Fetal Heart Rate Test

Antepartum Fetal and Heart Rate Testing: Current Status Michael Emerson, M.D.

1

I.

Fetal movement (FM) Counting A.

B.

C.

D.

Physiologic background 1.

Healthy fetuses move 20-30 times per hour

2.

Fetal movement occupies 10-15% of 24h day

3.

Active fetus has high likelihood of good result

4.

Feud compromise may follow decreased FMs

Basis for maternal perception of FM 1.

Inexpensive, simple No equipment needed

3.

May be done in home, office, hospital

4.

Patient is engaged in her care process

Application 1.

Each fetus acts as own control

2.

Baseline record of activity can be established

3.

Clinical alerts: decreased or absent Fms

4.

Note: same conditions for each session

Follow up 1.

Supplement other forms of testing

2.

Decreased FM 6 more intensive testing

3.

NOTE: inform patient that absolute FM count may vary

considerably between sessions E.

II.

Limitations 1.

Limited numbers of good clinical trials

2.

Low sensitivity to prediction of acute distress

3.

Normal “slow” or “hyperactive” fetus?

4.

Variation in ability to educate the patient and compliance

Principles: Physiologic Bases A.

Fetal heart rate testing: applications 1.

NST: office/hospital possibly home. 20-30 min to t-2 h

2.

VAS similar to NST. 10-20 min

3.

ACTG: similar to NST CST: office/hospital. 20-30 min to 2-3 h

B.

Nonstress test (NST) 1.

What does the NST test? a)

Selected FHR baseline features (1) Accelerations with FMs (2) Baseline rate and variability (3) Decelerations: spontaneous

b)

Physiologic

c)

Pathophysiologic

d)

Brainstem function

e)

Hypoxia./acidosis

f)

ANS/reflex control

g)

Malnutrition

h)

Maturation of FHR

i)

Cord compression

The issue here of what I am going to talk to you about in the ensuing talks and one of the things that I think we are beginning to recognize here is that the modern era of fetal assessment now is well into its third decade. We have been the beneficiaries of a lot of technology and, I guess, as we sit back now and look through the retrospectoscope, what we really see is that a lot of the technology arrived before the basic work was done to tell us exactly what the best way was to apply this or not to apply it to clinical practice. So what I would like to do is basically espouse what I call honest principles. I don't sell cars but if I did, I would stand there as honest Dr. Devoe and tell you about four testing modalities this morning. We will cover the issue of fetal movement counting and heart rate testing in the first half and then in the second half we will talk about comprehensive biophysical testing and discuss Doppler. I would submit to you that perhaps one of the things we should really consider in the background of this whole process is using the observation of ultrasound growth as the context for really what is happening through the bulk of pregnancy. I am not going to discuss that because it hasn't really been formalized as a test because this is something that we do periodically. I think we all recognize now with managed care that the issue of doing multiple ultrasounds has to be regarded by some of us as a charitable exercise because maybe we will get paid for one or for two of them. But if you do anymore than that, they are really contributing to patient care and they are not contributing to your reimbursement. Nonetheless, the ability to accurately chart fetal growth and to use it as a context for gauging assessment strategies is extremely important and that should underlie any assessment plan for a fetus judged to be at high risk for compromise. Let's talk about fetal movement counting as our initial foray into assessment because this is something that is very easy to do. We know a lot about the background physiologically of fetal activity. We know that fetuses that are healthy move quite frequently, an average of 20 to 30 times an hour near term. We also know that if you took a 24 hour day and just counted the amount of time the fetus spends in actual movement, that is going to be a good couple of hours out of that day in a cumulative sort of way. We also know from the great body of data that have been collected that activity is tantamount to the very high likelihood of good outcome and that diminished activity may precede a compromised fetus. This is a study that is about 20 years old. I like to show it because what it illustrates, really, are a couple of points. First, these are maternal perceptions of fetal activity done over the last half of pregnancy for 12 hours. So imagine if you will, a context of having a pregnant woman sit around for 12 hours, not do anything else except track fetal activity. These are normal babies and you can see that there is a very wide range of normal. These are hundreds of movements that are being charted here and that apparently the fetal activity, by maternal perception, diminishes towards term. This is a reproduced finding. In fact, what happens towards term is not that the fetus stops moving or even decreases moving. It is just that maternal perception of movement is blunted by other things going on such as uterine activity. Mothers often have more difficulty sitting around and counting movements and they are more easily distracted. So in fact, if you were to do this by more objective ways such as ultrasound observation, this doesn't happen at all. But nonetheless, if you use maternal perception of movement, you have to be aware that that is what the mother reports and this is a real phenomenon. The reason, physiologically, that movement might be a good window into fetal condition is illustrated by this slide which is an animal experiment. What it shows here is that when you make a fetus hypoxic, long before this fetus becomes acidotic or asphyxiated, the hypoxic fetus will reach a critical threshold of oxygenation and then fetal movement will abruptly cease. You can see that it doesn't take very long when you get to a fetal pO2 of about 15 torr for fetal activity to virtually come to a halt. So that if you take a look at fetal movement assessment by counting, that the baby that is not moving or has been active and suddenly has stopped may not be a baby that has died but may be a baby that has become hypoxic.

2

j)

Circadian rhythms

k)

Placental insufficiency

l)

Behavioral state

m) CV and CNS anomalies 2.

Physiologic basis for NST a)

FM is normal, episodic phenomenon

b) Third trimester fetuses respond to FM with coupled accelerations (>90%) c)

Hypoxia, asphyxia, malnutrition reduce FMs. decrease coupling fewer accelerations

3.

