Evolution of Fetal Heart Rate Monitoring Charles W. Fisher Kitch Drutchas Wagner Valitutti & Sherbrook PC One Woodward Avenue Detroit, MI 48226 (313) 765-7900 charles.fisher@kitch.com
Charles W. Fisher co-heads the firm s birth trauma defense practice, specializing in the defense of cases involving all aspects of pregnancy and early neonatal malpractice claims. Mr. Fisher, along with others in the birth trauma practice, currently serves as national panel counsel to nationwide institutions, including Ascension Health, AIG Insurance Company, and the CNA Insurance Company. Mr. Fisher is also a consultant on OB protocols for hospitals and health systems throughout the state of Michigan. As part of his practice, Mr. Fisher frequently lectures to physicians and hospital executives across the country on various birth trauma defense topics. He is licensed to practice law in Michigan and Ohio, as well as the U.S. District Court of Eastern Michigan. Mr. Fisher was named a Michigan Super Lawyer in 2007-2012. Only five percent of the attorneys in the state are named Super Lawyers and in 2010, he was recognized as one of the Top 100 Lawyers in the State of Michigan. Mr. Fisher is AV Peer Review Rated by LexisNexis/Martindale- Hubbell. The AV Peer Review rating is the highest rating allowed. It attests to a lawyer s legal ability and professional ethics, and reflects the confidential opinions of the Bar and the Judiciary. Mr. Fisher received his J.D. from the Detroit College of Law, where he graduated summa cum laude and first in his class. He also earned a B.A. and M.A. with honors in Music Education from the University of Michigan. Prior to attending law school, Mr. Fisher worked in the field of education, teaching music.
Evolution of Fetal Heart Rate Monitoring Table of Contents I. Original Concepts...6 II. 1997 NIH Consensus and Post NIH Recommendations...7 III. NIH in 2008...8 IV. Has Fetal Monitoring Made a Difference?...10 V. Final Summation...11 Evolution of Fetal Heart Rate Monitoring Fisher 3
Evolution of Fetal Heart Rate Monitoring Historical Overview In the late 1960s and early 1970s, fetal heart rate monitoring began being introduced to hospital wide systems. By the middle to late 1970s, most hospitals were using electronic fetal heart rate monitoring. The original goal of the fetal heart rate monitor was to detect hypoxia early, provide an opportunity for intervention, and prevent death or neurological injury. This ideal was an objective but in fact was not proven at that time through any scientific research. When one reviews early writings on electronic fetal monitoring, there were several assumptions made. The first assumption was that the vast majority of cerebral palsy occurred during labor and delivery from asphyxia or difficult deliveries. The second assumption was that the event would transpire over a period of time often referred to as partial prolonged hypoxia, and that there would be an opportunity for intervention. A third component that was assumed is that the interpretation of the fetal monitor would be uniform among doctors. A fourth factor assumed was that various patterns on the monitor strip would correlate well with fetal acidosis. Early on, these assumptions were critical to the concept that fetal monitoring was going to prevent death or neurological injury, including cerebral palsy. However, the above stated concepts had never actually been proven, or even tested. The majority of cerebral palsy is now known to occur before labor and delivery, and may produce abnormalities on the monitor strip suggesting that damage has already occurred. In addition, it appears as though instead of partial prolonged hypoxia being a cause of neurological injury in labor and delivery, acute near total events, such as cord prolapse, uterine rupture and abruption are more likely events to produce brain damage in labor and delivery when it does occur. Because of this, one questions the ability to diagnose such an event, make a decision that there isn t going to be recovery on the fetal monitor strip from the sentinel event, and order/perform a cesarean section before damage occurs. Thus, the concepts upon which electronic fetal monitoring s use were based actually had never been proven before the universal use of it became endemic to most hospitals. It is clear that a uniform method of interpretation, agreement and consensus that is, as to what patterns correlated with significant acidosis and required emergency cesarean section was not developed then and possibly remains undeveloped today (except in a few end spectrum patterns). One of the major advantages of using electronic fetal monitoring is that one to one nursing would become unnecessary, and in the sense that the hospitals could save money on nursing staff, it seemed a valuable investment. Likewise, because of the believed concept that it would help babies, it also, combined with the reduction in costs for personnel, was readily adopted by institutions. What wasn t expected, was the eventual plethora of lawsuits in which hospitals and doctors were spending extraordinary amounts of money on litigation costs, indemnification, and the rising cost of insurance. Clark, et al, published an article wherein he pointed out that elective cesarean section on all patients would be less costly than submitting patients to labor and delivery and having a few brain damaged baby lawsuits result from deliveries at the hospital. In the meantime, as the cesarean section rate went from about 3 to 4 percent in the United States to about 30 to 40 percent at some institutions after electronic monitoring was introduced, it was noted that in Dublin, Ireland, the c-section rate stayed at 3 to 4 percent. Outcomes of babies remained identical between the two countries, which began to be a marker of such distinction that the value of electronic monitoring was questioned by a number of various authors. From that point until today, literally hundreds of articles have Evolution of Fetal Heart Rate Monitoring Fisher 5
