Background and objectives: By combining the techniques of metabolomics and computational biology, this research aims to explore the mechanism of metabolic dynamics in critically injured patients and develop a new early warning method for mortality.
Methods and study design: A prospective cohort study was conducted, group plasma samples of critically injured patients were collected for 1H-NMR metabolomics analysis. The data was processed with partial least squares regression, to explore the role of enzyme-gene network regulatory mechanism in critically injured metabolic network regulation and to build a quantitative prediction model for early warning of fast death.
Results: In total, 60 patients were enrolled. There were significant differences in plasma metabolome between the surviving patients and the deceased ones. Compared to the surviving patients, 112 enzymes and genes regulating the 6 key metabolic marker disturbances of neopterin, corticosterone, 3-methylhistidine, homocysteine, Serine, tyrosine, prostaglandin E2, tryptophan, testosterone and estriol, were observed in the plasmas of deceased ones. Among patients of different injury stages, there were significant differences in plasma metabolome. Progressing from T0 to T50 stages of injury, increased levels of neopterin, corticosterone, prostaglandin E2, tryptophan and testosterone, together with decreased levels of homocysteine, and estriol, were observed. Eventually, the quantitative prediction model of death warning was established. Cross-validation results showed that the predictive effect was good (RMSE=0.18408, R2=0.87 p=0.036).
Conclusions: Metabolomics approaches can be used to quantify the metabolic dynamics of patients with critically injuries and to predict death of critically injured patients by plasma 1H-NMR metabolomics.