NST interpretation a)

Reactive: accelerations + FMs

b)

Nonreactive (1) Accelerations present, too few

This is an acute observation and I would submit that one of the things that we will talk about with some of our other testing is that you very rarely ever see this scenario occur in actual practice. You think of the number of times that babies die suddenly or have an acute catastrophe. That is extraordinarily rare. So if you take a look at movement, and we will take a look here, the appeal of fetal movement counting is the fact that it is easy to do, the mother can do it, it can be done anywhere. What I like about it is it keeps the fetus on the screen even when the mother is not in your office. There are some other issues here and I think this is a very important point because there are quite wide variations in how much is normal for a given fetus and you really have to sort of set standards. That standard is set for each fetus as his or her own control and then the mother collects a baseline of recording and you set arbitrary thresholds for clinical alerts. I try to be more generic. Rather than have the mother take out a calculator and say, "Is this 50% less movement than I had?" If she is counting and she can see that the gradual trend in numbers is starting to go down, that is a signal for her to come in and have some additional testing done. If there is a session in which fetal movement counting is not present at all and she has extended that observation for up to an hour, that is a signal to come in as well. The counting perception maternal exercise should be done under the same conditions. Usually a good time to do it is after the mother has eaten and she is reclining or stretched out and is sort of taking it easy.

(2) Accelerations absent. FMs present (3) Accelerations, FMs absent 4.

5.

6.

Nonreactivity sequence a)

Decreased acceleration counts, amplitude

b)

Decreased FM counts

c)

Uncoupling accelerations and FMs

d)

No accelerations or FMs

e)

Spontaneous decelerations

Causes of nonreactive NST a)

Compromised fetus

b)

Behavioral state

c)

Immaturity

d)

Maternal diet/drugs

e)

Fetal anomalies

Testing conditions a)

Length of observation: 30-60 minutes needed for 1

acceleration in 95% of normal fetuses b)

Devoe, McKinzie, et al. Am JOb Gyn 1985 (1) Reactivity in 95% within 70 minutes (2) Nonreactivity (>90 minutes) and abnormal CST (95 %)

c)

Corollary: prolonged NR NST at term - should prompt

consideration of delivery d)

Effects of immaturity on FHR baseline (1) Lower amplitude accelerations (2) More frequent decelerations (3) Less coupling

The issue is, really, that this is a supplemental form of testing. In essence, decreased fetal activity is not in itself a diagnostic observation but rather should bring a patient into a closer observation scheme. I pointed out, and that is why I showed you this 12 hour count slide to begin with, that there is a lot of variations in this. We see this in the testing center all the time. We see fetuses that are more or less active. The same fetus, just done in different sessions. So trying to be absolutely dogmatic and absolute about using fetal movement counting, I think, perhaps is a little bit unjustified. I am going to show you a number of slides that look like this and I want to explain to you what this format is so that you understand what I am driving at with each of these presentations. If we take a look here, we are going to look at a number of RCTs or randomized control trials. The reason I like to do this is that when you talk about the evidence for using a modality for fetal assessment, when you really talk about that, you are talking about it as compared to what? Usually, when you are talking about a test, you are either comparing it to another test or you are comparing it to nothing. In this case, we are looking at trials that compared maternal counting of fetal movements with nothing. There are a couple of large trials that have been reported. These trials have been overseas. They have engaged thousands of patients and what is interesting here, these are done in routine populations. So one trial had approximately 65,000 enrollees. That happens to be, to my knowledge, the largest trial of any modality ever done in any of the testing schemes that we do. Sixty five thousand enrollees. What we see in this slide is the outcome of interest - the number of trials in which this is reported and the odds ratio. Let me briefly explain to you what an odds ratio is. It is simply a statement saying that the relative likeliness of this outcome occurring in one population as opposed to another population. It is a ratio of likelihood of outcome. When there is no difference between the two populations, the odds ratio is 1. If the outcome is less likely or diminished, then the odds ratio is less than 1 and if it is more likely, it is greater than 1. Just stating an odds ratio would be insufficient because what we look at here are number of observations and what we want to be able to do is to generalize these observations to other populations. One can compute a range of confidence for how likely this is going to be a truly significant finding by doing a 95% confidence interval. If this 95% confidence interval, which is this last column here, includes unity, then it is unlikely that this is a significant difference between the two groups. So that is how to interpret these tables that I am going to show you over the next couple of sessions.

(4) Standard reactivity takes longer to obtain e)

Pathologic tracing as significant as in term

If you take a look here, for example, this is fetal movement counting. Some

3

(1) Prolonged NR (2) Late decelerations (3) Absent variation (4) Severe variables f)

Maternal drugs (1) $-blockers, CNS depressants (2) Nicotine, cocaine

g)

Maternal diet (1) Hypoglycemia: decreased FM only if profound 2)

7.

Hyperglycemia: more FBM, less FM,

Sequential nonstress tests a)

Concept of using each fetus as its own control

b)

Devoe, et al. Am JOb Gyn. 1986 (1) 16 of 18 fetuses with eventual compromise showed abnormal trends in NSTs before tests fell below threshold for normalcy 2)

C.

Recommend testing under same conditions

Contraction stress test (CST), oxytocin challenge test (OCT) 1.

Basis a)

Response of FHR baseline to reduced or spontane-

ous uterine contractions (UCs) b)

FHR-UC association

c)

Influences (1) Contraction frequency, intensity (2) Maternal buffering capacity, position (3) Fetal oxygenation, acid-base levels

d) 2.

