Abstract
An aerosol microbial ingress test was specifically designed and used to create a predictive model in order to determine the maximum allowable leakage limit (MALL) of single-use systems (SUSs). The MALL is defined as the greatest leak size that does not pose any risk to the product. The procedure involved taking test samples of film material from single-use bags. As test samples, an ethylene vinyl acetate multilayer film (300 μm thick) and a polyethylene multilayer film (400 μm thick) were cut into 50 mm patches. Artificial defects of 1–100 µm were laser-drilled in the middle of each film patch. The patch was assembled on a holder and properly sealed. The test units were filled aseptically with culture media and placed inside an aerosol chamber. Various pressures were applied to the test unit to simulate the constraints that single-use systems may be subject to under real-world conditions. After an aerosolization cycle with spores of Bacillus atrophaeus, a minimum concentration of 106 CFU/cm2 was reached on the film surface. The test units were incubated for 14 days at 30°C–35°C and visually inspected for bacterial ingress. Thirty samples per defect size were tested. Logistic regression was used to indicate the MALL for a single-use system according to the required risk level. With this method, the probability of the occurrence or absence of ingress for a specific defect size was reported according to the experimental data. In addition to physical parameters, such as the pressure applied and the film material, the effect of the probabilistic nature of the microorganisms in determining the MALL is considered. Although finding an experimental model to predict the MALL for real-life process conditions was the ultimate objective, this paper also presents the microbial ingress test data obtained so far for two extreme conditions. Potential constraints, such as vibration, shock, acceleration, liquid movement, and pressure differentials, observed during normal usage were simulated using two extreme differential pressures, 0 mbar and 300 mbar. The estimated MALL for typical use-case conditions are 10–20 µm for storage applications and 2–10 µm for shipping conditions. The microbial integrity test method used in this article was able to detect bacterial ingress down to 3 µm defect size.
LAY ABSTRACT: As use of single-use systems (SUSs) is increasingly expanding into all process steps of commercial manufacturing, integrity failure can significantly impact drug safety, availability, and costs. Consequently, growing industry scrutiny on single-use system integrity (SUSI) is raising the need to develop good science behind reliable determinations of liquid leakage and microbial ingress as well as the appropriate physical integrity testing technologies. In the current study, microbial ingress testing by the aerosol method is used to determine the maximum allowable leakage limit (MALL) for SUSs. To define the MALL, it is generally assumed that a system or product will not show any microbial ingress or leakage at a certain defect size. Statistical analysis of the experimental data in this study indicated the MALL with probability at a certain defect size for each system. As a result, the method studied provides a more accurate way of predicting ingress and increasing safety down the line for drug manufacturers and patients alike.
- Single-use system integrity (SUSI)
- Maximum allowable leakage limit (MALL)
- Single-use system (SUS)
- Microbial ingress testing
Introduction
An important aspect of pharmaceutical product quality assurance is to demonstrate the integrity of a container and closure system throughout the product life cy-cle, as it is critical to ensure product sterility and safety (1⇓–3). Sterility tests performed on final products have certain limitations. Therefore, augmenting this method with alternative controls to confirm the integrity of the container and closure system as a component of the stability protocol for sterile products is recommended (3). These alternative tests include physical, chemical, or microbiological integrity tests. There are valuable studies in which physical and microbiological integrity tests were used to check the integrity of glass vials (4⇓⇓–7), food cans (8, 9) and even flexible bags like retort pouches (10⇓–12). Although these nondestructive, deterministic physical tests are capable of identifying the existence of defects, they are not able to directly measure the threshold defect size (11).
Because one of the major concerns associated with package integrity is identifying the maximum allowable leakage limit (MALL), validation tests are required to demonstrate the microbial integrity of packages up to a certain defect size.
Two main microbial integrity test methods have been used so far—a microbial liquid-immersion test, in which sterile culture media–filled containers are challenged by immersion in a liquid microbial suspension (7, 13⇓⇓–16), and a microbial aerosolization test, which is the focus of this paper. This latter method uses an airborne microbial suspension (nebulization) to challenge containers filled with sterile growth-support media (11, 12, 17⇓⇓⇓–21). In both methods, incubation is performed at a growth-promoting temperature, and cultures are subsequently checked for evidence of microbial growth.
