It is time to introduce ST analysis for fetal monitoring in the labor ward?
- G.H.A. Visser has stated explicitly that he has no conflicts of interest in connection with this article. Jørg Kessler received a lecture fee from Neoventa Medical one-one occasion.
Five randomized controlled trials have been published on intrapartum fetal heart rate monitoring with ST analysis of the fetal electrocardiogram, but the debate on its usefulness has not yet been ended. We consider ST analysis a useful and cost-effective addition to conventional fetal heart rate monitoring. We will provide support for this opinion by discussing the pathophysiology of ST changes in relation to fetal asphyxia, the results of the randomized controlled trials and numerous meta-analyses of these randomized controlled trials and trends in fetal outcome in hospitals in different countries following the introduction of the ST technology in the labor ward.
fetal blood sampling
fetal heart rate
individual participant data meta-analysis
randomized controlled trial
The clinical use of ST analysis (STAN) is based on a solid fundament of experimental data, of which only some key studies are mentioned in the following. Research on STAN started four decades ago with the observation of fetal electrocardiogram (ECG) changes during experimentally induced hypoxia in lambs, guinea pigs and cats . The observed T-wave elevation was dependent on the severity of hypoxia . Circulating fetal catecholamines were identified as the mediator between hypoxia and ST changes through beta-receptor stimulation and subsequent myocardial glycogenolysis . It could be shown that a differential threshold to intrauterine hypoxia was reflected in the fetal ECG: a T-wave increase was paralleled by lactacidemia, whereas a stable T wave indicated acid -base homeostasis . The investigation of the temporal relation between fetal ECG changes and cerebral function revealed an intact fetal cortical activity when hypoxia-induced T-wave changes occurred . This finding represented an important milestone for the later clinical use of STAN.
The concept of using ST interval changes of the fetal ECG as an adjunct to conventional cardiotocography (CTG) was then brought forward through technical refinements of the monitoring technology (automated ECG analysis), a structured educational program for the medical personal, followed by clinical use in selected European centers .
Cardiotocography abnormalities may occur early during high-risk labor. While a scalp electrode with STAN is applicable during latency or early in labor, fetal blood sampling (FBS) may not be feasible or may be technically challenging at that time.
Whereas the acquisition of the fetal ECG only requires one perforation of the fetal scalp, FBS often requires repeated incisions with the risk of infection and bleeding. The use of STAN is thus less invasive than FBS.
Information on significant ECG changes is available immediately. In contrast, the median decision-to-result interval in FBS is 16–18 min. Use of STAN is time-saving.
STAN supplies the user with continuous information on fetal cardiac metabolism, whereas FBS only gives information at the time on the acid-base homeostasis in peripheral fetal tissues.
Outcome of RCTs
Since 1993 five randomized controlled trials (RCTs) have been published on the use of ST-analysis in intrapartum fetal monitoring [5-11]. In the first trial a preliminary version this technology was used; three trials included more than 2300 patients and altogether more than 15 000 patients have been included. Table 1 summarizes the most important results. Overall, the use of ST resulted in a non-significant 20–30% reduction of severe metabolic acidemia in the umbilical artery (pH umbilical artery <7.05, base deficit >12 mmol/L in extracellular fluid), a 45% reduction in the use of FBS and a 10% reduction in operative vaginal deliveries. Most meta-analyses report more or less similar figures and this also holds for an individual participant data meta-analysis (IPDMA), which was based on the four more recent trials [12-14]. A meta-analysis from Sweden, also based on the latest four RCTs, showed only a 4% lower incidence of metabolic acidemia in the ST-arms . A lower incidence of acidemia at birth was only present in the three large trials in which FBS was 13–48% lower in the ST-arm as compared with the conventional fetal heart rate (FHR) monitoring arm. In the smaller trials, reduction in blood sampling was about 55% and acidemia at birth was non-significantly higher in the ST-arms. There was an inverse relation between reduction in acidemia and reduction in FBS, which may indicate that improved outcome only occurs if FBS is still used in a considerable number of cases. This may in particular have been true for the two smallest trials, with a lack of a proper learning effect given the small sample sizes and the largest reduction in FBS.
Table 1. Randomized controlled trials on the use of ST-analysis during labor; for data analysis and definitions of outcome parameters see Becker et al. .
| n ||2334||4966||1483||799||5681|
|RR metabolic acidosis||0.38 (0.14–1.07)||0.49 (0.24–0.97)||1.51 (0.43–5.34)||1.6 (0.53–4.86)||0.72 (0.43–1.19)|
|RR operative||0.9 (0.79–1.01)||0.88 (0.79–0.99)||1.03 (0.82–1.31)||0.98 (0.89–1.11)||0.96 (0.89–1.05)|
|RR FBS||0.81 (0.63–1.06)||0.87 (0.74–1.03)||0.45 (0.33–0.61)||0.44 (0.46–0.52)||0.52 (0.46–0.59)|
In the Swedish study a clear learning effect was found, with a further improved outcome in the second part of the trial . Similar findings were reported in the IPDMA of the four most recent trials, in which the decrease in metabolic acidemia was significant in the second part of the data set [RR 0.50; 95% CI 0.29–0.85 ]. This seems logical given that introduction of new technology usually requires a learning phase for clinicians to obtain adequate knowledge and gain confidence in the new assessment techniques. In the second part of the trial, outcome in the conventional arms had deteriorated, which also seems logical, at least if the first assumption is correct, that increased confidence in new technology may result in decreased confidence in the conventional technology. A better assessment of the contribution of new technology on outcome may be obtained by comparing the results of the conventional monitoring technique in the first half of the trial with the outcome in the ST-arm in the second half of the trial, although patient characteristics may have shifted in the course of a trial.
