Abstract
Currently, sterility testing in the pharmaceutical industry—a mandatory release test for all sterile drug products—takes an incubation time of at least 14 days and is based on liquid media according to the pharmacopoeias. The search is on for a rapid sterility test to reduce this rather long time frame. For this we have chosen the Millipore Milliflex Rapid Microbiology Detection System, which is based on solid nutrient media. As a prerequisite for the validation of this rapid sterility test, a solid nutrient medium promoting the growth of stressed and unstressed micro-organisms replacing tryptic soy broth and fluid thioglycollate medium from the traditional sterility test had to be found.
For this a wide variety of appropriate nutrient media were evaluated. After a prestudy with 10 different nutrient agar media, tryptic soy agar, Center for Disease Control (CDC) anaerobic blood agar, Schaedler blood agar, and Difco brewer anaerobic agar were tested in detail using a range of 22 micro-organisms (7 ATCC strains and 15 production site-specific strains). These strains were inoculated in their unstressed and in a stressed state. Stress was evoked by heat treatment and nutrient starvation in the case of the sporulating bacteria. This stress effect—resulting in deceleration in growth—was experimentally confirmed based on growth curve analysis. It was statistically evaluated which media and which incubation temperatures are best suitable.
The resulting data showed that Schaedler blood agar has the best growth-promoting properties among the agars tested and is going to be used in the rapid sterility test with the incubation temperatures 20–25 °C for aerobes, 30–35 °C for aerobes, and also 30–35 °C for anaerobic micro-organisms.
Introduction
The currently used sterility test performed according to United States Pharmacopeia (USP) 〈71〉 “Sterility tests” (1), European Pharmacopoeia (EP) 2.1.6 “Sterility” (2) and other relevant pharmacopoeias is a mandatory microbiological test to release sterile drug products. Today this traditional test uses two liquid nutrient media, tryptic soy broth (TSB) and fluid thioglycollate medium (FTM) with an incubation time of at least 14 days at 20–25 °C and at 30–35 °C, respectively. This long time frame is necessary to detect potentially present micro-organisms that are thought to be in a more or less stressed state due to the manufacturing process and/or the properties of the drug products.
As a consequence to this long time frame, detection of potential product contaminations and the respective corrective actions are delayed, resulting in an enlarged amount of affected product batches. Besides preventing these disadvantages of the traditional sterility test, the main benefit of an alternative rapid sterility test is the reduction of throughput time for sterile drug product release with a reduction of stock keeping costs, the possibility for a more flexible planning, and a lower risk for running out of stock.
The objective of the project, of which this study is the first step, was to establish a rapid sterility test due to the mentioned reasons. Therefore, commercially available rapid microbiological methods were evaluated for their ability to serve as basis for a rapid sterility test. Close proximity to the traditional test led to the decision to use a growth-based method with adenosinetriphosphate (ATP) bioluminescence as a principle of detection of contaminants. Available ATP bioluminescence systems are offered by different vendors. Systems with growth of micro-organisms in liquid nutrient media and detection of ATP in solution via luminometer make it difficult to differentiate between microbial signal and background. Therefore, a system for rapid sterility testing was chosen, where microcolonies are detected on a membrane filter incubated on a solid nutrient medium, so that no problems arise with signal-to-noise differentiation as is the case with the “liquid” systems. The system is additionally able to get information about the extent of contamination. Consequently, for the use of the chosen ATP bioluminescence system in a rapid sterility test a solid nutrient medium became necessary, one that should also enable the growth of stressed micro-organisms.
In this paper the evaluation of different solid nutrient media using a range of 22 micro-organisms consisting of 7 ATCC strains and 15 production site-specific isolates from environmental monitoring samples, bioburden, and sterility test failures is presented. In order to consider the fact that contaminants of drug products are supposed to be in a more or less stressed state, we established a procedure to stress micro-organisms in a defined way and included these stressed micro-organisms in our nutrient media evaluation.
