Application of lipid metabolism-based indicators in constructing prognostic prediction models for anemia with end-stage renal disease and genomics to explore gene-chemical-anemia relationships

Abstract

Abstract Background Patients on hemodialysis (HD) for end-stage renal disease (ESRD) have poor anemia and prognosis, and this retrospective study from a multicenter in China aimed to investigate the effects of anemia prediction and treatment attainment by constructing model. Methods 1652 patients with ESRD on maintenance hemodialysis (MHD) from September 2021 to June 2022 were selected. After screening the validated factors into the prediction model of random forest regression (RF), the interaction effect was subsequently validated by applying the boosted regression tree method (BRT) and generalized additive model (GAM), and finally the gene-chemical-disease triad was used to verify the potential mechanism of the main predictors. Results Patients with anemia were mainly affected by social cognitive function and renal burden in quality of life. Low levels of HGB under biochemical indicators synergistically predicted anemia onset in ESRD patients with low levels of TRF and high levels of GLU, Meanwhile, the combined effect of high MCHC and low WBC, high TC and high TSAT affected the effect of Hb compliance. The key chemical predictors of anemia are GLU, TC, HDL, Cr, etc., which are influenced by key genes such as EPO and TNF through lipid and atherosclerosis and other mechanisms of lipid metabolism and energy metabolism. Conclusions We developed models for predicting the onset of anemia and Hb attainment effects in ESRD patients and validated the potential mechanisms of their lipid metabolism-associated factors by establishing a gene-chemical-disease triad.