Pulse oximetry: evaluation of accuracy during outpatient general anesthesia for oral surgery

Pulse oximetry has been shown to be accurate under steady state conditions. In this study, the accuracy of four pulse oximeters are evaluated and compared during outpatient general anesthesia for third molar extractions. The oximeters evaluated are the Nellcor N-100, the Ohmeda 3700, the Novametrix model 500, and the Bird 4400 portable pulse oximeter.Ultralight general anesthesia for oral surgery presents a unique challenge for respiratory monitoring in that patients are often not intubated and commonly experience periods of hyper- and hypoventilation. Airway obstruction, apnea, and laryngospasm may occur easily and patients often vocalize and move during surgery. Because hypoxemia is the primary cause of morbidity and mortality during anesthesia, an accurate, continuous, and noninvasive monitor of oxygenation is critical to risk management.Twenty ASA class I and II patients underwent outpatient general anesthesia for third molar removal using nitrous oxide-oxygen, midazolam, fentanyl, and methohexital. Arterial blood samples were obtained at five-minute intervals during anesthesia, as well as any time a desaturation of >5% occurred, for measurement of arterial SaO(2) with an IL282 CO-Oximeter. These values were compared with simultaneously recorded saturations observed for each pulse oximeter. A total of 122 arterial samples were obtained over a range of PaO(2) from 52-323 mm Hg and observed saturations of 70-100%.The Bird 4400 portable pulse oximeter proved to be the most accurate and reliably predicted arterial saturation under these conditions (y = 1.03x - 2.8, r = 0.85). The Novametrix model 500 pulse oximeter also demonstrated a high degree of accuracy by linear regression analysis, but displayed the lowest correlation coefficient (spread of data points) overall (y = 0.97x + 2.8, r = 0.80.) The Nellcor N-100 pulse oximeter also proved to be highly accurate. (y = 1.05x - 4.1, r = 0.84.) In contrast, regression analysis of the observed saturations obtained with the Ohmeda 3700 pulse oximeter revealed that this unit significantly underestimated arterial saturation (y = 1.20x - 19.6, r = 0.83.)This study demonstrates that despite the rigorous conditions imposed by outpatient general anesthesia for oral surgery, three of the pulse oximeters tested were linearly accurate in predicting arterial oxyhemoglobin saturation over the range of 70-100%. The Ohmeda 3700 was found to significantly underestimate arterial saturation.