Abstract
The Milliflex Rapid System is used as a rapid microbiological method based on adenosine triphosphate (ATP) bioluminescence in the pharmaceutical industry to quantify the amount of micro-organisms present in water and in bioburden samples. The system can also be used for qualitative analyses, for example, to perform a rapid sterility test. This rapid sterility test has been successfully validated and implemented at Novartis and Sandoz. As the reagents used for the ATP bioluminescence reaction, which are directly sprayed on a micro-colony, disrupt the walls/membranes of the present cells to release ATP and therefore no intact cells for subsequent identification were believed to be present, the identification was supposed to be impossible until now.
During development and validation of a rapid sterility test with the Milliflex Rapid System, a possibility to identify contaminants was found. A method based on regrowth of the Milliflex Rapid-treated microbial cells and consecutive genotypic identification reproduced feasible and robust results. The data presented here show that sufficient recovery of the micro-colonies detected with the Milliflex Rapid System was reached with the test strains, except with Penicillium spec. The chosen micro-organisms represent the full spectrum of environmental isolates and ATCC strains, and it was shown that they are not destroyed after application of the reagents for the ATP bioluminescence reaction.
Overall, 22 stressed microbial strains were examined during the study.
LAY ABSTRACT: After Milliflex Rapid System detection, it was supposed that a subsequent identification of the contaminant is not possible. In this paper it is shown how contaminants can be identified in the rapid sterility test application.
1. Introduction
The Milliflex Rapid System is used as a rapid microbiological method based on adenosine triphosphate (ATP) bioluminescence in the pharmaceutical industry to quantify the amount of micro-organisms present in water and in bioburden samples. Pharmaceutical manufacturers are now realizing the potential and the advantages of introducing rapid methods into their routine release testing (1). During development and validation of a rapid sterility test with the abovementioned rapid microbiological method, it has also been necessary to find a method to identify contaminants in the case of a failure. The ability to identify the contaminating microbial strain in a sterility test failure is a necessity to investigate sterility test contamination of aseptically produced or terminally sterilized drug products, and is required by the guidelines (2–4). In the traditional sterility test, aliquots of the contaminated nutrient medium are transferred out of the respective sterility test canister or the direct inoculation liquid medium and processed further for subsequent genotypic identification using a system such as the Microseq System.
In the rapid sterility test a solid nutrient medium is used instead of the liquid media tryptic soy broth and fluid thioglycollate medium. Advantages of using a solid nutrient medium instead of a liquid nutrient medium were reported in 1973 (5). The second difference to the traditional sterility test is the fact that the incubation time of 14 days is lowered down to 5 days. The final readout of the three Milliflex Rapid membranes in the rapid sterility test is performed after 5 days of incubation: to do this the plates with the rapid sterility test membranes are taken out of the incubators (20–25 °C and 30–35 °C aerobic incubation) and from the anaerobic jar out of the 30–35 °C incubator (in the Milliflex Rapid sterility test only filterable samples can be tested). The incubation time for the rapid sterility test was determined during a prestudy for the validation of the alternative sterility test. This time is dependent on the worst-case micro-organism. The worst-case micro-organism was determined to be stressed Propionibacterium acnes, which can reproducibly be found with the Milliflex Rapid after 91 h of anaerobic incubation at 30–35 °C. These 91 h were rounded up to 4 full days and an additional day was added to account for variability and testing in the laboratory.
A visual check for microbial growth is performed on the plates. If no growth is observed then the membrane is removed to prepare the visually “sterile” membrane for Milliflex Rapid analysis. The membrane is dried for 5 min under laminar flow and then sprayed with ATP-releasing reagent and with bioluminescence reagent. The sprayed membrane is then transferred into the detection tower, in which a charge-coupled device (CCD) camera detects whether present light signals were emitted from the ATP bioluminescence reaction. The attached personal computer with the Milliflex Rapid software calculates the number of micro-colonies present and delivers a picture of the micro-colonies on the membrane.
The detection method used is different in the rapid sterility test: detection is not performed visually, it is performed through the light detection of the ATP bioluminescence reaction with a highly sensitive CCD camera. The CCD camera is able to detect micro-colonies that are not yet visible to the human eye.
