93

Frank et al, Noninvasive FECG FHR monitoring

J. Perinat. Med. 20(1992)93-100

Noninvasive fetal ECG mode fetal heart rate monitoring by adaptive digital filtering Thomas H. Frank1, Otis R. Blaumanis1, Serena H. Chen2, Roy H. Petrie3, Richard K. Gibbs1, Robert L. Wells1, and Timothy R. B. Johnson2

^erinatronics Medical Systems, Inc., Crofton, Maryland, U.S.A., 2Division of Maternal-Fetal Medicine, Department of Gynecology and Obstetrics, The Johns Hopkins School of Medicine and Hospital, Baltimore, Maryland, U.S.A., 3 Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, Missouri, U.S.A.

We describe an abdominal fetal ECG method of FHR monitoring using Adaptive Digital Filtering which obtains a valid noninvasive measure of FHR beat-to-beat variability.

1

Introduction

An accurate measure of instantaneous fetal heart rate (FHR) and FHR beat-to-beat variability (BTBV), is an important prognostic indicator of fetal well being [10]. Present techniques used to noninvasively measure FHR, i. e., Doppier ultrasound (US), are not capable of providing accurate FHR beat-to-beat variability [6]. The purpose of this paper is two fold. First, we report an evaluation of the accuracy of a new noninvasive abdominal fetal ECG (AFECG) mode of FHR monitoring, which utilizes Adaptive Digital Filtering (ADF). The evaluation compares the beat-to-beat fetal ECG R-R interval measurements, obtained from maternal abdominal ECG electrodes, with those R-R intervals simultaneously obtained from a direct fetal scalp electrode (FSE). Second, we provide a comparison of our evaluation of this noninvasive abdominal fetal ECG FHR monitoring technique with prior evaluations of Doppler US FHR monitoring techniques currently in use.

Curriculum vitae Dr. THOMAS H. FRANK was born in 1944. He has completed an A.B. degree in physics and the BSEE, MSEE, and Ph.D. in Electrical Engineering. He is currently President and Director of Research and Development at Perinatronics Medical Systems, Inc., Crofton, Maryland U.S.A. From 1982 to 1985 he served as Assistant Professor, Department of Obstetrics and Gynecology, University of Maryland School of Medicine, Baltimore, Maryland. Previously he conducted research in radar, sonar, and digital signal processing at the Westinghouse Electric Corp., Baltimore and Annapolis, Maryland U.S.A. for twelve years.

I

each and every heart beat, namely, FHRj = l/(Ti — ΤΜ), where the instantaneous FHRi is computed from the difference in time of the previous heart beat, Tj_b and the present heart beat Ti. The variation in the FHR has been described as BTBV, or short term variability. The clinical significance of BTBV is well documented [10]. BTBV is defined as the absolute value of the difference between two successive beat-toThe presentation of FHR on electronic FHR beat heart rate determinations. That is, BTBVi monitors should be instantaneous FHR and notBrought = |RtoMyou —Ri|, where: Rj represents the instantaby | University of Arizona averaged rate. That is, the rate is presented at neous FHR determination for the ith R-R interAuthenticated 1992 by Walter de Gruyter & Co. Berlin · New York