been written on the methods of interpretation, indications for c-section, results of such interventions, and the overall rate of cerebral palsy in the United State as correlated to electronic monitoring. Were it not for the intersection of medicine and lawsuits at great monetary exposure, the study of the evolution of fetal monitoring would be one of the most intriguing historical events in medicine. It leaves one questioning, as did Clark later on, why a device that appears to be about 99 percent wrong is even being used in medicine? (See Clark s article on Demographic Trends cited in the slide presentation). Below are three separate time frame explorations. They involve the original evolution and science behind fetal monitoring, matters that occurred after the 1997 NIH conference on fetal monitoring, and current information and recommendations that appeared in 2008 from the second NIH conference. Finally, there will be a brief discussion on whether electronic fetal monitoring is efficacious. There may be some rare instances in which the electronic fetal monitor actually has prevented a poor outcome. However, the fact that cerebral palsy has not been reduced at all, suggests that the rare and exceptional instance where a monitor has saved a baby might be so small and sparse, that it would be difficult if not impossible to say that there is scientific evidence that electronic fetal monitoring during labor and delivery prevents cerebral palsy I. Original Concepts Edward Hon, in 1969, wrote the Atlas of Fetal Monitoring. In his monograph, he made it clear that there are essentially two patterns that will occur during labor when there is developing hypoxia. Those two patterns that he listed were either prolonged variable decelerations or late decelerations. This is not difficult to understand, because labor presents a unique event that involves periodic stress on the baby as a test of the intrauterine environment. There are two basic reasons why you could develop hypoxia during labor, one being a cord accident, the other being a placental insufficiency problem. In both those situations, the first sign you would expect to see is a deceleration pattern, producing either variable decelerations of a significant or severe type, or late decelerations. Although it is common in lawsuits to claim that other abnormalities (decreased variability, tachycardia, fewer accelerations) are evidences of hypoxia, this does not fit into any of the original scientific investigator s concepts. From Hon, to Parer, to Schifrin, to Paul, to Freeman, and contained in a number of textbooks over the years, the most crucial element that you must have before any hypoxia can be declared, is a pattern of decelerations (or simply an outright sudden bradycardia). In fact, the electronic fetal monitor literature, as it originally was presented, contained no evidence that patterns in labor other than decelerations should be interpreted as hypoxia. Over a period of years, another concept was explored in detail and that was the significance of variability. Decelerations with progressive loss of variability would be, and is accepted as one of the first indications that the baby is starting to develop metabolic acidosis. It was found by Dr. Hon, and agreed to by other investigators, that it is not unusual for many babies to normally have some degree of hypoxia during labor and delivery because it is stressful. This stress was exhibited by decelerations, but did not necessarily represent any significant metabolic acidosis. Furthermore, the ph actually normally drops during labor and delivery. Thus, there is some mild hypoxia and some mild metabolic acidosis that occurs in most deliveries. It is for this reason that researchers began looking very carefully at the interpretation of decelerations in the context of baseline variability. Many, many cesarean sections were being done yet there was no improvement in overall outcome. In other words, doctors were regularly doing cesarean sections simply on the basis of decelerations without any acidosis. Freeman s text indicated that the most common reason for unnecessary cesarean sections was vari- 6 Medical Liability and Health Care Law March 2013