Sequence of events leading to positive test

Significance a)

CST reflection of 02, acid-base balance, placental reserve

b)

Positive CST may reflect (1) Fetal compromise (2) Maternal hypoxia, hypotension (3) Uterine hyperstimulation (4) Umbilical cord vulnerability

form of obstetric intervention was more likely in the group in which fetal movement counting was routinely done. You will see here that 1 is not included in this. It exceeds 1. You can see, however, that the stillbirth rate in the two populations, the tested and the untested was the same. The confidence interval includes 1 and goes from 0.7 to 1.35. Reporting decreased movement is much more frequent in the tested population because that is what was being looked at to begin with. There are more NSTs and more ultrasounds. So, what is the take from this? The take from this, in these two large populations, this one was 64,400 and this other study was 3,100, the take home is that routine counting in general OB populations led to more testing, more resource utilization, more interventions and, in the larger study, didn't appear to reduce the rate of stillbirth which is really one of the targets of antepartal testing to begin with. I think that as you take home this, you say, "Well, perhaps we should pay attention to this." In the studies that have focused on fetal movement counting in high risk pregnancies, I have not included here because they are small, their designs leave a lot to be desired. I can tell you that the small, relatively less robust, studies that have been done suggest that maternal perception of fetal movement counting is useful in a high risk pregnancy. But I say that with a caveat. It hasn't been studied as adequately as these trials have done here. I would also say that use of movement counting has certain limitations. It is not the best predictor of acute distress. There are quite wide ranges of activity. There are fetuses that are "slow" or less active that are perfectly normal and there are fetuses that are so active that it keeps the mother up 24 hours a day. We don't know whether or not this is associated with the so-called attention deficit disorders or hyperactivity in childhood because the followups haven't been done. But there are also patients who simply can't do this. We have patients in our population and you may even have it in yours, who simply can't do this. You give them a movement chart and it comes back with two movements on it. So there are patients for whom this is not a valid method because they won't comply. Let's turn to the issue of fetal heart rate testing because that is where we are going to be spending the bulk of this first talk on. As time has evolved, there have been more and more variations of fetal heart rate testing added to our armenatarium. I am going to talk to you about four of them but three of them are really variations on the same theme and that theme is the baseline observation of fetal heart rate. The outlier there is the contraction stress test or CST which involves the administration of either oxytocin or breast stimulation to produce uterine activity. But the non-stress test, the vibroacoustic stimulation test and the ACTG, which is the actocardiotocograph, are all tests that look at resting baseline heart rate. If we take a look here, again developmentally, we know that the centers that control the expression of fetal heart rate regulation are the last ones to mature during the development of the fetus. We also know from good evidence now that when the fetus becomes hypoxemic, the first physiologic marker that we can look at in the antepartal period to be affected is the reactivity of the fetal heart rate. Experimentally, some disturbing data that come from animal studies, but nonetheless under controlled circumstances, is that you can have a fetus that undergoes a significant hypoxic event with acidosis and with focal or generalized brain damage in which if the fetus survives, the heart rate may recover and appear to be normal. We occasionally, but fortunately, very rarely see this. In addition, if we take a look at a model that Bob Danyon developed in Canada and look at a chronic model of embolization of the fetal circulation over a period of time, the first thing that we lose in baseline fetal heart rate is reactivity. The very last parameter to go is heart rate regulation or rate regularity. That sequence is an important one to keep in mind as you are following a baby serially over time. If we go ahead and look at really what the NST tests do, and I think this is

4

III.

Interpretive Criteria

Table 2 CST Interpretive Criteria

Result

Description

Negative

No late deceleration(s) present on tracing with uterine activity that is adequate

Positive

Late decelerations present with most: (>½) of the UCs (unless hypertension

an important thing to keep in mind, it is not what you do during the test, it is what you are actually looking at. Physiologically, the test really deals with issues that lie at the level of the brainstem - the autonomic nervous system. Some maturational events which really color the baseline rate and the appearance of the baseline. There is a certain amount of circadian influence so this makes it important about the timing during the day when you do the test. Likewise, this is coupled with fetal behavior as the third trimester rolls on and pathophysiologically what we are asking of this test to tell us is, "Is the fetus hypoxic, acidotic? Is it malnourished?" Occasionally, we will luck out and we will see fetuses that have unexpected cord compression and placental insufficiency. Serendipitously, we may occasionally pick up the baby that has a significant anomaly because the test has a prolonged nonreactive appearance but that is really serendipitous and that is not really the rationale for doing the test. Likewise, we may pick up the odd fetal arrhythmia that requires further exploration with ultrasound.

present), even if uterine activity is less than adequate Suspicious

Adequate uterine activity present with some late deceleration(s), but does not meet criteria for a positive test

Hyperstimu

Late deceleration(s) present with or following exces-

lation

sive uterine activity

Unsatisfac-

Quality of tracing inadequate for accurate interpreta-

tory

tion, or adequate uterine activity cannot be achieved

IV.

Vibroacoustic Stimulation A.

B.

Physiologic basis 1.

Signal = broad-band (20-10K)

2.

SPL = 82 db in air, 110 db in water

EAL provides two components 1.

Vibratory (+)

2.

Acoustic (-)

C.

Shortens testing time

D.

Predictive accuracy is similar to standard NST

E

May be useful intrapartum

F.

No apparent adverse side-effects

G. Normal fetal responses to VAS signal 1.

95% will become reactive post-VAS

2.

85 5 will have increased FHR baseline 10 bpm,

>180 seconds

3.

Mean onset:

7.5 sec

4.

Mean duration:

600 sec

5.

Modal duration:

300 sec

6.

Median duration:

360 sec

H.

Post-VAS responses reflect change in state

I.

Post-VAS responses are not equivalent to spontaneous stimuli

J.

Failed VAS does not exclude normal outcome

K.

Safety remains to be firmly established

L.

