Benchmarking of Four Near Infrared Spectroscopy Devices for Long Time Use in Neonates

Using near-infrared spectroscopy (NIRS) mixed tissue saturation can be calculated by measuring the oxygen saturation of oxygenated and deoxygenated erythrocytes in the tissue. Quality of the calculated value is not only dependent on the exposure of the measured values in the calculation, but also on external factors such as artifacts. Main object of this study was to determine whether and how the measurement quality of different devices varies in their long-term use in premature infants.

In 54 measurements, each lasting 2 hours, 4 NIRS devices were attached in pairs on the forehead of 9 cardio-respiratory stable, spontaneous breathing premature infants. Pooled meta-analysis was used to compare the correlation between regional tissue saturation to the pulse oximetry saturation per device.

The pooled random effect of all Pearson's correlation coefficients was 0.490 (CI95: 0.403-0.568) with the NIRO 200, 0.575 (CI95: 0.463-0.668) with the INVOS 5100c, 0.712 (CI95: 0.640-0.772) with the Fore-Sight and 0.638 (CI95: 0.554-0.709) with the SenSmart X- 100.

In this trial, a significant correlation between the tissue saturation and pulsoxymetry saturation was observed. The tremendous variation range among the measurements showed, however, that the measurement quality can be severely affected by unrecognized artifacts, after excluding other possible causes. None of the devices had reliable artifact detection for long-term measurements in very small premature infants. Key words: Near-Infrared-Spectroscopy, premature infants, Benchmark Test, Long-term measurements.