies compared continuous CTG plus fetal blood sampling versus intermittent auscultation (Copenhagen 1985; Dublin 1985; Melbourne 1976; Pakistan 1989; Seattle 1987), six compared continuous CTG without fetal blood sampling versus intermittent auscultation (Athens 1993; Dallas 1986; Denver 1976; Melbourne 1981; New Delhi 2006; Sheffield 1978) and one study had three groups comparing continuous CTG with and without fetal blood sampling versus intermittent auscultation (Denver 1979). One study compared continuous CTG with fetal blood sampling versus intermittent CTG with fetal blood sampling (Lund 1994).
Participants were assessed as being at low risk of complications in three studies (Dallas 1986; Lund 1994; Melbourne 1981) and outcome data for women at low risk were available for one outcome, neonatal seizures, from one other study (Dublin 1985). Participants were assessed as being at high risk of complications in six studies (Denver 1976; Denver 1979; Melbourne 1976; New Delhi 2006; Pakistan 1989; Seattle 1987) including one study that specifically included women in preterm labour (28 to 32 weeks) and assessed outcomes for babies below 1750 g birthweight (Seattle 1987). The data for neonatal seizures in women at high risk of complications were available from one other study (Dublin 1985). Participants were assessed as mixed risk (mixture of women at high risk and low risk of complications) in four studies (Athens 1993; Copenhagen 1985; Dublin 1985; Sheffield 1978).
Five studies had overall caesarean section rates below 10% (Athens 1993; Copenhagen 1985; Dublin 1985; Melbourne 1981; Sheffield 1978). The highest overall caesarean section rates were reported in Pakistan 1989 (23.5%) and New Delhi 2006 (28%).
Four studies were excluded (see Characteristics of excluded studies).
Risk of bias in included studies
Included studies were assessed for methodological quality on the basis of selection (sequence generation and allocation concealment) and attrition bias (see 'Methods of the review' above) (Figure 1).
|Figure 1. Methodological quality summary: review authors' judgements about each methodological quality item for each included study.|
Allocation concealment was graded as 'low risk of bias' in three trials (Dublin 1985; Lund 1994; Melbourne 1976), as 'unclear risk of bias' in six trials (Copenhagen 1985; Denver 1976; Denver 1979; New Delhi 2006; Seattle 1987; Sheffield 1978), and as 'high risk of bias' in four trials (Athens 1993; Dallas 1986; Melbourne 1981; Pakistan 1989).
Blinding of participants and personnel was assessed as 'high risk of bias' in all 13 studies. Blinding of outcome assessment was assessed as 'unclear risk of bias' in all but one study where it was assessed as 'high risk of bias' (Athens 1993).
Incomplete outcome data
Attrition bias was graded as 'low risk of bias' in eight trials (Athens 1993; Copenhagen 1985; Denver 1976; Denver 1979; Dublin 1985; Lund 1994; New Delhi 2006; Pakistan 1989), as unclear risk of bias in three trials (Dallas 1986; Melbourne 1976; ; Sheffield 1978), as high risk of bias in two trials (Melbourne 1981; Seattle 1987).
This was assessed as 'unclear risk of bias' in all 13 studies as we did not have access to any of the trial protocols.
Other potential sources of bias
All 13 studies were considered at low risk for other potential sources of bias.
Effects of interventions
Continuous cardiotocography (CTG) versus intermittent auscultation (IA) (Comaprisons 1-8)
A total of 12 randomised trials were included in this comparison with over 33,000 women participating (Athens 1993; Copenhagen 1985; Dallas 1986; Denver 1976; Denver 1979; Dublin 1985; Melbourne 1976; Melbourne 1981; New Delhi 2006; Pakistan 1989; Seattle 1987; Sheffield 1978). Denver 1979 was a three-arm trial comparing continuous CTG alone, versus continuous CTG plus FBS versus intermittent auscultation.
For the infant
There was no significant difference in perinatal mortality between the groups. Risk ratio (RR) was 0.86 with, 95% confidence intervals (CIs) ranging from 0.59 to 1.24, n = 33,513, 11 trials, ( Analysis 1.1). The funnel plot analysis indicated no missing studies (Figure 2). The quality of the evidence for this outcome was assessed as moderate (Summary of findings table 1).
