Changes in the PR interval--fetal heart rate relationship of the electrocardiogram during fetal compromise in chronically instrumented sheep

End-to-end trained encoder-decoder convolutional neural network for fetal electrocardiogram signal denoising

OBJECTIVE

Non-invasive fetal electrocardiography has the potential to provide vital information for evaluating the health status of the fetus. However, the low signal-to-noise ratio of the fetal electrocardiogram (ECG) impedes the applicability of the method in clinical practice. Quality improvement of the fetal ECG is of great importance for providing accurate information to enable support in medical decision-making. In this paper we propose the use of artificial intelligence for the task of one-channel fetal ECG enhancement as a post-processing step after maternal ECG suppression.

APPROACH

We propose a deep fully convolutional encoder-decoder framework, learning end-to-end mappings from noise-contaminated fetal ECGs to clean ones. Symmetric skip-layer connections are used between corresponding convolutional and transposed convolutional layers to help recover the signal details.

MAIN RESULTS

Experiments on synthetic data show an average improvement of 7.5 dB in the signal-to-noise ratio (SNR) for input SNRs in the range of  -15 to 15 dB. Application of the method with real signals and subsequent ECG interval analysis demonstrates a root mean square error of 9.9 and 14 ms for the PR and QT intervals, respectively, when compared with simultaneous scalp measurements. The proposed network can achieve substantial noise removal on both synthetic and real data. In cases of highly noise-contaminated signals some morphological features might be unreliably reconstructed.

SIGNIFICANCE

The presented method has the advantage of preserving individual variations in pulse shape and beat-to-beat intervals. Moreover, no prior knowledge on the power spectra of the noise or the pulse locations is required.

A systematic review of cardiac time intervals utilising non-invasive fetal electrocardiogram in normal fetuses

BACKGROUND

Non-invasive fetal electrocardiogram (NIFECG) is an evolving technology in fetal surveillance which is attracting increasing research interest. There is however, only limited data outlining the reference ranges for normal cardiac time intervals (CTIs). The objective of our group was to carry out a systematic review to outline normal fetal CTIs using NIFECG.

METHODS

A systematic review of peer reviewed literature was performed, searching PUBMED,Ovid MEDLINE and EMBASE. The outcomes of interest included fetal CTIs (P wave duration, PR interval, QRS duration and QT interval) and a descriptive summary of relevant studies as well. The outcomes were grouped as early pre-term (≤ 32 weeks), moderate to late pre-term (32-37 weeks) and term (37-41 weeks).

RESULTS

8 studies were identified as suitable for inclusion. Reference ranges of CTIs were generated. Both PR interval and QRS duration demonstrated a linear correlation with advancing gestation. Several studies also demonstrated a reduction in signal acquisition between 27 and 32 weeks due to the attenuation by vernix caseosa. In this group, both the P wave and T waves were difficult to detect due to signal strength and interference.

CONCLUSION

NIFECG demonstrates utility to quantify CTIs in the fetus, particularly at advanced gestations. Larger prospective studies should be directed towards establishing reliable CTIs across various gestations.

Análise da variabilidade da frequência cardíaca em cordeiros da raça Bergamácia, do nascimento aos 35 dias de idade