The microbial ingress test can be performed without pressurization as well as with pressure or vacuum to simulate the constraints observed in different conditions, for example, transportation by airfreight (7, 11, 21, 22). This test can reflect “intended use” or “worst-case” conditions.
According to the severity of the case and the ambient conditions, various types of microorganisms have been used, including Brevundimonas diminuta, Escherichia coli and Serratia marcescens (5⇓–7, 11, 23). In the present study, Bacillus atrophaeus, a common reference microorganism used for microbial ingress testing by aerosolization, was utilized. The rationale behind this selection is explained in detail in the Materials and Methods.
Various studies have already reported on microtubes, microwires, pinholes, and orifices that were positioned in different parts of the intact container, such as on welds and folds, to representatively simulate physical defects (4⇓⇓–7, 11, 12, 20, 21).
However, the authors have selected laser-drilled holes to mimic the type of failure mode that most needed to be considered (24). Laser-drilled holes were also a more practical and controllable option among all of the artificial leak types that currently exist (25). Furthermore, evaluating the defects of welds and folds was beyond the scope of this paper.
Traditionally, microbial ingress testing is performed with a liquid-immersion method as described in USP <1207> (26) that entails exposing sterile packaging to the test microorganism in worst-case conditions (bacteria concentration, exposure time, and full immersion as a contact method with the challenge solution). However, it is expected that no flexible bulk container will be subjected to liquid immersion during storage and shipping. Therefore, microbial ingress testing by aerosolization is better suited to SUSs, as it is more representative of the real conditions of use for flexible packaging (11). In addition, Table I summarizes each MALL obtained for different products using liquid-immersion and aerosol methods in previous studies (7, 21, 22, 27). A MALL of 2 µm resulting from the integrity test performed by Keller is a good example that confirms that a microbial-challenge test using aerosolization can be as stringent as a liquid-immersion test.
MALLs Obtained With Microbial Ingress Testing by Liquid Immersion and Aerosol Methods in Different Studies
Although in the study conducted by Pethe et al. (28) the MALL achieved by the aerosol method is much greater than that obtained by the liquid-immersion method, the integrity test method proposed in Technical Report 27 (29) does not necessarily consider worst-case testing parameters. Therefore, to design an appropriate microbial ingress test using representative conditions, the purpose for which the container, that is, the bag will be used must be considered.
The objective of the present study is to describe an appropriate microbial testing method and its validation with the final goal of determining the MALL for SUSs during their product life cycles, including storage, mixing, and shipping after filling.
The SUSs under study were made of ethylene vinyl acetate (EVA) film and polyethylene (PE) film. The microbial integrity test method used by the authors was the aerosol test. This method, the equipment, and its qualification are presented in this report. In addition, this paper considers the different parameters that can affect microbial ingress by aerosol exposure, such as leak size, material thickness, pressure inside and outside the packaging, process conditions, type of microorganisms, and concentration of the inoculum.
Materials and Methods
Test Microorganisms
As challenge microorganisms, spores of B. atrophaeus (ATCC®9372TM) were used. This bacteria species was selected for two main reasons: First, it is spore-forming, with a small spore size of 0.1 to 0.4 µm. Second, B. atrophaeus spores are resistant to dry conditions and are not destroyed by aerosolization. In addition, under its former name of Bacillus subtilis, it is recommended to be used for bacteria ingress testing by liquid immersion in ISO 15747 (30).
Test Unit Assembly
As test samples, an EVA multilayer film (300 μm thick) and a PE multilayer film (400 μm thick), the materials used in Flexboy, Celsius, and Flexsafe bags (Sartorius Stedim Biotech), were each cut into 50 mm diameter patches.
To simulate a pinhole defect, the patches were laser-drilled in the center with different microhole sizes ranging from a nominal size of 1 μm to 100 µm. The orifice leak size of laser–drilled holes was calibrated by Lenox Laser using flow measurement (26). Laser drilling was done with a Lenox Laser.
A polypropylene patch holder secured the film patch on each screw cap with two silicon gaskets on both sides of the patch. To fill the test unit aseptically, a needleless connector was glued to the holder body. To permit air to circulate in the media, a silicon tube with a 0.2 µm filter was connected to the bottom of the holder. Using a 0.2 µm filter also kept the test unit sterile during integrity testing. The test unit is shown in Figure 1.
The test unit used to simulate a defect in a single-use system.
This test unit with a laser-drilled defect is representative of actual full-size bags with an artificial defect. Different aspects of this similarity and representativeness have been previously tested.
