Electronic fetal monitoring, cerebral palsy, and caesarean section: assumptions versus evidence

Abstract

Given evidence that cerebral palsy is not reduced by electronic fetal monitoring, Karin Nelson, Thomas Sartwelle, and Dwight Rouse ask why routine monitoring and related litigation continue to contribute to high rates of caesarean births

A third of babies born in the United States,¹ and a quarter in England,² are delivered by caesarean section. Rates of caesarean have risen substantially since the 1970s, and current rates exceed the optimum to minimise mortality and morbidity of mothers and infants.³

An important driver of the rising rate of operative deliveries is the use of continuous electronic fetal monitoring (cardiotocography) in labour. It was introduced to enable early identification of fetal asphyxia in the hope of preventing death or long term neurological morbidity, especially cerebral palsy. However, it has a high false positive rate and has not been shown to reduce cerebral palsy.⁴ We examine the evidence on its effectiveness and some unintended consequences of its use.

Electronic fetal monitoring and cerebral palsy

Cerebral palsy was once thought to be caused mainly by birth asphyxia. Early advocates of electronic fetal monitoring expected that early detection of asphyxia in labour would largely eliminate cerebral palsy. Studies using surrogate endpoints such as low neonatal pH and low Apgar scores encouraged that expectation. Although it was initially developed for use in births in which the fetus was thought to be at increased risk, electronic monitoring soon became a standard intervention in obstetrics. Clinical trials were not conducted before monitoring was introduced because it was considered unethical to deny the expected benefits to controls.⁵

Only two randomised trials, both published more than 25 years ago, have compared cerebral palsy rates in births monitored electronically or by intermittent auscultation. Cerebral palsy cannot be diagnosed at birth, so infants have to be followed for several years, until a diagnosis can reliably be made. In a trial in 13 079 births in Dublin, evaluation at age 4 years indicated that the rate of cerebral palsy was not lower in children whose births were monitored electronically.⁶ A multicentre randomised trial in the US compared 93 singleton, vertex presenting infants with birth weight ≤1750 g whose births were monitored electronically with 96 comparable children monitored by intermittent auscultation.⁷ At 18 months of age, the cerebral palsy rate was significantly higher in the electronically monitored group.⁷

As a predictor of cerebral palsy, abnormalities on electronic monitoring have a false positive rate of 99.8%,⁴ so almost all abnormalities are false positive results. Using additional observations, including depression of the ST segment of the fetal electrocardiogram, has not improved reliability or improved neonatal outcome.⁸

Lack of benefit is supported by the fact that as electronic monitoring became widely adopted in the 1970s, rates of caesarean section rose about fivefold but the cerebral palsy rate was unchanged.^{9 10} Nor were rates of perinatal death, intrapartum stillbirth, neonatal death, low or very low Apgar scores, need for special neonatal care, or neonatal death lower with electronic monitoring.¹¹

Although evidence of lack of effectiveness for preventing cerebral palsy is consistent, the effect of electronic monitoring on risk of intrapartum death is less certain. Rates of perinatal death were already falling when electronic monitoring was introduced, and randomised trials showed no decrease in perinatal deaths.¹¹ However, most observational studies are compatible with the conclusion that electronic monitoring reduces the risk of intrapartum or neonatal death,¹² and it is less expensive than one-on-one auscultation.¹³ The obstetric societies of the United States, Canada, Australia, and New Zealand acknowledge that electronic monitoring provides no long term benefit for children.¹⁴

Why doesn’t it work?

Use of electronic fetal monitoring to prevent cerebral palsy was based on several erroneous assumptions. Firstly, fetal heart rate decelerations are commonly associated with peripheral chemoreflex rather than compression of the fetal head or umbilical cord, and may be relatively unthreatening.¹⁵ Furthermore, studies consistently show that most cases of cerebral palsy in babies born at or near term are not caused by birth asphyxia but associated with congenital malformations, fetal growth restriction, intrauterine exposure to infection or inflammation, and other unknown factors.^{11 16 17}

A test that identifies abnormalities in a high percentage of births will inevitably produce a high false positive rate if used to identify an uncommon outcome such as cerebral palsy, guaranteeing an “arithmetic of failure.”¹⁸ Tightening criteria to reduce false positives would further lower the test’s already low sensitivity.⁴

Reliability of interpretation of fetal heart rate tracings is also low. Experienced obstetricians had only mediocre agreement with one another in reading monitoring tracings, and when shown the same tracing months later agreed with their previous interpretations even less well.¹⁹ The United States Preventive Services Task Force rated effectiveness of electronic fetal monitoring as grade D, the lowest grade possible.¹²

Rise in caesarean births

All relevant randomised clinical trials have shown that use of electronic monitoring is associated with many more interventions in labour, surgical vaginal deliveries, and caesarean births.¹¹ A high proportion of caesarean deliveries are performed partly or wholly in response to non-reassuring electronic traces.^{20 21} In the United States each year, 900 000 women have a primary caesarean delivery, of whom at least 80% (720 000) attempt vaginal birth.²² Of those attempting vaginal birth, 86% (620 000) have electronic monitoring.²³

In randomised clinical trials comparing electronic monitoring with intermittent auscultation, electronic monitoring increased the risk of caesarean by 63% (relative risk=1.63, 95% confidence interval 1.29 to 2.07).¹¹ This suggests that, if electronic monitoring had not been used in those 620 000 women, 240 000 fewer of them would have had caesarean deliveries. Because 80% of women who have had a caesarean do not attempt vaginal delivery in a subsequent pregnancy, many more caesarean deliveries can be attributed to electronic monitoring.

