Cell Tissue Bank (2014) 15:277–286 DOI 10.1007/s10561-014-9455-8
ORIGINAL PAPER
Validation of an alternative microbiological method for tissue products Susanne Suessner • Simone Hennerbichler Stefanie Schreiberhuber • Doris Stuebl • Christian Gabriel
•
Received: 1 August 2013 / Accepted: 30 April 2014 / Published online: 9 May 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract According to the European Pharmacopoeia sterility testing of products includes an incubation time of 14 days in thioglycollate medium and soya-bean casein medium. In this case a large period of time is needed for product testing. So we designed a study to evaluate an alternative method for sterility testing. The aim of this study was to reduce the incubation time for the routinely produced products in our tissue bank (cornea and amnion grafts) by obtaining the same detection limit, accurateness and recovery rates as the reference method described in the European Pharmacopoeia. The study included two steps of validation. Primary validation compared the reference method with the alternative method. Therefore eight bacterial and two fungi test strains were tested at their preferred milieu. A geometric dilution series from 10 to 0.625 colony forming unit per 10 ml culture media was used. Subsequent to the evaluation the second part of the study started including the validation of the fertility of the culture media and the parallel testing of the two methods by investigating products. For this purpose two product batches were tested in three independent runs. Concerning the validation we could not find any aberration between
S. Suessner (&) S. Hennerbichler S. Schreiberhuber D. Stuebl C. Gabriel Austrian Cluster for Tissue Regeneration, Red Cross Blood Transfusion Service of Upper Austria, Krankenhausstraße 7, 4017 Linz, Austria e-mail: [email protected]
the alternative and the reference method. In addition, the recovery rate of each microorganism was between 83.33 and 100 %. The alternative method showed noninferiority regarding accuracy to the reference method. Due to this study we reduced the sterility testing for cornea and amniotic grafts to 9 days. Keywords Sterility testing Validation alternative method Cornea Amniotic membrane Quality control Tissue bank
Introduction Sterility testing is an essential issue in the processing and release of tissue products. Sterility of primary materials, application of established procedures for sterilization as well as the short shelf life of many tissue grafts are of great concern in the release process (Montag-Lessing et al. 2010). Due to reports of tissuetransmitted infections implementation of more federal regulations have led to an increased safety of tissue transplants (Wang et al. 2007). Associations of bacterial infections with allografts have involved bones, tendons, dura mater, corneas, heart valves, skin and pericardium (Eastlund and Strong 2003). Several steps had been implemented reducing microbial contamination: criteria for donor selection, minimization of interval between death and
123
278
tissue removal, monitoring and control of microbial contamination of air and surfaces, use of aseptic surgical technique during dissection and removal of tissues (Eastlund 2006). Since 1910 amniotic membrane has been used in various clinical applications, reconstructive surgery, in therapy of burns and chronic ulcers (Matthews et al. 1982; Davis 1910). The transplantation of amnion grafts plays an important role in the treatment of ophthalmic disorders such as persistent corneal epithelial defects with ulceration or corneal burns (Lee and Tseng 1997; Dua et al. 2004). The source material for amniotic grafts may not always be sterile (Aghayan et al. 2012) and bacterial contamination is dependent on the way placental membranes are retrieved. Placental membranes obtained by caesarean sections have less burden of pathogenic organism than such of normal vaginal deliveries (Adds et al. 2001). Gram-positive bacteria are the most frequent microbial causes for infections of amniotic membrane transplants in ocular surgery (Marangon et al. 2004). Other important tissue products used in ophthalmologic surgery are corneal grafts. The main indications for keratoplasty combined with corneal transplantation are keratoconus, pseudophakic or aphakic corneal edema (Edwards et al. 2002). Corneal grafts are cultivated in culture media containing antibiotics to minimize the risk to transplant bacterial contaminated products (Dichtl et al. 2010). The main bacterial contaminants of donor corneas are skin germs, primarily Staphylococcus epidermidis (Pardos and Gallagher 1982). The European Pharmacopoeia characterizes the regulations for sterility testing (European Directorate for the Quality of Medicines & HealthCare 2010). The appropriate culture medium for detection of anaerobic bacteria is liquid thioglycollate medium, for aerobic bacteria and fungi soya-bean casein digest medium is recommended. The principle of these growth-based methods is a change in turbidity indicating the presence of viable micro-organisms. The incubation time is defined to be 14 days. Given to a relative short shelf life of some tissue grafts it is too long, meaning that cultivated corneal transplants with a shelf life of 28 days can only be transplanted 14 days upon receipt of negative microbial testing result. A faster microbial method for sterility testing increases the available pool of negative-tested tissue products and therefore
123
Cell Tissue Bank (2014) 15:277–286
improvement in terms of adequate and assured product supply can be guaranteed. In recent years several approaches were proven as a rapid alternative method to the common growth-based methods (Parveen et al. 2011). A possible option are automated blood culture systems which are suitable for microbiological testing of tissue preparations (Schroeter et al. 2012). The aim of the present study was the validation of an alternative qualitative microbiological testing method regarding shortened incubation time compared with the reference method of the European Pharmacopoeia, from which validation criteria for such alternative methods of microbiological quality control were used. The concept of our alternative method was only a change in terms of incubation period of microbiological testing while maintaining the described methods.
Materials and methods Study design As the aim of the alternative method was to reduce the incubation time for sterility testing, the liquid culture media were read out in an exact time interval to get a positive or negative result. For the comparison of the two methods the study was performed in two steps (Fig. 1). At first a comparison between the reference method and the alternative method was performed to define the specificity, limit of detection, relative accuracy, recovery rate and robustness of the alternative method. Subsequently the second validation step was done, validating the routinely processed tissue preparations (cornea, amniotic grafts and cryomedium) and comparing them with the obtained results of the first step of this study. Therefore two different lots of each tissue preparation were validated in three independent runs. A main issue of the first part of this validation was the visual assessment of the liquid culture media. Exact points of time for direct reading had to be defined (Fig. 2). All bouillons were mixed thoroughly on a vortex before evaluation. After the first runs it became evident that no positive results were observed after 18 and 24 h. Hence, the first reading was done at 39 h in the following runs.
Cell Tissue Bank (2014) 15:277–286
279
Fig. 1 Study design for validation of the alternative microbiological method for sterility testing
Fig. 2 Time lapse for evaluation of liquid culture media. h hours, d day
18 h
39 h
24h
66 h
48h
d5
d4
d7
d6
d9
d8
d11
d10
d13
d12
d14
h = hours; d = day
Test strains and methodology Ten different American type culture collection (ATCC) test strains were chosen to be used in this validation study representing a wide range of microorganisms (Table 1). Six of these ten test strains were chosen according to the European Pharmacopoeia Chapter 2.6.1. The test strains were reconstituted according to the manufacturer’s instructions. Then the required dilution could be produced. Each test strain was incubated in its preferred milieu (Table 1). Aerobic microorganisms were inoculated in liquid soya-bean casein medium (Biomerieux, France) and incubated at room temperature (20–25 °C). The anaerobic test strains were incubated at 30 °C in liquid thioglycollate medium (Biomerieux, France) due to another validation (unpublished data).
Turbidity of liquid culture medium indicated microbial growth. No visible changes yielded in a negative result. To assure negative result agar cultivation was performed. After the 14-day incubation trypcase-soya agar (Biomerieux, France) for the aerobic microorganisms and a Schaedler agar (Biomerieux, France) for the anaerobic test strains were plated and cultivated. The trypcase-soya agar plates were incubated for 2 days at 30 °C and Schaedler agar plates for 7 days at 37 °C in anaerobic atmosphere.
