The search for an adequate back-up test for intrapartum fetal heart rate (FHR) monitoring has spanned the entire history of this technology. Although randomized controlled trials (RCTs) of electronic FHR monitoring (EFM) have been disappointing, the expectations may have been unrealistic.

The hope that this technology would lower the rate of cerebral palsy from intrapartum asphyxia has not materialized. There is, however, reasonable good evidence that EFM does decrease the rate of perinatal mortality, particularly early infant death because of hypoxia. Two analyses, 1 a large population analysis and 1 metaanalysis of 9 RCTs and at least 1 RCT have confirmed this reduction in death rates. Clearly, before EFM, intrapartum deaths were relatively common, occurring in approximately 2-3 per 1000 laboring patients, and now such deaths are rare. Nevertheless, the price we pay for this technology is high. It is a very difficult and hard task to master technology. All studies that have evaluated EFM have shown an increase in cesarean delivery rates in the electronically monitored patients. EFM has created a malpractice nightmare for obstetricians because of the unrealistic expectation that we can prevent all cases of neonatal encephalopathy with early operative intervention in hypoxic babies. Perhaps if we examine what the technology can and cannot do, these results with EFM should not be totally surprising. EFM is highly reliable (sensitive) for the detection of fetal hypoxia. Virtually no significantly hypoxic baby will have a normal (category I) tracing. But the problem with EFM is that many factors other than hypoxia cause abnormal tracings. As well illustrated in a subsequent study of the National Institutes of Child Health and Human Development guidelines for the categorization of EFM patterns, 84% of patients will have category II or “indeterminate” FHR patterns. Thus, it can be said that EFM is an excellent screening test that reliably confirms the absence of hypoxia but that its specificity is poor. Like other screening tests that we are familiar with in obstetrics (such as screening for aneuploidy), when a test result is normal (as with a category I FHR tracing or a negative serum screen for aneuploidy), we are comfortable not doing any further intervention. However, when the test result shows an increased but not definitive likelihood of abnormality (as with a category II FHR tracing or a positive serum screen for aneuploidy), a back-up test is needed. For a serum screen, this is a diagnostic test (amniocentesis or chorionic villus sampling) that gives a definitive answer. With EFM, there is therefore a desperate need for such a back-up test. Indeed, we have been searching for such a back-up since the inception of this technology.

Initially, scalp pH monitoring was the back-up test. Although still used in certain parts of the world, scalp pH monitoring has disappeared largely from use in the United States, mainly because of its technical difficulty and unreliability. Other technologies (such as continuous pH or pO ² monitoring) have not been perfected. Fetal pulse oximetry seemed initially to hold great promise, given its impact in adult and pediatric care, especially anesthesiology; however, the technology never gained favor, probably because it was introduced before it was optimized technically.

The most recent technology available that shows promise is the STAN technology (Neoventa Medical, Gothenburg, Sweden) or analysis of the ST segment of the fetal electrocardiogram waveform. Previous randomized trials from Europe suggest that this modality may hold promise to provide us with the back-up test that we need for EFM ; one of the trials suggested that ST analysis could also reduce the rate of metabolic acidosis.

In this issue of the Journal, Schuit et al have performed a sophisticated “individual participant data metaanalysis (IPDMA)” focusing on the RCTs of the use of the STAN technology. As the article points out in its introduction, previous metaanalyses that have used aggregate data have been limited in their abilities to perform accurate analyses of RCTs as comparison of endpoints are quite difficult. RCTs vary in the way endpoints are described, what variables are included, what ranges are included in the variables, and, in some cases, whether variables are included at all. By gathering the actual raw data from previous RCTs, the authors of this study have been able to do much more sophisticated and probably more valid analyses and to draw conclusions regarding how well this new technology could help us with the limitations of EFM.

The conclusions of the individual participant data metaanalysis appear to be disappointing, because no differences were found between monitoring with and without ST analysis in the rates of metabolic acidosis and the rates of cesarean deliveries, although there was a reduction in operative vaginal deliveries and admissions to the neonatal intensive care unit in the ST-analysis group. One finding that is most interesting relative to practice in the United States was a reduction in the need for scalp sampling. Because scalp sampling is not in use in the United States and because scalp sampling results are often normal in the face of an indeterminate or category II FHR tracing, this practice should avoid some cesarean deliveries. Thus, it is possible that the finding of reduced need for scalp sampling could translate into a reduction in cesarean deliveries in countries such as the United States where scalp sampling is not used.

ST analysis does seem to have some other appealing aspects. Unlike EFM alone, which requires a great deal of sophistication and training to be proficient, the STAN technology has the interpretation of the electrocardiogram waveform done by the computer, and “STAN events” are the outputs that alert the clinician to the likelihood of fetal hypoxia. These, in combination with the FHR pattern, create (in theory) a more accurate and hopefully simpler and more uniform means of improving the specificity or likelihood of real hypoxia when an abnormality is alerted. However, the algorithms for management with STAN events still require a sophisticated ability to interpret FHR patterns.

STAN technology does have disadvantages. It cannot truly perform as a back-up test because (in theory) it needs to be on the patient and to establish a baseline before any hypoxia exists and thus must be used on all patients or at least all “high risk” patients rather than just those with a nonreassuring or category II tracing. Although often a FHR pattern reverts back and forth between category I and II tracings, STAN technology can certainly be applied in such cases. Also, perhaps because the vast majority of patients will ultimately have a category II tracing, this is not a great limitation. Category II is so heterogeneous that the majority of these patients really do not need a back-up test with reasonably well-trained clinicians. Unlike EFM, which was introduced without real randomized clinical trials, ST analysis has been subjected to multiple randomized trials. Whether the knowledge from these randomized trials applies specifically to the way EFM is used in the United States is also questionable and certainly requires testing. The use of scalp pH measurements in both groups in the RCTs of ST analysis in Europe, as previously mentioned, make it difficult to know how this translates into practice in the United States. Another variable that is different in the United States and Europe is that midwives provide the majority of intrapartum care in Europe and presumably in the existing RCTs; this is not the case in the United States.

Currently, a large RCT of ST analysis vs EFM is ongoing in the United States by the National Institute of Child Health and Human Development Maternal Fetal Medicine Units Network and, hopefully in the near future, will provide an answer as to how this technology plays out in this country. Whatever the case, from the current article in the Journal, it is apparent that the optimal solution for the much needed back-up test for EFM has not yet been found and that the need still desperately exists.