Limitations of autocorrelation in fetal heart rate monitoring

Fetal heart rate monitors that use autocorrelation of the ultrasonic fetal signal usually produce a cleaner fetal heart rate record than that obtainable with conventional ultrasonic fetal monitors. However, since the autocorrelation function will emphasize any periodic signal originating from the fetus or the mother, in clinical situations the resultant fetal heart rate tracing may contain spurious data. To illustrate the limitations of the autocorrelation technique in fetal monitoring, we compared the autocorrelated ultrasound fetal heart rate records from 23 patients in active labor with the simultaneously recorded direct scalp fetal electrocardiogram tracings. The results indicate that every hour of recording contained, on the average, five instances in which data were missing (range, 0 to 13), four in which data were added (range, 0 to 23), and seven instances in which data were absent for greater than 1 minute (range, 0 to 26). The potential problem of misinterpretation of autocorrelated fetal heart rate data is discussed.