|Figure 2. Funnel plot of comparison: 1 Continuous CTG versus intermittent auscultation, outcome: 1.1 Perinatal mortality (primary outcome).|
The use of continuous CTG monitoring in labour halved the risk of neonatal seizures (RR 0.50, 95% CI 0.31 to 0.80, n = 32,386, nine trials ( Analysis 1.2). The funnel plot indicated no missing studies (Figure 3) and the quality of the evidence was assessed as moderate (Summary of findings table 1). This reduction was consistent across the trials and subgroups, although the incidence of neonatal seizures varied considerably between trials. In the two largest trials of 14,618 women (Dallas 1986) and 12,964 women (Dublin 1985), the incidence of neonatal seizures in the intermittent auscultation groups was 0.04% and 0.4% respectively ( Analysis 1.2). In the two trials of high quality reporting data for this outcome (Dublin 1985; Melbourne 1976), the risk of neonatal seizures was RR 0.40, 95% CI 0.21 to 0.77 ( Analysis 8.2).
|Figure 3. Funnel plot of comparison: 1 Continuous CTG versus intermittent auscultation, outcome: 1.2 Neonatal seizures (primary outcome).|
There was no difference in the incidence of cerebral palsy (average RR 1.75, 95% CI 0.84 to 3.63, n = 13,252, two trials, random-effects, ( Analysis 1.3). The quality of the evidence was assessed as moderate (Summary of findings table 1). The data on cerebral palsy are heavily influenced by one small trial (Seattle 1987) that randomised only very preterm babies (less than 32 weeks) and assessed outcomes for 173 babies of birthweight less than 1750 g with a cerebral palsy rate of 19.5% in the CTG group compared with 7.7% in the controls (RR 2.54, 95% CI 1.10 to 5.86). The other trial in this comparison (Dublin 1985) showed no significant difference in the incidence of cerebral palsy (RR 1.20, 95% CI 0.52 to 2.79, n = 13,079) with a cerebral palsy rate of 0.18% in the continuously CTG group and 0.15% in the intermittently monitored group.
There was no difference in the incidence of cord blood acidosis between the groups ( Analysis 1.6). The quality of the evidence was assessed as very low, mainly due to severe heterogeneity and limitation in design of many of the included studies (Summary of findings table 1).
For the mother
There was a significant increase in the caesarean section rate in the CTG group (average RR 1.63, 95% CI 1.29 to 2.07, 18,861, 11 trials, ( Analysis 1.4). However, the quality of this evidence was assessed as low, mainly due to severe heterogeneity and limitation in design of many of the included studies (Summary of findings table 1). Risk difference in the caesarean section rate was 5% (95% CI 2 to 8%), with two-thirds of the data coming from Dublin 1985, where the overall caesarean section rate was 2.3%. In addition, the funnel plot indicated the possibility of missing studies (Figure 4). It appears that the risk of having a caesarean section was not influenced by the quality of trials ( Analysis 8.4, interaction test P = 0.31).
|Figure 4. Funnel plot of comparison: 1 Continuous CTG versus intermittent auscultation, outcome: 1.4 Caesarean section (primary outcome).|
Although numbers needed to treat to benefit or harm (NNTB/NNTH) analyses remain controversial in the context of meta-analysis and should be interpreted with caution, we have calculated that there will be one additional caesarean section for every 44 women monitored continuously (95% CI 26 to 96). This calculation is based on the pooled caesarean section rate of 3.6% (337/9313) in the IA group from this meta-analysis. However, in most settings caesarean section rates are likely to be much higher. Assuming a caesarean section rate with intermittent auscultation of around 15%, there would be an additional CS for every 11 women monitored (95% CI 7 to 23).
Continuous CTG was also associated with an increase in instrumental vaginal birth ( Analysis 1.5). The funnel plot indicated that some studies might be missing (Figure 5). The quality of this evidence was assessed as low, mainly due to severe heterogeneity and limitation in design of many of the included studies (Summary of findings table 1). There was no difference identified in the use of any pharmacological analgesia ( Analysis 1.7), with the quality of the evidence assessed as low (Summary of findings table 1).
|Figure 5. Funnel plot of comparison: 1 Continuous CTG versus intermittent auscultation, outcome: 1.5 Instrumental vaginal birth (primary outcome).|
For the infant
There was no evidence of any other benefit or harm for the babies in terms of hypoxic Ischaemic encephalopathy ( Analysis 1.8), Apgar scores ( Analysis 1.10), or admission to neonatal intensive care unit ( Analysis 1.12).
For the mother
Women in the continuous CTG group were more likely to have a caesarean section for abnormal fetal heart rate and/or acidosis ( Analysis 1.15) and less likely to have a spontaneous vaginal birth, Analysis 1.17).There was no difference in the use of epidural analgesia ( Analysis 1.18). The use of fetal blood sampling was reported in two trials (Copenhagen 1985; Dublin 1985) with significantly more sampling tests performed in the continuous CTG group ( Analysis 1.13). There were no reported data suitable for analysis for the use of non-pharmacological methods for coping with labour, amniotomy, perineal trauma, inability to adopt preferred position in labour, dissatisfaction in labour and postpartum depression.