Download Date | 6/6/15 2:51 AM

94

Frank et al, Noninvasive FECG FHR monitoring

val and RM represents the instantaneous FHR for the previous R-R interval. Several earlier AFECG techniques have been evaluated and compared with FSE generated FHR data. LEVENTHAL et al [8] reported a method of continuous FHR monitoring employing the fetal ECG obtained from electrodes placed on the maternal abdomen. They qualitatively reported that their determinations of BTBV were precise and without the artifactuality of ultrasonic or phonocardiographic methods. Previous AFECG techniques [2] have been ineffective in blocking out the maternal ECG, which is necessary to prevent calculating the maternal heart rate rather than the FHR. NAGEOTTE et al [9] quantitatively evaluated FHR BTBV derived from fetal R-R intervals using an abdominal ECG technique in patients undergoing antepartum FHR monitoring. Their computer study, which made allowances for the fetalmaternal ECG coincidence by excluding these intervals, was thus only able to analyze 13% to 61% (± 2 SD) of the FHR record with a mean of 37% of the FHR data reported. We [3] have described the development of a prototype electronic fetal monitoring system which utilizes an AFECG technique with ADF. This system optimizes a set of ADF algorithms for the cancellation of the interfering maternal abdominal ECG component and muscle noise even in the face of maternal-fetal ECG coincidence to provide for an accurate measurement of fetal R-R

intervals for the computation of instantaneous FHR and FHR BTBV (figure 1). In 1980, with the then current noninvasive FHR monitoring techniques, SOLUM [11] compared the percentage of acceptable high quality FHR tracings without artifacts using phonocardiography (23.4%), abdominal ECG (55.2%), and ultrasound FHR techniques (85.9%). Several investigators have evaluated the accuracy and precision of FHR derived from Doppler US FHR monitoring techniques [1, 4, 5, 6, 7]. Technical enhancements, including autocorrelation and range gating, have been made to improve both the quality and the cosmetic appeal of Doppler US FHR tracings [4]. LAUERSEN et al [5] evaluated the accuracy of a Doppler US FHR monitor utilizing range gating compared with an FSE technique in high-risk patients. BOEHM et al [1] conducted a detailed study of the distribution of R-R interval differences between direct fetal scalp ECG electrode and Doppler US FHR data for first and second generation (autocorrelation) Doppler monitoring techniques. LAWSON et al [6, 7] completed two studies comparing different Doppler US detection systems with FSE detector systems. They concluded that "measurements of BTBV by Doppler are unlikely to be accurate and there seems no point in using them as an index of FHR variation" [6]. With this background, we present an evaluation of the accuracy of the AFECG technique utiliz-

Figure 1. An electrocardiographic record of the composite ECG signal recorded from the pregnant maternal abdomen. The upper tracing displays the raw ECG signal consisting of maternal (M) ECG QRS complexes and fetal (F) ECG QRS complexes. The lower tracing shows the processed result after Adaptive Digital Filtering (ADF) of the maternal ECG signal. (Scale: standard ECG speed of 25 mm/sec with increased sensitivity of Brought to you by | University of Arizona 4 uvolts/mm). Authenticated Download Date | 6/6/15 2:51 AM

J. Perinat. Med. 20 (1992)

Frank et al, Noninvasive FECG FHR monitoring

ing ADF and a comparison with previous Doppler US studies.

2 Materials and methods Five women, in labor, with uncomplicated pregnancy between 38 and 42 weeks gestation who were being monitored with the direct FSE for clinical indications, were recruited to participate in this evaluation with appropriate institutional approval. Four chest ECG electrodes and three abdominal ECG electrodes were used to collect maternal thoracic and abdominal ECG signals. Abdominal palpation by Leopold's maneuvers determined fetal position for placement of an abdominal electrode at the fetal head and another at the buttock. Fetal ECG signals, totaling 3298 seconds, or 55 minutes, (and consisting of approximately 8000 R-R intervals) were collected from these subjects using the Perinatronics prototype FHR monitor with ADF in the labor and delivery units at the clinical sites. Digital recordings were made from these ECG signals, collected during labor from the direct FSE and maternal thoracic and abdominal ECG electrodes required for the noninvasive AFECG technique. These recordings of the amplified ECG signals were made using a DEC PDF ll/ 23 minicomputer equipped with a 12-bit analogto-digital converter system, sampling each ECG signal at a rate of 500 samples/second. The PDP 11/23 computer system was subsequently used to play back the FSE and noninvasive ECG data into two Perinatronics prototype FHR monitors using a 12-bit digital-to-analog converter.

