Manage scale-up and technology transfer risk with better process understanding
Introduction
Pharmaceutical industry has seen a continuous evolution in the manufacturing processes. While the development of a product or a process at a lab or kilo-lab scale has its own importance and challenges, taking it through the pilot scale up to the production scale is altogether a different thing. Some of the challenges may be easily recognized, such as:
- The transport phenomena (e.g. mass transfer, heat transfer) at the production scale are completely different than those at the lab or kilo-lab scale. Hydrodynamic behavior changes at the production scale.
- The equipment available for the large scale production may be completely different than the one in which the process was developed. e.g. the type of an impeller for the experimental purpose in a crystallizer may not be present in the actual scale operation. This is more evident in the case of the organizations that run multiple product campaigns.
- Often, the scale up is carried out merely by maintaining the geometric ratios across the scales. While this may work in some cases, a closer look would reveal that there is an overdesign at one end, or the desired throughput is not obtained.
All these challenges point out toward a common solution: a more systematic and predictable way to design a process or an equipment. Over the past decade or so, the traditional “quality-by-testing” approach of trial-and-error is being consciously replaced, partly or fully, by a more scientific approach across all aspects of pharmaceutical engineering – discovery, development and manufacturing. This “Quality by Design (QbD)” paradigm makes it imperative that the entire development lifecycle be carried out in a systematic, science-based and quality risk-driven manner, so that the quality of the final product is no longer left to chance. A litany of tools and supporting technologies to support QbD have been identified – chief amongst them being Process Analytical Technology (PAT), statistical DoE, risk assessment tools and modeling/simulation. In this contribution, we will focus on how modeling-driven methods can be effectively used to support QbD in scale up and technology transfer.
Approaches for Scale-Up and Technology Transfer in Pharmaceutical Industry
At the heart of any process is the equipment. The geometric configuration of an equipment influences the final product in a great way. Equally important are the operating conditions. For instance, in a stirred bioreactor, the viable cell count (VCC) is a complex function, not only of the impeller type and diameter, but also of the rotational speed, along with many other factors, such as temperature and pH. In the QbD terminology, a process can be characterized in terms of critical process parameters (CPPs), critical material attributes (CMAs) and critical (product) quality attributes (CQAs). In general, CQAs are influenced by CPPs and CMAs. The ranges of values within which these CPPs and CMAs are maintained to achieve a target range of CQA values, is called the “Design Space”.
A fairly common scale up/scale down and technology transfer approach in the pharmaceutical industry focuses on using data-driven relationships between CPPs and CQAs obtained from lab and pilot scale experiments. Response Surface Methods (RSM) and/or multivariate statistical analysis (MVA) methods are commonly employed in this approach. It is also common practice to use measurable CPPs (agitator RPMs, fill levels, solids loadings, gas flow rates etc.) as the basis for such relationships. Using such CPP-CQA relationships is a flawed approach to scale-up: because there are scale and geometry dependent aspects that are inherent in them that prevent a state of true process understanding from being achieved. In fact, the ICH-Q8 R2 guidelines clearly state that scale independent quantities are preferred when describing ‘Design spaces’ for a given process.
An alternative approach, that considers the effect of operating variables (and material attributes) on the processing micro-environment is recommended in this article. The processing micro-environment aspects are accounted for by introducing the concept of Critical Process Metrics (CPMs). Examples of CPMs include shear rates (at various locations in a stirred vessel), mixing time scales (macro, meso and micro), gas-liquid mass transfer coefficients (in gas-liquid contactors), tablet residence times (in pan coaters) and dimensionless spray flux (in granulators). From these examples, it is clear that CPMs are unique to a given unit operation. CPM-CQA relationships can therefore be used to characterize a given unit operation in an entirely scale independent manner. It is also possible to derive asset-specific CPP-CPM relationships using computational and/or first-principles based methods. For instance, for a given combination of geometrical configurations, RPM, liquid volume and physical properties, the shear rate (which is a CPM) in a stirred vessel can be predicted using computational fluid dynamics or other means. Thus, using a combination of CPP-CPM and CPM-CQA relationships, activities related to technology transfer (scale up, scale down, equipment fit analysis etc.) can be conducted in a manner that is at once physically meaningful and operationally seamless.
This methodology is described using case studies in the next few articles: