Effectiveness of ambulatory non-invasive fetal electrocardiography: impact of maternal and fetal characteristics

Abstract

Introduction: Non-invasive fetal electrocardiography (NIFECG) has potential benefits over the computerized cardiotocography (cCTG) that may permit its development in remote fetal heart-rate monitoring. Our study aims to compare signal quality and heart-rate detection from a novel self-applicable NIFECG monitor against the cCTG, and evaluate the impact of maternal and fetal characteristics on both devices.

Material and methods: This prospective observational study took place in a university hospital in London. Women with a singleton pregnancy from 28 + 0 weeks' gestation presenting for cCTG were eligible. Concurrent monitoring with both NIFECG and cCTG were performed for up to 60 minutes. Post-processing of NIFECG produced signal loss, computed in both 0.25 (E240)- and 3.75 (E16)-second epochs, and fetal heart-rate and maternal heart-rate values. cCTG signal loss was calculated in 3.75-second epochs. Accuracy and precision analysis of 0.25-second epochal fetal heart-rate and maternal heart-rate were compared between the two devices. Multiple regression analyses were performed to assess the impact of maternal and fetal characteristics on signal loss.

Clinicaltrials: gov Identifier: NCT04941534.

Results: 285 women underwent concurrent monitoring. For fetal heart-rate, mean bias, precision and 95% limits of agreement were 0.1 beats per minute (bpm), 4.5 bpm and -8.7 bpm to 8.8 bpm, respectively. For maternal heart-rate, these results were -0.4 bpm, 3.3 bpm and -7.0 to 6.2 bpm, respectively. Median NIFECG E240 and E16 signal loss was 32.0% (interquartile range [IQR] 6.5%-68.5%) and 17.3% (IQR 1.8%-49.0%), respectively. E16 cCTG signal loss was 1.0% (IQR 0.0%-3.0%). For NIFECG, gestational age was negatively associated with signal loss (beta = -2.91, 95% CI -3.69 to -2.12, P < 0.001). Increased body mass index, fetal movements and lower gestational age were all associated with cCTG signal loss (beta = 0.30, 95% CI 0.17-0.43, P < 0.001; beta = 0.03, 95% CI 0.01-0.05, P = 0.014; and beta = -0.28, 95% CI -0.51 to -0.05, P = 0.017, respectively).

Conclusions: Although NIFECG is complicated by higher signal loss, it does not appear to be influenced by increased body mass index or fetal movement. NIFECG signal loss varies according to method of computation, and standards of signal acceptability need to be defined according to the ability of the device to produce clinically reliable physiological indices. The high accuracy of heart-rate indices is promising for NIFECG usage in the remote setting.

Keywords: ambulatory monitoring; computerized cardiotocography; fetal heart rate monitoring; non-invasive fetal electrocardiography; signal loss; signal quality.

FIGURE 1

The novel, self‐applicable non‐invasive fetal electrocardiography (femom) placed on the maternal abdomen.

FIGURE 2

Raw non‐invasive fetal electrocardiography trace after noise filtering and processing, displayed on the four channels as derived from PC‐based software. Maternal (mpeak) and fetal (fpeak) R waves are marked by blue and red dots, respectively.

FIGURE 3

Raw non‐invasive fetal electrocardiography trace with high amplitude noise (electrical interference). Maternal (mpeak) and fetal (fpeak) R waves are marked by blue and red dots, respectively, but fetal R waves are evidently difficult to identify as masked by noise.

FIGURE 4

Scattergram showing signal loss (%) on the Y‐axis, against gestational age (GA) (weeks) on the X‐axis, in both non‐invasive fetal electrocardiography (NIFECG) (E16 signal loss) and cCTG. Empty blue circles depict overall NIFECG signal loss, and filled red diamonds are cCTG signal loss. The line of best fit in the NIFECG group has a stronger negative correlation with gestational age compared with the cCTG group.