MP009: Reliability and interchangeability of measures of two tissue oximeters in healthy volunteers

Abstract

Introduction: Near-infrared spectroscopy (NIRS) is a non-invasive, continuous and painless method of monitoring oxygen saturation of hemoglobin in any given superficial tissue. Given that hemodynamic instability can affect the oxygen saturation, NIRS could prove to be an interesting tool in quantifying tissue oxygenation, consequently guiding clinical management. The aim of this study was to compare the reliability of two commonly used tissue oximeters, the INVOS 5100c from Covidien and the Equanox 7600 from Nonin. We postulated the Equanox (a more recent tissue oximeter) would have a better reliability than the INVOS. As a secondary outcome, we evaluated whether the measures given by the two oximeters were comparable. Methods: The study population was composed of healthy adult volunteers. Three measurements were taken at six different sites on both sides of the body in a randomized order. Two different sensors were used for each measure. From these measures, two intra-class correlations (ICC) - one inter-sensor and the other intra-sensor - were calculated for each device and compared using the Fisher’s r-to-z transformation method. An additional inter-device ICC was also calculated. We considered ICCs over 0.75 as an indicator of good reliability, while ICCs under 0.40 were considered to represent poor reliability. The sample size was calculated based on the calculation of a unidirectional confidence interval for a parametric ICC. Expecting a 0.75 ICC value, we concluded that 53 participants needed to be recruited in order to attain 80% power and a range of 0.1 towards the low values. Results: Fifty-three healthy volunteers (27 men and 26 women) with a mean age of 31 years (standard deviation 10) were recruited. We found no differences between the repeatability of the INVOS and the Equanox for both inter and intra-sensor reliability (ICC=0.94 (95% confidence interval (CI) 0.86-0.97) versus ICC=0.92 (95%CI 0.86-0.95), p=0.42 and ICC=0.94 (95%CI 0.89-0.96) versus ICC=0.96 (95%CI 0.93-0.98), p=0.21, respectively). However, when compared directly, we found that the readings produced by the two oximeters varied considerably (ICC 0.18 (95%CI -0.10 to 0.43). Conclusion: When taken individually, both tissue oximeters displayed good inter and intra-sensor reliability. However, they oximeters displayed poor inter-devices agreement, their readings varying considerably amongst each other.