RESUMO A atividade do sistema nervoso autônomo sobre o coração pode ser verificada pela variabilidade da frequência cardíaca (VFC), método que quantifica e assim infere sobre a modulação autonômica cardíaca, refletindo o funcionamento do sistema nervoso autônomo. Durante o período neonatal, diferenças significativas no que se refere à maturação autonômica são descritas em diversas espécies. Embora a espécie ovina venha sendo utilizada como modelo experimental em diversos protocolos em neonatologia e cardiologia em humanos, estudos descritivos sobre a VFC utilizando animais saudáveis e não sedados são escassos na literatura. O objetivo do estudo foi descrever a VFC em cordeiros da raça Bergamácia durante os primeiros 35 dias de idade. Foram realizados exames eletrocardiográficos em 20 cordeiros da raça Bergamácia a partir do nascimento e semanalmente até o 35° dia de idade. A VFC foi analisada pelo intervalo RR normal (batimentos sinusais), pela frequência cardíaca, pelo índice de tônus vasovagal (iTVV), pela raiz quadrada da média do quadrado das diferenças entre intervalos RR normais adjacentes (RMSSD) e pela raiz quadrada da somatória do quadrado das diferenças dos valores individuais em relação ao valor médio, dividido pelo número de iR-R em 90 segundos - VFC global (RMSM). Todos os parâmetros se alteraram ao longo das semanas. A frequência cardíaca média elevou-se entre o nascimento e os primeiros sete dias de idade, com decréscimo nas quatro semanas subsequentes, sendo o menor valor encontrado aos 35 dias de idade (145,63±37,80bpm). Entre 21, 28 e 35 dias de idade, o iTVV elevou-se significativamente, o RMSM a partir do 28º dia, e, aos 35 dias, o RMSSD, reflexo da ativação parassimpática, exibiu diferenças em relação aos momentos subsequentes. O início da predominância parassimpática, refletida nos índices da VFC, particularmente o iTVV, ocorre aos 21 dias de idade. A partir dos 35 dias de idade, os índices RMSM e RMSSD podem ser utilizados como marcadores fidedignos das mudanças nos efeitos simpático e parassimpático sobre o coração de cordeiros.

Comparison of ECG-based physiological markers for hypoxia in a preterm ovine model

BACKGROUND

Current methods for assessing perinatal hypoxic conditions did not improve infant outcomes. Various waveform-based and interval-based ECG markers have been suggested, but not directly compared. We compare performance of ECG markers in a standardized ovine model for fetal hypoxia.

METHODS

Sixty-nine fetal sheep of 0.7 gestation had ECG recorded 4 h before, during, and 4 h after a 25-min period of umbilical cord occlusion (UCO), leading to severe hypoxia. Various ECG markers were calculated, among which were heart rate (HR), HR-corrected ventricular depolarization/repolarization interval (QTc), and ST-segment analysis (STAN) episodic and baseline rise markers, analogue to clinical STAN device alarms. Performance of interval- and waveform-based ECG markers was assessed by correlating predicted and actual hypoxic/normoxic state.

RESULTS

Of the markers studied, HR and QTc demonstrated high sensitivity (≥86%), specificity (≥96%), and positive predictive value (PPV) (≥86%) and detected hypoxia in ≥90% of fetuses at 4 min after UCO. In contrast, STAN episodic and baseline rise markers displayed low sensitivity (≤20%) and could not detect severe fetal hypoxia in 65 and 28% of the animals, respectively.

CONCLUSION

Interval-based HR and QTc markers could assess the presence of severe hypoxia. Waveform-based STAN episodic and baseline rise markers were ineffective as markers for hypoxia.

Fetal magnetocardiography: development of the fetal cardiac time intervals

OBJECTIVES

To analyse the physiologic development of fetal cardiac time intervals throughout gestation using fetal magnetocardiography (FMCG).

METHODS

FMCG data of 163 uncomplicated pregnancies (19th and 42nd gestational week) were analysed. Mean value, standard deviation, minimum and maximum of the duration of the P-wave, the QRS-complex, the PR and the QT-interval were plotted against gestational age.

RESULTS

QRS-complex, P wave and QT-interval showed a significant lengthening between the 20th and 42nd gestational week. The mean of the QRS complex raised from 36+/-4.7 ms (week 21-24) up to 48+/-5.2 ms (> or =37th week), (p=0.0001). The mean of the P-wave was between 47+/-5.9 ms (week 21-24) and 53+/-9.5 ms (> or =37th week), (p=0.05) and the mean of the QT-interval was 198+/-18 ms (week 21-24) and increased up to 244+/-23.9 ms (> or =37th week), (p=0.009). The PR-interval did not show a correlation with gestational age.

CONCLUSION

FMCG provides sufficient information about all parts of the fetal cardiac conduction system from the 19th gestational week on. It offers the possibility to analyse the shape and the duration of the PQRST-complex.

