Salivary Biomarkers for Classifying Acute Physical Fatigue: A Comparative, Exploratory Study

Abstract

Introduction: Physical fatigue is a key determinant of operational readiness in tactical athletes. Hormonal, immune, and enzymatic biomarkers have been proposed for fatigue assessment, but reliability can be affected by external factors. Therefore, this study aimed to compare targeted stress-related biomarkers, proteins identified via untargeted salivary proteomics, and a combined panel of both for classifying acute physical fatigue. Methods: Ten recreationally active adults (6M, 4F) completed a fatiguing protocol within a 3-day sample collection window. Saliva samples were analyzed for targeted biomarkers via commercial immunoassays and proteins via untargeted liquid chromatography-mass spectrometry. Targeted, protein, and combined panels were identified using nested leave-one-subject-out cross-validation with FDR correction and stability selection, then classified via ridge-penalized logistic regression. Models were extended across the full protocol to evaluate classification of fatigue onset, rest, and recovery states. Results: A candidate identified protein (C1S) achieved slightly higher classification accuracy (80%) relative to a panel of targeted biomarkers (IgA and UA 70%), with superior sensitivity but similar specificity. In comparison, the combined panel, spanning immune (IgA), oxidative/metabolic (UA), complement (C1S), and cellular stress-related (ANG) outperformed both single-platform models with 90% accuracy, sensitivity, and specificity. Conclusions: Improved classification accuracy observed with the inclusion of candidate salivary proteins provide preliminary support for non-invasive proteomic monitoring of fatigue in tactical settings, and in particular the combination of complementary targeted and proteomic panels to capture a more complete physiological signature of acute fatigue. Future studies should validate these findings in larger, more diverse populations and assess applicability for chronic fatigue monitoring.