Notwithstanding the caution regarding NNTB/NNTH calculations, when the risk of neonatal seizures is around 3 per 1,000, 667 women have to be continuously monitored during labour to prevent one such seizure (95% CI 484 to 1667). There is an opposite effect on caesarean section. Assuming a 3.6% caesarean section rate with IA, there would be 15 more caesarean sections in this cohort associated with preventing one neonatal seizure. However, if caesarean section with IA is higher (15%), 61 extra caesarean sections would be associated with preventing one neonatal seizure.
Continuous CTG versus IA (Subgroup: pregnancy risk status - high/low/unclear or both - Comparison 2)
Of the 12 studies that compared continuous CTG with intermittent auscultation, six included women at increased risk of complications (Denver 1976; Denver 1979; Melbourne 1976; New Delhi 2006; Pakistan 1989; Seattle 1987), three included women at low risk of complications (Dallas 1986; Melbourne 1981; Sheffield 1978) and three studies included both groups of women or did not specify (Athens 1993; Copenhagen 1985; Dublin 1985). There was a significant difference in the impact of CTG monitoring on caesarean section rate depending on the risk status of women (P = 0.004; I
Subgroups analysis by onset of labour (spontaneous/induced/unclear or both - Comparison 3)
None of the included trials provided separate data for spontaneous and induced labours. Hence, there is no information to determine if there might be a difference in the impact of CTG for women in spontaneous labour compared with those with induction of labour.
Subgroup analysis by gestational age (preterm/term/unclear or both - Comparison 4)
Of the 12 studies that compared continuous CTG with intermittent auscultation, one included only preterm labours (Seattle 1987). Three studies included only term labours (Copenhagen 1985; Melbourne 1981; Sheffield 1978) and eight studies included both or did not specify (Athens 1993; Dallas 1986; Denver 1979; Denver 1979; Dublin 1985; Melbourne 1976; New Delhi 2006; Pakistan 1989). We found no evidence of a difference between the subgroups.
Subgroup analysis by number of babies being monitored (singleton/twin pregnancy/unclear or both - Comparison 5)
Eight studies included only singleton pregnancies (Athens 1993; Dallas 1986; Denver 1976; Melbourne 1981; New Delhi 2006; Pakistan 1989; Seattle 1987; Sheffield 1978) and four included both singleton and twin pregnancies or did not specify (Copenhagen 1985; Denver 1979; Dublin 1985; Melbourne 1976). There was a significant subgroup effect for the rate of neonatal acidosis (P = 0.04; I
Subgroup analysis by access to fetal blood sampling (FBS) during labour (Comparison 6)
Six studies offered FBS alongside the CTG (Copenhagen 1985; Dublin 1985; Melbourne 1976; Melbourne 1981; Pakistan 1989; Seattle 1987), five studies did not use FBS (Athens 1993; Dallas 1986; Denver 1976; New Delhi 2006; Sheffield 1978) and one study randomised to three groups, CTG with FBS, CTG alone and IA (Denver 1979).
There was a significant subgroup effect on instrumental vaginal birth with apparently more instrumental deliveries (P = 0.04; I
Subgroups by parity (primiparous/multiparous women/unclear or both - Comparison 7)
None of the studies included only primiparous women, one study included only multiparous women (New Delhi 2006) and 11 studies included both primiparous and multiparous women (Athens 1993; Copenhagen 1985; Dallas 1986; Denver 1976; Denver 1979; Dublin 1985; Melbourne 1976; Melbourne 1981; Pakistan 1989; Seattle 1987; Sheffield 1978). As only one of these studies reported results based on the parity of the women involved, so it was no possible to perform a meaningful subgroup analysis.
Continuous CTG versus IA (Subgroup: high/low/unclear quality of studies - Comparison 8)
Of the 12 studies that compared continuous CTG with intermittent auscultation, two were considered to be of high methodological quality (Dublin 1985; Melbourne 1976), four studies where considered to be of low methodological quality (Athens 1993; Dallas 1986; Melbourne 1981; Pakistan 1989) and for six studies the methodological quality was unclear (Copenhagen 1985; Denver 1976; Denver 1979; New Delhi 2006; Seattle 1987; Sheffield 1978). It appears that in a high-quality trial, there was less neonatal acidosis compared with low-quality trials (P = 0.04; I
Continuous CTG versus intermittent CTG (Comparison 9)
Lund 1994 involved 4044 high-risk pregnant women and found no significant differences in any of the eight outcomes specified in this meta-analysis.