Ventricular preexcitation sensitive to flecainide in late stage chick embryo ECGs: 2,3,7,8-tetrachlorodibenzo-p-dioxin impairs inotropic but not chronotropic or dromotropic responses to isoproterenol and confers resistance to flecainide

ECGs free of movement artefacts were obtained without anesthesia in 16- to 18-day-old chick embryos close to hatching and used to study the effect of the environmental toxin 2,3,7, 8-tetrachlorodibenzo-p-dioxin (TCDD) on cardiac rhythm and conduction. The ECGs of normal late stage chick embryos exhibited short PR intervals, frequent nonisoelectric PR segments, delta waves, and inverted T waves. Those ECG characteristics are found in patients with the Wolff-Parkinson-White syndrome (WPW) in which they reflect ventricular preexcitation associated with the use of accessory conduction pathways and arrhythmias. Isoproterenol (30 microg/egg) did not alter the ECG preexcitation characteristics. Flecainide, a sodium channel blocker used clinically to suppress WPW accessory pathway activity, at 0.5 to 5 mg per egg diminished the preexcitation and caused atrioventricular (AV) block, supporting the use of accessory pathways together with AV-nodal conduction in normal late stage chick embryos. The findings challenge the dogma that accessory pathways are entirely replaced by AV conduction pathways in late fetal development. TCDD, at 1-2 nmol per egg for 48 h, did not affect heart rate, the increase in heart rate by isoproterenol, or the ECG characteristics, suggesting that short-term TCDD treatment did not affect sinus node function or cardiac conduction. The latter results taken together with prior findings indicate that TCDD differentially impairs the inotropic and lusitropic effects but not the chronotropic or dromotropic effects of isoproterenol. In TCDD-treated embryos, flecainide, tested at 5 mg per egg, caused much less inhibition of preexcitation or production of AV block than in the untreated or solvent-treated controls. The resistance to flecainide represents a new TCDD effect consistent with the reported increase of cardiac myocyte [Ca(2+)](i) by TCDD treatment.

The fetal electrocardiogram: relationship with acidemia at delivery

OBJECTIVE

This study was undertaken to analyze the relationship between the fetal electrocardiogram and arterial pH and base excess at delivery.

STUDY DESIGN

In the labor wards of three teaching hospitals a retrospective observational study was conducted on fetal monitor recordings of 679 women for whom there was an indication for continuous fetal monitoring during labor. These women had been recruited as part of either observational studies or a prospective randomized trial related to the Nottingham fetal electrocardiographic project. Fetal heart and uterine contraction data were obtained with the Nottingham fetal electrocardiographic analyzer. Morphologic and time interval analyses of the fetal electrocardiogram were performed. Evaluation was carried out for the last half hour before delivery. Main outcome measures were umbilical arterial pH and base excess at delivery.

RESULTS

The study demonstrated a relationship between time interval analysis of the fetal electrocardiogram and a low umbilical arterial pH and base excess at delivery. Analysis of the morphologic characteristics of the fetal electrocardiogram (ST segment and T-wave height) showed no significant relationship.

CONCLUSIONS

Time interval analysis of the fetal electrocardiogram during labor is related to relative acidemia at delivery.

The PR interval-fetal heart rate relationship during repetitive umbilical cord occlusions in immature fetal sheep

OBJECTIVE

To evaluate the relationship of the PR interval and fetal heart rate during repetitive umbilical cord occlusions in immature sheep fetuses.

STUDY DESIGN

In seven chronically cannulated immature sheep fetuses [gestational age 90.6 days (mean)], we analyzed continuous fetal electrocardiogram recordings during repetitive cord occlusions for 2 out of every 5 min until fetal mean arterial pressure dropped to 50% of baseline value. PR interval-fetal heart rate correlation coefficients (Pearson) was measured on consecutive blocks of 2.5 min. R-values of the baseline and the repetitive occlusion period were compared by Fisher's exact test.

RESULTS

Repetitive cord occlusions resulted in acidosis and hypotension. Two fetuses died at the end of the repetitive occlusion period. Four out of seven fetuses showed a significant change from a negative relationship between the PR interval and fetal heart rate during baseline to a predominantly positive relationship during the repetitive occlusion period.

CONCLUSION

In immature fetal sheep, a change from a negative relationship between the PR interval and fetal heart rate to a predominantly positive relationship between the PR interval and fetal heart rate was observed in four out of seven fetuses following the initiation of repetitive umbilical cord occlusions.

