CTG monitoring is in use during a woman’s labour. The fetal heart rate pattern is now abnormal. As her maternity professional, you know that the positive predictive value of the CTG is low, and want to avoid caesarean section unless it is going to improve the fetal outcome. And you don’t want to misidentify when caesarean section is a great idea and as a result have an avoidable poor outcome for the fetus / baby. What can you do?
Historically, the most common option in this scenario was to collect a sample of capillary blood from the fetal scalp and test the pH. The aim was to distinguish the fetus who is acidotic because of hypoxia from one who is not, making it possible to recommend surgical birth only when acidosis has been confirmed. In more recent years, lactate testing of the sample rather than pH testing has become more common, mostly because the equipment is cheaper, more transportable, and the amount of blood needed is much less. There are regional differences in how often fetal blood sampling is used, and the draft NICE guidelines suggest that it should be abandoned as evidence that it improves outcomes is absent.
Analysis of the ST segment of the fetal ECG (STAN) relies on an association between fetal acidosis and the shape of the ST segment (Rosen, et al., 1976). STAN might then be a useful adjunct to CTG monitoring. So far, STAN hasn’t been shown to reduce the caesarean section rate compared to CTG monitoring alone or improve perinatal outcomes other than a small reduction in the incidence of acidosis at birth (Blix, et al., 2015). It does seem reasonable to test whether STAN is a better addition than fetal blood sampling when the fetal heart rate pattern is abnormal.
Back in 2005, a research team from Denmark set out to look at this issue, and collected data until 2012 (Victor, et al., 2022). The death of a senior member of the research team slowed progress, but their work has now been published. Women with a singleton, term, cephalic presenting fetus who developed an abnormal fetal heart rate pattern (according to FIGO guidelines) and who had a normal fetal pH (>7.25) determined by fetal blood sampling were offered randomisation to either ongoing fetal blood sampling or STAN. CTG monitoring was continued in both groups after placing a fetal spiral electrode. STAN data were collected for women randomised to the fetal blood sampling group but were not available to clinicians during the woman’s labour. Fetal blood sampling in the STAN group was restricted to situations when the ECG signal quality was poor. The main outcome measure was cord blood gas acidosis (pH < 7.05 and base excess of -10 or less), and a power calculation based on a 50% reduction in this outcome informed the size of the study population. The study enrolled slightly below this target (1,013 women rather than the 1,025 planned).
There were no statistically significant differences in the rates of metabolic acidosis in the two groups (1.5% with STAN + CTG and 0.8% with fetal blood sampling + CTG, RR 1.96, 95% CI 0.49 – 7.79), but the trend towards a higher rate with STAN use is concerning. This amounts to an additional one baby with acidosis for every 143 fetuses monitored with STAN. There were no differences in low Apgar scores, admissions to the nursery, or a composite score that included all these outcomes.
There were statistically significant differences in instrumental vaginal birth (34.0% with STAN + CTG and 25.8% with fetal blood sampling + CTG, RR 1.32, 95% CI 1.08 – 1.60). Rates of episiotomy for fetal indications were more than doubled in the STAN group, and there was a 38% increase in the use of caesarean section for fetal indications. The use of caesarean section for other indications was lower in the STAN group (16.7% with STAN + CTG and 25.4% with fetal blood sampling + CTG, RR 0.66, 95% CI 0.51-0.84). As a consequence, there was no difference in the total caesarean section rate.
Of the 507 women randomised to ongoing fetal blood sampling, 45.4% had one or two additional samples tested after the initial sample performed at enrolment, 28.8% had three to five additional samples and 3.4% had six to ten samples. 22.4% of women had none. A high number of women in the STAN group had one or two additional pH samples tested (50.6% – not statistically significant to the fetal blood sampling group), but only 11.4% of the STAN group had more than two tests and 38% had none performed. While the authors correctly claimed that STAN use reduced the use of fetal blood sampling, I was surprised by how often it continued to be used. This seems to imply that issues with the quality of the ECG signal required for STAN were very common, but this was not commented on by the authors.
I do wonder whether there may have been a link between the use of high numbers of fetal blood samples and the higher caesarean section rate for non-fetal indications in the fetal blood sampling group. I can’t recall ever having done more than three for the one woman in labour, and the thought of doing ten is – disquieting… Fetal blood sampling isn’t a pleasant experience, so women and clinicians may have sought other reasons for caesarean section to reduce the use of repeated pH tests.
Blix and colleagues’ recent population study (2022) looking at the outcomes of widespread use of STAN in Norway did not collect data about metabolic acidosis, but reported a worrying rise in the rate of babies with an Apgar score of less than seven at five minutes of age (I wrote about this paper recently). The trend towards higher rates of acidosis seen in this research suggests a question that requires a larger population to address.
People interested in reducing the incidence of severe perineal trauma might be interested to note that STAN was associated with higher rates of two interventions known to increase the incidence of severe perineal trauma – namely episiotomy and instrumental birth. Efforts at reducing pelvic floor trauma might be more effective if they consider the impact of fetal monitoring practices on this outcome.
We still lack convincing evidence that adding fetal blood sampling to CTG monitoring improves outcomes. Adding STAN to the mix as well appears to increase surgical trauma for women while possibly allowing worse perinatal outcomes. While this new research should not be interpreted as providing a vote of confidence for fetal blood sampling, it should raise a few eyebrows regarding the use of STAN technology.
Blix, E., Eskild, A., Skau, I., & Grytten, J. (2022). The impact of the introduction of intrapartum fetal ECG ST segment analysis. A population study. Acta Obstetrica et Gynecologica Scandanavica, in press, 1-10. https://doi.org/10.1111/aogs.14347
Rosen, K. G., Hökegård, K. H., & Kjellmer, I. (1976). A study of the relationship between the electrocardiogram and hemodynamics in the fetal lamb during asphyxia. Acta Physiologica Scandinavica, 98(3), 275-284. https://doi.org/10.1111/j.1748-1716.1976.tb10312.x
Victor, S. F., Bach, D. B. B., Hvelplund, A. C., Nickelsen, C., Lyndrup, J., Wilken-Jensen, C., Scharff, L. J., Weber, T., Secher, N. J., & Krebs, L. (2022, Jun 14). Cardiotocography combined with ST analysis versus cardiotocography combined with fetal blood sampling in deliveries with abnormal CTG: a randomized trial. Archives of Gynecology & Obstetrics, in press. https://doi.org/10.1007/s00404-022-06649-3