Virtual Presentation SRB Virtual Awards 2020

Elemental Metabolomics to identify pregnancies at risk of fetal growth restriction and preeclampsia (#10)

Daniel McKeating 1 , Tu'uhevaha Kaitu’u-Lino 2 , Joshua Fisher 3 , Stephen Tong 2 , Teresa MacDonald 2 , Sue Walker 2 , William Bennett 1 , Anthony Perkins 1
  1. Griffith, Gold Coast, QLD, Australia
  2. Mercy Perinatal, Melbourne, VIC, Australia
  3. The University of Newcastle, Hunter Medical Research Institute, Newcastle, NSW, Australia

Micronutrition has been shown to correlate with detrimental, placental related, pregnancy outcomes such as preeclampsia (PE), and fetal growth restriction (FGR). The consequences of both disorders cause life-long complications for mother and child. This project aims to use elemental metabolomics to identify relationships between the concentrations of key trace elements and pregnancy risk factors. Using bioinformatic algorithms elemental metabolomic profiles will determine if nutritional variations observed in plasma enables early prediction of pregnancy complications.

Plasma samples were from a nested case cohort study from a prospective collection of 2000 samples (the Fetal Longitudinal Assessment of Growth Study) at the Mercy Hospital for Women, Melbourne Australia. Following sample extraction, inductively coupled plasma mass spectrometry (ICP-MS) was used to measure 28 elements in plasma from 328 patients preceding diagnosis of any pregnancy complications (n=193 control, n=97 FGR, n=44 PE).

The ratio of Na+/Ca2+ provided the best means of differentiation for PE, with a receiver operating characteristic (ROC) curve area under curve (AUC) of 0.704 when comparing to controls. Univariate ROC analysis of FGR samples observed that utilising a combination of Na+ and Cu2+ was able to provide preliminary predictive capability of 0.621 – 0.628 AUC. Combing the univariate analysis, with the use of random forest algorithm, allowed for the creation of multivariate predictive models for both complication groups. Multivariate modelling was able to accurately predict 75% of women who would later develop PE (34 of 44), and 63.4% of those who would develop FGR (59 of 97).

This study successfully applied elemental metabolomics to a moderately sized cohort of maternal samples from 36-week pregnant women. The predictive capacity of generating micronutrient profiles with ICP-MS for gestational disorders holds significant potential as a non-invasive assessment of maternal and fetal health, with a possible novel intervention strategy of targeted micronutrient supplementation.