Metabolic characteristics and diagnostic biomarkers of malnutrition in maintenance hemodialysis patients based on WGCNA analysis

Objective Malnutrition is a common and serious complication in maintenance hemodialysis (MHD) patients, significantly impacting their prognosis. However, the metabolic characteristics and diagnostic biomarkers of the nutritional status of MHD patients remain inadequately explored. Methods We compared the clinical baseline characteristics between 100 MHD patients and 60 healthy controls. According to the Modified Quantitative Subjective Global Assessment (MQSGA) scores combined with clinical biochemical indicators, the MHD patients were divided into three groups: the well nourished group (Group A), the mild to moderate malnutrition group (Group B), and the severe malnutrition group (Group C). We then analyzed the differences in MQSGA scores and clinical characteristics among these three groups. Untargeted plasma metabolomics profiles were analyzed to identify differential metabolites and dysregulated metabolic pathways. Additionally, weighted gene co-expression network analysis (WGCNA) was employed in this study to screen key metabolic modules and potential biomarkers associated with malnutrition in MHD patients. Finally, the sensitivity and specificity of the core metabolites were evaluated using receiver operating characteristic (ROC) curves, and the reliability of candidate metabolite results was validated by colorimetric methods. Results Clinical data results showed that compared with normal population, the MHD patient group exhibited decreased expression of nutritional indicators (BMI, Alb, PA, Hb) and bone metabolism-related indicators (Ca, VitD, Klotho). Furthermore, as the severity of malnutrition increased, the levels of these indicators gradually decreased. Conversely, the levels of Scr, BUN, P, CRP, iPTH, and FGF-23 gradually increased. Metabolomics results identified multiple differential metabolites among the three groups, and KEGG analysis revealed enrichment in steroid biosynthesis and the arginine-proline pathway. In addition, revealed significant alterations in both steroid biosynthesis and arginine-proline metabolism modules, characterized by markedly reduced levels of 25(OH)D 3 , Calcidiol, and Pregnanediol. Conversely, the concentrations of L-glutamate and trans-4-hydroxy-L-proline were significantly elevated ( P < 0.05). Moreover, the results of ROC curve analysis combined with validation tests indicate that the above five metabolites may have good diagnostic performance for MHD patients with malnutrition ( P < 0.05). Conclusion Our findings highlight the metabolic derangements underlying malnutrition in MHD patients and propose novel biomarkers for early diagnosis. These insights may facilitate targeted interventions to improve clinical outcomes.