FIRST-TRIMESTER NUCLEAR MAGNETIC RESONANCE-BASED METABOLOMIC PROFILING INCREASES THE PREDICTION OF GESTATIONAL DIABETES MELLITUS

Objective: To investigate whether nuclear magnetic resonance (NMR)-based metabolomic profiling of maternal blood can be used for first-trimester prediction of gestational diabetes mellitus (GDM).

Study design: This was a prospective study of 20,000 women attending for routine pregnancy care at 11-13 weeks' gestation. Metabolic profiles were assessed using a high-throughput NMR metabolomics platform. To inform translational applications, we focused on a panel of 34 clinically validated biomarkers for detailed analysis and risk modelling. All biomarkers were used to generate a multivariable logistic regression model to predict GDM. Data were split using random seed into a 70%-30% training and validation set, respectively. Performance of the multivariable models was measured by receiver operating characteristic (ROC) curve analysis and detection rates (DRs) at fixed 10% and 20% false positive rates. Calibration for the combined risk model for all GDM was assessed visually through a figure showing the observed incidence against the predicted risk for GDM. A sensitivity analysis was carried out excluding the 64 women in our cohort who were diagnosed with GDM prior to 20 weeks' gestation.

Results: The concentration of several metabolomic biomarkers, including cholesterols, triglycerides, fatty acids, and amino acids, were different in the group that developed GDM, compared to those without GDM. Addition of biomarker profile improved the prediction of GDM provided by maternal demographic characteristics and elements of medical history alone (area under the receiver operating characteristics curve [AUROC] 0.790 and detection rate (DR) of 50%, 95%CI 44.3%-55.7%, at 10% false positive rate (FPR) and 63%, 95%CI 57.4%-68.3%, at 20% FPR, to 0.840, 56%, 95%CI 50.3%-61.6% and 73%, 95%CI 67.7%-77.8%, respectively). The performance of combined testing was better for GDM treated by insulin (ROC AUC 0.905, DR 76%, 95%CI 67.5%-83.2%, at 10% FPR and DR 85%, 95%CI 77.4%-90.9%, at 20%) than GDM treated by diet alone (ROC AUC 0.762, DR 47%, 95%CI 37.7%-56.5%, at 10% FPR and DR 64%, 95%CI 54.5%-72.7%, at 20% FPR). In the calibration plot there was good agreement between the observed incidence of GDM against that predicted from the combined risk model. In the sensitivity analysis excluding the women diagnosed with GDM prior to 20 weeks' gestation, there was a negligible difference in ROC AUC in comparison to the results from the entire cohort combined.

Conclusions: Addition of NMR-based metabolomic profiling to risk factors can provide first-trimester prediction of GDM.