QbD in API Manufacturing

With a quality-by-design approach, robust processes consistently deliver quality product. Small drug manufacturers looking to adopt a quality-by-design (QbD) approach aim to scientifically determine product and process characteristics that will meet specific criteria set after careful analysis of the intended drug application. These critical quality attributes (CQAs) of the final drug product (and often of the API) include the physico chemical properties and performance behavior of the formulated drug (drug substance). The manufacture of small-molecule APIs often requires the production of multiple intermediates using different processes. To consistently achieve these CQAs, it is necessary to ensure that each of these processes is robust. Critical process parameters (CPPs) are, therefore, important components of an effective control strategy for ensuring that the API process consistently delivers product with the appropriate CQAs.

Linking CQAs with CPPs
While true CQAs are found in the final specification for a drug product, there are several different aspects of a formulated drug, and particularly the API, that are affected by process parameters. These aspects, therefore, must be linked to the CPPs for each manufacturing process required for the preparation of an API, according to Chris Senanayake, vice-president of chemical development for Boehringer Ingelheim Pharmaceuticals. These characteristics include, but are not limited to, total purity, individual impurity levels, the polymorph(s), solvates, pH, water content, and particle size.

For the API, the most common CQAs are related to impurities from chemical processes and the solid-state properties (e.g., particle size, polymorphism) of the drug substance, according to Gert Thurau, group leader, Technical Regulatory at Roche. It is essential to control impurities for the safety of the drug product, while control of the solid-state properties of the drug substance is necessary to achieve consistent product properties, including bioavailability. “Once these attributes have been defined, it is possible to define what is important to the process and build a process that comes closer to quality by design and is, therefore, highly robust,” he says. The assay and impurity profile of an API are essentially inter-related and are important because low impurity levels translates to safety for patients, while a high assay indicates appropriate efficacy, according to Vishwanath Nadig, associate director of quality assurance for Dr. Reddy’s Custom Pharmaceutical Services (CPS). Nadig adds that moisture content is also important because, in some cases, the presence of moisture may accelerate degradation of the drug substance.

These CQAs should be selected such that if the final product has these attributes, then it will work for its intended purpose. CPPs should then be selected because they have a relationship to the final CQAs and, when controlled, will assure the manufacture or a product with the right CQAs, according to Senanayake. “CPPs are the process parameters that affect the critical quality attributes of the API, and it is important to determine the range for each critical process parameter expected to be used during routine manufacturing and process control,” adds Mark TePaske, director of regulatory affairs, quality, and compliance for Cambrex. Most importantly, it is essential to establish appropriate CPPs for each individual process step because they will ultimately affect the quality of the drug, which in turn can impact efficacy, according to Nadig. “Establishing and meeting CPPs and CQAs allows quality to be built into the process and affords robustness beyond release testing,” adds TePaske.  He also notes that well-defined CPPs ensure that a process operates in a state of control and in a validated state, which is crucial, because process validation is a cGMP requirement.

The definition of the CQAs for the final drug product and API is the first step in a comprehensive risk-based approach to product and process development, according to Thurau. “The starting point of the analysis is typically the quality target product profile, which leads to defining the critical quality attributes for the final product, and in turn, the critical quality attributes for the drug substance. Since the most important CQAs of the final API are the process-related impurities, which mostly result from the starting and raw materials, it is essential to understand the chemistry of the individual process steps,” he explains. In addition to risk assessment, Senanayake adds that prior knowledge, designed experiments, and regulatory expectations all play a role in determining CQAs and CPPs.

Risk-assessment approaches

Once CQAs are defined, typically during the initial scale-up phase, it is necessary to identify the appropriate CPPs. “At Boehringer Ingelheim, we believe that using a design of experiment (DOE) approach is the best method for demonstrating the interdependency of CPPs and the CQAs for individual process steps,” Senanayake says. Along with general risk assessments, the use of design failure mode and effects analyses (FMEA) is also effective for defining critical parameters of drug products, drug substances, and production processes, according to Nadig. “When applied appropriately, these two approaches are sufficient for defining the important attributes,” he says.

While orthogonal experiments can also be used, Senanayake notes that with DOE, knowledge is gained about the design space for each process that cannot be obtained using screening experiments, although they are useful when determining certain aspects of processes, such as the most effective solvents or catalysts. TePaske adds that once defined, critical attributes should be affirmed experimentally and control limits established, with processes, procedures, equipment and/or facilities revised as needed.