Materials and Methods
Creation of Cryocollection of ATCC Strains and Production Site-specific Strains
Coming from either lyophilized culture (ATCC strains) or directly from plate (production site-specific strains) the micro-organisms (for a complete list see Table I) were cultivated in either liquid TSB or on solid media (for example, on Sabouraud dextrose agar) over an appropriate time period (genotypic identification of strains was performed using the MicroSeq System, Applied Biosystems, Foster City, California, USA). The culture was then centrifuged at 800 × g for 20–30 min, and the pellet was resuspended in protective medium (Oxoid-CM67 containing 15% glycerine). The culture was diluted to 10–100 colony-forming units (CFU) per 100 μL (the number of CFU was checked on solid media) and filled into 2 mL Nunc CryoTube™ Vials; cryocultures were stored at −80 °C.
List of Micro-organisms; in Red All Micro-organisms Mentioned in the Sterility Test Chapters of the Pharmacopoeia are Listed, the Additional Micro-organisms in Black were Chosen to Broaden the Range, HK-WST means Environmental Strain Isolated at the Novartis Pharma AG Stein Sterile Manufacturing Site
Determination of Strain-specific Stress/Kill Conditions
In a first step, stress was evoked by the application of UV light (240–250 μW/cm2), heat (50–70 °C in a water bath), or by incubating the micro-organisms in a dilution series of a parenteral drug product with microbicidal properties (Voltaren, Novartis, Stein, Switzerland) for 1–10 min each, taking aliquots every minute. Stress was directly applied to a fluid suspension of the tested micro-organisms in a range around 100 CFU.
For monitoring of the survival rate, the aliquots were plated on tryptic soy agar followed by incubation. Results were counted out after 2–7 days (depending on strain, for example Escherichia coli and Aspergillus brasiliensis have to be counted out after 2 days the latest, and Propionibacterium acnes can't be counted out earlier than after 7 days of incubation). The time necessary to lead to a decrease of the initially inoculated amount of CFU of more than 50% was determined. The Gram-positive sporulating bacteria were “stressed” by achieving a higher spore content. This sporulation was triggered by nutrient depletion and storing an overgrown microbial culture at 2–8 °C for more than 6 days.
Determination of Stress Effect on Growth Rate by Optical Density (OD) Measurement (λ = 600 nm)
Overnight cultures of the tested micro-organisms were stressed by applying the stress parameter settings that were determined in the ≥50% reduction study as mentioned above. TSB or FTM was inoculated with approximately 3–4 × 106 CFU per milliliter and analyzed over a time frame of up to 8 h by taking aliquots to measure optical density each hour. Incubation of the aerobic strains took place on a shaking table. The stressed cultures were compared to unstressed cultures regarding growth rate.
Growth Promotion Study of Selected Nutrient Media
In the growth promotion study, 7 ATCC strains (including Aspergillus brasiliensis ATCC 16404 (formerly Aspergillus niger ATCC 16404), Bacillus subtilis ATCC 6633, Candida albicans ATCC 10231, Clostridium sporogenes ATCC 11437, Pseudomonas aeruginosa ATCC 9027, Staphylococcus aureus ATCC 6538—as required from the pharmacopoeias—plus Escherichia coli ATCC 8739) and furthermore 15 micro-organisms that were isolated from the Novartis Pharma Stein AG sterile manufacturing site in Stein, Switzerland were used (see Table I).
The chosen range of micro-organisms consists of yeasts and molds, Gram-positive sporulating bacteria, Gram-negative bacteria, Gram-positive cocci and Gram-positive rods (both aerobic and anaerobic micro-organisms). Especially the production site-specific isolates represent known contaminants in manufacturing areas (environmental monitoring, bioburden samples) and were also isolated from sterility test failures in the past.
For the nutrient media evaluation study, Difco brewer anaerobic agar, Schaedler blood agar, Caso-Agar (tryptic soy agar) and Center for Disease Control (CDC) anaerobic blood agar were tested with 22 strains in a stressed and unstressed state.
These nutrient media were chosen following a preselection study. The following solid nutrient media were used for this preselection study using a growth promotion test with 10 strains in an unstressed state:
-
FTM-A (fluid thioglycollate medium containing additional 10g/L agar, leading to an end concentration of 1.075% agar), Amimed, (Allschwil, Switzerland)
-
Brain–heart infusion agar, γ-irradiated, Heipha (Eppelheim, Germany)
-
Difco brewer anaerobic agar, γ-irradiated, Heipha
-
Reasoner's 2A (R2A) agar, Oxoid, (Basingstoke, United Kingdom)
-
Schaedler blood agar, γ-irradiated, Heipha
-
Caso-Agar (tryptic soy agar), γ-irradiated, Heipha
-
Columbia agar 5% blood, Biomerieux, (Marcy l'Etoile, France)
-
CDC Anaerobic Blood Agar, γ-irradiated, Heipha
All γ-irradiated media were supplemented accordingly to sustain the irradiation process (9–20 kGy).
