Effect of peripheral perfusion on accuracy of pulse oximetry in children.

Study Objective: To examine the effect of perfusion on accuracy of two pulse oximeters in children and to determine thresholds of perfusion below which these pulse oximeters become inaccurate or cease to function.

Design: Prospective, observational clinical study.

Setting: Operating room of a large university hospital.

Patients: 19 children 10 years of age or less, who were scheduled for general anesthesia with placement of an intraarterial catheter.

Interventions: A radial artery catheter, laser Doppler probe, skin temperature sensor, and band probes of two oximeters, Ohmeda 3700 (Boulder, CO) and Nellcor N200 (Hayward, CA), were attached to the same hand. Baseline pulse oximeter and Doppler readings were obtained with simultaneous hemoximetry (AVL Model 912 CO-Oxylite, Roswell, GA), skin and esophageal temperatures, total hemoglobin, and transduced arterial pressure. Readings of all parameters (n = 94) were obtained during periods of low perfusion or by occluding the upper arm to 70% to 100% of systolic pressure.

Measurements And Main Results: Bias (SpO2-SaO2) of each oximeter is compared to each perfusion variable (age, weight, core and skin temperature, hemoglobin concentration, pulse pressure, and percent flow by laser Doppler) to determine effect on accuracy. Data were analyzed using backward multivariate linear regression, Pearson correlation coefficients, and independent paired t-test. p < 0.05 was considered significant. Less than 2% bias is seen with either oximeter (Nellcor 1.55 +/- 2.33, Ohmeda 0.78 +/- 2.25). Independent predictors of bias for each machine include weight (r = -0.376; p < 0.001) and pulse pressure (r = 0.250; p = 0.021) for the Nellcor, and weight (r = -0.390; p < 0.001), percent flow by Doppler (r = 0.220; p = 0.035), and core temperature (r = 0.307; p = 0.003) for the Ohmeda. However, using predetermined thresholds for each variable, only skin temperature below 30 degrees C is identified as a significant predictor of oximeter inaccuracy.

Conclusions: At the parameters explored in this study, the selected seven perfusion variables (age, weight, core and skin temperature, hemoglobin concentration, pulse pressure, and percent flow by laser Doppler) have little effect on accuracy of pulse oximetry in children.