Efficacy trials are limited

If we take a look at the selected baseline features of which we are interested in, acceleration of movements is the hallmark of the NST. We know that variation in rate has some importance but it is very difficult with an eyeball to quantify variation. You can do qualitative assessment of variation. You can call it minimal or decreased or absent or average or increased but again the reproducibility of doing it visually is a problem and decelerations occasionally do occur spontaneously and disturb us. If we take a look here, physiologically we have a very good basis for using this test. It has been well established that the movements and heart rate accelerations are tightly coupled in the last trimester of pregnancy and that as the fetus becomes hypoxic or asphyxiated or malnourished, movements will diminish and coupling between accelerations and movements will start to disappear. I think if we take a look here, there will be some published standards during the course of the year that I think will appear in three different sites that are the result of a consensus committee in which I was a participant that will look into the issues of definitions of accelerations and decelerations and so forth. But I can assure you that you don't really have to worry about this 15 beat business. Because if you get to 15 beats over baseline, you are going to be there for an average of 20 seconds or greater so you don't have to get out your little ECG caliper that you used as a medical student and get down and measure the amount because it is really a non-issue. In fact, in the earlier part of the third trimester, most younger fetuses really produce many more 10 beat accelerations than do 15. Discounting the fact that gestational age is a factor, I think, is shortsighted. If we take a look at interpreting NSTs, I am not going to burden you with what is reactive and what is not reactive because many of you trained in places different from my training experience, had different professors and different resident colleagues and different colleagues. But rather I will just say that we all use a definition that embraces the concept that there are movements and with these movements are accelerations. It brings up the issue that nonreactivity really is a graded phenomenon. You can have accelerations but they don't meet your arbitrary threshold or they are not there but movements are still occurring or none of the above. If you take a look here, in reality, here is a test that nobody should have any trouble with. Even from the back of the room, you can see that there are lots of accelerations. The baseline has a lot of variability in it. People say, "Well. It's an external tracing. How do you measure variability?" The second generation Dopplers actually introduced much less jitter than their first generation predecessors. I would submit to you that crude assessment of baseline variability is certainly quite possible on an external trace. You just can't quantify it so easily by eye. But this has the appearance of normal variability. Here, if we take a look, we can see a different situation here. This is a troublesome kind of test to get because you would have to be up close to see the little blips that occur here periodically. That is ultrasound jitter but here is a baseline rate that is about 140 that there is no reactivity. The

5

V.

Actocardiotocography (ACTG) A.

Combined recording of Doppler-derived FHR and FMs in same time

VI.

B.

Commercial units now available

C.

Evaluated as potential for extending capability of NST

D.

Studies are very limited

Test Selection/Diagnostic Values A.

NST vs CST 1.

Contraindications a)

NST: none

b)

CST (1) Third trimester bleeding (2) Premature rupture of membranes (3) Hypersensitivity (4) Possible previous uterine surgery

2.

3.

Applications a)

NST: safer to perform

b)

CST: potential hazard of UCs

Similar

baseline variability here appears to be diminished but I would also suggest to you that occasionally we have mothers that operate in this baseline rate. So I can't even assure you that this is fetal and I think that is a very important observation to make. An apparently prolonged nonreactive strip like this may not even by a fetal strip. It is always good when you see this and it persists that you don't make the assumption that you are looking at a nonreactive baby necessarily. Here is the sequence and this is the sequence I alluded to earlier. What happens when you become progressively hypoxic is you decrease your acceleration counts and amplitude, you decrease the movements you uncouple and eventually you start seeing spontaneous accelerations and that series of events has been very well worked out in the laboratory. If we take a look at testing conditions, this is extremely important. Because if you are going to do a test on a baby and you are going to look at that baby over time, one of the things that you want to keep in mind is the fact that if you are going to compare test 1 to test 2 or test 2 to test 3, that it is "apples and oranges" unless you have done the test the same way. You have to have standards and your standard will be how long you observe, the time of day in which you do it and whether the mother has eaten or fasted. What we see here is that if you take a normal baby of term, that almost all normal babies will produce at least one acceleration within an hour. So that should be a minimum standard for reactivity and that, in fact, they will do it generally within 70 minutes and that if at term, you produce a lower rate of accelerations than that, your likelihood of having a followup CST which is positive is extraordinarily high. That suggests that a prolonged nonreactive test ought to be a consideration of term for proceeding on to getting this baby delivered.

a ) Test length: if nipple-stimulation used

4.

5.

b)

Specificity

c)

Negative predictive value

Varies with interpretive criteria a)

Sensitivity

b)

Positive predictive value

Note: few prospective studies of sufficient size exist to establish clear-cut advantage of either approach

6.

Most recent study suggests that NST and nipple stimulation CST are virtually equivalent predictors of outcome in similar obstetric populations managed in a similar manner

B.

Diagnostic values 1.

Specificity, negative predictive values are excellent with

both tests 2.

Sensitivity varies with criteria, population tested (see

Table 3) 3.

False-negative test is uncommon and usually results from nonpredictable events (see Table 4)

4.

False-positive rate is extremely variable but in most studies approximates 50%

5.

Conclusions: both tests are most useful in determining health but

I mentioned immaturity and I should come back to this issue because we have been very successful in pushing the envelope of fetal salvage well back into the third trimester. There are many institutions now that if you give them a "rompin-stompin" 575 gram 24 weeker, they will many days later produce a slightly impaired neonate that goes home on oxygen and maybe has some visual problems. But the very fact is that there is plenty of precedent for looking at a high-risk pregnancy at a very early window of the third trimester. If you are going to do that, you have to accept that these babies are different from babies at term. That is an important thing to keep in mind. You can't use term standards for acceleration amplitude. It is lower. They very frequently will produce little brief decelerations of movement. That is normal. There is less coupling of movement and accelerations. That is normal and if you are going to use standard term reactivity for these babies, what you are going to find is it takes longer to get a reactive NST. You can eventually get it but you are going to have to have that mother on the monitor for maybe an hour, an hour and half, two hours. On the other hand, if you see a prolonged nonreactive trace, if you see lates, if you see totally absent variation or severe variables, this is a bad trace no matter how old the baby is. I think if we take a look now, there are other things we need to bear in mind when we do testing. It is a matter of just getting a good inventory. We see a number of patents that are on beta blockers. Beta blockers will reduce your resting baseline just like they do in adults. It will reduce your ability to accelerate your heart rate just like they do in adults. So a patient who is taking beta blockers, which are very popular drugs for the treatment of chronic hypertension in pregnancy, you probably are going to have to use a test other than heart rate to assess the well being of the baby. Nicotine and cocaine, two drugs of abuse, are both very notorious for raising the heart rate. When you raise the baseline heart rate, your ability to vary that heart rate is also diminished and the amplitude of your decelerations is going to be decreased.