Nutrient media to be tested were inoculated with the micro-organisms (stressed and unstressed state) with an approximate amount of 10–100 CFU. The experiment was conducted using five replicates for each incubation parameter (incubation parameters are 20–25 °C and 30–35 °C for aerobic incubation, 30–35 °C for anaerobic incubation). CFU were visually counted after 2–7 days of incubation.
Resulting raw data were assigned to 6 groups for each micro-organism:
-
20–25 °C aerobic incubation—stressed micro-organism
-
20–25 °C aerobic incubation—unstressed micro-organism
-
30–35 °C aerobic incubation—stressed micro-organism
-
30–35 °C aerobic incubation—unstressed micro-organism
-
30–35 °C anaerobic incubation—stressed micro-organism
-
30–35 °C anaerobic incubation—unstressed micro-organism
The tested nutrient media were grouped together in subgroups.
The traditional sterility test was taken as a control in this experiment. In the traditional sterility test, one can only document the day/hour that growth first became visible for the different strains (22 stressed strains, 22 unstressed strains). As one cannot compare quantitative data to qualitative data (CFU compared to presence/absence result together with timeline of first visual growth observation in liquid media), this aspect was not mentioned in this paper also because another important part of the validation of a rapid sterility test is the limit of detection and the specificity study, where a direct comparison with low-level inoculum (using the 22 strains, stressed and unstressed) is made between the traditional sterility test and the rapid sterility test.
Statistical Analysis of Data
For the statistical analysis of the raw data, the following methods implemented in Minitab® Release 14.20 Statistical Software were used.
To compare means of two groups, a t-test was used. To compare more than two groups, analysis of variance (ANOVA) was used, which compares means and results in a P-value (3). As a cut-off a P-value of 0.05 was used—meaning the null hypothesis is rejected, when the P-value is below 0.05 corresponding to a confidence limit of >95%.
Every agar that showed no significant difference to the highest mean value (including the one having the highest mean) was rated with one credit. An agar that showed a significant difference to the highest mean value got no credit. The credits of the groups were summed up and the sums of the groups were compared. The three incubation parameters, 20–25 °C and 30–35 °C for aerobic incubation and 30–35 °C for anaerobic incubation, and the two different stress states for each micro-organism (stressed and unstressed) led to the formation of six groups for each micro-organism (multiplied by 22 strains). An ANOVA of each of these groups was performed; credits for good growth-promoting properties were summarized for the 22 micro-organisms for each agar. One credit was given to the group/agar if it achieved the highest count. Groups/agars without significant difference to this highest count agar also gained one credit.
Outlier analysis was performed using Dixon's test (4–6).
General description of the Milliflex Rapid Microbiology Detection System (7)
Samples under test (diluted in 100 mL of rinsing fluid to assure adequate distribution of micro-organisms on membrane) are filtered over Milliflex Rapid filter using Milliflex PLUS pump; a possible contamination is trapped on the filter membrane (pore size: 0.45 μm).
Following filtration, the filter is applied to a solid nutrient medium using the Milliflex system, in which the sterile Milliflex filter clicks on the Milliflex cassette and the funnel breaks off. Using this Milliflex system to transfer the membrane onto the nutrient medium, there is no risk of secondary contamination through membrane handling. Because the cassettes are closed tightly, the risk for secondary contamination during incubation is low. By incubating the filters on the Milliflex cassettes, growth of potential contaminant(s) to microcolonies can take place.
At the end of the incubation time, filters are separated from the Milliflex cassettes and applied onto the AutoSpray Station inside the laminar flow hood. All the steps after incubation are not critical steps regarding secondary contamination because microcolonies have to have a certain size (rule of thumb: 10–100 yeast cells, 1000 bacterial cells) which is necessary for detection in the Milliflex Rapid system. If single environmental or operator-derived microbial cells were added to the membrane during the steps after incubation, they would not be detected due to the lack of incubation and therefore the lack of a certain amount of ATP necessary for detection.
