Current biopharmaceutical potency testing methods are time consuming, labor intensive, and—on a sterile manufacturing line—a potential source of contamination, says the team behind a new software-based alternative.
Drug potency is a critical quality attribute (CQA). Medicines that are too weak or too strong are at best ineffective and at worst dangerous. In theory, biopharmaceutical manufacturing processes are designed to address this by ensuring products are consistent and meet defined potency specifications. In practice, however, inherent variability in raw materials and culturing means there is always a risk that products will be outside the desired potency range.
To try and mitigate this risk, biopharmaceutical manufacturers regularly test the potency of samples taken from the production line and—if an out-of-spec (OOS) result is detected—adjust the process accordingly.
Unfortunately, while effective, current potency testing methods can be challenging to carry out, according to researchers at Bayer Pharmaceuticals in California.
“Traditionally, the evaluations regarding potential process setting adjustments are made through manual, ad hoc calculations that can be error-prone, and inefficient, often requiring the valuable time of subject matter experts (SMEs).
“Therefore,” the authors write in a new study, “there is a need to streamline these adjustments leveraging statistical modeling workflow and software tools, allowing for more accurate and efficient decision-making processes.”
Time saving
The Bayer team’s alternative consists of a framework for OOS risk assessments implemented as a cloud-based software application. The software is designed to automate data acquisition, pre-processing, statistical calculations visualization, and reporting.
The idea is to model information against both historical data and a range of process parameters—everything from thawing processes through to formulation buffer concentration—and identify potency deviations as efficiently as possible.
The authors describe their software as a “digital assistant” to help experts evaluate potency targets in sterile fill-finish processes for therapeutic proteins. Its key advantage, they say, is the ability to link out-of-spec results to specific process parameters.
The approach shows real promise, according to the authors, who write that, “Overall, the use of the application has resulted in a reduction of the total effort required to conduct such assessments.
“Specifically, key stakeholders in the manufacturing science and technology (MSAT) department estimate the overall effort reduction from a week to about an hour to acquire the data and perform the statistical analysis. In addition to efficiency improvement, consistency, and accuracy are other key benefits.”
The Bayer team also suggested their approach could be applied more widely, writing: “The statistical framework along with the cloud pipeline and interactive user interface can be easily re-repurposed and adjusted as needed to accommodate the evaluation of OOS risk for other CQAs, for example, total protein.”
