Conventional linear criteria with sample entropy in the interpretation of reactive antepartum fetal heart rate tracings.

Objective: To compare the test duration times and rate of nonreactive results between the conventional linear reactive criteria (CLRC) and the modified nonlinear reactive criteria (MNRC) in electronic fetal heart rate (FHR) monitoring. The MNRC are the CLRC with the addition of approximate entropy or sample entropy.

Methods: One thousand women with singleton pregnancies between 30 and 37 weeks' gestation were selected. They visited Hanyang University Hospital for non-stress tests (NSTs). In these patients, FHR tracings were recorded for 50 min including the first and second 20 min of NST and 10 min of vibroacoustic stimulation test. All patients had a reactive NST with CLRC in these antepartum FHR tests. The rate of nonreactive tests and test duration times of these linear and nonlinear criteria were calculated and compared.

Results: The rate of nonreactive tests within a 20- and 40-min period with the MNRC (6.8 and 1.2%) was significantly lower than with the CLRC (35.2 and 10.5%, respectively; p < 0.001). To meet the reactive criteria, the mean durations of the MNRC and the CLRC were 11.47 (95% CI, 11.11-11.84) and 18.45 (95% CI, 17.79-19.12) min, respectively (p < 0.001). The MNRC led to a significant reduction in testing time of 37.8% compared to the CLRC. The rate of reactive tests using only sample entropy (0.20), calculated over 5-min segments of FHR tracings, was 88.3% (95% CI, 88.0-88.7). It was close to the results using the MNRC (93.2%) in the 20-min time window.

Conclusion: The use of the MNRC significantly reduced the test duration times and decreased the rate of nonreactive tests when compared to the CLRC. It may be useful in providing more rapid and accurate assessment of the fetus in electronic FHR monitoring.