Primary validation The first part was the validation of the alternative method in relation to the reference method including
123
280
Cell Tissue Bank (2014) 15:277–286
Table 1 Test strains and incubation conditions Test strains
Strain ID
Incubation condition
Staphylococcus aureus subsp. aureus
ATCC 6538a
Aerobic
Bacillus subtilis subs. spezizenii
ATCC 6633a
Aerobic
Pseudomonas aeruginosa
ATCC 9027a
Aerobic a
Clostridium sporogenes
ATCC 11437
Candida albicans
ATCC 10231a
Aerobic
Aspergillus brasiliensis
ATCC 16404a
Aerobic
Escherichia coli
ATCC 8739a
Aerobic
Pseudomonas fluorescens
ATCC 49838b
Aerobic
Streptococcus pyogenes
ATCC19615c
Aerobic
Propionibacterium acnes a
ATCC 11827
b
Anaerobic
Anaerobic
Bioball
b
Epower microorganism
c
EZ CFU one step-microorganism
testing of specificity, limit of detection, relative accuracy, recovery rate and robustness. Specificity At first the specificity of the alternative method had to be determined. According to the guidelines (European Directorate for the Quality of Medicines & HealthCare 2010) the specificity of an alternative qualitative method is the ability to detect the required range of microorganisms that may be present in the tested sample. For this purpose three runs with ten test strains were performed. Each test strain was inoculated at a concentration of 10 colony forming units (CFU) in 10 ml culture media. All applied microorganisms had to be detected, thus implying a visual recognizable turbidity in 100 % of the bouillons.
show reproducible growth in the chosen microbiological method. To get enough comparable data the test strains were inoculated in each dilution for at least nine times. For this step a geometric dilution series of each microorganism was prepared. Therefore in preliminary tests we have cultivated the various dilutions in triplicates on agar plates (trypcase-soya agar plates for aerobes and fungi and Schaedler agar plates for anaerobes). The calculated average of the obtained CFUs acted as control of the inoculated microorganisms for the dilution series (data not shown). The primary validation was based on these findings. Depending on the test strain the lower limit of the dilution varies. All test strains were diluted from 10 CFU per 10 ml culture media to 0.625 CFU per 10 ml culture media. The suspensions were inoculated in the equal aerobic or anaerobic culture media and incubated at the appropriate temperature, respectively incubation condition. Evaluation was done at the time points indicated in Fig. 2. Relative accuracy and recovery rate Accuracy of the alternative method was determined based on the findings of the detection limit regarding number of CFU and incubation time. Experiments were carried out for at least eight times for each test strain to determine relative accuracy. In a next step the recovery rate of the reference method for positive results was determined. For this validation step each test strain was tested for a minimum of ten times at the limit of detection. Based on these findings we identified the final time point of the alternative method showing 100 % consensus compared with the reference method.
Limit of detection
Robustness
After determining the specificity the detection limit of the alternative method had to be investigated. The detection limit is defined as the lowest detectable number of microorganisms in a sample under the stated experimental conditions. The limit of detection refers to the number of microorganisms present in the original sample before any incubation steps. The level of inoculation must be adjusted until at least 50 %
Another parameter to define and evaluate was the robustness of this qualitative microbiological method. Therefore the incubation temperatures were changed (liquid soya-bean casein medium at 30 °C and liquid thioglycollate medium at room temperature). Testing for robustness was carried out in duplicates with a fixed inoculation concentration of 10 CFU per test unit.
123
Cell Tissue Bank (2014) 15:277–286
281
Table 2 Limit of detection in respect of incubation time p value
Limit of detection (LOD) Reference method
Alternative method
CFU per 10 ml
Incubation time (days)
CFU per 10 ml
Incubation time (days)
LOD50a
Staphylococcus aureus
1.250
14
1.250
4.0
1.448
0.662
Pseudomonas aeruginosa
2.500
14
2.500
2.75
2.993
0.496
Bacillus subtilis subs. spezizenii
1.250
14
1.250
1.625
1.749
0.573
Escherichia coli
0.625
14
0.625
7.0
0.692
0.923
Aspergillus brasiliensis
1.250
14
1.250
7.0
1.916
0.914
Candida albicans Clostridium sporogenes
1.250 2.500
14 14
1.250 2.500
2.75 1.625
1.677 2.841
0.199 0.415
Pseudomonas fluorescens
1.250
14
1.250
1.625
0.934
0.822
Streptococcus pyogenes
2.500
14
2.500
7.0
1.864
0.885
Propionibacterium acnes
1.250
14
1.250
7.0
0.625
0.580
Calculated two-tailed t test
a
Calculated with probit analysis for alternative method
Secondary validation For the last part of the study routine tissue products not suitable for transplantation due to quality control (QC) aspects were analysed. Placentas after caesarean section were collected with informed consent from donating mothers and stored in transport solution with antibiotics. Amniotic membrane was peeled off the placenta and washed three times in phosphate buffered saline. For sterility testing amnion biopsies with 1 cm were used in this validation. Medium used for cryoconservation (human albumin 50 %, DMSO 10 %, RPMI 40 %) of amnion grafts was also validated in this study. Bulbi for the preparation of corneal grafts were retrieved from multi organ and non-heart-beating donors. After disinfection with iodine dilution and rinsing with phosphate-buffered saline (three times) cornea processing was performed according to the Standard Operating Procedure. For sterility testing a swab of the iris was used afterwards. All tissue products were tested for HIV, HBV, HCV, HAV, TPHA and CMV. Collection and further processing were performed under GMP-conditions and according to the Austrian Tissue Safety Act, based on EC Directives 2004/23/EC, 2006/17/EC, and 2006/86/EC (Rommel et al. 2013; Dichtl et al. 2010). The first part of this second validation dealt with the fertility of the medium in presence of the product.
If at all, pharmaceutical products are contaminated at a low level and growth of the microorganism is necessary to obtain detection. Therefore it must be proven that the product does not inhibit the growth of microorganisms under the test conditions (European Directorate for the Quality of Medicines & HealthCare 2010). Inoculums are separately added at a maximum of 100 CFU for each tested microorganism into the portion of medium containing the product. For this purpose three independent runs with two different tissue product batches were used. The quantitative recovery of the microorganism should at least be 70 % (limit of detection 70 = LOD70). As mentioned the concentration should not be more than 100 CFU, the chosen primary dilution for the particular microorganism was the detection limit (Table 2). In case that the determined result did not reach the acceptance criterion an equal higher concentration of the test strain had to be tested. The second part of the final validation included the parallel analysis of the alternative method and the reference method in presence of the product. Concerning this point two different lots for each product were tested in triplicate. As acceptance criterion the results must be correlated with the reference method described in the European Pharmacopoeia. For this purpose the CFU concentration of the detection limit was used.
123
282
Additionally, negative controls as well as positive controls were carried out for each test strain and product. Sterility testing of the product for 14 days was considered as negative control. Each test strain was spiked in the appropriate bouillon in absence of the product to assure microbial growth. Statistical analysis Statistical analysis was conducted with the use of SPSS 20 software (SPSS Inc, Chicago IL, USA). Limit of detection was defined as the lowest inoculated number of microorganism with a growth rate of minimum 50 % and a minimum of 5 positive results. LOD50 (limit of detection 50 %) was calculated for the alternative method by means of probit analysis. LOD50 is determined as the concentration when a minimum of 50 % of the inoculated microorganisms show growth and therefore a positive result can be detected. Two-tailed t test was performed to detect the difference between the two methods regarding limit of detection (LOD). Accuracy of the alternative method was defined via relative accuracy. The relative accuracy is the degree of concordance between the response obtained by the alternative method and the reference method. Non inferiority tests (two-tailed fisher’s exacter test) were carried out to compare the alternative method (reduction of incubation time) to the reference method. P values of less than 0.05 were considered statistically significant.