able decelerations as the vast majority of those decelerations were of a benign nature, not reflecting any significant hypoxia or acidosis. Eventually, and as followed up by Acog Bulletins from 1977 onward, the concept became stronger and stronger about the requirement of decelerations to prove hypoxia, and the importance of variability in evaluating a deceleration pattern. In Acog Bulletin #207, it was specifically indicated that decreased variability without decelerations is not likely to be due to hypoxia, thus, ensuring and endorsing the concept over the years that had developed the first sign of possible hypoxia is the development of a deceleration pattern, however, as long as the variability of the strip was present you had reassurance that this was not acidosis. Importantly, physicians thought that this would decrease the incidence of cesarean section, and more importantly, provide real evidence that cesarean section was making a difference in specific cases where decelerations with absent variability existed. However, the continued investigation of causation, comparing rates of cerebral palsy to fetal monitoring and crash c-section, repeatedly proved that the rate was not falling and fetal monitoring was not producing what they expected. Because of the many perceived issues in fetal monitor interpretation and efficacy, including lack of uniformity in definitions or agreement as to management, and the obvious concern about the high cesarean section rate without evidence of improvement, there was a consensus meeting in 1997 at the NIH to try to find some solutions. II. 1997 NIH Consensus and Post NIH Recommendations In 1997, although it was intended to be a consensus meeting, there actually was very little consensus. The gurus of fetal monitoring all agreed on what was a completely normal strip. Secondly, they agreed that when you have a deceleration pattern with completely absent variability, this is ominous and delivery needs to be implemented fairly expeditiously. However, the major gap in agreement was virtually everything that existed outside of these two categories, and this included the vast majority of monitor strips, many of which had abnormalities but which were not specific enough for the physicians to reach a consensus. Thus, they indicated that those patterns and their management were controversial and that more investigation was needed. These are the patterns however that many lawsuits have been based on. The consensus group set forth a number of definitions hoping that there would be more uniform interpretation for further study. If there was uniform identification, the thought was that they could then determine efficacy. Post 1997 NIH consensus, Acog Bulletin #70 2005 was produced. What is intriguing in that bulletin is even after the definitions in 1997, panels of experts were given the same monitor strip and still disagreed on interpretation. In only 22 percent was there agreement as to the interpretation. Two months later, the same physicians looked at the same strips, and 21 percent read the fetal heart rate pattern differently from their original reading. They also found that knowing the outcome definitely affected the reading of the strip, and they stated that the reading therefore was not reliable. By 2005, a number of studies cited in this bulletin revealed that the incidence of cerebral palsy still had not been reduced. In the meantime, the often used term of fetal distress was abandoned for what probably is a more nebulous concept, non-reassuring. Non-reassuring was not well defined as a consensus, and depending upon which author you read, the definitions differed, or there were no real definitions. This provided fertile ground, of course, for lawsuits in which plaintiffs experts would find something non-reassuring. It was an easy retrospective argument and continues to be an easy argument that an interpretation was misread because the baby ends up with brain damage. The lack of a unified and agreed upon definition Evolution of Fetal Heart Rate Monitoring Fisher 7
for non-reassuring was a major omission especially in the context of the NIH publication and consensus meeting in 1997. The NIH made it very clear there was absolutely no agreement on patterns between the normal and completely ominous patterns, however, one could regularly categorize many of the in between patterns as non-reassuring which was often done in medical records as well. This was used to great advantage by plaintiffs in lawsuits. In other words, it doesn t sound very logical to define something as non-reassuring and then fail to perform a c-section. There are quite a number of variant patterns that are non-reassuring (i.e. mild variable decelerations) which are in-between patterns but require no cesarean section. In one effort to help clarify the patterns on monitor strips, fetal pulse oximetry was introduced. What was then proven is that in some of the most ominous looking patterns visually, the baby was not acidotic or hypoxic at all, and the fetal pulse oximetry proved that this was true. Unfortunately, physicians still became concerned when seeing some of these patterns, and did c-sections anyway. This eventually resulted in an abandonment of the fetal pulse oximetry because the c-section rate was not decreasing. In reality however, this was unfortunate in the medical/legal sense. If you did continue labor with a fetal pulse ox on, you had unequivocal evidence that hypoxia was not present even with decelerations. It would have provided a significant piece of evidence in the lawsuit that it was reasonable to continue labor in the face