Background

Despite almost universal fetal monitoring during labor in recent decades, debates over its role and benefits persist among practitioners, stakeholders, and in the proliferation of obstetric lawsuits. There is a widespread agreement that “perinatal asphyxia,” is a leading cause of mortality and long-term morbidity among survivors . Further, the frequency of cerebral palsy (CP), neonatal seizures or neonatal encephalopathy in term infants, and allegations of negligence have not diminished , and substandard care during labor is responsible for a significant amount of preventable neurological injury .

The most vulnerable areas for preventable injury include the failure to respond to an abnormal CTG, the management of labor, and the timing of cesarean delivery. Importantly, these identified areas have persisted over time, geography, and, to some extent, initiatives to ameliorate the problems .

Assigning both preventability and timing of injury to labor and proving such a causative relationship in a courtroom are complicated by diverse, even divergent criteria that tend to dismiss both the events of labor and delivery and the CTG . The monographs from the American College of Obstetricians and Gynecologists (ACOG) among others maintain that an acute intrapartum event may only result in injury if there are immediate and severe signs of neonatal compromise, and that the ultimate CP is of the spastic quadriplegic or dyskinetic type . Requiring severe fetal asphyxia by means of critically low pH values in the umbilical artery (7.0 or lower, base deficit of −12 or greater), sustained low Apgar scores, and other criteria to establish a timing in labor disregard those with “asphyxial” assault earlier in labor or with perinatal injury from mechanical, ischemic, thromboembolic, or infectious causes that do not result in severe fetal acidemia or problems with neonatal adaptation. When such cases appear in courtrooms, however, they are supported by clinical, CTG, and neuroradiological assessments of the timing and mechanism of injury consistent with intrapartum, not antepartum, injury . Such deliberations call into question the pronouncements of “authoritative sources” and the guidelines for interpreting the CTG. The aim of this paper, therefore, was to review the use of the much contested CTG in litigation and in the understanding of the timing and mechanism of perinatal injury.

The precepts of CTG monitoring

We have used CTG in labor for the past 50 years to help prevent hypoxic neurologic injury during labor . In theory, severe or deteriorating hypoxia (not well defined) will prompt the “rescue” of the fetus from its hostile situation before the development of an injury . By this approach, CTG has dramatically reduced the risk of intrapartum fetal death attributable to intrapartum hypoxia , but not affected long-term handicap. At the same time, CTG is implicated in raising the cesarean delivery and operative vaginal delivery rates especially for “fetal distress” – many of which are considered unnecessary because “objective criteria” of hypoxia (low pH and low Apgar score) are not met .

The CTG is currently used as an instrument of “rescue” where we accept a “certain amount” of asphyxia (abnormal tracings/abnormal pH) before rescuing the fetus . We find this fundamental precept wanting and legally unhelpful. The goal of CTG monitoring must be to keep the fetus out of harm’s way by limiting wherever possible asphyxial/ischemic, mechanical, and infectious threats early in their course while at the same time maximizing fetal homeostasis to deal with such threats and avoiding the need to “rescue” the fetus.

The most prevalent theory of causation of perinatal brain injury derives from the experiments of Windle, Myers, and their colleagues . In these models, progressive asphyxiation results in impaired cardiac output ultimately causing diminished cerebral flow (ischemia) and neurological injury . The differing patterns of brain pathology are related to the severity and time course of the asphyxial insult . Clinically, most neuroradiological lesions (hypoxic–ischemic injury) represent infarction in one or more areas of the arterial circulation with sometimes significant differences in symmetry. In some, a focal infarction subscribes to the distribution of a major cerebral vessel, whereas in others injury results from venous infarction . Catastrophic events tend to involve the basal ganglia and hippocampus, areas with the highest blood flow, which control vital functions. These babies are often severely affected both at delivery and at follow-up . Although some infants with obvious encephalopathy may have no lesions, the majority of neuroradiological lesions involve white matter or cortex (or both) where the newborn course may be less dramatic . The affected neonate may even escape immediate detection or motor deficits only to be compromised by serious cognitive impairments later in life .

Cowan et al. showed that >90% of newborns with encephalopathy had evidence of perinatally acquired ischemic lesions on a magnetic resonance imaging (MRI) performed within the first 2 weeks after birth, irrespective of the severity of symptoms . In addition to injuries associated with hypoxic–ischemic encephalopathy (HIE), intracranial hemorrhage (ICH) and focal cerebral infarction appear with increasing frequency . Of 33 such infants reported by Takenuochi and Perlman, 13 (about 40%) were initially asymptomatic, and they were admitted to the well-baby nursery.