Results Primary validation Specificity of the alternative method was demonstrated by incubation of the ten test strains depicted in Table 1. Aerobic microorganisms (bacteria and fungi) were inoculated in triplicates in liquid soya-bean casein medium at 20–25 °C and anaerobic bacteria were inoculated in triplicates in liquid thioglycollate medium at 30 °C. A positive result defined as growth promotion of all three test replicates for each microorganism was obtained. In the next step the limit of detection (LOD) was determined for the reference as well as the alternative method. LOD was calculated in two different ways: The results for the limit of detection were at least five
123
Cell Tissue Bank (2014) 15:277–286 Table 3 Relative accuracy of the alternative method compared with reference method p valueà
Test strain
Relative accuracy (%)a
Staphylococcus aureus
100.0
1.000
Pseudomonas aeruginosa
92.9
0.999
Bacillus subtilis subs. spezizenii
57.1
0.192
Escherichia coli
100.0
1.000
Aspergillus brasiliensis
100
1.000
Candida albicans
85.7
0.695
Clostridium sporogenes
85.7
1.000
Pseudomonas fluorescens
100.0
1.000
Streptococcus pyogenes
100.0
1.000
Propionibacterium acnes
100.0
1.000
à
Calculated with Fisher’s exacter test (two-tailed)
a
Relative accuracy (%) = (number of positive agreements ? number of negative agreements) * 100/ (number of positive agreements ? number of negative agreements ? number of positive deviation ? number of negative deviation) between alternative and reference method
replicates and also a minimum of 50 % of the replicates were positive are depicted in Table 2. The compared alternative and reference methods did not differ significantly (p value [0.05, Table 2). An important independent performance parameter for qualitative methods is LOD50, the level at which 50 % of the replicates are positive. The LOD50 for the alternative method was calculated from the results obtained by definition of limit of detection (Table 2). Relative accuracy was determined for the alternative method on the same terms evaluated by the limit of detection (Tables 2, 3) and comparison with the reference method was carried out. No statistically significant difference was found. For the reference method the recovery rate of positive results was determined on the basis of limit of detection (Fig. 3; Table 2). In nine of ten runs Streptococcus pyogenes showed a positive result. Bouillons inoculated with Pseudomonas aeruginosa, Candida albicans and Clostridium sporogenes revealed turbidity in 71 %, indicating a positive result. 70 %, respectively 69 % positive results were found for the microorganisms Pseudomonas fluorescens and Staphylococcus aureus. 8 of 13 replicates were positive for Aspergillus brasiliensis. Bacillus subtilis subs. spezizenii, Escherichia coli and Propionibacterium acnes showed the lowest recovery rate with 54, 57 and 56 %, respectively.
Cell Tissue Bank (2014) 15:277–286
Fig. 3 Reference method—recovery rate of positive results
The aim was to define the exact time point for the alternative method when at least the same frequency of recovery compared with the reference method was achieved. Therefore we assembled the frequency of positive results obtained of each visual evaluation as defined by time lapse (Fig. 2). Exactly the same rate of positive results was obtained for the alternative method compared with the reference method (results depicted in Fig. 3). The results obtained from this comparison revealed to be the definite timeframe for incubation (Fig. 4). The evaluation of robustness of the alternative method was proven by switching incubation temperature. Due to the results obtained from testing of relative accuracy and recovery rate incubation time for all test strains were scheduled for 8 days. Aerobic culture media were incubated at 30 °C in darkness and anaerobic culture media at 20–25 °C on light exposure. All test strains except P. fluorescens showed positive results. Secondary validation The second part of validation consisted of two phases. In phase one fertility of the medium in presence of
283
amniotic grafts, amnion cryomedium and swab of the iris was tested. The results for this validation step showed that higher concentrations as determined from the limit of detection had to be inoculated in the presence of the different products (Table 4). After definition of the appropriate inoculation concentration in presence of the product the recovery rate of positive results was examined. Therefore two lots of individual products were tested in triplicates. The recovery rates ranged between 83.3 and 100 % as displayed in Table 4. In the last step of our validation procedure we tested two different tissue preparation lots of the three above mentioned products in triplicates with the reference method and the alternative method. The only difference between the two methods was the period of incubation. 100 % consensus was found in this parallel validation. No growth could be detected in negative controls. All spiked controls showed a turbidity of the bouillon. No difference was found between the two tested microbiological methods.
Discussion and conclusions The need of some types of human tissue allografts for clinical applications increased steadily in the recent years whereas the number of donations was not rising equally e.g. shown with cardiovascular tissue (de By et al. 2012). In spite of implementation of GMP and GTP guidelines for tissue banking sterility of human allografts is essential for a positive clinical outcome. Monitoring of clean rooms at rest as well as in operation during processing of tissue products enhance the microbial safety of tissue products. Due to short-shelf life of some tissue products acceleration of microbiological testing, optimizing process
Fig. 4 Incubation time of the alternative method at same recovery rate (Fig. 3) of reference and alternative method based on the tested limit of detection (Table 2). Compared with the same accuracy obtained with the reference method
123
284
Cell Tissue Bank (2014) 15:277–286
Table 4 Required number of CFU per 10 ml culture media (LOD70) and recovery rate in the presence of the product Test strains
Amnion graft CFU per 10 ml
Cryomedium CFU per 10 ml
Iris-swab CFU per 10 ml
LOD70a
LOD70a
Recovery rate (%)b
LOD70
Recovery rate (%)b
Recovery rate (%)b
Staphylococcus aureus
1.25
100
2.50
100
1.25
83.3
Pseudomonas aeruginosa
2.50
100
2.50
100
2.50
100
10.00
100
1.25
100
2.50
83.3
Escherichia coli
1.25
100
1.25
100
1.25
83.3
Aspergillus brasiliensis
1.25
83.3
1.25
83.3
5.00
100
Candida albicans
2.50
83.3
2.50
100
1.25
83.3
Clostridium sporogenes
5.00
100
2.50
83.3
2.50
100
Pseudomonas fluorescens
5.00
100
10.00
83.3
1.25
100
Streptococcus pyogenes
2.50
83.3
2.50
83.3
2.50
83.3
Propionibacterium acnes
2.50
100
1.25
100
1.25
83.3
Bacillus subtilis subs. spezizenii
a
Limit of detection 70 %: dilution at which at least 70 % of the results were positive
b
Recovery rate (%): percentage of positive results in presence of the product
workflow and faster product release motivated us to perform the present validation study. As expected, results of specificity testing showed 100 % positive results with each test strain attributable to the used liquid culture media proposed in the European Pharmacopoeia. Six of these ten test strains are suggested in the Chapter 2.6.1 of the European Pharmacopoeia to be part of the microorganisms tested for growth promotion. Additionally P. acnes and E. coli were included in this study as these are common bacterial contaminants of amniotic membrane products (Adds et al. 2001). As previously reported, streptococcus can be one of the most frequent cornea contaminant (Farrell et al. 1991). Therefore we incorporated S. pyogenes as a representative of this family due to its recommendation in the European Pharmacopoeia (2.6.27) in our validation study. In Chapter 2.6.27 of the European Pharmacopoeia, describing the microbiological control of cellular products with the growth promotion test, limit of detection is achieved getting a positive result after inoculation of 10–100 viable microorganisms. For testing of the limit of detection for both methods we started our validation with 10 CFU and serial dilutions downwards due to partly low microbial burden of allografts. Same levels for limit of detection were found for all test strains within 7 days with a maximum number of 2.5 CFUs per 10 ml culture media. No statistically significant difference was
123