Improved outcome following the introduction of ST technology
A few centers have published and/or presented data on neonatal outcome in the years following the introduction of ST technology in the labor ward. In the University Hospital of Mölndal, Sweden, outcomes were prospectively monitored over a period of 7 years . In these years the use of ST technology increased from 29 to 69% of labors and the incidence of cord metabolic acidemia fell from 0.72 to 0.06% with no significant changes in cesarean and vaginal operative delivery rates. The cord blood sampling rate increased from 81 to 96% and therefore does not explain the striking fall in severe acidemia. The reduction in acidemia was associated with a significant reduction in the number of cases with a prolonged response time between CTG+ST indications to intervene in delivery, and the ability of the staff to identify and act on preterminal and “unstable” FHR patterns at the onset of a recording. The authors conclude that a paradigm shift in the outcome of delivery occurred, related to a high rate of CTG+ST usage and the application of structured CTG analysis.
A similar study was published from Bergen, Norway. In a period of 5 years, monitoring with CTG+ST increased from 20 to 33% of labors . In these cases the incidence of severe metabolic acidemia decreased from 1.4 to 0.3% and absolute numbers from 13 to five cases per year. The overall incidence of cesarean deliveries fell slightly in this period. Finally, 5 years of data from St. Georges Hospital, London, UK, showed that the use of ST technology with intensive process-based training in CTG interpretation and competency testing, was associated with a fall in severe metabolic acidemia from 1.35 to 0.76% and in neonatal deaths from 1.7 to 1.3/1000, with a decrease in emergency cesarean deliveries from 15 to 9% .
Large observational studies have also investigated the clinical performance of STAN either without  or with a low frequency (1.4 and 2.4%) of additional FBS [17, 20]. The metabolic acidosis rate in these studies was not different from those reported in the RCTs. In contrast to the use of conventional CTG alone, a recent study has shown that the risk of an adverse neonatal outcome after monitoring with CTG and ST-waveform analysis was dependent on the time between an indication of hypoxia and delivery .
Observational studies do not prove that STAN by itself is effective in improving outcome, but that combined with intensive training, the outcome definitely improves.
The use of ST technology during labor results in a 20–30% reduction in severe metabolic acidemia (ns) in a 45% reduction of FBS and in a 10% reduction of operative (vaginal) deliveries. Two studies have shown that the use of this technology is more or less cost-effective [22, 23]. So, what could be the reasons for not using this assessment technique? The only one we can think of is a policy of non-invasive (abdominal) monitoring of FHR. However, in the case of spontaneous rupture of membranes or induction of labor by amniotomy and oxytocin or prostaglandins, invasive FHR monitoring seems to be the method of choice, guaranteeing optimal signal quality without limiting maternal activity with belts around the abdomen.
There may be several reasons why ST technology improves fetal outcome (Table 2). First, experimental studies on pathophysiology have shown a clear relationship between changes in the ST segment and fetal stress and distress. Secondly, increasing medico-legal issues related to intrapartum FHR monitoring and discussions on how to lower cesarean delivery rates have stimulated a renewed interest in fetal monitoring and in new assessment techniques. This may well result in better interpretation of monitoring results and in more appropriate actions. In this respect it is interesting to note that outcome in the control arm of the largest (Dutch) RCT was much better than predicted, i.e. an incidence of severe metabolic acidemia 1.2% as compared with a predicted percentage of 3.5% , which was based on earlier local data. The strict guidelines, training and systematic FHR classification that come with the ST technology may further improve fetal assessment. Last but not least, the STAN provides the doctor with alerts in case of ST changes. This forces the caregiver to assess the clinical situation and to decide on whether to act. From earlier studies on intrapartum medico-legal issues it has become clear that doctors tend to wait in making decisions despite the presence of abnormal heart rate patterns [24, 25]. Moreover, inter-observer variation and the decision to act or otherwise are much lower in case of STAN alerts compared with classification and interpretation of FHR patterns alone . The progressive improvements in outcome reported by centers using ST-technology are most likely the result of a combination of all these factors.
Table 2. ST-analysis of the intrapartum fetal electrocardiogram; why would it work.
|● Scientific (experimental) evidence has shown a relationship between ST changes and medico-legal issues|
|● Increasing cesarean delivery rates stimulate the doctor to look for new monitoring techniques|
|● New assessment techniques result in more attention and better assessment of fetal heart rate patterns|
|● Strict guidelines, training and systematic fetal heart rate classification come with ST technology|
|● (Significant) ST events (alerts) force the doctor to act|
|● Subjective visual interpretation of fetal heart rate patterns|
|● The human factor|
There remain some weaknesses in fetal monitoring using STAN. The most important one is the subjective visual classification of FHR patterns, with its large inter-individual variation, which has to be done to evaluate the significance of ST-changes. The second weakness remains the human factor. In a discussion on 11 “false-negative” STAN cases it appeared that it was not so much the technology that failed but rather the doctor who did not act despite the presence very abnormal FHR patterns or significant ST events . Finally, criteria for significant ST-alerts may, with increasing experience, be modified.
In conclusion, RCTs on the introduction of ST-technology in the labor ward have shown a moderate improvement in outcome with a reduction of interventions. With continued use and introduction of strict training programs, outcome may well continue to improve. It is time to introduce this technique in the labor ward.