In vivo measurements of the influence of the skin on cerebral oxygenation changes measured with Near Infrared Spectrophotometry (NIRS)

Goal: To investigate the influence of skin on the accuracy and precision of regional cerebral oxygenation measurements using CW-NIRS and to reduce the inter individual variability of NIRS measurements by normalization with data from an extra wavelength. Method: Three piglets (7.8-9.3 kg) were anesthetized, paralyzed and mechanically ventilated. Receiving optodes were placed over the left and right hemisphere (C-3, C-4 EEG placement code) and one emitting optode on C, position (optode distance= 1.8cm). Optical densities (OD) were measured for 3 wavelengths (767, 850, 905 nm) (OXYMON) during stable normoxic, mild and deep hypoxemic conditions (SaO(2)=100%, 80% and 60%) of one minute in each region. This was repeated 3 times: all optodes with skin (condition 1); one receiving optode directly on the skull (2); emitting and also receiving optode on the skull (3). The absolute cO2Hb, cHHb, ctHb concentrations ( mumol/L) were calculated from the OD's and changes with respect to the SaO(2)=100% condition were estimated. Because ODs varied over a large range, the light intensity was externally attenuated to adapt to the range of the spectrophotometer. The data were then corrected for these attenuation effects and for pathlength changes caused by skin removal using the OD at the independent wavelength (lambda=975nm). Results: Removal of the skin resulted in an increase of the absorption values (average 0.25 OD in condition 2 and 0.42 OD in condition 3 with respect to condition 1). The change from normoxic to medium, and to deep hypoxic conditions produced a decrease Of cO(2)Hb (- 15, and -29 mumol/L, respectively), an increase in cHHb (+ 16, and +35 mumol/L) and in ctHb (+ 1, and +5 mumol/L). Total skin removal yielded an extra change in (C)O(2)Hb (-5, -1 mumol/L), cHHb (+8, +9 mumol/L), and ctHb (+3, +8 mumol /L). The coefficient of variability of the absolute concentration changes was considerably decreased by the normalization of densities by the density obtained at 795 nm. Conclusion: Skin and subcutaneous layers influence the regional oxygenation measurements but the estimated concentration changes are dominated by changes of the oxygenation levels in the brain. Inter individual variability can be considerably reduced by the normalization.