The AutoSpray Station sprays the ATP-releasing reagent and afterwards the bioluminescence reagent onto the membrane. The ATP-releasing reagent leads to a lysis of the cells, and the bioluminescence reagent adds Luciferin and Luciferase.
The reaction chemistry shown here is the basis of Milliflex Rapid detection (8):
where
-
ATP = Adenosine Triphosphate
-
AMP = Adenosine Monophosphate
-
hv = emitted photon
-
λ = wavelength
Fast transfer of the sprayed membrane from the AutoSpray Station to the Milliflex Rapid Detection Tower is necessary in order to catch the maximum of emitted photons (validated transfer time between 0 s and 90 s). The possible light signals are detected with a charge-coupled device camera, and the Milliflex Rapid software calculates the amount of microcolonies.
Evaluation of Background Noise of Used Media in the Milliflex Rapid System
To evaluate the background noise of the used nutrient medium, 100 mL of rinsing fluid (Fluid A and Fluid D tested) are filtered over the Milliflex Rapid filter. The filter is applied to the nutrient medium under test. Incubation at 30–35 °C resembles worst-case conditions (at higher temperatures more medium diffuses through membrane, which could disturb the reaction). The time for incubation used here to determine background generated by nutrient media was an incubation for 5 days. A 5-day time period resembles the incubation time in the rapid sterility test. Membranes are handled and read out as described above.
Make and Model Numbers
| Description | Product Code/Manufacturer |
|---|---|
| Milliflex Rapid Detection Tower | MXRPDT110 Millipore |
| Milliflex Rapid Computer | MXRPCOMPT Millipore |
| Milliflex Rapid software, version 2.0.0 | MXRPSFT01 Millipore |
| Milliflex PLUS Pumps with Pump heads | MXPPLUS01 Millipore |
| AutoSpray Station Kit | MXRPSPRKT Millipore |
| Milliflex Rapid Image Analyser | MXRPANA01 Millipore |
| IQV40 Image Processor PC board | MXRPIVQ40 Millipore |
| Nebulizer (Position 1 and 2) | MXRPSPRKT Millipore |
Reagents and Buffers
-
Cleaning and decontamination kit of Autospray Station (Millipore)
-
Standard Reagent Kit (Millipore), which contains standard bioluminescence reagent, reconstitution reagent, and ATP-releasing reagent
-
Fluid A, diluting and rinsing fluid (in-house preparation)
-
Fluid D, diluting and rinsing fluid (in-house preparation)
Material
-
Milliflex Filtration Funnel MXHVWP124 (Millipore)
-
empty Milliflex media cassettes for solid media, MXSMC0120 (Millipore); filled with nutrient media and gamma-irradiated (9–20 kGy), Heipha
-
Millipore absorbent pads for AutoSpray Station, AP1004700
Results
Determination of Strain-specific Stress Conditions
After evaluating different methods to stress cells—which could be, for example, hypotonic or hypertonic stress—irradiation (such as UV light, gamma irradiation, microwaves), ultrasound, heat, low temperature, or chemical stress (evoked, for example, by chlorine or pharmaceutical drug products), three stress factors were chosen for a first part of the study: application of heat (50–70 °C), dilution in a parenteral drug product (Voltaren, Novartis) and UV light. For each of the 22 strains, the time necessary to reduce the initial inoculum by 50% was determined. The results are shown in Table II. Figures 1–3 give examples of the gained data for Moraxella osloensis, E. coli, and Acinetobacter lwoffii, where a 50% reduction time of 4, 3, and 2 min can be deduced.
List of Micro-organisms Including Their Defined Stress Parameters
M. osloensis data for UV treatment over a time frame of 10 min, ≥50% reduction after 4 min (3 runs).
E. coli data for 60 °C heat treatment over time frame of 10 min, ≥50% reduction after 3 min (3 runs).
A. lwoffii data for parenteral drug product treatment over time frame of 10 min, ≥50% reduction after 2 min (3 runs).
Observations of changed colony morphology (colonies grew slower, but regained normal colony size later on) were made in some cases of the tested micro-organisms. For example, Kocuria rhizophila (formerly Micrococcus luteus) showed a decrease in colony size after heat treatment (Figure 4a and 4b).
a. K. rhizophila heat treated (70 °C, 3 min) compared to Figure 4b: untreated K. rhizophila plated on tryptic soy agar, 3 days of incubation at 30–35 °C.
