Computerized analysis of fetal heart rate variability using the matching pursuit technique as an indicator of fetal hypoxia during labor

Abstract

Objective. To determine whether the computerized analysis of fetal heart rate variability with the new matching pursuit technique can indicate fetal distress during labor.

Study design. Eighty women were studied during the intrapartum period with external cardiotocography. In all cases, cord arterial pH and 5-min Apgar Scores were evaluated. Six cases that presented large segments of missing data were excluded from the study. The remaining 74 women were divided into two groups; 32 women with normal (Group A) and 42 women with non-reassuring FHR tracings (group B). Group B was divided in subgroup BI, including 24 women with pH > 7.20, and BII, including 18 women with pH < 7.20. In order to evaluate the FHR fluctuations, in different frequency ranges, we applied an adaptive time-frequency method, called Matching Pursuit. We estimated the power of the FHR signal in four frequency ranges.

Results. The 5-min Apgar Scores were significantly lower in both subgroup BI and subgroup BII (p = 0.003 and p = 0.003 respectively). The Low Low Frequency (LLF) parameter appears to recognize better the cases with lower pH (sensitivity 78.5%, specificity 52.3%) than the cases with non-reassuring FHR (66.6%, 56.2). The sensitivity and specificity of the Very Low Frequency (VLF) parameter were 72.2% and 59% respectively in recognizing the cases with lower pH and 64.2% and 53.1% in recognizing non-reassuring FHR.

Conclusion. Fetal hypoxia during labor can be recognized using the MP technique for the analysis of FHR signal power in the VLF and LLF frequency ranges. Since the analysis is feasible in real-time, it can be a useful tool for the intrapartum evaluation of fetal well-being.

Keywords:

• Matching pursuits
• fetal heart rate variability
• fetal hypoxia