Introduction
At a recent European Union conference a senior inspector with one of the foremost regulatory agencies in Europe remarked, “What is wrong with BigPharma? We (the regulators) have opened the door and even drawn you the map and yet you still do not adopt rapid microbiological methods”.
As the person responsible for achieving the first multiple country approval for a rapid microbiological method (RMM) in Europe for finished product testing it left me bemused as well. Canvassing attendees at the meeting where all the major pharmaceutical companies were represented it soon became apparent that the “barriers” to implementation were predominantly internal. The perceived issues were:
1) Lack of real benefit in reduction of product release cycle time, such that the financial investment in RMM equipment and consumables were not justified.
2) for existing products filing a variation to introduce an RMM was not considered a high priority by the regulatory group
3) “Fear of failure”, what happens if the regulatory body does not approve the variation? Answer – fix the problem or maintain existing methodology based on 200 year old technology.
4) Resistance to introduction of RMM from microbiologists within a company. The reduction of the skills set needed to quantify microorganisms to mere “button pushing” was seen to be a threat to the continued employment of the microbiologist. Our chemistry colleagues must surely have faced the same dilemma with the introduction of fully automated HPLC equipment yet this does not seem to have impeded introduction.
Within the EU at least, manufacturers seem to have ignored one key point. The marketing authorisation holder (licence) has a legal obligation to utilise new technology where it provides improvement in for example analytical technology. In fact recent EU legislation (1) cites failure to introduce such advances as one of the areas which can attract censure and financial penalties.
Regardless of the other internal barriers the lack of cycle time reduction is also difficult to reconcile. Generally traditional microbiological methods provide results in the 2 day (for example microbial limits) to 14 day (sterility testing) timeframe. RMMs can provide comparable results within 24 to 48 hours. Certainly in the USA one company has already achieved FDA approval for the use of an RMM in place of a conventional sterility test saving c.10days of inventory hold. What is even more important to the author is the ability RMMs give to investigate microbiological issues within a meaningful timescale. Both U.S. and EU cGMPS talk to the contemporaneous and immediate nature of investigations and yet a 2 to 14 day delay for those in which microbiology is involved is tolerated. This renders many such investigations as exercises in memory recall for the people involved rather than meaningful investigations.
As an example, excursions in environmental monitoring in an aseptic processing area using traditional methods will not be realized until 2 – 5 days after the event. What were personnel doing during this sampling period? What was the status of product? How long did this “out of control” situation last? Is the whole batch at risk or just a portion? What is the data based rationale for a part release decision?
Admittedly not all these questions will be completely resolved by the introduction of RMMs but they do offer two distinct advantages. Firstly, results in a time-frame that should make reconstruction of events easier and secondly the ability to take more frequent samples in such an area providing a more comprehensive picture of the microbiological status of an aseptic area.
In some respects the major regulatory authorities could force the industry to adopt RMMs by insisting that for all new approvals in which microbiological methods are utilized, that the license is “non-approvable” unless RMMs are used. As previously indicated in the EU this would ensure that MA holders honour their legal responsibilities to use up to date technologies.
Selection of an RMM
A number of guides to the numerous RMMs available have been published (2, 3, 4, 5) and new proposals for RMMs appear regularly in the trade press. The PDA Technical Report No 33 “Evaluation, validation and implementation of new microbiological methods” 2000 provides an excellent overview of the types available as does the “quick guide” published by the author (6).
Selection of the most suitable technique needs a careful consideration as to the type of result required ~ quantitative or qualitative; the acceptable limit of detection ~ 10; 100; 1000 etc. or more colony forming units; whether subsequent identification of isolated organisms is required i.e. the method is non-destructive; how rapid is rapid and what accuracy may need to be sacrificed to achieve a result in a few hours versus 18-24 hours. It should also be noted that the adoption of an RMM may demonstrate an apparent rise in the number of micro-organisms enumerated. It has long been accepted that there is no such thing as an ideal media i.e. one which will cultivate all potential isolates or indeed ideal incubation temperature. With some RMMs the nature of the enumeration for example the “tagging” of a viable cell regardless of whether it was culturable by traditional methods will demonstrate increased numbers. This situation was recognised by the MHRA (UK Health Agency) who stated “for water used in pharmaceutical production, limits may have to be adjusted to compensate for increased sensitivity”.
There has also been considerable debate in the EU on VNC, viable non culturable organisms. These are often damaged or subtly changed organisms which will not be enumerated by traditional methods but will be counted by many of the RMM methods.
This potential for increased counts has also been a concern for some companies who are fearful that the higher counts seen with RMMs may be seen as representing a public health risk for their products by the regulators. In the authors experience this is not the case and regulatory bodies recognise that these levels have likely always been present and have not presented a known risk. There is some suggestion that VNC’s have been implicated in infection outbreaks which may challenge the current pragmatic view. The article by Cundell (7) provides a number of other factors to consider in the selection process.
In these days of cost sensitivity the capital cost and the ongoing consumables cost can be significant for RMMs. It is perhaps not surprising that the RMM equipment suppliers consider consumables as a legitimate ongoing profit generator. Nevertheless in the authors experience the selection process cost benefit calculation should involve the Company’s financial personnel. They will often identify the “hidden” costs of traditional methods. For example depreciation of autoclaves, calibration and qualification costs, disposal costs, support staff, all need to be factored into the costs of traditional media manufacture plus the records infrastructure. These costs are minimal with “plug in and play” consumables.