Summary of main results
The main reason for the introduction of continuous intrapartum cardiotocography (CTG) monitoring in clinical practice was a belief that it would reduce rare but devastating outcomes - perinatal death and neonatal hypoxic brain injury in otherwise healthy babies. However, this review found no statistically significant difference in perinatal deaths between pregnancies monitored during labour with continuous CTG compared to those monitored those intermittent auscultation. The overall quality of evidence that underpins this conclusion has been judged as 'moderate' (Summary of findings table 1). It does, however, seem unrealistic to expect that any intrapartum intervention in modern maternity care will result in a statistically significant improvement in perinatal deaths. In order for a trial to test a realistic hypothesis that continuous CTG can prevent one death in one thousand births (0.1%), more than 50,000 women would have to be randomised. It is, therefore, more logical to concentrate on short- and long-term childhood morbidity. Unfortunately, very few clinically relevant neonatal outcomes have been reported consistently in all trials.
For decades, low Apgar scores have been used as a surrogate measure for birth asphyxia and subsequent adverse neurodevelopmental outcomes. Recent evidence has confirmed a strong association between low Apgar score and cerebral palsy in both low and normal birthweight infants (Lie 2010). This review found no evidence that use of continuous intrapartum CTG monitoring has an impact on Apgar score. However, there were very few babies with clinically significant low Apgar scores in studies that assessed this outcome. Therefore, potentially important differences between the two groups cannot be ruled out.
Hypoxic ischaemic encephalopathy, a more robust measure of hypoxic brain injury, was only reported in one study (Athens 1993). In the absence of any meaningful long-term follow-up data, the impact of continuous CTG monitoring on a neonate can only be evaluated based on the data from two clinically important outcomes, i.e. neonatal seizures and cerebral palsy.
For both neonatal seizures and cerebral palsy, the majority of data are provided by Dublin 1985. At first glance, the data appear contradictory. There is a significant reduction in neonatal seizures in the continuous CTG group, but no impact on cerebral palsy. If anything, the rates of cerebral palsy appear to be higher in the continuous CTG group, although the pooled result did not reach statistical significance. This apparent increase in cerebral palsy in children monitored by CTG comes from Seattle 1987. However, the results from this study, the only study of CTG monitoring during preterm labour, are not statistically significant using 99% confidence intervals. In addition, this study excluded infants with a birthweight of more than 1750 grams (34% of randomised cohort) which may be a source of bias. Given that all other outcomes in this trial, including caesarean section rates, neonatal seizures and deaths were almost identical this may have been a chance finding and should be interpreted with caution.
It is now generally accepted that cerebral palsy is more often caused by antepartum, rather than intrapartum, events (Palmer 1995). It may, therefore, be unrealistic to expect that intrapartum interventions will have the capacity to achieve a significant reduction in cerebral palsy. There are, clearly, some cases of cerebral palsy that are a direct consequence of intrapartum hypoxic injury. These cases are very rare, and even systematic reviews of randomised trials are unlikely to have sufficient power to test intrapartum CTG as a method to reduce cerebral palsy caused by acute and avoidable intrapartum events.
The reduction in seizures associated with continuous CTG monitoring is important, but has to be interpreted cautiously in the absence of good quality long-term follow-up data. It has been suggested that seizures may be a "sentinel event" of a peripartum adversity that does not necessarily always manifest itself as hypoxic encephalopathy (Dennis 1978; Derham 1985, Keegan 1985; Lien 1995; Spellacy 1985). When asphyxia, infection, brain malformations and metabolic causes are excluded, some neonatal seizures are associated with cerebral infarction or neonatal stroke (Estan 1997; Lien 1995). Although the underlying causes are not well understood, neonatal seizures may have long-term consequences other than cerebral palsy. One longitudinal study found that some babies who had neonatal seizures were classified as normal at five years and had normal overall intelligence in adolescence as assessed by IQ tests, but had some abnormal results on detailed neuropsychological testing (Temple 1995). Clearly, there is a need for comprehensive long-term follow-up of the randomised cohorts that is not limited to extreme adverse outcomes such as cerebral palsy, but also includes more subtle neuropsychological assessment.
The results of this review demonstrate that continuous CTG monitoring leads to an increase in caesarean sections. Such an effect of continuous CTG is clinically plausible as CTG monitoring leads to more interventions (e.g. fetal blood sampling, amniotomy) and more diagnoses of presumed fetal compromise for which emergency caesarean section is seen as the only safe management option. However, the overall quality of evidence for this outcome was judged as 'low' (Summary of findings table 1) and, therefore, the observed increase has to be interpreted cautiously. It is noteworthy that size and direction of the effect on caesarean section was consistent for pre-specified subgroups, including high quality trials and trials where clinicians had access to intrapartum fetal blood sampling. Subgroup interaction test was only significant (I
There was some evidence that labour was more painful in the continuous CTG group, but the statistically significant increase in the 'need for any analgesia' included general anaesthesia. It is, therefore, likely that this difference was caused by an increase in the number of caesarean sections, rather than necessarily more painful labour. Women do report more pain when lying on their backs during labour. At the times when the studies in this review were undertaken (between 1976 and 1994), women in the intermittent auscultation group may well also have been on their backs and not using mobility and positions to help them with their labours. There were no data from the trials included in the review to allow any analysis of this potential confounder.