In addition, assembling the laser-drilled film patch on a small holder instead of attaching it to the actual package enables the bacteria ingress test to be performed by aerosolization for more than 30 samples in each run. This improves the statistical results and is a key benefit of the method presented by the authors.
The fully assembled test unit, including the patch, was air leak-tested using a Sartorius Sartocheck 4 plus Bag Tester at 300 mbar pressure, with a pressure stabilization time of 240 s and a test time of 240 s.
This test is required to check the integrity of the entire assembly besides the defect itself and to confirm that the microhole is not blocked in the process.
The test units were gamma-irradiated at a dose between 25 kGy and 50 kGy. To check whether irradiation had an impact on the integrity of the patches and the microhole behavior, 10% of the gamma-irradiated test units were air leak tested again using the Sartocheck device with the same testing parameters applied.
Aerosol Chamber and Aerosolization Cycle
The exposure chamber with a volume of approximately 1 m3 is made of stainless steel. It has two air inlet ports and four stainless steel ducts for aerosol circulation. A support plate with 36 slots and plate holders is located inside the chamber for positioning the samples.
A volume of spore suspension corresponding to a minimum of 4 × 1011 spores was placed in the reservoir of the liquid nebulizer. This quantity used in equipment performance qualification is the minimum quantity needed to attain 106 CFU/cm2 on the surface of the film patch in each test unit.
To achieve a microbial concentration of a minimum of 106 CFU/cm2 on the surface of each test unit, the following parameters were applied and controlled during the aerosolization cycle. The pressure was 3.5 ± 0.5 bar inside the liquid nebulizer; the temperature was 20 ± 5°C and the humidity was 50 ± 10% inside the aerosol chamber. To measure and control the temperature and humidity, two probes were used to connect the thermometer and hygrometer to the aerosol chamber.
The previously mentioned spore challenge of 106 CFU/cm2 on the film surface was selected to represent worst-case conditions with regard to usual applications for SUSs. Even though these systems are typically used in ISO Class 7 conditions, the ISO Class 8 surface contamination specification was used and augmented by an additional six logs (31). The spore concentration was verified on two film patches, which were placed on an empty holder and exposed to an aerosol cycle based on the qualified method.
Microbial Ingress Procedure
Each routine test examined a maximum of 34 test units with defined defect sizes and one positive control representing a compromised sample that had a large defect (3 mm). The compromised sample was exposed to aerosol to confirm that the aerosolization process was able to cause bacterial growth inside the holder. One nonexposed negative control was used to challenge the aseptic conditions and to visually compare growth after incubation to evaluate the results. In addition, one exposed negative control was used to check the integrity of the test unit during the entire process, from preparation to incubation.
Test units, the compromised sample, and the exposed negative controls were aseptically filled with tryptic soy broth (TSB; Merck), pressurized, exposed to aerosol, and removed from the aerosolization chamber for microbial ingress testing; after individual protection, they were incubated for 14 days at 30°C–35°C. The nonexposed negative control was left under a laminar flow hood during aerosolization and incubated later with the other test units.
Finally, the presence or absence of growth was determined by visual inspection and compared with that of the nonexposed negative control. The growth-promotion test was conducted by spiking B. atrophaeus (10–100 CFU) in the exposed negative control after the incubation period to verify that the media had retained their growth-promoting characteristics throughout the entire process.
During its life cycle, each SUS may be exposed to different pressures. Therefore, to qualify and assess the effect of different pressure levels applied to SUSs, apart from atmospheric pressure, the film patches were tested under different use-case pressure conditions representing storage, shipping, and mixing. This can be used later to establish a predictive model to determine the MALL under use-case conditions. However, so far testing has only been performed at atmospheric pressure and 300 mbar as the extreme points used to build the predictive model.
The sequence of the routine runs for these testing pressures are presented in Figure 2.
The sequences done for each run of microbial ingress testing by aerosolization.
Qualification
Qualification of the chamber and aerosolization process was performed according to the classic three-step plan of installation qualification (IQ), operational qualification (OQ), and performance qualification (PQ).
