Validation of rapid microbiological enumeration methods: contamination control strategies to reduce risk, optimize yields and improve quality.
Testing for microbial contamination in pharmaceutical production processes is necessary to monitor product quality. However, traditional microbiological methods are slow and require several days to obtain results.
Matrices like water and in process or finished products are usually tested for microbial contamination using membrane filtration and incubation. The membranes are incubated at least three days on Tryptic Soy Agar (TSA), or on R2A (1) agar for at least five days, as specified by the European Pharmacopoeia (Ph. Eur.) (2); (3), allowing microorganisms to grow and yield visible colonies or turbidity.
The rapid microbial detection system tested here (Milliflex[R] Quantum, EMD Millipore, Billerica, MA, USA) is a non-destructive alternative to traditional methods. Direct fluorescent staining makes it possible to detect cells before they become visible to the human eye using digital imaging and earlier results minimize warehousing space, reduce personnel time and allow direct intervention in manufacturing.
When using a Rapid Microbiological Method (RMM) to monitor for contamination, it is necessary to confirm that the results are reliable and repeatable. This validation process was performed following the guidance documents Ph. Eur. 5.1.6, USP <1223> and PDA Technical Report 33 (4); (5); (6). Many papers have been published on choosing and validating an RMM (7); (8); (9); (5), but validation case studies are rare. This article provides an overview of our validation of a number of rapid microbial monitoring techniques, discusses regulatory expectations and explores challenges experienced during the process.
PRINCIPLE AND WORKFLOW
The system evaluated is a growth based technology that requires membrane filtration. After a shorter-than-traditional incubation period, media cassettes are removed and each membrane is transferred into a pad soaked with 2 mL of staining solution and incubated for 30 minutes at 32.5 [degrees]C [+ or -] 2.5 [degrees]C.
Viable microorganisms are labeled using Fluorassure proprietary based reagents. These are based on non-fluorescent substrates that release free fluorochrome into the cytoplasm when enzymatically cleaved. Given that only viable cells (i.e. endospores, vegetative forms etc.) accumulate fluorescein and perform this cleavage, only those cells are stained.
Fluorescent micro-colonies are counted using the micro-colonies are counted using the Milliflex[R] Quantum Reader (EMD Millipore) or a photo displayed on a PC screen. The reader LEDs emit light in a determined wavelength and excites fluorescine metabolized by microorganisms which then emits fluorescent light. With the Quantum Reader it is possible to mark the colonies and decrease the chance of miscounting. Since the method is non-destructive, membranes can be reapplied to growth media after the enumeration. This allows microorganisms to be identified in case of sample contamination.
STUDIES PRIOR VALIDATION
System qualification parameters for the validation were defined; namely, optimal incubation time, the fluorescence reagent quantity and the possibility of microorganism's identification after the fluorescence staining and method specificity.
To evaluate incubation time, tests were run with ATCC[R] strains as well as four other isolates. Since it is impossible to know which strains a non-sterile sample includes, the optimum incubation time was based on every value obtained from the triplicates, in which all tested strains (including environmental stressed) had recovery rates (Milliflex[R] Quantum counts *100/traditional counts) of at least 70%.
Results showed significant reduction in the optimum incubation time, which consists of 24 hours for TSA and 48 hours for R2A agar samples. The prestudies showed that in membranes with less than 2.0 mL fluorescence reagent, not all microorganisms were stained, especially those on the border. Therefore the established optimal fluorescence reagent amount is at least 2.0 mL.
The stained membranes were reincubated for colony identification by Gram-Staining, API[R] (BioMerieux, Marcy I'Etoile, France) and PCR. Membrane colonies were of the same species as the control colony without staining; therefore adding fluorescence reagent does not affect genetic or enzymatic identification of the microorganisms.
To verify that the method does not produce false-positives, a test with sterile latex beads (1.1 pm, Sigma-Aldrich, MO, USA) was performed. It showed that that the Quantum system detected the target panel of microorganisms without staining this kind of particles.
