E D I T O R I A L
A D V I S O R Y
B O A R D
sPecial eDitionn cleaninG valiDation iii
Gamal amer, P.D.
PCI, Pharmachem International
Validation and Process Associates, Inc. louis a. anGelucci, iii Foster Wheeler Corporation GeorGe n. Brower Analex Corporation Kenneth G. chaPman Drumbeat Dimensions, Inc. Dennis christensen Consultant roBert c. coleman US Food & Drug Administration shahiD Dara Independent Consultant DaviD r. Dills Medtronic Xomed michael Ferrante Catalytica Pharmaceuticals Patricia stewart Flaherty Bayer Corporation roBerta D. GooDe Consultant cynthia Green Northwest Regulatory Support Daniel harPaz, P.D.
william e. hall, P.D. Hall & Associates elDon henson Boehringer Ingelheim Animal Health Jay h. KinG LifeScan, a Johnson & Johnson Company John G. lanese, P.D. The Lanese Group, Inc. BarBara mullenDore AstraZeneca roBert a. nash, P.D. St. John’s University charlie neal, Jr. BE&K toD e. ransDell Bio-Rad Laboratories melvin r. smith Independent Consultant roBert w. stotz, P.D. Validation Technologies, Corporation eric D. veit Johnson & Johnson DaviD w. vincent Validation Technologies, Inc.
J O UR N AL M I S SI O N The Journal of Validation Technology is a peer-reviewed publication that provides an objective forum for the dissemination of information to professionals in FDA-regulated industries. TheJournal’s Editorial Advisory Board reviews
all submissions to current ensure that theystandards, have been and researched thoroughly, reflect industry are not promotional in nature. The Journal will not publish articles which have not been approved by the Board.
4
Institute of Validation Technology
Editor and Publisher Glenn Melvin Vice President Terri Kulesa Production Director Edward Eick Associate Publisher Brandon Melvin
Disclaimer:
Any reproduction of the contents of this publication in whole or part is strictly prohibited without permission. Views and conclusions expressed in articles herein are those of the authors. The publisher accepts no responsibility for the accuracy of information supplied herein or for any opinion expressed. No liability can be accepted in anyway. The information provided does not constitute legal advice. change of aDDress:
Notices should be sent promptly. Provide new address, including zip code or postal code. submissions:
Manuscripts are welcomed. Please call for editorial guidelines. reprints:
Reprints of all articles in this issue are available. Call 561-790-2025 for more information.
PO Box 6004 Duluth, MN 55806 Telephone: 218-723-4977 U.S. only: 888-524-9922 Fax: 218-723-9308 or E-Mail:
[email protected] wb s: .. ISSN 1079-6630
CONTENTS T A B L E
sPecial eDition
n
O F
cleaninG valiDation iii
equiPment cleaninG valiDation: microBial control issues . . . . . . . . . . . . . . . . . . . . . . 6 by
Destin A. LeBlanc, M.A.
cleaninG valiDation: maximum allowaBle resiDue: question anD answer. . . . . . . 13 by
William E. Hall, Ph.D.
DeveloPment oF total orGanic carBon (toc) analysis For DeterGent resiDue veriFication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 by
James G. Jin and Cheryl Woodward
total orGanic carBon analysis For cleaninG valiDation in Pharmaceutical manuFacturinG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 by
Karen A. Clark
DeterGent selection – a First critical steP in DeveloPinG a valiDateD cleaninG ProGram .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 by
Mark Altier
analysis cleaninG valiDation samPles: what methoD? . . . . . . . . . . . . . . . . . . . . . . . . . 35 by
Herbert J. Kaiser, Ph.D., Maria Minowitz, M.L.S.
control anD monitorinG oF BioBurDen in Biotech/Pharmaceutical cleanrooms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 by
Raj Jaisinghani, Greg Smith and Gerald Macedo
a c leaninG v aliDation P roGram F or t he e liFa s ystem. . . . . . . . . . . . . . . . . . . . . . . . . . . 56 by
LeeAnne Macaulay, Jeff Morier, Patti Hosler and Danuta Kierek-Jaszczuk, Ph.D.
a cleaninG valiDation master Plan For oral soliD Dose Pharmaceutical manuFacturinG equiPment ..................................... by
S U N O B
61
Julie A. Thomas
ProPoseD valiDation stanDarD — vs-3
.........................................
Special Edition: Cleaning Validation III
71
5
Equipment Cleaning Validation: Microbial Control Issues By Destin A. LeBlanc, M.A. Cleaning Validation Technologies
v
T
he PDA spring conference was held in Las Vegas, Nevada in March 20, 2001. The conference showcased cleaning validation, residue limits, bioburden, microbial limits, and sanitization. This paper is based on a presentation at that conference. The initial focus of regulatory documents relating to cleaning validation for process equipment
}…it
is becoming more common for regulatory authorities to cite manufacturers for deficiencies related to microbial control in cleaning validation programs.~
in pharmaceutical manufacturing involved measuring residues of the drug active and the cleaning agent. For example, the introduction to the Food and Drug Ad ministration (FDA) guidance document on clean ing validation1 states: “This guide is intended to cover equipment cleaning for chemical residues only.” While admitting that microbial re sidues are beyond the scope of the guideline, that guidance document further states, “microbiological aspects of equipment cleaning should be considered,” particularly with reference to preventive measures so that microbial proliferation does not occur during storage. The 2 European PIC/S document, that was issued several years later, does explicitly mention microbial sire dues. In Section 6.2.1, contaminants to be removed include “the previous products, residues of cleaning agents as well as the control of potential microbial con6
Institute of Validation Technology
taminants.” However, Section 6.7 of this document that covers “Micro biological Aspects” focuses exclusively on the same issue discussed in the FDA guidance document, namely the issue of preventing microbial proliferation during storage. As a practical matter, microbial residues on equipment surfaces are part of the contaminants that should be reduced to an acceptable level;
that acceptable level being what is safe for the manufacture of the subsequently manufactured product. Unfortunately, very little has been written on what is a safe level for microorganisms following cleaning and/or sanitation.3,4 Part of the reason for this is that microbial residues are significantly different from chemical residues. Chemical residues are “inert” in the sense that it is easy to calculate (especially using scenarios of uniform contamination in the subsequently manufactured product) the potential levels and effects of those chemical residues in the subsequently manufactured product should they be transferred to that subsequently manufactured product. With microbial residues left after the cleaning process, the situation is somewhat different. Because microorganisms are living organisms, those left as residues on equipment may change in number after the cleaningprocess, but
Destin A. LeBlanc, M.A.
before the manufacture of the subsequently manufactured product. Those microbes transferred to the subsequently manufactured product may also change in number after they are incorporated into the subsequently manufactured product in the manufacturing step. This change may be a significant reduction in bioburden, either due to drying of the equipment or due to a preservative in the finished drug product, for example. This change may also involve rapid proliferation, either due to suitable growth conditions in wet equipment during storage, or due to suitable
the cleaned equipment. However, many times this does not include any assessment as to the effect of that unchanged bioburden level on the subsequently manufactured product. This paper will address issues coveringproaches ap to control of microorganisms in process equipment, setting of acceptance limits, sampling techniques, and approaches toproviding acceptable documentation.
Microbial Control Measures
growth conditions in the finished drug product. Or, Control measures to reduce the bioburden on they may result in no significant change in microbial cleaned process equipment include control of biolevel, because the bioburden was due to bacterial burden of raw materials, the cleaning process itself, spores (that will survive readily in dried equipment), or because the }Some companies will measure the subsequently manufactured product was a dry product (with low water change in microbial levels on activity). Therefore, knowing the equipment surfaces during storage levels of microorganisms left on the equipment following cleaning does of the cleaned equipment. However, not necessarily give one the full many times this does not include any story of the potential hazards of those microbial residues. Additional inforassessment as to the effect mation is required to assess those of that unchanged bioburden potential hazards. Why has microbial evaluation during cleaning of process equiplevel on the subsequently manufactured product.~ ment been a little discussed topic? Part of the reason is that it is not a significant problem in process manufacturing. Yes, it could conceivably be a problem if a separate sanitizing step, and drying of the equipment following cleaning. Bioburden of raw materials cleaning and storage were inadequate. However, for the most part, cleaning and storage of pro cess equip- includes the active, excipients, water,and any processment, in so far as it applies to microbial residues, ing aids. In many cases, the manufacturer may have probably is done relatively well in most pharmaceu- little control over the bioburden of raw materials other tical manufacturing facilities. On the other hand, it is than to accept a specification by the raw material supbecoming more common for regulatory authorities to plier. The most critical raw materials probably will be cite manufacturers for deficiencies related to micro- natural products, in which there may be considerable bial control in cleaning validation programs. One variation in the levels and types of microorganisms. in bioreason for this seeming anomaly is that while firms A solid monitoring program to control coming are adequately controlling microbial contamination of burden of raw material is necessary. If there could be process equipment, there may be little documentation significant variation in bioburden, then that should to support this. This lack of documentation includes be addressed in the cleaning validation Performance any measurement of microbial residues during the Qualification (PQ) trials. At least one PQ trial should cleaning validation and/or during routine monitoring. utilize the worst-case incoming bioburden of raw Some companies will measure the change in micro- materials to demonstrate adequate cleaning and microbial levels on equipment surfaces during storage of bial control under those conditions. Special Edition: Cleaning Validation III
7
Destin A. LeBlanc, M.A.
A second means of microbial control is the cleaningconditions (“initial” bioburden, time, temperature, and process itself. The conditions of aqueous cleaning humidity). Needless to say, if the chemical sanitizing are often hostile to microbial survival. These ditions con step is performed immediately prior to manufacture of include high temperature (commonly 60-80ºC), pH the subsequently manufactured product, then removal extremes (>11 and <4), and the presence of oxidizers of the sanitizer chemical residues to an acceptable level (such as sodium hypochlorite in biotechnology manu- should also be demonstrated. facture). In addition, the presence of surfactants in the A fourth consideration for control of microorcleaning solution can assist in providing good physical ganisms is drying the process equipment surfaces removal of microbes (without necessarily killing them).following the final rinse. Drying the surfaces will Good cleaning is also beneficial to microbial control infurther reduce the levels of vegetative organisms on that chemical residues left behind can provide physia the surface. In addition, drying will assist in preventcal “microbial trap” to allow microorganisms to surviveing microbial proliferation during storage. Drying even in the presence of chemical sanitizers. Those can be achieved by heated air, heated nitrogen, or chemical residues left behind might also serve as a by rinsing with alcohol. In all cases, the process can nutrient source that allows microbes to proliferate dur-be assisted by application of a vacuum (to speed the ing improper storage. Based on the author’s experience,evaporation of the water or, in the case of an alcohol in most cases, effective control of microorganisms in rinse, of the alcohol itself). pharmaceutical process equipment can be achieved Limits for Microbes with the use of an effective cleaning process, without the need for a separate chemical sanitizing step. In some cases, a separate sanitizing step may be As mentioned earlier, it is possible to reasonably necessary. This may include sanitation by steam or by predict levels of chemical residues in subsequently chemical sanitizers. Suitable chemical sanitizers for manufactured products based on the levels present on 5,6 With microorganisms, it is posprocess equipment include sodium hypochlorite (chlo- equipment surfaces. rine bleach), quaternary ammonium compounds, alco- sible to measure levels on equipment surfaces; howhol (ethyl or isopropyl), hydrogen peroxide, and per- ever, the effect of those residues will depend on what acetic acid. It should be noted that, with the exception happens to those microorganisms once they come in of alcohol and hydrogen peroxide, additional rinses contact with the subsequently manufactured product. would be necessary to remove any chemical residues Areas that may have to be evaluated include the species of the sanitizer from the equipment. Those chemical (including the so-called “objectionable” organisms), residues may also have to be evaluated as residues to type of organism (vegetative bacteria versus bacterial be measured in the cleaning validation protocol. For spore, for example), the presence of preservatives in that such chemical treatments, it is not an expectation that subsequently manufactured product, the water activity the equipment be sterile. Unless the final rinse is with of the subsequently manufactured product, as well as sterile water, microorganisms will be reintroduced any subsequent sterilization process performed on that into the equipment from the use of Water-for-Injection product. As a general rule, if the water activity is less (WFI) or purified water as the final rinse. than 0.6, then it can be expected that microorganisms Some companies will use an alternative to sanitizing will not proliferate (although they may continue to sur7 Water activity is a physicalimmediately after cleaning. This usually involves sani- vive without reproducing). presses the water vapor tizing after storage and immediately before use. This chemical measurement that ex may be used in situations where it is difficult to control pressure above the test sample as a fraction of the water microbial recontamination or proliferation during stor- vapor pressure of pure water at the same temperature age. It should be noted that control of storage condi- as the test sample. For aqueous products with a neutral tions, if possible, is preferable. The practice of relying pH, microbial proliferation can generally be expected solely on a separate sanitizing step immediately before unless there is a preservative in the product. If there manufacture should be discouraged . If this is practiced, is a possibility of microbial proliferation because the then the sanitization step should be shown to be effec- product is unpreserved and neutral, then that should be tive in reducing bioburden under theworst-case storage addressed in setting limits. 8
Institute of Validation Technology
Destin A. LeBlanc, M.A.
Three methods to set microbial limits will be ment surfaces are not the only source of bioburden. addressed. The first (Case I) involve limits where theOne must also consider the raw materials themselves, subsequent product does not allow microbial prolif-as well as the primary packaging, as potential sources eration and is not subject to any further sterilizationof microorganisms. The best way to deal with this process. The second (Case II) involves subsequentlyissue is to develop information on the bio burden of the manufactured products that are terminally sterilized. raw materials and the primary packaging, and factor The third (Case III) involves subsequently manufac- these into the limits calculation. For example, if one tured products that are processed aseptically. were dealing with an oral liquid, one might calculate the contribution from the raw materials (assuming Case I Limits the upper limit bioburden for each raw material) as a If the subsequently manufactured product does not maximum of 27 CFU/g. At the same time the contribuallow microbial proliferation, then the determination tion from the primary packaging is determined to be 3 of acceptable microbial limits in the cleaned equip- CFU/g. Therefore, the amount allowed from equipment ment can be calculated using the same principles used surfaces would be 70 CFU/g (100 minus 27 minus 3). for chemical residues with one important exception. An additional safety factor should be used to account This process involves first determining the accep- for the significant variability in microbiological enutance limit in the subsequently manufactured product. meration. An appropriate factor may be on the order This limit is typically given in Colony Forming Units of 5. Therefore, in this case, the limit (in CFU/g) that (CFU) per gram of product. Once this is determined, would be allowed solely due to the cleaned equipment then the limit per surface area of equipment (assum- surfaces would be 14 CFU/g (obtained by dividing 70 ing uniform contamination) can be calculated based by 5). Higher safety factors also could be considered. on the batch size of the subsequently manufactured These numbers are given for illustration purposes only. product and the equipment surface area. It should be realized that the contribution percentage How is the limit in the subsequently manufactured allowed from cleaned equipment would vary dependproduct determined? For chemical residues, it is baseding on the contributions from the raw materials and the on dosing information for actives or toxicity formation in primary packaging. for cleaning agents. Such concepts cannot be directly Once the limit in the subsequently manufactured applied to microbes. Fortunately, there are two good product allowed from the cleaned equipment sursources of information relating to levels of microorgan-faces is determined, the next step is to determine the isms in products. One isthe manufacturer’s own Quality limit per surface area (CFU/cm2). This is calculated Control (QC) specifications for the product, thatmay exactly as it would be for chemical residues: include a limit for bioburden in the product. A second source is information given in the proposed UnitedLimit per surface area = LSP x MBS SA States Pharmacopeia (USP) <1111> relating to “Microbial Attributes of Nonsterile Pharmacopeial where Articles.”8 Examples of those limits are given below: LSP = Limit in the subsequent product Solid oral: ≤1000 CFU/g MBS = Minimum batch size Liquid oral; ≤100 CFU/g SA = Product contact surface area Topicals: ≤100 CFU/g In the example above, if the batch size is 200 kg 2 Note: Although these limits were discussed and and the product contact surface area is 260,000 cm , then the microbial surface limit of the cleaned equipproposed in the Pharmacopeial Forum, these specific recommendations were not adopted officially ment is: as part of the 24th edition of the USP. 2 Unfortunately, this is where the one exception to Limit per surface area = (70 CFU/g)(200,000g) = 54 CFU/ cm the conventional treatment arises. When one looks at (260,000 cm2) the bioburden in a finished drug product, the equipSpecial Edition: Cleaning Validation III
9
Destin A. LeBlanc, M.A.
If sampling were done with a typical contact plate of 25 cm2, this would correspond to a limit of over 1300 CFU per contact plate. Since it is reasonable to count a maximum of only 250 CFU on a typical contact plate, this would clearly be in the TNTC (too numerous to count) category. Needless to say, this will vary with the limit in the subsequently manufactured product, the portion allowed from cleaned surfaces, the safety factor used, batch size, and the shared surface area. However, under most reasonable scenarios, the calculated limit due to microorganisms on the cleaned
ment surfaces where the subsequently manufactured product is aseptically produced. This case is slightly different from Case II in that it is the equipment itself, and not the product, which is subsequently sterilized. This case is relatively straightforward, because the microbial limits on the surfaces of cleaned equipment are established based on the assumed bioburden of the equipment surfaces for sterilization validation of that equipment. No information on batch sizes or surface areas is necessary. The assumed bioburden for the sterilization validation can be used directly for limit
equipment surfaces will be significantly above what purposes. The only adjustment may be the incorporashould be (and can be) achieved by proper cleaning. tion of a safety factor (to accommodate normal variaAs a general rule, a good cleaning process should tion in microbiological enumeration). produce surfaces that contain no more than 25 CFU 2 per contact plate (<1 CFU/cm ). When failures occur, Measurement Techniques generally they will be gross failures, with counts generally above 100 CFU per-plate. Conventional tools used for microbial enumeration from surfaces can be used. These include rinse water Case II Limits sampling (usually with membrane filtration), swabThis involves setting limits for cleaned equipment bing (with desorption of the swab into a sterile soluwhen the product subsequently manufactured in that tion and then a pour plate count), and use of a con tact equipment is to be sterilized. In this case, the microbial plate. The choice of recovery medium and incubation limit in the subsequently manufactured product can be conditions is usually dictated by the expected organestablished based on the assumed bioburden of that isms. As a general rule, the initial focus is on aerobic product at the time of sterilization. In other words, any bacteria. However, if anaerobic bac teria or molds/ validated sterilization process depends on an assumed yeasts are suspected problems, these should be also bioburden of the item being sterilized. That assumed evaluated. bioburden then becomes the limit in the subsequently One issue that does not translate directly from manufactured product. Once that limit in the subse- chemical residue measurements is the idea of deterquently manufactured product is established, then the mining percent recovery using the sampling method. calculations are the same as for Case I – a certain por- In the measurement of chemical residues, the target tion of that total limit is allowed from cleaned equip- residue is spiked onto a model surface and the quanment surfaces, a safety factor is applied, and then the titative percent recovery is determined. The amount limit per surface area is calculated using the minimum recovered as a percent of the amount spiked is considsubsequent product batch size and the product contact ered the sampling method percent recovery. Per cent surface area. It is significant that this issue is actually recoveries in chemical sampling measurement are addressed in the FDA’s cleaning validation guidance generally above 50 percent. This percent recovery is document; that states: then used to convert an analyzed sample value; for example, if a chemical residue measured by a swab“…it is important to note that control of biobing technique gives 0.6µg of residue, then with a 50 burden through adequate cleaning and storage of percent recovery, this actually represents the possibilequipment is important to ensure that subsequent ity of 1.2 µg being on that surface. This concept cansterilization or sanitization procedures achieve not be applied directly to microbiological sampling. the necessary assurance of sterility.”9 The reason for this is partly theinherent variability in microbiological testing. If one measured 10 CFU in Case III Limits one test and 5 CFU in a duplicate test (a 50 percent This third case involves setting limits on equip- difference), one would be hard pressed to say that 10
Institute of Validation Technology
Destin A. LeBlanc, M.A.
those numbers are significantly different. In addition, limits should be included in the validation protocol, how would one actually measure the percent recovery and measured as part of the three PQ trials. One in a microbiological test? If a model surface is spiked should also include the absence of “ob jectionable” with a specific number of a certain bacterium, and organisms as part of the acceptance criteria. then that surface is allowed to dry and is sampled, To deal with processes for which cleaning validajust the process of drying might cause a low recovery tion has already been completed, but for which no of bacteria (due to the dying of vegetative bacteria by microbial evaluation has been done, there are two drying). In addition, what species of bacteria would strategies available. The objective of each is to develbe used for the recovery study? op documentation that the cleaning process consisIt is recognized that microbiological sampling tently provides equipment surfaces with acceptable methods may understate the number of microbes on bioburden. One option is to perform a cleaning a surface (indeed the concept of a CFU, that may validation PQ, measuring only bioburden on surfaces for comparison to calculated acceptance limits. The other option }One issue that does not translate is to initiate a routine microbiological monitoring program as part of directly from chemical residue the monitoring of cleaning. This measurements is the idea of may involve something as simple as monitoring the bioburden in the determining percent recovery final rinse water to demonstrate conusing the sampling method.~ sistency. This data, combined with product QC data on bioburden, may satisfy the need for adequate docucontain any number of bacteria, also clouds the issue). mentation. There are two ways to view such an issue. One is to One should also consider one’s motivation for make it clear that whatever variation exists in measur-wanting to obtain assur ance that the bioburden is ing microorganisms on surfaces is probably equally an acceptably low after cleaning. If the im petus for action issue when one sets limits based on product limits or is due to lack of data, one should resist the impulse to sterilization bioburden limits. Therefore, the variabili- immediately add a sanitizer into the cleaning program. ty issue becomes a “wash.” The other perspective is to The focus should be on developing data to demonstrate account for such variation by choosing extremely high the sufficiency of the current cleaning process. Adding safety factors. In the calculation example for Case I, a separate sanitizing step only complicates matters by a factor of 5 was used as a safety factor. Even if that adding additional residue concerns. If the impetus for safety factor were increased to 10 or 20, the calculated action is due to observed high microbial counts on acceptance limits would have still been ex tremely equipment surfaces or (more likely) in manufactured high, and still beyond what one should achieve with a product, then it is important to determine by careful well-designed cleaning program. investigation whether that unacceptable contamination is due to issues with the cleaning process, with storDocumentation Strategies age, or to both. In such a case, a separate sanitizing step should only be added if the data fully support it. How these issues will be addressed will depend on the stage of the cleaning process development. For a Conclusion new process being designed, the best strategy is to prepare a calculation of microbial limits, and then design Bioburden on cleaned equipment is an importhe cleaning process to meet those acceptance criteria. tant concern in the cleaning process. Fortunately, Included in that evaluation should be any change in most aqueous cleaning processes, properly designed, bioburden (in particular, any increase or proliferation) should provide low and acceptable bioburden levels cess. on storage of the equipment. The micro bial acceptance on equipment surfaces following the cleaning pro Special Edition: Cleaning Validation III
11
Destin A. LeBlanc, M.A.
Proper drying and storage should provide assurance that microbial proliferation does not occur be fore the manufacture of the subsequently manufactured product in that equipment. Any scientifically justified determination of acceptable bioburden levels, particularly for non-sterile products, is generally far higher than what should be achieved in conventional practice. This is becoming more of a regulatory and compliance issue, not because microbial contamination is a widespread problem, but rather because pharmaceutical manufacturers may lack appropriate documentation to support their practices. This can easily be remedied by a separate validation protocol to address microbial issues, or by routine monitoring to demonstrate consistency.o
About the Author Destin A. LeBlanc, M.A., is with Cleaning Validation Technologies, providing consulting in the area of pharmaceutical cleaning validation. He has 25 years experience with cleaning and microbial control technologies. He is a graduate of the University of Michigan and the University of Iowa. He can be reached by phone at 210-481-7865, and by e-mail at
[email protected].