Results and Discussion
A total of 30 patches per defect size were evaluated in different experimental runs using the aerosol method of the microbial ingress test. However, based on the prediction of having microbial ingress at a defect size of around 20 µm for atmospheric pressure, fewer samples were tested for a 2 µm defect. The rationale behind the selection of the number of test samples is explained in the Appendix. For all runs, the negative controls (exposed and nonexposed) demonstrated that no contamination occurred during preparation of the test units. The results obtained from compromised samples show that test units with a defect allowed microbial growth. The growth-promotion test demonstrated that the growth-inducing characteristics of the media were maintained up to the end of the incubation period. The spore concentration on the film patch used for verification of the concentration (film patch without a defect and on an empty holder) was a minimum of 106 CFU/cm2 for each run. The aerosolization parameters were compliant with the cycle validated in PQ. Because all acceptance criteria were met, the results were conclusive.
In most studies conducted in the field of integrity testing, a specific defect size is commonly reported as the MALL for individual systems (5, 12, 21, 23, 32). Yet in the case of microbial ingress testing, it is not realistic to sharply define the transition between the actual absence or occurrence of microbial ingress because of the variability of microorganisms in terms of their size, shape, spore-forming capability, motility, and environmental growth conditions.
In addition, there are experimental limitations in considering all of the physical parameters that can affect the microbial integrity of a SUS. Therefore, to report a MALL for a SUS, it is more appropriate to talk about the probability of the occurrence or absence of ingress for a specific defect size. The binary nature of the results makes it easier to use logistic regression to calculate and predict the probability of having growth or no growth for a certain defect size based on the experimental data.
The results of the microbial ingress testing by aerosolization for the PE and the EVA films with a defect size between 2 μm and 100 μm at atmospheric pressure are reported in Table II.
Number of Tested Samples that have Shown Bacterial Ingress at Certain Defect Sizes Compared to the Total Number of Tested Samples for PE and EVA Films at Atmospheric Pressure
For the EVA film at an applied pressure of 0 mbar, one single incident of bacterial growth was reported at 40 μm, and for samples with a defect size <40 μm, no growth was reported.
For the PE film at an applied pressure of 0 mbar, one single incident of bacterial growth was reported at 20 μm, and for samples with a defect <20 μm, no growth was reported.
Figure 3 shows the probability of bacterial ingress as a function of defect size based on statistical analysis of the experimental data at an applied pressure of 0 mbar for PE and EVA films.
The bacterial growth or absence of growth is shown with gray circles as a function of defect size for 0 mbar of applied pressure. The probability of bacterial ingress expressed as a function of different defect sizes (solid curve) with a 95% confidence interval of (0.5%, 3.4%) and (0.97%, 6.4%) for (A) PE and (B) EVA respectively (dashed curves).
Logistic regression indicates that the probability of bacterial ingress for the PE film at a defect size of 20 µm is 1.3%. For a defect size of 15 µm, even though no bacterial ingress was reported experimentally, this probability is ∼0.89%.
Similar results for the EVA film indicate that the probability of bacterial ingress at a defect size of 40 µm is 2.5%. For a defect size of 30 µm, even though no bacterial ingress was reported experimentally, this probability is ∼1.3%.
Therefore, determining the MALL for a SUS needs to be statistically evaluated.
According to prior studies, it is necessary to have liquid in the defect pathway for microbial growth to occur (29, 32, 33). The threshold pressure required to initiate liquid flow through a microchannel is an inverse function of the microchannel diameter size.
Performing a microbial ingress test at higher pressure provides the opportunity to search for a smaller MALL at the same time. For 300 mbar, the defect sizes are therefore chosen in a smaller range.
Table III shows the results of the microbial ingress test for PE and EVA films with a defect size between 1 µm and 10 µm at an applied pressure of 300 mbar. Statistical analysis of the preceding results is shown in Figure 4.
The bacterial growth or absence of growth is shown with gray circles as a function of defect size for 300 mbar of applied pressure. The probability of bacterial ingress expressed as a function of different defect sizes (solid curve) with a 95% confidence interval of (3.4%, 13%) and (11.6%, 25.6%) for (A) PE and (B) EVA respectively (dashed curves).
Number of Tested Samples that have Shown Bacterial Ingress at a Certain Defect Size Compared to the Total Number of Tested Samples for PE and EVA Films at 300 mbar Pressure
Two incidents of positive growth out of the 30 samples with 2 µm defects and nine incidents of positive growth out of the 30 samples with 3 µm defects are reported for PE and EVA films, respectively. Similar to the evaluation of the test results at atmospheric pressure, statistical analysis shows that the probability of bacterial ingress for a defect size of 2 µm in the PE film is 6.8%. The same statistical analysis of the EVA film shows that the probability of bacterial ingress for a defect size of 3 µm in the EVA film is 17.5%, although the probability of bacterial ingress for a 1 µm defect in the PE film and a 2 µm defect in the EVA film is around 4.4% and 13.6%, respectively.