We have prespecified several subgroups that could have been expected to influence the direction and size of the differences compared with results when all trials are considered together. We were conscious that any differences between subgroups and overall results would have to be interpreted with extreme caution (Rothwell 2005). With this proviso, we found no subgroup differences of clinical importance, but the number of trials and women in subgroups was relatively small.
Overall completeness and applicability of evidence
Clearly, the lack of long-term follow-up data and inadequate reporting of the data according to the clinically important subgroups is regrettable and limits the applicability of the evidence.
Quality of the evidence
The overall quality of the evidence can be best described as low to moderate and has been summarised in the Summary of findings table 1.
Potential biases in the review process
Our selection of outcomes in general and primary outcomes in particular might have been influenced by our knowledge of the published literature and the first Cochrane review on this topic (Thacker 2001).
Agreements and disagreements with other studies or reviews
Some large cohort studies suggest much more profound benefit on neonatal morbidity and mortality (Chen 2011). Some observational data also suggest the benefit of fetal blood sampling during labour in cases of suboptimal CTG, (Stein 2006) but we found no evidence that the increase in caesarean section rate was greater if fetal blood sampling was unavailable; nor did access to fetal blood sampling influence the difference in neonatal seizures or any other prespecified outcome.
Implications for practice
Translating the evidence from this review into clinical practice poses significant challenges. One would hope that the quality of CTG equipment, interpretation and training have improved over the years making the external validity of much of the data included in this review questionable. In most studies included in this review, intermittent auscultation was carried out according to the strict protocols in a hospital setting with quick recourse to continuous monitoring and intervention if required. In some trials, most notably the Dublin trial (Dublin 1985), intact fetal membranes were ruptured at the earliest opportunity to confirm absence of meconium and women were provided with one-to-one care from a midwife. This monitoring package differs significantly from practices in some modern birth settings (including, for example, stand alone midwifery units) where artificial rupture of membranes is avoided as long as possible and where mobilisation and normality are promoted. In addition, one-to-one care by a midwife, or a nurse-midwife, seems hard to implement in many healthcare settings and is likely to be an important contributory factor for effectiveness (or lack of it) of both types of fetal heart rate monitoring.
With this proviso, women should be informed that continuous CTG during labour is associated with a reduction in the incidence of neonatal seizures, has no obvious impact on cerebral palsy or perinatal mortality but is associated with an increase in the incidence of caesarean section and instrumental vaginal births. The adverse affects of operative births are well described albeit that longer term morbidity data are less available than shorter term morbidity data. The possible long-term effects of preventable neonatal seizures remain unknown. Women also need to be informed of the loss of mobility associated with the use of continuous CTG in labour.
Women, practitioners and policy makers should consider carefully the absence of evidence that continuous CTG monitoring has a different impact on caesareans section and neonatal seizures in low- and high-risk populations and that there is an absence of evidence from trials included in this review of a beneficial effect for fetal blood sampling.
The risk benefit debate will continue to focus on caesarean section and neonatal seizures. Given the perceived conflict between the risk for the mother (increased caesarean section and instrumental vaginal delivery rate) and benefit for the baby (decreased incidence of neonatal seizures), it is difficult to make quality judgments as to which effect is more important. The issue of effectiveness is particularly important. CTG advocates will continue to argue that lack of clear long-term benefit for the child is not proof that intermittent auscultation is safe. However, it would seem reasonable to base clinical decisions on the evidence we currently have rather than on unknown risks of unknown quantity. Obviously, the risk-benefit assessment will vary between individuals, policy makers and healthcare settings. The real challenge is how best to convey this uncertainty to women and help them to make an informed choice without compromising the normality of labour.
Implications for research
The question remains as to whether future randomised trials should measure efficacy (the intrinsic value of continuous CTG in trying to prevent adverse neonatal outcomes under optimal clinical conditions) or effectiveness (the effect of this technique in routine clinical practice).
Along with the need for further investigations into the long-term effects of operative births for women and babies, much remains to be learned about the causation and possible links between antenatal or intrapartum events, neonatal seizures and long-term neurodevelopmental outcome, bearing in mind the changes in clinical practice over the intervening years (one-to-one-support during labour, caesarean section rates). The large number of babies randomised in this review will now have reached adulthood, and could potentially provide us with a unique opportunity to clarify if a reduction in neonatal seizures is something inconsequential that should not greatly influence women's and clinicians' choices, or if seizure reduction leads to long-term benefits for babies. Defining meaningful neurological and behavioural outcomes that could be measured in large cohorts of young adults poses huge challenges.
