Antenatal foetal heart rate assessment was introduced into clinical medicine before clear evidence of any benefits had been reported. Ad hoc definitions were used to define normal and abnormal recordings resulting in a high false-positive rate for foetal compromise.
The understanding of the foetal states resulted in an improved physiologically based assessment of the antenatal tracings and allowed their classification as (i) reactive – 2 accelerations in 10 min within a recording period of 120 min, (ii) unreactive – no accelerations seen in 120 min of tracing and (iii) decelerative – the presence of repetitive decelerations on an otherwise unreactive trace. This classification reduces the high rate of false-positive traces associated with recording times of less than 40 min. Traces performed on pregnancies before 32 weeks predict clinical outcome, but need to be interpreted in light of the fact the many foetuses will not show a mature reactive pattern.
Highlights
- •
Despite the antenatal CTG being used daily, no trials have identified clinical benefits.
- •
Ad hoc definitions have resulted in the unreactive trace having a high false-positive rate.
- •
Understanding foetal state changes is essential for antenatal CTG interpretation.
- •
The CTG trace should not be classified as unreactive until 120 mins have elapsed.
- •
The length of the admission CTG should be redefined to avoid a high false-positive rate.
Introduction
The monitoring of foetal welfare before labour using the electronic recording of heart rate dates back the early 1960’s. This was only a few years after Hon recorded the heart rate in labour using his foetal scalp electrode . The original antenatal devices used microphones placed on the maternal abdomen, which were heavy and often gave an unreliable signal . The recordings were however sufficient to allow the determination of baseline variability and documentation of features that appeared to be associated with foetal compromise, namely decreased variability and recurrent decelerations.
On the basis of the diagnostic promise, a number of manufacturers developed recording systems based on Doppler ultrasound technology . Originally, the machines produced a relatively unprocessed signal which, although an improvement on the microphone, still over-exaggerated the variability of the baseline over that recorded via a scalp electrode. The advent of autocorrelation, a technique for noise reduction, and signal averaging in later models manufactured in the 1970s allowed for a reliable signal to be recorded, so that the traces produced via Doppler and scalp electrode varied little in baseline variability.
Early studies assessed the physiology and pathophysiology of heart rate parameters. Normal heart rate was variously described as being between 110–120 beats per minute (bpm) and 150–160 bpm with the baseline declining with increasing gestation between 30 and 40 weeks . At any given time, the rate was determined to be a reflection of the intrinsic rate of the heart and autonomic input from the sympathetic and parasympathetic nervous system. Other external influences that were noted to affect the heart rate were cord occlusion, foetal urination, eye rubbing, grunting, and hiccoughing as well as significant acute foetal hypoxia, which caused a decrease in rate; and movement, mild hypoxic stress, maternal temperature and tachycardia, which caused an increase in rate .
Despite the lack of a full understanding of the foetal heart rate patterns, or trials of effectiveness of antenatal recording, the advent of the commercial monitors resulted in their widespread use in low- and high-risk pregnancies. Varying definitions of what constituted a normal or abnormal trace were published along with disagreements on the ideal length of recording .
The degree of disparity of definitions used in antenatal CTG testing was highlighted by Devoe et al in 1985 . Most reports used a test length of 20 min; however, some used 30 or 40 min. The number of accelerations for reactive test was variously determined to be 2 in 10 minutes or 2, 3, or 5 in 20 min. The number of decelerations that were required to define a pathological trace ranged from 1–3/h. Finally, although normal baseline variability was universally defined as being more than 5, a rate of less than that was not universally accepted as abnormal especially if it was the only abnormal feature of the trace . Schifrin et al noted that the nonreactive CTG, that is the trace with no defined accelerations and a diminished baseline variability during the recording period, had false-positive rates of 50–80% for foetal compromise, and, on the basis of fact that up to 33% of recordings were unreactive , their conclusions about the antenatal CTG were that ‘the applicability of the test is questioned by many, in part due to its inability to precisely define adverse outcome’.
To determine foetal reserve in the ‘unreactive’ trace more accurately, American and European units investigated the use of the contraction stress test (CST), whereby oxytocin was administered by IV infusion or by intranasal spray during the test to invoke uterine contractions . Unfortunately, decelerations seen after the oxytocin-induced contractions had a 33% false-positive rate for foetal compromise , leading to the technique being largely abandoned outside North America.
