Spatial filtering of the fetal electrocardiogram

One of the basic problems in the analysis of the rhythm and the shape of the fetal ECG (FECG) is the interference caused by the electrical activity of the maternal heart (MECG). Methods devised for suppressing the maternal signals in the abdominally recorded leads obviously have to rely on differences in the characteristics of both kinds of signals. In the past various differences have been employed for this purpose such as differences in wave shape, frequency content and source location. The method presented here is based on the latter principle. The electrical sources of the fetal and maternal heart are clearly separated in space. By employing multiple abdominal leads, a linear combination of the recorded signals can be formed which acts effectively as a spatial filter. As such the filtering characteristics depend exclusively on the separation inside the human body (the electrical volume conductor) and, as such, can be expected to be independent of the temporal aspects of the recorded signals. In particular, the obtained filter is unaffected by the occurrence of ectopic beats in either the fetal or the maternal heart. The determination of the required filter (weighting coefficients) is carried out by using the technique of singular value decomposition (SVD) of the data matrix representing the sampled input signals. The filter effectively suppresses the maternal component in the filter output and optimizes the fetal component. The properties of this filter are demonstrated.     

Maternal and fetal ECG alterations during and after infusion of fenoterol

Maternal and fetal ECG alterations in a series of patients infused with fenoterol during the early stage of parturition have been investigated. Analysis of MECG revealed changes similar to those found during sympathetic stimulation, i.e. an early and steep ST rise and an increase or decrease in T as a sign of sympathetically altered de- and repolarization of the myocardium. FECG alterations were less impressive, probably due to a reduced placental transfer of fenoterol from the mother to the fetus.     

Variability analysis of fetal heart rate signals as obtained from abdominal electrocardiographic recordings

The present paper introduces an original method of digital signal processing for an automatic analysis of non-invasive abdominal ECG recordings on pregnant women starting from the 25th week of gestation. The procedure has been implemented on a DEC-VAX 750 digital computer at the Department of Electrical Engineering, Polytechnic of Milano and the signals are recorded at the Department of Obstetrics and Gynecology "L. Mangiagalli", University of Milano, Italy. The experimental results presented in here are still preliminary as only few cases have been considered up to now (about 20) and the goal of the paper is mainly focused on the algorithmic aspects of the whole procedure implemented in the computer and on the approach of heart rate variability (HRV) signal analysis both in the mother and in the fetus. Abdominal ECG lead processing is illustrated starting from the step of maternal (M) and fetal (F) QRS recognitions through linear digital filtering (derivative and low-pass FIR filter, Weber-Cappellini window) and weighted averaging techniques synchronized with maternal QRS's. Figure 1 a shows the original abdominal lead; figure 1 b the filtered signal for MQRS recognitions; figure 2 a the template of maternal cardiac cycle as obtained after the averaging operation synchronized with the instants of MQRS occurrence. The subtraction of the template results in the abdominal lead shown in figure 1 c in which the contribution of MECG is practically entirely reduced even in the case of MQRS and FQRS overlapping.(ABSTRACT TRUNCATED AT 250 WORDS)     

Technical progress in fetal electrocardiography--a review

The fetal electrocardiogram (FECG) was first demonstrated 80 years ago, but initial progress in research into the subject was slow and limited by the technical equipment available. As improved amplifiers became available, the detection of the waveform became easier, but observation of waveform morphology was still difficult because of background noise. The signal-to-noise ratio of the original FECG waveform was improved considerably with the use of directly applied fetal electrodes, and the development of digital computer techniques has allowed signal enhancement. It is now possible to monitor the FECG waveform continuously in labor and reliably produce waveforms with all of the components clearly defined.     