References 1. FDA. “Guide to Inspections of Validation of Cleaning Processes.” 1993. 2. Pharmaceutical Inspection Cooperation Scheme. Recom mendations on Cleaning Validation. Document PR 1/99-2. Geneva, Switzerland. April 1, 2000. 3. A.M. Cundell. Microbial Monitoring. Presented at the 4th IIR Cleaning Validation Conference, October 20-22, 1997. (http:// microbiol.org/files/PMFList/clean.ppt, accessed May 29, 2001). 4. S.E. Docherty. “Establishing Microbial Cleaning Limits for Nonsterile Manufacturing Equipment.”Pharmaceutical Engineering. Vol. 19 No. 3. May/June 1999. Pp. 36-40. 5. G.L. Fourmen and M.V. Mullen. “Determining Cleaning Validation Acceptance Limits for Pharmaceutical Manufact uring Operations.” Pharmaceutical Technology.Vol. 17 No. 4. 1993. Pp. 54-60. 6. D.A. LeBlanc. “Establishing Scientifically Justified Ac ceptance Criteria of Finished Drug Products.”Pharmaceutical Technology. Vol. 19 No. 5. October 1998. Pp. 136-148. 7. R.R. Friedel. “The Application of Water Activity Measurements to Microbiological Attributes Testing of Raw Materials Used in the Manufacture of Nonsterile Pharma ceutical Products.” Pharmacopoeial Forum. Vol. 25 No. 5. September-October 1999. pp. 8974-8981. <1111> Microbial Attributes of Nonsterile Pharmacopoeial Articles (proposed). Pharmacopoeial Forum. Vol. 25 No. 2. March-April 1999. Pp. 77857791. 9. FDA. “Guide to Inspections of Validation of Cleaning Processes.” 1993.
8.
12
Institute of Validation Technology
Article Acronym Listing CFU: FDA: PQ: QC: USP: WFI:
Colony Forming Units Food and Drug Administration Performance Qualification Quality Control United States Pharmacopeia Water-For-Injection
Cleaning Validation: Maximum Allowable Residue Question and Answer
W
e are involved in the production of soft gelatin }…sometimes capsules and tablets in the many our newly built facility. Our products consist of at least 17 minerals possible and multivitamins in a single procombinations duct, while other products consist of the same ingredients having some of products and quantity (in MG) varying with the previous one. In some products, equipment would some vitamins are not present. Iwant result in so many to know how to conduct a cleaning validation study of each product. studies that the Again, I want to know which ingrecompany would dients I have to check after cleaning of the equipment to determine the never be able to residues?
The choice of which ingredient in a multi-ingredient product should serve as the focus of the cleaning validation is often a difficult one for vitamin and mineral products. For classical pharmaceutical products, the choice is usually based on choosing the most potent ingredient, or the least water soluble ingredient, or a combination of these two factors. For vitamins and minerals the choice may be more difficult because of the many ingredients present in the formulation and the relatively small amounts present. Coupled with these difficulties is often the difficulty in assaying the very small amounts of complete them active residues that might be presduring a • What will the limit be for the microent after cleaning. My suggestion bial contamination for the cleaning would be to identify an ingredient for reasonable validation studies, and what will be which there is a good sensitive assay period of time.~ available. For example,if one of the the rationale for the same? • If I’m using som e cleaning agent, ingredients happens to show good then what rationale is used for detectable levels of fluorescence keeping the limit the same? (e.g., riboflavin, folic acid, and certain B vitamins show good fluorescence) in water, then this material Thank you for your question. It is a very good could be selected as the “marker” material, and could one because it represents cleaning from the serve as the ingredient to focus on during the analysis point of view of a manufacturer of vitamins and min- of the rinse samples. In the case of vitaminsand minerals, which in some countries, are considered drugs, erals, it may be necessary, and even highly desirable, and in other countries, are considered as “nutraceuti- to take this approach because of the extremely low cals,” an important and emerging part of our business. levels of residuespresent after cleaning. It may also The first specific question you asked related to be possible to examine equipment in a dark room with how to conduct a cleaning validation for each prod- the use of an ultraviolet light to identify areas of equipuct, and how to select which ingredient to check ment that are not cleaned sufficiently (an enhanced after cleaning to verify that the cleaning is adequate. visual examination), again utilizing the known fluo-
A:
Special Edition: Cleaning Validation III
13
William E. Hall, Ph.D.
rescent behavior of certain vitamins. A brief study will cases, I would suggest that you refer to the Internet, need to be carried out to determine if this approach is and conduct a search on the toxicity or potency of these appropriate and adequate for your particular situation. materials. You may be surprised to find that a vitaI would suggest that you not try to con duct cleaning min, such as folic acid, is quite potent in terms of its validation for every product. The reason I say that medical effect and dosage. is because sometimes the many possible combinaThe limits for these products can be calculated tions of products and equipment would result in so by allowing a certain small fraction of vitamins or many studies that the company would never be able minerals to carry over to each dose of the following to complete them during a reasonable period of time. product. Again, you will need basic information, such If, for example, you have 50 products, and each could as the medical dosage of the initial product, the batch be run on ten (10) different pieces of equipment, then size and dosage of the next or subsequently manufacyou would need 500 studies to cover all the possible tured product. In terms of the safety factor, i.e., the combinations and permutations. That is simply too factor that is used to reduce the allowable dosage, I th for vitamin much of a resource and cost issue for the average suggest that you use a factor of 1/100 th company to face. It would be much better to divide and mineral products. A factor of 1/1000 is often your products into groups or families, and choose one used for pharmaceuticals, but I feel a more generous or two representatives from each group to conduct full factor of 1/100th is appropriate for vitamin and mincleaning validation. The assumption is that you can eral products. You could refer to some of the articles pick some “worst-case,” most difficult to clean, potent published in the Journal of Validation Technology for products from each group. The first step is to divide the details of how to calculate specific limits. the products into groups. I don’t know the names and Your last question related to what rationale should ingredients of the products your company manufactur- be used for the cleaning agent itself. The basic ers; however, you did mention that some products are requirement is that you be able to provide data that moved vitamin products and others are mineral products. So I demonstrates that the cleaning agent itself is re think there would be two major groups – vitamins and during the cleaning process, usually by the final rinse. minerals. Then each of these groups might be further You will need to go through the same rationale for divided, if necessary. For example, in the vitamin cat- the product residue limits, i.e., establish a scientific egory you may have some products that contain water basis or justification that shows that the most potent soluble vitamins, and some that contain fat soluble ingredient in the cleaning agent is reduced to a medivitamins. So now we have three (3) major groups cally insignificant level. It is beyond the scope of this (water soluble vitamins, fat soluble vitamins, and answer to go into the mathematical details of how to mineral products). So you begin to see our approach. calculate this data, but again the details can be found It might be that if you have vastly different types of in the various articles published in theJournal of mineral products you might want to also further divide Validation Technology. You will need to know about that group into smaller groups. In any event, you want the ingredients in your cleaning agent, as they are to have probably four (4) to ten (10) products in each typically multi-ingredient formulations, just like our group, and then pick a worst-case representative from pharmaceutical products, and you will need to get that each group. So by choosing this “grouping approach,” information from your supplier of cleaning agents. you have reduced the work from a very large resource The good news is that if you use the same cleaning agent and cleaning procedure for many products, then requirement to a doable or achievable project. The choice of the worst-case representative shouldyou only have to do a single cleaning validation study o be based on a combination of aqueous solubility and(three runs) for the cleaning agent. potency. The potency can be determined for some products by determining the amount present in the product from the label or package insert. SometimesThis answer was provided by an Editorial Advisory this may be a little confusing for vitamin products Board Member, William E. Hall, Ph.D. Dr. Hall be because the amounts are listed in units instead of reached by phone at 910-458-5068, or by fax at 910quantitative amounts, such as milligrams. In these 458-1087, and by e-mail at
[email protected]. 14
Institute of Validation Technology
Development of Total Organic Carbon (TOC) Analysis for Detergent Residue Verification By James G. Jin and Cheryl Woodward Boehringer Ingelheim Pharmaceuticals, Inc.
v
T
he 1993 FDA Guideline for concluded that the visual detection cleaning validation states of foam was the best method for the }…the that the removal of deterdetergents they tested.4 The method gent residues should be evaluated biotechnology and of visual detection of foam is only and there should be no or very low effective for foaming detergents, pharmaceutidetergent levels left after cleaning.1 but is invalid for low foaming detercal industry has gents. From a user’s point of view, Currently, the pharmaceutical industry employs varieties of detergents this paper documents that TOC is an become for cleaning and different cleaning effective and quantitative method increasingly validation programs. Many comfor detergent residue verification. panies have not included detergent interested in residue evaluation as part of their Total Organic Carbon the use of Methodology cleaning validation programs mainly due to unavailability of effective TOC [Total methodologies or lack of awareness TOC is a non-specific method for Organic Carbon] the compound analyzed. However, of the requirement by management. In the late 1970s, Total Organic analysis is sensitive to very as an analytical TOC low levels of 0.002-0.8 ppm carbon, Carbon (TOC) analysis had been tool in cleaning depending on whether the sample is used for monitoring water quality in pharmaceuticals and environmental a water sample or a swab sample. validation controls. More recently, the biotechCurrently, two major oxidation techprograms.~ nology and pharmaceutical industry nologies dominate the TOC market: combustion and Ultra Violet (UV)/ has become increasingly interested persulfate. There has been debate in the use of TOC as an analytical tool in cleaning validation programs. TOC analy- about which technique is better suited for TOC testing sis has been used as an analytical tool for cleaning since the late 1980s. The major differences for each 2,3 validation in the biotechnology industry for years. technique5 are described in Figure 1,and give the user Westman and Karlson recently conducted acompari- appropriate information to make an informed decison study for different analytical methods – visual sion as to which technique better serves their needs. detection of foam, pH, conductivity measurements, The best TOC oxidation technology is the one and TOC for detergent residue evaluation. They that meets the application and analytical needs of the Special Edition: Cleaning Validation III
15
James G. Jin
Figure 1
tp f t og cb t od Combustion Combustion UV/Persulfate
D t Thermal Conductivity Detector (TCD) Coulometric Non-Dispersive Infrared Detector (NDIR)
a rg (toc) 0.5 – 100% 1 – 100% 0.002 – 10,000 mg/L
off md AOAC 955.07 ASTM D4129 USP 643
Heated Persulfate
NDIR
0.002 to 1,000 mg/L
USP 643
Combustion
NDIR
0.004 – 25,000 mg/L
USP 643
UV/Persulfate
Membrane/Conductivity
0.0005 – 50 mg/L
USP 643
UV
Conductivity or NDIR
0.0005 – 0.5 mg/L
USP 643
user’s situation. The UV/Persulfate method meets precision and accuracy requirements for low-level calibration check standards such as 0.5 ppm carbon in detergent residue evaluation. However, if capturing the particulate organic matter in the TOC value is important, then combustion would be the better oxidation technology. The instrument we chose is a Tekmar-Dohrmann Phoenix 8000 with the UV/Persulfate oxidation technique.
degradation of all carbon species to carbon dioxide, water, and other oxides of heteroelements. The UV light alone induces breakdown of many carbon species with the persulfate providing additional help to attack compounds difficult to oxidize. The radical reactions are aggressive and indiscriminate in their attack. S O -2 → SO -1 + R → H O + CO 2
8
4
2
2
Chemistry of Oxidation and Total Organic The NDIR is constructed in such a way as to be Carbon Analysis of UV/Persulfate sensitive and selective for carbon dioxide present Wet chemistry oxidation of carbon compounds utilizes two chemical reactions to complete the analysis. A 21 percent solution of phosphoric acid is utilized in converting inorganic carbon species. Acidification of the sample allows for attack on inorganic species such as carbonates and bicarbonates to convert them to carbon dioxide. This, along with any dissolved carbon dioxide in the sample is then sparged out, and either exhausted to vent or routed to the Non-Dispersive Infrared detection (NDIR) for quantification when analyzing for Inorganic Carbon (IC) or TOC by difference (TC-IC).
in the gas flow. An infrared beam from the source is passed through a chopper and down the sample chamber to a dual chamber detector. Each chamber is filled with carbon dioxide and is separated by a thin membrane. Varying intensity of the light hitting the cell causes fluctuation in temperature and thus the pressure of the gas inside the detector. This causes the membrane to deflect, which is ultimately read as a millivolt output signal from the detector.
Detergent Evaluation
Three detergents (CIP-100, CIP-200, and Sparquat 256) were tested both in-house using the Tekmar H+ + CO -2 → H O + CO Dohrmann Phoenix 8000 TOC Analyzer and at a contract lab, Quantitative Technologies Inc. (QTI), Persulfate is used to do the rest of the oxidation to verify the total amount of organic carbon in each chemistry that is required for analysis. Sodium persul- detergent at its srcinal concentration. The method fate, at a concentration of 10 percent, and phosphoric and instrument used at QTI was a Perkin-Elmer CHN acid, five percent are added to the UV chamber for Analyzer 2400. This experiment was performed to analysis. The persulfate species in the presence of make a comparison between our instrument and the UV light breaks down at a weak oxygen-oxygen instrument in a qualified contract laboratory for inforbond yielding two radicals per molecule. These radi- mation purposes only. One detergent (Chlor-Mate) cals start chain reactions that ultimately lead to the was tested in-house and compared with the available 3
16
2
2
Institute of Validation Technology
James G. Jin
vendor’s specification. The TOC results for all the detergents are shown inFigure 2. The differences between the in-house and QTI results with respect to the TOC assay for CIP-100 and CIP-200 are 5.0 percent and 9.6 percent, respectively. These differences are relatively low compared to the 20 percent recovery criteria during recovery studies. The difference between the in-house and QTI results with respect to the TOC assay for Sparquat 256 is 28.4 percent. The in-house result was reviewed and no error was noted in the performance of the test-
and TX2418 show acceptable results with respect to result consistency. The average of the seven TOC results from TX2412 and TX2418 found inFigure 3 is 0.8327 ± 0.1860 ppm carbon. The variation is acceptable compared to the acceptance criterion of three ppm carbon. These two swabs with the same material were selected to be our TOC swabs (cut to 5x5 cm2) for detergent residue verification. The TX3340 TOC cleaning validation kit including Eagle EP Picher 03464-40mL clear vials, Tex wipe® TM TX714L-large SnapSwabs , and blank vial labels
ing procedure. The major differences may be due to may be chosen since it is specially de signed for TOC instrument and testing method variations. The result swabbing purposes. for Chlor-Mate is within the vendor’s specification. Detergent Recovery Evaluation from Stainless Steel Surface
Swab Selection It has been known for years that polyester is a suitable material for TOC swabbing analysis. Over 20 different kinds of polyester swab samples were received from The Texwipe Company LLC. Five of them were chosen for TOC evaluation based on sample design and the convenience for use. The purpose of this experiment was to select a type of swab that has little TOC background interference and with consistent TOC results over time. Ultra purified water with 0.05 to 0.08 ppm carbon was used for swab analysis. The TOC results obtained from our TOC analyzer are shown in Figure 3. Swabs TX761 and TX741A showed increasing TOC results from 0.0813 to 0.9692 ppm carbon and from 0.1724 to 1.1246 ppm carbon over five days, respectively. Swab TX700 showed an unacceptably high TOC result of 46.1991 ppm carbon at the beginning of the experiment, and was therefore not tested further. None of these swabs are suitable for our TOC analysis. Both polyester wipers AlphaSorb® HC TX2412
Ten stainless steel templates were spiked with detergent solution and swabbed using the polyester wipers AlphaSorb® HC TX2418 (5x5 cm2) for the detergent recovery study. The spiking and swabbing procedures were the same as those used for drug substance recovery studies. Forty mL of ultra purified water was added to each test tube as the extraction solution, vortexed about one minute, and then sonicated for five minutes for testing. The results are shown in Figure 4. The recoveries for CIP-100, CIP-200, and ChlorMate are over 80 percent and no correction factor is necessary. For Sparquat 256, a correction factor of 0.61 will be used. For example, if a result of 0.5 ppm carbon is obtained from the TOC analyzer, the final reported result would be 0.82 (0.5 ÷ 0.61) ppm carbon. Detergent Recovery Evaluation from Non-Stainless Steel Surfaces
The aforementioned study was repeated using non-stainless steel templates. Two or three non-stain-
Figure 2
t og cb r f Dg e Dg
mf/l
idf CIP-100
Vestal Convac lot 211097
F BiPi* 4.0208 ± 0.0139%
CIP-200
Convac lot 213915
2.4986 ± 0.0114%
Sparquat 256
ISSA (lot: n/a)
Chlor-Mate
WestAgro lot®J8G0489AR
t og cb r
14.0232 ± 0.9336% 1.29% ± 0.0086%
toc r F qti/vd 4.22% 2.26% 18.0% 1 – 1.5%
*Boehringer Ingelheim Pharmaceuticals, Inc.
Special Edition: Cleaning Validation III
17
James G. Jin
Figure 3
t og cb r (pp c) f sb s swb Dep
toc/tw hu h o 2
toc/Fu hu h o 2
Polyester Alpha 0.0813 swab TX761 ± 0.0041 Polyester Alpha 0.1724 swab TX741 A ± 0.0144 Polyester wipers 1.1665 ® AlphaSorb ± 0.0406 HC TX2412
0.3221 ± 0.0853 0.2509 ± 0.0068 0.6091 ± 0.0490
Polyester 0.7406 ® AlphaSorbwipers ± 0.0056 HC TX2418 Polyester Alpha 46.1991 swab TX700 ± 8.0761
0.7269 ± 0.0297
toc/oe Dy h o
toc/tw Dy h o
2
0.3926 ± 0.0166 0.5330 ± 0.0250 0.8602 ± 0.0264
2
0.9410 ± 0.0288 0.8091 ± 0.0200 0.7535 ± 0.0328
(1)
N/A
toc/Fe Dy h o 2
0.9692 ± 0.0299 1.1246 ± 0.0394 0.9723 ± 0.0668
(1)
(1)
N/A
N/A
N/A
N/A
N/A
N/A
® 1. Polyester wipers AlphaSorb® HC TX2412 and polyester wipers AlphaSorb HC. TX2418 is same material cut to different sizes.
less steel templates were spiked with each detergent Figure 4 solution and swabbed using the polyester wipers t og cb r AlphaSorb® HC TX2418 (5x5 cm2). The results are r f s shown in Figure 5. For CIP-100 and CIP-200, the recoveries from s sf each non-metal surface are over 80 percent. There- Dg P nb P fore, no correction factor is needed with respect to r f r sp sdd the TOC recovery. For Sparquat 256, the recoveries D vary with different surfaces. The correction factors are as follows: CIP-100 111.7 30 5.92 For Delrin surface: For Glass surface: For Nylon surface: For Lexan surface:
correction factor = 0.74 correction factor = 0.75 correction factor = 0.43 correction factor = 1.0
Evaluation of Detergent Residue After Rinsing
The purpose of this experiment was to evaluate:
∂The suitability of the Acceptance Criterion (AC) of three ppm carbon ∑The effect of detergent concentration on detergent residue after rinsing ∏Recovery of detergent from different surfaces with and without rinsing πRinsing efficiency and rinse time Four detergents (CIP-100, CIP-200, Sparquat 256, and Chlor-Mate) were used in both a concentrated form and at a working concentration of 0.5 oz/gal. Approximately one mL of detergent solution 18
Institute of Validation Technology
CIP-200 Sparquat 256
92.4 61.0
10 20
4.10 8.47
Chlor-Mate
99.1
10
2.76
Note: Results were automatically corrected for the instrument blank effect.
Figure 5
t og cb r r f n-s s sf Deege lex De G ny sufe Pee Pee Pee Pee reey reey reey reey CIP-100
106.9
113.8
107.6
127.0
CIP-200
90.3
92.3
97.4
93.2
Sparquat 256
83.3
74.0
75.1
42.5
was pipetted and spiked onto the templates with different materials of construction and dried with ventilation under a hood in the research and devel-
James G. Jin
opment manufacturing area for a minimum of four hours. The templates were swabbed per standard swabbing procedure either before or after rinsing, using the polyester wipers AlphaSorb® HC TX2412 cut to 5x5 cm2. The rinse was first conducted using tap water and then purified water United States Pharmacopoeia (USP), both at room temperature and with a slow flow rate of approximately 2.7 L/ min. Two different rinse times (30 seconds and 60 seconds) were evaluated for different detergents on different templates to simulate the final rinse step in
The Tekmar Dohrmann Phoenix 8000 TOC analyzer was easily able to detect the non-rinse samples with the results of 3.911 ppm carbon, 2.0928 ppm carbon, and 10.0868 ppm carbon for CIP-100, CIP200, and Sparquat 256, respectively. The results indicate that the AC of three ppm carbon is still high for detergents CIP-100, CIP-200, and Sparquat 256. The AC of one ppm carbon is acceptable. There were no differences in detectable residue for all four detergents (both concentrated and at 0.5 oz/gal) on stainless steel after a 30-second tap water rinse fol-
our manual cleaning process. The recovery results are reported in Figure 6.
lowed by a 30-second purified water, USP rinse. Delrin was chosen for a typical material of construc-
Figure 6
t og cb r Dg rd b rg smpe idef
ce
tempe
CIP-100 CIP-100 CIP-100 CIP-100 CIP-100 CIP-100 CIP-100 CIP-100
0.5 oz/gal 0.5 oz/gal Concentrated 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal
SS a SS a SS a Delrin Delrin Nylon Glass Lexan
No rinse 30”/30” 30”/30” 30”/30” 60”/60” 30”/30” 30”/30” 30”/30”
CIP-200 CIP-200 CIP-200 CIP-200 CIP-200 CIP-200 CIP-200 CIP-200
0.5 oz/gal 0.5 oz/gal Concentrated 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal
SS a SS a SS a Delrin Delrin Nylon Glass Lexan
No rinse 30”/30” 30”/30” 30”/30” 60”/60” 30”/30” 30”/30” 30”/30”
0.5 oz/gal 0.5 oz/gal Concentrated 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal 0.5 oz/gal
SS a SS a SS a Delrin Delrin Nylon Glass Lexan
No rinse 30”/30” 30”/30” 30”/30” 60”/60” 30”/30” 30”/30” 30”/30”
0.5 oz/gal Concentrated
SS SS
Sparquat Sparquat Sparquat Sparquat Sparquat Sparquat Sparquat Sparquat
256 256 256 256 256 256 256 256
Chlor-Mate Chlor-Mate Notes:
a a
re tme
30”/30” 30”/30”
b b b b b b b
b b b b b b b
b b b b b b b
b b
ae swbbed 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm
2
100 cm 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm
2
100 cm 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm 100 cm
2
100 cm 100 cm
2
2 2 2 2 2 2 2
2 2 2 2 2 2 2
2 2 2 2 2 2 2
2
toc reu (ppm c)d 3.9111 Less than blank Less than blank Less than blank Less than blank 0.6682 0.0001 Less than blank 2.0928 Less than blank Less than blank Less than blank Less than blank 0.7720 0.0133 Less than blank 10.0868 c 0.2693 c Less than blank Less than blank Less than blank 0.3866 c Less than blank Less than blank Less than blank Less than blank
a. Stainless steel. b. 30”/30” or 60”/60” – rinse time in seconds, tap water/purified water United States Pharmacopoeia (USP). c. Result without correction factor.
Special Edition: Cleaning Validation III
19
James G. Jin
tion and 30/60 seconds were chosen for evaluation of the rinse time. There was no difference in detectable residue for CIP-100, CIP-200, and Sparquat 256 on the Delrin surface after 30-second and 60-second rinse times. The results also show that it is more difficult to remove residues of CIP-100, CIP-200, and Sparquat 256 from a Nylon surface than from other materials.