Improving the definitions and predictive value of the antenatal CTG
Early work on both the contraction stress test (CST) and the non-stress test (NST) determined that the presence of accelerations as defined as an increase in the baseline rate of more than 15 bpm for more than 15 s were an assurance of foetal wellbeing whether accompanied by a perceived foetal movement . Studies that attempted to define whether the movements required for a trace to be classified as reactive was 2, 3, 4, or 5, in 10–20 min determined that 2 in 10 minutes had as good a predictive value as any other definition. The early data reported a rate of foetal mortality of 3/1000 within 24 h of a reactive test .
The major issue with the antenatal test, as outlined above, has been to obtain a robust definition of an abnormal trace, and in particular, the unreactive test with its apparently high false-positive rate. From the early to mid-1980s, it became apparent that the foetal heart in the term foetus did not accelerate continuously, but rather it went through periods of acceleration that occurred between periods where the CTG had a normal baseline, no accelerations, and where the variability of less than 5 bpm .
Physiologically, the reason for the periods of nonreactivity is that the normal foetus near term spends periods of time in quiet sleep where baseline rate is normal, and baseline variability as recorded is often less than 5, and at times less than 2, with no accelerations to a foetal movement. Nijhuis therefore observed that CTG interpretation, especially in the antenatal period, could not be divorced from an understanding of the foetal behavioural states .
The quiet sleep (1F) state was denoted with the foetus in non-REM sleep where the foetal heart rate had diminished variability on a normal baseline and episodic foetal movements of the foetal limbs were not always accompanied by accelerations. The active sleep (2F) state is associated with the CTG that many recognise as reactive, that is, a tracing showing a normal heart rate with variability on the baseline of 5–25 and accelerations that accompany foetal body movements ( Figure 1 ). The active awake (4F) state where the foetus has its eyes open and is moving continuously is associated with a varying heart rate where the baseline is hard to define due to a multitude of accelerations amongst foetal body movements that are sometimes described as excessive ( Figure 2 ). A 3F quiet awake state involving continuous eye movements associated with increased variability but no accelerations were also described, but it is not commonly recognised .
The issue therefore has long been how to detect the foetus in a normal quiet sleep pattern from the one with a nonreactive pattern due to a hypoxia. Brown and Patrick assessed the issue and determined that although the majority of foetuses will naturally transition from quiet sleep to active sleep within 40 min, a proportion could take 120 min. These data go a long way to explain the false-positive data surrounding the unreactive trace that has a recording period of less than 40 min.
It is an unfortunate fact that the early trials undertaken to determine the benefits of the antenatal CTG in high-risk patients either did not indicate the period of recording needed to define an unreactive trace , or limited tracing to 30 min , or in one case recorded ‘up to 40 minutes’ but only performed the test once a week . Not surprisingly, the conclusion of the trials was that concealing or revealing the CTG tracings performed under these conditions to the treating clinician did not affect the outcome of a pregnancy. Similarly, the admission CTG in early labour that uses a trace of 20 min to assess foetal wellbeing does not consider the number of foetuses that will be in quiet sleep for that period of time. Clearly, any publication on the antenatal CTG that does not allow for the normal period of quiet sleep in the normal foetus and labels a trace as unreactive after 20, 30, or 40 min is flawed, and cannot be used to as evidence of the benefit or otherwise of the antenatal CTG in clinical practice. It is unfortunate that no published trial of the use of the antenatal CTG has considered the duration of the 1F state in the normal foetus.
Alternate approaches to the assessment of the unreactive trace have been to use foetal vibroacoustic stimulation , or computerisation of the CTG assessing anomalies in the short-term foetal heart rate variation (STV) that is altered by foetal hypoxia but not quiet sleep .
Unfortunately, the use of vibroacoustic stimulation can cause a foetal reaction in both the quiet sleep and the hypoxic state , and refinements like the assessment of habituation to the stimulus have not translated easily to the clinical arena.
