|| Checking for direct PDF access through Ovid
We determined whether two different devices for measuring near-infrared spectroscopy (NIRS)—the INVOS 4100 and the NIRO 300—produce similar cerebral oxygenation data during the CO2 challenge test. Nineteen patients anesthetized with sevoflurane, 67% nitrous oxide in oxygen, and fentanyl were studied. A series of measurements of regional cerebral oxygen saturation (rSo2), measured by the INVOS 4100, and tissue oxygen index (TOI), measured by the NIRO 300, were performed in the following conditions: 1) normocapnia (Paco2, 35–45 mm Hg); 2) hypocapnia (Paco2, 25–35 mm Hg); 3) normocapnia; and 4) hypercapnia (Paco2, 45–55 mm Hg). Hemodynamic variables, including arterial blood gases and cerebral blood flow velocity, were measured at the same time with transcranial Doppler. The values and percentage changes of rSo2 and TOI were compared by using regression analysis and Bland and Altman analysis. The rSo2 showed a significant positive correlation with TOI (r = 0.58, P < 0.01). The percentage change of rSo2 also showed a significant positive correlation with the percentage change of TOI during the CO2 challenge (r = 0.85, P < 0.01). Bland and Altman analysis revealed a bias of −0.5% with 2 sd of 15.6% when comparing the rSo2 value with the TOI value, and it showed a bias of −3.4% with 2 sd of 15.2% when comparing the percentage change of rSo2 with the percentage change of TOI, indicating unacceptable disagreement of these data. These results indicate that cerebral oxygen saturation and its relative change during the CO2 challenge may vary depending on the type of NIRS used. Because the measurement technique and algorithm were different in each device, we should carefully consider the clinical application of the values produced by NIRS.