95

Each FHR monitor which was used in the assessment was equipped with an output jack to provide timing pulses which were generated at the monitors' detection of fetal ECG signals. These FHR monitor output timing pulse sequences were then used in the subsequent computer comparisons. Pairs of timing pulse sequences, one generated by each of the two monitors used in both the control and study analyses, were simultaneously acquired by a second PDP 11/23 computer, and later analyzed by a FORTRAN computer program developed for this study. The computer analysis program precisely determined corresponding pairs of fetal ECG RR intervals from these two sets of pulse sequences. These pairs of R-R intervals were subsequently converted to separate measures of instantaneous FHR and BTBV. The difference in these two fetal heart rates, on a beat-to-beat basis, was compared for the study group and the control group using the computer analysis program. The statistical analysis measured all R-R intervals, the mean R-R interval difference, standard deviation of the interval difference, and provided a complete distribution of the R-R interval differences at 1 bpm intervals between pairs of FHR instruments. The analysis program utilized the fetal pulse sequence generated by the FSE equipped FHR monitor as the reference for comparison with the AFECG fetal detection timing sequence. The analysis program, like the FHR monitor, does not report instantaneous FHR or BTBV if the instantaneous change in rate is greater than 20 bpm. The analysis program also calculates the percentage of R-R intervals in the data that are reported.

Two sets of computer comparisons, referred to below as the control and study analyses of the 3 Results instantaneous FHR and BTBV, were made using the recorded ECG data. In the control analysis We have examined a comparison of R-R inter(FSE-FSE), the recorded FSE data were re- vals from a pair of prototype Perinatronics FHR played through two Perinatronics FHR monitors monitors to provide an effective method of exeach equipped for a direct FSE input. A com- pressing the accuracy of the AFECG FHR monparison of the outputs of these monitors served itoring technique with ADF compared to the as the control in order to rule out differences, if direct FSE monitoring technique. Table I lists any, between FSE equipped instruments per se the length of each subject's FHR record statisprior to an effective comparison of the AFECG tically analyzed, average FHR of each record, technique with the FSE technique. In the study mean R-R interval difference, standard deviation analysis (AFECG-FSE), the recorded AFECG of the R-R interval difference of the study group data were replayed into one AFECG monitor and control group for each subject, and the perequipped with ADF, and simultaneously, the re- centage of R-R interval data reported. The mean corded FSE data were replayed into a second difference between the R-R intervals of the study FSE monitor. A comparison of the outputs of and control group were both virtually zero. A Brought to you byof | University Arizonathe study group's these two monitors served as the study analysis. summary the dataofshows J. Perinat. Med. 20 (1992)

Authenticated Download Date | 6/6/15 2:51 AM

96

Frank et al, Noninvasive FECG FHR monitoring

Table I. Statistical analysis of R-R interval differences of the study group (AFECG-FSE) and the control group (FSE-FSE) Subject No.

001 002 003 004 005 Summary

Record Length

Average FHR

Mean R-R Interval Difference

S. D. of Interval Difference

Percent Reported

(sees)

(bpm)

Study (bpm)

Control (bpm)

Study (bpm)

Control (bpm)

Data

788 394 980 404 732 3298

175 114 138 144 139

0.0067 0.0409 0.0204 0.0433 -0.0553 0.0056

-0.0038 0.0194 -0.0026 0.0008 -0.0009 0.0005

1.94 1.03 1.34 1.26 1.50 1.50

1.08 0.74 0.65 0.72 0.63 0.79

70 85 59 90 74 72

Table Π. Percentage of abdominally collected fetal ECG R-R interval differences of instantaneous measurements lying within an accuracy of 1 bpm and 2 bpm of a direct fetal scalp electrode standard Subject

Within 1 bpm

No.