observed between the two methods regarding limit of detection. For C. sporogenes, P. fluorescens and Bacillus subtilis a period of incubation of 39 h was sufficient to obtain the same accuracy as the reference method. In 2007 Schmidt et al. have demonstrated that Bacillus spp has a mean time to positive signal between 24 and 55 h which confirms our finding in this study (Schmidt et al. 2007). P. acnes is a slow growing bacterium with a prolonged lag phase and in our study a detection limit of 7 days was defined. A comparable inoculated concentration for P. acnes was tested in platelet concentrates in a study resulting in 173 h (7 days and 4 h) incubation as time to earliest bacterial detection substantiate our observation (Nussbaumer et al. 2007). The overall incubation time has to be expanded to at least eight 8 days due to the accuracy finding for this microorganism. A recovery rate of less than 50 % recovery was found in the first 7 days of incubation. P. acnes showed the lowest recovery rate with 56 % on day 7 as well as on day 14. This phenomenon is likely due to the poor growth performance as mentioned above and confirmed by the work of Schmidt et al. Results of robustness testing of the alternative method indicated that a mixing-up of incubation temperature of aerobic and anaerobic culture media is not of great concern. Only P. fluorescens, a bacterium which pathogenicity for human is debatable and known to be tricky in regard to incubation
Cell Tissue Bank (2014) 15:277–286
temperature (Von and Weinstein 1971) showed no growth at changed incubation temperatures. In phase one of the secondary validation study the fertility of the media in presence of amniotic grafts, cryomedium or iris-swab was tested. The results indicated that presence of amniotic membrane caused a higher required inoculation concentration to get positive results in more than 70 % due to the known anti-microbial property of amniotic membrane. Amniotic membrane, respectively amniotic epithelial cells have the ability to produce ß-defensins and 2 lowmolecular-masse elastase inhibitors, namely secretory leukocyte proteinase inhibitor and elafin (King et al. 2007; Buhimschi et al. 2004). The bacterial growth in the presence of the product was excellent. A 100 % recovery showed P. aeruginosa with all three kinds of product. The other nine test strains had a recovery rate of 5/6 or 6/6 tests in the various tested materials. Sensitivity at the same level was achieved with a period of incubation of 8 days. Due to the findings we hypothesized that an incubation time of 8 days is not inferior to an incubation time of 14 days as required in the European Pharmacopoeia for the three tested product components. Mc Donald et al. tested 15 microorganisms (10 CFU mL-1) in thioglycollate broth in 2002. Bacterial growth was detected within a maximum of 4 days (McDonald et al. 2002). These findings were comparable with us as we used a lower CFU concentration in our study. The performed positive and negative controls imply that the obtained study results are valid and the reduction for the period of incubation is acceptable. Risk–benefit analysis has also been prepared within validation of alternative microbiological method according the European Pharmacopoeia. Based on the purpose the benefit and the risk of an alternative method were evaluated. Advantages of the proposed microbiological method are earlier product release date associated with lower product storage time again higher product quality and more products in stock for release. Optimization of sterility testing workflow can be achieved with a decreased period of incubation. Potential risk with implementation of the alternative method could be false-negative results based on the shorting of incubation. Based on our findings we evaluated that a reduction of the incubation time would not result in such a higher risk. All tested
285
bacteria and fungi were detected in the appropriate concentrations with the chosen period of incubation. In conclusion, we suggest a period of incubation for microbiological testing of amniotic grafts, cryomedium for amnion processing and iris-swab of 9 days owing to the findings of this study and taking into account the differences in time point in processing and visual evaluation for microbial growth. Despite this decrease in microbiological testing time further efforts are necessary to enhance the performance of sterility testing of tissue grafts and therefore to improve tissue safety and minimize adverse events due to bacterial contamination. Acknowledgments We would like to thank Anja PeterbauerScherb for her input during manuscript preparation.
References Adds PJ, Hunt C, Hartley S (2001) Bacterial contamination of amniotic membrane. Br J Ophthalmol 85:228–230 Aghayan HR, Goodarzi P, Baradaran-Rafii A, Larijani B, Moradabadi L, Rahim F, Arjmand B (2012) Bacterial contamination of amniotic membrane in a tissue bank from Iran. Cell Tissue Bank 14:401–406 Buhimschi IA, Jabr M, Buhimschi CS, Petkova AP, Weiner CP, Saed GM (2004) The novel antimicrobial peptide beta3defensin is produced by the amnion: a possible role of the fetal membranes in innate immunity of the amniotic cavity. Am J Obstet Gynecol 191:1678–1687 Davis JW (1910) Skin transplantation with a review of 550 cases at the Johns Hopkins Hospital. Johns Hopkins Med J 15:307 de By TM, Parker R, Delmo Walter EM, Hetzer R (2012) Cardiovascular tissue banking in Europe. HSR Proc Intensive Care Cardiovasc Anesth 4:251–260 Dichtl M, Gabriel C, Hennerbichler S, Seitz B, Priglinger SG (2010) EU konformes Hornhautbanking - eine Bestandsaufnahme Hornhautbank Linz. Spektrum der Augenheilkunde 24:166–173 Dua HS, Gomes JA, King AJ, Maharajan VS (2004) The amniotic membrane in ophthalmology. Surv Ophthalmol 49:51–77 Eastlund T (2006) Bacterial infection transmitted by human tissue allograft transplantation. Cell Tissue Bank 7:147–166 Eastlund T, Strong DM (2003) Infectious disease transmission through tissue transplantation. Adv Tissue Bank 7: 51–131. In: Phillips GO (ed). World Scientific Publishing Company, Singapore Edwards M, Clover GM, Brookes N, Pendergrast D, Chaulk J, McGhee CN (2002) Indications for corneal transplantation in New Zealand: 1991–1999. Cornea 21:152–155 European Directorate for the Quality of Medicines & HealthCare (2010) European Pharmacopoeia, 7th edn, vol 1
123
286 Farrell PL, Fan JT, Smith RE, Trousdale MD (1991) Donor cornea bacterial contamination. Cornea 10:381–386 King AE, Paltoo A, Kelly RW, Sallenave JM, Bocking AD, Challis JR (2007) Expression of natural antimicrobials by human placenta and fetal membranes. Placenta 28:161–169 Lee SH, Tseng SC (1997) Amniotic membrane transplantation for persistent epithelial defects with ulceration. Am J Ophthalmol 123:303–312 Marangon FB, Alfonso EC, Miller D, Remonda NM, Muallem MS, Tseng SC (2004) Incidence of microbial infection after amniotic membrane transplantation. Cornea 23:264–269 Matthews RN, Faulk WP, Bennett JP (1982) A review of the role of amniotic membranes in surgical practice. Obstet Gynecol Annu 11:31–58 McDonald CP, Rogers A, Cox M, Smith R, Roy A, Robbins S, Hartley S, Barbara JA, Rothenberg S, Stutzman L, Widders G (2002) Evaluation of the 3D BacT/ALERT automated culture system for the detection of microbial contamination of platelet concentrates. Transfus Med 12:303–309 Montag-Lessing T, Stormer M, Schurig U, Brachert J, Bubenzer M, Sicker U, Beshir R, Spreitzer I, Loschner B, Bache C, Becker B, Schneider CK (2010) Problems in microbial safety of advanced therapy medicinal products. Squaring the circle. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 53:45–51 Nussbaumer W, Allerstorfer D, Grabmer C, Rheinschmidt M, Lin L, Schonitzer D, Lass-Florl C (2007) Prevention of transfusion of platelet components contaminated with low
123
Cell Tissue Bank (2014) 15:277–286 levels of bacteria: a comparison of bacteria culture and pathogen inactivation methods. Transfusion 47:1125–1133 Pardos GJ, Gallagher MA (1982) Microbial contamination of donor eyes. A retrospective study. Arch Ophthalmol 100:1611–1613 Parveen S, Kaur S, David SA, Kenney JL, McCormick WM, Gupta RK (2011) Evaluation of growth based rapid microbiological methods for sterility testing of vaccines and other biological products. Vaccine 29:8012–8023 Rommel N, Rohleder NH, Gabriel C, Hennerbichler S, Bauer F, Mucke T, Kolk A, Loeffelbein DJ, Wolff KD, Kesting MR (2013) Secondary correction of posttraumatic orbital wall adhesions by membranes laminated with amniotic membrane. Br J Oral Maxillofac Surg 51:e224–e229 Schmidt M, Karakassopoulos A, Burkhart J, Deitenbeck R, Asmus J, Muller TH, Weinauer F, Seifried E, WaltherWenke G (2007) Comparison of three bacterial detection methods under routine conditions. Vox Sang 92:15–21 Schroeter J, Wilkemeyer I, Schiller RA, Pruss A (2012) Validation of the microbiological testing of tissue preparations using the BACTEC blood culture system. Transfus Med Hemother 39:387–390 Von GA, Weinstein J (1971) Pathogenic significance of Pseudomonas fluorescens and Pseudomonas putida. Yale J Biol Med 44:265–273 Wang S, Zinderman C, Wise R, Braun M (2007) Infections and human tissue transplants: review of FDA MedWatch reports 2001–2004. Cell Tissue Bank 8:211–219
ORIGINAL PAPER
Validation of an alternative microbiological method for tissue products Susanne Suessner • Simone Hennerbichler Stefanie Schreiberhuber • Doris Stuebl • Christian Gabriel
