BACKGROUND: The reliability of pulse oxymetry probes when applied to the finger or toes may be compromised in certain patients. Other sites less subject to mechanical interference or a pathophysiologic decrease in pulse amplitude have been sought. In the patients with moderate defect (N=20) in pulmonary function test, we examined the accuracy of buccal and digital SpO2 (oxygen saturation of pulse oxymetry) monitoring. METHODS: SpO2 probe was placed firmly in the corner of the patient's mouth. Buccal and finger SpO2 and radial SaO2 (arterial oxygen saturation) were measured before the induction of anesthesia. The agreement between SaO2 and each SpO2 were calculated with the method outlined by Bland and Altman. RESULTS: Buccal SpO2 was higher than finger SpO2, but finger SpO2 agreed more closely with SaO2 (buccal; 97.9+/-1.89, finger; 94.5+/-2.48, radial; 93.73+/-2.73%). CONCLUSIONS: We conclude that buccal SpO2 monitoring may offer alternative when other sites aren't available. But, we suggest that buccal SpO2 should be further evaluated for the accuracy.