After all prestudy aspects were considered, the parameters required were: accuracy, precision, range, linearity, limit of quantification, specificity and robustness.
Several of these do not translate as well from chemistry to microbiology. In the linearity test, some concentrations needed to be set within a specific range to show correlation between methods. Because of microbiology's high variability, it is not possible ensure an exact target concentration.
To illustrate this we leveraged linearity and limit of quantification (LOQ) of the Milliflex[R] Quantum method. This new method has a normal distribution (Anderson Darling) and a recovery rate of at least 70%. Results were proportional to the target concentration (0.8 [greater than or equal to] a [greater than or equal to] 1.2) and had a good correlation coefficient ([R.sup.2] [greater than or equal to] 0.9) between the two sets of data demonstrated with all tested strains, except for Aspergillus brasiliensis (Figure 1). Similarly, the Quantum Method did not always meet the criteria of the LOQ test, especially in the trials analyzing Aspergillus brasiliensis. This implies that the lowest number of microorganisms accurately counted using the Milliflex[R] Quantum method was sometimes higher than that of the traditional method.
[FIGURE 1 OMITTED]
These observations do not mean that the tested method cannot detect fewer CFUs than the traditional one, or that its results are disproportionate to the target concentration. There was no statistical equivalence (T-test: p [greater than or equal to] 0, 05) in these cases because the Milliflex[R] Quantum system delivers a higher microbial count. This happens due to the system's ability to detect micro-colonies before they merge into a large unit, the detection of some stressed microorganisms that are not always seen using the traditional method, and the filamentous morphology of some microorganisms (Figure 2) such as Aspergillus brasiliensis (ATCC 16404), which may hide other CFUs due to its dispersed mycelia.
[FIGURE 2 OMITTED]
The average number of CFUs was significantly higher with the Milliflex[R] Quantum versus traditional system when membranes were prepared under the same conditions (Figure 2), demonstrating that the Milliflex[R] Quantum method is more accurate than the traditional one. However, the two methods are not statistically equivalent. Although a more sensitive method is normally accepted over the compendial one, it is essential to review existing acceptance criteria.
This holds true for RMM technologies that do not need microbial growth to detect cells, which can potentially lead to the detection of viable but non-culturable (VBNC) microorganisms, increasing the CFUs counted even more. In the past, these damaged microorganisms have not been detected by traditional methods, but are counted by many of the RMMs making it crucial to evaluate acceptance criteria and action levels.
Precision parameter makes a case for attention to current specifications and acceptance limits for the validation of alternative microbiological methods. The study showed that the Milliflex[R] Quantum method is precise (coefficient of variation (CV) [greater than or equal to] 30% and a confidential interval with p [greater than or equal to] 0, 05) and not larger than the traditional method except for those strains in lower concentration (10 CFU or less). It occurs because biological samples are not homogeneous suspensions which can be prepared with a known cell number per unit. At this low inoculum level microbiological variability is high and is affected by varying factors such as sample distribution error, cell morphology or metabolic activity.
This variability must be taken into account in order to set up the CV. Chapter 5.1.6, section 3-3-2 of the European Pharmacopoeia (4) establishes the acceptable CV between 10 and 15%, although the validation example at the end of this chapter is based on a CV between 15 and 30%. In these cases the specifications are so strict that even the traditional method does not comply.
When establishing a CV under real conditions, the number of CFUs per plate should be considered, as they are in the USP (5). Higher counts should present less variation, while lower counts may bring more. Therefore, further discussions and a future guideline revision of requirements and acceptance limits should be considered.
VALIDATION OF INTENDED USE
Validation was also performed under routine conditions in batches of filterable matrices with real bulk and water samples. Particular bulks used were chosen to cover various types of active ingredients (vitamins, corticosteroids, antibiotics, proteins), different dosage forms (injection solutions, nose and eye drops), and those that can restrict growth because of the presence of conservation. Several rinses were performed according to internal regulations.
