An evaluation of electronic fetal monitoring with clinical validation of ST waveform analysis during labour
Dissatisfaction with the electronic recording of fetal heart rate and uterine contractions (the cardiotocogram or CTG) has resulted in a search for new techniques of monitoring the fetus during labour. It is important that each method has a sound physiological and pathophysiological basis, that a model for the interpretation of changes is elucidated and that each method is thoroughly evaluated before introduction into clinical practice. Analysis of the ST waveform of the fetal electrocardiogram (FECG) is the most advanced of the new techniques under investigation. Experimental studies have shown that elevation of the ST waveform occurs with a switch to myocardial anaerobic metabolism and a negative waveform occurs during direct myocardial ischaemia. Human observational studies have suggested that a combination of ST waveform and CTG analysis may improve the specificity of intrapartum monitoring and reduce unnecessary intervention. A high quality FECG signal is necessary for waveform analysis. The FECG can be recorded from a scalp electrode (FSE) during labour. The suitability of 5 commonly available FSEs for ECG waveform analysis was compared. Single spiral FSEs had the most favourable physical and electrical properties and produced the best quality signals in a randomised clinical trial of 50 fetuses in labour. Intervention rates and neonatal outcome in labours monitored with CTG alone were compared with those monitored with the combination of ST waveform analysis plus CTG (ST+CTG) in a randomised clinical trial of 2434 high risk labours in a large district general hospital over an 18 month period. There was a 46% reduction in operative intervention for fetal distress in the ST+CTG group (p<0.001, OR 1.96 [1.42-2.71]). There was a trend to less neonatal metabolic acidemia (p = 0.09, OR 2.63 [0.93-7.39]) and fewer low five minute Apgar scores (p = 0.12, OR 1.62 [0.92-2.85]) in the ST+CTG arm. All recordings were reviewed retrospectively, blind to outcome and the CTG classified as normal, intermediate or abnormal according to the trial protocol. There was no significant difference in the proportion of recordings in each category between the trial arms. Operative intervention in the ST+CTG arm was significantly reduced in recordings classified as normal and intermediate by the review (12/1043 ST+CTG arm versus 48/1066 CTG arm, p <0.001). Three patterns of ST+CTG change were identified. 1. Normal CTG, persistent stable ST waveform elevation. These fetuses had good outcome and a significantly higher mean pH (7.29) and lower base deficit (1.1 rnmol/1) at delivery. The raised ST waveform may reflect sympathoadrenal stimulation from the general arousal of labour or a response to mild but compensated hypoxaemia and is in keeping with experimental data. 2. CTG abnormal, progressive elevation in ST waveform. All cases occurred towards the end of second stage. These fetuses had a significantly lower mean pH (7.05) and higher base deficit (7.6 mmol/1) than all other groups. This combination identified fetuses who were developing a metabolic acidosis as a result of significant hypoxia. 3. Abnormal CTG and a negative ST waveform. All cases with persistently negative waveforms were depressed at birth, required resuscitation and had low arterial pHs (where available). This high risk group probably had depleted myocardial glycogen reserves and suffered direct myocardial hypoxia, as seen in animal studies. These findings indicate that ST waveform analysis can discriminate CTG change during labour, the combination can result in a reduction in unnecessary intervention and has the potential to more accurately identify fetuses at risk of neonatal morbidity. The term 'monitoring' implies a degree of automatic surveillance but this is not the case as CTG and ST+CTG records are subjectively interpreted, frequently by junior, inexperienced staff. The retrospective review of cases in the trial revealed significant errors in the use of fetal blood sampling and the interpretation of both CTG and ST+CTG recordings during the study. The feasibility of representing expert clinical knowledge in a decision support tool to provide consistent, accurate interpretation of the CTG was demonstrated in two clinical studies. The full potential of ST+CTG analysis may only be achieved with some degree of automatic data processing and interpretation. The randomised trial also demonstrated the lack of appropriate measures of neonatal outcome with which to judge the effectiveness of fetal monitoring. Analysis of cord artery and vein blood gas status at delivery can provide useful information about fetal oxygenation prior to delivery but currently the information is poorly used, if at all. Use of erroneous data, inappropriate measures of 'acidemia', failure to distinguish between respiratory and metabolic components and unphysiological expectations about relationships to other measures of neonatal outcome were some of the problems highlighted. The use of generic terminology such as 'birth asphyxia' or 'acidosis' which have varying definitions has caused much confusion and should be avoided. There is unlikely to be one 'gold standard' measure of neonatal condition at delivery.