2 = 0.79 mL of Chlor-Mate to be left on 100 cm of equipment surface after cleaning, respectively.
allowed to be left on GMP equipment surfaces. In our detergent residue verification program, the AC for each detergent residue left on equipment surfaces depends on the sensitivity of the instrument used for analysis. This means we must set a low AC that isstill quantifiable and applicable. Toxicity of the detergent is not a concern at these trace amounts de tergent level. Effects on human health from sidue re left on equipment surfaces should be insignificant at a low concentration such as 0.5 oz/gal and with a routine rinse procedure. Our objective in this program is to demonstrate that we are able to verify whether or not the detergent residues are removed to an acceptable low-level we can achieve. Therefore, the AC should be established as close to the instrument’s level of detection as possible. We tighten the initial limit of three ppm carbon to AC = 1.0 ppm carbon (net reading automatically corrected with blank by the instrument in a 40 mL solution), which is less than two times the blank baseline. The AC can also be expressed as AC ≤ 10 ppb carbon/ cm2. This AC is practical and verifiable. The significance of the 1.0 ppm carbon AC for each detergent can be explained inFigure 7. We can see from the above calculations that AC = 1.0 ppm carbon means, for all detergents at 0.5 oz/gal, that we allow the maximum of 1 ÷ 3.92 = 0.26 mL of CIP-100, 1 ÷ 2.44 = 0.41 mL of CIP-200, 1 ÷ 13.68 = 0.07 mL of Sparquat 256, and 1 ÷ 1.26
results shown in Figure 6 that all the residues are easily removed by a 30-second tap water rinse followed by a 30-second purified water, USP rinse with very low spray rate. Verification rather than validation is currently required by the 1993 FDA,Guide to Inspections of Validation of Cleaning Proceduresdue to the fact that detergent residue is less significant than drug substance residue left after cleaning.
Detergent Residue Verification Program
Our detergent verification program is designed to be a one-time verification for each detergent used. This was based on the rinse experiment and the assumption that our routine rinsing procedures performed by well trained operators are sufficient to Acceptance Criterion for Detergent Residue remove detergent residues to the level of less than There is no universal AC for detergent residue the AC. This assumption has been verified from the
Summary The detergent residue verification program has been successfully established using the Tekmar Dohrmann Phoenix 8000 TOC analyzer. This paper has shown the program development, and presents critical data to support the detergent verification reports for each detergent used. The instrument Installation Qualification (IQ), Operational Qualification (OQ), system calibration, and the TOC analysis method development were performed but not discussed in this paper. The polyester wipers AlphaSorb® HC TX2412 and TX2418 cut to 5x5 cm2 have been selected as the swabs for sampling detergent residue from equipment surface for TOC analysis. The AC for the detergents CIP100, CIP-200, Sparquat 256, and Chlor-Mate with respect to TOC has been established as AC ≤ 10 ppb carbon/cm2. Two different rinse times, 30 seconds and 60 seconds, were evaluated. The results show
Figure 7
sgf f t og cb r f Dg 0.5 /g ciP-100 1 mL at 0.5 oz/gal diluted to 40 mL 1.0 ppm C per 100 cm2 corresponding to 20
ciP-200 3.92 ppm
0.26 mL
Institute of Validation Technology
spqu 256 2.44 ppm
0.41 mL
13.68 ppm 0.07 mL
c-me 1.26 ppm 0.79 mL
James G. Jin
that 30-second/30-second rinse time (30-second rinse with tap water and then 30-second rinse with purified water, USP) is sufficient to remove the detergent residues from different material templates including stainless steel, Delrin, Glass, Nylon, and Lexan to a level below the AC. The correction fac tors were determined based on the results of the recovery studies and will be used by analytical sciences to report the final o TOC results for the detergent residue verification.
About the Authors James G. Jin is Chairman of the Cleaning Validation Committee for Boehringer Ingelheim Pharma ceuticals, Inc., which is responsible for cleaning validation program development and implementation. He has more than ten years experience in pharmaceutical science and business arenas. He can be reached by phone at 203-798-5309. Cheryl Woodward is Associate Director of Research and Development (R&D) Manufacturing, for Boehringer Ingelheim Pharmaceuticals, Inc. She is responsible for all aspects of GMP manufacturing for clinical supplies and has over 18 years experience in the pharmaceutical and related industries. She can be reached by phone at 203-798-5367.
References
1. FDA. Guide to Inspections of Validation of Cleaning Procedures. July, 1993. 2. Jenkins K.M., Vanderwielen A.J, Armstrong J.A, Leonard L.M, Murphy G.P, Piros N.A. 1996. “Application of Total Organic Carbon Analysis to Cleaning Validation.” PDA.Journal of Pharmaceutical Science and Technology. 50. Pp 6-15. 3. Guazzaroni M., YiinB., Yu J., 1998. “Application ofTotal Organic Carbon Analysis for Cleaning Validation in Pharma ceutical Manufacturing.”American Biotechnology Laboratory. September. Pp. 66-67. 4. Westman L., Karlsson G., 2000. “Methods for Detecting Residues of Cleaning Agents During Cleaning Validation.” Research Article, Vol. 54, No. 5. September/October. 5. Furlong J., Booth B., Wallace B. 1999. “Selection of a TOC Analyzer: Analytical Considerations.” Tekmar-Dohrmann Application Note. Vol. 9.20.
Special Edition: Cleaning Validation III
21
Total Organic Carbon Analysis for Cleaning Validation in Pharmaceutical Manufacturing By Karen A. Clark Anatel Corporation
v
I
n the pharmaceutical industry, specific methods like TOC is that Good Manufacturing Practice they cannot identify exactly what }TOC analysis (GMP) requires that the cleanthe residue material is. Depending can be adapted on the chosen cleaning process and ing of drug manufacturing equipment be validated.1 Many different established acceptance limits, a to any drug validation techniques can demonnon-specific method may be all that strate that the manufacturing equipcompound or is needed to validate the process. ment is cleaned and essentially free TOC analysis can be adapted cleaning agent from residual active drug substancto any drug compound or cleanes and all cleaning agents. ing agent that contains carbon and that contains Common analytical techniques is “adequately” soluble in water. carbon and is in the validation process include Studies have been conducted to High Performance Liquid Chromdemonstrate that TOC methods can ‘adequately’ atography (HPLC), spectrophotomalso be applied to carbon containing etry Ultraviolet/Visible (UV/Vis) soluble in water.~ compounds that have limited water and Total Organic Carbon (TOC). solubility, and recovery results are HPLC and UV/Vis are classified as equal to those achieved by HPLC.6 specific methods that identify and measure appropri- TOC methods are sensitive to the parts per billion ate active substances. TOC is classified as a non- (ppb) range and are less time consuming than HPLC specific method and is ideal for detecting all carbon- or UV/Vis. United States Pharmacopoeia (USP) containing compounds, including active species, TOC methods are standard for Water-for-Injection excipients, and cleaning agent(s).2,3,4,5 and Purified Water,7 and simple modifications of The disadvantage of specific methods, particular- these methods can be used for cleaning validation. ly HPLC, is that a new procedure must be developed for every manufactured active drug substance. This Methodology development process can be very time consuming and tedious, plus important sampling issues must TOC analysis involves the oxidation of carbon and alsoperformed be considered. In addition, HPLC analyses be in a relatively short time period must after sampling to avoid any chemical deterioration of the active substance. Finally, the sensitivity of HPLC methods can be limited by the presence of degradation products. Of course the disadvantage to non22
Institute of Validation Technology
the of oxidation the resulting carbon exist, dioxide. A number detection of different techniques including photocatalytic oxidation, chemical oxidation, and high-temperature combustion. In this study, an Anatel A-2000 Wide-Range TOC Analyzer, equipped with an autosampler, was used. The Anatel A-2000 Wide-
Karen A. Clark
Range Analyzer measures TOC in accordance with American Society for Testing and Materials (ASTM) methods D 4779-88 and D 4839-88. It measures TOC directly by adding phosphoric acid to the water sample to reduce the pH from approximately two to three. At this low pH any inorganic carbon that is present is liberated as COinto a nitrogen carrier gas and is directly measured by a non-dispersive infrared (NDIR) detector. Any remaining carbon in the sample is assumed to be TOC. A sodium persulfate oxidant is then added to the sample, and in the presence of UV 2
radiation, the remaining carbon is oxidized to CO . The amount of CO generated is then measured by the NDIR to determine the amount of TOC srcinally present in the water. For equipment cleaning validation there are two types of TOC sampling techniques. One is the direct surface sampling of the equipment using a swab. The second consists of a final rinse of the equipment with high-purity water (typically <500 ppb TOC) and collecting a sample of the rinse for analysis. In general, direct surface sampling indicates how clean the actual surface is. This study demonstrates how to develop and validate a TOC method to measure a variety of different organic residues on stainless steel surfaces. Performance parameters tested include linearity, method detection limit (MDL), limit of quantitation (LOQ), accuracy, precision, and swab recovery. 2
2
Linearity TOC analysis should provide a linear relationship between the measured compound concentration and the TOC response of the analyzer. We evaluated four different types of cleaning agents for linearity:
∂ CIP-100® (alkaline) ∑ CIP-200® (acidic) ∏ Alconox® (emulsifier) π Triton-X 100 (wetting agent) Results are shown in Figures 1-4. Correlation coefficients ranged from 0.9787 to 0.9998. Alconox and Triton-X 100 have a tendency to foam, depending on the concentrations that are analyzed and this foaming phenomena can have a negative effect on the accuracy of the TOC result (reduced R 2). Three
Figure 1
l f ciP-100 9000 ) 8000 b p 7000 (p C 6000 O 5000 T d 4000 e r u 3000 s a e 2000 M 1000 0 0
y=39.254x + 1.462 R2=0.9997
50
100 150 200 250 CIP 100 Concentration (ppm)
Figure 2
l f ciP- 200 9000 y=19.132x + 51.042 ) 8000 R2=0.9998 b p 7000 (p C 6000 O 5000 T d 4000 e r u s 3000 a e M 2000 1000 0 0 100 200 300 400 500 CIP 200 Concentration (ppm)
Figure 3
l f a ) m p p ( C O T d re u s a e M
45 y=0.0355x + 1.1983 40 R2=0.9787 35 30 25 20 15 10 5 0 0 200 400 600 800 1000 Alconox Concentration (ppm)
Special Edition: Cleaning Validation III
23
Karen A. Clark
Figure 4
Figure 7
l f t-x 100 12500 ) b p10000 (p C 7500 O T d e r 5000 u s a e 2500
y=415.76x + 16.997 R2=0.9982
M
0 0
5
10 15 20 25 Triton-X 100 Concentration (ppm)
Figure 5
l f s 12000 ) b10000 p p ( C 8000 O T d 6000 e r u 4000 s a e M 2000
y=1.003x + 45.185 R2=0.9996
Figure 6
l f v ) b p p ( 6000 C O T 4000 d e r u s a 2000 e M
y=0.8758x + 62.133 R2=0.9998
0 0
24
8000 ) 7000 b p (p 6000 C O 5000 T d 4000 e r u 3000 s a e 2000 M 1000
y=0.9287x + 30.8 R2=0.9998
0 0
2000 4000 6000 8000 Endotoxin Concentration (ppb)
representative examples of active substances were also tested for linearity: an excipient (sucrose), an antibiotic (vancomycin), and endotoxin. Results are shown in Figures 5-7. All three compounds demonstrated excellent linearity with correlation coefficients (R2) ranging from 0.9996 to 0.9998.
Method Detection Limit and Limit of Quantitation
0 0 2000 4000 6000 8000 10000 12000 Sucrose Concentration (ppb)
8000
l f ed
2000 4000 6000 8000 Vancomycin Concentration (ppb)
Institute of Validation Technology
We determined the Method Detection Limit (MDL) by measuring the TOC response of the method blank. A method blank consists of the sampling vial, swab, and recovery solution. In this study, the recovery solution was low TOC (< 25 ppb) water. Ten pre-cleaned vials were filled with the low TOC water. One swab was placed in each vial (Texwipe Alpha Swab TX761; tips cut off). Solutions were vortexed and allowed to stand for one hour prior to analysis. Four replicates from each vial were analyzed. The four replicates from each of the ten blank vials were averaged. These ten values were averaged again and a standard deviation was calculated. The standard deviation was multiplied by the Student t number for n-1 degrees of freedom (3.25 for n=10), at 99% confidence levels to determine the method detection limit. The MDL was calculated to be 50 ppb. The Limit of Quantitation (LOQ) was calculated by multiplying the MDL by three. A value of 150 ppb was obtained (see Figure 8).
Precision and Accuracy
Karen A. Clark
Figure 8
cd toc ag f 10 Bk v v nb 1 2 3 4 5 6
ag toc (ppb) 58 72 75 93 79
To demonstrate the precision and accuracy for this TOC method, a representative solution of CIP-100 as 1000 ppb, or oneppm as carbon, was analyzed sequentially ten times. This carbon concentration was chosen to evaluate these method parameters because, in general, TOC residual limits are typically around one ppm. Results are listed inFigure 9. At this TOC level, the precision was ± 1% and the accuracy was ± 5%.
102 7
60
8 9
83 67
10
Swab Recovery
54 74.3 Average 15.5 Standard Deviation 50 ppb MDL (Student t, n=10) 151 ppb LOQ
Figure 9
cd a d P f 10 rp f 1pp ciP100 s cb v nb
md toc (ppb)
1 1 1
1041 1025 1039
1 1 2 2 2 2 2
1057 1054 1034 1042 1048 1054 1055 Average Standard Deviation % CV (precision) % Recovery based on 1 ppm C (accuracy)
Stainless steel plates were used in the swab recovery test to simulate manufacturing equipment. One side of each plate was spiked with a solution of active substance or cleaning agent. The plates were allowed to completely dry overnight at room temperature. A Texwipe alpha swab TX761 was moistened with low TOC (< 25 ppb) water and the spiked plate surface was swabbed both vertically and horizontally. The swab end was cut off, placed into a vial to which we added 40-mL of low TOC water. The vial was capped tight, vortexed, and allowed to stand for one hour prior to analysis. The same volume of each solution that was spiked onto the plates was separately spiked directly into 40-mL of low TOC water and analyzed. The percent recoveries of the different substances are listed in Figure 10. Reported values are the average of three individual swab samples for each substance. The swab recoveries varied between 79.3% to 95.9%
Conclusion 1045 10.5 1.0% 105%
This study demonstrates that TOC analysis is suitable for measuring organic residues on stainless steel surfaces, and that it is a reliable method for cleaning validation as demonstrated by surface residue recoveries of 79%-96%. This methodology
Figure 10
rp ep f sb r f cg ag d a sb sb CIP-100 Sucrose
pp c f spk spkd sdd s pp c fP 1710 94.5 2663 2112 79.3 1810
% r
% rsD
1.8 4.9
Vancomycin
661
634
95.9
3.0
Endotoxin
902
736
80.0
2.8
Special Edition: Cleaning Validation III
25
Karen A. Clark
shows that low limits of detection, excellent linearity, precision, and accuracy can be obtained. All of these TOC results, with the exception of Alconox and Triton-X 100, were generated using the same TOC method, making TOC analysis a low cost and less time consuming alternative for cleaning validation. o
About the Author Karen A. Clark is a Product Manager at Anatel Corporation. She has over 15 years experience in the pharmaceutical/biotechnology industry focusing on drug formulations, analytical methods development and validation, and GLP/GMP laboratory management. Clark holds a B.S. in Biochemistry from Millersville University and an M.S. in Chemical Engineering from the University of Colorado. She can be reached by e-mail at
[email protected] or at Anatel Corporation, 2200 Central Avenue, Boulder, CO 80301.
References 1. FDA. Current Good Manufacturing Practice Regulations, 21 CFR 211.220. 2. Baffi, R. et al. 1991. “A Total Organic Carbon Analysis Method for Validating Cleaning Between Products in Bio pharmaceutical Manufacturing.” Journal of Parenteral Science and Technology 45, no. 1: 13-9. 3.
Jenkins, et al. 1996. “Application ofTotal Organic Carbon PDA Journal of PharmAnalysis K.toM.Cleaning Validation.” aceutical Science and Technology 50, no. 1: 6-15. 4. Strege, M.A. et al. 1996. “Total OrganicCarbon Analysis ofSwab Samples for the Cleaning Validation of Bioprocess mentation Fer Equipment.”BioPharm (April). 5. Guazzaroni, M.et al. 1998. “Application ofTotal OrganicCarbon Analysis for Cleaning Validation in Pharmaceutical Manufacturing.” American Biotechnology Laboratory 16, no. 10 (September). 6. Walsh, A. 1999. “Using TOC Analysis for Cleaning Validation.” Presented at The Validation Council’s Conference on Cleaning Validation, 27 October, Princeton, New Jersey. 7. USP 23, Fifth Supplement, 15 November 1996.
26
Institute of Validation Technology
Advertisement
What are you are you working on today?
We’d like like to hear about it… The Journal of Validation Technology and the Journal of GXP Compliance are looking for authors to submit case studies on contemporary validation and compliance topics and issues. If you would like to submit an article for consideration or request a copy of editorial guidelines, please fill out this form and fax it to:
Institute of Validation Technology – Editorial Department. Phone: (218) 740-7028 • Editorial Department Fax: (218) 740-6308. n: ___________________________________________________________________ til: _____________________________________________________________________ copy: ________________________________________________________________ addr: _________________________________________________________________ ciy: ____________________________________________________________________ s: ____________________________________________________________________ Po Bx 6004 Dh, mn 55806 Ph: 218-740-7028 u.s. oy: 888-524-9922 Fx: 218-740-6308 e-m: f@h. wb s: .h.
zip: ______________________________________________________________________ coury: ________________________________ tlpho:_______________________ Fx: _____________________________________________________________________ e-mil: ___________________________________________________________________ topi: ____________________________________________________________________
Special Edition: Cleaning Validation III
27
Detergent Selection – A First Critical Step in Developing a Validated Cleaning Program By Mark Altier Ecolab, Inc.
v
T
he FDA recognizes the importance of effective cleaning and sanitizing protocols as a proactive measure in preventing cross-contamination in the pharmaceutical and cosmetic industries: 21CFR 211.67: “Equipment and utensils shall be cleaned, maintained, and sanitized at appropriate intervals to prevent malfunctions or contamination that would alter the safety, identity, strength, quality, or purity of the drug product beyond the official or other established requirements.”
}This
article will discuss the key factors that must be addressed when selecting a
28
good first step. Following this step, laboratory testing is required to determine the exact nature of the potential contaminant. Next, identification and testing of various cleaning chemistries against the potential contaminant is performed to determine which detergent type is best suited for contaminant removal. The next step is to return to the manufacturing site, test the cleaning chemistry, and optimize the program. This approach provides a sound, scientific rationale for the detergent selection and lays a firm foundation to the formal cleaning protocol, once developed. This article will discuss the key factors that must be addressed when selecting a detergent. Each factor will be discussed in detail and examples are given when appropriate. The roles
detergent. factor willEach be discussed in detail and examples are given when appropriate. The roles of laboratory testing and plant optimization are also addressed.~
In order to comply with this regulatory requirement, sound cleaning and sanitizing protocols must be developed and followed. One of the most critical components of any cleaning program is detergent selection. Different processes and potential contaminants may require different detergents that are appropriate for the application. In certain cleaning applications, a neutral foaming detergent might be appropriate, whereas in others, a non-foaming alkaline detergent is desirable. The choice of Institute of Validation Technology
detergent for a given application should be based on sound, scientific reasoning. A sound rationale for detergent selection begins at the manufacturing site, where the process and cleaning program will take place. A full evaluation of the pro cess, cleaning strategies, potential contaminant levels, and available utilities is a
Mark Altier
of laboratory testing and plant optimization are also residue types will fall into one of the following three addressed. categories: organic, inorganic, or a combination of The Five Factors for Determining these. Most potential contaminants are a combina-
Detergent Suitability Figure 1 There are five key factors that must be addressed when determining which detergent is most suitable for a cleaning application. These are:
c rd tp P id og rd
∂ Nature of the residue (or potential contami- Eudragit nant) Acetaminophen ∑ Surface to be cleaned ∏ Method of application π Role of water ∫ Environmental factors
Carbopols
ig rd Titanium Dioxide Zinc Oxide Iron Oxide
Albuterol Sulfate
Calcium Carbonate
Neomycin Sulfate
Inorganic Salts
Water/Oil – Oil/Water
Silicon Dioxide
All five of these factors must be addressed when Emulsions developing a cleaning program. Failure to address Glyburide any of these issues in sufficient detail can result in a less than desirable cleaning program and could place the successful completion of the cleaning validation tion of organic and inorganic components. Com mon at serious risk. residue types in the pharmaceutical industry are given in Figure 1. The Nature of the Residue A number of powerful analytical instruments are A residue can be defined as any unwanted matter available that can provide tremendous insight into or potential contaminant on the surface of the ject ob the nature and composition of almost any unknown or equipment being cleaned. Oftentimes, what is potential contaminant type. Some of the more useful referred to as a “residue,” is in fact a finished prod- tools include: uct, drug active, or other component that is produced using the process equipment that is being cleaned. • Fourier Transform Infrared Spectroscopy The terms “residue,” “contaminant,” and “potential (FTIR) contaminant” will be used interchangeably through• Energy Dispersive X-Ray Spectroscopy (EDS) out this article. • Scanning Electron Microscopy (SEM) Determination of the nature of a residue is a funda• Compound microscopic imaging mental component in the development of any clean ing • Nuclear Magnetic Resonance imaging (NMR) program. In some cases, the exact nature and com• Inductively Coupled Plasma detector (ICP) position of a residue is known. For example, if the • Atomic Absorption Analyzer (AA) residue is a finished product, the exact composition and physical properties are almost always known. Often, a combination of two or more of these tools However, the identity and nature of the sidue re may is required to provide a full picture of a potential be completely unknown if the re sidue is composed contaminant in question. For example,Figure 2 and of an intermediate, byproduct, or result of thermal, Figure 3 are typical images generated to help charchemical, or other degradation of a previously known acterize unknown potential contaminant samples. substance. This type of analysis is invaluable in determining the The nature of the potential contaminant plays a exact residue type and breakdown of the organic and central role in determining what type of detergent inorganic portions of a residue. is most appropriate for the application. Individual Figure 2 is an FTIR image of an unknown re sidue. residues require different detergent chemistries. All This characterizes and gives a general breakdown of Special Edition: Cleaning Validation III
29
Mark Altier
Figure 2
Ftir s f uk sp. t a id P f ak d ad P cp d Pb ig c .35 2 .2 1 3 6 1
.30 .25 e c n a rb o s b A
5 .7 5 4 5 1
6 6 . 4 5 4 1
7 0 . 6 9 3 1
4 .0 1 1 3 1
.20 6 .8 8 5 9 2
.15 .10
1 7 . 3 3 9 3
.05
8 .3 6 2 9 2
6 8 . 4 5 8 2
5 .5 7 8 2 2
6 .6 7 5 2 2
6 .8 1 1 2 2
9 9 . 5 6 1 2
8 8 . 5 5 1 2
7 9 . 3 3 0 2
6 2 . 1 5 2 1
4 6 . 4 4 0 0 1 1
1 .5 7 7
6 5 4 . 0 8 9
4 2 .9 4 9 6 4 0 2 . 0 7 8
7 0 . 3 1 0 2
0 3500
3000
2500 2000 Wave Number (cm-1)
1500
1000
Figure 3
eDs s f uk sp t a cf P f ig cp s s, a, d i, add og cpd
s t n u o C
600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0
C
0.000
30
Fe Si
Al O Fe
1.000
Mg
Fe
2.000
Institute of Validation Technology
3.000
4.000 Key
5.000
6.000
7.000
8.000
Mark Altier
the organic portion of the residue. FTIR imaging gives valuable insight into the functional groups that may be present in the organic component of a sidue. re Figure 3 confirms the presence of inorganic ma terial and identifies the specific inorganic components present in an unknown sample. This information is useful when determining which chelant or surfactant family is most suitable for re moving or tying up the free metal ions and other inorganic material. Combined, FTIR and EDS imaging can give a complete picture of most unknown residues. These
may tolerate high pH, but may not tolerate chlorine or chlorides. It is important to have a clear understanding of how the substrate being cleaned will interact with the detergent system, otherwise serious damage to equipment surfaces can result. A SEM image, shown in Figure 4, is a stainless steel surface that has been pitted by using an in compatible detergent. The prospective customer in this case felt that the residue was becoming more tenacious with time and was using higher detergent concentrations to remove the residue.
analyses provide the information needed to select a group of detergent chemistries that are formulated and known to be effective against the residue type.
