Abstract
BACKGROUND: Blood lactate (BL) is a critical biomarker for assessing anaerobic metabolism and fatigue. Sweat lactate (SWL) and sweat rate (SWR) have been explored as non-invasive alternatives, but their capacity to estimate BL dynamics after short-term high-intensity exercise remains unclear. PURPOSE: This pilot study aimed to evaluate whether BL dynamics can be predicted using a regression model based on the time-series patterns of SWL and SWR measured by wearable sensors. METHODS: Five healthy male athletes (three sprinters and two endurance runners) performed a 30-second Wingate anaerobic test. SWL and SWR were continuously monitored using a wearable electrochemical sensor and a ventilated capsule-type sweat rate meter. Capillary BL was sampled for 30 minutes post-exercise. RESULTS: BL showed a delayed peak at 6.4 ± 1.2 min, while SWL and SWR exhibited biphasic responses. The second SWL peak (7.5 ± 2.2 min) aligned with the BL peak. Although peak-based correlations were not significant, Pearson correlations using time-series data revealed strong associations (r = 0.501-0.933 for SWL; r = 0.515-0.805 for SWR; all p < 0.001). A multivariate regression model using both variables predicted BL with high accuracy (R² = 0.763, RMSE = 1.612, MAE = 0.995, p < 0.001). CONCLUSION: These findings support the feasibility of a regression-based approach using sweat-derived time-series data to non-invasively estimate BL dynamics after high-intensity exercise.