•
Received: 1 August 2013 / Accepted: 30 April 2014 / Published online: 9 May 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract According to the European Pharmacopoeia sterility testing of products includes an incubation time of 14 days in thioglycollate medium and soya-bean casein medium. In this case a large period of time is needed for product testing. So we designed a study to evaluate an alternative method for sterility testing. The aim of this study was to reduce the incubation time for the routinely produced products in our tissue bank (cornea and amnion grafts) by obtaining the same detection limit, accurateness and recovery rates as the reference method described in the European Pharmacopoeia. The study included two steps of validation. Primary validation compared the reference method with the alternative method. Therefore eight bacterial and two fungi test strains were tested at their preferred milieu. A geometric dilution series from 10 to 0.625 colony forming unit per 10 ml culture media was used. Subsequent to the evaluation the second part of the study started including the validation of the fertility of the culture media and the parallel testing of the two methods by investigating products. For this purpose two product batches were tested in three independent runs. Concerning the validation we could not find any aberration between
S. Suessner (&) S. Hennerbichler S. Schreiberhuber D. Stuebl C. Gabriel Austrian Cluster for Tissue Regeneration, Red Cross Blood Transfusion Service of Upper Austria, Krankenhausstraße 7, 4017 Linz, Austria e-mail: [email protected]
the alternative and the reference method. In addition, the recovery rate of each microorganism was between 83.33 and 100 %. The alternative method showed noninferiority regarding accuracy to the reference method. Due to this study we reduced the sterility testing for cornea and amniotic grafts to 9 days. Keywords Sterility testing Validation alternative method Cornea Amniotic membrane Quality control Tissue bank
Introduction Sterility testing is an essential issue in the processing and release of tissue products. Sterility of primary materials, application of established procedures for sterilization as well as the short shelf life of many tissue grafts are of great concern in the release process (Montag-Lessing et al. 2010). Due to reports of tissuetransmitted infections implementation of more federal regulations have led to an increased safety of tissue transplants (Wang et al. 2007). Associations of bacterial infections with allografts have involved bones, tendons, dura mater, corneas, heart valves, skin and pericardium (Eastlund and Strong 2003). Several steps had been implemented reducing microbial contamination: criteria for donor selection, minimization of interval between death and
123
278
tissue removal, monitoring and control of microbial contamination of air and surfaces, use of aseptic surgical technique during dissection and removal of tissues (Eastlund 2006). Since 1910 amniotic membrane has been used in various clinical applications, reconstructive surgery, in therapy of burns and chronic ulcers (Matthews et al. 1982; Davis 1910). The transplantation of amnion grafts plays an important role in the treatment of ophthalmic disorders such as persistent corneal epithelial defects with ulceration or corneal burns (Lee and Tseng 1997; Dua et al. 2004). The source material for amniotic grafts may not always be sterile (Aghayan et al. 2012) and bacterial contamination is dependent on the way placental membranes are retrieved. Placental membranes obtained by caesarean sections have less burden of pathogenic organism than such of normal vaginal deliveries (Adds et al. 2001). Gram-positive bacteria are the most frequent microbial causes for infections of amniotic membrane transplants in ocular surgery (Marangon et al. 2004). Other important tissue products used in ophthalmologic surgery are corneal grafts. The main indications for keratoplasty combined with corneal transplantation are keratoconus, pseudophakic or aphakic corneal edema (Edwards et al. 2002). Corneal grafts are cultivated in culture media containing antibiotics to minimize the risk to transplant bacterial contaminated products (Dichtl et al. 2010). The main bacterial contaminants of donor corneas are skin germs, primarily Staphylococcus epidermidis (Pardos and Gallagher 1982). The European Pharmacopoeia characterizes the regulations for sterility testing (European Directorate for the Quality of Medicines & HealthCare 2010). The appropriate culture medium for detection of anaerobic bacteria is liquid thioglycollate medium, for aerobic bacteria and fungi soya-bean casein digest medium is recommended. The principle of these growth-based methods is a change in turbidity indicating the presence of viable micro-organisms. The incubation time is defined to be 14 days. Given to a relative short shelf life of some tissue grafts it is too long, meaning that cultivated corneal transplants with a shelf life of 28 days can only be transplanted 14 days upon receipt of negative microbial testing result. A faster microbial method for sterility testing increases the available pool of negative-tested tissue products and therefore
123
Cell Tissue Bank (2014) 15:277–286
improvement in terms of adequate and assured product supply can be guaranteed. In recent years several approaches were proven as a rapid alternative method to the common growth-based methods (Parveen et al. 2011). A possible option are automated blood culture systems which are suitable for microbiological testing of tissue preparations (Schroeter et al. 2012). The aim of the present study was the validation of an alternative qualitative microbiological testing method regarding shortened incubation time compared with the reference method of the European Pharmacopoeia, from which validation criteria for such alternative methods of microbiological quality control were used. The concept of our alternative method was only a change in terms of incubation period of microbiological testing while maintaining the described methods.
Materials and methods Study design As the aim of the alternative method was to reduce the incubation time for sterility testing, the liquid culture media were read out in an exact time interval to get a positive or negative result. For the comparison of the two methods the study was performed in two steps (Fig. 1). At first a comparison between the reference method and the alternative method was performed to define the specificity, limit of detection, relative accuracy, recovery rate and robustness of the alternative method. Subsequently the second validation step was done, validating the routinely processed tissue preparations (cornea, amniotic grafts and cryomedium) and comparing them with the obtained results of the first step of this study. Therefore two different lots of each tissue preparation were validated in three independent runs. A main issue of the first part of this validation was the visual assessment of the liquid culture media. Exact points of time for direct reading had to be defined (Fig. 2). All bouillons were mixed thoroughly on a vortex before evaluation. After the first runs it became evident that no positive results were observed after 18 and 24 h. Hence, the first reading was done at 39 h in the following runs.
Cell Tissue Bank (2014) 15:277–286
279
Fig. 1 Study design for validation of the alternative microbiological method for sterility testing
Fig. 2 Time lapse for evaluation of liquid culture media. h hours, d day
18 h
39 h
24h
66 h
48h
d5
d4
d7
d6
d9
d8
d11
d10
d13
d12
d14
h = hours; d = day
Test strains and methodology Ten different American type culture collection (ATCC) test strains were chosen to be used in this validation study representing a wide range of microorganisms (Table 1). Six of these ten test strains were chosen according to the European Pharmacopoeia Chapter 2.6.1. The test strains were reconstituted according to the manufacturer’s instructions. Then the required dilution could be produced. Each test strain was incubated in its preferred milieu (Table 1). Aerobic microorganisms were inoculated in liquid soya-bean casein medium (Biomerieux, France) and incubated at room temperature (20–25 °C). The anaerobic test strains were incubated at 30 °C in liquid thioglycollate medium (Biomerieux, France) due to another validation (unpublished data).
Turbidity of liquid culture medium indicated microbial growth. No visible changes yielded in a negative result. To assure negative result agar cultivation was performed. After the 14-day incubation trypcase-soya agar (Biomerieux, France) for the aerobic microorganisms and a Schaedler agar (Biomerieux, France) for the anaerobic test strains were plated and cultivated. The trypcase-soya agar plates were incubated for 2 days at 30 °C and Schaedler agar plates for 7 days at 37 °C in anaerobic atmosphere.
