Evaluating Continuous SpO2 as a Lactate-Threshold Marker: Implications for Wearable Sports-Monitoring Systems

Continuous peripheral oxygen saturation (SpO2) monitoring can describe systemic oxygenation during exercise, but its value for identifying blood lactate threshold (LT) depends on the timing of recognizable signal changes. Two independent cohorts of 18 healthy adults completed the same incremental cycling protocol with forehead or fingertip pulse oximetry. We characterized SpO2 at key landmarks and throughout exercise and evaluated three changes: a decrease exceeding individual early-exercise variation, a rapid recent decline, and a sustained transition to a steeper decline. All detectors used only measurements available at each update. Event times were compared with Dmax and modified Dmax LT, distinguishing the sustained transition’s estimated breakpoint from its later detection. The first decrease typically preceded LT by 13–16 min. Rapid recent declines also showed substantial timing errors; requiring steeper declines reduced detection coverage. Sustained transitions were detected in 14/18 forehead and 13/18 fingertip recordings. Detection times were closer to LT than breakpoints, but median absolute detection errors remained 3.09–4.25 min across cohorts and LT definitions. Only 4–5 of 18 recordings per cohort had detection times within ±2 min of LT. Timing accuracy remained limited under alternative processing and detection settings. In a separate trajectory comparison, the fingertip cohort showed a larger SpO2 decline. Separate adjustments for exercise frequency and attained workload per kilogram of body mass reduced the estimated cohort difference, but the design could not isolate anatomical-site or sensor-configuration effects. Continuous SpO2 therefore characterized exercise-related oxygenation changes, but these three features had limited accuracy as stand-alone LT markers in healthy adults under this cycling protocol.