This reaction chemistry (6) forms the basis of the Milliflex Rapid detection:
(ATP, adenosine triphosphate; AMP, adenosine monophosphate; hv, emitted photon; λ, wavelength)
The ATP-releasing reagent contains chemical components that lead to cell rupturing to release ATP from the microbial cells. This ATP release is the indicator for a microbial contamination. As all present cells were said to be ruptured after the reagent treatment, it was assumed that no intact cells could be used for further identification and investigation of the root cause of a rapid sterility test failure. Because not being able to investigate would be an inacceptable situation, a method to circumvent this problem had to be found.
Opposite to what was supposed to occur, a robust and reliable method was found using “regrowth”—a sprayed rapid sterility test membrane was brought back to the solid nutrient medium and a subsequent incubation period showed results of regrowing colonies that had been sprayed and detected with the Milliflex Rapid System. The first part of the study shows that more than 95% of the expected contaminants that could lead to a rapid sterility test failure do not need to be detected using the Milliflex Rapid System after the rapid sterility test incubation time of 5 days, as they can be read out visually due to fast growth. These stressed representatives of the broad spectrum of contaminants were monitored in hourly intervals and showed their exact time to visual detection, which is far below the rapid sterility test incubation time of 5 days. In order to consider the fact that contaminants of drug products are supposed to be in a more or less stressed state, a procedure to stress micro-organisms in a defined way was established and included these stressed micro-organisms in all parts of the study (7).
In a second part of the study six stressed representatives; one of each group of micro-organisms for the validation of the rapid sterility test was treated and detected using the Milliflex Rapid System. Following this approach, which used different and considerably lower incubation times than used for rapid sterility testing and also the 5 day incubation time, micro-colonies were quantified and subsequently treated for regrowth. Lower incubation times were used to obtain even smaller micro-colonies, which resemble a worst-case scenario. In the regrowth approach an inoculated rapid sterility test membrane is transferred back onto the solid nutrient medium, the rapid sterility test medium (RSTM), after it had been sprayed and detected using the Milliflex Rapid System. RSTM is the nutrient medium used during validation and routine operation of the rapid sterility test. This nutrient medium was evaluated in a growth promotion study using a broad spectrum of stressed and unstressed micro-organisms and was demonstrated to promote better growth than the liquid media used in the traditional sterility test (7).
In a third aspect it was shown that the orientation of the Milliflex Rapid image obtained during detection can be directly compared to the image obtained after regrowth of the colonies on the used membrane, which can be visually detected and photographed.
All findings of this study might be used for other applications of the Milliflex Rapid and presumably not only for the rapid sterility test application.
2. Material and Methods
2.1. Expendables
Milliflex Filter Funnel Unit, Millipore MXHVWP124 (polyvinylidene difluoride [PVDF] membrane)
Nutrient media in Milliflex cassettes, MXSM C0120, filled with RSTM by Heipha (Germany), gamma-irradiated (9–20 kGy) product code: 051
Anaer indicator strips, BioMerieux
Transfer units, for example, Millipore transfer units TZA000010 or Sartorius Sterisart NF 16470 GB
2.2. Further Material
Steritest Compact pump, Millipore
3 Milliflex PLUS pumps, Millipore
Fluid A—sterile, in-house
Milliflex Rapid System, Millipore
Water bath
Stop watch
Float device for water bath
Sterile forceps
Eppendorf vial
Eppendorf pipettes
Tritainers (anaerobic jars from Trilab System)
Trilab System (system that fills Tritainers with anaerobic gas)
2.3. Micro-Organisms Used
2.4. Methods
The required quantity of cryovials containing micro-organisms in the inoculum range of 10–100 colony-forming units (CFU) is defrosted from the −80 °C freezer. The microbial suspension is transferred from the cryovial into a marked Eppendorf vial.
The micro-organisms are stressed according to the stress protocol (i.e., Propionibacterium acnes, 60 °C, 1 min in a float device in a water bath; see Table I for stress parameter and micro-organism list [7]).
