The Research On Calibration Strategy And Quality Monitoring Of Plasma Alcohol Precipitation By Using Near Infrared Spectroscopy

In recent years,with the development of China's biomedicine industry and people's increasing attention to drug use and health,the public's demand for drug product quality is getting higher and higher.At present,there are some problems in our country's blood products.For examples,the plasma comprehensive utilization degree is not high and the blood products production quality monitoring method lags behind.These all restrict the blood products product quality upgrades.Therefore,the development of intelligent production lines to process analytical technology(PAT)as the main quality control means to improve product quality has become a research focus of the pharmaceutical industry.Near infrared spectroscopy(NIRS),as a kind of PAT,has been applied to the quality analysis of blood products.However,the model calibration strategy,the design of the stable spectrum acquisition preprocessing system for plasma alcohol precipitation process analysis and quality monitoring control has not been reported.Therefore,this paper carries out the following research on the above issues:(1)Research on a new calibration strategy based on design of experiment(DoE)with directional and random optimization.Methods:four common and important preprocessing steps are considered:baseline correction,scattering correction,smoothing correction and scaling correction.After the best pretreatment strategy was selected by DoE combined with Bootstrap method and factorial analysis,the influence of variable selection methods VIP,CARS,UVE and CC on the blood product model was compared.Results:general rules of NIRS for modeling in the complex system of blood products for drugs were revealed,the model quality was improved and the model evaluation parameters were improved,and the progress of intelligent production process of blood products for drugs was promoted.(2)Research on the intelligent judgment system for the end points of liquid reaction.Methods:a high efficiency on-line filter,steady pressure and steady flow defrothing system were built.The influence of secondary cast filter elements with different materials and different pore sizes was investigated by using alcohol precipitation plasma as the experimental material.The feasibility and reliability of the system were investigated by using ethanol solution with different concentrations as the experimental material.Results:this chapter builds tosses type terminal intelligent judgment system.Alcohol precipitation processes that to eliminate the influence of bubble measurement process becomes more stable.External validation of the model and validation of terminal intelligent judgment system of alcohol precipitation processes show that the spectral quality good,model prediction precision,accurate process monitoring,it is proved the feasibility and reliability of the device.(3)On-line detection studies of plasma alcohol precipitation process.Methods:this part combines the calibration strategy with the intelligent judgment system,and applies it to the determination of endpoints and the content determination of ethanol,albumin and globulin in the actual plasma alcohol precipitation process.Results:the RMSEP,Rp and RPD values of the optimal ethanol model were 1.79g·L-1,0.9796 and 4.95,respectively.The RMSEC,RMSECV,RMSEP,Rc,Rcv and Rp values of the albumin model were 1.4677g·L-1,2.0172g·L-1,2.1385g·L-1,0.8708,0.7372 and 0.9133,respectively.The RMSEC,RMSECV,RMSEP,Rc,Rcv and Rp values of globular protein model were 0.8656g·L-1,1.3786g·L-1,0.8534g·L-1,0.9751,0.9355 and 0.9835,respectively.Finally,the end points were determined by using a simple system to predict the content of ethanol in a complex system and the content of albumin and globulin in alcohol precipitation process.