Risk-Stratified Predictive Models of Pregnancy Loss
Relevance. Given the substantial prevalence of pregnancy loss (20-46 % of conceptions), there is an urgent need to establish risk-prediction models for miscarriage to facilitate targeted preventive interventions.Aim. To identify predictors of miscarriage and to develop prognostic models of miscarria...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | Russian |
| Published: |
Numikom LLC
2025-07-01
|
| Series: | Эпидемиология и вакцинопрофилактика |
| Subjects: | |
| Online Access: | https://www.epidemvac.ru/jour/article/view/2258 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Relevance. Given the substantial prevalence of pregnancy loss (20-46 % of conceptions), there is an urgent need to establish risk-prediction models for miscarriage to facilitate targeted preventive interventions.Aim. To identify predictors of miscarriage and to develop prognostic models of miscarriage development based on risk assessment.Materials and methods. We conducted a case-control study involving 282 women: 152 cases of spontaneous pregnancy loss (≤21 weeks gestation) and 130 controls with term deliveries. The assessment protocol included: a standardized 77-item questionnaire, validated Russian version of the Perinatal Anxiety Screening Scale (PASS-R, 31 items). Laboratory analyses comprised: molecular biological methods: quantitative PCR analysis of vaginal microbiota, microbiological assays: oropharyngeal swabs with semi-quantitative microbial growth assessment, fecal bacteriological examination for dysbiosis. Statistical analysis was performed using R software (v4.3.1), incorporating: univariate and multivariate logistic regression, ROC curve analysis (AUC) for model performance evaluation.Results and discussion. Multivariate analysis identified statistically significant predictors of pregnancy loss, enabling development of two predictive mathematical models with strong discriminatory capacity. Key risk factors included: socio-demographic predictors (maternal age ≥35 years (aOR = 11.1; 95 %CI:1.46–238; p = 0.043), leadership occupation (OR = 8.92; 95 %CI:2.93–31.6; p < 0.001). Protective factor: normal body mass index (18.6–25 kg/m2; OR = 0.03; 95 % CI:0.001–0.32; p = 0.008). Clinical-anamnestic factors: medically indicated abortion history (OR = 8.07; 95 % CI:1.50–55.8; p = 0.021), smoking during pregnancy (OR = 6.06; 95 % CI:1.45–33.4; p = 0.022). Microbiological markers: severe intestinal dysbiosis (OR = 9.51; 95 % CI:2.37–64.7; p = 0.005). The developed models demonstrated excellent predictive performance, with high sensitivity and specificity (AUC 0.86 and 0.79).Conclusion. The developed pregnancy loss risk prediction models, incorporating comprehensive socio-demographic, clinical, and microbiological predictors, demonstrate high specificity for spontaneous abortion risk stratification (AUC 0.861 and 0.790). These models show potential for refinement, validation, and eventual clinical implementation. |
|---|---|
| ISSN: | 2073-3046 2619-0494 |