aseptic processing: a review of current industry...

126 Pharmaceutical Technology OCTOBER 2004 www.pharmtech.com

James Agalloco is thepresident of Agalloco &Associates, PO Box 899, BelleMead, NJ 08502. JamesAkers, PhD, is the presidentof Akers Kennedy & Associates.Russell Madsen ispresident of The WilliamsburgGroup, LLC.

*To whom all correspondenceshould be addressed.

Aseptic Processing: A Review ofCurrent Industry PracticeJames Agalloco, James Akers, and Russell Madsen

DIG

ITA

L V

ISIO

N

This article reviews currentindustry practices andregulatory expectations for theaseptic processing of steriledrugs. It provides comparisonsand outlines points of tensionbetween curent manufacturingtechnology and capabilities withregulatory “requirements” forthis important activity.

n 1988, the Parenteral Drug Association(PDA) published a position paper onaseptic processing in response to intenseinterest in aseptic processing in the in-

dustry at that time and in partial responseto the publication of the Food and DrugAdministration’s (FDA’s) 1987 guidelineon aseptic processing (1, 2). The intent ofthis review is similar to that of its 1988predecessor: to identify and discuss thecurrent capabilities of aseptic processingtechnology.

The improvements to aseptic process-ing operations described in the 1988 po-sition paper were characterized in theopening section of that document as “evo-lutionary.” The improvements in asepticprocessing technology that have occurredsince 1988 would perhaps be better char-acterized as “revolutionary.” We believethat the practices described in 1988 are,by and large, as valid now as they werethen. We also believe that products man-ufactured in compliance with the under-lying principles outlined in the 1988 doc-ument are still inherently safe.

The aseptic manufacture of sterile prod-ucts is perhaps the most difficult challengefaced within the healthcare industry. Asep-tic processing requires the careful appli-cation of microbiological contaminationcontrol principles to exclude infectiousorganisms from sterile products. We reaf-firm our belief stated in 1988 that, “themajor variable in the control of asepticprocessing arises not from the steriliza-tion processes, the cleanroom, or the fil-tration processes that are so often the sub-ject of technical papers and regulatoryguidelines, but rather from the workforceitself ” (1). Industry findings since 1988

confirm the earlier statement that human-borne contamination is the most criticalrisk factor in aseptic processing. Numer-ous industry surveys and technical arti-cles published since that time are in ac-cordance with this statement. (3–8).

Since the publication of the 1987 guid-ance, firms have continued to implementnew technologies and aseptic processingimprovements to better control human-borne contamination. At the same time,the industry has implemented expandedmicrobial-test regimens and more com-prehensive process simulation testing toensure that aseptic processing systemshave adequate process capability and thatthis capability is consistent and as repro-ducible as possible within the technicalconstraints that are inherent in the meas-urement of aseptic performance.

This article describes the improvementsin conventional and new technologies thathave occurred since the late 1980s. It alsodescribes the improvements that havebeen made in aseptic process validationand control. Finally, the paper discussesthe technical limitations that still exist inthe evaluation of aseptic processing andraises concerns regarding the increasingregulatory tendency to ignore the exis-tence of those limitations.

The condition of asepsisSince the earliest days of parenteral man-ufacturing it has been recognized thatmany drugs and biologics would not with-stand a physical sterilization process intheir final container. Currently, a majorityof parenterals and other products labeledsterile are manufactured using aseptic pro-cessing (8). It is therefore appropriate to

I


review the concepts of asepsis and steril-ity before embarking on a discussion ofkey process requirements and validationprinciples for aseptic processing to clarifysome of the misconceptions that have creptinto industry and regulatory belief overthe years. A vital element in process designand validation is a definition of the end-point. And it is not possible to establish anend-point and associated process controlparameters without careful considerationof the target and the capability of theprocess.

We believe that industry and regulatorycommunities currently find themselvescaught on the horns of a dilemma regard-ing aseptic processing. The dilemma arisesdirectly from the seemingly logical but sci-entifically flawed notion that it is possibleto prove that each and every lot of asep-tically manufactured product containsonly “sterile” material. Sterile is, if takenin its most absolute meaning, a word thatoffers no room for uncertainty in meas-urement or outcome. Sterile means free ofany viable organisms (6). This condition,however, is not now and never has beenpossible to prove in aseptic processing.

This dilemma is not new—it has beenrecognized for decades and should neverhave left our collective understanding.What has changed since the publicationof PDA’s 1988 position paper is an under-standing of the inherent limitations of anytechnology built around the exclusion ofmicroorganisms rather than the physicaldestruction of those microorganisms. It

seems to us that in the past most indus-try scientists and the regulatory commu-nity clearly understood that absolute steril-ity in aseptic processing was not possibleand therefore did not reflect a reasonableexpectation of outcome. It is now our be-lief that this appreciation is no longer aswidespread as it once was.

In the past, the perspective of processcapability for aseptic processing, and hencethe requirements for process control, weremore solidly grounded scientifically. AsBlock points out, there are pitfalls in usingthe word sterile in an absolute manner inpractical applications (9). For example,we have long understood that process fil-ters that produced safe product could notbe demonstrated to yield sterile productin the “absolute” meaning of that word(10). In other words, we can always iden-tify organisms that could penetrate anyfilter that is practical for use in aseptic pro-cessing. Absolutism is dangerous even inphysical sterilization because it presumesthat we can have discovered every exist-ing form of microbial life.