Study (%)

Control (%)

Study (%)

Control (%)

001 002 003 004 005 Summary

81.8 97.6 95.5 96.4 95.5 92.6

86.8 99.2 96.7 99.0 98.8 95.4

96.3 78.7 99.1 99.1 98.0 98.2

99.3 99.8 99.9 100.0 99.9 99.8

Within 2 bpm

standard deviation (SD) of the R-R interval difference is 1.50 bpm and the control group's SD is 0.79 bpm.

contained artifacts, resulting in a similar lack of agreement in the study group. In viewing all of the subject data, less than 2% of the R-R interval differences used to compute instantaneous FHR exceeded 2 bpm for the control and study groups.

Table II provides a comparison of the control group and study group fetal R-R interval differences within one-beat and two-beat intervals. These results illustrate that the noninvasive AFECG FHR monitoring technique using ADF 4 Discussion agreed with the direct FSE technique to within True instantaneous FHR and BTBV are impor1 bpm for 92.6% of the reported R-R intervals tant parameters in assessing fetal well-being and and to within 2 bpm for 98.2% of the reported fetal distress. Until now, these have been moniR-R intervals for the approximately 8000 paired tored accurately only by the fetal scalp electrode fetal R-R intervals analyzed for the five subjects and by this means, have been related to imporof this study. tant clinical correlates [10]. Initial enthusiasm for As shown in table II, less than 5% of all the AFECG techniques has not been supported by measured R-R interval differences were greater clinical or research use in recent years because than 1 bpm for all subjects other than No. 001 the basic technique of the AFECG method [8] is in both the control and study groups. For data frequently ineffective in blocking out the materset No. 001, more than 10% of the measured R- nal ECG to prevent calculating the maternal R interval differences, in both the control and heart rate rather than the fetal heart rate. The the study group, had errors which exceeded results of the present study, as indicated in table BroughtItoand you table by | University of Arizona II, indicate that there is no clinically 2 bpm. This suggests that the FSE data may have Authenticated Download Date | 6/6/15 2:51 AM

J. Perinat. Med. 20 (1992)

Frank et al, Noninvasive FECG FHR monitoring

97

significant difference in the instantaneous FHR with ruptured membranes during labor. BOEHM and BTBV obtained by the FSE technique and et al [1] reported a Doppler US comparison, the AFECG mode with ADF. There is some without an FSE control, in which they analyzed similar variation among subjects utilizing both 10,455 R-R intervals of data [approximately 87 techniques. minutes] collected from 15 randomly selected US is intrinsically limited in its capability to ;>ubJects in labor with ruptured amniotic memobtain true instantaneous FHR and BTBV. The branes *nd an F^m Place· The & and that Doppler closure. US signals are further limiting in that f ^^ ™* m"ge gatmg pr ™ft* "JfSS B V Wlth an error of 9 8 b m they can only sense the mechanical motion of ° ™ ™ ' P ^ SD> 11J table m the fetal heart (or valve) either toward or away ^ '· from the US beam - and not side to side. The data m table χγ allow direct comparison of For comparison, we have reviewed distributions the accuracy of the AFECG technique using of R-R interval differences from two similar ADF with LAUERSEN'S study [5] of Doppler US studies which examined the accuracy of Doppler using range gating and BOEHM'S study [1] of ultrasound FHR monitoring techniques. LAUER- Doppler US with autocorrelation. The AFECG SEN et al [5] reported a Doppler US — FSE monitoring technique with ADF agreed with the comparison in which they analyzed 362 minutes direct FSE technique while Doppler range gating of data collected from eight high risk patients or autocorrelation techniques are inaccurate.