List of Micro-Organisms Including Their Defined Stress Parameters. “HK-WST” Indicates Environmental Isolate from the Novartis Pharma Stein AG Site (7)
Fifty milliliters sterile Fluid A is filled into the Milliflex Filter funnel, the volume of the stressed strain under test (10–100 CFU) is added, and another 50 mL of Fluid A is poured into the funnel. The inoculated, 100 mL Fluid A is filtrated through the membrane.
The membrane is clicked onto the RSTM cassette. The anaerobic plate is transferred into the Tritainer and then it is filled with anaerobic gas by placing it under the Trilab System. The plates are either incubated at 20–25 °C aerobically, 30–35 °C aerobically, or 30–35 °C anaerobically (dependent on the micro-organism).
2.4.1. Aspect 1—Time to Visual Detection:
To prove that most of the micro-organisms under test (see Table I for list of micro-organisms) already show visible growth before the end of the rapid sterility test incubation time of 5 days, photographic monitoring on hourly intervals is performed.
In this experiment, most stressed bacteria are tested at 30–35 °C aerobic incubation. Exceptions are Clostridium sporogenes and Propionibacterium acnes—these are incubated at 30–35 °C anaerobic incubation. Yeasts and molds are incubated aerobically at 20–25 °C. With hourly photographs the incubation time necessary until visual colony stage occurs is determined. Aspect 1 uses every strain in triplicate, and two test runs are performed per strain.
2.4.2. Aspect 2—Micro-Colony Detection and Regrowth Strategy:
In a second part of the study, six stressed representatives listed in Table II are treated, incubated, and detected using the Milliflex Rapid System. The six representatives are chosen based on the following: All 22 micro-organisms for the validation work for rapid sterility testing are grouped into five groups (yeasts/molds, Gram-positive sporulating bacteria, Gram-negative bacteria, Gram-positive cocci, Gram-positive rods; see Table I ). From each group one representative is chosen. As only aerobic and one microaerophilic representative are part of the study, an obligate anaerobic bacterium, Clostridium sporogenes, is additionally chosen. Adding Clostridium sporogenes allows the analysis of regrowth also under anaerobic conditions.
Incubation Parameters and Micro-Colony Incubation Time of Used Micro-Organisms
The second part of the study uses the chosen six strains in triplicate, and two test runs are performed per strain. In this experiment different and considerably lower incubation times than used for rapid sterility testing have to be found for the micro-colonies. The micro-colony and its incubation time for reproducible Milliflex Rapid detection and quantification have to be defined for every strain of the listed micro-organisms. The following prerequisites have to be fulfilled to define the strain-specific micro-colony stage and its incubation time:
The micro-colony stage (detected with Milliflex Rapid System) has to have an incubation time that is lower than the incubation time to reach visual colony level (detected with human eye). Guidance is given by the results gathered in aspect 1, where the time to visual detection is determined.
A micro-colony is visually not detectable; the detection is performed using the Milliflex Rapid System.
The Milliflex Rapid System has to detect reproducible counts at the described incubation time.
The produced counts of the Milliflex Rapid System have to show at least 70% recovery compared to visually counted, not sprayed (not Milliflex Rapid reagent-treated) membranes.
The determination of the strain-specific micro-colony stage is necessary for the regrowth experiments to simulate worst-case micro-colony stages (listed in Table II ). The determination takes place by detection at reasonable time points, in some cases at hourly intervals. In the regrowth approach (second aspect of the study) an inoculated rapid sterility test membrane (10–100 CFU) is transferred back onto the solid nutrient medium (RSTM) after it had been sprayed and detected using the Milliflex Rapid System.
For aspect 2, the strain-specific incubation times necessary for Milliflex Rapid detection are used. The tested micro-organisms, incubation parameters, and micro-colony incubation times for aspect 2 are detailed in Table II .
Regrowth procedure: The membranes that have been analyzed in the Milliflex Rapid System are instantly again attached on a new RSTM nutrient media cassette. This reattachment has to be done taking special attention to not applying air bubbles between membrane and agar. Milliflex Rapid detection and reattachment of the membrane take place in a laminar flow hood using aseptic working technique. The membrane on RSTM nutrient medium is incubated until colonies are visible (5 days for yeasts/molds, 3 days for bacteria with the exception of Propionibacterium acnes, which requires 5 days for regrowth) and the visual read-out is performed on the regrown colonies.
