Objective: To evaluate the use of metabolomics for the first-trimester detection of maternal metabolic dysfunction and prediction of subsequent development of early-onset preeclampsia (PE).
Study design: This was a case-control study of maternal plasma samples collected at 11-13 weeks' gestation from 30 women who had subsequently developed PE requiring delivery before 34 weeks and 60 unaffected controls. Nuclear magnetic Resonance (NMR) spectroscopy was used to identify and quantify metabolomic changes in cases versus controls. Both genetic computing and standard statistical analyses were performed to predict the development of PE from the metabolite concentrations alone as well as the combination of metabolite concentrations with maternal characteristics and first-trimester uterine artery Doppler pulsatility index (PI).
Results: Significant differences between cases and controls were found for 20 metabolites. A combination of four of these metabolites (citrate, glycerol, hydroxyisovalerate, and methionine) appeared highly predictive of PE with an estimated detection rate of 75.9%, at a false-positive rate (FPR) of 4.9%. The predictive performance was improved by the addition of uterine artery Doppler PI and fetal crown-rump length (CRL) and with an estimated detection rate of 82.6%, at a FPR of 1.6%.
Conclusion: A profound change in the first-trimester metabolite profile was noted in women who had subsequently developed early-onset PE. Preliminary algorithms appeared highly sensitive for first trimester prediction of early onset PE.