Table ΠΙ. Statistical results comparing the accuracy of this Abdominal Fetal ECG FHR monitoring technique with two Doppler ultrasound monitoring techniques using the Invasive Direct Fetal Scalp electrode FHR monitoring technique as the control S. D. of differences in FHR (bpm)

Data reported

FHR Monitoring technique

FHR Record Length (min)

Mean difference in FHR (bpm) Study

Control

Study

Control

Abdominal ECG Range gated Doppler [Ref. 5] Autocorrelation Doppler [Ref. l]

55 362

0.0056 -0.3

0.0005 -0.1

1.50 9.8

0.79 5.8

/n/ \

V/o;

72 85-100

7.2

2.5

Table IV. Comparison of the abdominal fetal ECG FHR monitoring technique using ADF with two Doppler Ultrasound FHR monitoring techniques. The table shows the percentage of accurate measurements made to within a precision of 1 bpm and 2 bpm. FHR Monitoring technique

Abdominal ECG Range gated Doppler [Ref. 5; fig. 7E] Autocorrelation Doppler [Ref. l] J. Perinat. Med. 20 (1992)

FHR Measurements made to within 1 bpm

to within 2 bpm

Study

Control

Study

92.6 44.8

95.4 60.9

98.2 64.6

99.8 97.2



70.9



46.4

Brought to you by | University of Arizona Authenticated Download Date | 6/6/15 2:51 AM

Control

98

Frank et al, Noninvasive FECG FHR monitoring

The accuracy of the AFECG technique with ADF allows it to be clinically useful to determine BTBV noninvasively. For the data analyzed, this noninvasive maternal AFECG monitoring technique with ADF has been able to accurately report BTBV for 72% of the FHR record without averaging the presentation of instantaneous FHR.

The significance of this study is that a new noninvasive maternal AFECG technique utilizing ADF can provide a highly accurate continuous clinical record of instantaneous FHR and BTBV. Further clinical testing needs to be conducted on a larger more varied antepartum clinical population.

Abstract

Beat-to-beat variability (BTBV) of the fetal heart rate (FHR) is considered an indication of the neural integrity and is an important prognostic indicator of fetal well-being. We report the initial evaluation of a recently developed abdominal fetal ECG (AFECG) mode of FHR monitoring using Adaptive Digital Filtering (ADF) to accurately obtain BTBV noninvasively. Five women in labor at term were monitored with the direct fetal scalp electrode (FSE) and simultaneously with the AFECG using ADF. A computer analysis of 3298 seconds (55 minutes) of data provided a one-to-one comparison of the R-R intervals. One analysis of the direct FSE data with a second simul-

taneous analysis from the same electrode, to serve as control, was compared with the noninvasive AFECG data. The study group has a standard deviation of only 1.50 bpm compared to 0.79 bpm for the control group. The AFECG method agrees with the direct FSE method within 1 bpm for 92.6% of the reported R-R intervals and within 2 bpm for 98.9% of the reported intervals. This new noninvasive AFECG technique with ADF provides a continuous record of instantaneous FHR and BTBV that may be relied upon to provide an accurate continuous clinical record. The reliability of the technique has yet to be determined over a wide range of subjects.

Keywords: Adaptive digital filtering, beat-to-beat variability, fetal electrocardiogram, fetal heart rate, fetal monitoring.

