Прогнозирование преждевременных родов на основании клинико-анамнестических данных и состояния микробиоты влагалища

Abstract

Early detection of a high risk of preterm birth (PB) provides an opportunity to prevent it. Objective. To develop a mathematical prognostic model to assess the risk of PB based on a comprehensive analysis of clinical and medical record data and the state of the vaginal microbiota. Patients and methods. The study enrolled 199 pregnant women, of whom 41 had PB and 158 had full-term birth. All patients underwent clinical and laboratory observations. The vaginal microbiota was examined in the first trimester by a real-time polymerase chain reaction using the Femoflor-16 test system. Discriminant analysis was used to develop a prognostic model. Results. A method was developed to predict PB by calculating the predictive index for preterm birth (PIPB). If the PIPB value is >0, the risk of preterm birth is low; if <0, the risk is high. The sensitivity and specificity of the method are 70.7% and 81.65%, respectively, and its efficiency is 79.4%. The developed prognostic model allows early identification of pregnant women at risk for PB based on the information on occupational hazards, past medical history of combined hormonal contraception, gravidity and a history of pregnancy loss, data on the course of current pregnancy and characteristics of the vaginal microbiota. Conclusion. The proposed predictive index is based on available diagnostic criteria and can be widely used in clinical practice. Key words: vaginal microbiota, preterm birth, Femoflor-16, prediction, real-time polymerase chain reaction, discriminant analysis