Early Serum Metabolism Profile of Post-operative Delirium in Elderly Patients Following Cardiac Surgery With Cardiopulmonary Bypass

BACKGROUND: Cardiac surgery with cardiopulmonary bypass (CPB) is considered to be one of the surgical types with the highest incidence of post-operative delirium (POD). POD has been associated with a prolonged intensive care and hospital stay, long-term neurocognitive deterioration, and increased mortality. However, the specific pathogenesis of POD is still unclear. Untargeted metabolomics techniques can be used to understand the changes of serum metabolites in early POD to discover the relationship between serum metabolites and disease. MATERIALS AND METHODS: The present study recruited 58 elderly patients undergoing cardiac surgery with CPB. Serum was collected within the first 24 h after surgery. The Confusion Assessment Method (CAM) and ICU-CAM assessments were used to identify patients who experienced POD. All patients with normal post-operative cognitive assessment were included in the non-POD groups. Moreover, we collected serum from 20 healthy adult volunteers. We performed untargeted analyses of post-operative serum metabolites in all surgical groups, as well as serum metabolites in healthy non-surgical adults by using liquid chromatography mass spectrometry (LC/MS) and analyzed metabolic profiles and related metabolites. RESULTS: The probability of POD after cardiac surgery were 31%. There were statistically significant differences in post-operative mechanical ventilation time, ICU stay time and post-operative hospital stay between POD and non-POD group (P < 0.05). And ICU stay time was an independent risk factor for POD. The analysis revealed that a total of 51 differentially expressed metabolites (DEMs) were identified by comparing the POD and non-POD group, mostly lipids and lipid-like molecules. Three phosphatidylinositol (PI) were down-regulated in POD group, i.e., PI [18:0/18:2 (9Z, 12Z)], PI [20:4 (8Z, 11Z, 14Z, 17Z)/18:0], and PI [18:1 (9Z)/20:3 (8Z, 11Z, 14Z)]. The receiver operating characteristic (ROC) curve analysis showed that three kinds of PI metabolites had the highest area under the curve (AUC), which were 0.789, 0.781, and 0.715, respectively. Correlation analysis showed that the expression of three PIs was negatively correlated with the incidence of POD. CONCLUSION: Our findings suggest that lipid metabolism plays an important role in the serum metabolic profile of elderly patients with POD in the early post-operative period. Low serum lipid metabolic PI was associated with incidence of POD in elderly following cardiac bypass surgery, which may provide new insights into the pathogenesis of POD.