Why is high caesarean delivery rate a problem?

Surgical delivery increases risks to mothers, immediate and long term.^{24 25} In a population based case-control study based on nationwide surveys in France, maternal death (adjusted for confounders) was 3.6 times more common after caesarean than vaginal delivery.²⁵ Operative deliveries increase risks of maternal haemorrhage, infection, and thromboembolism and of respiratory depression in the neonate. Complications in subsequent pregnancies include a high rate of repeat caesarean delivery, abnormally invasive placentation with potential for catastrophic haemorrhage, and uterine rupture.²⁶

It seems intuitively likely that electronic monitoring occasionally enables rescue of a threatened fetus, even if such events are too rare to register in randomised trials or cerebral palsy rates. But is this enough to justify the known harms from increased caesareans? Constantine and Saade, reviewing evidence on electronic fetal monitoring, concluded that “The evidence is overwhelming that continuous EFM ... has overall caused more harm than good.”¹²

Caesarean sections are among the most common surgical procedures in the United States. Higher medical care costs accompany a high section rate: an uncomplicated vaginal delivery typically costs $9000 to $17 000 (£7000-£14 000; €8500-€16 000), while an uncomplicated caesarean section ranges from about $14 000 to $25 000 or more.²⁷ Even a modest reduction in surgical deliveries would contribute to lowering healthcare costs.

Ethical problems arise when doctors use ready recourse to surgical intervention to protect themselves against litigation, thereby placing their own interests above those of their patients. And there is the injustice to medical care givers of using bad science in courtrooms to charge them with harms they did not cause.

Litigation

Failure to perform a timely caesarean delivery is a leading claim in obstetric malpractice litigation,^{28 29} although health workers are seldom sued for doing one. Obstetricians are sometimes told, “The only caesarean section you will regret is the one you didn’t perform.” ³⁰

In a survey of US obstetricians three out of four had had at least one professional liability claim, and most had had more than one.³¹ Many of the obstetricians polled had altered their practice in response, increasing the number of caesarean deliveries, decreasing their availability to high risk patients, or decreasing the number of deliveries performed.

Proper use of electronic fetal monitoring arises in most cerebral palsy claims³² despite the evidence that such monitoring is irrelevant to prevention. Plaintiffs’ experts can usually find something worrisome in any monitoring strip. Importantly, knowledge of neurological abnormality in a child with alleged birth injury influences interpretations of monitoring tracings and judgments about the appropriateness of the clinical care provided.^{18 33 34}

“Birth injury” lawsuits are among the most expensive of claims. Birth injury cases accounted for half of British NHS litigation costs in 2013, almost 20% of the total budget for maternity services.³⁵ Settlement of tens of millions of dollars or pounds are not rare, with most of the money going on lawyers, experts, and court costs.³⁶ The current medicolegal approach is expensive, irrational, and highly inefficient.^{20 36}

Achieving consensus

In an effort to reach a more evidence based approach to electronic fetal monitoring, cerebral palsy, and the rate of surgical births, we suggest a two pronged approach. Firstly, there is a strong need for consensus about the evidence. A major step in this direction would be a review by an impartial expert task force with focus on a narrow question: does electronic fetal monitoring in labour reduce risk of cerebral palsy? Such a task force would include evidence brought forward by all concerned parties and might be participated in and supported by national and professional healthcare organisations, lawyers, and consumer groups. Clear agreement among all relevant professional organisations would improve teaching and practice.

A second step is to tackle the problem of litigation because this influences practitioner behaviour. Testimony of expert witnesses often does not acknowledge best medical evidence, and an unwarranted assumption that surgical delivery in response to electronic monitoring can prevent cerebral palsy underlies many lawsuits for birth injury.^{32 36} Courts in the US, UK, and many other countries require that the testimony of expert witnesses is both reliable and relevant in order to be admissible. But it falls to trial judges, often scientifically unschooled, to assess whether the experts’ reasoning and methodology are compatible with good science and applicable. A consensus document reviewing the relevant medical evidence on electronic fetal monitoring would help to provide judges with contemporary evidence by which to decide whether evidence from the plaintiff’s expert was admissible. Such challenges would occur one at a time, but once one plaintiff expert was excluded that precedent would beget others.

With decades of evidence that electronic monitoring is ineffective in preventing cerebral palsy but contributes to a heightened rate of surgical births, that evidence should be made clear by relevant professional societies to medical practitioners and trainees, lawyers, and the public. Consistent with the guidelines of birth related professional societies,^{37 38} clinical practice should move towards use of intermittent auscultation in deliveries not thought to be at special risk. For such low risk births, practitioners should be asked to provide justification for use of electronic fetal monitoring.

Our current high rate of caesarean births is an important women’s health issue. Any serious discussion of reducing our high rate of surgical deliveries should include consideration of the role of electronic fetal monitoring and the evidence for its role in preventing birth asphyxia and subsequent cerebral palsy, and related litigation in maintaining the high rate of caesarean births.

Key messages.

• Electronic fetal monitoring during labour was introduced with an expectation that it would reduce cerebral palsy

• Randomised trials and time trends in population based data establish that it has not done so

• Electronic monitoring is linked with higher rates of caesarean deliveries, with their higher risks and costs

• Litigation relating fetal monitoring to cerebral palsy has also increased despite evidence that monitoring does not aid in its prevention

• Control of the high rate of caesarean sections and litigation requires better understanding of the evidence among all healthcare professionals and the public