[Computer-assisted analysis of intrapartum cardiotocodynagraphs]

Abstract

Objective: To develop a numerical index by quantifying characteristic cardiotocography parameters in order to determine the intrauterine state of the fetus sub partu with sufficient accuracy to improve the diagnostic potential of cardiotocography. This score is then computerized for offline and online analysis.

Methods: The last 120 minutes of 406 directly recorded intrapartal cardiotocograms (CTG) were evaluated with a magnifying glass on the basis of conventional criteria. The mean basal frequency, the amplitude and the oscillation frequency of each individual baseline minute, the decelerations with depth, duration and type classification as well as the accelerations were determined. The Apgar score and the measurements of blood gases in the umbilical arterial and venous blood served as outcome parameters. The actual pH in the umbilical artery (UA) was the key variable. The newly developed CTG score was converted into a complex programming language after meticulous statistical testing and prepared for offline and online use.

Results: A score which is based on the CTG characteristics basal frequency level, baseline fluctuation and total decelerations area was developed on the basis of about 40 000 baseline minutes and more than 8 000 analysed decelerations from 406 fetuses which were all vaginal deliveries with a rate of acidosis (pH,UA < 7.100) of 9.8 % (deliberate selection). This score correlates highly significantly (r = -0.5752) with the actual umbilical arterial pH (p < 0.001). The correlation with the Apgar score (1 min) is r = -0.398, p < 0.001. The positive likelihood ratio (score 9) is 11.0 (95 % confidence interval 7.4-16.5) indicating that the test is very reliable. The score could be prepared for offline analysis of electronically registered CTGs with the assistance of a mathematician, so that the obstetrician's "eye work" can be dispensed with entirely. This program written in "C" also enables online evaluation of fetal heart rate patterns over time and a representation of the output parameters with acoustic and optical warning display options appropriate for the delivery room.

Conclusions: Quantitative registration of fetal heart rate phenomena during delivery can be accomplished successfully by means of a score. The score correlates highly significantly with the actual pH value (UA) and the Apgar score (1 min). It is suitable both for offline and online analysis of intrapartal CTGs, and offers the prospect of quantitative fetal online monitoring.