When we use the word sterile—whetheron a label or in a guidance document,standard, or regulation in the pharma-ceutical industry—we must consider thatthe meaning of the word cannot be takenin a literal, absolute manner becausedoing so does nothing but ensure failure.While suggesting the acceptance of par-tial or incomplete sterility may be ananathema to those conditioned to thinkthat parenteral manufacturing must bemicrobiologically perfect, it is our viewthat only by consideration of asepsis in acareful, scientific manner can we avoidunreasonable standards and correspond-ingly impractical regulatory enforcement.It is important to ensure that we as an in-dustry are not trapped by semantics orimpossible-to-achieve standards borne ofmisunderstanding.

Perhaps industry chose poorly by label-ing aseptically manufactured products as“sterile.” Little if any harm has come fromthis choice, however. In fact, it can be ar-gued that the standards currently in placein industry clearly recognize the seman-tic tension between sterile and aseptic. Nocurrent standard calls for zero contami-nation in media-fill process simulationtests or in environmental monitoring, al-though several documents—including

PDA’s Technical Report No. 22—suggestthat zero contamination is an appropri-ate target (6, 11, 12). We believe it essen-tial to recognize the validity of zero as atarget but at the same time remind indus-try and the regulatory community thatthis target may not be consistently anduniformly attainable because no asepticenvironment or aseptically producedproduct is provably sterile.

The industry’s approach to validation of aseptic manufacturingIn no other segment of the pharmaceuti-cal industry is the control of manufactur-ing processes as critical as in the produc-tion of aseptically produced products. Therecognition of the criticality of theseprocesses has led to the continued devel-opment of advanced production and qual-ity assurance systems. Firms have contin-ued to develop and implement morerigorous methods for the validation ofaseptic processes (13).

The resource and staffing requirementsfor validation programs within the indus-try have continued to increase during thepast decade and a half. These increasedvalidation efforts are comprehensive, in-cluding both prospective validation andongoing validation maintenance. Themanagement of a sound validation pro-gram requires the participation of special-ists from various academic backgroundsin technical and administrative disciplines.Validation costs across the industry havecertainly increased, and as might be ex-pected the most costly of all operations tomaintain in a validated state is asepticmanufacturing. Only part of the increasedcosts can be attributed to the increasedtechnical complexity of aseptic operations.Substantial portions of the increased costsare a direct result of increased regulatoryexpectations. The following sections ofthis article present examples of the typesof programs that make up the aseptic pro-cessing validation effort.

Sterilization validation and qualification ofsterilizers. Comprehensive engineeringqualifications are conducted on each newsterilizer to ensure that the design andfunctional specifications are met. Thesequalification activities are used as the basisfor formalized change control programsthat support the continued appropriate-ness of the sterilizer over time. In addi-

Table I: Rank-ordered sources ofmicrobial contamination in aseptic processing (3, 8).

1986 2001Personnel contaminants 1 1Human error 2 2Nonroutine activity 3 4Aseptic assembly 4 3Mechanical failure 5 5Improper sanitization 6 7Material transfers 6 8Surface contaminants 7 7Airborne contaminants 7 6Routine APAactivity 7 7Failure of 0.2 filter 8 8Failure of HEPA 9 8Improper sterilization 10 9Other 10


tion, the maintenance procedures and cal-ibration requirements for all process con-trol devices are developed, as are the nec-essary standard operating procedures forthese activities.

Following completion of the qualifica-tion of the sterilizer, process validation isundertaken to ensure that the sterilizationprocess complies with regulatory require-ments. These studies are designed to en-sure that the conditions achieved through-out each sterilizer load result in thedelivery of the appropriate lethality for theprocess. Equally important, manufactur-ing and quality assurance controls are de-signed to demonstrate that the sterilizeroperates reliability and reproducibly underintended-use conditions during its oper-ational life. Validation typically includesan adequate margin of safety to allow forprocess variation.

After the prospective validation hasbeen completed, change control systemsare established to ensure that sterilizersare maintained in a state of control. PDAcommented in the 1988 position paperthat “revalidation” studies that are essen-tially repeats of prospective validationstudies on a periodic basis are not neces-sary. Unfortunately, PDA’s position in thismatter has not gained the broad accept-ance it deserves.

In addition to periodic performancechecks, calibration, and maintenancechecks, firms enforce rigorous programsto ensure that no change is made to anyaspect of the sterilization process or ster-ilization equipment without a full inter-disciplinary evaluation. These change con-trol programs are now far broader in scopeand much more connected to the regula-tory function within firms than they werein 1988. In the United States, firms followregulations that define requirements re-porting process changes to FDA. Similarpractices are required outside the UnitedStates (12). Regulations defining report-

ing responsibilities for sterilizationprocesses are far more clearly delineatedthan they were in 1988.

We are also concerned about the in-creasingly common practice of repeatingvalidation studies on a more or less an-nual basis. There is no reason to believethat a sterilizer operating within validatedparameters for physical operation shouldrequire periodic biological challenge test-ing. Modern steam and dry-heat steriliza-tion equipment is robust and reliable.There is no scientifically valid reason torepeat biological challenges when theprocess control and engineering data in-dicate that the sterilization parameters andequipment function are consistentlywithin target values. The imposition ofunnecessary validation or revalidation re-quirements based upon unreasonable fearrather than scientifically meaningful analy-sis results in wasted resources and reducedproductivity, with no benefit to end users.

