Validation

In addition to the minor aspects of validation, a bioburden validation programme has principally to demonstrate that the procedures routinely used are capable of:

• Accurately and reproducibly enumerating low concentrations of organisms contained in, or, in the case of devices, on, the surfaces of samples or products

• Detecting low levels of specific objectionable organisms in products

• Adequately neutralizing any antimicrobial activity associated with the product and that any chemical inactivator employed is not itself toxic

The principle of these validation procedures is simply that the sample is inoculated with a known number of challenge organisms, and subjected to the routine method for TVC or detection of objectionable organisms. The sample is considered to possess no antimicrobial activity and the process is validated if a minimum designated proportion of the inoculum is recovered, or, in the case of qualitative testing, the objectionable organism is detected. The detailed methods are covered in the EP (2.6.12 and 2.6.13), Effectiveness of culture media and validity of counting method and Nutritive and selective properties of the media and validity of the test respectively. The USP describes validation in <61>, Preparatory Testing, and in <1227>, Validation of Microbial Recovery from Pharmacopeial Articles.

Obtaining reliable TVC values from medical devices, complex formulations or sparingly soluble raw materials is much more difficult than demonstrating accuracy and reproducibility in counting simple aqueous suspensions of pure cultures. It is clearly desirable that an operator is capable of demonstrating the latter skill before attempting the former. What constitutes an adequate level of competence in debatable, but Sandle* has described a simple statistical analysis of pour plate data that can be used for technician qualification. An alternative "rule of thumb" is that replicate dilution and plating of a bacterial (but not necessarily a fungal) suspension, should yield a coefficient of variation of <10% in the resulting colony counts.

The number of colonies counted on a plate influences the accuracy of the recorded result, and the pharmacopoeias indicate that between 30 and 300 colonies on a standard 9-cm plate is appropriate for most bacteria and Candida albicans. Since this range is not optimal for all environmental isolates, however, or for many yeasts and molds or for counts on 47-mm diameter membrane filters, it is necessary to validate the countable range. A procedure for doing so is in USP <1227>.

Routine challenge organisms to be used for validation are described in the pharmacopoeias, although it would be appropriate to use additional or alternative organisms that

• Are regularly isolated during environmental monitoring

• Regularly constitute a significant fraction of the raw material or finished product bioburden

• Are critical to the process with which the bioburden sample is associated

Pure culture rather than mixed culture inocula are recommended by the USP for each of the organisms selected, but the EP method requires a mixed inoculum.

The compendial recommendations are inconsistent on the number or concentrations of organisms to be inoculated into the product or the percentage recovery values considered acceptable. The EP directs that suspensions should contain about 100 CFU/ml, while the PDA recommend both a low level (<100) and a high level (103-104) inoculum. The USP requires 1 ml of a 1000-fold dilution of an overnight culture to be added to the first dilution of the product (100 ml). The concentration of an overnight culture of many bacteria is approximately 109 CFU/ml, so the USP requirement corresponds to a final inoculum concentration of approximately 104 CFU/ml. These recommendations apply both for quantitative and "absence of" testing, so it is worth reemphasizing that in qualitative tests it is often much easier to detect an objectionable organism from an inoculum of 104 than from 102 organisms. Thus the EP validation is, in this respect, more rigorous than that of the USP. The FDA would expect a low inoculum in the region of 10 to 100 CFUs.

If there is no reason to suspect that the sample will exhibit any intrinsic antimicrobial activity, it is sufficient just to inoculate it with a known number or concentration of the selected challenge organisms and demonstrate that the

*http://www.pharmig.org.uk/pages/news/issue5/big.html.

minimum designated percentage is consistently recovered (usually in three batches). The percentage recovery required varies: the PDA and USP direct that a minimum of 70% is acceptable, the EP states that the count "should differ by not more than a factor of five" from the value derived from the inoculum, and the American Society for Testing and Materials recommends statistical analysis to identify significant differences.16 The EP acceptance of 20% recovery is out of line with industry practice and regulatory expectations.

