The increasing prevalence of chronic kidney disease (CKD) and its terminal stage, end-stage renal disease (ESRD), raises the importance of an accurate, early, and point-of-care method to diagnose and monitor patients. Saliva is a potential point-of-care diagnostic biofluid for its simple collection and ability to reflect systemic health status. This study investigated salivary spectral signatures in ESRD patients and their diagnostic potential compared to healthy controls. Saliva samples were collected from 24 ESRD patients undergoing hemodialysis and 24 age/sex-matched healthy controls. The dried saliva samples were analyzed using Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) spectroscopy in the 4000-400 cm⁻¹ range. Chemometric analyses, including Principal Component Analysis (PCA) and Partial Least Squares Discriminant Analysis (PLS-DA), were applied to preprocessed spectra to identify discriminatory spectral features and establish classification models. Second derivative spectroscopic analysis of ATR-FTIR spectra revealed distinctive spectral patterns in dried ESRD saliva samples, including characteristic peak shifts observed in both the amide I secondary structures (from 1636 cm-1 in controls to 1629 cm-1 in ESRD) and carbohydrate (from 1037 cm-1 in controls to 1042 cm-1 in ESRD) regions. PCA demonstrated clear clustering patterns across key biological spectral regions, including the lipid CH stretching region (3000-2800 cm-1), the fingerprint region (1800-900 cm-1), and their combination (3000-2800 cm-1 + 1800-900 cm-1). PLS models based on the fingerprint region achieved optimal diagnostic performance (87.5-100% accuracy, 75-100% sensitivity, and 100% specificity). Biochemical markers associated with ESRD revealed variations in lipids, protein, sugar moieties, carbohydrates, and nucleic acids, reflecting the underlying pathological changes in CKD, with the most prominent band at ∼1405 cm-1. ATR-FTIR analysis of dried saliva demonstrated potential as a non-invasive diagnostic tool for ESRD. This approach could complement existing diagnostic methods, particularly in resource-limited settings or for frequent monitoring requirements.
Read full abstract