PAT SUPPLEMENT


Case Study II: Review of the historical in-process control data from a cell cultivation process using multivariate data analysis The scope of this work is to improve the understanding and the reduction of the batch to batch variability in terms of antibody yield for a cell cultivation process. Specifically, the upstream manufacturing process (seed and main bioreactor) is discussed herein. All the batches in scope of the MVDA model provided drug


substance batches of the same quality. Nevertheless, some variability in the antibody yield at the end of the harvesting phase was observed. MVDA has then been used to better understand the differences in yield at the end of the harvesting phase. Furthermore, it has been used to investigate whether there is any correlation between the upstream behaviour of the cell culture and the quality attributes of the molecule produced. The modelling approach used to develop the MVDA models .


includes time-dependant variables as described by S. Wold et al1


Particularly, MSPC and batch-level multivariate statistical process control were used here. The process variables, typically recorded during the different phases of the upstream manufacturing process, were included in the MVDA model. The batches that produced the highest amount of antibody at the end of the harvesting phase were


MVDA contributes significantly in a structured way to evaluating and visualising data stemming from lab and pilot scale ”


“ During process development,


defined as reference (golden). All other operations in scope of the present study have been assessed by comparing them against the MVDA model, based on reference batches. As an outcome of this work, reference trajectories based on the golden batches were established for the most important process variables, namely viable cell density, medium feeding rate and cell aeration rate. Drug substance release attributes have also been modelled to establish whether there is any correlation between the behaviour of the cell culture and the quality attributes of the molecule produced. Figure 3 (page 8) is a MSPC chart generated from the cultivation


data for the so-called golden batches (golden MSPC model). Cross- validation identified two principal components scores (i.e. t1 and t2) for this phase. The first principal component, which captures most of the variability within the dataset, is shown in Figure 3 on page 8. The explained variability is about 90 per cent. The green line is the average or expected process signature for the golden batches. The red curves (±3 standard deviations from the golden average) represent the variability of the golden batches and therefore lines correspond to the statistical process limits. The golden batches were used to define reference trajectories


for the different cultivation variables. All the non-golden batches were assessed by comparing them against the golden MSPC model to better understand why these batches showed lower productivity. In summary, the most influential variables appeared to be viable cell density, medium feeding rate and aeration rate through the sparger.


www.europeanpharmaceuticalreview.com


European Pharmaceutical Review Volume 16 | Issue 6 | 2011


PROCESS ANALYTICAL TECHNOLOGY


Discover the benefi ts of Synchronized PAT Solutions.


For Sartorius PAT is much more than just an expensive sensor integrated in a process. PAT is a risk based strategy for design- ing, analyzing and controlling manu facturing processes.


Based on the defi nition of the Critical Quality Attributes (CQAs) of intermediates or the fi nal products the Critical Process Parameters (CPPs) are identifi ed. With a broad range of process analyzers, process control and software tools Sartorius’ PAT offers solutions to measure and control these CPPs to optimize your process.


SYNCHRONIZED PAT SOLUTIONS www.sartorius.com/pat turning science into solutions


Sartorius Stedim Biotech USA +1.800.368.7178 | Europe +49.551.308.0


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17