Microprocessor based waveform analysis of the fetal electrocardiogram during labor

An increase in T wave amplitude of the fetal ECG (FECG) has experimentally been correlated to elevated catecholamine levels and myocardial glycogenolysis. The FECG changes have also been described during human delivery. The present aim was to investigate whether these ECG changes could be reproduced in an easily handled real time microprocessor system, and to correlate them to biochemical and clinical data. During 40 deliveries the FECG signal was transferred to a microcomputer system for real time averaging of the FECG. There was a high capacity of the system to reproduce the ST waveform changes though the averaging procedure reduced the QRS magnitude by 10%. With a normal umbilical artery pH (greater than or equal to 7.25) the highest T/QRS ratios recorded during each delivery was 0.26 +/- 0.19 (mean +/- S.D.). With lowered pH (less than 7.25) the T/QRS increased to 0.33 +/- 0.10 (P less than 0.02). A similar difference between the two groups was seen when the T/QRS ratios from the last hour before birth were compared; 0.13 +/- 0.08 and 0.18 +/- 0.05, respectively (P less than 0.01). Changes in the ST waveform with T/QRS greater than or equal to 0.30, ST segment alterations, or negative T waves appeared during 40% of the deliveries, however, 30% were short standing changes (less than 30 min). Intermediate CTG changes during at least 30 min occurred in 41% and the pattern was classified as abnormal in 18%. Using the scalp electrode as signal source, the ECG analysis could add further information to the routine CTG recording on the fetal condition during delivery.     

The potential distribution generated by the fetal heart at the maternal abdomen

The pathways along which the electrical currents generated by the fetal heart are conducted to the surface of the maternal abdomen are not known. As a consequence, in recording the fetal electrocardiogram (FECG) it is hard to predict where electrodes should be placed in order to obtain an optimal signal. The amplitude of the FECG varies with gestation, and there is a large interindividual variability in the amplitude of the FECG and in the optimal recording site among subjects within the same gestational age. Attempts have been made to explain these phenomena in terms of volume conduction. In this research the complete potential distribution on the maternal abdomen is studied in connection with the geometrical configuration of the electrical source (fetal heart) and the volume conductor (surrounding tissues). For a small group of pregnant women the abdominal FECG is recorded simultaneously in 32 leads during a period of about one minute, once every two weeks from 20 weeks of gestation onwards. A spatial filtering technique which combines information of all 32 leads is used to provide a trigger of the fetal QRS complexes. Using this trigger, an average fetal complex is constructed for each lead by time coherent averaging, after subtraction of the maternal contribution. These average fetal complexes are combined to plot the complete potential distribution generated by the fetal heart at the maternal abdomen (fetal body surface map, FBSM) at any given time instant during the fetal cardiac cycle. At these recording sessions the geometry is carefully quantified by making transverse scans every 2 cm with a compound echo scanner. The contours of fetal head and body, the placenta and the uterus are manually drawn on hardcopies of the video display images. Real time echoscopy is used to support the identification of the geometry. The contours are fed into a computer using a graphics tablet. The three dimensional surfaces of fetus, placenta and uterus are separately represented by a triangulation of the respective contour lines. Figures 5 and 6 show an example of the triangulated representation of the recorded geometry. Figure 7 shows the average fetal complexes of an individual at 26 weeks of gestation, plotted at the site where they have been recorded.(ABSTRACT TRUNCATED AT 400 WORDS)     

Intrapartum fetal monitoring by electrocardiography

To study the diagnostic values of intrapartum fetal electrocardiography (FECG) and cardiotocography (CTG), 68 patients in labor were monitored at random. One-min Apgar score was chosen as the gold standard. The specificity (87.1%) and accuracy (86.8%) of FECG were found better than those of CTG (54.5% and 57.4% respectively). Fetal hypoxia and acidosis was firstly manifested by shortened P-R interval, FHR deceleration and increase in the T wave amplitude. It suggested that FECG can be a reliable diagnostic method following CTG screening.     