fall short of good diagnostic methods for

determining disease

What about the issue of diet? You really have to make a mother pretty hypoglycemic to induce a nonreactive state in the fetus. On the other hand, if the mother is very hyperglycemic, the fetus will breathe more and do body movements less. So again, one has to bear in mind that glycemic status

6

Table 4 Diagnostic Values (%) of NST and CST for Perinatal Morbidity

Study

No of

Test

Sen

Speci-

False-

False-

Pa-

sitivi

ficity

Positive

Negative

tients

ty

Devoe

297

CST

33

85

98

1

Weingo

381

CST

60

94

87

1

ld

566

CST

22

91

96

1

Keane

390

CST

43

85

85

4

Free-

297

NST

33

79

98

1

man

795

NST

67

63

97

1

Devoe

367

NST

80

83

94

0

Evertso

566

NST

33

81

97

1

may be important, especially so when we are evaluating mothers that have diabetes. Sequential tests, I iterate this point again, because it is just like the concept of maternal perception of movement. If you are following a baby over time, it is quite important to maintain a good ability to see what that baby has looked like from session to session. Because one of the things that you would like to be able to pick up is the baby who is less well today than he or she was a week ago. I think in looking at the concept of the fetus' own control, if you just keep a record of strips or representative segments of strips and you array those so you can see those each time the patient comes in, you will get a better idea of when the baby starts looking not so well and that is the point at which you are going to consider intervention. Now, let's get to the RCT evidence for NSTs because it is a little bit disappointing. You know, here is a test that has been the mainstay of our fetal testing armamentarium for many, many years. In the United States, the NST has been applied since the mid '70s so we are talking about well over two decades now of clinical use and yet there really have been only four trials that have focused on the NST and all of these have been done overseas. All of them have been done in the United Kingdom.

n Mende nhall Keane Table 5 Cumulative Reports of Perinatal Deaths Following Normal Tests Less Than 7 Days

Reported Mortality

Uncorrected Mortal-

Corrected Mortality

ity Rate

Rate

Cumulative

6.2/1,000

2.5/1,000

(NST):8,0433

10.3/1,000

4.2/1 ,000

Cumulative

8.4/1,000

3.5/1,006

(NST):2,15490 Cumulative (CST1:4,62690 Exclusions for congenital malformations, cord prolapse, sepsis, immaturity

You can see a couple of things here and we will get back to the number issues. If you take a look at these trials, what you see here in terms of outcomes, and we will focus on the outcomes you are most interested in which are the issues of intrapartum problems, low APGAR score, abnormal neurologic outcomes and perinatal death, what you see here is that in the trials that used NST versus nothing, NST versus nothing didn't appear to give better outcomes. Spuriously, it appeared to actually make it more likely that you would have a perinatal death. But again, I think the issue here is if you take a look at the actual raw data, the numbers of perinatal deaths were so small, you can see the confidence interval is extraordinarily wide so that you probably can't generalize those data. Again, if we take a look at when NSTs were used or revealed to clinicians, the outcomes were almost the same as when NSTs were not used or were concealed and this direction was consistent. The bottom line here, and the one take home I got from reviewing these data was that NSTs probably shouldn't be your only well being test. That is why I began with the context that fetal growth profiles have got to be considered in the overall context and the NST becomes a component of fetal assessment but certainly cannot be the only hook on which you hang your entire well being scheme. What I do want to point out and perhaps I should backtrack on this slide for a moment is these trials were reported in the '80s. They reflect obstetrical practice of that era. The size of the trials, one of these trials had more than 500 entries and I did a little analysis based on the kinds of outcomes we are likely to want to be able to detect and prevent and you would need to have a trial of 15,000. I can tell you sort of parenthetically, because it is water under the bridge and I don't bear any malice or animus to anybody in Washington, that I proposed such a trial to the NIH in the late '80s saying that if they really wanted to do it right this is what they would have to do. They didn't give it a high enough score to get funded. So it is very unlikely that we are going to get this kind of trial ever done in the United States which is too bad. But I think we also need to move on and say, "Well. These are the best data we are going to have and the NST probably should not be the sole test for assessing the well being of the baby."

How many people have this in their practice? I see a smattering of hands. It looks like about, by rough calculations, 12%. If you work at a VA hospital… actually I work at a VA hospital, I don't do any OB over there oddly enough… they have very few pregnant veterans. But I do gynecology over at the VA and you see these a lot because the VA takes care of a lot of individuals who got hooked on smoking and lost their larynxes. So the electronic artificial larynx sort of serendipitously got into our practice because it has some properties that are useful for stimulating the fetus. What it has really are two properties. It has a sound property which is very distinctive and it has a vibratory property. The sound property is a broad band signal. When you record this signal with a hydrophone - a microphone adapted to a water environment - you put this inside the uterus of a pregnant woman near term, you register a sound pressure level of 110 dB. How loud is 110 dB? Does anybody here have teenage children? Do they listen to hard rock? One hundred ten dB is Metallica on