Afterwards the recovery of the CFUs from the read-out by eye and from the Milliflex Rapid System are compared and have to reach the acceptance criterion of 70% recovery (amount of regrown visual colonies with visual examination compared to the count determined by the Milliflex Rapid System). The regrown strains are subsequently genotypically identified using the Microseq to confirm the identity of the inoculated strain.
2.4.3. Aspect 3—Orientation of the Milliflex Rapid Image Compared to Regrown Rapid Sterility Test Membrane:
Aspect 3 of the experimental study is the analysis of the orientation of the micro-colony on the Milliflex Rapid image compared to the orientation of the visually regrown colony on the same membrane. The goal of the study is to show that a secondary contamination that is potentially added after the rapid sterility test incubation time (during detection) can be identified as secondary contamination. This can be achieved by comparison of the Milliflex Rapid image and the location of the regrown colony. For this aspect membranes of stressed Clostridium sporogenes are inoculated with 10–100 CFU, incubated at 30–35 °C anaerobically for 26 h (see Table II ), and detected using the Milliflex Rapid System. Only one strain is used for aspect 3, as this test is regarded as strain-independent. The orientation of the membrane is recorded using a mark on the “12 o'clock” insertion of the membrane. The mark on the membrane is always used as an indicator for inserting the membrane at exactly the same position in the detection tower. After the readout, the membranes are transferred back onto a new RSTM cassette. The membranes are again incubated until colonies are visible (3 days for Clostridium sporogenes) and the visual read-out with additional photograph is performed on the regrown colonies.
Afterwards the photograph of the regrown colonies and the Milliflex Rapid image are compared to analyze the orientation of the regrown colonies. For this the “12 o'clock mark” is used to determine the orientation on the Milliflex Rapid image and on the photograph.
3. Results
3.1. Aspect 1—Time to Visual Detection
This aspect shows how many micro-organisms of the chosen range are found at a visual colony level before the end of the rapid sterility test incubation time of 5 days. As the cause for contamination expected to be found in a sterility failure (traditional and rapid sterility test) can be divided into fast-growing micro-organisms and slow-growing micro-organisms, most contaminations are identified during readings before the end of the 5 day incubation time by visual observation of the sterility test membrane. The exact times until detection by eye or detection with Milliflex Rapid System are shown in Table III (in this study only the rapid sterility test was experimentally analyzed).
Incubation Times To Determine Visual Colony Level
Table III shows that only stressed Propionibacterium acnes grows later than the 5 days of incubation to visual colony level. Only this micro-organism needs to be detected with the Milliflex Rapid System; all other micro-organisms in the list are detected at visual colony level.
3.2. Aspect 2—Micro-Colony Detection and Regrowth Strategy
In the following tables the results for the time to micro-colony detection and the regrowth strategy are presented. Each table shows the recoveries gathered for the comparison of the counts of the images. The acceptance criterion for the recovery is 70%.
Stressed Propionibacterium acnes (incubation time for micro-colony stage: at least 91 h or maximum 5 days at 30–35 °C anaerobic incubation):
In test run 2 a high count was detected due to an operator mistake. Generally it is not easy to control the inoculum of stressed environmental isolates from cryo cultures precisely. An alternative reason is that the Milliflex Rapid System is said/known to not count accurately >70 CFU, thereby underestimating in comparison to visual counting.
The acceptance criterion of 70% recovery of regrown colonies compared to the gained count in Milliflex Rapid detection is fulfilled in all cases for stressed Propionibacterium acnes. Microseq result: All regrown colonies were identified as Propionibacterium acnes.