Zusammenfassung

Differenz sowie die Standardabweichung der Intervallunterschiede zwischen Untersuchungs- und KontrollWir berichten über den Einsatz einer kürzlich entwik- gruppe bestimmt. Somit konnten wir die Verteilung kelten Methode, mit der über ein abdominales fetales der R-R-Intervall-Differenzen bei den beiden MethoECG (AFECG) mit adaptiver digitaler Filterung den zur Überwachung der fetalen Herzfrequenz über(ADF) die „beat to beat"-Variabilität (BTBV) genau blicken. Im Mittel gab es keine Unterschiede hinsichtund noninvasiv aufgezeichnet wird. Gleichzeitig woll- lich des R-R-Intervalls zwischen Untersuchungs- und ten wir diese neue Überwachungstechnik mit dem her- Kontrollgruppe. Tab. I zeigt die Zusammenfassung des kömmlichen Monitoring mittels Doppler-Schall ver- Datenmaterials, woraus hervorgeht, daß jedoch in der gleichen. Bei 5 Frauen unter der Geburt wurden die Untersuchungsgruppe die Standardabweichung (SD) fetalen ECG-Signale aufgezeichnet, wobei insgesamt der R-R-Intervall-Unterschiede 1.50 bpm beträgt, 3298 sec oder 55 min (das entspricht ungefähr 8000 R- während sie in der Kontrollgruppe bei nur 0.79 bpm R-Intervallen) zur Auswertung kamen. Wir benutzen lag. Tab. II zeigt den Vergleich von Untersuchungseinen FHR-Monitor mit ADF (Perinatronics proto- und Kontrollgruppe hinsichtlich der R-R-Intervalltype) und simultan eine direkte Ableitung vom fetalen Differenzen innerhalb eines Intervalls von einem oder Skalp (FSE). Mit der computergestützten Analyse des zwei Schlägen. Die Ergebnisse verdeutlichen, daß die Datenmaterials wurde ein direkter Vergleich der R-R- noninvasive AFECH-Technik mit ADF zur ÜberwaIntervalle möglich. Die direkt vom Skalp abgeleiteten chung der fetalen Herzfrequenz mit der direkten AbWerte wurden mit den Daten aus dem noninvasiven leitung (FSE) bei Intervallen von l bpm in 92.6% der AFECG verglichen; eine zweite simultane Analyse von beobachteten R-R-Intervalle und bei Intervallen von der direkten Elektrode am Skalp diente als Kontrolle. 2 bpm in 98.2% der R-R-Intervalle übereinstimmt, Die Unterschiede zwischen den fetalen Herzfrequenzen wobei hier, wie oben erwähnt, ca. 8000 R-R-Intervalle bei der „beat to beat"-Registrierung wurden durch bei 5 Patientinnen analysiert wurden. Wie Tab. I und Vergleich von Untersuchungs- und Kontrollgruppen II zeigen, gibt es klinisch keine signifikante Differenz unter Anwendung einer computergestützten Analyse bei der momentan FHR und BTBV zwischen der FSEermittelt. Für die statistische Bearbeitung wurden alle und AFECG-Technik mit ADF. Für weitere Vergleiche Broughtbenutzten to you by |wir University of Arizona Datenmaterial aus 2 ähnlichen Studien, R-R-Intervalle gemessen, die mittlere R-R-IntervallNoninvasive Überwachung der fetalen Herzfrequenz mittels ECG und adaptiver digitaler Filterung

Authenticated Download Date | 6/6/15 2:51 AM

J. Perinat. Med. 20 (1992)

99

Frank et al, Noninvasive FECG FHR monitoring mit der die Genauigkeit des FHR-Monitoring mit Doppler-Schall erfaßt werden sollte. Diese früheren Studien zeigen, daß der Doppier mit Autokorrelation ein Maß für die FHR-BTBV mit einer Irrtumswahrscheinlichkeit von 7.2 bpm (l SD) liefert und daß der Doppler-Schall die FHR-BTBV mit einer Fehlerwahrscheinlichkeit von 9.8 bpm (l SD) liefert (Tab. III und IV). Die AFECG-Überwachungstechnik mit ADF stimmt mit den Werten der direkten Ableitung (FSE) überein, während der Dopplerschall, auch mit Auto-

korrelation nicht so exakt arbeitet. Die Genauigkeit der AFECG-Technik mit ADF erlaubt eine klinisch sinnvolle, noninvasive Bestimmung der BTBV. Diese Studie kann somit zeigen, daß die neue noninvasive AFECG-Technik mit ADF genaue Werte bei der kontinuierlichen Aufzeichnung der aktuellen FHR und BTBV liefern kann. Es sind weitere klinische Studien notwendig, wobei ein größeres Kollektiv mit einer breiteren Variation untersucht werden sollte.

Schlüsselwörter: Adaptive digitale Filterung, „beat to beat"-Variabilität, fetale Herzfrequenz, fetale Überwachung, fetales Elektrocardiogramm.