Semantics, unfortunately, have come to play a distracting role in this area . HIE, for example, is widely used to refer to neonatal encephalopathy with evidence of “severe” hypoxia/acidemia during parturition, whereas “neonatal encephalopathy” describes infants with abnormal neurological findings of any etiology . Numerous other terms have been used despite considerable variation in how they are defined .

Most fetuses injured during labor are not severely asphyxiated although that has not, in many cases, excluded the diagnosis of HIE . In 51,519 cases, Yeh et al. found that the majority of infants with neonatal encephalopathy or death had a pH of >7.00. They confirmed that neonates with “hypoxia” may not develop acidemia, and that even catastrophic intrapartum events may fail to show cord acidemia or obvious acute insult. Mechanical effects on cerebral circulation may also cause injury, and normal cord gases are not incompatible with brain hypoxia from this mechanism . To exclusively equate patient safety and the opportunity to improve outcome simply by using the CTG as a surrogate of asphyxia/acidemia and by only studying those with a low pH at delivery assures that the effort will fall short of the objective of insuring “safe passage” for the fetus or for the defendant in the courtroom.

The precepts of CTG monitoring

We have used CTG in labor for the past 50 years to help prevent hypoxic neurologic injury during labor . In theory, severe or deteriorating hypoxia (not well defined) will prompt the “rescue” of the fetus from its hostile situation before the development of an injury . By this approach, CTG has dramatically reduced the risk of intrapartum fetal death attributable to intrapartum hypoxia , but not affected long-term handicap. At the same time, CTG is implicated in raising the cesarean delivery and operative vaginal delivery rates especially for “fetal distress” – many of which are considered unnecessary because “objective criteria” of hypoxia (low pH and low Apgar score) are not met .

The CTG is currently used as an instrument of “rescue” where we accept a “certain amount” of asphyxia (abnormal tracings/abnormal pH) before rescuing the fetus . We find this fundamental precept wanting and legally unhelpful. The goal of CTG monitoring must be to keep the fetus out of harm’s way by limiting wherever possible asphyxial/ischemic, mechanical, and infectious threats early in their course while at the same time maximizing fetal homeostasis to deal with such threats and avoiding the need to “rescue” the fetus.

The most prevalent theory of causation of perinatal brain injury derives from the experiments of Windle, Myers, and their colleagues . In these models, progressive asphyxiation results in impaired cardiac output ultimately causing diminished cerebral flow (ischemia) and neurological injury . The differing patterns of brain pathology are related to the severity and time course of the asphyxial insult . Clinically, most neuroradiological lesions (hypoxic–ischemic injury) represent infarction in one or more areas of the arterial circulation with sometimes significant differences in symmetry. In some, a focal infarction subscribes to the distribution of a major cerebral vessel, whereas in others injury results from venous infarction . Catastrophic events tend to involve the basal ganglia and hippocampus, areas with the highest blood flow, which control vital functions. These babies are often severely affected both at delivery and at follow-up . Although some infants with obvious encephalopathy may have no lesions, the majority of neuroradiological lesions involve white matter or cortex (or both) where the newborn course may be less dramatic . The affected neonate may even escape immediate detection or motor deficits only to be compromised by serious cognitive impairments later in life .

Cowan et al. showed that >90% of newborns with encephalopathy had evidence of perinatally acquired ischemic lesions on a magnetic resonance imaging (MRI) performed within the first 2 weeks after birth, irrespective of the severity of symptoms . In addition to injuries associated with hypoxic–ischemic encephalopathy (HIE), intracranial hemorrhage (ICH) and focal cerebral infarction appear with increasing frequency . Of 33 such infants reported by Takenuochi and Perlman, 13 (about 40%) were initially asymptomatic, and they were admitted to the well-baby nursery.

Semantics, unfortunately, have come to play a distracting role in this area . HIE, for example, is widely used to refer to neonatal encephalopathy with evidence of “severe” hypoxia/acidemia during parturition, whereas “neonatal encephalopathy” describes infants with abnormal neurological findings of any etiology . Numerous other terms have been used despite considerable variation in how they are defined .

Most fetuses injured during labor are not severely asphyxiated although that has not, in many cases, excluded the diagnosis of HIE . In 51,519 cases, Yeh et al. found that the majority of infants with neonatal encephalopathy or death had a pH of >7.00. They confirmed that neonates with “hypoxia” may not develop acidemia, and that even catastrophic intrapartum events may fail to show cord acidemia or obvious acute insult. Mechanical effects on cerebral circulation may also cause injury, and normal cord gases are not incompatible with brain hypoxia from this mechanism . To exclusively equate patient safety and the opportunity to improve outcome simply by using the CTG as a surrogate of asphyxia/acidemia and by only studying those with a low pH at delivery assures that the effort will fall short of the objective of insuring “safe passage” for the fetus or for the defendant in the courtroom.