Finally the willingness of highly skilled microbiologists to accept methods which can be seen to negate the microbiological “black arts” they have been trained to perform cannot be overlooked. Certainly many RMMs require little microbiological knowledge and can be performed by non-specialist staff. The answer to this fear is to deploy skilled microbiologists to do what they should be doing i.e. spending time in the manufacturing areas looking at potential sources of contamination; advising on the design and cleaning of areas; looking at operational practices and so on.
Method Validation
Whole treatise have been written on the topic of validating RMMs and the author can do no better than to point the reader at the PDA Technical Report No 33, USP Informational Chapter 1227 (8) and 1233 (9) and Ph Eur 5.1.6 (10). It is strongly recommended that where RMMs are to be deployed for finished product testing that a meeting for technical review is requested from the appropriate regulatory agency. This will enable a clear understanding of the regulatory expectations to be achieved and lessen the chances of the filing/variation being rejected.
Setting aside the topic of equipment qualification for which the RMM supplier should be expected to provide a robust package requiring minimal additional work by the purchaser, the principle validation burden concerns the methodology itself. It is perhaps worth reflecting that traditional methods in use for > 200 years have not been validated to modern acceptance criteria and would in fact be expected to fail some of these criteria. For example the basic premise of all plate count studies is that each colony arises from a single cell, how often is this presumption correct?
Not withstanding the historical situation, for the substitution of an RMM for quantitative assessment the validation expectations would be
- Accuracy
- Precision
- Specificity
- Limit of quantification
- Linearity
- Range
- Ruggedness/Robustness*
- Equivalence/comparability
[* such data should be available from the supplier of the method]
It is easy when carrying out a method validation to become ensnared in the process itself without due consideration of the outcome. For example, statistically speaking, if a technician wished to claim with 90% probability that two methods used to enumerate a suspension containing 30 cfu with acceptance criteria of no more than a 10% difference, then 63 replicates would be required. There is scant regulatory guidance on what constitutes agreement although Ph Eur 2002 (2.6.12) states that where a limit of 102 is given then results up to 5x102 would be considered compliant and USP24/NF19 (1231) states the error on a plate count enumeration for a count of 3 cfu from a 10-1 dilution is 58%.
Comparability between traditional and RMM is a complex topic. As has already been mentioned due to the nature of the method RMMs might be expected to give “enhanced” recovery rates. It should also be noted that many statistical methods for method comparability are based on an assurance that the “concentration” was accurately known and that the “analyte” was normally distributed in the sample. Both of these situations may not apply for a microbiological method and those microbiologists with many years of bench work to their credit will attest to the difficulty of obtaining homogeneous suspensions of low levels of micro-organisms (particularly fungi).
Nevertheless techniques such as relative standard deviation (coefficient of variation) can be used but acceptance criteria of up to 30% should not be unexpected. F tests for comparison of standard deviations can also be used. Correlation coefficients can be used to determine linearity with an acceptance criterion of greater then 0.9. Finally t-tests can be applied if we assume that data is normally distributed.
As previously the microbiologist promoting the use of RMMs should be part of a team including a statistician but should not be surprised if universal agreement on the correct statistical treatment for results is difficult to achieve!
Conclusion
There appears to be little external barriers to the application and approval of an RMM. Such usage is well aligned with the Process Analytical Technology (PAT) initiative, honours the licence holders’ responsibility to deploy up to date methods and most importantly provides microbiologists with results within a timescale whereupon they can make reasoned judgements on results achieved. On the presumption that this is an attractive proposition to a Company’s microbiologist, then from the authors experience it appears that in-house management and regulatory personnel need also to be convinced of the merits of RMMs. Certainly with the bourgeoning of the biotech industry, many of whose products are aseptic injections, the continued use of the traditional sterility test would appear totally anachronistic and must surely be displaced by the routine acceptance of appropriate RMMs in the not too distant future.
Note: The views presented in this article are those of the author and should not be construed as being representative of Wyeth.
References
1. Commission Regulation EC 658/2007
2. K. O. Habermehl. Rapid Methods and Automation in Microbiology and Immunology. Springer-Verlag. 1985.
3. N. Halls. Microbiological Contamination Control in Pharmaceutical Clean Rooms. ERC Press 2004.
4. Rapid Microbiological Methods in Pharmaceutical Industry. Edited by M. C. Easter 2003.
5. Encyclopaedia of Rapid Microbiological Methods. Edited by M. J. Miller. DHI Publications 2005.
6. Rapid Microbiological Methods Quick Guide. Sue Horwood Publishing Ltd. 2003.
7. A. M. Cundell. American Pharmaceutical Review.
8. USP Informational Chapter 1227. Validation of Microbial Recover from Pharmacopoeial articles.
9. USP Chapter 1223 (Draft) 2003. Validation of alternative microbiological methods. Pharmacopoeial forum, vol. 29 (1) pp256-264.
10. Ph Eur 5.1.6. Alternative Methods of Control of Microbiological Quality. Pharm Europa Vol.1m No.4, Oct 2004
Stewart holds degrees in Applied Biology and Microbiology as well as a Masters in Strategic Quality Management. He has occupied increasingly senior roles in the Pharmaceutical Industry spanning a career of almost 40 years. He has worked in Quality Control and Quality Assurance as well as validation, production and logistics. He is current Chairman of PharMiG a UK special interest group for Pharmaceutical microbiology, sits on the British Pharmacopoeia Expert Advisory Group for Antibiotics and the British Standards LBI/030 committee for Cleanroom Technology. He has retained a life long working interest in Aseptic manufacturing in all its guises. He is currently Director of Quality for Wyeth UK