Limitations of ST analysis in clinical practice: three cases of intrapartum metabolic acidosis

Sir, Westerhuis et al. described three cases of neonatal encephalopathy associated with severe intrapartum metabolic acidosis, which were not detected by fetal ST segment ANalysis (STAN).1 In previous trials, babies monitored with STAN, who were born with undetected metabolic acidosis had good Apgar scores and neonatal outcome,2,3 suggesting that such acidosis was not severe enough to impair neonatal adaptation in those babies and that the three recent cases1 probably had other coincidental factors at play. The authors and several other expert commentators highlighted gaps in the STAN clinical guidelines but appeared to overlook the critical role of noxious intrapartum factors such as fetal sepsis, maternal pyrexia, oxytocin infusion/tachysystole, birth trauma, and meconium. There is compelling experimental and human epidemiological evidence that the risk of neonatal encephalopathy and cerebral palsy increased exponentially following the combined exposure of the fetus to infection and birth asphyxia. Infection/inflammation sensitises the fetal neurological tissue to injury by lowering the threshold at which hypoxia exerts neuronal cell apoptosis.4 The magnitude of such synergistic injury is far in excess of what might be expected if the inflammation or hypoxia was operating alone.5 In a meta-analysis of studies in full-term infants, clinical chorioamnionitis was associated with a 4.7-fold increase in the risk of cerebral palsy (95% CI 1.3–16.2),6 whereas exposure of the fetus to a combination of infection and intrapartum hypoxia dramatically increased the risk of spastic cerebral palsy (OR 78; 95% CI 4.8–406) and spastic quadriplegic cerebral palsy (OR 367; 95% CI 19–1974) compared with hypoxia alone.7,8 The fetal injuries in cases 1 and 2 were likely to have been exerted by a synergistic interaction between fetal inflammation, evidenced by group B streptococcal sepsis, histological chorioamnionitis, funisitis and maternal pyrexia, and intrapartum hypoxia from intermittent asphyxial insults of contractions aided by syntocinon infusion. The lack of decelerations with contractions on the cardiotocograph (CTG) traces shown in figure 1A, B suggests that there was no significant hypoxia at that stage, as was confirmed by the acid–base status with pH of 7.29, unlike in figure 2B–D in case 2, where monitoring was discontinued when there was evidence of emerging hypoxia. This, however, does not exclude a nonhypoxic fetal injury driven by inflammation or by very modest and relatively innocuous levels of hypoxia-ischaemia on a background of fetal systemic inflammatory response. It is pertinent that the frequency of contractions in the CTG segments displayed in figure 1A–D ranged between six and seven every 10 minutes. At present, no tool or modality of intrapartum fetal monitoring has been shown to detect infection/inflammatory or synergistic injuries or highlight their timing of onset. Therefore, the real monitor is the clinician, who is best placed to exercise judgement as to when the cumulative risk of all the risk factors acting in context becomes untenable.9 Research is urgently needed to address the critical question of whether fetal infection/inflammation modify its metabolic response and/or cardiac electrical signalling in such a way that key information is missed by computerised ST segment analysis? The implications of such research for intrapartum care and monitoring will go far beyond STAN and impact a significant proportion of nonhypoxicischaemic neonatal encephalopathies. The weakest part of the clinical application of STAN monitoring is the visual interpretation of the CTG. This was evident in the three reported cases1 and was typical of our experience at St George’s Hospital, London. For example, the reduced fetal heart rate variability observed in all three cases suggested impaired fetal autonomic nervous system function and should have prompted action. The continuing work to digitally analyse and quantify reduced fetal heart rate variability is welcome and timely10 and will potentially reduce observer error in characterising the abnormal CTG. j