The classification of CTG patterns

Guidelines for the use of CTG patterns have not been consistent in their recommendations, and interpretation has been shown to be of a low standard in clinical practice and in allegations of malpractice . The CTG classification from the ACOG identifies Categories I–III according to their perceived likelihood of fetal acidemia with no attempt made to provide a physiological or a “neurological/behavioral” assessment of the fetus. Nor do these guidelines use the fetus as its own control (baseline rate and variability) or provide a specific assessment of the evolution of patterns or the “recovery” from decelerations.

Although the severity of acidemia increases with increasing severity of the patterns, the classifications are poor predictors of fetal acidemia or subsequent injury . In the ACOG classification, Category I represents a “normal” CTG pattern without the threat of acidemia. Reasonably, with recurrent accelerations with fetal activity and alternating sleep/wake or rest activity cycles and moderate variability and stable baseline rate, one is surely entitled to the inference of a normal fetal behavior with appropriate neurological responsiveness – and absent injury – a crucial determination . Category III is “abnormal” with significant, but far from universal, threat of acidemia. The broad Category II, which appears in 80–90% of tracings, is “non-diagnostic” with a low risk of acidemia. Both acidemia and injury may be seen with Category II tracings, and clinicians may be unable to separate Category II from Category III . Compared with the first stage of labor, Category I tracings decrease, and Category II and III tracings increase during the second stage of labor. Increasing the time in Category II in the last 2 h of labor is associated with an increased short-term newborn morbidity .

Numerous modifications of this classification have been offered for the purposes of refining the prediction of “acidemia,” and for mounting a more selective intervention strategy . Waiting to intervene until a tracing fulfills the criteria for a Category III (or an “Orange/Red” condition in the five-tier system) may satisfy requirements for an “indicated” cesarean, but it may be too late to prevent an injury or provide timely neuroprotection for the newborn. The time needed to prevent fetal injury once variability is lost is unknown, but on occasion, the previously normal fetus may suffer ischemic injury so rapidly and without warning that even the most expedited intervention will not prevent an injury .

In some malpractice cases, the fetus presents with an abnormal tracing on admission that reflects a previous neurological insult . Such fetuses are unlikely to profit from cesarean delivery as long as ongoing asphyxia/ischemia can be avoided.

In the previously normal fetus, decelerations appear before changes in the baseline rate or variability and changes in the baseline rate in response to decelerations generally precede changes in the baseline variability . If hypoxia and injury are to be prevented, intervention (but not necessarily delivery) should reasonably occur long before the sustained changes in rate and/or variability.

Ultimately, the notion of using CTG patterns to “rescue” the distressed fetus fails to satisfy fundamental concepts of preventive care. Although severe CTG abnormalities may justify intervention, because they have evolved from Category I tracings, they also increase the exposure to litigation in the event of a poor outcome. Clinically or in a courtroom, safeguarding the mother and fetus by keeping them out of harm’s way in the first place appears to be a more easily assimilated principle of care than deciding, with limited guidance, where on the spectrum of CTG abnormality intervention will take place.

CTG – technical issues

Poor-quality tracings or unintentional recording of the maternal heart rate, especially during the expulsive stage of labor, reduces the ability to recognize threatened fetal homeostasis . The frequency of this clinical problem is unknown although, despite publicity, it remains a frequent allegation in malpractice cases that is difficult to refute.

Uterine activity and the forces on the fetal head

Uterine contractions (UCs) decrease oxygen availability to the placenta and fetus by reducing the flow in the uterine artery branches that traverse the myometrium . UCs also increase fetal intracranial pressure (ICP) potentially to a level of two to four times the associated intrauterine pressure (IUP). Even normal strength contractions can increase ICP and reduce cerebral blood flow (CBF) at least briefly even in the absence of systemic hypoxia or acidosis and without significant changes in the systemic fetal pO ₂ or in cardiac performance .

The increased pressures on and within the skull compared with the IUP are due to the anatomy of the fetal skull and the resistance offered by the cervix as well as by muscular and bony pelvic structures . Additional influences potentially diminishing both placental and/or intracranial blood flow during labor and delivery are described in Table 1 .