Figure 5 shows the threshold defect size above which microbial ingress can occur as a function of applied pressure. The threshold pressure required to initiate liquid flow through a known defect is an inverse function of the defect size. Therefore, the existing experimental data for the PE film are fitted to the predictive model using nonlinear fit. Based on the application-specific pressure conditions, MALL can be predicted with a certain probability using this model.
Threshold defect size as a function of pressure applied. The dashed line is the fitted line to the PE film experimental data according to the predictive model. Two ovals show the estimated position of the defect size for certain applications.
Correspondingly, the MALL given for low-pressure storage conditions is significantly higher, likely in the range of 10–20 µm, compared to the one associated with high-pressure shipping conditions, in the range of 2–10 µm. To improve the validity of the model, testing with additional pressure data points will be performed in a future study.
Summary and Conclusion
This paper examines the effect that defect size, as represented by laser-drilled holes, and applied pressure have on microbial ingress. A novel integrity test system was designed to be representative of a SUS made from PE or EVA multilayer films. These test systems were challenged by microbial aerosols to determine their MALL.
The experimental results obtained for PE and EVA films with various defect sizes at different applied pressures are provided and discussed. The number of test samples showing bacterial growth, that is, ingress, was determined. According to USP < 1207> (26), the greatest leak size that does not pose any risk to the product is considered the MALL.
The statistical approach discussed in the present paper allows only the probability of predicting the MALL for a certain defect size to be determined.
The probabilities of microbial ingress of 0.89% and 1.3% at atmospheric pressure and 4.4% and 13.5% at 300 mbar for PE and EVA calculated in this study are in line with the level of integrity assurance compared with the probability of 10% microbial ingress given in USP < 1207> (26) for the MALL in primary packaging. This establishes a MALL of 10–20 µm for controlled storage applications and a MALL of 2–10 µm for more aggressive shipping applications.
In conclusion, the microbial ingress test by the aerosolization method described in this paper should provide a more accurate prediction of SUSI level. Although artificial leaks, such as those created in this study, may not be viewed as completely representative of actual leaks encountered in SUSs, the authors believe that this predictive model and test scheme can be used to develop and implement a physical test method for improving safety for pharmaceutical manufacturers and patients alike. The limit of detection for the physical test method should be on the same level as the MALL determined here.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
Acknowledgments
The tests in this study were performed by Confarma, an independent laboratory, and the test results were validated by the authors. The latter wish to express their sincere gratitude to Confarma for providing technical assistance and insights.
Appendix
There are several studies in the field of integrity testing with different numbers of samples tested for each defect size. We used 30 samples per defect size; in the study done by Keller (27), seven samples per defect size were used, and in the study by Moghimi (13), 250 samples per defect size were used. However, the rationale behind these numbers is not clear or relevant, so we would like to explain how we proceeded to answer the following questions:
How many samples are appropriate for this kind of integrity test study?
How will these numbers affect the results and how can we explain these effects?
There are planning and performing tools like DoE or 3Pod to design the optimal number of experiments required to achieve the appropriate levels of statistical power and sensitivity.
In our study, owing to the binary nature of the data we had, we were not able to use DoE. For 3Pod, we needed some input to design the rest of the experiments. We started with 30 patches per defect size; however, this number of samples per existing defect size is far above what 3Pod predicts for the rest of the experiment.
In this case, the probability approach, which is based on experimental data, can be used to explain the results and predict the probability of growth occurring for any specific defect size that has not been examined before. This approach also shows that the growth for a certain defect size detected in a limited quantity of tests cannot exclude the possibility that growth would never occur. In other words, this transition from the occurrence of growth to its absence for a certain defect size is not an absolute phenomenon, rather it is based on probability.
Furthermore, we did the following step by step:
We performed 30 tests per each defect size; we examined how many of them exhibited bacterial ingress and growth. We reported these results as is.
We fitted our data with logistic regression, which is appropriate for binary data.
From there, we calculated the probability of growth occurring for a certain defect size.
- © PDA, Inc. 2019