Validation of sterilization procedures forcomponents, containers, and closures thatenter the aseptic processing areas. In 1988,PDA asserted that the sterility assurancelevel (SAL) of aseptically produced prod-ucts was “generally considered to be 103

or more,” while physical sterilization tech-nologies used in component preparationresulted in an SAL of “at least 106.” (Itshould be noted that PDA correctly pre-sented the SAL value as a positive expo-nent, not to be confused with the proba-bility of nonsterility concept use to assessrisk associated physical sterilization meth-ods, which is a fraction of one and there-fore carries a negative exponent.) We be-lieve that the sterilization of componentsis a very minor risk factor in aseptic pro-cessing, just as it was in 1988.

The predominant cause of contamina-tion in aseptic processing remains thesame as it was in 1988: activities per-formed by personnel in direct support ofthe aseptic process. Not only are person-

nel the major source of contamination,their actions serve to distribute the organ-isms within the environment. In PDA sur-veys on aseptic processing conducted in1986 and again in 2001, industry opinionsregarding the sources of contaminationin aseptic processing were virtually un-changed (see Table I). Table I also showsthat the seven most prevalent sources ofcontamination—identified as most likelyto contribute microbial contaminationduring an aseptic process—all are relatedto activities performed by operators andare virtually unchanged in relevance dur-ing the 15 years between these surveys.

We also have seen shifts in the technolo-gies applied for the sterilization of com-ponents. Since 1988, environmental andoccupational safety concerns have greatlyreduced the utilization of ethylene oxide(EtO) for the sterilization of plastics andother non–heat stable components. Ofcourse, EtO is still used and is still an ex-tremely efficacious gas sterilant.

However, several other technologieswith similar levels of sterilization efficacyhave been introduced or have expandedin their use. Both gamma and beta (elec-tron beam) sterilization were used in 1988,but they are more widely used today.Lower-energy radiation technologies arecoming into increasing use in sterilizationof heat-labile materials. Also, new gas- orvapor-sterilization methods have been in-troduced. These include vapor phase hy-drogen peroxide and chlorine dioxide.

In some respects, the methods used forcomponent sterilization in the manufac-ture of glass and elastomeric stopper con-tainer systems have changed little since1988. Moist-heat sterilization continuesto be the most widely used sterilizationmethod, and validation technology andefficacy have changed little since 1988.Validated moist-heat sterilizationprocesses continue to result in an ex-tremely safe outcome.

Dry-heat depyrogenation continues tobe the method of choice for the steriliza-tion and depyrogenation of glassware. In-dustry develops and validates depyrogena-tion processes to yield at least a three-logreduction of reference-standard bacterialendotoxin. The resistance of endotoxin toheat is at least an order of magnitudegreater than even the most resistant spore,therefore a three-log or more reduction

Table II: Media-fill survey comparison—acceptance criteria (19).Criteria 1980 1986 1992 1996 2001

,0.05% 3 11.5% 7 13.2% 4 7.5% 6 12.5%0.05–0.09 2 7.7% 5 9.4% 12 22.6% 9 18.8%

0.10% 4 25.0% 19 73.1% 36 67.9% 51 92.6% 33 68.8%0.11–0.20% 4 25.0% 1 3.8% 1 1.9%0.21–0.30% 3 18.8% 1 3.8% 5 9.4%.0.30% 5 31.2%


in endotoxin results in a very low proba-bility of nonsterility, far lower in fact that1026 (SAL . 106).

The most significant process risks asso-ciated with component handling have al-ways been the aseptic handling requiredin the initial setup of component supplysystems such as parts hoppers and feedchutes poststerilization and in keepingaseptic processing operations supplied dur-

ing manufacturing. In 1988, supplyingcomponents was largely a manual process.The widespread adoption of continuouswashing and depyrogenation of glasswareusing “tunnel” systems began in the late1970s and was well underway in 1988. Atthat time, many aseptic filling operationsrequired operators to supply the filling linewith depyrogenated glass. Today, contin-uous glassware washing and depyrogena-

tion without the need for human interven-tion to supply the filling machine is stan-dard practice in all but the lowest through-put operations. The incrementalimprovements in operational efficiency ofthe continuous glassware-processing sys-tems have led to higher levels of through-put, while human interventions for main-tenance, correction of jams, or removal offallen containers have been reduced or insome cases essentially eliminated.

Similarly, a great deal of engineering ef-fort has been invested in developing stop-per processing systems that minimizehuman intervention. Continuous stopperprocessors with continuous automatedsupply have been introduced. Also, inbatch component feed operations, ad-vanced aseptic technologies such as re-stricted accesses barriers (RABS) or isola-tors equipped with rapid transfer port(RTP) technologies have all but eliminatedthe risk of human-borne contamination.

Of course, many aseptic processing op-erations continue to rely on human oper-ators to supply components using aseptictechniques. Some significant improve-ments have been made in cleanroomclothing during the past decade and a half.These improvements have led to saferhuman-scale cleanroom operations. Thefiltration properties of cleanroom cloth-ing have improved, for example, and newhood, goggle, and face-mask systems haveresulted in less skin exposure, better seal-ing, and improved operator comfort.

The 1988 PDA paper clearly stated “Thethreat of contamination in aseptic pro-cessing arises not from the sterilizationprocesses, the cleanroom, or the filtrationprocesses which are so often the subjectof technical papers and regulatory guide-lines, but rather from the workforce itself”(1). As a consequence, equipment suppli-ers, cleanroom suppliers, and the pharma-ceutical industry itself have reacted to thischallenge, systematically reducing the riskof contamination. The continued imple-mentation of automation and advancedaseptic processing technologies will con-tinue to incrementally reduce risks asso-ciated with human-borne contaminationin aseptic processing.