A close look at the surface revealed that the surface was actually being pitted by the detergent, providing microscopic crevices where the residue was able to harbor during the cleaning cycle. This problem was Surfaces To Be Cleaned aggravated by the fact that the customer continued to Different substrates (i.e., productcontact surfaces, increase the detergent concentration, which accelersuch as stainless steel, glass, or plastic) will interact ated the rate and degree of corrosion, and provided differently with the contaminant and the de tergent the residue with even more locations to harbor during system. Some materials, such as glass, and alumi num, the cleaning cycle. are not tolerant to high pH systems. Other substrates These images clearly demonstrate the problems
Figure 4
mp c f s s sf cd b ipp Dg s. i s Bd rg 1000 t mgf
Special Edition: Cleaning Validation III
31
Mark Altier
that can be caused by improper detergent selection. CIP application, as both create high-shear and thus In this case, the customer was advised to discontinue are prone to foam formation. The result of this is a the use of the incompatible detergent, and a compat- detergent solution that foams out-of-process or CIP ible detergent chemistry was identified and tested. vessels, cavitates pumps, and pro vides inefficient The customer was also required to replace or repair surface coverage when sprayed on the inside of a vesdamaged equipment. sel through a spray ball. Con versely, a high foaming When developing a cleaning protocol, it is neces- detergent is desirable in a manual application, as this sary to identify all components of the process that will gives the operator a visual indication of where the be exposed to the cleaning chemical(s). This includes detergent solution has been applied to the surface. equipment surfaces, gasket materials, nozzles, piping, Some cleaning application technologies exist pumps, etc. It is also important to consider surfaces that are widely used in other industries, but have that will be exposed to the vapor phase of the cleaning not taken hold in the pharmaceutical industry. These solution, such as overhead spaces in enclosed vessels application methods include: and pipes. A common mistake is to concentrate only on items that will have direct contact with the liquid • Thin film cleaning solution, neglecting the vapor phase. • Stabilized foam generators • Built, solvated detergents (Generally RecogMethod of Application nized As Safe [GRAS]) There are several common methods of applying a detergent to equipment surfaces. Some are more Discussion regarding these application methods common than others in the pharmaceutical industry. are outside of this article and will not be addressed. Some of the more common methods of application in the pharmaceutical industry include: The Role of Water In general, 95-99% of a cleaning solution is • Clean-in-Place (CIP) composed of water. It is important to know the • Clean Out-of-Place (COP) purity level of the water being used for cleaning and • Manual scrubbing/wiping sanitizing. In many pharmaceutical applications, the • High and low pressure spray water being used for cleaning and sanitizing is high • Soaking/immersion purity water. However, this is not the case in every application and in these cases, knowing and underEach of these application methods dictate cer- standing how the purity level of the process water tain desirable or undesirable detergent properties. affects the cleaning process is critical. Some of the For example, a high pH detergent is ideal in a CIP factors that can affect the cleaning process include application where little, or no direct contact is made water hardness, pH, metals, salts, and microbial conbetween the detergent and the operator. In a manual tamination. Refer to Figure 5. application, however, a high pH detergent creates Of the factors listed above, waterhardness has the a significant safety risk to an operator handling the most significant impact on cleaning and sanitizing detergent concentrate and use solutions. In a manual solutions. Water hardness can be classified as temapplication, a neutral or mildly alkaline detergent porary or permanent hardness. Temporary hard ness (pH 7.0 – 10.0) is much more desirable as it sig- indicates the presence of bicarbonates of magnesium nificantly reduces the risk for accidental chemical or calcium. Both of these compounds are readily burn to the operator’s eyes, skin, and mucous mem- water soluble and can be present at high levels. When branes. heated, these compounds react to form the carbonate Other detergent characteristics, such as foam salts, which are water insoluble. Permanent hardness properties, are important considerations in light of the refers to a condition where the chloride or sulfate method by which the cleaning solution will be applied salts of magnesium and/or calcium are present in the to a surface. A moderate-to-high foaming detergent is water. These compounds are also very water soluble, not desirable when used in an agitated immersion or but are unaffected by temperature. 32
Institute of Validation Technology
Mark Altier
Figure 5
tp w ip t c ip cg P cp
c F
Pb cd
Barium Sulfate
BaSO
Scale
Carbon Dioxide
CO
Calcium Bicarbonate
Ca(HCO )
Calcium Sulfate
CaSO
4
Corrosion
2
3
2
Scale and Corrosion Scale and
4
must not fall below four. If a strong acid or alkaline detergent is used, the pH restrictions could be violated. In this case, choosing a neutral, mildly alkaline or mildly acidic detergent may be the solution. However, in some cases, a strongly acidic or alkaline detergent might be required to effectively re move the potential contaminant from equipment surfaces. If a strong alkaline detergent is re quired, the cleaning cycle could be designed to include an acid rinse. The acid rinse will help reduce the amount of rinse water quired re to neutralize residual alkalinity in the system, will help
remove any inorganic residues, and can be captured and mixed with the alkaline wash water to neutralize. In general, detergents will have the greatest Manganese Mn Scale impact on pH and phosphate levels. Relative to the Magnesium Bicarbonate Mg(HCO ) Scale residue load, detergents generally have little im pact Magnesium Chloride MgCl Scale and Corrosion on BOD, COD, TOC, or solids levels. Magnesium Sulfate MgSO Scale and If effluent restrictions exist, these should be Corrosion addressed in the early stages of the development of Oxygen O Corrosion a cleaning program to avoid compounded problems Sodium Chloride NaCl Corrosion later on when the cleaning protocol is implemented. Silica Si Scale At this point, five key factors that should be r Suspended Solids Deposit and considered when selecting a detergent to be used Corrosion as a part of a validated cleaning program have been Both temporary and permanent hardness cause discussed. Once these factors are addressed and an problems in alkaline solutions, as they both pre- appropriate detergent chemistry is identified, laboracipitate in high pH solutions and cause scaling on tory testing should be done to verify that the chemequipment surfaces. Water hardness is responsible istry is effective against the potential contaminant. for scaling, film formation, excessive detergent con- Other cleaning parameters such as cleaning time, sumption, and formation of precipitate. Water hard- temperature, and concentration can be evaluated in ness can be addressed by installing a water softening the laboratory as well. system, or by using a detergent that is formulated to Laboratory Testing handle hard water. Iron
Corrosion Scale
Fe
3
2
2
4
2
Cleaning studies conducted in the laboratory can be designed to closely mimic the actual applicaMany pharmaceutical plants have some type of effluent restrictions mandated by local municipalities, tion method, such as a CIP system, or they can be or by the plant’s internal effluent treatment facility. Common factors that must be considered are pH, Figure 6 phosphate levels, Biological Oxygen De mand (BOD) w hd or Chemical Oxygen Demand (COD) loading, Total (rpd cco ) rg Organic Carbon (TOC) levels, and solids levels. In Environmental Factors
3
hd
many cases, the correct choice of detergent help reduce the impact on components of the can effluent stream that are a concern. For example, if phos phates are a concern, a detergent that contains low levels of phosphate can be used. Another example is a situation where the pH of the effluent must not exceed 10 and
G P G
P P m (PPm)
Soft
0 – 3.5
0 – 60
Moderately Hard
3.5 – 7.0
60 – 120
Hard
7.0 – 10.5
120 – 180
Very Hard
>10.5
>180
Special Edition: Cleaning Validation III
33
Mark Altier
designed to stress the system to differentiate between similar cleaning chemistries. An example of the latter is a designed study that removes all mechanical action from the system, forcing the chemistry type and concentration, thermal energy, and contact time to act on the residue. This approach is especially effective in differentiating between similar chemistries that appear to be equally effective when applied using some type of mechanical action. An important component of designed cleaning studies is the preparation of the residue being tested. Typically, the residue is applied to a 304 or 316 stainless steel coupon and the treated coupon is then subjected to the cleaning solution. The application of the residue to the coupon is critical to obtain results that can be directly applied to the actual system in the plant. For example, a manufacturing process may involve a heating step that causes some of the finished product to “bake” onto a vessel side wall. To obtain results that are applicable to this situation, the residue should be applied to the coupon surface, heated, and then allowed to bake for an equivalent amount of time as is experienced in the actual process. If this is not done, the results of the study will have little relevance to thedevelopment of a cleaning program aimed at removing a baked on residue from equipment surfaces. Prior to implementing any cleaning studies, a set of success criteria must be established. Once the cleaning studies have been completed, quantitative measurements against the success criteria should be made. Based on the results of this work, a finaldetergent chemistry recommendation is derived. Ideally, the detergent chemistry should meet or exceed all established success criteria. The end result of the laboratory work will be a scientifically sound recommendation of detergent chemistry and other important cleaning parameters such as cleaning time, temperature, and concentration. This is the basis for the overall cleaning program that will be tested at the production facility. The importance of performing preliminary laboratory testing is that it provides a sound, scientific rationale of why the selected chemistry is appropriate for the cleaning application.
Plant Optimization 34
Institute of Validation Technology
Once a cleaning chemistry has been identified and verified in the laboratory and other cleaning parameters such as cleaning time, temperature and concentration have been established, testing and optimization must be carried out at the production plant. Initial optimization and testing is usually done on a pilot scale, prior to scaling up. The results of the laboratory cleaning studies should be used as a guide or a starting point for the optimization process at the plant site.
Conclusion Process cleaning is an integral component of any pharmaceutical process. The five key factors that must be addressed to help identify a detergent when developing a cleaning program have been defined and discussed. The interaction of these factors with each other and with the development of a cleaning program must be understood. Laboratory testing is critical for documenting the appropriateness of the detergent selection for the cleaning application. Plant optimization is a final critical step prior to starting the validation process at the production facility. When these steps are taken, a complete, scientifically sound approach to the development of a cleaning program can be documented.o
About the Author Mark Altier is a Principal Chemical Engineer for Ecolab Inc., where he manages their pharmaceutical and cosmetic programs. Mark has worked for Ecolab for seven years and has held positions in quality assurance, process engineering, and research and development. He can be reached at 651-306-5876, by fax at 651-552-4899, or by e-mail at
[email protected].
Analyzing Cleaning Validation Samples: What Method? By Herbert J. Kaiser, Ph.D. and Maria Minowitz, M.L.S. STERIS Corporation
v
C
leaning validations are very and cons of each technique will be difficult to perform. They examined. Validating the methods can be made easier if an } This article will will be discussed, as well. The refappropriate method for analyzing erences included with this paper can the samples is used. The method describe various be used to provide more in-depth used should be based on the previinformation to the reader and act as analytical ously established residue limits of guides to the available literature. the active and cleaning agents. There Choosing the appropriate anatechnologies are many choices of analytical techlytical tool depends on a variety of niques that can potentially be used. available for use, factors.5,6 The most important facThis article will describe various tor is determining what species or particularly for analytical technologies avail able for parameter is being measured.7 Is it use, particularly for cleaning agent cleaning agent res- an organic compound or inorganic residues. References are provided to compound? The next question is guide the reader to more in-depth idues. References measurement. How is this cominformation. are provided to pound going to be measured? Is it Cleaning validation in the phargoing to be swabbed from a surface maceutical industry is of critical guide the reader to or determined from a rinse water importance. 1,2,3 There are many sample? If it is going to be swabbed more in-depth analytical techniques available that from a surface, where will thisswabcan be used in cleaning validations.4 bing occur? Another important facinformation.~ The choice of the technique used tor in choosing an analytical tool is in analyzing a particular sample is establishing the limits of the resivery important in cleaning validadue. The limit should always be 8,9 tion. The technique must be appropriate for measur- established prior to selecting the analytical tool. ing the analyte at and below the acceptance residue The limits should not be established solely based on limit. Today’s analytical chemist has a wide vari- detection limits of a particular method. Yet, another ety of techniques available for use. These choices important factor in choosing an analytical tool is include specific and nonspecific methods. Many whether or not the method can be validated. If the methods are complementary to each other. The pros method can’t be validated, then another technique Special Edition: Cleaning Validation III
35
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
needs to be chosen.
Sampling Technique The sampling technique plays a large role in determining which analytical technique to use. Some techniques are more applicable for swab samples, and other techniques are more applicable for rinse water sampling. The acceptable sampling techniques include direct surface sampling (swab) and rinse water samples.10 The rinse water sample is a direct measure of potential contaminants, but the analysis should not just be a compendial test for water. Rinse water analyses should be directed toward responses peculiar to the possible contaminants. A questionable form of sampling is placebo sampling. The placebo method sampling is when the product, not containing the active ingredient, is processed in the specific piece of equipment. This is analyzed for any active that may have been picked up from the equipment. A problem with placebos is the potential lack of uniformity. The contaminant may not be evenly distributed throughout the placebo. Another problem is the analytical power of the tools that are used to analyze the samples. The residue levels may be extremely low if in fact the contaminant is evenly distributed throughout the sample. The use of placebos is only acceptable if used with swab or rinse water data. Therefore, placebos are generally not used because of the additional work involved. Another important factor to consider in choosing an analytical method is the type of residue being analyzed. Residues can be drug actives, formulation components, cleaning agents, organic, inorganic, water soluble, water insoluble, particulate, microbial, and/or endotoxins. If the residue being detected is a drug active, and the method used for detection is the same method that is used for qualitycontrol purposes of the final formula, it must be established that the active has not changed its chemical nature during the cleaning process. That is, it must be established that the active is still detectable and quantifiable using the analytical method. This can easily be established by performing forced degradation studies. Exposing the active to the cleaning compound at an elevated temperature and then analyzing that sample will help determine the compatibility of the cleaner with the active. If the active has indeed changed its chemical nature during the cleaning process, a new technique 36
Institute of Validation Technology
will need to be established for its analysis. Limit of Detection and Quantitation Before choosing a method, some definitions need to be established. The Limit of Detection (LOD) is the lowest amount of a compound that can be detected. The Limit of Quantitation (LOQ) is defined as the lowest amount of a compound that can be quantified. The LOD is usually lower than the LOQ, but is never higher. The LOD should never be used to establish residue acceptance limits. The residue acceptance limit should be well above the LOQ so that it can be accurately quantitated. Specific and Nonspecific Methods
A specific method is a method that detects a unique compound in the presence of potential contaminants. Some examples of specific methods are High Performance Liquid Chromatography (HPLC), ion chromatography, atomic absorption, inductively coupled plasma, capillary electrophoresis, and other chromatographic methods. It should be noted that HPLC is not inherently specific. What is meant is that the conditions in an HPLC measurement can usually be adjusted to separate out known potential contaminants. Nonspecific methods are those methods that detect any compound that produces a certain response.Some examples of nonspecific methods are Total Organic Carbon (TOC), pH, titrations, and conductivity. A very interesting and sensitive nonspecific technique is dynamic contact angle.11 Titrations may be specific for acids or bases, but they are not specific for particular acids or bases. There are, however, specific titrations for classes of surfactants.12 Interferences A good nonspecific strategy that could be followed is to first identify possible interferences. These interferences can be either positive or negative. The nonspecific property is then measured, and the residue is calculated as if all of the measured property is due to that residue. For example, if the cleaning agent was the analyte and TOC was themethod used, all of the TOC would be assumed to have come from the cleaning agent and calculated as such.This would then provide a worst-case upper-limit value. There are many possible sources of interferences. Cleaning agents and compounds can be a source of
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
interferences, for example. Active agents and their byproducts, water system components, maintenance materials, and the atmosphere can all be sources, as well as people, if samples are not handled properly. The materials used to perform the analytical method can also be a source of interference. For example, if a swab that has a high TOC value is used to sample, it could increase the level of TOC detected. For specific methods, there should be no interference if the method is properly designed. Again, it should be stressed that the method must be able to fol-
is no need for heating or cooling the detector. While there are many advantages of UV detectors, there are also some significant disadvantages. UV detectors cannot detect all types of compounds and therefore are not considered to be universal. All compounds do not have chromophores. This is particularly true of surfactants that are used in the pharmaceutical industry. Dirty cells, air bubbles, and the use of gradients can affect baseline drift and detection capability. The limits of detection can be higher than other detector types due to background
low the analyte after exposure to the cleaning vironen ment. It is necessary to establish that the cleaning environment or the cleaning process does not change the analyte. For nonspecific methods (which measure a nonspecific property), any compound with the property that is introduced into the sample will interfere. For example, if the method being used is TOC, atmospheric carbon that may enter the sample could cause interference. With all nonspecific methods, there is a need to identify potential sources of interference.
interferences.
High Performance Liquid Chromatography The first technique that will be discussed is HPLC. Almost every pharmaceutical company has an HPLC instrument. HPLCs utilize a variety of detectors. These include ultraviolet (UV), fluorescence, electrochemical, refractive index, conductivity, evaporative light scattering, and many others. The ultraviolet detector is by far the most common. However, Evaporative Light Scattering Detection (ELSD) may be the most appropriate detector for cleaning agents. We will discuss the use of both UV and ELSD detectors in depth. n
• Ultraviolet Detectors There are many advantages of using UV detectors. Many compounds have chromophores and therefore, they can be easily detected by UV. Many instruments are equipped with diode array spectral capabilities. This allows for easy detection of impurities or potential contaminants within peaks. Ultraviolet detection usually requires no additional reagents or post column or pre-column reactions. UV detectors are not harmful to the sample, if that is important. They are generally inexpensive and readily available. Also, molar absorptivities are generally not affected by temperature and therefore, there
• Evaporative Light Scattering Detection In ELSD, the compound is separated on an HPLC column as usual, and then enters a nebulizer that is combined with a gas stream and passed through a heated column. The heated column evaporates the mobile phase leaving the solid analyte in the column. The solid analyte then passes through a detector that consists of a laser or light source. The laser or light source is scattered when it hits the solid analyte. The detector then picks up this scattering. There are many advantages associated with evaporative light scattering detectors. ELSD is claimed to be universal. It is called universal because it can detect any type of compound. ELSDs are simple, versatile, and rugged in use. Since it is a mass detector, all compounds produce similar responses. Additionally, there is no baseline drift due to mobile phase effects. There are two primary disadvantages of ELSD. First, there is a very limited choice of buffer salts that can be used. Recall that the mobile phase is evaporated or removed, leaving the analyte. Any buffers that will not evaporate will also produce solid particles that will then be detected and cause interferences. The second disadvantage is that the nebulizer and detector must produce consistent particle sizes. This requires careful cleaning and monitoring of the nebulizer. Actives and Detergent There are many types of residues that can be analyzed using HPLC techniques. These include both actives and detergent residues. When dealing with detergent residues, it is important to identify what is being analyzed: surfactant, builder components, Special Edition: Cleaning Validation III
37
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
chelating agents, etc. The separation and quantitation of surfactants at low levels is difficult, at best. Industry literature is full of references for surfactant analyses using HPLC. The vast majority of techniques described in the literature are for the deter13,14 mination of surfactants in concentrated products. Therefore, the limits of quantitation and the limits of detection are rather high. There are also references for the analysis of surfactants related to the environment.15,16 In environmental analysis, the sample is pre-concentrated so that the limits of quantitation are very low. The pre-concentration can be up to one thousand fold.
} The
coating simply to make it more rugged. All common detection techniques (UV, fluorescence, etc.) can be used in capillary electrophoresis detection. The capillary itself serves as the detector cell. A small portion of the polyimide coating is scraped off prior to use, and the bare portion of the capillary is placed in the light path. This detection is different from that seen in HPLC because the detection occurs while the separation is taking place, rather than after separation has been completed. Using a Z-cell can increase the sensitivity of the technique. This is accomplished by
TOC is then computed by
Suggested Reading subtracting the inorganic carbon Authors Lin, et. al., compared concentration from the total the analysis of anionic, cationic, and amphoteric surfactants concarbon concentration taining n-dodecyl groups using of the sample.~ HPLC and capillary electrophoresis.17 They found that HPLC was best for all classes of surfactants, especially for for- using a special accessory that bends the capillary, mulated surfactants. Authors Carrer, et. al., utilized causing the source radiation to penetrate lengthwise ELSD for amphoteric type surfactants.18 Amphoteric through the capillary rather than a cross-sectional surfactants are a class of surfactants that display sampling. This, in effect, increases the path length cationic behavior in an acidic solution and anionic of the cell. The Z-cell can be used in all types of CE behavior in an alkaline solution. The lowest cali- where UV detection is used. bration standard that they utilized was 50 ppm, but CE can be used for many different types of analythey probably could have gone much lower. Authors ses. Surfactants can be determined quite readily using Guerro, et. al., obtained a limit of quantitation of this technique.20,21 However, detection limits typically 0.49 ppm for alkyl polyethylene glycol ethers using are higher than with HPLC. This can be overcome by pre-concentrating the samples on the capillary itself. ELSD.19 A voltage is applied to the capillary in a manner that n Capillary Electrophoresis allows the compounds to collect at one end of the An interesting method of analysis is Capillary capillary without flowing through to the detector. An Electrophoresis (CE). There are manydifferent types advantage that capillary electrophoresis holds over of CE. Capillary Zone Electrophoresis (CZE) is by HPLC is the ease with which indirect detection can far the most common. CE instrumentation is fairly take place. Indirect detection is where a highly UVsimple, consisting of a high voltage source, a capil- absorbing material is included in the mobile phase. lary, and a detector. The high voltage source is used As the analyte is eluted or travels along the capillary to apply a potential across two solutions. One of through the detector, a negative peak is seen for the the solutions contains the analyte, and the potential analyte. This typically is done for compounds that disapplied to the solutions causes the analyte to migrate play low UV absorption. In addition to being useful through the capillary, through the detector, and for the analysis of surfactants, capillary electrophore22 into the other solution. The column or capillary is sis can be used to analyze organic acids, inorganics, 23 typically composed of fused silica with a polyimide and trace drug residues. coating. The diameter of the capillary is typically 25-75µm in diameter. The capillary has a polyimide Suggested Reading 38
Institute of Validation Technology
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
Vogt, et. al., provided a good overview of the separation of cationic, anionic, and nonionic surfactants using capillary electrophoresis.24 They indicated that one can easily adjust the parameters of the separation to coelute or separate oligomers. Coelution of the oligomers increased the sensitivity at the ex pense of increasing the potential for coeluting positive interferences. Direct UV detection could be used for UV-absorbing materials and indirect or non -UV absorbing materials. Heinig, et. al., utilized micellar electrokinetic cap-
CE of 4.0 ppm; and for HPLC, they obtained a limit of quantitation of 5.0 ppm. n Total Organic Carbon TOC is used widely in the pharmaceutical industry.31,32,33 The TOC is determined by the oxidation of an organic compound into carbon dioxide. This oxidation can occur through a number of mechanisms depending on the instrument being used. Some typical methods are persulfate, persulfate/UV oxidation, and direct combustion. The carbon dioxide that is produced from these oxidations is either
illary chromatography for the separation of non-ionic 25 Howalkylphenol polyoxyethylene type surfactants. ever, the use of this method was limited because of insufficient peak resolution and relatively low detection sensitivity. Heinig, et. al., also compared HPLC and CE analyses of surfactants.26 The surfactant types they studied were linear alkylbenzenesulfonates, nonylphenolpolyethoxylates, cetylpyridinium chloride, and alkylsulfonates. For the CE analyses, they utilized UV detection either in the direct or indirect modes, depending on the nature of the surfactant. For the HPLC analyses, they utilized either direct UV detection or conductivity detection. Anionic surfactant samples were pre-concentrated one thousand fold through the use of solid phase extraction. This allowed for detection limits in the parts per billion range to be obtained. Kelly, et. al., utilized CE with indirect detection to determine sodium dodecylsulfate concentrations.27 They also indicated that it is important to look at the absorption of the surfactants onto filters if the samples are indeed filtered prior to analysis. This is most important in dilute solutions. Filtering large volumes of sample can minimize this. Again, appropriate studies need to be done to determine if this indeed is a problem. Altria, et. al., examined the use of CE in the analysis of sodium dodecylbenzenesulphonate.28 They obtained a limit of quantitation of 0.6 ppm and a 0.3 ppm limit of detection. They utilized direct UV detection. Shamsi, et. al., utilized CE with indirect detection for the determination of cationicand anionic surfactants.29 The authors obtained limits of detection of 0.25 and 0.5 ppm, respectively. Heinig, et. al., also utilized CE in the analysis of cationic surfactants using indirect UV detection.30 They compared this with HPLC. They obtained a limit of quantitation for
measured using conductivity or infrared techniques. Instruments generally measure the inorganic carbon content of a sample. The inorganic carbon consists of carbon dioxide, bicarbonate, and carbonate. They then determine the total carbon content of the sample. The TOC is then computed by subtracting the inorganic carbon concentration from the total carbon concentration of the sample. There are two primary advantages associated with TOC. The first is that it does not take long to develop a method. There are not a lot of variables in the actual analysis. The second advantage is that it is relatively quick. A third potential advantage (which can also be a disadvantage) is that it will detect and analyze any compound containing carbon. As with most techniques, there are disadvantages in using TOC. A significant disadvantage is that the compound or the analyte must be water soluble. This does not mean that the compound must be soluble in the hundreds of parts per million range but soluble in the low parts per million range. Another disadvantage is that organic solvents cannot be used. If organic solvents were used, the TOC of the solvents would be measured instead of the residue. There are also many sources of contamination that can occur using TOC. These sources can include the atmosphere, the swab itself, personnel, and many other sources. Methods developed using TOC should be written to include controls and blanks to identify or account for possible contamination. For example, a common source of contamination is the technique used to cut the handles of the swabs so that they fit into the TOC vials. Many times, the scissors or utensils are not clean enough for TOC use. This introduces contamination into the sampling vial when the swab is cut. Excipients Special Edition: Cleaning Validation III
39
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
Some methods/techniques can be used in certain situations to complement each other. Examples include TOC and HPLC. Consider the case of a drug in the presence of excipients. The excipients are very soluble in water while the drug active has ex tremely low solubility in water. The excipients contribute to the TOC values because they are very soluble in water; however, the drug active does not show up in the TOC analysis. An HPLC analysis is performed to monitor the loss of the drug. The excipients are removed much faster from a surface during cleaning
cleaners.39,40,41 Most cleaners contain sodium and/or potassium. The ion chromatography detection technique of suppressed conductivity is more sensitive to potassium than it is to sodium. Very low levels of cleaning agent can be detected using this technique. This assumes that the rinse water used contains no potassium. Ionizable organic acids are also readily quantitated using ion chromatography. This includes chelating agents that are often found in cleaning compounds.
than the drug active is removed. In this case, TOC analysis is not a good stand-alone method. It is, however, a good complement for the HPLC assay. The TOC analysis enables the analyst to see what water soluble matter is left behind, if any.