Stressed Penicillium spec. (incubation time for micro-colony stage: 35 h at 20–25 °C):
* Regrowth of Penicillium spec. after Milliflex Rapid detection was documented, but could not be counted. The detected amount of around 67 to 108 micro-colonies (detected and counted with the Milliflex Rapid) was most likely “sprayed” onto the outer rim of the Milliflex Rapid membrane and regrew only on the outer ring (see Figure 1). The acceptance criterion of 70% could not be fulfilled in the case of Penicillium spec., when the membrane was sprayed in the micro-colony stage. Microseq result: All regrown colonies were identified as Penicillium spec.
Left pictures show the Milliflex Rapid images; the pictures on the right show photographs of the regrown ring of Penicilliumspec.
Stressed Bacillus clausii (incubation time for micro-colony stage: 32 h at 30–35 °C aerobic incubation):
The acceptance criterion of 70% recovery of regrown colonies compared to the gained count in Milliflex Rapid detection is fulfilled in all cases for stressed Bacillus clausii. Microseq result: All regrown colonies were identified as Bacillus clausii.
Stressed Acinetobacter lwoffii (incubation time for micro-colony stage: 22 h at 30–35 °C aerobic incubation):
The acceptance criterion of 70% recovery of regrown colonies compared to the gained count in Milliflex Rapid detection is fulfilled in all cases for stressed Acinetobacter lwoffii. Microseq result: All regrown colonies were identified as Acinetobacter lwoffii.
Stressed Kocuria rhizophila (incubation time for micro-colony stage: 45 h at 30–35 °C aerobic incubation):
The acceptance criterion of 70% recovery of regrown colonies compared to the gained count in Milliflex Rapid detection is fulfilled in all cases for stressed Kocuria rhizophila. Microseq result: All regrown colonies were identified as Kocuria rhizophila.
Stressed Clostridium sporogenes (incubation time for micro-colony stage: 26 h at 30–35 °C anaerobic incubation):
In Figure 2 two different phenomena can be observed: In the red and green circles there is one Milliflex Rapid count each but two colonies in the red and green circle in 2b have regrown. This gives evidence that the Milliflex Rapid cannot distinguish between two colonies, if colonies lie close together. In the yellow boxes two Milliflex Rapid counts can be observed. These micro-colonies did not regrow in the photographic image 2b. This means that not every sprayed micro-colony has the ability to regrow. In the end it can be observed that 100% recovery occurs. One hundred percent recovery can be explained due to the summary of no regrowth of two colonies and twice the fact that merged colonies were found, which could not be counted as four single micro-colonies by the Milliflex Rapid System.
Exemplary comparison of membrane #1 (test run 1) of Clostridium sporogenes: 2a) shows the Milliflex Rapid image, 2b) shows the photograph of the regrown colonies. 2c) is an edited Milliflex Rapid image merged with the photograph.
In Figure 3 it can again be observed that not all micro-colonies have the capability to regrow. Overall in membrane #2 three micro-colonies did not regrow. Two micro-colonies (in the red circles) have not regrown due to an underlying air bubble. This air bubble was added after the membrane had been sprayed with the Milliflex Rapid reagents and was reattached to a new RSTM medium cassette. The observed finding stresses the importance of reattaching the membrane to the agar properly. In the yellow box another micro-colony detected in the Milliflex Rapid System did not regrow. Overall the acceptance criterion was met, as the recovery is in this case was 88%.
Exemplary comparison of membrane #2 (test run 1) of Clostridium sporogenes: 3a) shows the Milliflex Rapid image, 3b) shows the photograph of the regrown colonies, 3c) is an edited Milliflex Rapid image merged with the photograph.
In Figure 4 one can again observe two different phenomena: In the red circle there is one Milliflex Rapid count more than in the regrown colony photograph. This micro-colony did not regrow in the photographic image 4b. This means that not every sprayed micro-colony has the ability to regrow. In the yellow box, two micro-colonies were counted as one by the Milliflex Rapid System. In the blue boxes there are three regrown colonies, which were presumably counted as two micro-colonies by the Milliflex Rapid System. This again gives evidence that the Milliflex Rapid System cannot distinguish between two colonies if colonies lie close together. In the end if both observations are summarized, the count between Milliflex Rapid count and regrown image is nearly equal (105% recovery).