Process filter validation. The basic prin-ciples involved in product filtration aregenerally the same today as they were in1988 (15). Firms continue to carefully as-


sess fluid bioburden and use these data toreaffirm the appropriateness of their fil-tration systems. Firms generally reassessbioburden on a periodic basis and clearlydefine hold times and the temperaturesof materials that reduce the opportunityfor microbial survival. The highly special-ized nature of filter validation generallyrequires considerable participation by fil-ter vendors.

Of course, there have been improve-ments made in filter technology and in itsprocess control since 1988, including

• improved materials of constructionof filter cartridges that have allowedfor greater use of in situ sterilization

• using preassembled cartridge filtersrather than manually assembled fil-ter disk sets throughout the industry,thereby reducing the opportunity for

errors in assembly• improved vent and gas filters with

greater strength and reliability• increased usage of reliable sterilize-

in-place technology on both vent andprocess filters, which further reduceaseptic connections and manipula-tions

• improved automated systems for insitu evaluation of filter integrity.

Process simulation testing. It is our beliefthat PDA’s 1996 technical report that fo-cused on the validation of aseptic process-ing is still the most comprehensive trea-tise on process simulation or media-filltesting of aseptically manufactured prod-ucts (6). The reader is referred to that doc-ument for detailed guidance regarding thedesign of media-fill tests and the interpre-tation of results.

Much of what was described as stan-dard industry practice in the 1988 posi-tion paper is still suitable today. Evolvingregulatory doctrine for media-fill test re-quirements has become increasingly bur-densome. Unfortunately, however, this hasnot increased end-user safety. Particularconcerns include media-fill container re-jection issues, length of fill and numberof units filled, personnel qualification, andrevalidation practices (16).

During the past few years, FDA has ap-plied considerable pressure regarding therejection of units from the media-fill pop-ulation that is incubated and inspected.FDA’s concern in this regard is under-standable: The media fill should not be bi-ased by the removal of containers a firmbelieves might have been compromised,therefore yielding an unwelcome result.The media-fill test must always be a sci-entifically valid evaluation of the asepticprocess, and, as such, there can be no roomfor artificial biasing of the outcome to-ward success.

On the other hand, it is unreasonablefor regulators to hold that all media-filledunits—particularly those that would be re-jected because they lack container/closureintegrity—should be incubated even as aseparate population. Incubation of con-tainers that normally would be rejected forlack of container/closure integrity accom-plishes nothing. We agree that units shouldnot be rejected from the media-fill testpopulation for cosmetic defects only, evenif they would normally be rejected in prod-


uct manufacturing.It is our opinion that far too much has

been made of this issue. It is possible toascertain whether a media fill is represen-tative in terms of rejects by simply com-paring the normal lot rejection rate forcontainer/closure integrity with that ofthe media-fill test. It is clear that the re-ject rate should not be higher in mediafills than in normal production runs of

comparable size. On the other hand, it isequally clear that they should not be ex-pected to be consistently lower either. Thealignment of intervention practices forproduction and media fills should ensurethis consistency of performance.

In 1988 the number of units incubatedby most firms in a media-fill test was al-most always 3000 or slightly more. Thenumber of media-filled units since 1988

have increased since the introduction ofhigher throughput aseptic processing fill-ing systems. In operations with fill speedsat least 200 units/min, the duration of amedia fill in which the target populationwas only 3000 would result in a media fillthat might last substantially less than 30min—not including set up. However, re-quiring media-fill tests that are a high per-centage of the total number of units filledin a batch is not necessary to assessprocess capability. Media-fill populationsof more than 10,000 units are rarely, ifever, required even for high-throughputoperations.

Companies also conduct longer dura-tion fills to test operator fatigue. Duringthe past 25 years, industry has conductedmedia-fill tests under a wide variety of con-ditions, including so called “piggy-back”media fills done at the end of a normalproduction fill. There is no evidence indi-cating that operator fatigue is a factor inenvironmental control, media-fill outcome,or product safety (17). Nevertheless, thereis no need to fill enormous quantities ofmedia for this sole purpose of evaluatingfatigue because the more automated anaseptic operation is the less likely fatiguewill be an issue in asepsis. Each firm shouldperform risk analysis to ensure that theirmedia-fill tests are representative evalua-tions of their processes, adjusting media-fill sample size accordingly.

There should be no fixed requirementfor each operator to participate in amedia-fill test before being admitted toaseptic production work. Abundant meansexist to qualify personnel for aseptic op-erations without the requirement for atleast one media-fill test. Each employeecan be evaluated in terms of gowning ef-fectiveness, and laboratory simulationscan be used to evaluate their aseptic tech-nique (6, 8). In addition, operators can becomprehensively trained on equipmentoperations and relevant operating proce-dures and work instructions. Critical per-sonnel, including those required to per-form equipment set-up and critical asepticassembly, should be required to success-fully participate in a media-fill test beforetaking up their work assignment.

There is no need to conduct more thanone media-fill test per operational shiftper year. More-frequent media fills on val-idated production lines are unnecessary.


It is also unnecessary to test each containertype each year. A firm should develop arationale for their container/closure sys-tem selection on the basis of a carefulanalysis of risk (6).