A product may interfere with the fluorescence signal due to a quenching effect (10), which may decrease the fluorescence intensity under a specific wavelength. Since this could cause false negative results in the release of a product using the Milliflex[R] Quantum system, this possibility was investigated as well.
The Milliflex[R] Quantum system had a recovery of at least 70% in comparison to the traditional method. This implies that the products do not induce a quenching effect and therefore are suitable to be tested with the Milliflex[R] Quantum system.
Given advancing technologies, there is a need for industry to embrace RMMs, which benefit both business and public health. Sharing and dialoguing with suppliers, manufactures and regulatory bodies reduces workload and redundant tests, thereby facilitating successful uptake of RMM test systems. There is an opportunity to encourage flexible regulation by presenting the knowledge gained through the application of scientific validation.
Because RMM benefits manufacturing, many initiatives and guidelines to facilitate its adoption have been developed. Efforts of the USA (5) and Eur. Ph. (4) commissions as well as PDA (6) and other regulatory bodies are appreciated in the implementation of alternative microbiological technologies. Much has been done, but further discussions and revision of acceptance criteria should be considered.
(1.) Reasoner, D. J., Geldreich, E. E.: A new medium for the enumeration and subculture of bacteria from potable water. Appl. Environm. Microbio. 49 (1985) 1-7.
(2.) European Pharmacopoeia 7 (2011) Chapter 2.6.12--Microbiological examination of non-sterile products: microbial enumeration tests.
(3.) European Pharmacopoeia 7 (2011) Chapter 2.6.13--Microbiological examination of non-sterile products: Test for specified microorganisms.
(4.) European Pharmacopoeia 7 (2011) Chapter 5.1.6--Alternative Methods for Control of Microbiological Quality.
(5.) United States Pharmacopeia. (2006) Chapter <1223> Validation of Alternative Microbiological Methods.
(6.) Parenteral Drug Association, Technical Report No. 33, 2000: Evaluation, Validation and Implementation of New Microbiological Testing Methods, PDA Journal.
(7.) Green, S. Industry strategy case study E: How to select, validate, and implement a rapid microbiology method and get it approved--A true story. American Pharmaceutical Review 10 (5), p. 102-107.
(8.) Miller, M. J. RMMs and the regulatory environment. Rapid Micro Methods (online) http://rapidmicromethods.com/files/regu-latory.html (accessed on April, 2011).
(9.) Sutton, S. Counting Colonies. Pharmaceutical Microbiology Forum 2006, 12 (9).
(10.) Lottspeich, F; Engels, W. J (eds). Bioanalytik. Spektrum Akademischer Verlag, Heidelberg, 2005; p 164 -168.
By Natalia Picioli Gealh, State University of Maringa, Maringa PR, Brazil and Dr. Michael Rieth, Merck Serono, Darmstadt, Germany
About the authors:
Natalia Picioli Gealh studied Biology (Major Microbiology) in the State University of Maringa, Brazil. After that, she worked as an intern at Merck KGaA in Germany in the QC-Department. She was responsible for developing and executing a method validation for a rapid microbiological enumeration method. She is currently enrolled in a master of bio and environmental process engineering at the Rhein Main University of Applied Sciences, Germany. +49 0176 37603760; firstname.lastname@example.org
Dr. Michael Rieth studied Biology at the Gottingen University (Major Microbiology), Germany. After earning his PhD, he worked for several life science companies. Since 1999 he has been at Merck KGaA, in Germany as the Head of Microbiological Quality Control. ++49 (0) 6151-72-4448 email@example.com
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||CELL CULTURE|
|Author:||Gealh, Natalia Picioli; Rieth, Michael|
|Date:||Oct 1, 2012|
|Previous Article:||A leaner, greener cleanroom: implementing efficient, contemporary cleanroom solutions can help shorten timelines.|
|Next Article:||Perfecting parenteral product handling: an interview with Ryan Hawkins, Vice President & Chief operating officer, cook pharmica, Bloomington, IN.|