The mechanical complications confronting the macrosomic fetus are well known. Molding of the fetal head occurs in response to the compressive forces of UC and the birth canal. The pliability of the fetal skull bones allows the sutures to override, thereby reducing the intracranial volume, but potentially increasing the ICP in return . “Excessive molding” anticipates cephalopelvic disproportion (CPD) and other adverse consequences . Early amniotomy increases the amount of cranial molding, which, if exaggerated, may produce “early” decelerations and ischemic and hemorrhagic lesions in the fetal brain . In an ultrasound flow study , early decelerations appeared in 67% of cases with absent end-diastolic flow velocity, and in 100% with reversed diastolic flow with normal fetuses. “Head compression” may also cause variable decelerations, especially in the second stage of labor .

It seems necessary to revise the notion that “early” or “variable” decelerations are invariably benign (meaning “not associated with hypoxia or acidemia”) or that variable decelerations only indicate umbilical cord compression and that either may be tolerated indefinitely. Rather, persistent early or variable decelerations, in either phase of labor, should prompt an evaluation of the feasibility of safe vaginal delivery with an assessment of the descent, position, and molding of the fetal head .

The pattern of pushing may influence the forces on the fetal head . During the passive phase of the second stage, pushing is withheld. Generally, there is some progress in the descent, but no deterioration of the scalp pH. With coached, Valsalva-based, maternal pushing, closely spaced, exaggerated peaks may prolong the exposure of the fetus to increased compressive forces, potentially diminishing the time for recovery and causing the pH to fall . Alternatively, non-Valsalva-pushing strategies with open glottis and slowly developing transient peak pressures not only appear to subject the fetus to less head compression, but they may reduce the frequency of decelerations and improve Apgar scores and umbilical pH values at delivery . The fetus that enters the second stage already compromised may deteriorate rapidly during this time – again, emphasizing the role of pushing during the second stage in the causation of injury.

As the duration of the second stage of labor increases, so do the risks of operative delivery, adverse maternal and fetal outcome, and trauma to the fetus . The features most prognostic of adverse outcome appear to be the duration of pushing without progress in the descent and the recovery pattern of any deceleration that pushing induces. It appears reasonable, for example, to permit a 5-h (combined passive and active) second stage as long as there is normal, true progress in the descent of the presenting part (distinguished from an increasing caput) and a reassuring tracing. The recommendation that intervention must await pushing for 3 or more hours without progress in the descent appears to be quite a riskier circumstance .

Despite normal uterine activity, the occiput posterior (OP) position is associated with decelerations, prolonged labor, marked molding, failed instrumental delivery, cesarean delivery, and oxytocin administration , and it is an independent risk factor for fetal trauma, subsequent CP and low mental scores in the offspring, and medicolegal challenge . Manual rotation from the OP position to a more favorable position may reduce some of the stresses on the fetal head, and it may diminish the frequency of operative delivery .

Operative vaginal delivery is a risk factor for mechanical/traumatic/ischemic injury to the scalp and brain, including subgaleal hemorrhage . Of those injured in association with ventouse delivery, the majority appear to have been injured in the second stage – before the applications of the device – determined by the CTG pattern .

Excessive uterine activity

For the purpose of identifying XSA, the ACOG guidelines define “tachysystole” as a frequency of contractions >5 in 10 min, averaged over 30 min . The frequency of contractions alone, however, is an inadequate assessment of uterine activity and an insufficient guideline for the control of oxytocin. The hypoxemic effects of XSA appear long before 30 min, especially in the second stage, and, as defined, it is found to be associated with an increased risk of adverse neonatal outcome especially in oxytocin-stimulated labors .

Diminished relaxation time, not frequency, appears to be a more sensitive metric for the recognition of XSA than does contraction frequency . Bakker et al. and others have found that uterine rest of at least 60 s between moderate to strong contractions appears necessary to allow sufficient time for reperfusion and adequate oxygen delivery to the fetus . Often, especially in malpractice cases where XSA is commonplace, the evaluation of uterine activity seems based on the notion that as long as the CTG pattern is normal, contractile forces cannot be “excessive.” This unfortunate perspective gives more control of the oxytocin infusion to the CTG pattern and less to the details of the contraction pattern or progress in labor . XSA does not reliably enhance progress in active labor or facilitate recovery should the fetus respond adversely . The resilient fetus is usually able to tolerate considerable amounts of even XSA in the 1st stage of labor . As labor progresses further, however, XSA is clearly associated with adverse neonatal outcome, decreased umbilical artery pH at the time of delivery, abnormal CTG patterns, and even neonatal encephalopathy .

These various pitfalls in properly assessing uterine activity and determining when it is excessive account for the high prevalence of “tachysystole,” for the lack of a consistent response to its appearance and for the widespread allegations of oxytocin abuse in medicolegal cases involving adverse neonatal outcomes . The European recommendation to limit the frequency of contractions to 7 in 15 min seems far more physiological and safe than the 5 in 10 limit in the United States .