Investigation by molecular method of potentially non-compliant aseptic process simulation-media fill test (APS-MFT) and sterility test

An aseptic processing operation should be validated using appropriate microbiological growth medium in place of the product. This process simulation, also known as “media fill test”, normally closely simulate the same exposure that the product itself will undergo. The sealed containers filled with the medium are then incubated to detect microbial contamination. Results are then interpreted to assess the potential for a unit drug product to become contaminated during operations (Guidance for industry, Sterile drug products produced by aseptic processing, cGMP, 2004).

Sterility tests are limited in their ability to detect contamination because the small sample size typically used and the need for micro-organisms to be able to grow (lag phase and growth phase) under the conditions of the test. In APS-MFT, Trypto Casein Soja Broth (TSB) is mainly used, while in Sterility Tests (ST) the culture media used are Trypto Casein Soja Broth (TSB) and Thioglycollate Broth (TB), to allow microbial contaminant growth. They are now incubated for 14 days with intermediate and final readings, at 20-25 °C for 7 days and then 30-35°C for seven days for APS, and for ST at 20-25°C for TSB and 30-35°C for TB, respectively. These media and their corresponding incubation time and temperatures were chosen to maximize recovery of potential contaminants.
Both tests rely on the assumption that one contaminating cell would grow and result in media turbidity (unverified assumption).

Hence, contamination is detected based on turbidity of the media (initally limpid) detected by human vision (subjective) after incubation in appropriate conditions. Compliance or non-compliance is determined by the non-detection or the detection of a turbidity, respectively. Some of the main concerns faced in the interpretation of the tests are atypical turbid medium (faint turbidity) and no detection of viable micro-organisms after sub-culturing from turbid medium issued either from APS-MFT or ST. These question the microbiological origin of the turbidity instead of a non-microbial (i.e chemical) origin, and led to consider the test as “potentially non-compliant” (Microbial Data Deviation [MDD] or Deviation or Out Of Specification, depending on the procedure in use) until further investigated. A molecular method for investigation of these “potentially non-compliant” tests, based on the detection of bacterial and fungal DNA by Polymerase Chain Reaction (PCR) has beendeveloped and validated. The rational of the method, its format, the validation criteria and the results of more than 300 investigations of both APS-MFT and ST are reviewed.


1. Rational
As bacteria or fungi multiply in a liquid culture medium, the medium becomes turbid, or cloudy with cells. The premise is that a turbid liquid culture medium is turbid enough to be visible for the human vision when cell suspension reach a concentration of more than 1 x 106 cells/mL (106-107 cells/ mL for bacteria) or more than 1 x 104 cells/mL (104-105 cells/mL for fungi). This is supported by classical microbiology literature (e.g. Tortora et al., “Microbiology an introduction”, 7th edition, 2002, Pearson Education Inc. editor, Chapter 6 page 179 ; Sutton S., “Rapid Sterility Testing”, The sterility test, J. Moldenhauser Ed., 2011, page 19).

A practical reference for turbidity assessment is the Mac-Farland scale. McFarland Standards are used to standardize the approximate number of bacteria in a liquid suspension by comparing the turbidity of the test suspension with that of the McFarland Standard. A McFarland Standard is a chemical solution of barium chloride and sulfuric acid; the reaction between these two chemicals results in the production of a fine precipitate, barium sulfate. When shaken well, the turbidity of a McFarland Standard is visually comparable to a bacterial suspension of known concentration as indicated in table 1.



Hence, the lowest MacFarland Standard (0.5) corresponds to an approximate bacterial suspension that is ten times higher than the minimal bacterial suspension visible using human vision. This is of importance if MacFarland standard would be used during qualification of operators to interpretation of sterility test results (it is not enough!).

In figure 1, the visual aspect of the corresponding turbidity is shown against a striped background.



The shape and size of the bacterial and fungal cells (yeasts) are varied and influence greatly the correlation between observed turbidity and approximate cell concentration. For instance, at the same cell concentration, yeasts cells that would be 10 times bigger than bacterial cells will generate roughly a ten times higher turbidity than bacterial cells.

In order to test the hypothesis that the turbidity of a liquid medium could be originating from micro-organism’s cells being present in suspension at the required high concentrations we have thus developed a PCR method to detect bacterial or fungal DNA. The method is not a growth-based method and both viable and non-viable cells could be detected with low limit of detection.


2. PCR Method for the detection of bacterial and fungal DNA
The principle of the method is to collect the cells in suspension by centrifugation, then get rid of the microbial residual DNA in solution and finally extract the nucleic acids from cell pellets. Then the nucleic acid extracts are submitted to PCR amplification using universal primers for either bacterial DNA or fungal DNA (two separate PCR) and optimized conditions. Appropriate controls are included in order to determine any possible inhibition of the PCR step related to the matrix (external control made up with the sample spiked with calibrated suspension of bacteria and fungi), and appropriate negative controls (including the same lot of culture media as the one used for MFT or ST), according to our standard operating procedure.