A case of fetal complete heart block recorded by magnetocardiography, ultrasonography and direct fetal electrocardiography

Fetal magnetocardiograms (FMCGs) were recorded in a case of fetal complete heart block (CHB) from the 30th to the 37th week of gestation using the multichannel SQUID system (Hitachi, Japan). M-mode ultrasonography and direct fetal electrocardiography using needle electrodes revealed fetal CHB. We identified independent fetal P-waves and QRS complexes in the FMCG recorded in the 32nd week of gestation when the fetal atriums were close to the FMCG sensor. We also recorded FMCG P-waves in the 37th week of gestation when the fetal heart was larger. Fetal heart position and size are important for obtaining a useful FMCG. To establish FMCG as a diagnostic tool of fetal arrhythmia, comparative studies with FECG are needed.     

Fetal electrocardiographic monitoring during labor in relation to cord blood levels of the brain-injury marker protein S-100

BACKGROUND: Cord artery protein S-100 levels at birth are potential markers of brain damage after asphyxia. Our aim was to investigate if S-100 levels were elevated in neonates with indirect signs of asphyxia during birth. S-100 levels in cord blood were studied in relation to cardiotocography (CTG) and fetal electrocardiography (FECG) changes during birth and to acidemia in umbilical blood. MATERIAL AND METHODS: This case-control study was performed in parallel to a large randomized controlled trial (RCT) studying FECG at birth. Protein S-100 samples were collected from 103 neonates at birth and related to the CTG and ECG changes during labor and to pH in umbilical blood. RESULTS: Protein S-100 was significantly higher in neonates with umbilical artery blood pH7.05. Furthermore, neonates with preterminal CTG patterns showed increased S-100 levels compared to neonates with normal CTG. Neonates having significant CTG and ECG changes, leading to intervention according to clinical guidelines, showed significantly higher S-100 levels compared to neonates without such indication of intervention. CONCLUSION: A relation exists between S-100 in umbilical blood at birth, acidosis and pathological patterns in CTG and FECG during labor.     

The relationship between intrapartum fetal heart rate disturbances, electrocardiographic changes and umbilical venous troponin-I (T-I)

OBJECTIVE: To analyse the relationship between intrapartum fetal heart rate disturbances and electrocardiographic changes and umbilical venous troponin-I (T-I), concentrations as well. MATERIAL AND METHODS: 14 fetuses were continuously surveyed by CTG and ECG recordings in the first and second stage of labor, using STAN S-21 analyser. After birth, umbilical venous blood samples were collected for determination of acid-base balance, base excess and troponin-I concentrations. RESULTS: From among monitored fetuses, in 8 cases the CTG and FECG patterns were correct. Vaginally delivered neonates were born in good clinical status, with normal acid-base balance, base excess and T-I < 0.3 ng/ml. In 3 cases abnormal CTG patterns were observed with early decelerations but FECG patterns were correct. Vaginally delivered neonates were born in good clinical status, with normal acid-base balance, base excess and T-I < 0.3 ng/ml. In 2 cases abnormal CTG patterns were observed with variable decelerations but FECG patterns were correct. Vaginally delivered neonates were born in good clinical status, with normal acid-base balance, base excess and T-I < 0.3 ng/ml. In one case abnormal CTG pattern were observed with late decelerations but FECG patterns was correct. The pregnancy was terminated by caesarean section because of fetal distress. The neonate was born in good clinical status with normal acid-base balance and base excess but T-I concentration was increased (1.5 ng/ml).     

胎儿心电图诊断胎儿宫内窘迫的临床价值

本文总结了>36孕周高危妊娠的胎儿心电图(FECG)共30例。同时分别作NST和B超监测,并以上述两项结果作对照。经过分析认为:FECG是反映胎儿在于宫内是否正常的一种客观检查指标。异常FECG提示胎儿宫內缺氧,FQRS>0.05秒或FST段改变时要注意脐带绕颈和羊水过少;而FQRS振幅>30μV时要注意羊水过少或胎儿体重≥3500g,并认为FQRS>0.05秒,振幅>30μV及FST段改变,可作为异常FECG的重要指标。其检出胎儿异常率和敏感性与NST和B超比较有显著差异(P<0.01)。     