Primary validation The first part was the validation of the alternative method in relation to the reference method including
123
280
Cell Tissue Bank (2014) 15:277–286
Table 1 Test strains and incubation conditions Test strains
Strain ID
Incubation condition
Staphylococcus aureus subsp. aureus
ATCC 6538a
Aerobic
Bacillus subtilis subs. spezizenii
ATCC 6633a
Aerobic
Pseudomonas aeruginosa
ATCC 9027a
Aerobic a
Clostridium sporogenes
ATCC 11437
Candida albicans
ATCC 10231a
Aerobic
Aspergillus brasiliensis
ATCC 16404a
Aerobic
Escherichia coli
ATCC 8739a
Aerobic
Pseudomonas fluorescens
ATCC 49838b
Aerobic
Streptococcus pyogenes
ATCC19615c
Aerobic
Propionibacterium acnes a
ATCC 11827
b
Anaerobic
Anaerobic
Bioball
b
Epower microorganism
c
EZ CFU one step-microorganism
testing of specificity, limit of detection, relative accuracy, recovery rate and robustness. Specificity At first the specificity of the alternative method had to be determined. According to the guidelines (European Directorate for the Quality of Medicines & HealthCare 2010) the specificity of an alternative qualitative method is the ability to detect the required range of microorganisms that may be present in the tested sample. For this purpose three runs with ten test strains were performed. Each test strain was inoculated at a concentration of 10 colony forming units (CFU) in 10 ml culture media. All applied microorganisms had to be detected, thus implying a visual recognizable turbidity in 100 % of the bouillons.
show reproducible growth in the chosen microbiological method. To get enough comparable data the test strains were inoculated in each dilution for at least nine times. For this step a geometric dilution series of each microorganism was prepared. Therefore in preliminary tests we have cultivated the various dilutions in triplicates on agar plates (trypcase-soya agar plates for aerobes and fungi and Schaedler agar plates for anaerobes). The calculated average of the obtained CFUs acted as control of the inoculated microorganisms for the dilution series (data not shown). The primary validation was based on these findings. Depending on the test strain the lower limit of the dilution varies. All test strains were diluted from 10 CFU per 10 ml culture media to 0.625 CFU per 10 ml culture media. The suspensions were inoculated in the equal aerobic or anaerobic culture media and incubated at the appropriate temperature, respectively incubation condition. Evaluation was done at the time points indicated in Fig. 2. Relative accuracy and recovery rate Accuracy of the alternative method was determined based on the findings of the detection limit regarding number of CFU and incubation time. Experiments were carried out for at least eight times for each test strain to determine relative accuracy. In a next step the recovery rate of the reference method for positive results was determined. For this validation step each test strain was tested for a minimum of ten times at the limit of detection. Based on these findings we identified the final time point of the alternative method showing 100 % consensus compared with the reference method.
Limit of detection
Robustness
After determining the specificity the detection limit of the alternative method had to be investigated. The detection limit is defined as the lowest detectable number of microorganisms in a sample under the stated experimental conditions. The limit of detection refers to the number of microorganisms present in the original sample before any incubation steps. The level of inoculation must be adjusted until at least 50 %
Another parameter to define and evaluate was the robustness of this qualitative microbiological method. Therefore the incubation temperatures were changed (liquid soya-bean casein medium at 30 °C and liquid thioglycollate medium at room temperature). Testing for robustness was carried out in duplicates with a fixed inoculation concentration of 10 CFU per test unit.
123
Cell Tissue Bank (2014) 15:277–286
281
Table 2 Limit of detection in respect of incubation time p value
Limit of detection (LOD) Reference method
Alternative method
CFU per 10 ml
Incubation time (days)
CFU per 10 ml
Incubation time (days)
LOD50a
Staphylococcus aureus
1.250
14
1.250
4.0
1.448
0.662
Pseudomonas aeruginosa
2.500
14
2.500
2.75
2.993
0.496
Bacillus subtilis subs. spezizenii
1.250
14
1.250
1.625
1.749
0.573
Escherichia coli
0.625
14
0.625
7.0
0.692
0.923
Aspergillus brasiliensis
1.250
14
1.250
7.0
1.916
0.914
Candida albicans Clostridium sporogenes
1.250 2.500
14 14
1.250 2.500
2.75 1.625
1.677 2.841
0.199 0.415
Pseudomonas fluorescens
1.250
14
1.250
1.625
0.934
0.822
Streptococcus pyogenes
2.500
14
2.500
7.0
1.864
0.885
Propionibacterium acnes
1.250
14
1.250
7.0
0.625
0.580
Calculated two-tailed t test
a
Calculated with probit analysis for alternative method
Secondary validation For the last part of the study routine tissue products not suitable for transplantation due to quality control (QC) aspects were analysed. Placentas after caesarean section were collected with informed consent from donating mothers and stored in transport solution with antibiotics. Amniotic membrane was peeled off the placenta and washed three times in phosphate buffered saline. For sterility testing amnion biopsies with 1 cm were used in this validation. Medium used for cryoconservation (human albumin 50 %, DMSO 10 %, RPMI 40 %) of amnion grafts was also validated in this study. Bulbi for the preparation of corneal grafts were retrieved from multi organ and non-heart-beating donors. After disinfection with iodine dilution and rinsing with phosphate-buffered saline (three times) cornea processing was performed according to the Standard Operating Procedure. For sterility testing a swab of the iris was used afterwards. All tissue products were tested for HIV, HBV, HCV, HAV, TPHA and CMV. Collection and further processing were performed under GMP-conditions and according to the Austrian Tissue Safety Act, based on EC Directives 2004/23/EC, 2006/17/EC, and 2006/86/EC (Rommel et al. 2013; Dichtl et al. 2010). The first part of this second validation dealt with the fertility of the medium in presence of the product.
If at all, pharmaceutical products are contaminated at a low level and growth of the microorganism is necessary to obtain detection. Therefore it must be proven that the product does not inhibit the growth of microorganisms under the test conditions (European Directorate for the Quality of Medicines & HealthCare 2010). Inoculums are separately added at a maximum of 100 CFU for each tested microorganism into the portion of medium containing the product. For this purpose three independent runs with two different tissue product batches were used. The quantitative recovery of the microorganism should at least be 70 % (limit of detection 70 = LOD70). As mentioned the concentration should not be more than 100 CFU, the chosen primary dilution for the particular microorganism was the detection limit (Table 2). In case that the determined result did not reach the acceptance criterion an equal higher concentration of the test strain had to be tested. The second part of the final validation included the parallel analysis of the alternative method and the reference method in presence of the product. Concerning this point two different lots for each product were tested in triplicate. As acceptance criterion the results must be correlated with the reference method described in the European Pharmacopoeia. For this purpose the CFU concentration of the detection limit was used.
123
282
Additionally, negative controls as well as positive controls were carried out for each test strain and product. Sterility testing of the product for 14 days was considered as negative control. Each test strain was spiked in the appropriate bouillon in absence of the product to assure microbial growth. Statistical analysis Statistical analysis was conducted with the use of SPSS 20 software (SPSS Inc, Chicago IL, USA). Limit of detection was defined as the lowest inoculated number of microorganism with a growth rate of minimum 50 % and a minimum of 5 positive results. LOD50 (limit of detection 50 %) was calculated for the alternative method by means of probit analysis. LOD50 is determined as the concentration when a minimum of 50 % of the inoculated microorganisms show growth and therefore a positive result can be detected. Two-tailed t test was performed to detect the difference between the two methods regarding limit of detection (LOD). Accuracy of the alternative method was defined via relative accuracy. The relative accuracy is the degree of concordance between the response obtained by the alternative method and the reference method. Non inferiority tests (two-tailed fisher’s exacter test) were carried out to compare the alternative method (reduction of incubation time) to the reference method. P values of less than 0.05 were considered statistically significant.
