Detecting preeclampsia with a multiple protein serum test: Assay and algorithm development

Abstract Background Preeclampsia is a common cause of maternal and neonatal mortality, and to date there is no definitive, diagnostic test available. We aimed to develop a test to detect preeclampsia using biomarkers representing different pathogenic pathways of disease. Methods A multi-stage development process was used to identify, prioritize, and assess performance of the biomarkers. Samples from symptomatic, asymptomatic, and subjects diagnosed with preeclampsia were screened to produce an algorithm with eight proteins and one clinical biomarker in order to report a binary output of “positive” or “negative”. Test performance was reported on a development cohort, and analytical validation was completed for the underlying protein assays. Results The algorithm resulted sensitivity of 90.6% (CI: 83.1%-95.0%), specificity of 78.4% (CI: 73.9%-82.4%), negative predictive value (NPV) of 96.9% (CI: 94.3%-98.4%), and positive predictive value (PPV) of 52.7% (CI: 45.1%-60.2%) for classification of presence or absence of preeclampsia. Conclusion A panel of biomarkers representing different physiological pathways in preeclampsia measured in serum of symptomatic women provides a sensitive binary classification of presence or absence of disease.