1999 OPEN FORUM Abstracts

ACCURACY OF THE NONIN ONYX 9500 PULSE OXIMETER FOR THE MEASUREMENT OF ARTERIAL OXYGEN SATURATION.

David C. Shelledy, PhD, RRT, Paulla K. Smith, BS, CRT and Rebecca A. Downing, BS, CRT. The University of Texas Health Science Center at San Antonio, San Antonio, Texas.

INTRODUCTION: The Nonin Onyx 9500 finger pulse oximeter (Nonin Medical, Inc., Plymouth, MN) is a miniature (3.3´3.3´5.7 cm), lightweight (60 gm) device for the measurement of arterial oxygen saturation (SpO2). We compared the performance of the Nonin Onyx 9500 system to oxygen saturations (SaO2) obtained via arterial blood co-oximetry using the IL 682 (Instrumentation Laboratory, Lexington, MA) and SpO2 determined using the Criticare Systems Inc. (CSI) model 503 pulse oximeter (Criticare Systems, Inc., Milwaukee, WI).

Methods: Sixty-five hospitalized patients requiring arterial blood gas analysis were assessed for SpO2 using the Nonin Onyx 9500 and CSI 503. SpO2 values and arterial blood samples were obtained concurrently. Mean SpO2 and SaO2 values were compared using the t-test for dependent samples. Pearson-product moment correlations were calculated and regression equations were evaluated to determine the slope and intercept of the regression line for paired values for SpO2 and SaO2. Means, standard deviations (SD) and ranges were also calculated for the differences between the paired results between instruments to provide measures of bias and imprecision. Limits of agreement were then calculated as the mean difference +/- 2 SD.

Results: Mean (SD) oxygen saturations obtained using the three instruments were:

There were no significant differences in mean oxygen saturations obtained using the Nonin Onynx 9500 and IL 682 (t = 1.41; p = .16). There were significant differences for the Nonin when compared to the CSI 503 (t = 2.21; p = .03) and the CSI 503 when compared to the IL 682 (t = 2.59; p = .01). There were significant correlations between values obtained using the Nonin when compared to the IL 682 (r = .83; p < .00001) and the CSI 503 (r = .91; p < .00001) and significant correlations between the CSI 503 and IL 682 (r = .76; p < .00001). Regression equations were:

Nonin Onyx SPO2 = 3.2627 + .96135 ´ IL 682 SaO2 Nonin Onyx SPO2 = 13.612 + .85881 ´ CSI SpO2

The mean (SD) difference for the paired results (bias) for the IL 682 and the Nonin was .3692 (2.11) with a range of -6 to +9%; for the CSI 503 and Nonin the mean difference was -.4615 (1.69) with a range of -6 to +3% and for the IL 682 and CSI 503 the mean difference was .8308 (2.59) with a range of -7 to + 12%. The limits of agreement between the IL 682 and Nonin were -3.85 to +4.59; IL 682 and CSI 503: -4.35 to +6.01 and CSI 503 and Nonin: -3.84 to + 2.91.

Conclusions: There was no significant difference between mean arterial oxygen saturations as measured by the Nonin Onyx 9500 pulse oximeter and the results of cooximetry using the IL 682. SpO2 results were significantly correlated with measured SaO2. There were small, but statistically significant differences between mean oxygen saturation values obtained using the CSI 503 and both the Nonin Onyx 9500 and IL 682, however, these differences were probably not clinically important.

OF-99-197