1. Expansion of Process Characterization Study
Conventional process characterization approaches and studies enable the accumulation of product and process understanding. However, it is observed that routine process characterization cannot be performed for all unit operations, such as those where a precise scaled-down model cannot be established (e.g., viral inactivation, formulation blending processes).
Worst-case condition study serves as an extension of conventional process characterization in such scenarios. Defined in PDA Technical Report No. 60, a worst-case condition refers to a set of process conditions encompassing the upper or lower limits of process parameters, or conditions that pose the maximum potential risk of process or product failure compared to ideal process conditions; notably, these conditions do not necessarily lead to actual product or process failure.
In process characterization, worst-case condition studies can be applied for the following purposes:
Perform process characterization (PC) under worst-case conditions to build product and process knowledge;
Conduct linkage studies under worst-case conditions to determine process robustness.
2. Principles of Worst-Case Condition Study
The strategy of worst-case condition study is to operate unit operations that affect specific Critical Quality Attributes (CQAs) under worst-case conditions, and evaluate whether the output CQAs remain within the normal acceptable process range.
If the results fall within the acceptable range of CQAs, it verifies that the process is feasible across the entire operating range of the unit operation.
If the results exceed the acceptable range of CQAs, it is necessary to narrow the acceptable range of one or more process parameters for the unit operation.
3. Determination of Worst-Case Conditions
Worst-case conditions of unit operations can be determined via two approaches:
Determination Based on Engineering Knowledge of Unit Operations
For instance, in the low-pH viral inactivation unit operation:
From the perspective of viral inactivation efficacy, high pH, short hold time, low temperature and high protein concentration are defined as worst-case conditions.
From the perspective of product quality (e.g., aggregate formation), low pH, long hold time, high temperature and high protein concentration represent worst-case conditions.
Determination Based on Process Characterization Study
Process characterization can identify process parameters impacting specific CQAs, and the setting direction of process parameters can be predicted via regression models for worst-case conditions.
Case Study: Process characterization was conducted for the Protein A affinity chromatography step, with investigated parameters including load capacity, wash volume, elution pH and flow rate.
A regression model for Host Cell Protein (HCP) derived from PC results is established as follows:
HCP (ppm) = 5823 – 227.4×Load – 585.6×pH – 2.1×Flow Rate + 4.4×Load²
Accordingly, the worst-case conditions for HCP removal in the Protein A affinity chromatography step are defined as low load capacity, low pH and low flow rate; wash volume shows no impact and can be set at the central point.
4. Process Characterization Under Worst-Case Conditions
Taking the low-pH viral inactivation step as an example:
For product aggregate characterization: Set conditions as pH 3.2, protein concentration 35 g/L and temperature 25°C, and investigate the impact of different process hold times on aggregate formation. The acceptable range of the aforementioned process parameters is determined based on the acceptable limit of aggregates.
For viral inactivation validation: Set worst-case conditions as pH 4.0, protein concentration 35 g/L and temperature 15°C, and conduct process characterization to verify robust XMuLV inactivation within a process hold time range of 60–180 minutes.
The process design space for the low-pH viral inactivation step can be finalized based on the above two sets of worst-case condition studies.
5. Linkage Study Under Worst-Case Conditions
When certain CQAs are affected by multiple unit operations, linkage studies under worst-case conditions are required to demonstrate the robustness of the overall manufacturing process.
Worst-case conditions for each unit operation are determined following the aforementioned methods. Some unit operations possess inherent process robustness, i.e., no process parameters exert impact on the targeted CQAs; such unit operations can be operated at target conditions in linkage studies.
If the final product achieves acceptable quality after linkage experimental study, the validity of the design space across the entire process is confirmed. If unacceptable product quality is obtained, further constraints on process parameters shall be implemented via the following measures:
Narrow the acceptable range of one or more process parameters in single or multiple unit operations to prevent CQAs from exceeding their Critical Quality Attribute Target Ranges (CQA-TR);
Leverage the understanding of parameter magnitude of influence and interaction effects to prohibit the use of parameter combinations that may yield unacceptable outcomes in future production. The latter approach allows greater operational flexibility in commercial manufacturing, yet requires defining the correlation between CQA-impacting process parameter values and establishing a well-defined process to properly manage cumulative or sequential variations of process parameter targets within the PALM framework.
6. Optimization of Worst-Case Condition Study
The conventional worst-case linkage method described above is highly conservative, as it combines worst-case parameter settings across the entire manufacturing process. While linkage study under full worst-case conditions has theoretical value in proving the process capability to deliver acceptable product quality, its practical applicability is limited. Running all process parameters at their worst-case setpoints simultaneously is unrealistic for routine commercial production.
Therefore, optimized worst-case condition study methodologies are proposed to achieve more scientific and rational process assessment:
Unit Operation Worst-Case Linkage Study: A linkage study to evaluate the worst-case combination of process parameter setpoints for a single unit operation. The collected stream from the unit operation operated at worst-case setpoints is subjected to downstream processing under representative process conditions with target parameter setpoints. This study is repeated for all relevant product quality attributes.
Model-Based Linkage Calculation: Complementary to experimental unit operation worst-case linkage study. This method adopts the worst-case combination of process parameter setpoints for a single unit operation, and employs process models to predict downstream processing outcomes under representative process conditions with target parameter setpoints.
Challenging Study: In the absence of elevated impurity levels in upstream harvest streams, challenging studies on process-related impurities can be applied to demonstrate the process robustness for the removal of specific impurities (e.g., residual DNA).
Fine Purification Step Skip Study: By conducting experimental trials with one specific unit operation skipped, this study verifies the sufficient removal of targeted CQAs and thereby confirms overall process robustness.
7. Summary
Worst-case condition study serves as a critical extension of conventional process characterization, enabling in-depth acquisition of product and process knowledge. With defined worst-case conditions, both process characterization studies and linkage studies can be implemented. Results from linkage studies indicate overall process robustness and provide feedback for rational design of process parameters。