It should be noted that the risk-assessment methods used today are much more rigorous than the processes used in the past. “Historically, process and CQA definition was driven by the process development chemist, and the depth and breadth of the investigations were chemist-dependent,” observes TePaske. As a result, process development was primarily viewed as a laboratory exercise.  He goes on to note that current approaches are more disciplined, and studies are defined by multidisciplinary project teams. “These teams review and reconcile findings and risk assessments, and the concerns of the production, materials, quality, engineering, and other departments involved in manufacturing are integrated into the development process.  In effect, CQA definition and process development have evolved from a laboratory exercise to a global preparation for process validation and commercial GMP manufacturing.” He concludes that this team-oriented approach is the best practice because it integrates the needs of all of the disciplines involved in commercial manufacturing and virtually eliminates the possibility that CQAs and CPPs that cannot be performed in the available facilities and equipment will be defined.

Connecting CPPs to CQAs
Based on the results of the quality risk assessment, various reaction and work-up conditions can be investigated, and the products from each process step can be carefully analyzed with regard to their impurity profiles and other properties, according to Thurau. “For example, the structures of the impurities can be determined, and consequently the process conditions can be adjusted to minimize the impurities and/or their fate in downstream steps can be determined. The result is the elimination of the impurities or the determination of safe limits for them,” he explains. He stresses that it is important in this part of the process to be systematic, and to consistently maintain a link between the CPPs being evaluated and the drug product and drug substance CQAs.

Barriers to developing CPPs
While there are many benefits to a QbD approach and the establishment of CPPs that are linked to final product and drug substance CQAs, there are challenges to implementing such an approach. For Senanayake, the time and resources required remain the biggest barrier to obtaining a useful understanding of a process. “These barriers must be overcome, though, because having thorough knowledge of a process once it has reached commercial manufacturing is critical for success, and therefore, it is necessary to take the time to fully understand the design space,” he asserts.  Thurau agrees that the major limitations are practical in nature and can typically be overcome. He does, however, note that there are some open questions on the definitions and differences in the applicability of the concept by various health authorities. “As registrations for products developed and manufactured using the QbD approach become more detailed and explicit about criticality, the hope is that the increased knowledge will lead to more flexible approaches to change management (do and tell) as opposed to the current set approach (tell and do).

Both Nadig and TePaske point to the need for participation of knowledgeable team members from across different functions of the organization as a challenging aspect of QbD and the establishment of effective CQAs and CPPs. “The multidisciplinary approach requires up-front involvement of disciplines whose primary job responsibility has historically been manufacturing and requires a new way of thinking,” says TePaske. He also adds that the risk-assessment approach often focuses on what can be done, which can stifle innovation.

Practical use of CPPs

Despites these barriers and limitations, many companies have recognized the value of a QbD approach and the importance of linking critical process parameters for individual process steps with the critical quality attributes of small-molecule APIs and the corresponding formulated drug products. Boehringer Ingelheim, for example, is monitoring process robustness at each development stage and focusing on understanding how this robustness translates to the commercial scale using batch histories, according to Senanayake. In addition, the company is trying to complete DOEs on the important steps in its manufacturing processes. “Our success to date is demonstrated by our validation efforts and the control charts we have established for individual steps,” he says.

At Dr. Reddy’s CPS, when defining CPPs for each process step, a cross-functional team meets to discuss the key sensitivities of the process parameters and material attributes, which are first weighted in importance, and then tolerances on the variables are established, according to Nadig. DOE screening is then performed to identify the critical parameters, followed by optimization and definition of the design space to understand what will happen to the CQAs under different parameter settings. “Importantly, we are trying to engage senior members from cross-functional teams during these pre-planning meetings so that an exhaustive risk assessment is conceptualized and a mitigation plan worked out,” he comments. The company also emphasizes the importance of clear communications, participation with the customer, and giving weight to all stakeholder concerns throughout the process.

Potential consequences
The consequences of not establishing CPPs that are effectively linked to CQAs can be severe. “If the proper controls for an intermediate step in the manufacturing process are not in place, then it is possible for a material to be produced that does not meet specifications for unknown reasons. Such a situation can in turn result in a recall if the intermediate is carried through the entire commercial process. In addition, without proper controls, processes often deliver poor quality product or product of inconsistent quality,” asserts Senanayake. In addition, according to Thurau, if inappropriate process parameters or in-process material attributes are established, the quality of the final product may appear high, but it is the result of testing quality into it, and not of process design and process understanding. Finally, TePaske stresses that operation in a state of control is fundamental to process validation and continued process performance in a validated state.