Determination of Stress Effect on Growth Rate by Optical Density (OD) measurement
As it is important to cause stress in the surviving 50% of the inoculated microbial cells, it had to be confirmed in the second part of the stress study that not only a “killing” of 50% of the cells occurred and the surviving cells stayed in a normal viable state; it had to be shown that the stressed micro-organisms have a reduced growth rate. A monitoring by OD measurement (optical density at λ = 600 nm) over a time frame of approximately 8 h showed that a heat treatment led to a deceleration in growth of the remaining micro-organisms. The other tested stress factors, UV light and incubation in the chosen parenteral drug product, only killed micro-organisms (with the used parameter settings)—the surviving micro-organisms showed growth characteristics as if they were not stressed at all. Therefore, for the nutrient media evaluation and the validation of the rapid sterility test, heat was chosen as the representative stress factor for the used strains.
Gained data for the differently treated inocula (untreated, heat treated, treated with UV light, diluted in a parenteral drug product, time/parameter used from 50% reduction experiments) showed a slightly different growth curve in the normal plot of the OD-measured data (Figure 5).
Optical density data plotted for the untreated E. coli culture, the heat-treated E. coli, UV-treated E. coli, and Voltaren-treated E. coli.
Plotting the data logarithmically shows the influence of stressing more precisely and makes the data independent of the amount of inoculum. Figure 6shows the effect of heat treatment on the growth curves of E. coli. There was no difference in growth observed to the untreated reference when the bacterial cultures were stressed by either UV light or by parenteral drug product treatment (with the parameter settings determined in the 50% reduction experiments).
Optical density data plotted for the untreated E. coli culture, the heat-treated E. coli, UV-treated E. coli, and Voltaren-treated E. coli.
The comparison of the slopes showed that only a heat treatment (parameters determined in the 50% reduction experiments) has an influence on the growth rate of micro-organisms. As UV light and treatment with a parenteral drug product (both “kill factors” and not “stress factors”) had no substantial influence on the growth rate, these factors were not used during nutrient media evaluation and further validation experiments.
In the case of E. coli, the growth slope is reduced for approximately 3 h, meaning this micro-organism needs around 3 h to resuscitate (Figure 7).
Comparison of untreated E. coli growth curves and heat-treated E. coli; over 3 h the stressed E. coli culture has a slope of 0.2313, then it rises back to the normal slope of the untreated culture.
S. aureus shows a reduced growth slope over the time frame of 4 h (Figure 8).
Comparison of untreated S. aureus growth curves and heat-treated S. aureus; over 4 h the stressed S. aureus culture has a slope of 0.1071, then it rises back to the normal slope of the untreated culture (after 4 h the slope is 0.4034).
The yeast C. albicans showed an even longer enduring stress. The growth slope stayed reduced for more than 8 h, and overnight the stressed culture regained the normal growth slope (Figure 9).
Comparison of untreated C. albicans growth curves and heat-treated C. albicans; over more than 8 h the stressed C. albicans culture has a slope of 0.0419.
Because heat stress is not feasible for Gram-positive sporulating bacteria, a higher spore content in the bacterial suspension was evoked as an indicator of stress, which was verified by the growth rate determination experiments (Figure 10).
Comparison of untreated B. pumilus growth curves and starved B. pumilus (nutrient depletion for 7 days at 2-8 °C), over 5 h the stressed B. pumilus culture has a slope of −0.0056, then it rises back to the normal slope of the untreated culture.
All micro-organisms were tested in the described manner. Difficulties were observed in three cases. The molds Penicillium and Aspergillus grow as mycelium in liquid medium; therefore no exact OD measurement was possible. As data are available for C. albicans and for all tested bacteria we assumed that Penicillium and Aspergillus behave in the same manner. Therefore the tested parameters are taken for the nutrient media evaluation. In the other case, for P. acnes (this strain grows better in FTM), the parameters for a ≥50% reduction that were determined by plate count could not be reproduced in the OD measurement experiment. A time lag in growth was only observed when stress was omitted for a time frame longer than the parameters from the ≥50% reduction experiment. P. acnes possibly showed different results in the OD measurement due to higher oxygen stress. For OD measurement, aliquots were taken every hour, leading to a certain oxygen stress for P. acnes.
