epr315 microbiology rt layout 1 29/06/2015 11:32 page 1 in ... · in-depth focus: microbiology...

Considering the upcoming revision of EU GMP Annex 1 for manufacture of sterile medicinal products, how would you propose the issue of different data forms(colony-forming units vs. bio-counts) be addressed so dueconsideration is made for rapid/real-time methods thatinclude electronic data ideal for trending?Herber: Clear guidance on limits seems to make life easier but it doesnot reflect the real world. We should see less guidance on specific limitsas technologies change. Annex 1 could put more emphasis on theexpectation of a risk-based approach to develop such limits – no matterif you use a ‘traditional’ or a modern microbial method. Although it is unrealistic to expect that Annex 1 can provide all sorts of limits, itcould state that the authors are aware of the limitations of the

traditional methods and that they see the value that modern methodscan bring as equal, at least.

Jones: The instantaneous type of viable cell detector has a solid nicheapplication for troubleshooting the cause and locations of microbialexcursions. Their use in monitoring clean room performance is growing as the industry and regulators adjust to the new ‘cell count’.The devices and methods have a great advantage in terms of detectingevents, however this type of testing still needs to be tied to moretraditional methods for surfaces and personnel and to close outinvestigations with the availability of an identification. Suitablevalidation guidelines and definition of the new baseline will givecompanies more options in this area.

IN-DEPTH FOCUS: MICROBIOLOGY ROUNDTABLE

68 European Pharmaceutical Review VOLUME 20 ISSUE 3 2015

ModeratorJames Drinkwater, Chairman, Pharmaceutical and Healthcare Sciences Society

Dr Ulrich Herber Sr European Product

Manager of the Charles River Endotoxin

and Microbial Detection service

David Jones PhD, Director of

Technical Services, Rapid Micro Biosystems

Paula Kelly Manager, Eurofins

Lancaster LaboratoriesProfessional Scientific

Services

Dr. Jasmin Grigat Product Manager

Microbiology, Lab Products& Services, Sartorius

Stedim Biotech GmbH

Rocco Petrizzo TSI Contamination Control

Application SpecialistEurope, Middle East

& Africa©

Jez

per /

Shu

tter

stoc

k.co

m

epr315 Microbiology RT_Layout 1 29/06/2015 11:32

Kelly: The change to rapid microbiologicalmethods (RMM) must be proven to beequivalent to or better than the traditionalmethod. Several considerations must betaken into account, including comparingdata from both methods over time, as well asperforming a comparison study between theRMM method against the traditional methodto demonstrate a correlation between bio-count and colony-forming unit (CFU)calculating accuracy, precision, specificity,detection limit, quantification limit, linearity,range, ruggedness and robustness. It is alsoimportant to establish new acceptancecriteria. Limits can be changed as long as anappropriate scientifically sound rationale is provided.

Grigat: The detection limits for RMM areoften more sensitive than the classicalgrowth-based methods (GBM) (of whichsome are RMM) which are limited due to availability of selective mediaand the reproducibility status of microorganisms etc. With a databasefrom comparison studies between GBM vs a defined RMM, a correlationbetween CFUs and bio-counts could be set up and new, potentiallylower alert and action levels easily be defined. The added value of amore sensitive method not distinguishing between viable/non-viable,dead/alive, culturable/non-culturable still remains unclear. Does thedetection of non-culturable, non-viable microorganisms need to beconsidered as critical as the detection of culturable ones?

Petrizzo: CFU reports relative levels of contamination and theoretically,CFUs can arise from a single viable cell but it is more likely that theyarise from an aggregate of ‘like cells’. Differences such as thosebetween zero and one are not wide enough to account for variability.Differences less than 10-fold may not be significant, particularly whenoperating at the limit of detection, as wetypically do in ISO 5. Sensitivity at single celllevel is needed with discrimination betweenviable and non-viable microorganisms andpossibility to detect viable but not culturablemicroorganisms. Different sampling tools andmeasuring systems will offer flexibility in the application of theappropriate technology for each sampling point. The USP 1116 is a goodway to go.

Is there a need to harmonise the approach on incubation regimes and should such guidance be provided in the revised GMP Annex 1? Jones: The change to one media and serial incubation may be driven bycost rather than science. This method can also delay the time todetermine a result due to the longer incubation required. While theincubation strategy does vary by site, serial incubation using a low tohigh temperature does seem to be the most popular method. Some sites also use a 28°C single temperature as a balance point and

claim good recovery. It may be appropriate to have some looseguidelines, e.g., serial low to high with a minimum of two days at eachtemperature or three to five days with a single temperature method,validated of course.

