Environmental and genetic risk factors for preterm birth: interplays with stressful events during pregnancy.

Abstract

Background: Preterm birth (PTB) etiology remains poorly understood. Our aim was to investigate the relation of environmental factors and specific gene polymorphisms involved in PTB in the context of stressful life events during pregnancy.

Methods: Parental sociodemographic and obstetric data as well as genetic variants of 1263 preterm newborns were analyzed. Logistic regressions were used to identify shared environmental and genetic risk factors for PTB and stressful life events. A Lasso Ridge logistic regression with cross-validation was used to select the best predictors of maternal stress. Associations were evidenced through Bayesian networks.

Results: Starting from a great number of variables, our model was processed and reduced until it allowed to visualize only two environmental factors (alcohol intake and chronic hypertension) along with three SNPs rs66911171 (CR1), rs854552 (PON1), rs4966038 (IGF1R) and two interactions rs854552 x rs4966038 (PON1xIGFR1) and rs5742612 x rs1942386 (IGF1xPGR) related to PTB and maternal stress.

Conclusion: Machine learning techniques allow us to identify two environmental factors, three genetic markers, and two interactions related to PTB in the context of stressful life events. Findings of this exploratory study contribute to the understanding of the complex pathways relating maternal stress and PTB.

Impact: An analysis of environmental factors and preterm birth specific gene polymorphisms in the context of stressful life events during pregnancy is presented. Alcohol intake and chronic hypertension along with SNPs of CR1, PON1, IGF1R and two interactions PON1xIGFR1 and IGF1xPGR are shown as related to preterm birth in the context of stressful life events. This research could help in developing targeted interventions and preventive strategies for at-risk populations. The study emphasizes the potential of machine learning to interpret biological and social interactions affecting health outcomes.