Data should also be collected from this cohort of women and babies, whilst the medical records still exist, to describe, where possible, the women's mobility and positions during labour and birth, to clarify if these might impact on outcomes. Research should also address the possible contribution of the supine position to adverse outcomes for the baby, and address the question of whether the use of mobility and positions can reduce the already low incidence of neonatal seizures and improve psychological outcomes for women.
The review authors would like to acknowledge the support of Mrs Sonja Henderson, Review Group Co-ordinator, and Ms Lynn Hampson, Review Group Trials Search Co-ordinator, in the preparation of this review. We wish to acknowledge the contribution from Mark Turner, Neonatologist from The University of Liverpool, for contributing to the discussion on the importance of adverse neonatal outcomes.
We acknowledge S Thacker, D Stroup and M Chang, who prepared the first version of this Cochrane review under the tItle 'Continuous electronic heart rate monitoring for fetal assessment during labor' (Thacker 2001).
The National Institute for Health Research (NIHR) is the largest single funder of the Cochrane Pregnancy and Childbirth Group. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR, NHS or the Department of Health.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Index terms
Appendix 1. Search strategies
In the 2006 update of this review, we searched CENTRAL (The Cochrane Library 2005, Issue 4), MEDLINE (1966 to December 2005), CINAHL (1982 to December 2005) and EMBASE (1974 to December 2005) using the search strategies below:
1. Labor Obstetric/
2. Delivery Obstetric/
2. Fetal Monitoring/
3. intrapartum near monitor*
4. fetal near surveillance
5. 1 or 2
6. 3 or 4 or 5
7. 6 and 7
1. exp Labor, Obstetric/ or exp Delivery, Obstetric/
2. exp Fetal Monitoring/
3. (intrapartum adj2 monitor$).ti,ab.
4. (fetal adj surveillance).ti,ab.
5. randomized controlled trial.pt.
6. exp Controlled Clinical Trials/
7. controlled clinical trial.pt.
8. 2 or 3 or 4
9. 5 or 6 or 7
10. 1 and 8 and 9
1. exp Clinical Trials/
2. clinical trial.pt.
3. (clinic$ adj trial$1).tw.
4. ((singl$ or doubl$ or trebl$ or tripl$) adj (blind$3 or mask$3)).tw.
5. randomi?ed control$ trial$.tw.
6. exp Random Assignment/
7. random$ allocat$.tw.
9. Quantitative studies/
10. allocat$ random$.tw.
12. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11
13. exp Fetal Monitoring/
14. (fetal adj2 monitor$).tw.
15. (intrapartum adj2 monitor$).tw.
16. (labor or labour).tw.
17. exp Childbirth/
18. 13 or 14 or 15
19. 16 or 17
20. 12 and 18 and 19
2. double blind procedure/
3. crossover procedure/
4. intermethod comparison/
5. single blind procedure/
6. clinical study/
7. controlled study/
8. randomized controlled trial/
9. (clin$ adj2 trial$).tw.
10. ((singl$ or doubl$ or trebl$ or tripl$) adj2 (blind$ or mask$)).tw.
11. exp clinical trial/
15. labour$ or labor or laboring.af.
18. fetal with monitor$
19. fetal adj surveillance
20. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14
21. 16 or 17 or 18 or 19
22. 20 and 21 and 15
We also searched for grey literature by searching Dissertation Abstracts (1980 to December 2005) and National Research Register (December 2005) databases, using monitoring terms identified above, adapted for each database. We chose not to repeat these additional searches for the update as they yielded no additional studies in the original search.
Appendix 2. Methods used in previous versions
We developed the methods of the review in light of the advice contained in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005).
We considered all identified randomised and quasi-randomised controlled trials involving a comparison of continuous CTG, with and without fetal blood sampling, with (a) no fetal monitoring, (b) intermittent auscultation of the fetal heart rate with a Pinard stethoscope or hand-held Doppler ultrasound device or (c) intermittent CTG.
One review author (Declan Devane (DD)) ran the additional search strategies. Each potentially eligible trial identified by the search strategy was obtained as a full-text article and independently assessed for inclusion by Zarko Alfirevic (ZA) and Gill Gyte (GG). There were no disagreements regarding eligibility for inclusion that needed to be resolved by discussion with DD. We did not encounter problems with language or missing information requiring classification as 'Study awaiting assessment' (RevMan 2003).
Quality assessment of included studies
Two review authors (GG and DD) independently assessed the quality of all included trials, namely selection and attrition bias. With regard to performance bias, due to the differences in the modus operandi of the continuous CTG and intermittent auscultation, it is unlikely that clinicians or women will have been blinded to either intervention. Therefore, lack of blinding was not considered to undermine the validity of studies.
