Process Control On Near Infrared Spectroscopy In Human Fibrinogen Production Process

Human Fibrinogen(Fg) is the main component of plasma, content of up to2-4g/L. In the final stage of human coagulation reaction, it reacts with the human coagulation factor thrombin XⅢ and Ca2+to format the fibrin gel, the physical components of the blood surrounding where to stop the bleeding. In clinical practice, it has corrective action due to the lack of Fg clotting dysfunction, so it is mainly used for prevention and treatment of congenital Fg missing or deficiency, as well as due to severe liver damage, cirrhosis, disseminated intravascular coagulation, postpartum hemorrhage, major surgery, trauma, internal bleeding as a result of acquired Fg deficiency.As of September2014, Domesticmanufacturers that obtained Fg production approval and having the ability to produce it are no more than10. On the same time, since the approval of the new plasma stations is more difficult, lack of raw plasma to supply Fg, current domestic production cannot meet the needs of the market. Fg is also one of the drugs strategic reserved, the state need a certain reserve every year. Blood products have high cost and high risk, so control production processes, improve product yield, and ensure product activity for the development of the whole blood products industry is important.At present, the company is still in the clinical application stage of Fg production,lack the understanding of the critical control points, and process control part is weak. Therefore, control of critical control pointsis necessaryfor Fg production process.Near-infrared spectroscopy technology(NIRS) is one of the important process control tools, with the characteristics such as fast, does not destroy of the sample, having successfully used in all aspects of the pharmaceutical industry’s raw material screening, mixing, preparation, etc. But the application of the blood products analysis is still blank.This thesis selected the chromatography flowthrough process, the ultrafiltration process and the lyophilized preparation process in the Fg production process as the study objects, using NIRS combined with chemometric methods, established a process control method for Fgproduction process.Firstly, we selected flow through liquid of the chromatography process for the study. Based on the flow through liquid, we prepared a series of samples of different concentrations, whose number is60, acquired its near infrared spectroscopy by transmission module. In the process of model building, K-S method is first used to divide the sample into40calibration set and20validation set, then established Autoscale as a pretreatment method for spectrum, and selected Forward iPLS as the spectral variable selection method. The parameters of theeventual model are as follows:R2=0.995, RMSEC=0.1911, RMSECV=0.2245, RMSEP=0.1662. The method can test the protein content of the flow through liquid fast and online. Once used in the production, it can analyze the saturation in order to determine the best chromatography solution.Secondly, we collected the liquid after the ultrafiltration process, and then prepared a series of samples of different Fg concentrations. The number of the samples is54, and also acquired their near infrared spectroscopy by transmission module. In the modeling process, we established the first derivative+Autoscale as the pretreatment method for spectrums, and GA as the spectral variable selection method. The parameters are as follows:R2=0.999, RMSEC=0.3575, RMSECV=0.4894, RMSEP=0.5208, which has an excellent accuracy and precision. In the next step, we will collect a large number of liquid test samples of different purity of the original fiber in the production process, in order to establishNIR quantitative model which can directly predict the Fg purity, and realize direct preparation in the ultrafiltration process.Thirdly, we selected lyophilized completed products as the study object. We prepared60samples having different moisture, and then acquired their near infrared spectroscopy by integrating sphere diffuse reflectance module. The reference moisture was tested by Karl Fischer titration. In the modeling process, we examined pretreatment methods such as MSC, SNV, first derivative, second derivative, and use a variety of methods, compared the correlation coefficient method and iPLS two spectral variable selection method. Finally, we determined MSC+second derivative and the correlation coefficient method. The parameters are as follows:R2=0.881, RMSEC=0.4426, RMSECV=0.8370, RMSEP=0.6311. This result indicates that the near-infrared spectroscopy technology can effectively freeze-dried to a moisture content of fiber raw product were measured, confirming the effectiveness and feasibility of the method. This model can be used as basis for online monitoring of fiber original moisture content of freeze-drying process, and after freeze-drying equipment related to the transformation, we can realize the real-time monitoring of product moisture content by using optical fiber, and optimist the lyophilization process greatly.The innovations of this paper are:(1) Established the quantitative model for protein detection of the chromatography flow-through liquid for the first time;(2) Established the quantitative model for Fg detection of theultrafiltration process for the first time;(3) Established the quantitative model for Fg freeze-dried moisture content testing for the first time.