Media-fill tests are quite useful, but theyare not without limitations. Media-fill re-sults can lead a firm (or a regulatory in-spector) to conclude that an operation ismuch better or much worse than it actu-

ally is. It is important to remember that amedia-fill test is a snapshot in time and isnot always predictive of future outcomeor informative regarding previously man-ufactured product. Certainly a zero con-tamination target is appropriate andmedia-fill positives should occur ratherrarely. However, this does not mean thata single contaminated unit should be thecause of product quarantine or rejection.

Modern aseptic cleanrooms are outstand-ing, but they are not perfect. The quaran-tine or disposal of safe product because ofunwarranted concerns about “sterility as-surance” is wasteful and not scientificallysupportable. Documented improvementsin aseptic processing performance aresomewhat difficult to support in a quan-tifiable manner. One measure of industryperformance is the increasingly tighterlimits placed on media fills since 1980 (seeTable II).

That firms are voluntarily reducing theiracceptance criteria for media-fill contam-ination rates below 0.1% supports the con-tinued improvements that have been madeto aseptic processing. A comparable as-sessment can be found in USP, where arecommendation has been made that twoout of three media fills should be devoidof contamination (20). Coupled with thenear absence of documented evidence in-dicating the presence of actual microbialcontamination in sterile products, this sug-gests that sterile products manufacturedby aseptic processing are safer than ever.

Environmental systems and controlsEssential elements of the environmental con-trol and facility management program. Thebasic elements of an aseptic processing en-vironmental control program have notchanged since 1988 and still consist of

• a review of environmental factors thatgenerally include temperature, rela-tive humidity, air velocity, unidirec-tional air flow, HEPA filtration, andpressure differentials between roomsof different classification

• an evaluation of utility services thatcould affect microbiological safety orproduct quality

• a comprehensive microbiological andtotal-particulate monitoring system

• an evaluation of personnel gowningand materials transfer airlocks

• calibration, certification and preven-tive maintenance on critical facilitysystems and processing equipment

• training programs for personnel inboth aseptic technique and standardoperating procedures or work in-structions.

These systems should be subject to reg-ularly scheduled and unscheduled auditsand routine supervisory oversight andevaluation.


Cleanroom classification. The classifica-tion of each room or module within anaseptic processing area must be appropri-ate for its intended use. The highest levelof control will be directed to those areas,typically known as critical zones, in whichaseptic manipulation of uncovered con-tainers, closures, or components occurs.These areas are designed to comply withClass 5 of ISO 14644 (ISO Class 5 is func-tionally equivalent to traditional US Fed-eral Standard (FS) 209 E Class 100, and toEU Grade A)(12, 21, 22). These areas areequipped with total-coverage HEPA fil-tration, and unidirectional airflow is main-tained to the extent it is technically possi-ble to do so.

European and United States aseptic pro-cessing area zoning differs: In the UnitedStates, the area immediately adjacent tothe critical zone is typically Class 7 (FS 209Class 10,000). In Europe, this area is GradeB, for which there is no precise analog ineither the ISO or now withdrawn (re-placed by ISO 14644) FS 209E classifica-tion schemes. These classification schemes,

however, can be considered functionallyequivalent provided validation testing andongoing environmental controls indicatethat performance is in compliance withexisting guidelines and standards. Firmsmay use different approaches providedthey have a rationale based upon good sci-entific and engineering practice. It may bebeneficial to include a HACCP analysis tosupport a firm’s cleanroom design strat-egy for an aseptic operation.

Other parameters typically consideredin the design of an aseptic processing areaare direction of airflow, air balance, airchanges per hour, and air velocity. Thereare numerous engineering guidelines con-taining sound design recommendationsfor aseptic processing areas (23). Theserecommendations have changed very lit-tle since 1988, essentially because the de-sign features known to be important in1988 are still recognized as vital to goodcleanroom design today.

There has been a tendency for both airvelocity and uniformity of airflow to beoveremphasized in today’s regulatory en-

vironment. The 90-ft/min or 0.45 m/s airvelocity requirement is a reasonable andeffective target value. However, it is notreasonable to conclude that these valuesare in any way sacrosanct (24). In fact, be-fore the publication of the 1998 positionpaper, all references to air velocity wereremoved in FS 209C and do not appear inISO 14644-1, which replaced FS209E. Theadequacy of an air velocity and the closelyrelated specification of air changes perhour depends upon several factors, includ-ing total room volume and location of theHEPA filter or air entry point relative tothe work zone. Therefore, quite differentvelocities may provide similar levels ofperformance depending upon the designand usage of the facility.

One must also recognize that when airvelocity was included as a cleanroom de-sign specification, measurement of air ve-locity was taken approximately one footfrom the face of the HEPA filter. There isno basis to require that any specific air ve-locity be attained at the work surface. Thework surface, because it is oriented essen-


tially perpendicular to the direction of airsupply, will interfere with the airflow andmake the reproducible measurement ofair velocity at that level problematic. Thisis a reality of aseptic processing criticalzones that existed in 1988 and will con-tinue to exist in the foreseeable future.

Because the work surface is located per-pendicular to the direction of airflow andbecause the work surface cannot be made

aerodynamic, maintenance of strict uni-directional airflow at the work surface isnot possible and cannot be reasonably ex-pected. Visualization of airflow may bebeneficial for optimizing air movements;however, more meaningful evaluative in-formation can be obtained using theLjungqvist-Reinmüller method (25).