The format of the PCR test is described in figure 2. An example of the results obtained after agarose gel electrophoresis is presented in figure 3.



3. Validation of the PCR method for the detection of bacterial and fungal DNA
A global validation of the method of detection of bacterial and fungal DNA by using PCR with universal primers, with generic matrices (TSB and TB) according to ICH-Q2R for testing of impurities, limit assay recommendation for a qualitative test, has been performed. Criteria for the validation of the method were specificity and limit of detection.

The specificity of the method has been established both theoretically, against representative species of the main taxonomic groups (primers specificity), and practically, using reference strains of specified species of bacteria and fungi required for validation (Ph. Eur. § 2. 6. 1) and wild strains representative of the species found in pharmaceutical industries. The method for detection of bacterial DNA was specific of bacteria and the one for fungal DNA detection was specific for fungi.

The limit of detection (LOD) was determined using calibrated suspension of the specified species of bacteria and fungi required for validation (Ph. Eur. § 2. 6. 1) spiked in either TSB or TB. The LOD was of less than 1 x 103 cells/ mL for bacteria and less than 1 x 102 cells/mL for fungi, whatever the medium. This is roughly 100 times more sensitive than requested to test the hypothesis that turbidity is of microbial origin (more than 106-107 cells/ mL for bacteria or more than 104-105 cells/mL for fungi).

An example of the results obtained is presented in figure 3.

Legend: S molecular weight standard 100 bp lader. 1a and 1b : replicate PCR tests of sample 1 aliquote, 1T external control: PCR test of sample 1 aliquote spiked with calibrated suspensions of bacteria and fungi, T+ positive control, T- negative control medium (TSB), Bl: extraction blank, Blm: mix negative control. Same organization for sample 2 and sample 3.

The negative controls were compliant as well as the positive control. For all samples external controls (1T, 2T and 3T) were positive as expected demonstrating the absence of inhibition of the PCR in the test conditions with the samples.
For sample 1, replicate 1a and 1b were both positive for bacterial DNA amplification. To the contrary, sample 2 and sample 3 were negative. The results supported the hypothesis that the turbidity in medium sample 1 was of bacterial origin, but did not support the hypothesis for sample 2 and 3.



4. Results of more than 300 investigations of APS-MFT and ST
Out of 300 investigations of potentially non-compliant tests 58 % were negative for both bacteria and fungi detection. Hence the investigations were not supportive of a microbial origin of the observed turbidity. For some samples, further investigations have thus been performed to test the hypothesis of a chemical origin of the turbidity. This was confirmed for some of them using electronic microscopy and other methods. This stressed out the interest of investigations in case of “potentially non-compliant” tests, to figure out the origin of the observed or suspected turbidity. For the remaining 42 % of “potentially non-compliant” tests, the detection of either bacterial DNA (95% of cases) or fungal DNA (5% of cases) was positive. Hence, the investigations were supportive of a microbial origin of the observed turbidity and the non-compliance was definitely confirmed. From the positive amplification we were able to identify by comparative sequencing (genotypic method) the microbial species involved. The main species of bacteria found were mainly Gram-positive and few were Gram-negative. The results of identifications generated form positive detection of bacterial DNA in APS-MFT or ST are presented in figure 4. This permitted then to compare the species involved to the species recovered during environmental monitoring, sampling of gownings, etc…


Since the multiple consequences of any non-compliance for APS-MFT and ST, any suspected contaminated unit should result in an investigation. The need to distinguish between “Microbial Data Deviation” (MDD = not yet confirmed true product specification failure) and non-compliant sterility test (true product specification failure and all consequences for sterility assurance can now be adressed by this molecular method for bacterial DNA and fungal DNA detection.The method is used to test the hypothesis that an observed turbidity could be of microbial origin even despite positive growth confirmation after subculturing. This investigation method is absolutely not intended to replace the sterility test. It should be part of the investigations required when potentially non-compliant APS-MFT or ST are faced and will help robust scientific documentation of any MDD or true positive APS-MFT or ST.

We gratefully thank Thierry Bonnevay for the valuable review of the manuscript.

Share article

Capture D’écran 2019 01 23 À 09.34.06


For more than 30 years, Arnaud CARLOTTI has been passionate about industrial microbiology and production issues in sterile conditions, investigations of non-conformities, root cause analyzes, proposal and implementation of CAPAs.
As a recognized expert, he has published more than 30 scientific articles and book chapters to date (eg. 2018 PDA/DHI publishing “Contamination Control volume 5 chapter 14).