Heart rate variability and electrocardiogram changes in the fetal lamb during hypoxia and beta-adrenoceptor stimulation

Hypoxic changes in the ST waveform of the fetal electrocardiogram (FECG), with elevated T waves as the main response, have earlier been described in the fetal lamb as mediated via the beta-adrenoceptor cells, initiated by catecholamine release. A similar background to the increase in fetal heart rate variability (FHRV) during hypoxemia has been suggested. The aim of the present study was to elucidate this question and also to compare FHRV and changes in the ST waveform of the FECG as indicators of fetal distress. Twenty-six acutely exteriorized mature lamb fetuses were submitted to periods of graded hypoxemia. Fetal blood gases were measured and oxygen content was calculated. The FHRV was analyzed by a computer program and calculated as the differential index (DI) and the interval index (II). Eighty seconds of the recorded ECG coinciding in time with each fetal blood sample were analyzed using a PDP 11/40 minicomputer. The ECG signal was sampled at a rate of 1250 samples per second giving a resolution in time of 0.8 msec. Each QRS complex was detected with a semi-automatic program using a cross-correlation algorithm. QRS complexes that by the program were signalled to be doubtful were visually examined and either rejected or approved. Hence, the resulting list of R-R intervals was practically free from artefacts. From this list of R-R intervals the DI and II were computed according to YEH et al. [30]. Hypoxemia resulted in initially strongly increased DI from 3.1 +/- 2.5 to 17.7 +/- 13.8 (p less than 0.001) and in II from 1.88 +/- 0.65 to 3.77 +/- 2.06 (p less than 0.001) (Fig. 1). Obviously the change in oxygen content per se was strongly associated with the variability indices, as we found a strong correlation between delta oxygen content/min and delta DI/min (r = 0.81). In five fetuses the effect of prolonged hypoxemia on DI was studied (Fig. 2). After the initial rise from 1.6 +/- 1.1 to 31.2 /+- 9.0 (p less than 0.02) DI decreased to 19.4 +/- 20.7. A regression analysis showed a strong connection between DI and PaO2, (n = 93, T = - 5.34), whether there was no relation between DI and pH (T = - 1.85). There is strong evidence that hypoxemia and asphyxia induce an increase in the concentration of catecholamines in fetal blood [2, 23].(ABSTRACT TRUNCATED AT 400 WORDS)     

胎儿心电图ST段分析及其应用

胎儿心电图(fetal electrocrdiogram,FECG)是围生期胎儿监护的一种有效手段,随着其临床应用,所存在的不足也日渐显露。为减少其不足带来的弊端,用计算机进行FECG的ST段分析即STAN(ST analysis,STAN)应运而生。STAN的应用需要按照其进行的前提条件,确保信号质量,同时结合胎心率(fetal heart rate,FHR)监护的结果。在应用的过程中,根据STAN的显示以及连续性作出相应处理,及时提示胎儿宫内缺氧,以降低围生儿发病率及死亡率。     

Fetal electrocardiogram obtained with subcutaneous maternal electrodes. A new method

The low voltage of the fetal heart signals and the usual background interference are the main obstacles in obtaining clear fetal electrocardiographic tracings with a high signal-to-noise ratio. By the use of subcutaneous needle electrodes, impulses from the maternal abdominal wall can be short-circuited. Fetal heart potentials with higher amplitude and less background noise are thus recorded. Fetal viability, presentation, multiple pregnancy, and congenital arrhythmias were diagnosed antenatally by means of the fetal ECG with subcutaneous electrodes and were later correlated with the outcome of the pregnancy. More than two hundred tracings were performed with 98 per cent accuracy.     