Resume Surveillance par un filtrage digital ajustable du rythme cardiaque fatal par ECG fcetal non-invasif

Nous rapportons revaluation initiale d'une technique, recemment mise au point, de surveillance du rythme cardiaque foetal, au moyen de PECG foetal, par voie abdominale (AFECG), faisant appel a un filtrage digital ajustable (ADF) pour obtenir, de fagon fiable et non-invasive, la variabilite de battement ä battement (BTBV). Nous presentons, egalement, une comparaison de notre evaluation de cette technique de surveillance du RCF par ECG foetal, par voie abdominale, non invasive, avec les evaluations anterieures des techniques, d'usage courant, de surveillance du RCF, par US Doppier. Les signaux ECG foetaux, totalisant 3298 secondes, ou 55 minutes, (et, comportant, approximativement, 8000 intervalles R-R), ont ete recuillis chez cinq femmes en travail, ä terme, a l'aide d'un moniteur de RCF, prototype, Perinatronics, ayant un ADF, ainsi que, de fagon simultanee, une electrode au scalp foetal (FSE). Une analyse informatisee des donnees fournit une comparaison, une ä une, des paires d'intervalles R-R. Une analyse des donnees par FSE directe, avec une deuxieme analyse simultanee, ä partir de la meme electrode, comme cas controles, a ete comparee avec les donnees de l'AFECG non invasif. La difference entre ces deux rythmes cardiaques foetaux, sur une base de battement ä battement, a ete comparee pour le groupe d'etude et le groupe controle, ä l'aide d'un programme d'analyse informatique. Une analyse statistique a mesure la totalite des intervales R-R, la difference moyenne des intervalles R-R, la deviation standard de la difference des intervalles du groupe etudie et du groupe controle, et a fourni une distribution complete des differences d'intervalles pour des intervalles de l bpm, entre les paires d'enregistrement du RCF. La difference moyenne entre les intervalles R-R des groupes etudie et controle est virtuellement nulle. Un resume des des donnees, dans la tableau I, montre que la deviation standard (DS) du

groupe etudie de la difference des intervalles R-R n'est que de 1,50 bpm, en comparaison de 0,79 bpm, pour le groupe controle. Le tablau II presente une comparaison du groupe controle et du groupe etudie, pour les differences d'intervalles R-R foetaux, a Finterieur des intervalles de un et de deux battements. Ces resultats illustrent que la technique de surveillance non invasive du RCF par AFECG, utilisant un ADF est en accord avec la technique de la FSE directe pour 1 bpm, dans 92,6% des intervalles R-R etudies, et, pour 2 bpm, dans 98,2%, et cela sur, environ, 8000 intervalles R-R foetaux apparies, analyses ä partir des cinq sujets de cette etude. Comme cela est indique dans les tableaux I et II, il n'y a pas de difference, cliniquement significative, entre le RCF instantanee et la BTBV obtenue par la technique de la FSE et par 1'AFECG avec ADF. En comparaison avec notre ensemble de donnees, nous avons passe en revue les distributions des differences d'intervalle R-R, provenant de deux etudes similaires, ayant examine la fiabilite des techniques de surveillance du RCF par echographie Doppler. Ces etudes anterieures ont montre que le Doppler avec autocorrelation fournit une mesure de la BTBV du RCF avec une erreur de 7,2 bpm (l DS) et, que 1'echographie Doppler avec seuil fournit une mesure de la BTBV du RCF avec une erreur de 9,8 bpm (1 DS) (tableaux III et IV). La technique de surveillance de 1'AFECG, avec ADF, est en accord avec la technique de la FSE directe, alors que les techniques Doppler ä seuil ou avec autocorrelation ne sont pas precises. La precision de la technique de 1'AFECG avec ADF lui permet d'etre utile cliniquement pour determiner, de fagon non invasive, la BTBV. La signification de cette etude est qu'une nouvelle technique, non invasive, d'AFECG maternel, a l'aide d'un ADF, peut fournir un enregistrement clinique, continu et precis, du RCF et de la BTBV instantanes. D'autres essais cliniques sont ä realiser sur une population clinique, avant le travail, plus vaste et plus diverse.

