A Novel Multiomic AI Approach for Early Preeclampsia Prediction in Pregnancy

Preeclampsia (PE) is one of the leading causes of maternal-fetal morbidity and mortality. It is defined as systolic blood pressure (SBP) ≥ 140 and diastolic blood pressure (DBP) ≥ 90 in two consecutive measurements taken 6 hours apart, associated with proteinuria >300 mg/24 h or 2++ detected by a urine dipstick, or the presence of organ damage or intrauterine growth restriction (IUGR). Preeclampsia complicates 3-8% of pregnancies and is responsible for over 76,000 maternal deaths each year.

Scientific evidence suggests the existence of two distinct phenotypes of the condition: placental preeclampsia and metabolic preeclampsia. The first phenotype is caused by abnormal invasion of the maternal endometrium by the trophoblast, leading to preeclampsia often associated with early presentation in pregnancy, intrauterine growth restriction, and the need for delivery at early gestational ages. Fetal growth restriction due to likely placental insufficiency is defined by the Delphi criteria: estimated fetal weight (EFW) < 3rd percentile, or at least two of the following criteria: EFW < 10th percentile, a decrease in EFW by at least 40 percentiles even if above the 10th percentile, cerebro-placental ratio (CPR) ≤ 1 (5th percentile) or umbilical artery pulsatility index (PI) ≥ 95th percentile, uterine arteries with PI ≥ 95th percentile.

Beyond the hypothesis of abnormal placentation at the beginning of pregnancy, it is now recognized that maternal metabolic risk factors may cause placental malfunction later in pregnancy. The second phenotype is rooted in a metabolic basis, representing about 4% of hypertensive disorders of pregnancy (HDP), and depends on a maternal predisposition in patients with visceral obesity and metabolic syndrome. Visceral obesity is associated with a state of chronic low-grade inflammation, which contributes to insulin resistance, altered glucose homeostasis, and cardiovascular complications.

Metabolic preeclampsia occurs in patients with a pre-existing state of low-grade inflammation related to trunk obesity and metabolic syndrome, compounded by the inflammation and insulin resistance typical of pregnancy. Scientific evidence has shown that in the placentas of these patients, there is a higher density of tertiary villi compared to physiological pregnancies, with reduced intervillous spaces, resulting in hypoperfusion and oxidative stress. Differentiation between these clinical phenotypes can be identified during pregnancy by studying fetal growth as an index of placental function, as well as maternal cardiovascular adaptation to pregnancy in terms of hemodynamic parameters and body water, and finally by studying placental histology after delivery. Hypertensive disorders of pregnancy associated with intrauterine growth restriction (HDP-IUGR) and hypertensive disorders of pregnancy with appropriate-for-gestational-age fetuses (HDP-AGA) are distinguished.

Currently, starting from the first trimester of pregnancy, maternal cardiovascular and hemodynamic function can be assessed with a non-invasive and harmless method for both mother and fetus using the USCOM (Ultra Sonic Cardiac Output Monitor) system. This provides real-time data on numerous central and peripheral hemodynamic parameters such as cardiac output and stroke volume beat-to-beat. It allows measurement of cardiac output from both the right and left heart, systolic stroke volume (SV), systemic vascular resistance, and inotropic index. Its use in pregnancy has already been validated and will significantly enhance the quality of care provided to women with high-risk or pathological pregnancies.

However, there is still discordance among scientific societies regarding the classification of preeclampsia and its potential different clinical phenotypes, making a personalized clinical approach to this condition challenging. While diagnostic criteria have been codified by major national and international scientific societies, it is increasingly important to identify high-risk groups early on, not only to plan a close diagnostic follow-up but also to define appropriate therapeutic strategies based on the etiology.

Recently, a screening method at 11-13 weeks of gestation has been developed, capable of predicting 75% of pregnancies that will develop preterm preeclampsia (<37 weeks of gestation). This is based on a risk calculation algorithm that combines measurements of weight and height, mean arterial pressure measured with automated devices, blood sampling for PLGF levels, and Doppler ultrasound measurement of the mean pulsatility index (PI) of the uterine arteries. However, this screening can only predict a subset of patients who will develop preeclampsia <37 weeks and who may benefit from aspirin administration if taken at doses >100mg and before 16 weeks.

To date, it is still not possible to effectively predict and prevent preeclampsia manifesting >37 weeks, and the etiology of this serious obstetric condition remains a topic of debate and scientific research. Among emerging etiological hypotheses, numerous scientific publications support that an alteration in maternal immunity and immune tolerance underlies hypertensive disorders in pregnancy. Studies on animal models have shown that the activation and expansion of aberrant B cells can trigger inflammatory events leading to preeclampsia. IFN-γ produced by NK cells plays an essential role in spiral arterial remodeling in murine pregnancy. Studies on NK cell-deficient mice have shown defective placental vascular remodeling, characterized by narrow vascular lumens, thick vascular walls, and retention of vascular smooth muscle actin.

Recent discoveries suggest that changes in the maternal gut microbiota, a commensal microbial community capable of modulating the host's immune responses, underlie these immunological alterations. It has been widely described how changes in the diversity and composition of the host gut microbiota-a phenomenon called "dysbiosis," commonly induced by dietary changes or antibiotic treatment-affect systemic immune responses and can disrupt the balance between pro-inflammatory and anti-inflammatory activation. Gut dysbiosis can be associated with excessive weight gain during pregnancy and promote metabolic disorders such as gestational diabetes and preeclampsia, with risks of metabolic alterations in the newborn.

Based on these premises, this study aims to define the characteristics of the maternal gut microbiota in the two different clinical phenotypes of preeclampsia: placental preeclampsia, where the hypertensive disorder is associated with growth restriction, and metabolic preeclampsia, where the hypertensive disorder is associated with a fetus of appropriate weight for gestational age.