Suggested Reading In determining surfactants, an excellent review 42 concerning their analysis was done by Vogt, et. al.. They compared the use of HPLC, CE, ion chromatography, Liquid Chromatography-Mass Spectroscopy (LC-MS) and Gas Chromatography-Mass Spectroscopy (GC-MS). They also discussed pre-concentration of the samples. They compared the use of solid phase extraction, super critical fluid extraction, Soxhlet extraction, and steam distillation as means of pre-concentrating samples. They found, by far, that the best method was solid phase extractions for the pre-concentration of surfactants. They also examined the use of titrimetric methods of analysis for surfactants. For detecting anionics, substances like methylene blue, pyridinium azo, and triphenylmethane dye was used to complex the surfactants prior to photometric determination. Nonionics were determined indirectly by forming a cationic complex with barium. This complex was then precipitated by bismuth tetraiodide ion in acidic acid. The bismuth was then quantified by potentiometric titrations. Cationics were complexed with anionic dyes such as disulfine blue. Theile, et. al., brought up an excellent point that 43 This surfactants tend to concentrate at interfaces. can be a problem in extremely dilute solutions of surfactants. The surfactants can collect at the surface of the containers that they are stored in. This may cause errors in analysis. Proper controls in studies should be done to determine if this is a problem. The authors indicated that pre-concentration was required to determine very low levels of surfactant. Solid phase extraction was the best method for this. They were also able to obtain detection limits for linear alkylbenzenesulfonates of 2.0 ppb with fluorescence detection and 10.0 ppb using HPLC with
Suggested Reading Guazzaroni, et. al., examined the use of total organic carbon for the analysis of detergents, endo34 toxins, biological media, and polyethylene glycol. For detergents, they were able to obtain a 0.7 ppm limit of quantitation. Endotoxins were found to have a 0.2 ppm limit of quantitation. The biological media produced a total organic carbon limit of quantitation of 20.3 ppm; and the polyethylene glycol produced a 0.5 ppm limit of quantitation. They examined swab and rinse water recoveries. They were able to obtain 78-101 percent recoveries utilizing swabs, and 93 percent or better for rinse water recoveries. There are many examples in the literature that utilize ion chromatography as the method for analysis of surfactants.35 The surfactants have to be charged in order to be analyzed using ion chromatography, that is, only anionic or cationic surfactants can be detected. Pan, et. al., recorded limits of quantitation down to 0.5 ppm for linear alkane sulfates and sulfonates.36 Takeda, et. al., recorded a limit of quantitation of 0.1 ppm for dodecyl alkyl sulfates.37 Nair, et. al., separated different sulfate, sulfonate, and cationic type surfactants using ion chromatography with suppressed conductivity detection.38 They reported detection limits at less than 1.0 ppm.
Ion Chromatography In addition to its use for surfactants, ion chromatography can be used for the analysis of inorganics and other organic compounds present in n
40
Institute of Validation Technology
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
UV detection after pre-concentration. n Thin-Layer Chromatography There are many examples in the literature for the use of Thin-Layer Chromatography (TLC) for the qualitative determination of surfactants.44,45 Henrich described the TLC of over 150 surfactants in six different TLC systems.46 This was excellent for identification of the surfactants, but the author did not attempt to quantify the surfactants. Buschmann and Kruse combined diffuse reflection infrared spectroscopy and TLC, along with SIMS and TLC for sur-
50 This is based on phosphate (ATP) bioluminescence. the reaction of ATP with Luciferin/Luciferase. This technique is often used in biopharmaceutical facilities. It has extremely high sensitivity and a very high reproducibility. In many cases, the instruments can be used at the equipment site. This technique utilizes swabs for surface analyses.
factant identification.47 Although these techniques are tedious and time-consuming, there is no doubt that these methods could be developed into quantitative analyses. Novakovic has used high performance TLC for two generic drugs.48
conventional methods. A very sensitive method that may be applicable is Optically Stimulated Electron Emission (OSEE).51 The instrumentation for OSEE is fairly portable, and can be readily taken to tank side for analysis. The technique uses a probe that is placed against a surface, and a UV source illuminates and activates the surface. When some surfaces are exposed to UV light at certain wavelengths, electrons are emitted from the surface. The instrument measures the current that is produced. If even small amounts of residues are present on the surface, the current will be affected. The current can be affected either in a positive or negative way depending on the nature of the residue. This is an extremely sensitive technique. It can be used in either a qualitative or quantitative manner.
Other Techniques Other excellent techniques for inorganic contaminants, and in some cases actives, are Atomic Absorption (AA)49 and Inductively Coupled Plasma (ICP) atomic emission. These techniques can detect inorganic contaminants down to extremely low levels. Inorganic contaminants in a system are often ignored. These can come from rouge that forms in Water for Injection (WFI) systems. They can also come from the detergent utilized in cleaning the equipment.
Fourier-Transform Infrared Spectroscopy Fourier-Transform Infrared (FTIR) spectroscopy is never used as a stand-alone method for analyzing residues on equipment. This is because of the lack of portability of FTIR equipment and the semi-quantitative nature of the reflectance techniques used for these types of analyses. However, it is very useful in performing screening studies and in evaluating potential cleaning agents. This is done by soiling standard coupons with the cleaning agent, allowing them to dry, and performing static rinsing studies. These types of studies can indicate whether or not the cleaning agent is readily removed from surfaces. The height or area of a particular peak is measured versus the concentration of the standard coupon. n
nBioluminescence
Optically Stimulated Electron Emission In some cases, a company’s established limits of residue are so low that they cannot be detected by n
nPortable
Mass Spectrometer For those companies that require ultrasensitive measurements and identification of the residues, a technique has been developed – Lawrence Livermore National Laboratories has developed a portable mass spectrometer.52 The unit consists of a gun portion of the instrument that is connected with cables to vacuum pumps. The gun portion is held against the surface to be analyzed. A seal is formed, and the surface is heated to volatilize any compounds that are present. This instrument is used not only to measure how much of something is present, but also what that something is. This piece of equipment has been utilized in the aerospace industry. One drawback of the portable mass spectrometer is that it requires relatively flat surfaces. However, they are currently working on adaptors to be used on non-flat surfaces.
Bioluminescence is quite useful for biologicals. This type of analysis usually uses Adenosine Tri- Additional Techniques Special Edition: Cleaning Validation III
41
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
In the biopharmaceutical industry, a wide variA minimum validation requires two different ety of techniques are utilized.53 These include the analysts, instruments, columns (if chromatographic Enzyme-Linked Immunosorbent Assay (ELISA),54 method), days, and prepared standards and samthe Limulus Amoebocyte Lysate (LAL), and a wide ples.60 These are the “twos” of method validation. variety of protein determinations. These are all con- The point of any method validation is to show that taminant specific assays. For example, the LAL test the method can be utilized by different analysts and/ measures the level of endotoxins present. There is or laboratories, along with the ability to produce the also the anthrone assay that can be used to monitor same results. If a validated method is given to a labthe levels of carbohydrates on surfaces. These techniques are usually } For those companies that require ultraused in combination with TOC. The nonspecific techniques of sensitive measurements and pH, conductivity, and titrations identification of the residues, a can be used throughout all areas of pharmaceutical manufacturing. technique has been developed Obviously, these techniques are – Lawrence Livermore National most often utilized in rinse water monitoring. The conductivity and Laboratories has developed a pH of rinse water is typically moniportable mass spectrometer.~ tored and compared to the conductivity and pH of the water prior to introduction to the equipment. If acidic or alkaline materials are being measured, oratory, that laboratory must revalidate the method titration is a very useful technique. In some cases, for their laboratory. It is not sufficient or accurate titration can be more sensitive than performing to assume that another laboratory’s validation will TOC analyses. The sample size can be adjusted, apply in all laboratories. For example, if a surfactant and/or the normality of the titrant can be adjusted to is being quantitated, typically a low wavelength increase the sensitivity of the titration. is used in a UV detector for HPLC. UV detectors vary in their noise levels at these low wavelengths. Method Validation A detector used in one laboratory may have significantly less noise than a detector in another labIt is very important to scientifically establish the oratory. The second laboratory may not be able to residue limit prior to choosing the method of analy- detect at the same low level as the first laboratory. sis. This includes the limit in the analytical sample and the limit in the next product. This will ensure Coupons and Swabs that the method chosen will be able to detect and Coupons can be prepared for recovery studies quantitate the limit chosen. Once the technique for through the use of aerosol bottles available through analysis has been chosen, it is very important to vali- laboratory supply companies. A known weight perdate the method used.55-60 The validation of a method cent of a solution containing the analyte can be is very different than the validation of recovery. A sprayed fairly evenly over the surface of a coupon. validated method is one that is rugged and robust The coupon can be swabbed using a standard techenough to measure the residue limit established. nique. It does not matter how you swab the coupon, The validation of a recovery is to determine the as long as the complete surface is covered and that the amount that can be recovered from a surface. Again, coupons are swabbed the same way – each and every it should be stressed that these are two completely time. The type of swabs used in recovery studies must different validations. be the same as those used in the validation protocol. If this is a simulated rinse procedure, then the coupons “Twos” of Validation are rinsed and the rinse water is analyzed. 42
Institute of Validation Technology
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
For swabs, a desorption process is carried out. This can consist of simply shaking the sample vial or using an ultrasonic bath. The samples are then analyzed. Recovery studies are always done below acceptance limits in the test solution. This ensures that the limit will be (or can be)measured in the analysis. A recovery of greater than 80 percent is good. If the recovery is greater than 50 percent, it may be acceptable. However, if the recovery is less than 50 percent, questions arise and the source of the poor recovery should be investigated. A possible cause of
Detergents can be quantitated either using specific or nonspecific methods; however, care must be taken in choosing which component is measured. For example, a detergent may contain five percent of a surfactant and 20 percent of another organic ingredient. Assuming equal sensitivities of the analytical methods, the limit of quantitation of the whole detergent system will be lowered by a factor of four if the ingredient present in the greater amount is determined. If a nonspecific method (i.e., TOC) is used for
a poor recovery can be that the residue is being too tightly held by the swab. This can be investigated by spiking a swab with a known amount of residue, allowing it to dry, trying to desorb the residue, and following up with analysis. If the analyte is held too tightly by the swab, another type of swab material should be investigated. The recovery factor should be included in analytical calculations or in the acceptance limit calculation. It should not be included in both of the calculations.
the same system, a much lower limit of quantitation could be determined simply because there would be a tremendous amount of carbon present in the sample. In addition, if detergent systems are combined, such as in the case of adding a detergent additive to another product, the choice of a specific method would be made even more difficult. The question would be, “Which detergent do I determine?” A disadvantage of using a nonspecific method for the entire cleaning validation analysis is that, if there is a failure in the future, it would not be known where Containers the failure srcinally occurred. The failure could be The choice of containers used in the analysis of due to the active, excipients, detergent system, or samples is very important. It has been shown that, even an unknown source. in very dilute solutions, surfactants can adsorb onto the surfaces of sample vials. This will produce artiConclusion ficially low results in the analysis. This, however, There are many different analytical techniques typically only occurs in static systems. There is no need to worry about the adsorption of the surfactant available that can be used to detect residues. These on the walls of manufacturing equipment. This is range from simple titrations to more complex LCbecause the agitation that is involved in cleaning MS. The choice of technique should be based on removes the surfactants from the surfaces. This is what equipment is available, the type of residue, another matter in sample vials. Appropriate spiking and the scientifically established residue limit. It is studies should be performed to ensure that this phe- important for an analytical chemist to keep abreast nomenon is not occurring and will not interfere with of the literature and what techniques are available. the analytical method. This includes both HPLC or There are techniques available that will analyze any ion chromatography sample vials, as well as TOC residue at any level. At the end of the day, however, sample vials. This phenomenon is not limited to sur- it is always wise to choose the simplest technique factants. Proteins have been shown to adsorb readily that can be used to reach the desired goal.o onto glass surfaces. These proteins are much more difficult to remove from surfaces than surfactants.
About the Authors Specific Versus Nonspecific The choice of using a specific or nonspecific method can be difficult. If a drug active is highly 61 toxic, a specific method is always recommended.
Herbert J. Kaiser, Ph.D. is Manager – Hard Surface Products at STERIS Corporation. He has 18 years of experience in cleaning and surface technologies, which includes developing products and methods for the cleaning and analyzing of a wide variety of Special Edition: Cleaning Validation III
43
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
surfaces. Dr. Kaiser has developed a wide variety of 13. products for the healthcare, industrial, and pharmaceutical markets. He is the sole inventor listed in five United States Patents for various industrial treat14. ment schemes. Dr. Kaiser received his B.A. degree from St. Mary’s University in San Antonio, Texas, his M.S.(R) from St. Louis University, and his Ph.D. 15. from the University of Missouri. He is a member of the American Chemical Society and the Association for the Advancement of Medical Instrumentation. Dr. 16. Kaiser can be reached by phone at 314-290-4725, by fax at 314-290-4650, or e-mail at herb_kaiser@steris.
Potentiometric Sensors.” Analyst. Vol. 114. 1989. pp. 1435-1441. Schmitt, T. M. “HPLC Analysis of Surfactants.” Handbook of HPLC. Eds. Katz, E., Eksteen, R., Schoenmakers, P., Miller N. Chromatography Science. Series 78. Marcel Dekker, Inc.: New York. 1998. pp. 789-804. McPherson, B.P., Rasmussen,H. T. “Chromatographyof Cationic Surfactants: HPLC, TLC, and GLC.” Cationic Surfactants. Eds. Cross, J., Singer, E.Surfactant Science . Series 53. Marcel Dekker, Inc.: New York. 1994. pp. 289-326. Jandera, P. “HPLC of Surfactants and Related Compounds.” Liquid Chromatography in Environmental Analysis . Ed. Lawrence, J. Humana Press: New Jersey. 1984. pp. 115-167. Waters, J.“Analysis ofLow Concentrations ofCationic Sur factants in Laboratory Test Liquors and Environmental Samples.” Cationic Surfactants. Eds. Cross, J., Singer, Sur E. factant Science. Series 53. Marcel Dekker, Inc.: New York. 1994. pp. 235-256.
17. Lin, W., Lin, S., Shu,High-Performance S. “Comparison of Analyses of Surfactants in Cosmetics Using Liquid Chromatography and High Performance Capillary Electrophoresis.”Journal of Surfactants and Detergents. Vol. 3(1). 2000. pp. 67-72. 18. Carrer, G, Faccetti, E., Valtorta, L., et. al. “An Analytical Approach for the Determination of Betaines in Liquid Form ulation.” Rivista Italiana delle Sostanze Grasse. Vol. 76 (4). 1999. pp. 167-171. 19. Guerro, F., Rocca, J. L.. “RPLC Analysis of Alkyl Polyethyleneglycol Ethers Using Evaporative Light Scattering De tection.” Chimica Oggi. Vol. 13(4-5). 1995.pp. 11-15. 20. Heinig, K., Vogt, C., Werner, G. “Determination of Linear Alkylbenzenesulfonates in Industrial and Environmental Sample by Capillary Electrophoresis.”Analyst. Vol. 123. 1998. pp. 349353. 21. Heinig, K., Vogt, C. “Determination of Surfactants by Capillary Electrophoresis.”Electrophoresis. Vol. 20. 1999. pp. 3311-3328. 22. Oehrle, S. A. “Analysis of Low-LevelAnions in Water Extracts of Hard Disk Drive Heads by Capillary Electrophoresis.” Journal of Chromatography. Vol. 745. 1996. pp. 81-85. References 23. Altria, K. D., Hadgett, T. A. “An Evaluation of the Use of Galatowitsch, S. “The Importance of Cleaning Validation.” Capillary Electrophoresis to Monitor Trace Drug Residues Cleanrooms 2000. Vol. 14(6), pp. 19-22. Following the Manufacture of Pharmaceuticals.” ChromatoParenteral Drug Association. Points to Consider for Cleaning graphia. Vol. 40(2). 1995. pp. 23-27. Validation. Technical Report No. 29. 1998. 24. Vogt, C., Heinig, K. “Surfactant Analysis by Capillary LeBlanc, D. A. “Validated Cleaning Technologies for PharmaElectrophoresis.” Tenside Surfactants and Detergents. Vol. ceutical Manufacturing.” Interpharm Press: Denver. 2000. 35(6). 1998. pp. 470-475. Kanegsberg, B. and Chawla, M. “How Clean is Clean Enough?” 25. Heinig, K., Vogt, C., Werner, G. “Separation of Nonionic The Journal of Advancing Applications in Con tamination Surfactants of the Polyoxyethylene Type by Capillary Control 2000. Vol. 3 (9). pp. 9-12. Electrophoresis.” Fresenius Journal of Analytical Chemistry . Jenkins, K. M. and Vanderwielen, A. J. “Cleaning Validation: Vol. 357. 1997. pp. 695-700. An Overall Perspective.” Pharmaceutical Technology. Vol. 18. 26. Heinig, K.,Vogt, C.,Werner, G.“Separation of Ionicand Neutral 1994. pp. 60-73. Surfactants by Capillary Electrophoresis and High-Performance Gavlick, W. K., Ohlemeier, L. A., Kaiser, H. J. “Analytical Liquid Chromatography.”Journal of Chromatograhpy. Vol. 745. PharmaStrategies for Cleaning Agent Residue Determination.” 1996. pp. 281-292. ceutical Technology. Vol. 19. 1995. pp. 136-144. 27. Kelly, M. A., Altria, K. D., Clark, B. J. “QuantitativeAnalysis Kaiser, H. J., Tirey, J. F., LeBlanc,D. A. “Measurement of Organic of Sodium Dodecyl Sulphate by Capillary Electrophoresis.” and Inorganic Residues Recovered from Surfaces.”Journal of Journal of Chromatography. Vol. 781. 1997. pp. 67-71. Validation Technology. Vol. 6. No. 1. 1999. pp. 424-436. 28. Altria, K. D., Gill, I., Howells, J. S., et. al. “Trace Analysis of LeBlanc, D. A. “Establishing Scientifically Justified Acceptance Detergent Residues by Capillary Electrophoresis.”ChromaCriteria for Cleaning Validation of Finished Drug Products.” tographia. Vol. 40(9-10). 1995. pp.527-531. Pharmaceutical Technology. Vol. 22(10). 1998. pp. 136-148. 29. Shamsi, S. A., Danielson, N. D. “Individual and Simultaneous Fourman, G.L., Mullen, M. V. “Determining Cleaning Validation Class Separations of Cationic and Anionic Surfactants Using Acceptance Limits for Pharmaceutical Manufacturing Operations.” Capillary Electrophoresis with Indirect Photometric De tection.” Pharmaceutical Technology. Vol. 17. 1993. pp. 54-60. Analytical Chemistry. Vol. 67(22). 1995. pp. 4210-4216. LeBlanc, D.A. “RinseSampling for Cleaning Validation Studies.” 30. Heinig, K., Vogt, C., Werner, G. “Determination of Cationic Pharmaceutical Technology. Vol. 22. 1998. pp. 66-74. Surfactants by Capillary Electrophoresis with Indirect
com. Maria Minowitz, M.L.S., Information Associate at STERIS Corporation, has 10 years of experience in corporate research and development librarianship. She has been responsible for information management in the disciplines of chemistry, medicine, and engineering. Minowitz received her A.B. degree from St. Louis University and an M.L.S. from the University of Missouri-Columbia. She is a member of the Special Libraries Association, Midcontinental Chapter of the Medical Library Association, and the St. Louis Medical Librarians.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
11. Davies, Nunnerley, C. Profile S., Brisley, A. C.,in et. al. “Usethe of Photometric Detection.” Journal of Chromatography. Vol. 781. DynamicJ.,Contact Angle Analysis Studying 1997. pp. 17-22. Kinetics of Protein Removal from Steel, Glass, Poly tetra31. Jenkins, K. M., Vanderwielen,A. J., Armstrong, J. A., et. al. fluoroethylene, Polypropylene, Ethylenepropylene Rubber, and “Application of Total Organic Carbon Analysis to Cleaning Silicone Surfaces.” Journal of Colloid and Interface Science . Validation.” Journal of Science & Technology. Parenteral Drug Vol. 182(2). 1996. pp. 437-443. Association. 1996. Vol. 50. pp. 6-15. 12. Vyratas, K., Dvorakova, V., Zeman, I. “Titrations ofNon32. Biwald, C. E., Gavlick, W. K. “Use of Total Organic Carbon Ionic Surfactants with Sodium Tetraphenylborate Using Simple Analysis and Fourier-Transform Infrared Spectroscopy to
44
Institute of Validation Technology
Herbert J. Kaiser, Ph.D. & Maria Minowitz, M.L.S.
