Validity of pulse oximetry during exercise in elite endurance athletes

Eleven highly trained male cyclists [maximal aerobic power (VO2max) = 70.6 +/- 4.2 ml.kg-1.min-1] performed both high intensity constant load (90-95% VO2max) and incremental cycle exercise tests with arterial blood sampling to evaluate the accuracy of pulse oximeter estimates (%SpO2) of arterial oxyhemoglobin fraction of total hemoglobin (%HbO2). Three subjects also performed an incremental exercise test in hypoxic conditions (inspired partial pressure of O2 = 89, 93, or 100 Torr). Arterial %HbO2 was determined via CO-oximetry and ranged from 72 to 99%. Three Ohmeda 3740 pulse oximeters were used to estimate %HbO2, one on each ear lobe and a finger probe. The finger probe tended to provide the best estimate of %HbO2 during exercise: the mean %SpO2 - %HbO2 difference for 232 exercise observations was 0.52 +/- 1.36% (SD). Finger probe %SpO2 and %HbO2 were highly correlated [r = 0.98, standard error of the estimate (SEE) = 1.32%, P less than 0.0001]. The accuracy of pulse oximeters has been questioned during high-intensity exercise. When aerobic power was greater than 81% of VO2max (n = 75), the finger probe's mean error was -0.01 +/- 1.40%. Finger probe %SpO2 and %HbO2 were highly correlated (r = 0.97, SEE = 1.32%, P less than 0.0001). These results indicate that this pulse oximeter is a valid predictor of %HbO2 in elite athletes during cycle exercise.