Cardiotocography only versus cardiotocography plus PR-interval analysis in intrapartum surveillance: a randomised, multicentre trial. FECG Study Group

BACKGROUND

There is a need to improve the sensitivity and specificity of fetal monitoring during labour. We compared the gold standard, cardiotocography, with cardiotocography plus time-interval analysis of the fetal electrocardiogram in fetal surveillance. The aim was to find out whether time-interval analysis decreased the need for operative intervention due to fetal distress.

METHODS

We did a randomised, prospective trial in five hospitals in the UK, Hong Kong, the Netherlands, and Singapore. 1038 women undergoing high-risk labours were randomly assigned fetal monitoring by cardiotocography alone, or cardiotocography plus fetal electrocardiography (ECG). Outcomes measured were rates of operative intervention, and neonatal outcome. Analysis was by intention to treat.

FINDINGS

515 women were assigned management by cardiotocography, and 523 cardiotocography plus fetal ECG. There was a trend towards fewer operative interventions for presumed fetal distress in the time-interval analysis plus cardiotocography group (63 [13%] vs 78 [16%]), but this was not significant (relative risk 0.80 [95% CI 0.59-1.08], p=0.17). There was no significant difference between groups in the proportion of babies who had an umbilical arterial pH of 7.15 or less (51 [11%] vs 49 [11%]; 1.01 [0.7-1.47]), or in the frequency of unsuspected acidaemia (42 [9%] vs 35 [8%]; 1.17 [0.76-1.79]).

INTERPRETATION

The addition of time-interval analysis of the fetal electrocardiogram during labour did not show a significant benefit in decreasing operative intervention. There was no significant difference in neonatal outcome.

Do fetal electrocardiogram PR-RR changes reflect progressive asphyxia after repeated umbilical cord occlusion in fetal sheep?

The aim of this study was to determine whether there is a relationship between changes in PR-RR correlation of the fetal ECG and progressive changes in fetal acid-base status and blood pressure (BP) during repeated umbilical occlusion. Chronically instrumented fetal sheep at 126.8+/-0.6 d (mean+/-SEM) were randomized to receive 1 min of total umbilical cord occlusion either every 5 min for 4 h (1:5 group; n=8), or every 2.5 min until BP fell <2.7 kPa (20 mm Hg) on two successive occlusions (1:2.5 group; n=8). The PR-RR correlation was determined in 5- or 2.5-min intervals. Umbilical cord occlusion caused variable decelerations with initial sustained hypertension. In the 1:5 group BP remained elevated throughout, and there was little change in acid-base status (pH=7.34+/-0.07, base deficit=1.3+/-3.9 after 4 h). In contrast, after the third occlusion the 1:2.5 group showed progressive hypotension during occlusions, and severe progressive metabolic acidemia (pH 6.92+/-0.1, base deficit 17.0+/-4.7 mmol/L after the last occlusion). In both groups, the PR-RR relationship switched from positive to negative with the onset of occlusions, then reverted to positive after a variable interval. In the 1:2.5 group later reversion of the PR-RR to positive was associated with earlier and more prolonged hypotension during the middle and end of the occlusion series (p < 0.001). We conclude that the initial switch to a negative PR-RR relationship during repetitive umbilical occlusion was due to a reflex-mediated response unrelated to fetal acidosis or hypotension. Both stable well compensated fetuses and severely hypoxic, hypotensive fetuses subsequently showed a positive PR-RR correlation.

The fetal electrocardiogram

Advances in microprocessing technology have made fetal ECG analysis a feasible adjunct to fetal surveillance. Time interval and morphology changes of the FECG occur during fetal hypoxia. The use of these changes to detect a fetus at risk of intrapartum asphyxia awaits validation in terms of both future and ongoing clinical trials. Recognition of FECG changes during decelerations may improve the sensitivity of EFM. Antepartum FECG analysis has potential for the detection of a number of pathological fetal conditions, including intrauterine growth retardation, but remains hampered by low signal-to-noise ratios, rendering successful signal acquisition unreliable.