Determine Residues of Cleaning Agents on Surfaces.”Journal of AOAC International. Vol. 80. 1997. pp. 1078-1083. 33. Westman, L., Karlsson, G. “Methods for Detecting Residuesf o Journal of PharmaCleaning Agents During Cleaning Validation.” ceutical Science Technology. Vol. 54(2). 2000. pp. 365-372. 34. Guazzaroni, M., Yiin, B., Yu, J. L. “Application of Total Organic Carbon Analysis for Cleaning Validation in Pharmaceutical Manufacturing.” American Biotechnology Laboratory. Vol. 16(10). 1998. pp. 66-67. 35. Hoeft, C.E., Zollars, R.L. “DirectDetermination of AnionicSurJournal of Liquid Chromafactants Using Ion Chromatography.” tography. Vol. 17(12). 1994. pp. 2691-2704. 36. Pan, N., Pietrzyck, D. J. “Separation of AnionicSurfactants on Anion Exchangers.” Journal of Chromatography. Vol. 706. 1995. pp. 327-337. 37. Takeda, T., Yoshida, S., Ii, T. “Analysis of Sulfonate- and
Integrated Approach for Validating Cleaning Procedures in Biopharmaceutical Manufacturing Facilities.” Annals of the New York Academy of Sciences . Vol. 782. 1996. pp. 363-374. 54. Rowell, F.J., Miao,Z. F., Neeve, R.N. “Pharmaceutical Analysis Nearer the Sampling Point, Use of Simple, Rapid On-Site Immunoassays for Cleaning Validation, Health and Safety, and Journal of Pharmacy and Environmental Release Applications.” Pharmacology. Vol. 50. 1998. p. 47. 55. Seno, S.,Ohtake, S.,Kohno, H.“Analytical Validation inPractice at a Quality Control Laboratory in the Japanese Pharma ceutical Market.” Accreditation and Quality Assurance. 1997, 2(3), 140145. 56. Kirsch, R. B. “Validation of Analytical Methods Used in Pharmaceutical Cleaning Assessment and Validation.”Pharmaceutical Technology. 1998 (Analytical Validation Supplement). pp. 40-46.
Sulfate-Type Anionic by 441-444. Ion Chromatography.” 57. Chemistry Express . Vol. Surfactants 7(6). 1992. pp. 38. Nair, L. M., Saari-Nordhaus, R.“Recent Developmentsin Surfactant Analysis by Ion Chromatography.”Journal of Chroma58. tography. Vol. 804. 1998. pp. 233-239. 39. Weston, A. “Ion Chromatography in the Pharmaceutical Industry.” American Biotechnology Laboratory. Vol. 16(3). 59. 1998. pp. 32-33. 40. Murawski, D. “Ion Chromatography for the Analysis of Household Consumer Products.”Journal of Chromatography. 60. Vol. 546. 1991. pp. 351-367. 41. Masters, M. B. “Use of Ion Chromatography in Surfactant 61. Analysis.” Analytical Processing. London. Vol. 22(5). 1985. pp. 146-147. 42. Vogt, C. and Heinig, K. “Trace Analysis of Surfactants Using Chromatographic and Electrophoretic Techniques.” Fresenius Journal of Analytical Chemistry. Vol. 363. 1999. pp. 612-618. 43. Theile, B., Günther, K., Schwuger, M. “Trace Analysis of Surfactants in Environmental Matrices.”Tenside Surfactants and Detergents. Vol. 36(1). 1999. pp. 8-12, 14-18. 44. Bosdorf, V., Krüßmann, H. “Analysis of Detergents and Cleaning Agents with Thin-Layer Chromatography.”Fourth World Surfactants Congressional Asociacion Espanola de Productores de Sustancias para Aplicaciones Tensioactivas. Barcelona, Spain. Vol. 4. 1996. pp. 92-95. 45. Read, H. “Surfactant AnalysisUsing HPTLCand the Latroscan.” Proceedings of the International Symposium on In strumental High Performance Thin-Layer Chromatography. Third Edition. Ed. Kaiser, R. Institute of Chromatography. Bad Duerkheim. Federal Republic of Germany. 1985. pp. 157-171. 46. Henrich, L. H. “Separation and Identification of Surfactants in Commercial Cleaners.” Journal of Planar Chromatography — Mod. TLC. Vol. 5(2). 1992. pp. 103-117. 47. Buschmann, N., Kruse, A. “In-Situ TLC-IR and TLC-SIMS: Powerful Tools for the Analysis of Surfactants.”Comunicaciones Presentadias a las Jornadas del Comite Espanolde la Detergenteia. Vol. 24. 1993. pp. 457-468. 48. Novakovic, J. “Validationof a High Performance Thin-Layer Chromatographic Method for Trace Analysis for Some Generic Drugs Affecting Gastrointestinal Function.”Journal of AOAC International. Vol. 83(6). 2000. pp. 1507-1516. 49. Raghavan, R., Mulligan, J. A. “Low-Level (PPB) Determination of Cisplatin in Cleaning Validation (Rinse Water) Samples. I. An Atomic Absorption Spectrophotometric Technique.”Drug Device Industry Pharmaceutical. Vol. 26(4). 2000. pp. 423-428. 50. Davidson, C. A., Griffith, C. J., Peters, A.C., Fielding, L. M. “Evaluation of Two Methods for Evaluating Surface Cleanliness
Brittain, H. G. “Validation of Nonchromatographic Analytical Pharmaceutical Technology. Vol. 22(3). Methodology.” 1998. pp. 82-90. Ciurczak, E. “Validation of Spectroscopic Methods in Pharmaceutical Analyses.” Pharmaceutical Technology. Vol. 22(3). 1998. pp. 92-102. Swartz, M. E., Krull, I. S. “Validation of Chromatographic Methods.” Pharmaceutical Technology. Vol. 22(3). 1998. pp. 104-120. USP 23, United States Pharmacopoeia Convention.Rock ville, Maryland. 1995. Segretario, J., Cook, S. C., mbles, U C. L., et. al.“Validation of Cleaning Procedures for Highly Potent Drugs. II. Bisnafide.” Pharmaceutical Device Technology . Vol. 3(4). 1998. pp. 471-476.
– ATP Bioluminescence and Traditional Hygiene Swabbing.” Luminescence. Vol. 14. 1999. pp. 33-38. 51. Chawla, M. K. “Is It Clean?” Precision Cleaning. Vol. 8(6). 2000. pp. 36,38. 52. Meltzer, M., Koester, C., Steffani, C. “Criteria Evaluation for Cleanliness Phase 0.” Lawrence Livermore National Lab oratory. UCRL-CR-133199, PPG99-003. 1999. 53. Inampudi, P., Lombardo, S., Ruezinsky, G., et. al. “An
Special Edition: Cleaning Validation III
45
Control and Monitoring of Bioburden in Biotech/ Pharmaceutical Cleanrooms By Raj Jaisinghani Technovation & Greg Smith Encelle, Inc. & Gerald Macedo Med-Pharmex, Inc.
v his paper describes results of monitoring biotech clean-
}The
FDA rooms and a pharmaceutical has specific cleanroom equipped with an Electrically Enhanced Filtration (EEF) requirements and system that significantly reduces 1 airborne bioburden in cleanrooms. guidelines for bioThe EEF High Efficiency Partburden for iculate Air (HEPA) system traps various and kills bacteria and also im proves the filtration performance of a filter pharmaceutical media by two to three orders of operations and magnitude. In laboratory tests the EEF technology has been shown processes.~
T
to kill Staphylococcus epidermidis and Escherichia coli. These field test results support laboratory testing and show that basically there is no airborne bioburden in both a Class 10 room, with terminal HEPA in addition to the EEF, and in a Class 1000 room that utilizes only the EEF without any terminal HEPA filters. In the case of an old laboratory converted to a cleanroom, direct comparison of the EEF with respect to conventional 46
Institute of Validation Technology
HEPA fan filter units (FFUs) was possible. The results showed that at the same flow rate the EEF resulted in significantly lower bioburden as compared to the FFUs.
Background
The fundamental purpose of cleanrooms in the pharmaceutical, medical device, biotechnology, and hospital applications is to control the amount of bioburden due to both internal operations and transport from the air. From a particulate point of view, cleanrooms in these industries are classified and specified according to the same cleanroom standards (e.g., Federal Standard 209E) as in other industries. It is often assumed that the particle (total) concentrations will generally correlate to concentration of viable microorganisms. This may not always be valid. Hence, the concentration of viable organisms is also
Raj Jaisinghani, Greg Smith, & Gerald Macedo
directly measured – both at the work surfaces (or at the process) and in the air. Cleanrooms in these industries must meet separate standards for bioburden. The FDA has specific requirements and guidelines1 for bioburden for various pharmaceutical operations and processes. Similarly, the United States Pharmacopoeia (USP) and the European Union’s GMP2 guidelines give specific recommended limits for microbial contamination for each class of room. A cleanroom that meets the particle concentration requirements, but does not result in the desired
applied E. coli survived on the clean glass filter after four hours of airflow, keeping in mind that E. coli is not a hardy organism. Next about one gm colloidal of kaolin was added to theE. coli solution that was to be aerosolized. This time the recovery of E. coli was about 104 – 105 CFU/square inch of thefilter media. Similar tests with S. epidermidis recovered a little moreS. epidermidisthan with E. coli even without thecolloidal kaolin, due to the hardier nature ofS. epidermidis. With 1 gm of colloidal kaolin in the 25 mlS. epidermidis solution (in tryptic soy broth) the recovery of S. epider-
level of bioburden, willclearly be inadequate. midis was about 105 – 106 CFU/square inch of filter One of the main obstacles in achieving the required media. Tests with airflow continuing for seven hours bioburden levels is that the measurement of bioburden (following the aerosol) did not result in any significant is time consuming. Typically, bioburden measure- reduction in bacteria recovery. This result suggests ment involves sampling, incubation, and counting of that, even in normal environments, bacteria can surcolonies. This is a time consuming pro cess and thus vive or grow on the filters. As the trend towards using “real” time monitoring is not possible. Thus, it is not HEPA cleanroom filters for longer periods continues, always possible to relate higher incidents of bioburden the possibility of bacterial growth on the filter, and thus to operating events. Recently, however, ultraviolet the rise in the airborne bioburden, also increases. 3 (UV) fluorescence (cf. Seaver and Eversole , Pinnick et al.4) technology has made it possible to achieve real EEF Technology time monitoring of particles of biological srcin. This technology will find increasing use in the real time Jaisinghani7 has played a significant role in the monitoring of air in hospitals, cleanrooms, and mili- development and commercialization of EEF techtary nuclear, biological and chemical (NBC) warfare nology. The most recent version (seeFigure 1) of protection systems – as a real time supplement to the this technology7 maintains the filter under an ionstandard methods of determining bioburden. As this izing (as opposed to a simple electrostatic field) happens, more attention will be focused on cleanroom contamination control systems – currently mainly Figure 1 mechanical filtration. t’ ig One problem with mechanical filters is that under eeF tg certain conditions common bacteria caught on the Ionizing Filter filter can start growing on the filters, grow through Wires 5,6,7 the media, and start shedding into the room. The well-known case of theLegionnaire’s outbreak at the Flow veterans convention in Philadelphia has been attributed to this phenomena. In that case the filters were supDownstream posedly in a wet state. Generally, it is accepted that Ground Electrode bacteria is difficult to grow on clean glass fiber filter media, used in HEPA filters, under normal humidity Control conditions. However, since the function of these filIonizer Electrode ters is to capture all particulate contamination, filters Electrode eventually get dirty. The experiments conducted by Jaisinghani et al.8 show that very little contaminant is H.V. needed for growth ofStaphylococcus epidermidisand Supply Escherichia coli on HEPA glass filters. In their experi8 ments Jaisinghani et al. found that very little of the Special Edition: Cleaning Validation III
47
Raj Jaisinghani, Greg Smith, & Gerald Macedo
field. Another higher intensity ionizing field charges incoming particles, stabilizes the electrical fields, and increases the safety and reliability by ensuring that no spark over can occur towards the filter. This method provides two fundamental benefits:
confirm the initial viable concentration of bacteria. The rehydrated culture was then sprayed onto the filter using a Meinhard nebulizer, which was placed eight inches from the center of the filter. A control assay was performed to determine the amount of viableS. epidermidison the filter, without 1. Bacteria are killed as they pass through a first application of high voltage. The bacteria were sprayed high intensity ionizing field and then killed onto the filter as previously described, and the temperas they are subjected to continuous ionizing ature and humidity were monitored every 15 minutes radiation when they are trapped on the filter. for four hours or seven hours during which the airflow This inhibits growth of bacteria on the filter. was on. The relative humidity was held between 452. The same ionizing fields enable penetration reduction by about two to three orders of magnitude.
Since the cost of the additional electrical components is partially offset by the increase in filtration performance (either higher flow at the same pressure drop and filtration efficiency or lower pressure drop at the same flow and efficiency, as compared to mechanical filtration of the same size) this is a highly cost effective method for the control of bioburden.
Laboratory Evaluation of the EEF Jaisinghani et al.8 have demonstrated the bactericidal properties of the EEF under laboratory conditions. This study, conducted at Virginia Polytechnic Institute, is summarized in this section. Experimental Methods S. epidermidis was grown in Tryptic Soy Broth
55% using a Kaz steam vaporizer. At the end of each control run three pieces of the filter were extracted using a sterilized scalpel and forceps. The pieces of filter were approximately one square inch on the face of the filter, which when unfolded measured approximately 28 square inches of filter material. Filter pieces were removed from the center of the filter, directly above the center, and directly below the center. The samples were cut into small pieces and placed into 10 ml of sterile phosphate buffered saline (PBS) pH 7.4 in a Nasco Whirl-Pak bag. The bags were then processed in a Tek Mar Stomacher Lab-Blender 80 for one minute. Each sample was then serially diluted ten-fold to 10-2, 10-3, and 10-4, then spread on CBA plates to determine the number of viable bacteria per sample filter piece. Similar tests were then conducted by applying high voltage to the EEF. In addition to monitoring the temperature and humidity, the current was also monitored at fifteen minute intervals during the four or seven hour period of airflow with the applied high voltage on.
(TSB) to a concentration of 3 x 109 colony forming units (CFU)/ml. The culture was lyophilized Results and Discussion in Wheaton vials in 5 ml aliquots – 1.5 x 1010 CFU The results are summarized in Figure 2. In the per vial. All vials were stored in a desiccator at 4 absence of any voltage applied to the EEF unit (i.e., – 6ºC. control tests), viable bacteria were recovered from 5 Pleated 15.24cm by 15.24cm by 5cm (6” x 6” x one square inch of filter in the range of 1 x 10 CFU 2”) deep filters were first coated with colloidal kaolin to 2 x 106 CFU. Counts greater than about 3 x 106 and TSB using an Aztek airbrush. The airbrush cup CFU were too crowded to be accurately counted and was filled with 1g kaolin suspended in 25ml TSB were considered to be too numerous to count. When and sprayed onto a filter inside a laminar flow hood high voltage was applied for four hours, the majority and allowed to air dry before being used. The pleated of the bacteria were killed. The kill rate increased filters were placed in a miniature version of the EEF. with increased voltage or with the first applied field One vial of lyophilizedS. epidermidiswas resuspend- strength (applied voltage divided by the distance of ed with 1 ml of sterile distilled water. A small aliquot the ionizer wires from the control ground electrode – of this suspension was serially diluted ten-fold to -810 see Figure 1), V/d1. At a field strength (V/d1) of 4.2 and plated on Columbia Blood Agar (CBA) plates to kV/cm, there was no growth after 24 hours of incu48
Institute of Validation Technology
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Figure 2
eeF Bd t s ug S. epidermidis F
ib t
eeF ep t
h
h
c eeF
eeF Fd sg
ag c
(/d1) k/
c
#/s.
Control
24.00
0.00
0.00
1.00E+06
Control
24.00
0.00
0.00
1.02E+05
No additional growth After 24 Hours
EEF
24.00
4.00
4.64
00.0E+00
100% Killed
EEF EEF
24.00 24.00
4.00 4.00
3.99 4.24
3.44E+02 0.00E+00
99.93% Killed 100% Killed
EEF
24.00
4.00
4.50
0.00E+00
Some growth
EEF
24.00
4.00
4.20
0.00E+00
After 48 Hours
EEF
24.00
4.00
4.20
6.26E+03
EEF
48.00
7.00
4.20
5.44E+02
99.9% Killed
EEF
48.00
4.00
4.80
2.16E+02
99.95% Killed
EEF
48.00
4.00
4.20
3.51E+03
99.3% Killed
bation. After 48 hours, there was either no growth or small (in size and in number) colonies grown. These small colonies were identified asS. epidermidis, and were identical in biochemical profile as the isolate used in the tests. It was concluded that four hours at 4.2 kV/cm (V/d1) did not completely kill theS. epidermidis. If the bacteria were not all killed, some of them were damaged sufficiently so that no growth or very limited growth could occur after 24 hours incubation. When the ionizing time was increased to seven hours, over 99% of the bacteria (as compared to the control) were killed. When the applied field strength, V/d1, was increased to 4.5 kV/cm or higher, no growth occurred on any of the filter pieces except for one experiment. This exception may have occurred because the starting dose of bacteria for this experiment was three times higher than for the control and up to 10 times higher than for any other experiment. Nonetheless, there were still three to four logs of killing using an applied field strength, V/d1 of 4.5 kV/cm or higher, as compared to the control experiments. It should be noted that, in practice, bacteria caught on the filter are held within the ionizing field for an almost infinite amount of time, thus receiving an almost infinite radiation dosage. Hence, in practice, the killing efficiency should be higher even at lower field strengths. Similar results were obtained usingE. coli in a previ-
98.75% Killed
Figure 3
md 3001 eeF F
ous study conducted with the EEF at the University of Wisconsin.
Field Results in Cleanrooms Model 3001B or Model 1001B BIO PLUS® EEFs (Figure 3) manufactured by Technovation Systems, Inc., were used in the cleanroom discussed here. Special Edition: Cleaning Validation III
49
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Both models have a flow rating of about 3000 scfm without attached ductwork.
flow rate utilized here (22 fpm average air velocity) was on the lower end of flow normally used in Class Comparison to FFU in a Converted Cleanroom 1,000 cleanrooms. This room will simply be referred Jaisinghani et al.8 have conducted a field study com- to, henceforth, as“the Encelle cleanroom.” paring FFUs to an EEF in asmall laboratory converte d All processing and manufacturing conducted within to a cleanroom. This will be referred to as the “older the cleanroom areas are done aseptically. Workers are Encelle cleanroom.” (Encelle, Inc., Greenville, NC.) gowned in sterile coveralls, shoe covers, goggles, or Encelle had four conventional HEPA fan filter units face-masks with shields, hair nets, and sterile gloves. (FFUs) installed in this tissue culture laboratory, prior The Class 1K cleanroom and clean Class 10K surto replacing these with one Model 1001 EEF. One rounding zone are cleaned daily with a monthly rotation model 1001 provides HEPA filtered air at about the of sterile chemicals using cleanroom equipment and 3 same total flow (approx. 4250 m /h (2500 scfm) in this case) as the four FFUs. This allowed direct evaluation of the effect of EEF on the bioburden in the room, under field conditions. Airborne bioburden in the room was reduced by as much as 75% after switching to the EEF system. The airborne bioburden in the Class 10K 3 room was 0.021 cfu/ft3 (no process) and 0.392 cfu/ft (in process) after installation of the EEF filter. Figure 4 shows the Federal Standard 209E, USP, and European Union recommended airborne bioburden and particulate concentration for various class cleanrooms. Clearly, from a bioburden perspective Encelle’s older Class 10K room performs (at rest) at a level equivalent to a Class 100 room – without incurring the higher cost associated with building a Class 100 room. Most of this benefit should be attributed to the EEF filter system.
® trained personnel. Disinfectants include Hypochlor , ® Process LpH®, Process Vesphene , and as needed, treat® ments with a spore-killing agent called SporKlenz .
Sampling Methods
An environmental monitoring program has been designed to establish the standards and limits that are acceptable to the facility management and to regulatory agencies that will audit the manufacturing within the cleanroom environments. Daily activities for monitoring include temperature and pressure readings. Relative humidity is also reported on sampling days. Surface samples are collected on a routine basis using only sterile supplies. Surfaces monitoredclude in floors, equipment, walls, and ceilings. A five percent sheep’s blood agar plate (three inch diameter) is swabbed with a sterile, moist, cotton swab after New Class 1000 Biotech Tissue sampling various surface areas. Plates are labeled and Culture Cleanroom incubated at 37ºC, with five percent carbon dioxide for 72 hours. Colony forming units that grow are counted Facility Description and identified using standard microbial techniques. A 1,300 square foot Class 1,000 cleanroom and Particle counting is conducted biweekly or as 900 square feet ofClass 10,000 surrounding space was needed for monitoring during critical processes. A constructed at Encelle, Inc., Greenville, NC facility. It Biotest® APC Plus particle counter measures concen® utilized eight 3001B BIO PLUS EEF filters. The total tration at particle sizes of 0.3, 0.5, 1.0, and 5.0µm.
Figure 4
rd f ab Bbd f v c c c
P
10K
0.5 Um particles #/ft3 <10K
10K
CFU/ft3 <2.83
1K
0.5 Um particles #/ft3 <1K
1K
CFU/ft3
100
0.5 Um particles #/ft3
100
CFU/ft3
50
209e
eu
usP
<10K <2.5 <1K NA <100
Institute of Validation Technology
NA
<100 <0.028 at rest; <0.283 Process
<0.1 Process
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Data is collected in nearly 200 areas within the filtered-air areas. These areas are categorized by process and particle counts are reported to the facilities manager for evaluation and disposition if warranted. Data is downloaded via an RS-232 port for digital documentation of these counts. Microbial air sampling is performed in parallel with particle counting to provide data on airborne viable particulate counts and comply with Federal Standard 209E Cleanliness classes for cleanrooms ® and clean zones. A Biotest Centrifugal Air Sampler
to a Class 100 cleanroom in operation. It should be noted that the cleanroom certified as Class 100 at rest (i.e., without personnel). Similarly, the design Class 10,000 area is functioning as a borderline Class 1,000 in operation. It should be noted again, that the airflow rate used in this Class 1,000 cleanroom (22 fpm) was on the low end for a normal Class 1,000 cleanroom. The high performance of this room can be attributed to the high degree of ceiling coverage (i.e., Airflow is highly distributed throughout the room) which is an inherent ® feature of the Technovation BIO PLUS EEF system.
collects 500 liters of air in each location on sterile tryptic soy agar strips that are designed for this type Results – Airborne Microorganisms of sampler. Strips are labeled and incubated similarly Figure 7 shows the results of airborne microorto the surface agar plates. Classification and identifi- ganism monitoring in the Class 1,000 cleanroom. cation are performed using the standards described in The results clearly indicate that, from the perspecthe current edition of the USP. tive of airborne bioburden, the Class 1,000 area is USP standards9 for microbial growth follow in performing at a level that is to be expected for a Class Figure 5. 100 to Class 10 level. Note that (by comparing to the particulate data inFigure 8) the airborne bioburden Results – Particle Concentration does not correlate to the particulate concentration. The particle concentration measurements are During the months of December and January, the shown in Figure 6. sub-micron particulate concentration actually went up From the perspective of particle counts alone, thewhile the airborne bioburden was reduced. design Class 1,000 cleanroom is functioning very close Figure 8 shows the results of airborne microorganism monitoring in the Class 10,000 cleanroom. Figure 5 Note again that the bioburden does not relate to the particulate concentration. One probable reason usP sdd f Bbd for this is that the bacterium are larger than the size c of the sub-micron particles being monitored. Also c 100 note that on the average the Class 10K area performs r between a USP Class 100 and a USP Class 10K from 3 a bioburden perspective. Air 0.1 cfu/ft Surface Gowns
3 cfu/plate* 5 cfu/plate
c 10,000 r Air Surface Gowns
3 0.5 cfu/ft 5 cfu/plate (10 from floor) 10 cfu/plate
c 100,000 r Air Surface Gowns * 2 in2 surface
2.5 cfu/ft 20/plate 30/plate
3
Results – Surface Microorganisms Figure 9 shows the results of surface microorgan-
ism monitoring in the Class 1,000 cleanroom. The surface concentrations of bacteria are almost zero throughout the cleanroom suite. This can be attributed to: The cleaning protocols instituted at Encelle The use of the disinfecting compounds n The low airborne bioburden n n
Observations from the Encelle Cleanroom Monitoring
1. The Encelle cleanroom operates significantly Special Edition: Cleaning Validation III
51
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Figure 6
P c m e t c lb a P c (p f sz – > c 10,000 Dg ar apprx. sq. F.
