UC San Diego

Preeclampsia (PE) is a major cause of maternal and neonatal morbidity and mortality. Current methods for evaluation of patients with suspected PE do not reliably predict which patients will later develop PE nor distinguish between PE with and without severe features. The objective of this study was to identify candidate extracellular miRNA (ex-miRNA) biomarkers for early diagnosis and prognosis of PE in a cohort of subjects presenting for evaluation of suspected PE. Small RNA-seq libraries were created from maternal serum samples, prospectively collected from participants undergoing evaluation for suspected PE between 20-40 weeks gestational age. Measurements for bivariate biomarkers, consisting of ratios of pairs of ex-miRNAs were performed. 110 bivariate ex-miRNA biomarkers passed both discovery (48 cases, 34 controls) and verification (23 cases, 18 controls) criteria. Specific biomarkers differed in their patterns of expression among different categories of hypertensive disorders in pregnancy (HDP). An iterative machine learning method was then used to identify 3 bivariate miRNA biomarkers that, when applied serially, differentiated cases from controls with a sensitivity of 93% and a positive predictive value (PPV) of 55%, and additionally distinguished between PE cases of different severity. In an independent validation cohort of 11 cases and 7 controls, these three biomarkers had a sensitivity of 91% and a PPV of 85%. We have discovered, verified, and validated 3 bivariate ex-miRNA biomarkers, which, when applied serially, enable accurate early diagnosis of preeclampsia. Bivariate ex-miRNA biomarkers can distinguish between different subtypes of HDP.