Peripheral perfusion is often poor and barely pulsatile in patients undergoing prolonged major surgery. Hence, the arterial blood oxygen saturation (SpO2) readings from commercial finger pulse oximeters can become unreliable or cease when they are most needed. To overcome this limitation, the esophagus has been investigated as an alternative measurement site, as perfusion may be preferentially preserved centrally. A reflectance esophageal pulse oximeter probe, and a processing system implemented in LabVIEW were developed. The system was evaluated in clinical measurements on 49 cardiothoracic surgery patients. The SpO2 values from the esophagus were in good agreement with arterial blood oxygen saturation (SaO2) values obtained from blood gas analysis and CO-oximetry. The means (+/-SD) of the differences between the esophageal SpO2 and SaO2 results from blood gas analysis and CO-oximetry were 0.02 +/- 0.88% and -0.73 +/- 0.72%, respectively. In five (10.2%) of the patients, the finger pulse oximeter failed for at least 10 min while the esophageal SpO2 readings remained reliable. The results confirm that the esophagus may be used as an alternative monitoring site for pulse oximetry even in patients with compromised peripheral perfusion.