(T0930-02-09) A Streamlined Workflow for Visible Particulate Identification and Root Cause Mitigation in Biologic Drug Products

Purpose: The purpose of this project is to outline a streamlined workflow for the characterization, identification, and mitigation of visible particulates in biologic drug products. This process serves to save time and resources while maintaining the quality and relevance of data generation and interpretation. The approach aligns with the DMAIC (Define, Measure, Analyze, Improve, Control) methodology as a roadmap for Six Sigma, offering a structured path from particulate detection to root cause resolution. A case study is presented in which this workflow was applied to a biologic formulation exhibiting visible particulates. Particulates were observed after certain manufacturing steps and in frozen drug product vials post-thaw.

Methods: The process begins with visual inspection and microscopy imaging to assess particulate morphology. This initial assessment of the particulates, in addition to the context from where these particulates were first observed, are the first steps in defining the problem. Isolated particulates are then re-imaged and measured, followed by spectroscopic analysis—primarily Fourier-transform infrared (FT-IR) spectroscopy—to determine their biochemical composition and classify them as inherent, intrinsic, or extrinsic in nature. Imaging, measurement, and spectroscopic analysis are components of the Measure and Analysis steps of DMAIC. This characterization informs targeted follow-up studies to trace the origin of the particulates and suggest process modifications to prevent recurrence. While, in this case study, FT-IR was the primary tool for organic chemical identification, the workflow is adaptable. Raman spectroscopy is recommended for non-polar or symmetric molecules and scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDX) is useful for detecting trace metals or unexpected inclusions not visible under optical microscopy. Polarized light microscopy enables more detailed characterization of fiber types and aids in differentiating between protein and nylon when used alongside FT-IR spectroscopy. High-accuracy light obscuration (HIAC-LO) and microflow imaging (MFI) were used to assess particulate count, size distribution, and morphology in the mixing and settling studies that were performed to determine the mechanism of particulate formation and to assess viable process changes. These assays enabled the improvement and control of the issue by enabling process changes that eliminated particulate formation for future batches.

Results: Microscopy and FT-IR analysis revealed that the particulates were inherent, as their morphology and spectra were consistent with components of the drug product. Further analysis using HIAC-LO and MFI identified shear stress from mixing as the root cause of particulate formation. A settling study using HIAC-LO demonstrated that turning off the mixer at a specific stage preserved solution homogeneity while preventing particulate generation. Settling studies confirmed that the drug product remained homogeneous over the hold time without continuous mixing, supporting a process change.

Conclusion: Devising an integrated approach to identify and classify particulates led to timely actionable process improvements and control. The workflow exemplifies how analytical tools, when combined with a DMAIC framework, can drive efficient root cause analysis and enhance product quality. By defining and measuring particulate characteristics, analyzing their composition, and implementing improvements based on lab-scale studies, the process ensures robust control strategies that translate effectively to manufacturing scale.

Photomicrographs of white particulates in solution ~1 hour after thawing. FT-IR spectroscopy revealed that these particulates are consistent with lipid components of the drug product.

Microflow imaging (MFI) data suggests a significant impact of air entrainment on product aggregation and particulate formation during high mixing speeds.

Comparison of top, middle, and bottom of containers in a homogeneity assessment study. Settling trends were not identified over time. Thus, the final drug product was determined to remain homogeneous during the hold, enabling a process change for shutting off the mixers early. *T0 was homogenous and not sampled across different regions in the container.