Results Primary validation Specificity of the alternative method was demonstrated by incubation of the ten test strains depicted in Table 1. Aerobic microorganisms (bacteria and fungi) were inoculated in triplicates in liquid soya-bean casein medium at 20–25 °C and anaerobic bacteria were inoculated in triplicates in liquid thioglycollate medium at 30 °C. A positive result defined as growth promotion of all three test replicates for each microorganism was obtained. In the next step the limit of detection (LOD) was determined for the reference as well as the alternative method. LOD was calculated in two different ways: The results for the limit of detection were at least five
123
Cell Tissue Bank (2014) 15:277–286 Table 3 Relative accuracy of the alternative method compared with reference method p valueà
Test strain
Relative accuracy (%)a
Staphylococcus aureus
100.0
1.000
Pseudomonas aeruginosa
92.9
0.999
Bacillus subtilis subs. spezizenii
57.1
0.192
Escherichia coli
100.0
1.000
Aspergillus brasiliensis
100
1.000
Candida albicans
85.7
0.695
Clostridium sporogenes
85.7
1.000
Pseudomonas fluorescens
100.0
1.000
Streptococcus pyogenes
100.0
1.000
Propionibacterium acnes
100.0
1.000
à
Calculated with Fisher’s exacter test (two-tailed)
a
Relative accuracy (%) = (number of positive agreements ? number of negative agreements) * 100/ (number of positive agreements ? number of negative agreements ? number of positive deviation ? number of negative deviation) between alternative and reference method
replicates and also a minimum of 50 % of the replicates were positive are depicted in Table 2. The compared alternative and reference methods did not differ significantly (p value [0.05, Table 2). An important independent performance parameter for qualitative methods is LOD50, the level at which 50 % of the replicates are positive. The LOD50 for the alternative method was calculated from the results obtained by definition of limit of detection (Table 2). Relative accuracy was determined for the alternative method on the same terms evaluated by the limit of detection (Tables 2, 3) and comparison with the reference method was carried out. No statistically significant difference was found. For the reference method the recovery rate of positive results was determined on the basis of limit of detection (Fig. 3; Table 2). In nine of ten runs Streptococcus pyogenes showed a positive result. Bouillons inoculated with Pseudomonas aeruginosa, Candida albicans and Clostridium sporogenes revealed turbidity in 71 %, indicating a positive result. 70 %, respectively 69 % positive results were found for the microorganisms Pseudomonas fluorescens and Staphylococcus aureus. 8 of 13 replicates were positive for Aspergillus brasiliensis. Bacillus subtilis subs. spezizenii, Escherichia coli and Propionibacterium acnes showed the lowest recovery rate with 54, 57 and 56 %, respectively.
Cell Tissue Bank (2014) 15:277–286
Fig. 3 Reference method—recovery rate of positive results
The aim was to define the exact time point for the alternative method when at least the same frequency of recovery compared with the reference method was achieved. Therefore we assembled the frequency of positive results obtained of each visual evaluation as defined by time lapse (Fig. 2). Exactly the same rate of positive results was obtained for the alternative method compared with the reference method (results depicted in Fig. 3). The results obtained from this comparison revealed to be the definite timeframe for incubation (Fig. 4). The evaluation of robustness of the alternative method was proven by switching incubation temperature. Due to the results obtained from testing of relative accuracy and recovery rate incubation time for all test strains were scheduled for 8 days. Aerobic culture media were incubated at 30 °C in darkness and anaerobic culture media at 20–25 °C on light exposure. All test strains except P. fluorescens showed positive results. Secondary validation The second part of validation consisted of two phases. In phase one fertility of the medium in presence of
283
amniotic grafts, amnion cryomedium and swab of the iris was tested. The results for this validation step showed that higher concentrations as determined from the limit of detection had to be inoculated in the presence of the different products (Table 4). After definition of the appropriate inoculation concentration in presence of the product the recovery rate of positive results was examined. Therefore two lots of individual products were tested in triplicates. The recovery rates ranged between 83.3 and 100 % as displayed in Table 4. In the last step of our validation procedure we tested two different tissue preparation lots of the three above mentioned products in triplicates with the reference method and the alternative method. The only difference between the two methods was the period of incubation. 100 % consensus was found in this parallel validation. No growth could be detected in negative controls. All spiked controls showed a turbidity of the bouillon. No difference was found between the two tested microbiological methods.
Discussion and conclusions The need of some types of human tissue allografts for clinical applications increased steadily in the recent years whereas the number of donations was not rising equally e.g. shown with cardiovascular tissue (de By et al. 2012). In spite of implementation of GMP and GTP guidelines for tissue banking sterility of human allografts is essential for a positive clinical outcome. Monitoring of clean rooms at rest as well as in operation during processing of tissue products enhance the microbial safety of tissue products. Due to short-shelf life of some tissue products acceleration of microbiological testing, optimizing process
Fig. 4 Incubation time of the alternative method at same recovery rate (Fig. 3) of reference and alternative method based on the tested limit of detection (Table 2). Compared with the same accuracy obtained with the reference method
123
284
Cell Tissue Bank (2014) 15:277–286
Table 4 Required number of CFU per 10 ml culture media (LOD70) and recovery rate in the presence of the product Test strains
Amnion graft CFU per 10 ml
Cryomedium CFU per 10 ml
Iris-swab CFU per 10 ml
LOD70a
LOD70a
Recovery rate (%)b
LOD70
Recovery rate (%)b
Recovery rate (%)b
Staphylococcus aureus
1.25
100
2.50
100
1.25
83.3
Pseudomonas aeruginosa
2.50
100
2.50
100
2.50
100
10.00
100
1.25
100
2.50
83.3
Escherichia coli
1.25
100
1.25
100
1.25
83.3
Aspergillus brasiliensis
1.25
83.3
1.25
83.3
5.00
100
Candida albicans
2.50
83.3
2.50
100
1.25
83.3
Clostridium sporogenes
5.00
100
2.50
83.3
2.50
100
Pseudomonas fluorescens
5.00
100
10.00
83.3
1.25
100
Streptococcus pyogenes
2.50
83.3
2.50
83.3
2.50
83.3
Propionibacterium acnes
2.50
100
1.25
100
1.25
83.3
Bacillus subtilis subs. spezizenii
a
Limit of detection 70 %: dilution at which at least 70 % of the results were positive
b
Recovery rate (%): percentage of positive results in presence of the product
workflow and faster product release motivated us to perform the present validation study. As expected, results of specificity testing showed 100 % positive results with each test strain attributable to the used liquid culture media proposed in the European Pharmacopoeia. Six of these ten test strains are suggested in the Chapter 2.6.1 of the European Pharmacopoeia to be part of the microorganisms tested for growth promotion. Additionally P. acnes and E. coli were included in this study as these are common bacterial contaminants of amniotic membrane products (Adds et al. 2001). As previously reported, streptococcus can be one of the most frequent cornea contaminant (Farrell et al. 1991). Therefore we incorporated S. pyogenes as a representative of this family due to its recommendation in the European Pharmacopoeia (2.6.27) in our validation study. In Chapter 2.6.27 of the European Pharmacopoeia, describing the microbiological control of cellular products with the growth promotion test, limit of detection is achieved getting a positive result after inoculation of 10–100 viable microorganisms. For testing of the limit of detection for both methods we started our validation with 10 CFU and serial dilutions downwards due to partly low microbial burden of allografts. Same levels for limit of detection were found for all test strains within 7 days with a maximum number of 2.5 CFUs per 10 ml culture media. No statistically significant difference was
123