7

stage. This is why most rock musicians are deaf and this is why teenagers who listen to that music have transient deafness that usually persists until they are about 30 because this is very loud. Of course, the signal is very brief but this is loud and this will get your attention. In fact, this isn't what stimulates the fetus. What stimulates the fetus is the vibratory component. The reason that we know this is that the fetuses that have been stimulated this way that have congenital deafness respond to this stimulus. The stimulus uses a well known pathway through Meissner's corpuscles that when the baby is born and you stimulate that baby by rubbing it, you are stimulating Meissner's corpuscles and you get a response from that. That is the response that we get from putting on the vibratory stimulus. What you see is something like this. You get a nice transient tachycardic response that occurs very quickly after you apply the stimulus. This is what you expect to see and it works the same way in adults. When I am making rounds and I see a resident fall asleep, I just take the little vibroacoustic stimulator, put it right behind his ear and he or she is up in a matter of nanoseconds. They have occasionally done it to me when I fall asleep during rounds. So it works. It is not pleasant, but it works. The response is very quick. You should see a response within a few seconds. The mean duration of this response lasts for about 10 minutes so that occasionally you will get a fetus, not quite 100%, but mostly normal fetuses will respond to this stimulus and they will have an increased heart rate baseline. We have had one isolated case of a fetus that became tachycardic, got up to about 180. After two hours, the people in the testing center started getting a little concerned and they called me. I said, "Gee. That is pretty impressive." We took that baby up to the labor unit and the heart rate stayed up in the 170-180 range. Twelve hours later it was down to about 175. The next morning, it was down to 160 and this was the second baby we did with the stimulator. They asked me if I would consider ever doing this again and I said, "Well. That is an unusual response." We haven't seen it since but it has been reported. So occasionally, you have bizarre and unexpected sustained tachycardias in perfectly normal babies and it just proves that some babies have an on switch and they don't have an off switch. The other thing that we notice here is that what we are really seeing when we put this stimulator on is we are taking babies generally from a quiet state to an active state. This is part of the problem because the response to this stimulus is state dependent. Babies that are in quiet state, quiet sleep, non-REM, will respond predictably by a very rapid changeover into active state. Babies that are in active sleep don't consistently do this. If you take a look at the state distribution that babies have during the third trimester, there is more time spent in REM sleep than in non-REM sleep. That is why normal babies don't always predictably respond to this stimulus and that is where a failed acoustic stimulation doesn't exclude a normal outcome. The OSHA data on this are still somewhat sketchy. We have very few studies that have followed up babies that have been exposed. The limited data so far in neonates that have been followed in early infancy and up to ages 2 and 3 suggests that there is no hearing impairment or behavioral damage from applying this stimulus on repeated occasions. If we take a look at the advantages of vibroacoustic stimulation, you can get this response very quickly in most babies and it may also be useful intrapartum. We use this technique intrapartum to stimulate babies who are having abnormal heart rate tracings to try to exclude acidemia. A positive response is a pretty good indicator that you have got a pH over 7.25. But unfortunately, lack of response is not informative. These are the data from an earlier trial done by Carl Smith just comparing the increase in reactivity you get by applying this test as opposed to the nonstress test. If you have a busy testing unit, this can shorten your testing time and apparently give you comparable clinical results. Again, with respect to the randomized trials, there have been only a small number reported. If we take a look here, we see that in terms of reducing nonreactivity, it appears to be a very effective test. You can see it has about a 40% reduction in nonreactive tests. All the remaining outcomes of interest are really comparable between the NST standard and the vibroacoustic stimulation. Because we don't know it is safe, I am marketing this. I haven't figured out a way to get it in utero yet but I am working on that as my retirement project so that in case we do find out it is harmful we will be able to protect babies in the future. It is okay to laugh at that slide. I like that slide. I want to move onto the next of the variations on NST which is that actocardiotocograph. I abbreviate it to ACTG because that is a mouthful to say early on a Sunday morning. Most of the monitoring systems that are being vended today have this as a feature. This feature is actually a fairly neat feature and I am going to explain to you how it works and how it might help you in improving your antepartal assessment. What the point of this is is that we have actually a lot of background that has looked at indirect and direct ways of tracking fetal movement during the course of antepartal observation. This goes back into the late '70s. The use of toco units placed on the abdomen and the typical spiky little patterns you get when the fetus moves and perturbs the toco. But there is a lot of noise in that system and it is not a really truly high fidelity way of recording movement. About 10 years ago, Tim Wheeler in the United Kingdom did some work on the processing of the raw Doppler signal and was able to isolate low frequency Doppler signal disturbance from high frequency and low frequency disturbance of the signals associated with fetal movement. High frequency is associated with fetal heart rate. This principle was then applied and actually put into a number of developmental systems which now produce these actocardiotocographs. There are several systems out on the market that are available and most of these have appeared within the last five years. What basically is happening here is that the systems employ a filter which discriminates on the basis of frequency that Doppler shift associated with movement from that associated with heart rate. A detection algorithm is developed. This algorithm is generally based on comparing the output from the processed Doppler signal from what an observer sees on an ultrasound screen and there is a correction factor. The correction factor or the coefficient relies on the probability that if you have this disturbance for a certain amplitude and a certain length of time, this is movement as opposed to just random noise. Then, basically you print out on your trace here. This is an example of what one of these actocardiotocographs looks like. You see here, it doesn't take any great powers of observation to see that you've got lots of accelerations. On this channel here, you've got the little areas which are maternal perceptions of fetal activity that are being clicked and on the bottom here. You've got these blocks which are being printed out automatically by the monitor and these correspond to the low frequency Doppler signals that are being clustered together as fetal movements. What should strike you here is that you occasionally get the recording of a Doppler signal that is independent from the maternal perception and it suggests that this system will pick up movements that the mother misses. As you can see here, these are associated with accelerations where the mother wasn't reporting movement at