Dawes et al led the charge into the computerisation of the antenatal CTG to assess for foetal baseline heart rate variation, accelerations and decelerations as well as uterine contractions. The commercialised System 8000 also determined signal quality to ensure adequate recording. Although promising reports exist on the use of system in daily recording of the CTG , reports exist that show little difference between visual CTG interpretation and the interpretation provided by the System 8000 . Two papers were recently included in a meta-analysis of computerised CTG versus traditional antenatal CTG . Neither showed benefit in the outcomes of perinatal mortality, delivery by Caesarean section, any potentially preventable death, or Apgar scores of less than 7 at 5 minutes. The meta-analysis however suggested only an improvement in perinatal mortality. Unfortunately, one of the studies compared computerised traces of up to 60 min with ‘traditional’ traces lasting only 10 min, and, in the second paper, 4 babies died from congenital anomalies in the traditional CTG group compared to 1 in the computerised group. These issues make the meta-analysis unhelpful in determining any role for the use of the computerised CTG.
The outcomes associated with a trace defined as pathological are more defined. The tracing with repetitive decelerations either occurring spontaneously or associated with Braxton Hicks contractions with diminished variability on the baseline ( Figure 3, 4 ) is universally accepted as being associated with a high degree of foetal compromise in terms of hypoxia in the absence of a maternal metabolic or fetal congenital disorder. Similarly, the baseline bradycardia with absent variability has a universally poor outcome if the foetus is not delivered .
Baseline tachycardia in the antenatal trace is more commonly associated with maternal pyrexia or foetal infection than hypoxia ; thus, an uncomplicated increase in baseline rate should be investigated for a cause that may not be foetal compromise.
What about the isolated deceleration that occurs amongst the otherwise reactive trace ( Figure 5 )? Given the known time limited causes of foetal heart rate slowing like cord grasping and foetal urination, these decelerations must be considered ‘physiological’ and not pathological .
In light of the above data, how best to proceed? Clearly, despite the trial data stating that the antenatal CTG cannot be objectively shown to improvement pregnancy outcome, the test is still a daily feature in every obstetric service and must be perceived to be of some value especially in units where easy recourse to biophysical assessment of the foetus with ultrasound is not readily available .
Reported experience has shown that the reactive trace, where there are at least 2 accelerations within a 10-min recording period, is associated with a foetal mortality rate of less than 0.5% . Similarly, the pathological trace that shows repetitive decelerations or a baseline bradycardia has a very low false positive rate and clearly requires immediate assessment with ultrasound scanning or delivery of the foetus . The issues arises with the unreactive trace. In light of the foetal state data, it is clear that a normal foetus can remain in the 1F state for 90–120 min. Should 2 accelerations occur in the 10 min after 100 min of a trace which to that point was simply unreactive, the occurrence of the accelerations denotes the foetus as being normally oxygenated and is not acidaemic. Although data are reported showing that only acting on traces that are unreactive for more than 120 min reduces the false-positive rate to 1.5%, the use of this definition requires further investigation .
The advent of advanced neonatology has changed the intact survival rates of the neonates that are delivered before 32–34 weeks when foetal states are established. The ability of obstetric units to monitor premature pregnancies has provided a challenge. Despite a lack of evidence on benefit the CTG, it is often used in these situations to monitor an at-risk foetus exposed to various conditions including antepartum haemorrhage, preeclampsia, preterm premature rupture of the membranes and unexplained prematurity. Without clear data on what a normal CTG pattern is in extreme prematurity, there is a risk of unwarranted delivery and iatrogenic prematurity. No studies exist on the use of the CTG before labour in foetuses before 33 weeks. The only data come from reports based on observations performed during labour . Observational data however shows that between 28 and 34 weeks, the normal foetus develops a mature reactive pattern ( Figure 6 ) that resembles that of the foetus at term. Prior to that time, the tracings show a relatively low variability on a relatively high baseline heart rate ( Figure 7 ); however, these changes have the same predictive value as the mature pattern. Isolated accelerations are uncommon, rather, the heart rate increases for periods of 5–10 min during periods of foetal activity . The baseline may show very short episodes of slowing, which are often mistaken for decelerations. They are common and are variations of the baseline rate. Repetitive prolonged decelerations are however an indication to investigate foetal oxygenation with Doppler assessment.