The stress parameters heat and nutrient depletion are therefore used in the nutrient media evaluation. The chosen stress factors resemble the stress that possibly present micro-organisms are subject to in the production process of sterile drug products.
Nutrient Media Evaluation
The nutrient media evaluation study was done in two parts: The first preselection (growth promotion test with 10 strains, unstressed, tests on eight different agars) led to a reduction down to four media, which meant that only these media came into closer consideration (data not shown). FTM-A agar, brain–heart infusion agar, R2A agar, and Columbia agar 5% blood were excluded due to insufficient growth-promoting properties during this preselection.
In the nutrient media evaluation, the following four media were tested in detail: tryptic soy agar, CDC anaerobic blood agar, Schaedler blood agar, and Difco brewer anaerobic agar. The resulting data for each of the 22 strains, which were inoculated both in a stressed and in an unstressed state, were statistically analyzed.
In the 20–25 °C aerobic incubation group, Schaedler blood agar gathered 30 credits, the CDC anaerobic blood agar 27 credits, tryptic soy agar 24 credits, and Difco brewer anaerobic agar 17 credits (Table III). P. acnes and Cl. sporogenes gathered no credits, as they don't grow aerobically. Stressed A. brasiliensis, B. pumilus, B. sphaericus, and B. idriensis got 0 credits due to low counts on all tested agars (0–5 CFU).
Total Count of Credits and Counted CFU in Brackets for the Unstressed Micro-organisms (Left Side) and for the Stressed Micro-organisms (Right Side) Cultivated at 20–25 °C Aerobically
In the 30–35 °C aerobic incubation group Schaedler blood agar gathered 28 credits, tryptic soy agar 28 credits, CDC anaerobic blood agar 27 credits, and Difco brewer anaerobic agar 24 credits (Table IV). P. acnes and Cl. sporogenes gathered no credits, as they don't grow aerobically. Stressed B. idriensis and M. osloensis got 0 credits due to low counts (0–5 CFU).
Total Count of Credits and Counted CFU in Brackets for the Unstressed Micro-organisms (Left Side) and for the Stressed Micro-organisms (Right Side) Cultivated at 30–35 °C Aerobically
In the 30–35 °C anaerobic incubation group, CDC anaerobic blood agar gathered 25 credits, the Schaedler blood agar 25 credits, tryptic soy agar 21, and Difco brewer anaerobic agar 17 credits (Table V). Stressed B. idriensis got 0 credits due to low count (0–5 CFU). This means Schaedler blood agar has the best growth-promoting properties for the aerobic conditions, and in the anaerobic condition it gained the same amount of credits as CDC anaerobic blood agar.
Total Count of Credits and Counted CFU in Brackets for the Unstressed Micro-organisms (Left Side) and for the Stressed Micro-organisms (Right Side) Cultivated at 30–35 °C Anaerobically (Not All Micro-organisms Grow Anaerobically, Therefore No Count)
Therefore the Schaedler blood agar was demonstrated to be the most suitable agar for the 22 tested strains (stressed and unstressed) and for the three tested incubation parameters. The composition of Schaedler blood agar (9) (modified by Heipha) is as follows: Caseine peptone 10 g, soy flour peptone 1 g, meat peptone 2 g, meat extract 1 g, yeast extract 5 g, dextrose 2 g, NaCl 5 g, dipotassium hydrogen phosphate 2.5 g, agar bacteriological grade 15 g, sheep blood 50 mL; supplements: amino acids, buffer, hemin, vitamin K, pH 7.3 ± 0.2, gamma-irradiated: 9–20 kGy.
Differences between Aerobic Incubation at 20–25 °C and at 30–35 °C
A t-test was performed using the data from the nutrient media evaluation to show if there is a significant difference between the incubation temperatures 20–25 °C and 30–35 °C (aerobic incubation). The t-test was done on 40 groups consisting of all aerobic micro-organisms (without Cl. sporogenes and P. acnes) in both stressed and unstressed state. In 11 out of the 40 groups a significant difference occurred. Five times the incubation parameter 30–35 °C showed higher amounts of CFU (coming from the same inoculum) compared to the 20–25 °C group. The 20–25 °C group produced higher values in six cases. This shows that there is a significant difference in growth behaviour between the two temperature ranges.