8

that time. If you take a look, there really have been relatively few studies that have been done that have compared this kind of recording system to the standard NST. We did one of them published a couple of years ago in the Gray Journal and we can take a look here. These are the standard players in heart rate analysis for detection of fetal compromise. The fetal movement percent, which we arbitrarily cut at a threshold of 5% of time of observation, was approximately as good as NST reactivity or variability in terms of trying to predict the fetus that was likely to be compromised. Further, if we take a look here, the point here is that for mothers who have difficulty in perceiving movement, the Doppler processing appears to be a more sensitive means of doing that and, in fact, reports about twice as many movements as the mother does. I think secondarily, one of the things that is very nifty about this is that in a resting state, this Doppler processing system gives you additional information about behavior of the fetus. So you not only have the heart rate variability and the heart rate activity, but you also have movement which are all components that go into distinguishing one behavioral state from the next. I think it is important because you can start seeing babies using this technology that are nonreactive based on no movements and no accelerations from those that are nonreactive, who are moving but are not responding with accelerations to movement. I think that is a very important observation to make and I will tell you that although this technology has been out for a few years, there really still are no randomized control trials, to my knowledge, that have tried to explore this versus standard NSTs. This slide here just simply illustrates the approximate difference between mean perceived movements by the mother versus perceived movements recorded by the machine here. That is about a ratio of 2:1. I want to move on. How many folks continue to do CSTs in their practice? Just a show of hands. That looks like about 9%. I am sort of surprised because, in a way, that was the original fetal heart rate test in the United States. You are probably aware of that. Anybody that was in training like I was in the late '60s and early '70s knows that you've got your 1972 publication by Ray and Freeman that was the original description of the CST. But actually, if you want to go back even further, the uterine contraction stress test was really developed in South America and was imported to the United States and that started appearing in the late '60s. That is actually where this slide is taken from is some work by Pose and Posero which came out of a symposium in Uruguay. This was published about 30 years ago. What this observation is, and it is very important to understand what information is conveyed by the CST and probably why we shouldn’t totally drop this from our armamentarium. Because if you take a look here, under normal conditions, a moderate to strong contraction will reduce intervillous blood flow in the placental bed and as a consequence of that, the fetus will experience transient reduction in pO2. A healthy, well oxygenated fetus, this won't bother at all because the pO2 will still be well over the critical threshold of 18 torr. So your heart rate will not vary significantly with a contraction. You won't see any decelerations of any kind. At the other extreme is the fetus who is just barely making it. This is your chronically compromised, perhaps your IUGR fetus, just hanging in there. What you are seeing here is that it takes contractions of relatively modest intensity to produce late decelerations because the fetus already is marginally oxygenated and is probably becoming mildly acidotic. When you get into the marginally oxygenated, mildly acidotic stage, your myocardium simply doesn't work as well as a compensatory muscle and so you have late decelerations. You can't make that compensation to the need to drive output faster to supply vital tissue. Then we have the fetus that is sort of in between. This is a concept that is extremely important because… this applies to labor as well as it does to the antepartal period, we know that… and perhaps one of the things that we have seen here is the result that using this kind of approach to assessment - NSTs and CSTs - is that we have inadvertently or advertently, hopefully advertently, started to exclude a whole subgroup of patients from ever going through the process of labor that we would have included in years past. If you take a look at another phenomenon here, you have, of course, the baby that is chronically underoxygenated and this baby, given enough contraction intensity and frequency, will produce lates. Then you have a situation of the so-called decreased amniotic fluid post mature baby that we don't see very often anymore because people are now inducing at the drop of a hat to avoid the 42 week dilemma. But you still occasionally will see a patient that goes to 42 weeks and has diminished fluid. Here you have the phenomenon of cord compression and cord compression, secondary to oligohydramnios, produces generally variable decelerations and, in fact, if you have the baby that is mildly hypoxic and has decreased amniotic fluid, you may have both. I have brought with me some examples here. Here is simply a normal trace. Here you can see frequent contractions. I wanted to show this tracing. I like this trace because it illustrates another point. If you take a look over at the left hand side, what you see here is a baby that appears to be very quiet. This would be considered a nonreactive, negative CST, if you will, and then within a matter of moments here, this baby changes state and becomes much more active and has reactive accelerations. Just again, it is a basic issue of how long you let the test run but this is what normal babies do. They cycle in and out and this is a very healthy test. This is just a blowup of one segment of this test but just to point out to you that there should be a lot of irregularity of baseline heart rate associated with the CST as well. It shouldn't be totally smooth. So if you have a baby that has a totally smooth baseline here, even on an external trace, that is potentially a problem. Because remember what I said. The loss of variation, that is the last thing to go and if you get far enough down the line, you don't produce decelerations anymore. You just flatline. Here is a baby that I mentioned earlier. This baby has two things going on. It has sort of a potpourri of late decelerations and then periodically you have these nice little variable decelerations tossed in. Here is a baby that is postdate with oligohydramnios. So you see actually both phenomena. Here is a classic tracing here. You might say, "Where are the contractions?" They are there but I have to sort of give you a caveat here. I think this mother weighed about 350 or 360 pounds - had her own zip code - so one of the issues here is the fact that external monitoring of uterine contractions sometimes is an issue. But even if you couldn’t monitor them that well, what else could this be? Every time you have a contraction, you have got a late and you can see that the heart rate here, again, just barely recovers to baseline and then there are little efforts here to compensate with an increase in heart rate towards the end of the tracing. This is terrible. This is an awful tracing here. I hope you all don't see this with any frequency. This is a baby that is dying. What we have got here are not just lates, but look. This is an external trace, I might add. Where is the variability? Who says you can't tell whether variability is absent on an external trace? If we put an electrode on this baby, you can see this is a baby that is dying. You don't see the end of this trace, but it starts settling out into the west here. We have a number of different ways of getting at this same outcome. I bring this up to you because you may occasionally need to go back and do a CST. The indications are becoming much rarer because we have other modalities to add here but this is only one reason. Anytime you see, especially an unexpected, positive CST with late decelerations, you have to consider that there may be maternal issues going on here such as maternal hypobulimia or oxygenation. Those certainly need to be ruled out. Again, if your contractions are too frequent, the so-called hyperstimulation, you can make even a good baby look bad if you induce contractions without cease for a 10 minute window and cut off the intervillous flow for that long. Then periodically, you will get babies that throw pretty prolonged variables as well into the picture

