Comparison of oxygen saturation by pulse oximetry and co-oximetry during exercise testing in patients with COPD.

INTRODUCTION Measurement of oxygen saturation by pulse oximetry (SpO2) is frequently performed during exercise testing of patients with COPD to monitor for hypoxemia. The purpose of this study was to assess the accuracy and precision of pulse oximetry during exercise. We hypothesized that the SpO2 would more closely reflect oxygen saturation as measured by co-oximetry (SaO2) when it was corrected for carboxyhemoglobin (COHb). We also hypothesized that SpO2 would more closely reflect SaO2 when the pulse rate by oximeter was equivalent to the heart rate by ECG. Finally, we hypothesized that SpO2 would be a better measure of SaO2 at maximal workloads than at rest or submaximal workloads. METHODS Eight white men with severe COPD (mean +/- SD FEV1, 0.91 +/- 0.37) underwent progressive, symptom-limited exercise testing by cycle ergometry. SaO2 was measured from arterial blood at each workload using a co-oximeter. SpO2 and pulse rate were obtained by a pulse oximeter (Ohmeda 3700). Heart rate was continuously monitored by ECG. RESULTS Reliable oximetric values as determined by a dicrotic notch in each waveform and adequate signal intensity were obtained in all eight patients. SpO2 was a moderately accurate measure of SaO2 (bias, 1.7%; precision, 2.9). The bias actually increased (4.1%) when SpO2 was corrected for COHb. Accuracy of SpO2 was not improved when pulse rate by oximetry and heart rate by ECG were equivalent, nor was the accuracy improved at maximal workloads relative to submaximal workloads during the exercise test. CONCLUSION Oxygen saturation as measured by pulse oximetry (SpO2) in patients with COPD undergoing exercise testing is not sufficiently accurate to replace SaO2 as the gold standard for oxygen saturation.