There are two strategies available for validating the recovery of microorganisms from medical devices, but both have their drawbacks. The first approach is that commonly used for raw materials and finished medicines whereby the device is inoculated with a known number of CFUs which are dried (under HEPA-filtered laminar flow conditions) onto the surface and subsequently recovered by one or more of the following: swabbing; immersion and rinsing; ultrasonics; and glass beads.10,11 The problem with this is that many organisms, particularly Gram-negative bacteria, are susceptible to desiccation and may be killed by the drying process itself, so there is always doubt whether a recovery of less than 100% is due to inadequate removal or bacterial killing. This is unlikely to be such a problem with Gram-positive challenge organisms, especially sporeformers. The alternative method described in ISO 1173711 is termed validation using repetitive recovery. Here, the product is not artificially inoculated, but its naturally occurring bioburden is enumerated by subjecting the product to repeated cycles of the recovery procedure until no more organisms are removed. This process is time-consuming and, as with a deliberate inoculation (spiking) method that gives a low recovery, there is uncertainty at the end about whether there are yet more organisms to be recovered. The ISO 11737 suggests coating the surface of the device with molten agar which, when set and incubated, should permit residual organisms to give visible colonies. Producing a coating of uniform thickness and incubating in conditions that prevent the coating drying complicate this approach.

For some products possessing antimicrobial activity membrane filtration may not be an option and dilution of the sample or the use of chemical inactivators may be necessary. The EP lists some common inactivators in 2.6.13, but more comprehensive lists appear elsewhere.17 It is necessary both to demonstrate that the inactivator does effectively eliminate the antimicrobial activity, and that it is not toxic to microorganisms, so any validation process should therefore involve three viable counts. The first is for the inoculum suspension of the challenge organism, and the second and third are the organism in the presence of the inactivator with and without the product sample (testing inactivator effectiveness and toxicity respectively). The acceptance criterion is again not less than 70% of the control count recovered throughout a minimum 30-minute period of contact between the challenge organism, inactivator and sample. Because inactivator formulae frequently contain surfactants like lecithin and tween that disaggregate bacterial clumps, it is not uncommon for the recovery value to be >100%.

8 DOCUMENTATION AND DATA ANALYSIS

It is necessary both to document the procedures fully and to establish a system for recorded and trending the data, particularly quantitative bioburden determinations. The documentation for both routine methods of bioburden determinations and the associated validation processes should comprise standard operating procedures that record materials and equipment, methods and acceptance criteria, together with identification of personnel responsible for obtaining and approving the data. The account of the methods must include the details of all measurements and measuring equipment, incubation conditions, the composition and source of all diluents, inactivators and media, and the source and maintenance of reference organisms.

The manner and detail in which results are recorded demand careful consideration. The primary data, i.e., the individual plate counts, together with the calculation of means and the final bioburden value that is the product of the mean and appropriate dilution factors, should all be recorded. It is desirable that bioburden counts are recorded in a manner that then makes them amenable to charting, trend, and possibly statistical analysis. It is important that a numerical value is recorded if at all possible. If, for example, results were regularly recorded for water as <1 CFU/ml it would be difficult, if not impossible, to ascertain how, or whether, the quality was changing. Quantitative bioburden data may conveniently be recorded using statistical process control charts, the uses of which, in a pharmaceutical manufacturing context, are described by Ingram and Cochrane.18

The colonial morphologies of the major organisms that routinely constitute the bioburden are likely to be familiar to the personnel who regularly undertake the work, and it is useful to record a presumptive identification based upon visual recognition of colonies. This recognition should comprise part of the staff training program. Gram-negative organisms are often of particular interest or concern, either as potential pathogens, or as a source of endotoxins. Consideration should be given to the possibility of including in SOPs a statement that any presumptive Gram-negative isolates should be examined under Gram stain, and if still Gram-negative, identified to species level. Well-defined procedures also need to be in place describing the application of out-of-specification and atypical analytical results procedures and the manner in which results are transmitted for the purpose of batch release.

9 REFERENCES

1 ISO 11134. Sterilization of Health Care Products — Requirements for Validation and Routine Control — Industrial Moist Heat Sterilization, 1994.