Visualization (commonly called “smokestudies”) is subjective because human ob-

servers commonly judge the outcome with-out any clearly delineated criteria. Differ-ences of opinion regarding what constitutesgood or bad airflow are thus commonplace.Visualization has been overemphasized byregulators and should not take precedenceover objective data indicating that there arereliable aseptic conditions regardless of theobserved flow patterns.

Another measurement that has beenoveremphasized during the past decadehas been the uniformity of airflow fromHEPA filter to HEPA filter as well as acrossthe face of a HEPA filter. Expectations re-garding airflow should be realistic andmust consider the variable performancecharacteristics of HEPA filters as well asthe accuracy of the velocity measurement.The current frequency for HEPA filter re-certification in the industry is 6–12months and has proven quite adequate.Monitoring differential pressures acrossthe HEPA filters is enough to ensure thatthe filters are performing properly betweenrecertification events.

In 1988, PDA noted that cleanroom de-sign and operation guidelines should notbe applied too rigidly. This situation hasnot changed. No cleanroom design, oper-ation standard, or guideline is compre-hensive and most contain informationbased upon collective industrial experi-ence rather than unequivocal scientific orengineering facts. A suitable process is bestdefined by its validated ability to manu-facture product having the required qual-ity attributes. Evaluating processes requiresthoughtful analysis by experienced, well-trained technicians.

Environmental monitoring. Comprehen-sive environmental-monitoring programswere a general practice in the industry inthe late 1980s. Since then, these programshave become even more expansive (26).Unfortunately, there is no evidence thatthe environmental programs of today arefunctionally superior to those of 15 yearsago. They are certainly more costly and farmore time-consuming, but we do not be-lieve that they do a better job of assessingproduct safety.

The basic sampling methods and ap-proaches to environmental monitoringhave changed very little. What has changedis the amount of monitoring conductedand the actions expected by regulatory in-spectors in the event of excursions. Envi-


ronmental excursions, results that exceedaction levels, occur at a slightly lower fre-quency than they have in the past. This isdue to improved equipment design thathas reduced cleanroom population, bet-ter air movement, and probably most sig-nificantly improved cleanroom clothing.

We believe that the industry, in partthrough regulatory pressure, is now mis-using environmental monitoring to a sig-

nificant degree. The inherent uncertaintyand inaccuracy of environmental moni-toring are increasingly forgotten, and infar too many instances the results of sam-pling are considered something of a prod-uct release microbiological quality assay(24). In the most extreme cases, namelysamples required on so-called “critical sur-faces,” environmental monitoring has alltoo often evolved into a de facto product-

release sterility test. This is scientificallyinappropriate. We object to the applica-tion of Barr decision-like out-of-specifi-cation (OOS) test interpretation and res-olution in the event of environmentalmonitoring excursions. FDA’s OOS guid-ance was never intended for applicationto microbial testing, and this position hasbeen reaffirmed numerous times by theappropriate FDA personnel (27).

Environmental monitoring must takeinto account the realities of microbiolog-ical growth and recovery. The measure-ment accuracy of active air samples, oftenbut erroneously called “quantitative” airsamplers, is limited. A study by Reinmüllerand Ljungqvist found that the variabilityamong commonly used active air samplerscan exceed five-fold (25). Clearly then, set-ting an action level of 3 colony formingunits (cfu) and an alert level of 1 cfu is notlogical. In addition, microbiological enu-meration, particularly at low organismtiters, is prone to significant variability.Therefore, it is more scientifically reason-able to monitor trends by evaluating con-tamination incidence rates rather than byplacing stock in counts of colonies.

The environmental sampling planshould include representative sites in theISO Class 5 critical zone as well as the sur-rounding environment. These sites shouldbe sampled daily. Noncritical surfaces suchas walls and floors should be sampledweekly to ensure that the firm’s disinfec-tion program is performing adequately.Routine environmental sampling on per-sonnel should be limited to glove testing,although more-intensive personnel sam-pling on a quarterly basis may be desir-able in some cases.

Direct sampling of product-contact sur-faces such as parts hoppers or bowls andfilling needles may be conducted, but theresults should not be considered indica-tive of sterility or asepsis. Many firms con-sider sampling on these surfaces unnec-essary and we don’t consider data collectedon these sample sites as any more impor-tant than air sampling. In fact, recent stud-ies conducted by PDA demonstrate thatsurface contamination is a poor predictorof media-fill outcome (28).

It may seem logical to believe that any-thing less than perfect environmental re-sults indicate a loss of sterility assurance.However, this opinion is not scientifically


valid or supportable. It is unreasonable toexpect an aseptic environment contain-ing human operators to be sterile. Indus-try has long understood that the word ster-ile must be defined in a less than absolutemanner if it is to be used at all in the con-text of aseptic processing. Using environ-mental monitoring as a product-releasetest is inappropriate because sampling canneither prove nor disprove that contami-nation exists in any given unit or lot. Mon-itoring is today exactly what is was 15 yearsago: a method for assessing that theprocess control asserted in a cleanroom iswithin compliance with general industrystandards and that it is maintained withinprocess capability. The meaning of single-point excursions is unclear and should notbe overinterpreted.

Most important though, we must, as anindustry, focus on product safety ratherthan on an absolute and abstract expec-tation that a perfect sterile environmentmust exist. Sterile rooms do not exist, andthey never have. Rejection of productowing to environmental excursions even

on critical surfaces is a game of chance,rather than a valid quality control proce-dure. The current regulatory-driven ap-proach to environmental monitoring isone area in which aseptic processing hasnot improved since 1988.