Interpretation of the fetal ECG during labor: the effect of uterine contractions

This study was performed in order to investigate the fetal electrocardiogram (FECG) during uterine contractions associated with normal labor. Twenty-five patients with low risk pregnancy between 38-41 weeks gestation were studied during the active stage of labor. Both FECG and intra-uterine pressure are obtained in a conventional manner and are continually sampled into the computer. The FECG is averaged point-to-point, synchronized to the peak of the R-wave. This is performed by a QRS detection algorithm which is based on a digital analysis of slope, amplitude and width. A digital band-pass filter composed of cascaded high-pass and low-pass filters reduces false detections and permits the use of auto-adjustable low thresholds. A separate averaging is performed on the T-wave in order to prevent attenuation due to variable R-T interval. The T wave is subsequently aligned in time and position to the rest of the QRS complex. A significant increase was observed in the T/QRS amplitude ratio during the first half of the uterine contraction. Such an increase was also observed in the short and long-term FHR variability. No significant changes were observed in the other components of the FECG. In conclusion, by implementing a computer based system it is possible to analyse the FECG during labor. Based on this and previous studies it may well prove to be a sensitive indicator of fetal condition.     

On-line processing of the fetal electrocardiogram. A new direction for fetal monitoring

Digital filtering techniques can be applied to the recovery of the fetal electrocardiogram (ECG) from noise, using optimized digital filters matched to the frequency characteristics of the waveform. By forming a linear model of the fetal ECG, it is possible to measure time constants and characteristics on a real-time basis. The technique provides a powerful tool for the detection of subtle changes in the ECG, but the interpretation of these changes is critical to the application of the technique in fetal monitoring. Shifts in the ST segment and changes in T wave configuration are late indicators of fetal asphyxia, whereas the earliest indicators of stress are shortening of the PR interval and inversion of the normal positive relationship between the PR and RR intervals. The method appears to substantially enhance our ability to predict fetal stress at an early stage.     

Fetal electrocardiogram: ST waveform analysis in intrapartum surveillance

ST waveform analysis of fetal electrocardiogram (ECG) for intrapartum surveillance (STAN) is a newly introduced method for fetal surveillance. The purpose of this commentary is to assist in the proper use of fetal ECG in combination with cardiotocography (CTG) during labour. Guidelines and recommendations concerning CTG and ST waveform interpretation and classification are stated that were agreed on by the European experts on ST waveform analysis for intrapartum surveillance during a meeting in Utretcht, the Netherlands in January 2007.     

Techniques for the routine on-line processing of the fetal electrocardiogram

In order to routinely monitor the fetal electrocardiogram (FECG), during labor, in an on-line real-time situation, the following processes need to be undertaken: The FECG signal must be recovered from noise. The recovered signal or waveform must be measured with a high degree of precision in order to generate the timing intervals, areas and parameters that are of interest to the clinician. The FECG and measured parameters must be presented to the clinician in a meaningful and simple format. Control of the machine and its processing should be a user friendly operation. In the current work, a system is described that uses digital filtering to recover the FECG waveform from low frequency biological noise and associated artifacts. A linear model of the FECG waveform is used to evaluate some 18 different timing intervals of the FECG. The enhanced waveform and accompanying parameters are displayed upon an intelligent graphics terminal to achieve a user friendly operation of the system by clinical staff.     

Fetal ECG improvement due to selective signal addition

Electrocardiographic examination of cardiac potentials about the maternal torso has previously revealed fetal cardiac signals of relatively similar configuration and maternl cardiac signals of relatively dissimilar configuration. A selective combination of signals obtained from several abdominal sites largely attenuates random noise and the maternal ECG due to averaging of these incoherent signals. The relative improvement in fetal signal amplitude resulting from selective addition is demonstrated by obtaining a useful fetal ECG from maternal abdominal sites where signals were devoid of overt fetal energy.     

Diagnosis of congenital mitral and aortic stenosis from the fetal electrocardiogram.

A case of congenital mitral and aortic stenosis diagnosed from the fetal electrocardiogram (FECG) during labor is presented. Cardiologic work-up and autopsy findings in the neonate confirmed the diagnosis. This is the first case in which prenatal diagnosis of a specific anatomical lesion of the heart was made from the FECG and confirmed clinically as well as at autopsy.