Technology for early differential diagnosis of hypertensive disorders during pregnancy

Abstract

BACKGROUND: To date, no test provides sufficient sensitivity and specificity for the early diagnosis of severe preeclampsia. Meanwhile, severe preeclampsia is a condition that threatens the life of not only the mother, but also the fetus, and requires a solution to the issue of delivery. Therefore, the search for markers of severe preeclampsia is still relevant today.
 AIM: The aim of this study was to create a technology that allows for early differential diagnosis of hypertensive disorders during pregnancy based on a comprehensive analysis of echocardiographic data.
 MATERIALS AND METHODS: Based on the data collected in the Regional Clinical Hospital Perinatal Center, Chita, Russia in 2018-2021, the retrospective analysis of 112 cases of labor was carried out. The total sample was divided into five study groups: 30 relatively healthy women (group 1); 25 patients with chronic arterial hypertension (group 2); 21 patients with gestational arterial hypertension (group 3); 13 patients with moderate preeclampsia (group 4); and 23 patients with severe preeclampsia (group 5). The groups were formed in accordance with current clinical guidelines. Echocardiographic examination in all cases was carried out upon admission to the hospital. Statistical processing of the results was performed using the IBM SPSS Statistics Version 25.0 program.
 RESULTS: The technology for early differential diagnosis of hypertensive disorders during pregnancy is implemented based on a multilayer perceptron, the percentage of incorrect predictions being 20.5 %. The structure of the trained neural network included six input neurons: gestational age, left atrium size in the parasternal position, right ventricular size, interventricular septal thickness, systolic blood flow velocity, and pressure gradient in the pulmonary artery.
 CONCLUSIONS: Comprehensive analysis of echocardiographic data allows for early differential diagnosis of hypertensive disorders during pregnancy, while considering the result of neural network analysis as an additional criterion for severe preeclampsia. In the future, the use of this technology in clinical practice will not only optimize the tactics of managing patients with hypertensive disorders at admission to the hospital, but also reduce the incidence of adverse obstetric and perinatal outcomes.