Linear and nonlinear analysis of heart rate patterns associated with fetal behavioral states in the antepartum period

• DOI: 10.1016/j.earlhumdev.2006.12.006
• Publication Date: 2007-02-02T18:28:33Z
• Authors: Hernâni Gonçalves; João Bernardes; Ana Paula Rocha; Diogo Ayres-de-Ca...
• Keywords: Adult; Cardiotocography; Heart Rate, Fetal; 03 medical and health sciences; Fetus; 0302 clinical medicine; Nonlinear Dynamics; Pregnancy; Image Interpretation, Computer-Assisted; Linear Models; Humans; Female; Fetal Monitoring

Fetal behavioral states are important indicators of fetal physiology and pathology associated to typical fetal heart rate (FHR) patterns.To provide linear and nonlinear analysis of FHR patterns associated with fetal behavioral states regarding a better understanding of these states and patterns.Fifty FHR tracings from normal term pregnancies with a median duration of 40.3 min were acquired with the SisPorto 2.01 system for computerized analysis of cardiotocograms. Each tracing was divided into consecutive 10-minute segments and each segment was classified by two experts as pattern A, B, C or D.Linear and nonlinear indices were computed in each segment, namely mean FHR, long-term irregularity index (LTI), very low (VLF), low (LF) and high (HF) frequency spectral indices, approximate entropy (ApEn) and sample entropy (SampEn). Kappa statistic (kappa) and proportions of agreement (Pa) were used for assessment of inter-observer agreement. Bootstrap percentile confidence intervals and nonparametric statistical tests were calculated for statistical inference.Overall agreement between experts in pattern classification was good to excellent with values for kappa and Pa of 0.74 (95% CI: 0.64-0.94) and 0.94 (95% CI: 0.92-0.96), respectively. Most linear domain indices increased significantly with rising fetal activity whereas the opposite occurred with nonlinear indices, except for SampEn(2, 0.1). LF/(MF+HF) ratio also significantly increased with fetal activity, denoting an increased sympatho-vagal balance.Results support the hypothesis that entropy and linear variability indices measure different FHR features. FHR patterns associated with active sleep (B) and active wakefulness (D) evidenced more signs of autonomous nervous system activity, with sympatho-vagal imbalance, and less signs related to complexity or irregularity control systems than patterns associated with calm sleep (A) and calm wakefulness (C).