of abnormalities that might be called non-reassuring on a monitor stip. In 2008, the NIH held its second conference on fetal heart rate monitoring to try and reach a consensus again. In addition, there was another attempt to define and identify patterns specifically for more accurate study group comparisons. This NIH meeting was followed by another ACOG bulletin in 2009, Number 106 which is one of the current bulletins in effect. III. NIH in 2008 The NIH meeting in 2008 involved new terminology and redefined Categories of fetal monitoring patterns. By 2008, a number of additional articles had appeared regarding electronic monitoring and lack of efficacy. In fact, one of the articles was specifically entitled, Is Fetal Monitoring Salvageable? This of course was an interesting commentary on the fact that the original concepts of what fetal monitoring were supposed to be doing did not seem to be evidencing themselves, as the cerebral palsy rate continued to be unchanged and in fact was on the rise in both preterm and term infants. In 2008, the NIH defined three categories. A close look at the three categories in 2008 reveals that they directly resemble the same thing they found in 1997. The categories as define, I, II, III related to the normal strip (I), the completely ominous strip (III), and the indeterminate Category II strip which contained everything else. The statements in the NIH indicated that with Category II, no correlation had been shown yet with fetal acidosis or poor outcomes and therefore you could continue to watch and monitor patterns in Category II. It was appropriate at this time that the words non-reassuring were left out of the NIH and the eventual ACOG bulletin, and instead are more definitive identifications of patterns that fit into Category I, II and III were provided. Irrespective of the medical propriety of doing this, the medical/legal consequences was that for the first time, defendants in a lawsuit could identify and slot the abnormal patterns into either Category II or Category III. If it was Category II it was a good defense that cesarean section was not required at that point. In 2009, ACOG published the bulletin 106. The authors of the ACOG bulletin interestingly added something different than the NIH 2008 paper had stated. In ACOG 106, the management protocol for Category II was included as though it was Category III. In other words, the bulletin itself includes management of Category II and Category III together making no distinction between the two. Therefore, the management 8 Medical Liability and Health Care Law March 2013
distinctions made in the NIH 2008 were being changed somewhat by the ACOG bulletin in 2009. There is no explanation for this in the bulletin. Following up quickly after 2009 s Bulletin, ACOG published another bulletin on management of fetal heart rate patterns just a year later in 2010 bulletin 116. In this author s humble opinion, there were different things that it contained in the management section that heretofore had never been included or incorporated. For instance, tachycardia with decreased variability was clearly a Category II pattern as previously defined. However, the bulletin went on to indicate that with intrauterine resuscitation, if you could not improve the pattern, you must consider delivery (easily interpreted by Plaintiffs in a medical legal context as requiring a stat cesarean section). Thus, in 2010, one begins to question whether or not the recommendations as they exist in the 2010 bulletin are based upon scientific principles. Recall that the original developers of the electronic fetal monitoring clearly indicated, and repeatedly indicated, that the first sign of hypoxia in labor and delivery is a pattern of decelerations (not tachycardia, not decreased variability, and not absent accelerations). In fact, prior authors including Gabbe s textbook (Thomas Garite s chapter) made it very clear that those non deceleration patterns should not be considered signs of hypoxia in labor and delivery because they were not preceded by decelerations. In 2010, Acog Bulletin #116, the management protocol that could lead you to a possible emergency cesarean section does not even require a pattern of decelerations (you could just have decreased variability that doesn t improve, tachycardia with decreased variability that doesn t improve, or even no accelerations with decreased variability without improvement, and no decelerations at all in any of these situations). Thus, in 2010, one asks the question, where is the science to back up these recommendations that aren t just different than previous recommendations but almost the opposite? Not only has there been an absence of medical/scientific support for defining these non-decelerations patterns as hypoxia and acidosis in labor, perhaps even more importantly, there isn t a single study whatsoever that shows intervening with a nondeceleration pattern produces a different outcome. C-section is not without risk. Note that a mother whose baby has a tachycardia has maternal risk, including possible spread of infection if the tachycardia is caused by chorioamnionitis. Without any proven evidence that these patterns are hypoxic or acidotic, or causal on an acute preventable basis, Acog 116 seems to indicate