Exemplary comparison of membrane #3 (test run 1) of Clostridium sporogenes: 4a) shows the Milliflex Rapid image, 4b) shows the photograph of the regrown colonies, 4c) is an edited Milliflex Rapid image merged with the photograph.
The acceptance criterion of 70% recovery of regrown colonies compared to the gained count in Milliflex Rapid detection is fulfilled in all cases. Stressed Clostridium sporogenes was especially added to the study, because it should be shown that even under strictly anaerobic conditions regrowth after Milliflex Rapid detection takes place. It is therefore demonstrated that also anaerobic micro-organisms are capable of regrowing after Milliflex Rapid treatment. Microseq result: All regrown colonies were identified as Clostridium sporogenes.
3.3. Aspect 3—Orientation of the Milliflex Rapid Image Compared to Regrown Rapid Sterility Test Membrane
In Figure 5 every micro-colony and the corresponding regrown colony were given numbers. In the red and green circles the Milliflex Rapid System counted one micro-colony, but in the end two colonies regrew. In the yellow box two micro-colonies did not regrow in the photograph. Comparing the Milliflex Rapid-detected micro-colonies to the visual countable colonies shows that it would be possible to distinguish between a colony that was added as secondary contamination during Milliflex Rapid detection in the laminar flow and the primary contamination potentially found in the Milliflex Rapid sterility test. In the case of a nonsterile rapid sterility test result, it is expected that only a low level of micro-colonies will be detected and one additionally added colony can easily be differentiated from the original rapid sterility test failure. Using the mark on the “12 o'clock” orientation on the Milliflex Rapid membrane leads to the possibility that some further added secondary contaminations could be regarded as secondary contaminations. These secondary contaminations could theoretically be derived during the removal of the membrane from the RSTM cassette. This removal takes place after the 5 days of incubation time for rapid sterility testing inside a laminar flow (cleanliness zone C). After membrane removal, the membrane is handled by an operator using aseptic working technique, sprayed on the autospray station, and is subsequently manually inserted into the detection tower of the Milliflex Rapid System. During this step it is unlikely but can not be excluded in all cases that an operator-derived or environmentally derived micro-organisms is applied as a low-level, secondary contamination. In Figures 2⇑⇑–5 one can observe that the Milliflex Rapid System is not able to pick up merged micro-colonies as single micro-colonies. In contrast to that, counting by eye leads to higher counts because the human eye is indeed capable of differentiating between merged colonies.
Comparison of membrane 1 of stressed Clostridium sporogenes: 5a) shows the Milliflex Rapid image, 5b) shows the photograph of the regrown colonies, 5c) is an edited Milliflex Rapid image merged with the photograph.
4. Discussion
In the first part of the results for aspect 1 it was shown that most of the potential contaminants in rapid sterility test failures would be detected using visual examination of the rapid sterility test membrane after 5 days of incubation (in this study >95% with the chosen test strains). Only slow-growing, worst case micro-organisms would be detected using the Milliflex Rapid System. The largest part of the expected contaminating microbial strains would not be sprayed and could immediately be identified using the routine procedure for genotypic identification.
A certain percentage of slow-growing microbial strains that are detected in micro-colony state at the end of the incubation period of 5 days using the Milliflex Rapid System is expected. In the chosen in-house environmental isolates flora only one strain is considered a worst-case micro-organism, detectable in the micro-colony stage after 91 hours and still after 5 days of incubation. This strain is stressed Propionibacterium acnes. To simulate further micro-colonies from different strains, it was demonstrated that five additional strains representing the broad spectrum of micro-organisms could be partly regrown when sprayed in micro-colony state and lower incubation times than five days are used (the only exception was the regrowth of Penicillium spec.). The micro-colony state and corresponding incubation time is by definition the state before visual colony status that allows reproducible Milliflex Rapid counts. In this micro-colony stage a high enough incubation time is the critical parameter to reach sufficient cell mass to probably guarantee re-growth. The micro-colony state is unique to each micro-organism and ranges, for example, from 7.5 h for the fast-growing strain Escherichia coli, which was not part of this study, to 91 h for the slowest growing strain, stressed Propionibacterium acnes. A comparison of the time to visual colony detection and the time to micro-colony detection shows that the time micro-colony detection can not easily be predicted through the knowledge of the time to visual colony detection (see Table IV ). The time to micro-colony detection can be very close to the time for visual colony detection. On the other hand, it can be more than half the time to visual colony detection. These differences can presumably be explained by the strain-specific ATP content of the microbial cells, which is the principle of detection for the micro-colony using the Milliflex Rapid technology.