Advanced aseptic processing. Advancedaseptic processing can be defined as tech-nologies that through automation or en-vironmental separation actively or pas-sively reduce risk from human-bornecontamination. Because human-bornecontamination is really the only risk ofconsequence in human-scale cleanroomaseptic processing, it should be obviousthat technologies that reduce the likeli-hood of operators releasing microorgan-isms near open product or componentscan further improve the already impres-sive safety achieved in aseptic processing.

Advanced aseptic technologies includeisolators, restricted access barrier systems(RABS), blow– or form–fill–seal technolo-gies, and various types of machine au-tomation. The upsurge in the implemen-tation of these technologies is visible

throughout the industry (8). In 1988, thefirst isolator-based aseptic filling systemswere just being implemented, today thesesystems number in the hundreds. Blow–or form–fill–seal technologies have beenused in industry for more than 30 yearsand continue to undergo incremental im-provement. RABS systems provide ameans of upgrading existing aseptic pro-cessing systems by reducing the likelihoodof human-borne contamination in criti-cal operations. Examples of reduced riskthrough automation abound and includeloading of lyophilizers, component replen-ishment, and checking and adjustment ofcontainer fill weight.

In 1988, PDA cautioned that, “A dog-matic approach could stifle the develop-ment and implementation of technologywhich could markedly improve the SALof sterile products” (1). Unfortunately, nei-ther industry nor the regulatory commu-nity heeded PDA’s advice (29). In the caseof isolator technology, bad decisions madeby industry advocates and regulators havehampered implementation, particularly


in the United States.Industry advocate groups made the

mistake of setting a target of performancefor isolators equivalent to terminal steril-ization. This was a very unfortunate strat-egy because the demonstration of equiv-alence in terms of absolute sterility isimpossible. This mistaken target settingresulted in a plethora of validation expec-tations that were difficult to understandand implement. Targeting perfection re-sulted in the expectation of a perfectly ster-ile enclosure environment, perfect trans-fer technologies, and perfect systemintegrity. In the case of leaks, the actualmicrobiological significance of so-called“breaches” wasn’t considered and insteadtheoretical notions of perfection replaceda pragmatic approach to systematic tech-nological improvement.

We believe that advanced aseptic sys-tems have more than met the expectationswe had for them in 1988. Industry andregulatory authorities need to see thesetechnologies for what they are: an impor-tant incremental improvement in asepsisarising from reduced human-borne con-tamination risk. The target for validationof these systems should be improvementover conventional cleanrooms, not equiv-alence to terminal sterilization. It is illog-ical and therefore inappropriate for firmsto concern themselves with abstract ortheoretical risks that cannot be measured(30).

Validation techniques for these systemsshould not be appreciably different thanthose used for conventional cleanrooms,although their process capability is higher.Obviously this higher capability shouldbe reflected in the in-process control andvalidation acceptance criteria used forthese more technologically advanced sys-tems. However, we reiterate that perfec-tion is not currently attainable and thatwe lack the tools necessary to measure per-fection. It is wrong to allow perfection tobe the enemy of good.

SummaryWe believe that aseptic processing in ourindustry has improved markedly since1988, including

• improved cleanroom garments and abetter understand of modes of con-tamination

• improved cleanroom designs and op-


erational performance• more comprehensive employee train-

ing and qualification programs.• improved aseptic processing equip-

ment requiring fewer line interven-tions

• well-established validation programsincorporating sound change-controlpractices to ensure continuing relia-bility of the processes

• implementation of advanced contam-ination control technologies such asisolators, restricted access barrier sys-tems (RABS), and blow–fill–seal orform–fill–seal systems.

The industry as a whole has earned anenviable safety record in aseptic process-ing over the years. In 1988, aseptic tech-nologies operating in compliance with in-dustry and regulatory guidelines resultedin product that was unquestionably andunequivocally safe with respect to micro-bial contamination. Today, as in 1988, theweakest link in the sterile and aseptic prod-uct delivery pathway is at the level of ad-ministration rather than manufacturing.Logically, those concerned with patientsafety would seem better advised to focustheir energies where they are likely to havethe greatest effect.

There have, of course, been recent re-ports of firms releasing asepticallyprocessed products that purportedly lacksterility assurance. We do not doubt thatin very rare instances product has beenmade in a manner contrary to good con-tamination control practice. However, itis inappropriate to punish the many forthe sins of a very few. Those firms lackingthe commitment to manufacture “sterile”products in an appropriate manner shouldhave known better. Successful productionof “sterile” products using aseptic process-ing can certainly be made possible if theprecepts outlined within this article areascribed to.

Perfection in aseptic processing is notcurrently achievable, because an absolutedemonstration of unequivocal sterility as-surance is not possible. However, safety isboth possible and routinely attained, andfor that industry can be proud. Industryunderstands that continuous process im-provement in our performance, especiallyin the production of “sterile” products isnot just desirable, but essential. We haveno doubt that in the coming decades risks

from microbial contamination will be evenbetter controlled than they are today. Afterall, the only significant source of contam-ination is widely recognized, and there aresuperb methods for minimizing this riskeven further than is possible today.

References1. J. Akers et al., “A Review of Current Technol-

ogy in Parenteral Manufacturing,” J. Par-enteral Sci. Technol. 42 (2), 53–56 (1988).

