Accuracy of a third (Dolphin Voyager) versus first generation pulse oximeter (Nellcor N-180) in predicting arterial oxygen saturation and pulse rate in the anesthetized dog

Objective: To compare the accuracy of a 3rd (Dolphin Voyager) versus 1st generation pulse oximeter (Nellcor N-180). Study design: Prospective laboratory investigation. Animals: Eight adult dogs. Methods: In anesthetized dogs, arterial oxygen saturation (SpO(2)) was recorded simultaneously with each pulse oximeter. The oxygen fraction in inspired gas (FiO(2)) was successively reduced from 1.00 to 0.09, with re-saturation (FiO(2) 0.40) after each breathe-down step. After each 3-minute FiO(2) plateau, SpO(2) and pulse rate (PR) were compared with the fractional arterial saturation (SaO(2)) and PR determined by co-oximetry and invasive blood pressure monitoring, respectively. Data analysis included Bland-Altman (B-A) plots, Lin's concordance correlation factor (rho(c)), and linear regression models. Results: Over a SaO(2) range of 33-99%, the overall bias (mean SpO(2) - SaO(2)), precision (SD of bias), and accuracy (A(rms)) for the Dolphin Voyager and Nellcor N-180 were 4.3%, 4.4%, and 6.1%, and 3.2%, 3.0%, and 4.3%, respectively. Bias increased at SaO(2) < 90%, more so with the Dolphin Voyager (from 1.6% to 8.6%) than Nellcor N-180 (from 3.2% to 4.5%). The SpO(2) readings correlated significantly with SaO(2) for both the Dolphin Voyager (rho(c) = 0.94) and Nellcor N-180 (rho(c) = 0.97) (p < 0.001). Regarding PR, bias, precision, and accuracy (A(rms)) for the Dolphin Voyager and Nellcor N-180 were -0.5, 4.6, and 4.6 and 1.38, 4.3, and 4.5 beats minute(-1), respectively. Significant correlation existed between pulse oximeter and directly measured PR (Dolphin Voyager: rho(c) = 0.98; Nellcor N-180: rho(c) = 0.99) (p < 0.001). Conclusions and clinical relevance: In anesthetized dogs with adequate hemodynamic function, both instruments record SaO(2) relatively accurately over a wide range of normal saturation values. However, there is an increasing overestimation at SaO(2) < 90%, particularly with the Dolphin Voyager, indicating that 3rd generation pulse oximeters may not perform better than older instruments. The 5.4-fold increase in bias with the Dolphin Voyager at SaO(2) < 90% stresses the importance of a 93-94% SpO(2) threshold to ensure an arterial saturation of >or=90%. In contrast, PR monitoring with both devices is very reliable.