Circulating microparticle proteins obtained in the late first trimester predict spontaneous preterm birth at less than 35 weeks’ gestation: a panel validation with specific characterization by parity

Abstract

We have previously shown that protein biomarkers associated with circulating microparticles proteins (CMPs) obtained at the end of the first trimester may detect physiologic changes in maternal-fetal interaction such that the risk of spontaneous preterm delivery ≤35 weeks can be stratified. We present here a study extension and validation of the CMP protein multiplex concept using a larger sample set from a multicenter population that allows for model derivation in a training set and characterization in a separate testing set. Ethylenediaminetetraacetic acid (EDTA) plasma was obtained from 3 established biobanks (Seattle, Boston, and Pittsburgh). Samples were from patients at a median of 10-12 weeks' gestation, and the CMPs were isolated via size-exclusion chromatography followed by protein identification via targeted protein analysis using liquid chromatography-multiple reaction monitoring-mass (LC-MRM) spectrometry. A total of 87 women delivered at ≤35 weeks, and 174 women who delivered at term were matched by maternal age (±2 years) and gestational age at sample draw (±2 weeks). From our prior work, the CMP protein multiplex comprising F13A, FBLN1, IC1, ITIH2, and LCAT was selected for validation. For delivery at ≤35 weeks, the receiver operating characteristic (ROC) curve for a panel of CMP proteins (F13A, FBLN1, IC1, ITIH2, and LCAT) revealed an associated area under the ROC curve (AUC) of 0.74 (95% CI, 0.63-0.81). A separate panel of markers (IC1, LCAT, TRFE, and ITIH4), which stratified risk among mothers with a parity of 0, showed an AUC of 0.77 (95% CI, 0.61-0.90). We have identified a set of CMP proteins that provide, at 10-12 weeks gestation, a clinically useful AUC in an independent test population. Furthermore, we determined that parity is pertinent to the diagnostic testing performance of the biomarkers for risk stratification.