No Bioluminescent Background in Milliflex Rapid Detection
As the nutrient medium defined in this study will be used in the Milliflex Rapid Microbiology Detection System, a test was performed to determine if the nutrient medium causes a disturbing bioluminescent background. For this the media cassettes were incubated with the Milliflex Rapid membrane (MXHVWP124, Polyvinylidenefluoride membrane, pore size 0.45 μm), which had been rinsed with either 100 mL of Fluid A or Fluid D.
The incubation temperature was 30–35 °C, incubated for 5 days. Figure 11shows that there is no bioluminescent background coming from the Schaedler blood agar, meaning no disturbing ATP is left in the gamma-irradiated nutrient medium.
Background test of Schaedler blood agar: Milliflex Rapid membranes MXHVWP124 were incubated on Schaedler blood agar in Milliflex cassettes for 5 days at 30–35 °C. Figure 11a shows the Milliflex Rapid image; the membrane in this case was rinsed with 100 mL Fluid A. Figure 11b shows a membrane rinsed with 100 mL Fluid D.
Discussion
The advantage of the Milliflex Rapid System as a method for rapid sterility testing is the fact that it is a growth-based method and therefore with this aspect close to the traditional sterility test. In most microbiological tests in the pharmaceutical industry, growth-based methods are used because this is state-of-the-art in pharmaceutical microbiology.
The ATP bioluminescence assay, which is the assay used in the system for rapid sterility testing, has long been recognized as being “by far the most convenient and reliable method for estimating total microbial biomass in most environmental samples” (10).
The traditional sterility test uses FTM, which contains two phases—the lower phase of the liquid medium is an anaerobic phase, the upper phase contains oxygen for aerobic incubation. Since for more than 30 years the growth-promoting properties of FTM were questioned (11), we were now able to evaluate a new suitable medium for the rapid sterility test. We were now also in the good position to use a solid nutrient medium: it has been reported in the past that solid media have advantages over fluid media (12).
Stressed micro-organisms should be used for the validation of the rapid sterility test due to the fact that most contaminations in sterile products are expected to be caused by micro-organisms which are stressed, for example, following treatment with disinfectants or exposition to heat or dehydration. This means a stress representative of the production process had to be found. Micro-organisms (both stressed and unstressed) have to be detected in a sterility test. Therefore it has to be assured that growth-promoting properties of the used media are demonstrated even with stressed micro-organisms and also that the rapid sterility test system is able to detect the complete range of these micro-organisms. At first it seemed easy to stress micro-organisms, but it was shown that not all stress factors lead to an enduring stress on growth rates. Heat is known to cause damage on cytoplasmic membranes, proteins are denatured, and this leads to the death of some cells. UV irradiation causes mutations and a halt of DNA replication (13). The application of the used parenteral drug product lead to a chemical stress in the microbial cells—the antimicrobial properties of this specific parenteral drug product have been known for a long time. During this study it was confirmed that some stress factors, such as chemical stress caused by a specific parenteral drug product and radiation stress caused by the application of UV-light, decimate the amount of inoculated micro-organisms. In contrast to that, a heat treatment leads to an initially reduced growth rate in some of the tested micro-organisms and allows a challenge for the rapid microbiological method for use in the pharmaceutical industry, as micro-organisms are not only tested in their unstressed state but also in a stressed state. Heat-injured cells have already been reported to take a time to recover from 3 to 4 h—this was confirmed in this study (14). Heat stress is not useful for the sporulating bacteria; also, other stress parameters are not feasible because sporulating bacteria show wide resistance towards stress (15). For example, B. pumilus is used as a radiation indicator bacterium (16). In the case of the sporulating bacteria, a different stress factor was used—these were stressed by nutrient depletion and therefore a higher spore content was induced.