Comparison of Time to Visual Colony Detection and Time to Micro-Colony Detection
The results presented here for Bacillus clausii, Acinetobacter lwoffii, Penicillium spec., Kocuria rhizophila, Clostridium sporogenes, and Propionibacterium acnes (all used in stressed condition) show recovery rates over 100% when one compares the obtained Milliflex Rapid images to the regrown rapid sterility test membranes. The set acceptance criterion before starting the study was a recovery of 70% because it was expected that not all micro-colonies that had been treated using the Milliflex Rapid System would be able to regrow. The outcome of the study led to the result that presumably most sprayed micro-colonies are able to regrow using the approach presented here, if they were incubated long enough. After the rapid sterility test incubation time of 5 days, the present micro-colonies are expected to be in a stable and large micro-colony stage that allows regrowth of all known strains. The resulting recoveries lie higher than 100% in most cases. This indicates that the micro-colony stage/incubation time used is not ideal, as the Milliflex Rapid System does not pick up all micro-colonies present on the membrane using the short incubation time. It was shown that most of the micro-colonies detected by the Milliflex Rapid System are capable of regrowing to full colony size. It can be observed that the Milliflex Rapid System is not able to pick up merged micro-colonies as single micro-colonies. In contrast to that, counting by eye leads to higher counts because the human eye is indeed capable of differentiating between merged colonies. Regrowth of sprayed micro-colonies was shown to be possible in most cases, if proper adherence of the membrane to the nutrient medium is assured during the regrowth phase. A certain probability exists that Milliflex Rapid reagents lead to the fact that the micro-organism is not capable of regrowing on the same location of the membrane. The only strain that could not be identified on the same location is, to our knowledge, Penicillium spec. If this strain is the root-cause of the rapid sterility test failure, the strain would not be able to regrow on the same orientation/location on the membrane. As no investigation is possible in this case and possibly in other cases where lethal micro-colony destruction takes place, the aseptically produced or terminally sterilized batch of drug products would be rejected. In this case no complete investigation would be possible. It is possible that vegetative cells of Penicillium spec. are killed by the Milliflex Rapid reagent treatment, and only spores that were centrifuged to the edge of the Milliflex membrane by the AutoSpray Station movement survived, and only Penicillium spores have the capability to regrow.
The third aspect was experimentally tested to show that the image obtained using the Milliflex Rapid System can directly be compared to the regrown membrane. Every membrane tested that shows positive results in the detection of the rapid sterility test result is marked and put back onto RSTM to subsequently regrow. The used mark on the membrane corresponds to the upper part of the Milliflex Rapid image; they can be directly compared because membrane and Milliflex Rapid image are both circular. With the images presented here it could be shown that possibly derived additional contaminations, which could arise after Milliflex Rapid detection, could in most cases following the correct incubation time be judged as secondary contaminations that arose after the rapid sterility test incubation period. By using the possibility to regrow micro-organisms, the testing laboratory will be able to identify the contaminating micro-organism and will be able to conduct an investigation. The validation work presented here was conducted during the validation of the rapid sterility test using the Milliflex Rapid System. The rapid test was validated according to the guidelines on rapid method validation (9–11), and the regrowth study was performed additionally to this.
Conflict of Interest Declaration
The authors declare that no financial or non-financial competing interests are related to the manuscript presented.
Acknowledgements
Special thanks go to the Sandoz Microbiology laboratory in Kundl, Austria. Without your support we would not have been able to finish the study in time. We also thank colleagues from the Swiss Novartis Pharma Stein AG microbiology department involved in the project for their expertise and for critical reading of the manuscript. Last but not least: thank you to the RMM team!
- ©PDA, Inc. 2011
