Relationship Between Umbilical Cord Gas Values and Neonatal Outcomes

Abstract

OBJECTIVE: To evaluate the relationship between umbilical artery cord gas values and fetal tolerance of labor, as reflected by Apgar score. We hypothesized the existence of wide biological variability in fetal tolerance of metabolic acidemia, which, if present, would weaken one fundamental assumption underlying the use of electronic fetal heart rate (FHR) monitoring. METHODS: We conducted a retrospective cohort study of term, singleton, nonanomalous fetuses delivered in our institution between March 2012 and July 2020. Universally obtained umbilical cord gas values and Apgar scores were extracted. We calculated Spearman correlation coefficients and receiver operating characteristic curves for various levels of umbilical artery pH, base excess, and Apgar scores. RESULTS: We analyzed data from 29,787 deliveries. The statistical correlation between umbilical artery pH and base excess and both 1- and 5-minute Apgar scores was weak or nonexistent in all pH range subgroups (range 0.064–0.213). Receiver operating characteristic curve analysis suggested umbilical artery pH value of 7.22 yields the best discrimination for prediction of a severely depressed newborn (5-minute Apgar score less than 4), but sensitivity and specificity for this predictive value remains poor to moderate. CONCLUSION: The use of electronic FHR monitoring is predicated on a documented relationship between FHR patterns and umbilical artery pH, and an assumed correlation between pH and fetal outcomes, reflecting fetal tolerance of labor and delivery. Our data demonstrate a weak-to-absent correlation between metabolic acidemia and even short-term fetal condition, thus significantly weakening this latter assumption. No amount of future modification of FHR pattern interpretation to better predict newborn pH is likely to lead to improved newborn outcomes, given this weakness in a fundamental assumption on which FHR monitoring is based.

Reliability and agreement in intrapartum fetal heart rate monitoring interpretation: A systematic review

“…Second, the concepts of agreement and reliability were sometimes misused interchangeably, though they are two different concepts 7 . Among the statistical agreement measures presented in the included studies, only κ and quadratic weighted κ can be interpreted as reliability measures, as defined in Lord and Novick's 53,54 classical test theory. Further, reliability is a measure specific to the population studied and can only be generalized to populations with similar characteristics 7 .…”

Category II Intrapartum Fetal Heart Rate Patterns Unassociated With Recognized Sentinel Events

“…A category I FHR tracing simply indicates that the fetus is, at a particular moment in time, neither hypoxic nor experiencing an arterial pH below 7.15-7.2. 3 It says nothing at all about fundamental developmental neurologic integrity of the fetus; even the anencephalic fetus, representing perhaps the extreme of developmental central nervous system abnormality, may demonstrate normal FHR variability and accelerations. 24,25 The process of normal labor is one of repetitive hypoxia-inducing contractions, which leads to the development of significant metabolic acidemia.…”

“…3 This has occurred without any evidence that delivery based on any single or combination of FHR patterns reduces the rate of cerebral palsy or any other permanent long-term adverse neurologic outcome. 3,[6][7][8][9][10] This observation gives rise to a conundrum of major clinical significance. On the one hand, It is clear that certain FHR patterns are associated, albeit weakly, with lower umbilical artery pH values, that significant metabolic acidemia at birth may, on rare occasions be associated with hypoxic-ischemic encephalopathy, and that hypoxic-ischemic encephalopathy may lead to the later development of cerebral palsy.…”

Resistance to Change

“…It has attacked the very poor performance metrics for EFM that even its own key opinion leaders admit and essentially proposed that various kinds of risk factors and uterine contraction frequency be used to contextualize EFM data and expand the clinical utility of EFM-based predictive models [114][115][116]. Conceptually, this should have been seen as a relatively minor change but with important impact, i.e., recognizing that the effectiveness of a technology might be conditioned by other factors [117][118][119][120]. Nevertheless, the resistance to even considering a new approach has been considerable, sometimes publicly including personal invectives even on paper.…”