Mots-cles: Electrocardiogramme foetal, filtrage digital ajustable, rythme cardiaque foetal, surveillance foetale, variabilite de battement a battement. Brought to you by | University of Arizona J. Perinat. Med. 20 (1992)

Authenticated Download Date | 6/6/15 2:51 AM

100

Frank et al, Noninvasive FECG FHR monitoring

Acknowledgements: This work was supported in part by grants from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, U.S.A., Grant No. HL34898 and the National Science Foundation, Washington, D.C., U.S.A., Grant No. ISI-8860610. Remark: An abstract of this work was presented at the 10th Annual Meeting, Society of Perinatal Obstetricians, Houston, Texas, 1990.

References [1] BOEHM FH, LM FIELDS, JM HUTCHISON, AW BOWEN, WK VAUGHN: The indirectly obtained fetal heart rate: comparison of first-and-secondgeneration electronic fetal monitors. Am J Obstet Gynecol 155 (1986) 10 [2] COURTIN E, W RUCHAY, P SALFELD, H SOMMER: A versatile, semiautomatic fetal monitor for nontechnical users. Hewlett-Packard J (1977) 16 [3] FRANK TH, OR BLAUMANIS, RK GIBBS, RL WELLS: Adaptive filtering in ECG monitoring of the fetal heart rate. J Electrocardiol Suppl (1987) 108 [4] FUKUSHIMA T, CA FLORES, EH HON, EC DAVIDSON: Limitations of autocorrelation in fetal heart rate monitoring. Am J Obstet Gynecol 153 (1985) 685 [5] LAUERSEN NH, HM HOCHBERG, MED GEORGE: Evaluation of the accuracy of a new ultrasound fetal heart rate monitor. Am J Obstet Gynecol 125(1976)1125 [6] LAWSON G, GS DAWES, CWG REDMAN: A comparison of two fetal heart rate ultrasound detector systems. Am J Obstet Gynecol 143 (1982) 840

[7] LAWSON GW, R BELCHER, GS DAWES, CWG REDMAN: A comparison of ultrasound (with autocorrelation) and direct electrocardiogram fetal heart rate detector systems. Am J Obstet Gynecol 147 (1983) 721 [8] LEVENTHAL JM, WU BROWN, JB WEISS, MH ALPER: A new method of fetal heart rate monitoring. Am J Obstet Gynecol 45 (1975) 494 [9] NAGEOTTE MP, RK FREEMAN, AG FREEMAN, W DORCHESTER: Short-term variability assessment from abdominal electrocardiogram during the antepartum period. Am J Obstet Gynecol 145 (1983) 566 [10] PAUL RH, AK SUIDAN, S YEH, BS SCHIFRIN, EH HON: Clinical fetal monitoring VII. The evaluation and significance of intrapartum baseline FHR variability. Am J Obstet Gynecol 123 (1975) 206 [11] SOLUM T: A comparison of three methods for external fetal cardiography. Acta Obstet Gynecol Scand 59 (1980) 123 Received June 5, 1991. Accepted June 18, 1991. Dr. Thomas H. Frank Perinatronics Medical Systems, Inc. 2411 Crofton Lane, Suite 14B Crofton, Maryland 21114 U.S.A.

Brought to you by | University of Arizona Authenticated Download Date | 6/6/15 2:51 AM J. Perinat. Med. 20 (1992)

Noninvasive fetal ECG mode fetal heart rate monitoring by adaptive digital filtering.

Beat-to-beat variability (BTBV) of the fetal heart rate (FHR) is considered an indication of the neural integrity and is an important prognostic indic...
1MB Sizes 0 Downloads 0 Views