2. FDA, “Guideline on Sterile Drug ProductsProduced by Aseptic Processing” (1987).

3. J. Agalloco and B. Gordon, “Current Prac-tices in the Use of Media Fills in the Valida-tion of Aseptic Processing,” J. Parenteral Sci.Technol. 41 (4), 128–141(1987).

4. J. Agalloco and J. Akers, “Current Practicesin the Validation of Aseptic Processing 1992,”J. Parenteral Sci. Technol. 47 (2), supplement(1993).

5. J. Agalloco,“Microbiological Evaluation andMonitoring of Cleanroom Environments ”J. Parenteral Sci. Technol. 47 (4), 152-1541993.

6. J. Agalloco et al., “Process Simulation Test-ing for Aseptically Filled Products,” PDATechnical Report No. 22, PDA J. Pharma. Sci.Technol.Vol. 50, No. 6, supplement, 1996.

7. Agalloco, J., Akers, J., “Current Practices inthe Validation of Aseptic Processing – 1996,”PDA Technical Report No. 24, PDA J.Pharma. Sci. Technol. Volume 51, No. 2, sup-plement, 1997.

8. J. Agalloco, J. Akers, and R. Madsen, “Cur-rent Practices in the Validation of AsepticProcessing: 2001,” PDA Technical Report No.36, PDA J. Pharma. Sci. Technol. Vol. 56, No.3, 2002.

9. S.S. Block, “Definition of Terms,” in Disin-fection, Sterilization and Preservation, S. Block,Ed. (Lippincott, Williams and Wilkins, 5thed., Philadelphia, PA, 2001).

10. T. Meltzer et al., “Filter Integrity Testing inLiquid Applications Revisited: Part I,”Pharma.Technol. 25 (10), (2001) and Part 2Pharma. Technol. 25 (11), (2001).

11. USP ^1116& “Environmental Control of Asep-tic Processing Environments” (US Pharma-copeial Convention, Rockville, MD).

12. Annex 1, EU GMP Directive, “Manufactureof Sterile Medicinal Products” (1996).

13. “Sterilization in the 1990s,” in PDA/PMA Pro-ceedings of Joint Conference (1990).

14. J. Akers, J. Agalloco, and R. Madsen, “MoistHeat Sterilization: Myths and Realities,” PDAJ. Pharma. Sci. Technol. 52 (6), 1998.

15. J. Robertson et al., “Sterilizing Filtration ofLiquids,” PDA Technical Report No. 26, PDAJ. Pharma. Sci. Technol. 52 (1), 346–350(1998).

16. J. Agalloco, J.,“Intervention/Incubation Prac-tice” (under consideration for insertion inTR No. 22) in Proceedings of the PDA Spring2002 Meeting (2002).

17. A.M. Cundell et al., “Statistical Analysis of

Environmental Monitoring Data: Does aWorst Case Time for Monitoring CleanroomsExist?” PDA J. Pharma. Sci. Technol. 52 (6),326–330 (1998).

18. J. Agalloco and J., Akers,“Validation of Asep-tic Processing,” in Encyclopedia of Pharma-ceutical Technology, J. Swarbeck and J. Boy-lan, Eds. (Marcel-Dekker, New York, NY,2001).

19. J. Agalloco, “The State of Current IndustryPractice: Results from the PDA 2001 Surveyon Aseptic Processing,” in Proceedings of PDASpring 2002 Meeting (March 2002).

20. USP ^1116& “Microbial Evaluation of CleanRooms, Isolator Environments Used for theManufacture of Aseptically ManufacturedSterile Drug Products and Other ControlledEnvironments,” Pharmacopeial Forum 25 (3),8264–8279 (May–June 1999).

21. ISO 14644-1, “Cleanrooms and AssociatedClassified Environments” (2001).

22. Federal Standard 209E, 1992, withdrawn2001.

23. ISPE Guideline, Sterile Manufacturing Facil-ities (1999).

24. J. Akers and J. Agalloco, “EnvironmentalMonitoring: Myths and Misapplications,”PDA J. Pharma. Sci. Technol. 55 (3), 176–184(2001).

25. B. Reinmüller, “Dispersion and Risk Assess-ment of Airborne Contaminants in Pharma-ceutical Clean Rooms,” Bulletin No. 56 (RoyalInstitute of Technology, Building ServicesEngineering, 2001).

26. J. Moldenhauer et al., “Fundamentals of anEnvironmental Monitoring Program,” PDATechnical Report No. 13, PDA J. Pharma. Sci.Technol. 55 (5), supplement (2001).

27. J. Aydlett, Proceedings of PDA Seminar onAseptic Processing (2000).

28. J. Lindsay,“Experience in Media Fills at PDATRI,” poster presentation at PDA Spring 2002Meeting.

29. J. Akers and J. Agalloco,“Isolators: Validationand Sound Scientific Judgment,” PDA J.Pharma. Sci. Technol. 54 (2), 110–111 (2000).

30. J. Akers et al., “Design and Validation of Iso-lator Systems for the Manufacturing and Test-ing of Health Care Products,” PDA Techni-cal Report No. 34, PDA J. Pharma. Sci.Technol. 55 (5), (2001). PT

Please rate this article.On the Reader Service Card, circle a number:

348 Very useful and informative349 Somewhat useful and informative350 Not useful or informative

Your feedback is important to us.

aseptic processing: a review of current industry...

Documents