9

from a cord compression. The main reasons that we have sort of drifted away from using CSTs as our principal test, has this been the applicability issue and ease? CSTs are not tests that are usable on patients that have bleeding or patients that have ruptured membranes and we would probably be less inclined to use on patients with classical scars even though there is almost zero data to suggest that you could rupture a classical scar with a short blast of contractions on a CST. Nonetheless, we don't do it under those circumstances. It takes longer to do one. Even if you use breast stimulation, it takes about twice as long to do a CST as it does an NST. So, in essence, the CST has sort of been relegated to being a followup test rather than a primary test for fetal well being. Even the gurus of CSTs on the west coast have come around to this point of view over a period of 20 years of fighting and feuding. I show this slide here because occasionally you will see the baby that has a variable deceleration in antepartal testing. I get perhaps more calls and concerns about this than I do some of the really terrible things I have shown you already. The reason for this is that people don't know what to do with these variable appearing decelerations that are on an antepartal test. I can assure you that I have spent enough time looking at babies over continuous 24 hour periods at term who are perfectly normal to tell you that almost every normal baby, if looked at over an entire day, will do this. So, in and of itself, seeing something like this should not be a rush to judgment. How do you respond to a spontaneous deceleration that doesn't appear to be associated with a contraction? There are two things that you do. You extend the tracing out to see whether or not this is a trend, whether these are becoming more frequent or whether they disappeared, which is what you expect in a normal baby. Secondarily, you back it up with ultrasound observation. The most common concomitant pathologic finding with this is decreased amniotic fluid. The second most common is nuchal cord. The third most common that you see with this is simply a baby that is doing a lot of arching and extending. These are all things that you will need to have another modality to help you figure out. I want to get into the issue of indication for testing and test application. This slide here comes from a publication that we did about six or seven years ago that looked at the relationship of test outcome as a function of why you did the test. You may say, "That is a very basic and mundane sort of thing." But think about it for a moment. Babies that have the most common indications that we use for testing are diabetes, postdatism, growth retardation and hypertension. Those four indications in our testing unit are over 85% of all tests. Those are the "big four" as I call them. If you take a look at the big four, these babies are not all at the same risk for having bad outcomes and they don't all get to bad outcomes through the same route. If you look at this even more closely, what we just did was we broke down in a group of 1,000 consecutive patients what you got per indication for this test, the NST. What you see here is that from a standpoint of picking up babies that are likely to be compromised, the so-called sensitivity issue, the only groups in which sensitivity really was high at all were those that were IUGR and those mothers with hypertension. You can see that with postdatism, the sensitivity of this test was pretty low. That happens to be in many units the most common reason… in a normal practice, that would be the most common reason for your doing a test. That is the way it sorts out across the United States. The most common indication are mothers that go postdate because most of the other really serious complications either get sent to tertiary units or are managed and co-managed in consultation with a perinatologist. So that is what you are left with. You are left maybe with post dates. You might be left with a couple of mild chronic hypertensives and gestational diabetics but we don't even test gestational diabetics that have normal growth until they get to 40 weeks. What you see here is that these tests are good at excluding babies through high specificity who are unlikely to have compromise. The negative predictive value, which really tells you if your call is that the baby is going to be okay, how good is that? They are all very high. So it is important to understand that this is a test, as a single test, that is much more potent in babies that have birth restrictions and mothers with hypertension than it is in any other conditions that we test. We have an algorithm. I think I put that in your handout. I sort of hate cookbooks. My wife is a professional chef. She likes cookbooks. She always improves on them. I don't like them but you have to give people in a testing unit something to sort of hang their hats on and this is essentially what we do. But this particular algorithm here is based on a fairly vast clinical experience that has looked at outcomes and has worked its way back into a rational management scheme. We do about a 30 minute assessment because we figure at that point, we are going to account for behavioral state changes. We are very highly likely to see the babies that are going to be reactive, be reactive. Then we have steps here of extending the test. We don't extend an NST more than 90 minutes. This is at term and the reason for that is simply that our experience has been if you are not reactive within that time, "you ain't going to get reactive" and you need to go onto another modality. Then I think that the issue there is that if you are mature and you have prolonged reactivity, there isn't any reason that I can think of for keeping you pregnant. If you are immature, then you need to go to a backup test, either extended into a biophysical profile or a CST. The good news, and there is good news, even if you think about all the conditions that we do NST testing for, you think about all of these, what we have gotten from this is that the fact that babies have been subjected to this kind of scrutiny, gives you a corrected mortality rate that is almost one-fourth of that of the general OB population. The general OB population has a mortality rate, in the United States, of about 9.8 per 1,000. Here you are looking at 2.5 per 1,000. That is your expected mortality in a large general normal OB population with no high-risk factors whatsoever. So what you have really done through both the testing observation and the engagement in more intense care schemes is you basically have rendered your high-risk population to the same level of perinatal mortality as your low-risk population and I think that is a very important take home. It doesn't matter whether you use NST, vibroacoustic stimulation, CST, ACTG. It makes no difference. Any of these gives you that kind of good outcome. It is almost time for our coffee break. That is why I tossed this slide in here. This is what our coffee break looks like in my office. I think everybody is smiling at this point. Either that or we have been hyperstimulated by caffeine but I am going to stop at this point and we can have the lights back on and I will be very happy to entertain any questions.

10