COMPUTERIZED ANALYSIS OF FETAL HEART-RATE .1. DESCRIPTION OF THE SYSTEM (2CTG)

Objective: The object of this study was to develop an automatic system for the on-line analysis of fetal heart rate based on a low cost personal computer and allowing a reliable identification of both baseline and fetal heart rate patterns. Methods: An on-line method of baseline identification has been constructed on the implementation of the off-line method described by Mantel et al.6 Based on the determination of the baseline, accelerations and decelerations were identified, short- and long-term variability indices were measured, and fetal heart rate patterns (A to D) were identified. The results of the baseline fitting were evaluated on 20 tracings lasting 2 h. The on-line method developed was compared with the original off-line method by considering the agreement of a panel of five experts on baseline fitting and by regression analysis between the absolute values of baseline obtained. Tracings (169) lasting greater-than-or-equal-to 60 min from uneventful pregnancies at 31-42 weeks' gestation were also evaluated and the relationships between the fetal heart parameters were measured and both gestational age and fetal heart rate patterns were analyzed. Results: The panel of experts judged that in 70% of the cases, both methods fitted the baseline in an equivalent way. In the remaining 30% of the cases, experts frequently disagreed on considering the best fitting methods. Regression analysis showed a high coefficient of correlation (r = 0.92, P less-than-or-equal-to 0.001) between the baseline estimation of the two methods, with a maximum difference in baseline values of 4 beats per min (bpm) (median 2 bpm). Significant changes of baseline and the number and amplitude of accelerations were evidenced with advancing gestation. Significantly lower values of fetal heart variability were found during fetal heart rate pattern A when compared with pattern B. Conclusions: The automatic system of fetal heart analysis developed provides a reliable on-line identification of the baseline, a detection of accelerations and decelerations, and a measurement of fetal heart rate variability. Finally, by identifying the fetal heart rate patterns, it allows the study of fetal heart rate variability in relation to fetal behavior, a tool that may further validate this technique.