A key risk factor in the development of atherosclerosis is a high concentration of serum low density lipoprotein (LDL)-cholesterol. The main purpose of this study was to assess the LDL and high density lipoprotein (HDL) content in human serum by employing near-infrared (NIR) spectroscopy and multivariate calibration techniques. Initially a qualitative principal component analysis (PCA) based cluster model was generated to evaluate the feasibility of NIRS for classifying and identifying the LDL and HDL-cholesterol. Therefore TiO2 beads were used as an adsorbent for selectively immobilizing LDL and HDL-cholesterol and further analysing the incubated and washed samples via NIR diffuse reflection spectroscopy. A principle component regression (PCR) model of 24 LDL standards in a range from 500 - 3000 ppm (clinical value is 1500 ppm) and a partial least squares regression (PLSR) model of 25 HDL standards in a range from 100 - 1000 ppm (clinical value is 400 ppm) were computed. Furthermore, the wavenumber region between 4000 cm-1 and 7240 cm-1 was found comprising the main spectral information regarding the TiO2-LDL and TiO2-HDL composites. The regression coefficients (r) for LDL and HDL were > 0.99 (calibration curve) and > 0.97 (validation curve), respectively. The PCR model of TiO2-LDL showed a standard error of estimation (SEE) of 122.80 ppm and a standard error of prediction (SEP) of 121.15 ppm while the PLSR model of TiO2-HDL showed 47.70 and 47.14 ppm, respectively. In order to determine the concentration of HDL in real serum samples, LDL was removed by adding a precipitation reagent containing 10 mg/mL magnesium dextran-sulfate, followed by incubation and centrifugation. The pretreated serum samples were predicted by the PLSR model while the standard deviation (SD) from the reference to the NIR predicted values of six test samples in a concentration range from 500 - 2500 ppm showed < 10 %. These results indicate the usefulness of the NIR spectroscopy (NIRS) as a potential alternative or even supplementary clinical method for the quick determination of LDL and HDL in human serum.
Keywords: High density lipoprotein, low density lipoprotein, near-infrared spectroscopy, material enhanced, multivariate calibration, partial least squares regression
Rights & PermissionsPrintExport