that a c-section be done. Further changes developing over the years since 2008 to the present have been remarkable in terms of different authors recommending different ways to interpret monitor strips. We have authors recommending a whole different category system, or multiple other categories. We have authors recommending a colorcoding system for categories that is actually quite complex. We have authors that have suggested that the most important finding on electronic monitor strip is not the persistence of decelerations, but the squared area of the deceleration. We have authors that have looked at adjunct devices such as a fetal EKG to try and obtain better results. We have authors that have now basically suggested that with Category II patterns you have to call the surgical crew to be ready just in case, when in fact that would mean probably calling the surgical crew on about 90 percent of all cases in labor. Finally what is quite unique is the recent article by Shawn Blackwell who may have identified the most problematic issue of all. That is, he pointed out that there was very poor agreement, or inadequate agreement between obstetricians as to whether or not variability was absent, or still present, although diminished. This of course is the most crucial distinguishing factor between Category II and Category III and the distinguishing factor between the lack of metabolic acidosis and the presence of a dangerous acidosis. Such a study Evolution of Fetal Heart Rate Monitoring Fisher 9
is condemning of the ability of intra-observers to actually detect the most crucial point of the pattern that would distinguish continuing labor versus cesarean section. Why so many different authors suggesting so many different changes and ideas? There can only be two reasons. Number one, the interpretation of the strip is so subjective, that it is given to different interpretations except in the most extreme cases at either end. Thus it makes it extraordinarily unfair to any obstetrician to be criticized for interpretation on patterns that are considered the in between or indeterminate patterns. There simply is no real consensus on the interpretation of the vast majority of FHT patterns. Secondarily, it raises the suspicion that electronic fetal monitoring is cannot prevent poor neurological outcomes, either because they are already present before labor, or because the event is so acute that allowing time for interpretation of the strip, ordering a cesarean section and performing it, is not going to prevent the acute injury. It would seem that one or both of these concepts are true, and impact the ability of the health care provider to make a difference. As a final thought, it must be remembered that decelerations in labor are common, variable decelerations occur probably more than 80 percent of the time. The concept of the boy who cried wolf, certainly enters the thinking process here and the realities of fetal monitoring. Health care provides see abnormal patterns all the time with good outcomes! The concept that is completely abandoned in trial when you are simply focusing on one baby, is the near universal fact that vast majority of labors have decels, even significant ones, and yet outcomes are rarely abnormal. It was pointed out by Karen Nelson, even in those cases where there are late decelerations, about 98 percent of the babies turn out completely normal. If you are watching a test that 98 percent of the time is completely wrong, you become somewhat numbed to these patterns which are particularly frequent in a huge percentage of the laboring patients. Asking the clinician to distinguish the one or two percent (per Nelson) of those babies that would end up with a poor outcome on the basis of a pattern that is so often wrong is unfair to any clinician. IV. Has Fetal Monitoring Made a Difference? There have been many efforts over the years to explain away the reason why the cesarean section rate has not produced a reduction in cerebral palsy. There are regular theories posited by plaintiff s experts when they are confronted with this issue, all of which have been basically discredited. Often we heard that because more premature babies were being saved, the rate of cerebral palsy would not go down even though we are making a difference in the term infants. This has been disproven and particularly so by the CDC that divided up the rate of cerebral palsy in term and preterm babies showing that in term infants, the rate is continually gone up over the past 10 years. Another regular theory so posited by plaintiff s experts was that the number of babies that actually get brain damage in labor from hypoxia is small that it is impossible to do a study. Of course, if that is true, where is any scientific support that you are making a difference if you can t even study it? remember that Daubert is the plaintiff s burden of proof. However, this theory was also disproven by the Clark and Hankins article in the American Journal of Obstetrics and Gynecology in 2003. In an extraordinary article using statistical analysis they proved that this plaintiff theory was incorrect, and that you can detect differences that are this small. (a very complex article but very detailed). One of the points made in the Clark and Hankins article was that thirdworld countries don t have any worse cerebral palsy rates than we do, and they don t even do cesarean sections. A number of articles have addressed this whole issue of causal effect leading to a meta-analysis in the Cochran Database (updated in 2007) stating that there was no difference in the death rate or the cerebral 10 Medical Liability and Health Care Law March 2013
