The construction of a panel of serum amino acids for the identification of early chronic kidney disease patients

Abstract

Background: Serum creatinine, urea, and cystatin-c are standardly used for the evaluation of renal function in the clinic. However, some patients have chronic kidney disease but still retain kidney function; a conventional serum index in these patients can be completely normal. Serum amino acid levels can reflect subtle changes in metabolism and are closely related to renal function. Here, we investigated how amino acids change as renal impairment increases. Methods: Subjects were divided into three groups by renal function glomerular filtration rate: healthy controls, patients with chronic kidney disease with normal kidney function, and patients with chronic kidney disease with decreased kidney function group. We identified 11 amino acids of interest using LC-MS/MS on MRM (+) mode. Results: Statistical analysis indicated that alanine (ALA), valine (VAL), and tyrosine (TYR) decrease with renal function impairment, whereas phenylalanine (PHE) and citrulline (CIT) increase. We tried to construct a diagnostic model utilizing a combination of amino acids capable of identifying early chronic kidney disease patients. The accuracy, specificity, and sensitivity of the combining predictors were 86.9%, 84.6%, and 90.9%, respectively, which is superior to the reported values for serum creatinine, urea, and cystatin-c. Conclusion: Our data suggest that serum amino acid levels may supply important information for the early detection of chronic kidney disease. We are the first to establish a diagnostic model utilizing serum levels of multiple amino acids for the diagnosis of patients with early-stage chronic kidney disease.