Measurement of pH in whole blood by near-infrared spectroscopy

Whole blood pH has been determined in vitro by using near-infrared spectros copy over the wavelength range of 1500 to 1785 nm with multivariate calibra tion modeling of the spectral data obtained from two different sample sets. In the first sample set, the PH of whole blood was varied without controll ing cell size and oxygen saturation (O-2 Sat) variation. The result was tha t the red blood cell (RBC) size and O-2 Sat correlated with pH. Although th e partial least-squares (PLS) multivariate calibration of these data produc ed a good pH prediction cross-validation standard error of prediction (CVSE P) = 0.046, R-2 = 0.982, the spectral data were dominated by scattering cha nges due to changing RBC size that correlated with the PH changes. A second experiment was tarried out where the RBC size and OZ Sat were varied ortho gonally to the pH variation. A PLS calibration of the spectral data obtaine d from these samples produced a pH prediction with an R-2 of 0.954 and a cr oss-validated standard error of prediction of 0.064 pH units. The robustnes s of the PLS calibration models was tested by predicting the data obtained fi om the other sets. The predicted pH values obtained from both data sets yielded R-2 values greater than 0.9 once the data were corrected for differ ences in hemoglobin concentration. For example, with the use of the calibra tion produced from the second sample set, the pH values from the first samp le set were predicted with an R-2 of 0.92 after the predictions were correc ted for bias and slope. It is shown that spectral information specific to P H-induced chemical changes in the hemoglobin molecule is contained within t he PLS loading vectors developed for both the first and second data sets. I t is this pH specific information that allows the spectra dominated by PH-c orrelated scattering changes to provide robust pH predictive ability in the uncorrelated data, and visa versa.