Raman spectroscopy as an alternative rapid microbial bioburden test method for continuous, automated detection of contamination in biopharmaceutical drug substance manufacturing

• Published in: Journal of applied microbiology Vol. 135; no. 8
• Main Authors: Masucci, Erin M, Hauschild, James E, Gisler, Helena M, Lester, Erin M, Balss, Karin M
• Format: Journal Article
• Language: English
• Published: England 05.08.2024
• Subjects: biopharmaceuticals; test method; pharmaceutical microbiology; Rapid Methods; modelling
• ISSN: 1365-2672; 1365-2672
• DOI: 10.1093/jambio/lxae188

To investigate an in-line Raman method capable of detecting accidental microbial contamination in pharmaceutical vessels, such as bioreactors producing monoclonal antibodies via cell culture. The Raman method consists of a multivariate model built from Raman spectra collected in-line during reduced-scale bioreactor batches producing a monoclonal antibody, as well as a reduced-scale process with intentional spiking of representative compendial method microorganisms (n = 4). The orthogonal partial least squares regression discriminant analysis model (OPLS-DA) Area Under the Curve (AUC), specificity and sensitivity were 0.96, 0.99 and 0.95, respectively. Furthermore, the model successfully detected contamination in an accidentally contaminated manufacturing-scale batch. In all cases, the time to detection (TTD) for Raman was superior compared to offline, traditional microbiological culturing. The Raman OPLS-DA method met acceptance criteria for equivalent decision making to be considered a viable alternative to the compendial method for in-process bioburden testing. The in-line method is automated, non-destructive, and provides a continuous assessment of bioburden compared to an offline compendial method, which is manual, results in loss of product, and in practice is only collected once daily and requires 3-5 days for enumeration.