Kelly: Industry standards should be harmonised and guidance shouldbe provided in the revised GMP Annex 1 around this current debate.Clear guidance should be considered for dual incubation requirements(time, temperature and order of incubation), as well as dualmicrobiological media requirements. This approach should then alsobe harmonised with other regulators and outlined in the guidancedocumentation.

Grigat: Yes, this is definitely needed. TSA is a growth medium designed forbacteria whereas SDA is designed for mould and yeast recovery.

TSA needs to be incubated at 30-35°C which isthe optimal growing temperature for mostenvironmental bacteria. I believe that TSAshould be routinely used for environmentalmon itoring but for trending purposes, e.g., on aweekly base, since it also helps to monitor

spore-forming microorganisms. 20-25°C is the optimal growingtemperature for most mould and yeast species. Since SDA covers these organisms already there is no need to have a TSA incubation at 20-25°C. SDA is the medium which should be used for the routineanalysis at 20-25°C.

Petrizzo: There is no need to harmonise as this provides differentapproaches that manufacturers can use. Each method should bevalidated so that the media and incubation parameters will provide foroptimal recovery. Each aseptic manufacturer should consistentlyevaluate the growth promotion properties of media for a predefined listof organisms and must be able to prove that their microbial media are suitable to consistently recover environmental contaminants.

IN DEPTH FOCUS: MICROBIOLOGY ROUNDTABLE

VOLUME 20 ISSUE 3 2015 European Pharmaceutical Review 69

The change to one media and serial incubation may be driven by

cost rather than science David Jones

© K

ater

yna

Kon

/ Shu

tter

stoc

k.co

m


Plates should be tested for their ability to cultivate low numbers ofstandard bacteria and fungi. The reliable recovery of small numbers ofmicroorganisms is a prerequisite for successfully monitoring cleanrooms. When >50% of microorganisms fail to be detected for each species tested, the procedures utilised should be carefullyinvestigated and revised.

Biofilms in pharmaceutical water systems are notoriouslydifficult (or impossible) to remove once formed. Which microbiological methods can be applied to indicatethat conditions for biofilm formulation are developing? Jones: Rapid microbial results or in-line continuous monitoring bothallow the company a more detailed view of the ongoing status in thewater system. This allows actions to be taken earlier and hopefullyminimise the development of entrenched biofilms. These testingmethods need to start as soon as a new waterpipe is installed with adequate cleaningmethods to ensure no biofilm development.

Kelly: Several methods can be applied toindicate that conditions for biofilm formulationare developing, including total viable count using the membranefiltration technique, selective and chromogenic/fluorogenic media andmicrobiological isolation of Cryptosporidium, Giardia, Pseudomonasaeruginosa, Aeromonas hydrophila, Salmonella typhimurium,Campylobacter, Escherichia coli, Gram negative non-fermentingbacteria and coliforms. RMMs are also currently being developed to aidin the early detection of biofilm formation. It is also important to notethat biofilms are made up of more immobile microorganisms thanmobile ones. The above techniques may not detect the presence of abiofilm. As a result, a contamination control strategy approach must be considered.

Grigat: The formulation of biofilms is predominantly promoted by thegeometry of the devices and hosing connections as well as by

the materials used and stagnating water. To avoidbiofilms developing the distribution system has tobe constructed with no edges, rims or dead ends.Additionally, easy-to-clean and disinfectedmaterials should be used since regular cleaningand disinfection of the system and a constantcirculation of the water is a must to preventmicrobial growth. The choice of the system and awell-defined cleaning protocol are key for theprevention of biofilms. If indication of a biofilmformulation is given, it is already too late.

The identification of microorganisms, at least to species level, is a requirementin root cause investigations. Whatadvantages do you see in identifyingmicroorganisms during routinemonitoring other than just recordingCFU counts and how do you managerisks to avoid errors in identification?

Herber: Data collection for route cause investi gations should startbefore any alert/action level is exceeded. A past record of accuratelyidentified microorganisms from your routine EM program is invaluable.It provides more assistance in tracking the source of contaminationthan results coming from samples taken after the contamination event.My advice for managing risk regarding incorrect identifications is tochallenge the established technology by participating in aninterlaboratory test series which are unfortunately not popular in(bio)pharma microbiological QC or by sending a set of the mostfrequently occurring organisms to a service laboratory applying adifferent identification technology and to see how it performs againstyour approach.

Kelly: Microbial identifications during routine monitoring allows forestablishment of a baseline level for the typical microbial population

and helps confirm that microorganismsisolated are typical for that environment/system or indicative of no change in the level ofcontrol. Identifications can also assist in rootcause analysis of excursions/lab investigations,help to identify the top occurring isolates,

identify objectionable microorganisms, detect biofilms and aid in riskassessments. A robust training procedure for microbiologists is key.Pure cultures should be carefully isolated, stained and assessed. For automated systems, good preventative maintenance andqualification procedures are necessary. In addition, building a validatedreference library/database is essential to improving accurateidentification and avoiding errors.

