Adapted ST analysis during labor: relative versus absolute ST events, a case-control study

Background The value of ST analysis of the fetal electrocardiogram during labor to lower asphyxia and cesarean section rates is uncertain. Physiological variation of the electrical heart axis between fetuses may explain false alarms in conventional ST analysis (absolute ST analysis). ST events (alarms) based on relative T/QRS rises (relative ST analysis) correct for this variation and may improve diagnostic accuracy of ST analysis. Aims To compare the diagnostic accuracy of absolute and relative ST analysis with regard to fetal acidemia. Study design Retrospective case-control study. Subjects 20 healthy women with an uncomplicated pregnancy monitored with ST analysis during labor: 10 cases (umbilical cord artery pH < 7.05) and 10 controls (pH > 7.20). Outcome measures Sensitivity, specificity, positive and negative likelihood ratio. Results In 16 of the 20 patients a total of 54 absolute ST events were reported. Two reviewers classified the cardiotocograms; in cases 29% of the absolute ST events were significant, in the controls it was 19%. Relative ST analysis versus absolute ST analysis showed a sensitivity of 90% (55-100%) vs. 70% (35-93%), a specificity of 100% (69-100%) vs. 70% (35-93%), a positive likelihood ratio of infinity vs. 2.3 (0.8-6.5), a negative likelihood ratio of 0.1 (0.0-0.6) vs. 0.4 (0.2-1.2), and diagnostic odds ratio of infinity vs. 5.4 (0.8-36.9). McNemar showed no statistical significant difference between the sensitivity and specificity of the methods. Conclusions We observed higher positive and lower negative likelihood ratios for relative ST analysis in comparison to absolute ST analysis. In this small study we found no statistical difference. Relative ST analysis should be studied in a larger study.

Automated summary

The study compared the diagnostic accuracy of absolute and relative ST analysis in detecting fetal acidemia, with relative ST analysis showing higher sensitivity and lower negative likelihood ratio.