Multivariate calibration models for glucose and urea in bovine blood ultrafiltrate with near infrared spectroscopy

ANYL 324

Min Ren, Department of Chemistry, Optical Science and Technology Center, Iowa City, IA 52241 and Mark A. Arnold, mark-arnold@uiowa.edu, Department of Chemistry and Optical Science and Technology Center, University of Iowa, 230 IATL, Iowa City, IA 52242.
Near-infrared (NIR) spectroscopy offers a painless and convenient method for measuring glucose concentrations noninvasively, thereby enhancing frequent diabetes monitoring and delaying medical complications. The work presented here demonstrates the potential of noninvasive ultrafiltrate glucose measurements by analyzing samples collected from bovine blood.

Blood samples were passed through a KrosFlo hollow fiber ultrafiltration module coupled with a Materflex peristaltic pump. Standard solutions were prepared with randomized concentrations of glucose, urea and triacetin in blood ultrafiltrate. NIR spectra were collected over the combination spectral region with Nicolet Nexus 670 Fourier transform spectrometer. Single beam and absorbance spectra were used to build multivariate calibration models for glucose and urea.

Results demonstrate excellent calibration performance for measuring glucose and urea. Pure component selectivity analysis and correlations between pure component spectra and the corresponding PLS regression vectors indicate the calibration models built from single beam and absorbance spectra are based specifically on glucose spectral information.