Rapid sterility test products and rinsing fluid(s) are filtered through membrane filters and then the membrane filters are transferred onto solid nutrient media. One solid nutrient medium in the rapid sterility test should be equivalent to or better than the liquid TSB from the traditional sterility test; a solid nutrient medium/or the same nutrient medium should be equivalent to or better than the anaerobic phase of FTM (traditional sterility test), and a third solid nutrient medium or the same nutrient medium should be equivalent to or better than the aerobic phase of FTM. Therefore it is planned to incubate one solid nutrient medium under aerobic conditions at 30–35 °C and one solid nutrient medium under anaerobic conditions at 30–35 °C in the rapid sterility test to replace the two phases of FTM. This means to replace TSB and FTM there will be three filters and three solid nutrient media cassettes: one is incubated aerobically at 20–25 °C, two are incubated at 30–35 °C (one aerobically, the other one anaerobically). The nutrient media evaluation and statistical analysis revealed that Schaedler blood agar (now known as rapid sterility test medium, in short RSTM) will be applied in the validation and routine use of the rapid sterility test. With the background test it was shown that Schaedler blood agar is suitable for its intended use in the Milliflex Rapid System. It was shown that not gamma-irradiated Schaedler Blood Agar leads to the formation of a high background in Milliflex Rapid detection. The ATP contained in the Schaedler Blood Agar is therefore most likely destroyed by the gamma-irradiation process using 9–20 kGy.
During in house growth promotion test of each lot of used Schaedler blood agar (for validation and routine use of the rapid sterility test) a test for bioluminescent background is performed.
To show that there are no differences between plate count and count on membrane filters the transferability from solid media to membrane filtered microbial growth on solid media had to be shown. The membrane used during filtration could be a barrier for the diffusion of nutrients from the agar. The validation using plate count method was started without membrane filtration mainly due to practical reasons, as thousands of petri plates had to be inoculated, incubated and counted. Reports are available, that show that membrane filtered bacteria can even be better quantified using the membrane filtration technique (17). This means that a validation using plate count method is a worst case validation. One has to be aware that we are discussing a prestudy to the validation of a rapid sterility test. The limit of detection and specificity studies as a part of the method validation of the rapid sterility test are most important to show that all chosen micro-organisms grow on Milliflex membranes in the rapid sterility test. In this limit of detection/specificity study the traditional sterility test is directly compared to the rapid sterility test—it has to be shown that the rapid sterility test promotes equivalent or better growth than the traditional liquid media. In these studies the sensitivity of the system is tested directly on filters.
The stressed micro-organisms used for the nutrient media evaluation can and even should be used for validation of rapid detection systems, as we have to challenge rapid detection systems and also the traditional methods. A rapid sterility test validation should include “artificially stressed” micro-organisms, due to the fact that the sterility test samples under test in most cases never include any micro-organisms. During the validation of rapid methods for example for bioburden testing or water analysis one can use natural water/bioburden samples. These samples include “naturally stressed” micro-organisms. For this kind of validation we do not think it is necessary to use “artificially stressed” micro-organisms.
The t-test to analyze the differences between aerobic incubation at 20–25 °C and at 30–35 °C showed that there is a significant difference of growth between the two temperatures. As the difference was significant in 11 out of 40 cases it is necessary to validate both incubation temperatures in the rapid sterility test. These results point out the importance of both incubation temperatures not only for the traditional sterility test but also for the rapid sterility test.
Future projects for the rapid sterility test are the finalization of the validation following the guidelines in the pharmacopoeias (18, 19) and the PDA Technical Report 33 (20) and also the identification of contaminants on the Milliflex Rapid membranes.
Acknowledgements
We thank Dr. Dietmar Roczen (Heipha, Eppelheim, Germany) and Dr. Frederic Marc (Millipore, Molsheim, France) whose excellent support has been of value. We also thank the sterility test laboratory team involved in the project and our colleagues for their critical reading of the manuscript.
Footnotes
-
Conflict of Interest Declaration
The authors declare that no financial or non-financial competing interests are related to the manuscript presented.
- © PDA, Inc. 2010
Jump to section
Related Articles
Cited By...
- Rapid Sterility Test Systems in the Pharmaceutical Industry: Applying a Structured Approach to Their Evaluation, Validation and Global Implementation
- Validation of Milliflex(R) Quantum for Bioburden Testing of Pharmaceutical Products
- Investigation of the Detection Ability of an Intrinsic Fluorescence-Based Bioaerosol Detection System for Heat-Stressed Bacteria
- Validation of the BacT/ALERT(R) 3D System for Rapid Sterility Testing of Biopharmaceutical Samples
- Identification of Micro-Organisms after Milliflex Rapid Detection--A Possibility To Identify Nonsterile Findings in the Milliflex Rapid Sterility Test