3
)
11/14/99 11/14/99 12/1/99 12/1/99 12/31/99 12/13/99 1/6/00 1/6/00 1/28/00 12/28/00 0.3 m 0.5 m 0.3 m 0.5 m 0.3 m 0.5 m 0.3 m 0.5 m 0.3 m 0.5 m
Mechanical Corridor
235
1959
506
786
602
1163
903
801
899
852
512
Side Corridor
555
2240
642
4559
1035
3598
762
4058
687
7989
1422
Water Filtration Area
171
3649
2598
4657
2429
11406
4312
8350
4368
6378
1295
Labware Processing
212
343
102
981
719
1233
1798
243
326
618
530
Gowning Room
139
236
87
853
1063
1506
913
2874
2808
957
235
40
257
88
450
539
2461
2463
3887
6064
1451
595
482
671
683
1065
1187
1859
1123
2525
1014
765
Materials Pass Through
t s F – >
1312
Actual Area Classification c 1,000 Dg ar Specialized Cleanrooom
152
158
66
223
145
350
337
313
379
754
237
Formulations Mfg. Area
142
208
24
84
52
415
228
322
307
360
126
Product Testing Area
132
28
5
9
9
43
49
19
74
19
76
Product Finishing Area
127
264
26
129
49
474
139
400
179
472
184
Product Mfg. Area
145
293
83
250
87
534
180
639
698
829
Refrigerator/Freezer Storage 212
t s F – > Actual Area Classification
262
55
233
157
52
83
692
427
335
659
731 381
910
67
43
better than the design classification although the flow rate (average room velocity = 22 fpm) used is at the low end of what is normally used in such a cleanroom. This is due to the higher distribution of flow rate – an inherent feature of the BIO PLUS® filter system. 2. The airborne bioburden in both the Class 1K and 10K areas is lower than what would be expected for such rooms based on USP recommendations. The Class 1K room has the bioburden of what would be expected (based on USP) for a Class 100 room. Coupling this observation to the laboratory studies on the bactericidal properties of the EEF technology and the direct comparison with respect to conventional FFU HEPAs, it may be inferred that 52
832
Institute of Validation Technology
147
271
118
329
172
289
the low airborne bioburden is due to the BIO PLUS® EEF filters. 3. The surface bio contamination is almost non existent in the Class 1K cleanroom. This may be attributed to the good cleaning practices used at Encelle and due to the low airborne bioburden in the suite. 4. The airborne bioburden seems to be lower in the winter months, although the room temperature is held constant at 66ºF. This may be due to lower humidity in the winter months.
New Class 10 Pharmaceutical Cleanroom Facility Description
A 12’x20’ Class 10 cleanroom (including a
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Figure 7
ab Bbd e c 1,000 c n F’ mb a sp r cd B P cfg a sp Dg c 1,000
11/14/99 11/30/99 arg f a. f/f3 f/f3/24hr* f/f3/72hr* arg f a. f/f 3 f/f3/24hr* f/f3/72hr*
Device Testing
0
0
0
0
0
0
0
0
Coating
0
0
0
0
1
0.028
0.028
0.028
Formulations Device Manufacturing
0 3
0 0.085
0 0.085
0 0.085
0 2
0 0.057
0 0.057
0 0.057
Refrigeration
0
0
0
0
0
0
0
0
Isolation
3
0.085
0.085
0.085
15
0.425
0.425
0.425
Class 1,000 Average
1
0.028
0.028
0.028
3
0.085
0.085
0.085
Dg c 1,000
12/13/99 1/28/00 arg f a. f/f3 f/f3/24hr* f/f3/72hr* arg f a. f/f 3 f/f3/24hr* f/f3/72hr*
Device Testing
0
0.000
0
0
0
0.000
0
0
Coating
0
0.000
0
0
0
0.000
0
0
Formulations
0
0.000
0
0
0
0.000
0
0
Device Manufacturing
0
0.000
0
0
0
0.000
0
0
Refrigeration
0
0.000
0
0
0
0.000
0
0
Isolation
0
0.000
0
0
0
0.000
0
0
Class 1,000 Average
0
0.000
0
0
0
0.000
0
0
The time period refers to the incubation time in hours.
4’x12’ Class 10K gowning room) was constructed athypochlorite solution. ® MedPharmex in Pomona, CA using two BIO PLUS Model 3001B filters with eight terminal 2’x4’ HEPASampling Methods filters. The Model 3001Bs were used for the 12’x16’ Air sampling was done using Tryptic Soy Agar Class 10 inner room. The resultant average roomlocve (TSA) and Sabouraud Dextrose Agar (SDA). The ity in the Class 10 area was 24 fpm (4500 scfm). TheTSA values reflect total bacterial counts while the design specification for the room was Class 100. ThisSDA reflects molds and yeast, although it contains airflow was much lower than used in a Class 100 clean-no bacterial inhibitors. In some cases Rose Bengal room – normally, with conventional single terminalAgar (RBA) was used. This reflects a better value for HEPAs, at least 40 fpm average room velocity is usedmolds/yeast since the RBA contains bacterial growth in a Class 100 room. However, due to the double HEPAinhibitors. Surface monitoring was done using 24-30 filter system (each Model 3001B powered Airflowcm2 RODAC plates with TSA and SDA. The TSA through four terminal HEPAs) the cleanroom easilyplates were incubated for a minimum of 48 hours at classified as Class 10 as per Federal Standard 209E.32.5 +/- 2.5ºC while the SDA plates were incubated This resulted in significant energy savings. The roomfor a minimum of 72 hours at 22.5 +/- 2.5ºC. was validated for bioburden initially and then has been shut down since the facility is now being moved toResults – Airborne Microorganisms a new location. The facility was cleaned with 0.25% The gowning room was sampled in two zones Special Edition: Cleaning Validation III
53
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Figure 8
ab Bbd e c 10K a n F’ mb a sp r cd B P cfg a sp Dg c 10,000 11/14/99 11/30/99 3 arg f a. f/f f/f3/24hr* f/f3/72hr* arg f
arg f/f3 f/f3/24hr* f/f3/72hr*
Water Filtration Area
30
0.850
0.850
0.850
2
0.057
0.057
0.057
Side Corridor
3
0.085
0.085
0.850
3
0.058
0.085
0.085
Manufacturing Corridor Labware Processing
19 5
0.538 0.142
0.538 0.142
0.538 0.142
11 4
0.312 0.113
0.312 0.113
0.312 0.113
Gowning Room
0
0
0
0
3
0.085
0.085
0.085
Materials Pass Through
3
0.085
0.085
0.085
11
0.312
0.312
0.312
Class 10,000 Average
10
0.283
0.283
0.283
5.67
0.160
0.160
0.172
Dg c 10,000
12/13/99 1/28/00 arg f a. f/f3 f/f3/24hr* f/f3/72hr* arg f a. f/f 3 f/f3/24hr* f/f3/72hr*
Water Filtration Area
2
0.057
0.057
0.057
2
0.057
0.057
0.057
Side Corridor
2
0.057
0.057
0.057
0
0
0
0
Manufacturing Corridor
1
0.028
0.028
0.028
0
0
0
0
Labware Processing
0
0
0
0
0
0
0
0
Gowning Room
0
0
0
0
0
0
0
0
Materials Pass Through
2
0.057
0.057
0.057
0
0
0
0
1.17
0.033
0.033
0.033
0.333
0.009
0.009
0.009
Class 1,000 Average
The time period refers to the incubation time in hours.
while the Class 10 cleanroom was sampled in five zones. All plates (TSA and SDA) were negative (i.e., zero counts) in all the areas. The Class 10 area was also sampled using RBA and once again the results were negative – zero counts. Results – Surface Microorganisms
The surface measurements were made before and after cleaning the newly constructed cleanroom. The results are shown inFigure 10. The 0.25% Hypochlorite cleaning is obviously very effective in eliminating surface bacteria. Observations From the Medpharmex Cleanroom Validation
1. The new Encelle Class 1000 and the Med Pharmex Class 10 room have about the same airflow average velocity. From the particulate point of view the MedPharmex room operates at Class 10 simply 54
Institute of Validation Technology
because of the double HEPA filter system used. The MedPharmex cleanroom validates as a Class 10 cleanroom, although the airflow used was lower than what is normally used in a Class 100 room. 2. It should be noted that the MedPharmex room was simply validated and then shut down in order to move it to an adjacent facility, while the Encelle room is being continuously monitored and is operational. How ever, from the point of view of airborne bioburden, after the first month of operation the Encelle Class 1000 room operates at an equivalent level as the MedPharmex Class 10 room – with essentially zero airborne bacterial counts. The low bioburden benefit to Encelle (this Class 1000 room is operating at essentially zero airborne bioburden) may be attributed to the o bactericidal properties of the EEF system.
Raj Jaisinghani, Greg Smith, & Gerald Macedo
Figure 9
n F sf c s Dg cf 10,000 arg br f f/p gr 72 hr D – > 11/14/99 11/30/99 12/13/99 01/28/00 Water Filtration Area 0 0 2 0 Side Corridor 1 0 0 0 ManufacturingCorridor 0 0 0 0 Labware Processing 1 0 0 0 Gowning Room 0 0 0 0 Materials Pass Through 0 0 0 0 Dg cf 1,000 arg br f f/p gr 72 hr D – > 11/14/99 11/30/99 12/13/99 01/28/00 Device Testing 0 Coating 0 Formulations 0 Device Manufacturing 0 Refrigeration 0 Isolation 0 c 255
0 0 0 0 0 0 210
0 0 0 0 0 0 134
0 0 0 0 0 0 104
Figure 10
sf Bbd c 10/10K s Bfr cg tsa
Gg Table-gowning Wall-gowning c 10 Tank Fill Filter table Wall
sDa c
tsa c
afr cg sDa c
Greg Smith is facilities manager at Encelle, Inc. He holds a B.A. in Psychology from West Virginia University and a B.S. in Chemistry from East Carolina University. Smith and has assisted in the development medical devices has five years experience asof a hospital pharmacy aseptic compounding technician. He can be reached by phone at 252-355-4405 or by e-mail at
[email protected]. Gerald Macedo has a B.S. degree in Pharmacy and an M.S. in Pharmaceutical Sciences. He has over 30 years experience in pharmaceutical manufacturing, with extensive experience in the manufacture of sterile injectables. He has served as head of manufacturing, quality control, quality assurance, research and development, and regulatory affairs. Macedo currently heads Med-Pharmex, Inc., a pharmaceutical manufacturing company. He can be reached by phone at 909-593-7875 or by fax at 909-593-7862.
References
(c p 25 2 roDac P)
ar
the University of Wisconsin. Jaisinghani has extensive research experience in air and liquid filtration, colloid and aerosol science, fluid mechanics, heat transfer, and physical surface chemistry. He holds 10 patents and has many publications in technical journals and handbooks. He can be reached by phone at 804-744-0604, by fax at 804-744-0677, or by e-mail at
[email protected].
c
22
5
0
0
3
0
0
0
0
1
0
0
21
12
0
0
9
4
0
0
1
0
0
0
About the is Authors Rajan (Raj) Jaisinghani a chemical engineer with thirty years of research, product development, and business development experience. Jaisinghani holds a B.S. from Banaras Hindu University, India, and an M.S., with additional graduate work, from
1. FDA. “Guideline on Sterile Drug Products by Aseptic Processing.” Rockville, MD. 2. EU. 1998. “The Rules Governing Medicinal Products in the EU.” Good Manufacturing Processes 4. Luxembourg. 3. Seaver, M. and Eversole, J.D. 1996. “Monitoring Biological Aerosols Using UV Fluorescence.” Proceedings 15th Annual Meeting AAAR. October. Orlando, FL: 270. 4. Pinnick, R.G., Chen, G., and Chang, R.K. 1996. “Aerosol Analyzer for Rapid Measurements of the Fluorescence Species of Airborne Bacteria Excited with a Conditionally Fired Pulsed 266 nm Laser.” Proceedings 15th Annual Meeting AAAR. October. Orlando, FL. 5. Rhodes, W.W., Rinaldi, M.G., and Gorman, G.W. 1995. “Reduction and Growth Inhibition of Microorganisms in Commercial and Institutional Environments.”Environmental Health 12 (October). 6. Tolliver, D.I. 1988. “Domestic and International Issues in Contamination Control Technologies.”Microcontamination 6, no. 2 (February). 7. Jaisinghani, R. A. U.S. Patent 543,383. 4 April 1995. 8. Jaisinghani, R.A., Inzana, T.J., and Glindemann, G. 1998. “New Bactericidal Electrically Enhanced Filtration System for Cleanrooms.” Paper presented at the IEST 44th Annual Technical Meeting. April. Phoenix, AZ. 9. “Microbial Evaluation of Cleanrooms and Other Controlled Environments.” United States Pharmacopoeia, <1116>, p. 20992106.
Special Edition: Cleaning Validation III
55
A Cleaning Validation Program for the ELIFA System By LeeAnne Macaulay, Jeff Morier, Patti Hosler, & Danuta Kierek-Jaszczuk, Ph.D. Cangene Corporation
v
T
he Enzyme-Linked Immunofiltration Assay (ELIFA) provides high sensitivity of detection with rapid results. For this reason we developed a very sensitive, semi-quantitative ELIFA to determine IgA in therapeutic Win Rho SDF™ immunoglobulin. In the course of the development we no ticed that non-uniform and unusually high background (blank) responses, that occurred infrequently, greatly interfered with the test results obtained. We hypothesized that such background responses resulted from inadequate cleaning of the ELIFA apparatus. Accordingly, a cleaning program for the apparatus has been devised and validated. In this paper the results supporting the hypothesis will be presented, and the rationale and core aspects of the developed program delineated.
Cleaning Validation Programs for Research and Development?
}The
establishment of Cleaning Validation Programs (CVP) in the pharmaceutical industry is dictated by the regulatory requirements to develop and observe, in a fully documented way, effective cleaning procedures.~
The establishment of Cleaning Validation Programs (CVP) in the pharmaceutical industry is dictated by the regulatory requirements to develop and observe, in a fully documented way, effective cleaning procedures. Regulatory guidelines for validation of cleaning pro cesses1 are meant to sup56
Institute of Validation Technology
port individual CVPs and enforce compliance. The guidelines and programs may cover a plethora of different types of equipment but they usually refer to equipment used in the manufacture, processing, holding, filling, and packaging of raw materials, intermediate/final products, and associated components. The guidelines do not refer to equipment used in Research and Development (R&D), and to our understanding, there is no regulatory requirement for the development of CVPs for equipment used in these areas. The 2 is a Easy-Titer™ ELIFA system small, microtiter format compatible apparatus developed and manufactured by Pierce Chemical Company. As shown in Figure 1 (adapted from Product Instructions, Pierce Chemical Company) the apparatus utilizes a nitrocellulose membrane (NC) sandwiched between the sample application plate and vacuum collection chamber. Similar to the widely used Enzyme-Linked Immunosorbent
Assay (ELISA), theofELIFA an immunoassay well suited for testing multipleis sam3 ples over a range of serial dilutions. In the ELIFA, the immunological reaction between NC immobilized ligand and ligand-specific analyte in the test sample followed by an enzymatic reaction with a chromo-
LeeAnne Macaulay, Jeff Morier, Patti Hosler, & Danuta Kierek-Jaszczuk, Ph.D.
Figure 1
epdd v f e-t™ eliFa s
Thumb Screws
Sample Application Plate
Nitrocellulose
Clamp
Microtiter Plate Vacuum Relief Valve
Pump Tubing Port
Gaskets
Tubing
Position Stops (Acrylic Balls) Transfer Cannula
Collection Chamber
Membrane Support Plate
Guide Pins
genic substrate gives rise to colored dots. The color ELIFA CVP – Approaches and Hallmarks intensity of the individual dots varies proportionally to the amount of the analyte in the samples and dots A body of experience at Cangene with validaproduced by the samples devoid of analyte (blanks or tion5,6 or cleaning7 programs, as well as manufacbackground) are very pale or even colorless. We used turer’s cleaning instructions for the ELIFA system the ELIFA system to research and develop a screen-(Figure 2, adapted from Product Instructions, Pierce 4 The developed IgA ELIFA Chemical Company) was the foundation when develing assay for human IgA. will be used for testing of the licensed WinRho SDF™ oping the ELIFA CVP. Among others, the develtherapeutic, hence, itsperformance characteri stics need oped program addressed the following: to be established and validated. In pre-validation studies, however, we observed that the developed ELIFA nSpecific design of apparatus, its individual parts lacked reproducibilit y. The color of the blank dots varand accessories that require cleaning ied sometimes from experiment to experiment or, even nDisassembling and re-assembling the unit within the same experiment, from well to well. We also before and after cleaning observed that the color of dots produced by the replicated test samples occasionally varied. We postulated Figure 2 that the observed variability is a result of external concg f e-t™ tamination carried over from previous experiment(s). An inadequately cleaned ELIFA apparatus would then be the cause of obscured test results. We, therefore, decided to develop a CVP for the ELIFA apparatus before proceeding to assayvalidation.
eliFa s Clean all of the pieces to the Easy-Titer™ ELIFA System unit in a two percent PCC-54 solution and then rinse with distilled water. The unit may also be soaked in the PCC-54 solution to remove stains from the unit caused by the substrate solution. Special Edition: Cleaning Validation III
57
LeeAnne Macaulay, Jeff Morier, Patti Hosler, & Danuta Kierek-Jaszczuk, Ph.D.
nCleaning
operations procedures including cleaning vessels, agents and utensils nCompatibility of cleaning agents with equipment and assay nDecontaminating abilities of cleaning agents nSampling on cleaned equipment nAnalytical methods for monitoring of cleaning processes nStorage of cleaned parts n Inspection of apparatus for cleanliness before use nCleaning
tion with two ELIFA experiments; the standard IgA ELIFA and the mock ELIFA. Such a combination of analytical methods allowed for instantaneous monitoring of the effectiveness of the cleaning pro cess. 4 Standard IgA ELIFA method involved testing 96 replicates of a sample at a high (worst-case condition) IgA concentration, which were applied into 96 wells of the ELIFA apparatus. As expected, these experiments invariably produced highly colored dotsFigure ( 4). A tested cleaning regimen (procedure 1 or 2 in Figure 3) followed by the second mock experiment was then
nRecording
and documenting the cleaning procedures nEstablishing acceptance criteria, and nMaintaining cleaning records
executed. The mock experiment involved the use of the diluting buffer in lieu of a sample with high IgA concentration that was also applied into 96 wells of the ELIFA apparatus. Providing that the cleaning regimen was effective, the mock experiment should not produce colored dots, as there was no specific analyte Strategy for Validation of that could attach to the immobilized ligand to facilitate Cleaning Procedures subsequent enzymatic and color reactions. The results Two cleaning procedures (procedure 1 and 2 in obtained show that whereas Procedure 1 did not Figure 3) utilizing either enzyme or detergent-based remove the contaminants from preceding experiments cleaning agents were developed and tested in conjunc- well enough (Figure 5), procedure 2 was fully effec-
Figure 3
cg Pd 1 d 2 Pd 1
Pd 2
Disassemble bysampl first removing theplate, thumb screws on the the top unit of the e application then removing the application plate and top gasket, and finally unclamping the membrane support plate from the collection chamber.
Disassemble thetop unit removing the thumb located on the of by thefirst sample application plate,screws then removing the application plate and top gasket, and finally unclamping the membrane support plate from the collection chamber.
Rinse all parts for two minutes under running Reverse Osmosis (RO) water.
Rinse all parts for two minutes under running ROwater.
Immerse them into a vessel with two percentTERG-A- Immerse them into a vessel with five percent RBS10 solution ZYME (Alconox Inc., New York, NY, U.S.A.) solution and at 50ºC and wash for five minutes by agitating the vessel. wash for five minutes by agitating the vessel. Rinse all parts for two minutes under RO water.
Rinse all parts for two minutes under RO water.
Clean all 96 wells of the sample application plate with TOC swabs by dipping the swabs into the detergent solution, inserting them into wells once from top and once from the bottom, and swabbing the inner part of each well by turning the swab first to the right and then to the left.
Clean all 96 wells of the sample application plate with TOC swabs by dipping the swabs into the detergent solution, inserting them into wells once from the top and once from the bottom, and swabbing the inner part of each well by turning the swab first to the right and then to the left.
Clean all 96 wells of the top gasket in a similar way.
Clean all 96 wells of the top gasket in a similar way.
Raise all 96 cannulas on the membrane support plate and soak the plate for five minutes in the detergent solution.
Raise all 96 cannulas on the membrane support plate and clean them with TOC swabs by dipping the swabs into the detergent solution and swabbing the surface of individual cannulas and also spaces between cannulas and bottom gaskets.
Rinse each part and the spaces between the bottom gasket and the membrane support plate for two seconds under running RO water.
Rinse each part and the spaces between the bottom gasket and the membrane support plate for two seconds under running RO water.
58
Institute of Validation Technology
LeeAnne Macaulay, Jeff Morier, Patti Hosler, & Danuta Kierek-Jaszczuk, Ph.D.
Figure 4
Figure 5
iga eliFa r obd f t sp cg h iga c f 2µg/l
iga eliFa r obd f rpd t sp Dpd f h iga.
The experiment was performed in an apparatus cleaned with TERG-A-ZYME (Procedure 1).
Figure 6
tive (Figure 6). Procedure 2 was then validated, in two independent experiments performed by two analysts. It was shown that it invariably leads to results similar to those presented inFigure 6.
iga eliFa r obd f rpd t sp Dpd f h iga.
Assessment of the Effectiveness of the Validated Procedure The Total Organic Carbon (TOC) method is widely utilized in industrial CVPs as it measures low levels of carbon and is compatible with swab 4 sampling techniques. The standard IgA ELIFA followed by the validated cleaning procedure and swab sampling of the surface of three randomly selected wells were, therefore, used to assess the cleanliness of the apparatus by standard TOC. A procedure used at Cangene8 was followed. The results obtained confirm that the validated cleaning procedure was fully effective as the carbon concentration determined in
The experiment was performed in an apparatus cleaned with RBS (Procedure 2).
Figure 7
r F t og cb a ( ppb) smpe numbe
smpe rep nme 1
rep
rep 2
aege
sD
Pee
3
cv
1
Well B11
277.8
271.7
268.9
272.8
4.55
1.67
2
Well D8
235.4
234.3
225.9
231.8
5.20
2.24
3
Well G2
228.4
219.1
255.3
234.3
4.73
4
Water
185.2
165.4
167.4
172.7
10.9
8.03 6.29
Special Edition: Cleaning Validation III
59
LeeAnne Macaulay, Jeff Morier, Patti Hosler, & Danuta Kierek-Jaszczuk, Ph.D.
water extracts of test samples was only slightly great- a B.Sc. degree at the University of Winnipeg and received a diploma in Chemical and Bioscience er than that of water used for extraction Figure ( 7).
Implementing of the ELIFA CVP
Technology from Red River College in Winnipeg. She has experience in QC/QA Laboratories in the areas of microbiology and biochemistry.
