Application of near-infrared spectroscopy combined with multivariate analysis in monitoring of crude heparin purification process.

Ion-exchange chromatography is a widely used purification technology in the heparin manufacturing process. To improve the efficiency and understand the process directly, a rapid and equally precise method needs to be developed to measure heparin concentration in chromatography process. Here, two robust partial least squares regression (PLS-R) models were established for quantification of heparin based on the near-infrared (NIR) spectroscopy with 80 samples of adsorption process and 76 samples of elution process. Several variables selection algorithms, including correlation coefficient method, successive projection algorithm (SPA) and interval partial least squares (iPLSs), were performed to remove non-informative variables. The results showed that the correlation coefficient of validation (Rp) and the residual predictive deviation (RPD) corresponded to 0.957 and 3.4472 for adsorption process, 0.968 and 3.9849 for elution process, respectively. The approach was found considerable potential for real-time monitoring the heparin concentration of chromatography process.