observed between the two methods regarding limit of detection. For C. sporogenes, P. fluorescens and Bacillus subtilis a period of incubation of 39 h was sufficient to obtain the same accuracy as the reference method. In 2007 Schmidt et al. have demonstrated that Bacillus spp has a mean time to positive signal between 24 and 55 h which confirms our finding in this study (Schmidt et al. 2007). P. acnes is a slow growing bacterium with a prolonged lag phase and in our study a detection limit of 7 days was defined. A comparable inoculated concentration for P. acnes was tested in platelet concentrates in a study resulting in 173 h (7 days and 4 h) incubation as time to earliest bacterial detection substantiate our observation (Nussbaumer et al. 2007). The overall incubation time has to be expanded to at least eight 8 days due to the accuracy finding for this microorganism. A recovery rate of less than 50 % recovery was found in the first 7 days of incubation. P. acnes showed the lowest recovery rate with 56 % on day 7 as well as on day 14. This phenomenon is likely due to the poor growth performance as mentioned above and confirmed by the work of Schmidt et al. Results of robustness testing of the alternative method indicated that a mixing-up of incubation temperature of aerobic and anaerobic culture media is not of great concern. Only P. fluorescens, a bacterium which pathogenicity for human is debatable and known to be tricky in regard to incubation
Cell Tissue Bank (2014) 15:277–286
temperature (Von and Weinstein 1971) showed no growth at changed incubation temperatures. In phase one of the secondary validation study the fertility of the media in presence of amniotic grafts, cryomedium or iris-swab was tested. The results indicated that presence of amniotic membrane caused a higher required inoculation concentration to get positive results in more than 70 % due to the known anti-microbial property of amniotic membrane. Amniotic membrane, respectively amniotic epithelial cells have the ability to produce ß-defensins and 2 lowmolecular-masse elastase inhibitors, namely secretory leukocyte proteinase inhibitor and elafin (King et al. 2007; Buhimschi et al. 2004). The bacterial growth in the presence of the product was excellent. A 100 % recovery showed P. aeruginosa with all three kinds of product. The other nine test strains had a recovery rate of 5/6 or 6/6 tests in the various tested materials. Sensitivity at the same level was achieved with a period of incubation of 8 days. Due to the findings we hypothesized that an incubation time of 8 days is not inferior to an incubation time of 14 days as required in the European Pharmacopoeia for the three tested product components. Mc Donald et al. tested 15 microorganisms (10 CFU mL-1) in thioglycollate broth in 2002. Bacterial growth was detected within a maximum of 4 days (McDonald et al. 2002). These findings were comparable with us as we used a lower CFU concentration in our study. The performed positive and negative controls imply that the obtained study results are valid and the reduction for the period of incubation is acceptable. Risk–benefit analysis has also been prepared within validation of alternative microbiological method according the European Pharmacopoeia. Based on the purpose the benefit and the risk of an alternative method were evaluated. Advantages of the proposed microbiological method are earlier product release date associated with lower product storage time again higher product quality and more products in stock for release. Optimization of sterility testing workflow can be achieved with a decreased period of incubation. Potential risk with implementation of the alternative method could be false-negative results based on the shorting of incubation. Based on our findings we evaluated that a reduction of the incubation time would not result in such a higher risk. All tested
285
bacteria and fungi were detected in the appropriate concentrations with the chosen period of incubation. In conclusion, we suggest a period of incubation for microbiological testing of amniotic grafts, cryomedium for amnion processing and iris-swab of 9 days owing to the findings of this study and taking into account the differences in time point in processing and visual evaluation for microbial growth. Despite this decrease in microbiological testing time further efforts are necessary to enhance the performance of sterility testing of tissue grafts and therefore to improve tissue safety and minimize adverse events due to bacterial contamination. Acknowledgments We would like to thank Anja PeterbauerScherb for her input during manuscript preparation.
References Adds PJ, Hunt C, Hartley S (2001) Bacterial contamination of amniotic membrane. Br J Ophthalmol 85:228–230 Aghayan HR, Goodarzi P, Baradaran-Rafii A, Larijani B, Moradabadi L, Rahim F, Arjmand B (2012) Bacterial contamination of amniotic membrane in a tissue bank from Iran. Cell Tissue Bank 14:401–406 Buhimschi IA, Jabr M, Buhimschi CS, Petkova AP, Weiner CP, Saed GM (2004) The novel antimicrobial peptide beta3defensin is produced by the amnion: a possible role of the fetal membranes in innate immunity of the amniotic cavity. Am J Obstet Gynecol 191:1678–1687 Davis JW (1910) Skin transplantation with a review of 550 cases at the Johns Hopkins Hospital. Johns Hopkins Med J 15:307 de By TM, Parker R, Delmo Walter EM, Hetzer R (2012) Cardiovascular tissue banking in Europe. HSR Proc Intensive Care Cardiovasc Anesth 4:251–260 Dichtl M, Gabriel C, Hennerbichler S, Seitz B, Priglinger SG (2010) EU konformes Hornhautbanking - eine Bestandsaufnahme Hornhautbank Linz. Spektrum der Augenheilkunde 24:166–173 Dua HS, Gomes JA, King AJ, Maharajan VS (2004) The amniotic membrane in ophthalmology. Surv Ophthalmol 49:51–77 Eastlund T (2006) Bacterial infection transmitted by human tissue allograft transplantation. Cell Tissue Bank 7:147–166 Eastlund T, Strong DM (2003) Infectious disease transmission through tissue transplantation. Adv Tissue Bank 7: 51–131. In: Phillips GO (ed). World Scientific Publishing Company, Singapore Edwards M, Clover GM, Brookes N, Pendergrast D, Chaulk J, McGhee CN (2002) Indications for corneal transplantation in New Zealand: 1991–1999. Cornea 21:152–155 European Directorate for the Quality of Medicines & HealthCare (2010) European Pharmacopoeia, 7th edn, vol 1
123
286 Farrell PL, Fan JT, Smith RE, Trousdale MD (1991) Donor cornea bacterial contamination. Cornea 10:381–386 King AE, Paltoo A, Kelly RW, Sallenave JM, Bocking AD, Challis JR (2007) Expression of natural antimicrobials by human placenta and fetal membranes. Placenta 28:161–169 Lee SH, Tseng SC (1997) Amniotic membrane transplantation for persistent epithelial defects with ulceration. Am J Ophthalmol 123:303–312 Marangon FB, Alfonso EC, Miller D, Remonda NM, Muallem MS, Tseng SC (2004) Incidence of microbial infection after amniotic membrane transplantation. Cornea 23:264–269 Matthews RN, Faulk WP, Bennett JP (1982) A review of the role of amniotic membranes in surgical practice. Obstet Gynecol Annu 11:31–58 McDonald CP, Rogers A, Cox M, Smith R, Roy A, Robbins S, Hartley S, Barbara JA, Rothenberg S, Stutzman L, Widders G (2002) Evaluation of the 3D BacT/ALERT automated culture system for the detection of microbial contamination of platelet concentrates. Transfus Med 12:303–309 Montag-Lessing T, Stormer M, Schurig U, Brachert J, Bubenzer M, Sicker U, Beshir R, Spreitzer I, Loschner B, Bache C, Becker B, Schneider CK (2010) Problems in microbial safety of advanced therapy medicinal products. Squaring the circle. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 53:45–51 Nussbaumer W, Allerstorfer D, Grabmer C, Rheinschmidt M, Lin L, Schonitzer D, Lass-Florl C (2007) Prevention of transfusion of platelet components contaminated with low
123
Cell Tissue Bank (2014) 15:277–286 levels of bacteria: a comparison of bacteria culture and pathogen inactivation methods. Transfusion 47:1125–1133 Pardos GJ, Gallagher MA (1982) Microbial contamination of donor eyes. A retrospective study. Arch Ophthalmol 100:1611–1613 Parveen S, Kaur S, David SA, Kenney JL, McCormick WM, Gupta RK (2011) Evaluation of growth based rapid microbiological methods for sterility testing of vaccines and other biological products. Vaccine 29:8012–8023 Rommel N, Rohleder NH, Gabriel C, Hennerbichler S, Bauer F, Mucke T, Kolk A, Loeffelbein DJ, Wolff KD, Kesting MR (2013) Secondary correction of posttraumatic orbital wall adhesions by membranes laminated with amniotic membrane. Br J Oral Maxillofac Surg 51:e224–e229 Schmidt M, Karakassopoulos A, Burkhart J, Deitenbeck R, Asmus J, Muller TH, Weinauer F, Seifried E, WaltherWenke G (2007) Comparison of three bacterial detection methods under routine conditions. Vox Sang 92:15–21 Schroeter J, Wilkemeyer I, Schiller RA, Pruss A (2012) Validation of the microbiological testing of tissue preparations using the BACTEC blood culture system. Transfus Med Hemother 39:387–390 Von GA, Weinstein J (1971) Pathogenic significance of Pseudomonas fluorescens and Pseudomonas putida. Yale J Biol Med 44:265–273 Wang S, Zinderman C, Wise R, Braun M (2007) Infections and human tissue transplants: review of FDA MedWatch reports 2001–2004. Cell Tissue Bank 8:211–219