Studies were allocated a grade on the basis of allocation concealment as per criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005) i.e. (A) adequate, (B) unclear, (C) inadequate or (D) allocation concealment was not used. Approaches to allocation concealment considered to be clearly inadequate include: alternation, the use of case record numbers, dates of birth or day of the week, and any procedure that is entirely transparent before allocation, such as an open list of random numbers.
Due to inadequacies in reporting how losses of participants (e.g. withdrawals, dropouts, protocol deviations) were handled, the review authors were cautious about implicit accounts of follow-up. Given that study reports on attrition after allocation have not been found to be consistently related to bias, studies were not excluded on the basis of attrition. Studies were, however, graded for completeness of follow-up using the following criteria. For completeness of follow-up:
(A) less than 3% of participants excluded;
(B) 3% to 9.9% of participants excluded;
(C) 10% to 19.9% of participants excluded;
(D) more than 20% of participants excluded.
Two review authors (ZA and DD) independently extracted the data using predesigned data extraction forms, the fields of which had been agreed by all review authors. Study eligibility was verified again at the time of data abstraction or collection.
Additional information was extracted from the included trials by one review author (GG) and recorded in an additional table ( Table 2). The data include: (1) one carer to one woman support during labour; (2) labour induction; (3) the use of artificial rupture of membranes (ARM) in labour; (4) the use of oxytocin for augmentation of labour; (5) women's mobility during labour; (6) women's positions for giving birth; (7) women's views of labour and monitoring; (8) social and environmental context of trials; (9) experience of staff in CTG interpretation. These were considered to be factors that might impact on the comparison of outcomes. For example, it is unclear what impact the supine position, ARM and oxytocin use might have on the fetal heart rate patterns, and whether mobility in labour might reduce the use of such interventions.
We performed statistical analyses with the Review Manager Software (RevMan 2003). Dichotomous (or binary) outcomes were reported using the 'relative risk' summary statistic and their 95% confidence intervals. Continuous data were reported using the weighted mean differences and their 95% confidence intervals.
Denver 1979 was a three-arm trial with two experimental groups (CTG with and without access to fetal scalp sampling) and one control group (intermittent auscultation). Following statistical advice, we arbitrarily split the data for the controls into two equal groups and assigned them to each experimental arm. This approach ensured that there was no double-counting for controls when overall relative risks were calculated. An arbitrary decision was made when the number needing to be split was not an even number.
We used a fixed-effect model of meta-analysis for summarising the results of studies in the absence of substantial heterogeneity between trials. Where heterogeneity between trials was substantial a random-effects model was used. Measurements of heterogeneity were performed using the I-squared statistic, which is less affected by the number of trials in the analysis than the Chi-squared test. I-square of 30% to 50% suggests mild heterogeneity and more than 50% indicates substantial heterogeneity.
We planned subgroup analyses on the following a priori determined subgroups:
(a) low risk (absence of identifiable risk factors associated with increased in perinatal mortality and morbidity as defined by trialists);
(b) high risk for perinatal mortality and morbidity (as defined by trialists);
(c) spontaneous onset of labour;
(d) induction of labour;
(e) preterm (less than 37 + 0 weeks);
(f) term (greater than 37 + 0 weeks);
(g) singleton pregnancy;
(h) twin pregnancy;
(i) without fetal blood sampling (FBS) during labour;
(j) with FBS during labour;
We performed sensitivity analysis based on quality comparing high-quality trials with trials of lower quality. Given that study reports on attrition after allocation have not been found to be consistently related to bias, 'high quality' was, for the purposes of this sensitivity analysis, defined as a trial having allocation concealment classified as 'A' (adequate).
Ingemarsson, 30 March 2008
In this review you comment on the significant reduction in neonatal seizures associated with continuous cardiotocography rather than intermittent auscultation, but then put this in opposition to the increase in caesarean section. Yet, more caesarean sections are performed without clinical indication, on maternal `request' than are performed for threatening fetal hypoxia. Moreover, you stress that continuous cardiotocography is not associated with any beneficial effect on the risk of cerebral palsy, because 80%-85% of cases have an antenatal origin and therefore intrapartum CTG can not be expected to have a great impact on the overall figure.
A recent Swedish study (Lindström 2006) reported outcome at 15-19 years of age after moderate hypoxic-ischemic encephalopathy (Sarnat II with neonatal seizures in most cases). Of 43 children with moderate hypoxic-ischemic encephalopathy, 15 had cerebral palsy. Of the 28 children without encephalopathy, 20 had cognitive problems. Only 8 of the 43 children had no problem later in life. So, a halving in neonatal seizures with continuous cardiotocography seems to me, as an old obstetrician, to be a very good outcome.
