Research led by Mirvie Inc. has discovered a molecular signal in blood samples that can predict preeclampsia and related conditions months before clinical symptoms appear. Findings suggest a new approach to risk detection that may support precision-based strategies in maternal health.

Hypertensive disorders of pregnancy affect one in six pregnancies and are a leading cause of maternal morbidity and mortality in the United States, with incidence nearly doubling between 2007 and 2019.

Conditions present over a spectrum, ranging from cases that require extreme preterm delivery to those involving only mild elevations in blood pressure after delivery. Pathogenesis remains poorly understood. Multiple origins have been hypothesised, including viral infection, immune activation, and placental dysfunction.

Only 5% of pregnant individuals meet criteria for high risk, such as chronic hypertension or prior preeclampsia. For the remaining 95%, clinicians rely on combinations of moderately predictive factors, including race and socioeconomic status. Demographic factors are often difficult to apply clinically and may introduce risk bias. Current diagnostic tools are used only after symptoms appear.

In the study, "Molecular subtyping of hypertensive disorders of pregnancy," published in Nature Communications, researchers conducted a multi-center, prospective observational study to investigate whether molecular subtypes in maternal blood could improve early risk identification of hypertensive disorders of pregnancy.

A total of 10,745 pregnant individuals over the age of 18 were enrolled, with 9,102 samples included in the final analysis. Blood samples were collected between 17 weeks and four days and 22 weeks of gestation. Participants were recruited across 11 clinical sites and through mobile phlebotomy.

Researchers extracted cell-free RNA from maternal blood and generated transcriptomes for gene expression modeling. Samples were split by collection date into training and validation groups, with 5,399 used for model development and 2,829 held for testing. Molecular subtypes were identified through differential gene expression and validated across subgroups ranked by clinical severity.

Two distinct gene expression patterns emerged. One group showed high expression of placental genes, including PAPPA2, and was associated with early-onset or severe preeclampsia requiring preterm delivery. The other group displayed elevated immune-related transcripts and was linked to term preeclampsia and gestational hypertension.

Classifier performance for placental-associated hypertensive disorders reached an area under the curve of 0.88 in individuals of advanced maternal age with no high risk factors. Median lead times from blood draw to delivery were 108 days for placental-associated cases and 132 days for immune-associated cases.

PAPPA2 overexpression showed no association with spontaneous preterm birth, indicating specificity for hypertensive conditions. Individuals without high risk factors who tested positive for the placental-associated subtype delivered an average of 11.3 days earlier than test-negative individuals. Among neonates admitted to intensive care, the average stay was 14 days for test-positive cases compared to five days for test-negative cases.

Compared to current clinical guidelines, the RNA-based classifier showed stronger predictive performance. In individuals of advanced maternal age, the positive likelihood ratio was 15 times higher than USPSTF recommendations, and the negative likelihood ratio was substantially lower (0.12 versus 0.38). For placental-associated cases, the negative predictive value was 99.7%.

Findings support the use of a simple blood test during routine mid-pregnancy care to identify individuals at risk for hypertensive disorders of pregnancy before symptoms appear. Molecular subtypes were detectable in samples collected during the standard anatomy scan window and were independent of race or traditional clinical classifications.

The classifier performed most strongly in individuals without preexisting high risk factors, who represent the majority of pregnancies. Risk stratification based on molecular signals may improve adherence to preventive therapies and allow targeted enrollment in clinical trials. Results suggest the potential for precision-based interventions to reduce maternal and neonatal complications.