Grigat: For cases of false positive results in the final product releasetesting such as sterility testing or microbial limits testing or in case ofalert/action levels reached, the identification of microorganisms duringroutine monitoring would accelerate the root cause analysis a lot.However, I believe that the knowledge about the typically foundmicroorganisms (specified microorganism, house-germs) might be

70 European Pharmaceutical Review VOLUME 20 ISSUE 3 2015

Several methods can be applied toindicate that conditions for biofilm

formulation are developingPaula Kelly

© ro

bert

_s /

Shut

ters

tock

.com



VOLUME 20 ISSUE 3 2015 European Pharmaceutical Review 71


sufficient for that. Defined in-process critical control points which are monitoredmonthly for trending purposes could be astarting point.

To avoid risks in identification, theprocess from sample taking to the identi -fication itself needs to be very preciselydefined. Positive and negative controls are a must.

Petrizzo: Repetitive detection of the samemicroorganism indicates that a constantsource of contamination is present.Unexpected or exotic species may resultfrom contaminated raw materials oroperators recently exposed to a disease notendemic to the site of the manufacturer. Insuch cases corrective actions may be moreurgent. Identifications (IDs) can be per -formed routinely, but most companies willonly perform IDs with excursions or during their routine strategy forperforming them (e.g., quarterly). Risks in performing IDs depend onthe system used, how it was validated and the capability of theoperators. Multiple methods of identification and replicates arerequired to manage risks of errors in identification.

What do you consider to be the future for pharmaceuticalmicrobiology, given risk-based contamination control andmonitoring is an essential requirement for assurance ofproduct sterility and release?Herber: Data management will be key in the future of pharmaceuticalmicrobiology. More data will be generated almost in real time. We needto bring all the different bits and pieces(bioburden testing, identifications, endotoxindetermination) together and draw the rightconclusions. Testing will become moreautomated and will happen in a decentralisedmanner closer to production. The role of the QCmicrobiologist will change from someone running a lab to someoneinterpreting data. We will have robust and accurate methods, easy-to-manage interfaces and user-friendly data management systems. Thisapproach will enable us to mitigate risk.

Jones: Pharmaceutical microbiology will continue to be a criticalcomponent of quality control, but microbiologists must find innovativeways to provide results faster and with fewer resources. With greaterregulatory acceptance of new methods to test product sterility and themanufacturing environment, more labs will be adopting rapid methods.

Kelly: The implementation of risk-based contamination control andmonitoring is a more effective, compliant, meaningful, representativeand scientific approach for ensuring product quality assurance. Thedrive for risk assessments for assurance of product sterility and releasehas initiated cross-functional teams with QC, QA, Production,Engineering and Validation departments, working together to identify

risks and apply risk ratings. As a result, microbiologists are seen out onthe field a lot more and are gaining a better understanding of theproduction process and engineering systems and other departmentsare gaining a better understanding of microbiology and risks to productsterility. The next step is to harmonise this approach across theregulatory bodies.

Grigat: The key is to understand the product and its characteristics aswell as the complete production process, from raw material testing, in-process controls to final product release testing. Hazard analysis andcritical control point needs to set the right control points formonitoring, automated sample-taking and well-established data

tracking support and smart data exchangeinterfaces like LIMS need to integrate all resultsobtained. The gained helicopter view facilitatesan easy review and fast actions. Proactively,quality by design and process analyticaltechnology will be more important and help to

control the process. However, I expect that end product testing will notbe replaced entirely by process control testing.

Petrizzo: Pharmaceutical microbiology will remain important, however, conventional monitoring methods will become inappropriate.We may be able to justify reducing finished product micro tests as we enhance overall control strategies and move monitoring upstream. The only way to ensure product safety in aseptic processingis through thorough training and supervision and by minimising risks bylimiting critical zone-personnel contact. We must systematicallyeliminate contamination risks. Quantitative risk analysis is likely a more accurate metric for sterility assurance. Equipment is alreadyavailable allowing real-time, simultaneous, viable and non-viable counting. Microbiological quality needs to be built into the drugs by understanding contamination sources. RMM is here to support this. The future will be led by real-time release testing andcontinuous monitoring.

The gained helicopter view facilitates an easy review

and fast actions Rocco Petrizzo

© to

toja

ng19

77 /

Shut

ters

tock

.com


epr315 microbiology rt layout 1 29/06/2015 11:32 page 1 in ... · in-depth focus: microbiology...

Documents