First-trimester metabolomic detection of late-onset preeclampsia.

OBJECTIVE We sought to identify first-trimester maternal serum biomarkers for the prediction of late-onset preeclampsia (PE) using metabolomic analysis. STUDY DESIGN In a case-control study, nuclear magnetic resonance-based metabolomic analysis was performed on first-trimester maternal serum between 11(+0)-13(+6) weeks of gestation. There were 30 cases of late-onset PE, i.e., requiring delivery ≥37 weeks, and 59 unaffected controls. The concentrations of 40 metabolites were compared between the 2 groups. We also compared 30 early-onset cases to the late-onset group. RESULTS A total of 14 metabolites were significantly elevated and 3 significantly reduced in first-trimester serum of late-onset PE patients. A complex model consisting of multiple metabolites and maternal demographic characteristics had a 76.6% sensitivity at 100% specificity for PE detection. A simplified model using fewer predictors yielded 60% sensitivity at 96.6% specificity. Strong separation of late- vs early-onset PE groups was achieved. CONCLUSION Significant differences in the first-trimester metabolites were noted in women who went on to developed late-onset PE and between early- and late-onset PE.