2 FDA Center for Drug Evaluation and Research. Guidance for industry Q7A good manufacturing practice guidance for active pharmaceutical ingredients, 2001.

3 European Commission. The Rules Governing Medicinal Products in the EC, Vol 4: Good Manufacturing Practice (reproduced in the Rules and Guidance for Pharmaceutical Manufacturers and Distributors 2002 U.K. Medicines Control Agency), 2002.

4 Green, S., Randell, C. Rapid microbiological methods explained. This volume, 2004.

5 Wassall, M. Rapid enumeration and identification methods. In Industrial Pharmaceutical Microbiology: Standards and Controls (eds. N. Hodges, G. Hanlon), pp. 5.1-5.33. Euromed Communications, Haslemere, U.K., 2003.

6 Halls, N. Microbiological environmental monitoring. This volume, 2004.

7 Cundell, A.M. Comparison of microbiological testing practices in clinical, food, water and pharmaceutical microbiology in relation to the microbiological attributes of nutritional and dietary supplements. Pharmacopeial Forum, 28, 964-985, 2002.

8 Kuwahara, S.S. Microbiological based statistical sampling. In Microbiology in Pharmaceutical Manufacturing (ed. R. Prince) pp. 485-505. Parenteral Drug Association, Bethesda, MD and Davis Horwood International Publishing, Godalming, U.K., 2001.

9 Millar, R. Enumeration of microorganisms. In Handbook of Microbiological Quality Control: Pharmaceuticals and Medical Devices (eds R.M. Baird, N. Hodges, S. Denyer), pp. 54-68. Taylor & Francis, London, 2000.

10 Parenteral Drug Association. Bioburden recovery validation. Technical Report 21. Journal of Parenteral Science & Technology, 44, (6), 324-331, 1990.

11 ISO 11737. Sterilization of medical devices — Microbiological methods — Part 1: Estimation of population of microorganisms on products, 1995.

12 Pharmacopeial previews: <61> Microbial enumeration tests; <62> Microbiological procedures for absence of objectionable microorganisms; <1111> Microbiological attributes of nonsterile pharmacopeial articles. Pharmacopeial Forum. 25, 7761-7791, 1999.

13 Fels, P. An automated personal computer-enhanced assay for antimicrobial preservative efficacy testing by the most probable number technique using microtitre plates. Pharmaceutical Industry, 57, 585-590, 1995.

14 Hodges, N. Pharmacopoeial methods for the detection of specified organisms. In Handbook of Microbiological Quality Control: Pharmaceuticals and Medical Devices (eds. R. Baird, N. Hodges, S. Denyer), pp. 86-106. Taylor and Francis, London, 2000.

15 Bergey's Manual of Determinative Bacteriology, 9th ed. Williams & Wilkins, Baltimore and London, 1994.

16 American Society for Testing and Materials. Standard Practices for Evaluating Inactivators of Antimicrobial Agents Used in Disinfectant, Sanitizer, Antiseptic and Preserved Products, Document E 1054-91, 1991.

17 van Doorne, H. A basic primer on pharmaceutical microbiology. In Microbiology in Pharmaceutical Manufacturing (ed. R. Prince), pp. 71-123. Parenteral Drug Association, Bethesda, MD and Davis Horwood International Publishing, Godalming, U.K., 2000.

18 Ingram, M., Cochrane, T. Statistics and statistical process control in pharmaceutical microbiology. In Industrial Pharmaceutical Microbiology: Standards and Controls (eds. N. Hodges, G. Hanlon), pp. 5.1-5.33. Euromed Communications, Haslemere, U.K., 2003.

Was this article helpful?

0 0
Lower Your Cholesterol In Just 33 Days

Lower Your Cholesterol In Just 33 Days

Discover secrets, myths, truths, lies and strategies for dealing effectively with cholesterol, now and forever! Uncover techniques, remedies and alternative for lowering your cholesterol quickly and significantly in just ONE MONTH! Find insights into the screenings, meanings and numbers involved in lowering cholesterol and the implications, consideration it has for your lifestyle and future!

Get My Free Ebook


Responses

  • heike luft
    Are media fills required for clinical phase?
    1 year ago

Post a comment