(Summary of feedback from Ingemar Ingemarsson, March 2008)
Thank you for your comments. In our review, we feel we have clearly articulated the perceived conflict between our findings of increased caesarean section and instrumental vaginal birth and decreased incidence of neonatal seizures associated with continuous CTG when compared with intermittent auscultation.
We are unaware of any high quality evidence that demonstrates a higher rate of caesarean sections due to maternal ‘request’ than due to hypoxia. Caesarean sections for maternal ‘request’ is a complex issue and there are those who have argued that it is not a significant influencing factor on caesarean rates (Gamble 2007) Even if such evidence existed, we believe that this is addressing a different question from that in our review.
The focus of the quoted study by Lindström et al (Lindström 2006) is on neonatal encephalopathy. In our review, we highlighted that much remains to be learned about the causation and possible links between antenatal or intrapartum events, neonatal seizures and long-term neurodevelopmental outcome. Until and unless a causative link is made between neonatal seizures, we believe it reasonable to base clinical decisions on the evidence we currently have.
Panteghini, 30 September 2013
I have two comments about this review:
1) In the continuously monitored group the relative risk of perinatal mortality is lower rather than in the intermittently monitored group (RR 0.86). This result may be important for women when they choose which method of fetal monitoring to adopt during labour. Is it not more useful to present the absolute and relative risk, so the woman, her midwife and doctor can decide if these are significant to them or not? To consider a result significant only if it is statistically significant (and only if statistically significant at a given level of significance, such as 5%) is an arbitrary decision that needs to be shared with the woman and her clinical team.
2) An interesting question raised by this review is which method of intermittent auscultation is best. The review lumps together different types of intermittent auscultation; for example, auscultation during and after a contraction, and auscultation only after a contraction.
The review assesses the relationship between pH at birth and the method of foetal heart monitoring rate (intermittent or continuous) in two studies (Athens 1993, Dublin 1985), and does not find any difference between the two methods as regards neonatal pH at birth. It is interesting to note that in the Dublin trial, which used intermittent auscultation only after a contraction, the pH at birth was worse for woman allocated intermittent auscultation rather than continuous monitoring (RR 0.45, 95% CI 0.16 - 1.29). In contrast, in the Athens trial, which used intermittent auscultation during and after the contraction, pH at birth was better for woman allocated intermittent auscultation (RR 1.58, 95% CI 0.89 - 2.81).
The importance of decelerations during the contraction and their impact on foetal wellbeing is now well known. Therefore the National Institute for Clinical Excellence (NICE) (1) considers monitoring to be reassuring only if there are no decelerations. Some guidelines advise monitoring the foetal heart after a contraction (2), others during and after (3), and others again do not specify the timing of auscultation in relation to contraction (4). The review is appropriate in not drawing any conclusions about what is the best method of intermittent monitoring. We think that guidelines should state both that the mode of intermittent monitoring and the choice of one method rather than another is a grade C recommendation (personal opinion) (5) as, in the light of this review, we do not know which method of intermittent monitoring is best (although we could suppose that intermittent auscultation during and after a contraction may be better than auscultation only after a contraction for preventing low pH at birth).
(1) NICE. Intrapartum care, 2008; p219-220 Tables 13.1, 13.2.
(2) Royal College of Midwives. Evidence based guidelines for midwifery-led care in labour,2012.
(3) American College of Nurse and Midwives. Intermittent Auscultation for Intrapartum Fetal Heart Rate Surveillance. Journal of Midwifery and Women's Health, 2010; 55: 397-403.
(4) Association of Women's Health Obstetric and Neonatal Nurses. Fetal Heart Monitoring, 2008
(5) Danti L, Di Tommaso MR, Maffetti G, Carfagna M. Cardiotocografia. Milano 2010, Piccin editore.
Comment submitted by Marco Panteghini, September 2013
Last assessed as up-to-date: 31 January 2013.
Protocol first published: Issue 3, 2006
Review first published: Issue 3, 2006
Contributions of authors
Zarko Alfirevic (ZA) drafted the protocol. Declan Devane (DD) and Gill Gyte (GG) commented on all its sections.
ZA and GG assessed studies in respect of inclusion and exclusion criteria.
DD ran additional searches. ZA and DD extracted the data independently and double entered them into Review Manager. GG extracted additional descriptive information from included studies. All three authors wrote and agreed the final version of the review.
Declarations of interest
Medical Subject Headings (MeSH)
*Labor, Obstetric; Cardiotocography [*methods]; Cesarean Section [statistics & numerical data]; Heart Auscultation [*methods]; Heart Rate, Fetal [physiology]; Infant Mortality; Infant, Newborn; Randomized Controlled Trials as Topic; Seizures [prevention & control]
MeSH check words
Female; Humans; Pregnancy
* Indicates the major publication for the study