Alterations in the serum metabolome in patients with the COVID-19 Omicron variant and in recovered cases

Corona Virus Disease (COVID-19) has become a global public health crisis, and the Omicron variant has rapidly taken over as soon as it was detected Serum circulating metabolites can provide extensive insights into the pathogenesis and diagnosis of many diseases. We included 336 omicron variant cases (OC), 216 recovered cases (RC), and 380 healthy controls (HC) for untargeted metabolomics analysis and analyzed their serum metabolic profiles by liquid chromatography-tandem mass spectrometry. Principal component analysis, orthogonal partial least squares discriminant analysis, t-test analysis and false discovery rate were used to characterize the serum metabolites of OC and RC. In addition, a noninvasive diagnostic model for OC was developed using Receiver operating characteristic analysis. Finally, a correlation analysis was performed using data from our published articles. The results showed that compared with HC, five metabolites, including DL-stachydrine, D-(+)-pipecolinic acid, furazolidone, L-arginine and 5α-dihydrotestosterone glucuronide were significantly elevated and one metabolite, prenylcysteine, was significantly decreased in the serum of OC, and that the increase in L-arginine and the decrease in prenylcysteine led to impaired urea cycling and a high risk of developing atherosclerosis, respectively. These metabolites were not fully restored to healthy human levels in recovered cases. In addition, we constructed a noninvasive diagnostic model for distinguishing Omicron variant patients from healthy individuals based on the six differential metabolites, and achieved high diagnostic efficacy in both the discovery and validation cohorts. Finally, the results of the correlation analysis showed a strong correlation between the alterations in the oropharyngeal microbiome and serum metabolome and the clinical indicators in the omicron variant cases. This study was the first to characterize serum metabolites in OC and RC based on a large clinical cohort, and successfully constructed and validated a noninvasive diagnostic model for Omicron variant patients.