palsy rate with electronic monitoring and emergency c-section. The only difference that they could find with electronic monitoring and a cesarean section was reduced seizures in the nursery. Interestingly, the Cochran Database suggested that before doing a cesarean section, you should obtain informed consent which would include the fact that even if they do an emergency cesarean section, science has not shown that it prevents any injury. This is quite a reverse concept from what was previously suggested on causal prevention when fetal monitoring was first introduced. We also often hear the construct that there s a thirty minute standard of care for doing crash cesarean sections from decision to incision, suggesting that this would be preventative of insult. However, no one really knows how this time frame actually came about other than possibly from Dr. Hon s original publications (or others) that suggested he felt babies could withstand fetal hypoxia for about 30 minutes. Thus the concept developed that if you do the cesarean section in 30 minutes or less, you will prevent damage - another claim that has never been scientifically proven. Intriguingly, articles published by Murphy, et al and Ball, et al in February and August of 2007 showed that expediting delivery with operative vaginal (vacuum or forceps) in about 15 minutes produced no difference in outcomes as compared to cesarean sections done in about 30 minutes. Ball s data also compared operative vaginal delivery alone (vacuum/forceps) with failed operative delivery and cesarean section, as compared to just going to a cesarean section. There were no differences in outcomes. Therefore, we now have the scientific situation where we have not shown any proven evidence that we are reducing the incidence of cerebral palsy, and in comparing the deliveries done in 15 minutes to those done at 30 or 40 minutes, there is no difference in outcome. This seems to support the belief that the outcomes for whatever reasons are not being affected by emergent physician intervention. Does an acute injurious sentinel event occurring in labor happen so fast that we really can t diagnose it and do a c-section fast enough to prevent harm? In medicine, management should be based upon proven scientific methods and knowledge. Likewise in the court of law, the same equivalent standard should apply. From a purely statistical standpoint, there have been hundreds and hundreds of lawsuits that have been asserted and collected on based on a claim that the electronic monitor showed fetal hypoxia that required a crash cesarean section. How interesting it is to then have a plaintiff s expert proclaim that there aren t enough events of HIE during labor and delivery to even study the causation issue, while they go on to testify in their hundredth lawsuit that a c-section could have prevented injury. V. Final Summation The question when do I do a cesarean section? has really not been carefully answered in medicine and it is remains extraordinarily judgmental. Basing those decisions on a monitor strip that even now produces multiple articles on how to interpret it and different ways to interpret it, has to be discouraging to obstetricians. The current ACOG 116, has seemingly gone beyond the original medical science. Providing a nebulous exculpatory phrase that you may consider cesarean section in Acog 116, will mean (in a lawsuit) that you must consider and do a cesarean section. Concepts originating from Dr. Hon s 1969 Introduction to Fetal Monitoring have seemingly taken a backseat to extraordinarily complex and confusing regimens, codes, colors etc. In reality, most nurses and doctors will have difficulty following some of the current suggestions, especially in light of the fact that Category II strips pre-dominate labors with a variety of different abnormalities. In watching the evolution of the science of electronic fetal monitoring since its introduction in the late 60 s and 70 s, it is clear that it is not working in the way that was expected. It is and was self-evident that Evolution of Fetal Heart Rate Monitoring Fisher 11
there was a major problem with the interpretation of the fetal monitor given the incredible number of articles that have been written to address this. Will there ever be a uniform accepted method of interpretation? Legally, causation issues should be a part of the defense in all of these cases, including engaging Daubert challenges on these very issues. These issues must be part of the defense s preparation of their own fact witnesses as well. Defendants should not give away causation by ill prepared testimony. Medically, causation should be carefully studied, not presumed. Acog needs to take a significant look at this issue as being important as, or even more important than creating another interpretation and management bulletin. 12 Medical Liability and Health Care Law March 2013