Algorithm for predicting the course of the neonatal period at high risk of intrauterine infection

Abstract

DOI: 10.52705/2788-6190-2026-02-1
УДК 616.33-022.7-036.1-06-037:007

Objective of the study. To develop an algorithm for predicting the course of the neonatal period in cases of high risk of  intrauterine infection.
Materials and methods. We formed a complete database (clinical and laboratory indicators, anamnesis, course of pregnancy,
delivery results, data on the course of the neonatal period, and others) regarding 201 pregnant women who were carriers of pathogens of perinatally significant infections (Herpes simplex 1, 2, Staphylococcus aureus, Candida albicans, and Chlamydia trachomatis). Of these, 50.3% (101) of the women gave birth to healthy children, and 49.7% (100) gave birth to children with intrauterine infection. The age of women who gave birth to children with intrauterine infections was ±25,5 (23-30) years, and for women who gave birth to healthy children, it was ±25 (24-28) years; the distribution of the indicator in both groups differs from normal (DKS=1,43, p=0.033 and DKS=1,94, p=0.001, respectively), and no statistically significant differences were found (U=0,233, p=0,816). The complex of conducted studies included clinical, microbiological, virological, biochemical, and statistical methods.
Results. From the obtained data, it is evident that using the model, 79% (11/14) of all healthy newborns were classified
correctly; among newborns with Apgar scores from 0 to 7, this indicator equals 76%. In general, using the constructed model
on the test set, 77% (23/30) of all cases were correctly classified (concordance coefficient). The value of the Somers' D correlation coefficient (Somers' D), which reflects the relationship between the actual frequency and that predicted by the logistic regression equation, is 0,464, p=0,004. Analysis of clinical and laboratory data using the mathematical model of logit regression allowed forming a set of risk predictors for the development of intrauterine infection with a high level of concordance exceeding 90%. This formed the basis for developing a computer program for calculating the risk of birth of children with intrauterine infection (IUI).
Conclusions. Thus, we managed to develop a program for screening prenatal prediction of intrauterine infections, which includes clinical predictors of intrauterine infection and regulatory-transport proteins in the blood of pregnant women and the amniotic fluid of parturient women. The created program for calculating the risk of IUI in the early neonatal period is recommended to be applied from the moment signs of live birth are determined, from the 22nd week of pregnancy, the term of premature birth, after the second ultrasound screening.