The validated ELIFA cleaning procedure will Jeff Morier is a Senior Assay Development Techbecome part of awritten Standard Operating Procedure nologist at Cangene Corporation. He received his (SOP). Although addressing R&D instrumentation, the B.Sc. degree in Microbiology from the University SOP document will detail the activities that were con- of Manitoba. He has seven years experience in the pharmaceutical industry in the areas of QC microbiducted by adhering to industrial standards for cleaning ology, QA biotechnology, and R&D experience in the 1,9 validation. The document will also advise on safety validation of immunoassays of various formats. precautions, cleaning schedule, and assignment of Patti Hosler is a Technician at Cangene Corporation. responsibility for cleaning and storage of the cleaned She completed the first year of a B.Sc. degree proapparatus. The SOP document will be observed not gram at Brandon University and received a diploma in only when validating the performance of the IgA Chemical and Bioscience Technology from Red River ELIFA, but also during routine use of the ELIFA sys- College. She has seven years experience as a QA/QC tem. It will be the subject of a periodic evaluation and, laboratory technician in the food production industry. if deemed necessary, be updated and/or revised. Danuta W. Kierek-Jaszczuk is a Senior Research
Conclusions nA
comprehensive, credible CVP designed and developed at Cangene for the Easy-Titer™ ELIFA system has been shown to effectively remove contaminants and residues entrapped in the apparatus after the conclusion of the experiment(s) and/or subsequent cleaning. nThe CVP has been demonstrated to vastly reduce the analytical background of the IgA ELIFA, improve its signal to background ratio, increase the quality of the test results and may, therefore, be expected to notably support the upcoming assay validation. nThe CVP, by virtue of anti-viral and anti-bacterial properties of the RBS10, allows for simultaneous decontamination and sanitization of the ELIFA unit, thus facilitating its safe use with infectious samples. n The CVPs generated for R&D equipment that fulfill the standards of industrial cleaning validation not only improve the quality of the assays utilizing this equipment but may become vital components of assay validation. o
Scientist/Assay Development Supervisorat Cangene Corporation. She obtained her M.S. degree in Biology from the Nicolaus Copernicus University, and a Ph.D. degree in Agricultural Sciences from the Polish Academy of Sciences Institute of Genetics and Animal Breeding. She can be reached by phone at 204-275-4263, by fax at 204-269-7003, and by e-mail at
[email protected].
References 1. FDA. 1993. “Guideline to Inspection ofValidationof CleaningProcesses.” Office of Regulatory Affairs, USFDA, Washington, D.C. 2. Pierce Chemical Company. Product Instructions, Easy-Titer™ ELIFA System. Rockford, IL. 3. Paffard, S.M., Miles, R.J., Clark, C.R., and Price, R.G. 1996. “A Rapid and Sensitive Enzyme Linked Immunofilter Assay (ELIFA) for Whole Bacterial Cells.”Journal of Immunological Methods 192, no. 1–2: 133-6. 4. Morier, J.,Macaulay, L.,and Kierek-Jaszczuk, D.“Screening for the Presence of Human IgA in a Hyper Immune Product Using An Enzyme-Linked ImmunoFiltration Assay.” Poster Presentation at IBC Conference on Assay Development for Future High-Throughput Screening. 8 – 9 November 1999. Annapolis, MD. 5. Faurschou, A. 2000. General Procedure for Validation Program. SOP Document # 11.001.0001.RR. Cangene Corporation. Winnipeg, MB, Canada. 6. Alejo,M. and Faurschou,A. 1998. ProcessValidationQualification. SOP Document # 11.001.0002.RR. Cangene Corp oration. Winnipeg, MB, Canada. 7. Heise, R. and Poschner, E.1999. Manual Cleaning and Sanitizing Equipment. SOP Document # 2.010.0017.RR, Cangene Corporation. Winnipeg, MB, Canada. 8.
About the Authors LeeAnne Macaulay is a Technician at Cangene Corporation. She completed the first year towards 60
Institute of Validation Technology
Shinkarik, T. 1998. Surface# Sampling for Total Organic Carbon (TOC). SOP Document 500602.RR, Cangene Corporation. Winnipeg, MB, Canada. 9. Chudzik, G.M. 1998. “General Guide to Recovery Studies Using Swab Sampling Methods For Cleaning Validation.”Journal of Validation Technology5, no. 1: 77-81. 10. Pierce ChemicalCompany.ProductInformation, RBS. Rockford, IL.
A Cleaning Validation Master Plan for Oral Solid Dose Pharmaceutical Manufacturing Equipment By Julie A. Thomas McNeil Consumer Healthcare
v
W
ith the benchmark conValidations of Cleaning Processes stantly being raised, – July 1993.” Each of these will be }Often, many companies find discussed in greater detail in the seccompanies have tions below. that they are in perpetual validation mode. Often, companies have executed executed validations for equipment, n Objective cleaning, and processes, but the n Scope validations for documentation no longer stands up n Introduction to the latest in validation standards. n Responsibilities equipment, Although these validations are genn Philosophy erally complete and on file, there cleaning, and pron Methodology are many opportunities to improve cesses, but the doc- n Schedule both the supporting documentation and the execution. One way to Objective umentation ensure that your company’s policies This section should state the purand procedures regarding cleaning no longer stands validation are state-of-the-art is to pose of your cleaning master validaup to the latest tion plan and define whether you will assemble a multi-disciplined team from the appropriate manufacturing be revalidating current procedures or validation prospectively validating new ones. sites that can review and revise all Often, the plan will have provisions components associated with cleanstandards.~ ing validation. What follows are for both situations. excerpts from a Cleaning Validation Master Plan (the Plan) that was painstakingly comScope posed and has now become the standard for planning The scope needs to list exactly which aspects of valand executing cleaning validations at several manuidation will be covered in the document and to which facturing sites. An outline of the Plan contains the following seven types of products and/or processes the Plan applies. For elements, the concepts of which are taken directly example, “This document provides steps for planning, from the FDA publication, “Guide to Inspections of executing, and maintaining equipment cleaning validaSpecial Edition: Cleaning Validation III
61
Julie A. Thomas
tions for oral solid dose products at Your Company’s manufacturing facility in Your City, State.”
Introduction The introduction should let the reader know what elements will be addressed in the Master Validation Plan and why a formal plan is necessary. For instance, “This plan is intended to be a roadmap clarifying the course the Company will take as it plans and executes the cleaning validations required by current Good Manufacturing Practices (cGMP). This program describes and defines the various categories of cleaning validation, provides the necessary protocol elements, and offers guidance for unexpected results. Furthermore, it includes provisions for revalidation and monitoring and serves as a mechanism to organize and store critical information that supports the cleaning validation process.”
Responsibilities There are many departments and disciplines involved in planning for and executing a cleaning validation. It is necessary to list each contributing area and the associated tasks for which it is responsible. This serves to clarify roles and to ensure that tasks are not overlooked. Typically, representatives from Validation, Manufacturing, Quality Control, Engineering, and Research and Development (R&D) will be needed. The following are some examples of departmental responsibilities: Validation Specialist
• Review cleaning procedures • Assist the cleaning validation team in identifying equipment test sites for swab or rinse samples • Write cleaning validation protocols • Coordinate execution of the cleaning process with the appropriate departments and laboratories • Prepare the sampling schedule • Assemble the test data into final report form for approval
for accuracy and agreement with operating practices • Create and/or revise related SOPs and cleaning checklists • Perform cleaning processes per SOP as referenced in the validation protocol • Provide documented training for all personnel responsible for cleaning the equipment Quality Assurance
• Review and approve protocols and reports for conformance with cGMPs and internal procedures • Provide analytical technical support • Provide documented training for all personnel responsible for sample collection and testing • Collect analytical samples as specified in the protocol • Perform analytical testing using validated procedures • Label, package, and send out those samples that need to be analyzed by an external laboratory • Review and approve analytical results • Notify departments of test results Engineering
• Inform the affected department in advance of any anticipated change to the facility or equipment • Include all utilities and cleaning equipment in the calibration and maintenance program • Review and approve equipment drawings and surface area calculations Research and Development
• Provide swab and surface recovery data for active ingredients and cleaning agents • Validate analytical test methods for chemical and cleaning agent analyses • Transfer validated methods to the site QC laboratories and/or contract laboratories • Provide recommended cleaning procedures for new active ingredients and/or cleaning agents
Manufacturing
• Provide technical information for the development of protocols and reports • Review and approve protocols and reports 62
Institute of Validation Technology
Cleaning Validation Philosophy This section discusses the considerations you
Julie A. Thomas
have made in developing a comprehensive cleaning validation program, such as how to define equipment holding time, equipment storage time, and campaign length. In general, the philosophy section presents the Company’s position on what is being achieved by the cleaning validation and how it will be demonstrated. For instance, “Cleaning validation is required for all manufacturing and packaging equipment that comes into contact with the product or product components during production. Prior to validation, acceptance criteria will be developed for active ingredient and
preparatory and includes analytical methods validation, recovery studies, surface types, degradants, and methods transfer. There is a considerable amount of scientific activity that must be completed before the validation can begin. These steps are explored in the following sections.
cleaning agent residues. Verification of acceptable equipment holding time will be included as part of the validation. Holding time is defined as the time between the end of the last product manufactured and the start of the cleaning process. This will demonstrate that the cleaning procedure effectively removes residue after the equipment has remained idle fora specified period of time. Additionally, holding time will be evaluated to ensure storage conditions are adequate for a predetermined length of time. Storage time is defined as the time between cleaning completion and the next batch processed on the equipment. Campaign length will be determined jointly by Operations and R&D and validated with at least three iterations using the maximum number of batches or maximum length of time. This approach fully challenges the cleaning procedure by providing worst-case residues.”
Validation of the method should assess reproducibility, linearity, specificity, limit of detection (LOD), and swab and surface recovery. Other elements for consideration are the instrumentation, swabbing and dilution solvents, dilution volume, and sample handling and storage.2,3
Cleaning Validation Methodology To ensure all of the elements are in place for a thorough and successful validation, a chronological methodology should be followed. One such design is illustrated through the following four phases: development, planning, execution, and maintenance. (See Figure 1) In this section of the Plan, it is appropriate to include the number of sampling/testing iterations required for each piece of equipment and/or each analyte. (See Figure 2.) If you intend to reduce the number of tests required to validate cleaning after various products by using a grouping approach, it should be explained in this section.1
1. Analytical Methods Validation
Describe how the analytical methods will be developed and validated for active ingredients, degradants (if applicable), and cleaning agent residue.
2. Recovery Studies
Recovery studies evaluate quantitative recovery of chemical residue from both the surface to be sampled and the swab material to be used for sampling. The results confirm the appropriateness of the sampling method and material used. You should determine the minimum recovery criteria for each surface type and state that percentage in this section. For instance, you may want recovery values of at least 70% of actual readily soluble residues, but may choose a much lower recovery value for relatively insoluble proteins.4 Most important, you must provide data to justify the chosen value. 3. Surface Types
Since different surface types have different affinities, you may want to choose a few surface materials to represent the many product contact surfaces used in manufacturing. For oral solid dose manufacturing, you may determine that stainless steel, silicone, and polypropylene are the most abundant surfaces and that they also provide varying degrees of porosity. A matrix of all surface types and the representative material that will be used in recovery studies is appropriate. (See Figure 3) 4. Degradants
Development Phase The initial phase of the cleaning validation plan is
Many degradant products are more soluble in the cleaning solvent than are the active ingredients; there fore, you should determine the degree of degradant Special Edition: Cleaning Validation III
63
Julie A. Thomas
Figure 1
cg vd F Dg Dp P Analytical Method Validation
Analytical Method Development
• Degradant identification • Transfer
• Recovery • Surface types
Pg P Equipment • Sample site selection • Surface area calculation • Schematic
Analyte Selection and Acceptance Criteria
Protocol Development • Write • Approve • Train
• Active ingredient • Cleaning agent
Cleaning SOP • Write • Approve • Train
e P Protocol Execution • Clean • Sample • Test No Pass?
Incident Investigation
Yes
Validation Report • Write • Approve
m P Monitoring
Change Control
Revalidation
64
Institute of Validation Technology
Julie A. Thomas
Figure 2
cg i s r sp
t nb f i
Active Residue
cd
3
1 at maximum campaign length or maximum time period plus holding time. 2 at maximum campaign length or time period.
Cleaning Agent Residue
3
testing required based on the toxicity and solubility of potential degradants. Likewise, active ingredients should be exposed to the selected cleaning agent under normal usage conditions to determine if degradants are formed as a result of the cleaning process. 5. Analytical Methods Transfer
3 per cleaning procedure, per piece of equipment.
1. Equipment Information
This section should detail the methodology for providing specific equipment information. One option is to prepare a binder containing detailed surface area calculations, swab sampling sites (with justification), photos, and schematic diagrams for each piece of equipment. This binder can be maintained separately and used as an attachment to the cleaning validation protocol as needed.
In this section, you can state how sampling and analytical methods will be transferred from the R&D laboratories to the site QC laboratories and a) Sample Site Selection how the analysts conducting validation testing will Explain how you will select sampling sites to repbe qualified. Reference appropriate SOPs and/or resent the product contact surface area of the equipDevelopment Transfer Report. ment. One of the best sources of information is the operator who routinely cleans the equipment. He or she can certainly point out the areas they find most difficult to clean. Make the operator part of a larger The next phase of preparation isthe planning phase. team of experts to include representatives from This is a broad category that focuses on equipment Validation, QA, and Operations, and let the team information, analyte selection, acceptance criteria, determine the product contact surface areas that are cleaning procedures, and protocol development. At this most difficult to clean and those that are most reprepoint, you are starting to think about what equipment sentative of the equipment. Sampling these sites will will be included in the validation, which analytes will represent the entire equipment surface area using be chosen, and how youwill determine acceptance cri- the assumption that residue will be evenly distribteria. This leads to an in-depth review of the procedures uted over the equipment and that the most difficult and, finally, to protocol development. to clean locations will represent the worst case for residue removal. Include the basis for selecting each
Planning Phase
Figure 3
sf r m r sf: Material Used: To Represent:
316l s s
P
316L Coupon
Plastic Bulk Container
304 Stainless Aluminum Brass
Teflon
s Hose Rubber
Lexan HDPE
EPDM Neoprene
Special Edition: Cleaning Validation III
65
Julie A. Thomas
site. For example, sampling sites may be deemed to be the most difficult to clean, most difficult to dry, or of different material of construction. Swab sites
Figure 6 Kason Separator
Figure 4
sb s
can be indicated with either digital photographs or pare schematic diagrams labeled with the major secsuitable diagrams. (See Figure 4) tions of the equipment. (See Figure 6) The drawings b) Surface Area Calculation do not have to be to scale, but should appropriately An accurate surface area must be calculated for represent the equipment. If a schematic is not practical each piece or section of equipment. This can be (i.e., a packaging line), a photograph is acceptable. The done with manufacturer’s drawings, but should be intent is to depict the product contact surfaces that are confirmed by field measurements. If drawings are included in the calculations. This helps to ensure that not available, the equipment must be measured to the swab samples are taken from the intended location. determine surface area (seeFigure 5). Although not shown here, it is advisable to include the calcula- 2. Analyte Selection Analyte selection is required for active, excipient (possibly), and cleaning agent residues. Keep Figure 5 in mind that you are validating a cleaning procesf a dure, not a manufacturing process. In the situation sb nb a sbbd where the same cleaning procedure is used for many product formulas, there is an opportunity to select a 1 Screen/ring interface gasket representative analyte to cover multiple active ingredients and reduce the amount of testing. 2 Discharge port – inside of top circular area (top seam) 2 Total contact S.A. of Kason Separator (in )
Total contact S.A. of Filter Socks (in ) 2
a) Actives 3,171.2 15.6
tions with the schematic diagram in the equipment information binder mentioned above.
If several active ingredients are processed in a single piece of equipment, a marker active, or guiding substance, can be selected based on the active ingredient solubility in water, potency, previous production experience, and R&D studies. This reduces the number of studies required to validate the cleaning procedure.5
c) Schematic Diagram
To clearly illustrate each piece of equipment, pre-b) Excipients 66
Institute of Validation Technology
Julie A. Thomas
The removal of excipients can either be confirmed by visual inspection or through analytical testing. The approach should be stated here along with training requirements for individuals performing visual inspection.
calculated using a formula such as the No Observed Effect Limits (NOEL).8
recommendation of the cleaning agent manufacturer. Removal of volatile cleaning agents that do not leave a residue, such as isopropyl alcohol, may not need to be validated.
10 who will be executing the validation studies.
4. Cleaning Procedures
This section should indicate that cleaning procedures will be developed (or existing procedures reviewed) prior to the validation. It should also list the required c) Cleaning Agents elements for cleaning procedures, such as temperature, Testing for cleaning agent residue is essential but pressure, water quality, cleaning agent concentration, is often an area in which current cleaning validations spray nozzle location, etc., or it should reference where are deficient. For most cleaning agents, a marker these requirements can be found.9 Additionally, you compound can be selected for analysis based on the should describe the process for training the operators
5. Protocol Development
The next step is to write a cleaning validation protocol for each cleaning procedure that you intend to 3. Acceptance Criteria validate. The protocol should describe all documentation required to complete the cleaning validation. It The equipment must pass visual and olfactory inspection, where appropriate, as defined in should also present the rationale for using a marker the cleaning validation protocol prior to initiation active to cover validation for multiple products. For of swabbing.6 This is a critical step in the validation ease of review, include a matrix of the products and process that, if skipped, can lead to failed results. equipment that are covered by each validation, or reference where this information can be found. For a) Active Ingredient example, if there are three active ingredients processed Acceptance criteria for active ingredients should in Fluid Bed Granulator #1, indicate which active will be based on medical and pharmacological properties be used to represent the other two. Likewise, indicate and scientific information. Calculations using the which pieces of equipment will be used to validate maximum allowable carryover (MAC) and/or 10ppm Figure 7 formulas can be used.7 To ensure that all active contact surfaces are considep cg m ered in the carryover calculation, you may want to idena a a B a c cg tify equipment trains. Acceptance criteria are calculated ag a using the surface area from the entire equipment train; however, protocols are executed per each piece of equipFluid Bed GranX1 – – X ment. Equipment trains could be designated as follows: Fluid Bed Gran 2 – – – – Starch Kettle 1
Granulation – granulator system through the product container n Compression through printing – compression, film-coating, and printing phases n Packaging – product contact surfaces for each type of packaging line
–
–
–
X
n
b) Cleaning Agent
the removal of active ingredient and cleaning agent residues. (SeeFigure 7)
Execution Phase When all of the supporting documentation iscomplete, it is time to execute the plan. During the execution phase, you will complete the protocol, investigate any nonconformances that may have occurred, and write a report to summarize your findings.
Acceptance criteria for the cleaning agent marker should be based on toxicity, limit of detection of validated assay method, and/or data gathered during certification studies. Acceptance criteria can be 1. Protocol Execution
Special Edition: Cleaning Validation III
67
Julie A. Thomas
Typically, three iterations of cleaning, sampling, and testing using the same procedure are required. Acceptance criteria for all cleaning iterations must be met for both the active ingredient and the cleaning agent. Be sure to reference the procedure where a detailed description of the chemical swab preparation and sampling methods can be found. 2. Incident Investigation
This section explains how the Company will handle test failures and nonconformances during execution of the validation. Once the root cause of the failure has been identified, options are to addend the protocol or start over with a new protocol. For any incident that occurs during validation, document the investigation along with corrective and preventive actions. The incident report may contain elements such as: n n n n n n n n
Cleaning validation protocol number Incident report number Equipment model and location Initiator and date Incident description Root cause analysis Corrective actions recommended/taken Assessment of effect on product
3. Reports
Describe the report format and content that will be used to summarize the validation. Reference appropriate SOPs for detailed report information. An explanation of all deviations should be included in the validation report.
Maintenance Phase The final phase of the Plan should specify how you will maintain the conditions you have just validated. This includes periodic monitoring, using a control of change process, and potentially, revalidating. 1. Monitoring
This section details how you will ensure that the conditions used during validation remain in control during routine production. This is especially important for manual cleaning procedures, where 68
Institute of Validation Technology
repeatability is highly dependent on the quality and consistency of training. Monitoring should include, at a minimum, a review of changes made to the cleaning procedure or equipment, visual inspection of the equipment, and direct observation of employees executing the cleaning procedure. For some equipment, swab samples for active ingredients may be necessary in addition to the visual inspection and observation. Indicate the frequency that you intend to monitor the cleaning process. Reference the appropriate SOP for detailed requirements of the monitoring program. 2. Change Control
Indicate how changes will be managed to ensure the validated state is maintained. Any change in the facility, process equipment, cleaning procedure, cleaning agent, product formulation, or addition of new products to the equipment train should be documented and approved via the Change Control System. The change should be reviewed by the Validation Group, Operations, and QA, who will decide if revalidation is necessary. Reference appropriate SOPs.11 3. Revalidation
Indicate the criteria that will be used to determine the need for revalidation. Based on the nature of the change, a determination will be made as to which, if any, phases of the validation must be repeated. Reference where documentation of the revalidation will be filed.12,13
Cleaning Validation Schedule Prioritization
As is usually the case, all cleaning validations cannot commence at one time; therefore, it is necessary to set up a priority list. Some situations to consider are: Equipment shared between products containing different active ingredients n Equipment in contact with raw material with high potential for contamination n Unshared primary equipment currently in use with outdated validations n Unshared auxiliary equipment used for pron
Julie A. Thomas
duction with limited potential for product contamination
References 1.
Tactical Schedule
A proposed schedule, including equipment prioritization and target initiation dates, should be presented in this section. This gives an indication that you have contemplated the order of execution, and it also provides a tool that can be used to track your progress.
Summary There are many aspects of cleaning validation that must be carefully planned to guarantee a successful validation program. If you begin with a philosophy, this will set the stage for you to develop a structured approach. By dividing the approach into sections, such as development, planning, execution, and maintenance, it breaks down the project into manageable segments. To complete the Plan, generate a tactical schedule and begin monitoring progress towards your new and improved cleaning validation status. o
2. 3. 4. 5. 6. 7.
8. 9. 10. 11. 12. 13.
About the Author
Jenkins, K.M. and Vanderwielen, A.J. “Cleaning Validation: An Overall Perspective,” Pharmaceutical Technology, April 1994, p. 62. McCormick, P.Y. and Cullen, L.F., Pharmaceutical Process Validation, 2nd ed., edited by I.R. Berry and R.A. Nash, 1993, p. 334. Kirsch, R.B., “Validation of Analytical Methods Used in Pharmaceutical Cleaning Assessment and Validation,” Pharmaceutical Technology, Analytical Validation, 1998. Chudzik, G.M., “General Guide to Recovery Studies Using Swab Sampling Methods for Cleaning Validation,”Journal of Validation Technology, Vol. 5, No. 1, pp. 77-81. Hall, W.E., “Your Cleaning Program: Is It Ready for the PreApproval Inspection?” Journal of Validation Technology, Vol. 4, No. 4, August 1998, p. 302. Alvey, A.P. and Carrie, T.R., “Not Seeing is Believing – A Non-Traditional Approach for Cleaning Validation,”Journal of Validation Technology, Vol. 4, No. 3, pp. 189-193. Fourman, G.L. and Mullen, M.V., “Determining Cleaning Validation Acceptance Limits for Pharmaceutical Manufact ur ing Operations,” Pharmaceutical Technology, 17 (4), 1993, pp. 54-60. Hall, W.E., “Validation of Cleaning Processes for Bulk Pharmaceutical Chemical Processes,”Cleaning Validation An Exclusive Publication, p. 4. Hall, W.E., “Proper Documentation and Written Procedures,” Journal of Validation Technology, Vol. 4, No. 3, pp. 199-201. Tunner, J.,“Manual CleaningProcedure Designand Validation,” Cleaning Validation An Exclusive Publication , p. 28. PDA Technical Report No. 29, “Points toConsider for Cleaning Validation,” March 1998, p.43. Coleman, R.C., “How Clean is Clean?” Journal of Validation Technology, Vol. 2, No. 4, August 1996, p. 278. Jenkins, K.M.and Vanderwielen,A.J., “CleaningValidation: An Overall Perspective,” Pharmaceutical Technology, April 1994, p. 70.
Julie Thomas is Validation Manager at McNeil Consumer Healthcare in Round Rock, Texas. She has 14 years of experience in various aspects of solid dose pharmaceutical manufacturing. Most recently, she chaired a company-wide committee to enhance cleaning validation practices and procedures for all McNeil facilities. She can be reached by phone at 512-248-4470 or by e-mail at
[email protected].
This article presents only one alternative for preparing a Master Validation Plan. The views pressed ex in this article are strictly those of the author and in no way represent the view of McNeil Con sumer Healthcare, Johnson & Johnson, or this publication.
© Advanstar Communications Inc. All rights reserved.
Special Edition: Cleaning Validation III
69