DATA AND REVIEW Managing Aseptic Interventions James Ngalioco cerforming an aseptic process almost aways requires op- exators to manipulate sterile or sterilized products and components. These activities are universally considezed HE tw offer the greatest poteatal for introducing microbial ‘oistamnination (147. That potential has been recognized by ‘the US Food and Drug Administration inspectors and has led “to some substantial cmcerns raised in warning letters issued in ‘2000 (5-5). The measures taken fo respond to these concerns _ by the companies involved have implications for other firms peformipg aseptic processing. The PDA held a conference call in November 2001 with representatives of the Office of Com- pliance at FDA's Center for Drug Evaluation and Research (CDER) to address relevant issues and establish some ground rules forthe indi fol. Participants ta that conversa ‘tion were Joseph Farnulate and Richard Friedman of FDA, and This article outlinesa comprehensive ‘Eason rigged Manor, i peti represented " essence ofthat conversation was subsequently re- approach for organizing a firm's aseptic Ported in the PDA newsletter (72. ‘operations, including planning for routine This article outlines a comprehensive approach for organiz- and nonroutine interventions, establishing ing firm’ aseptic operations consistent with the results of that effective process simulations, and discussion and aligned with recently defined regulatory expec- determining which vials to incubate. {ations The approach was initially developed in the spring of 2000 to soit the specific needs ofa major manafsctarer of asep- tic products and subsequently was refined through application ina variety of settings at other firms. Hhaman interventions in aseptic processing Its appropriate to focus on human interventions performed ‘during aseptic processing because no other factor has the same potential for introducing contamination. Sterilization processes, ‘xvironmental sanitzation, room design, and heating and ver- fiation systems are all substantially les significant 2s sources F contamination. Personnel continuously shed microorgan- isms and particles to ther surroundings and gowning materi- 4s cannot contain the milions of organisms present om human skin. The proximity of personnel and thus substantial numbers James Agalloco is the eal potentially contaminating microorganisms to sterile mate- president of Agabioco & Associates, PO Box 899, Ballo Mead, NJ rials, components, and surfaces during the performance of in- (28502, te. 908.874.7556, terventions is largely unavoidable in staffed cleanrooms. iagatoco@a0l com. Restine and nonreatine interventions. Interventions in aseptic processing operations fll imto two main categories: routine and 56 Pamscra kag sc 205 wow. pharmtech.comData AND REVIEW nonroutine. Routine interventions are activities that ace inher- ‘Ent parts of the aseptic process and integral parts of every batch. ‘Typical routine interventions include: aseptic assembly of the equipment before use; {nitial product connection ar introdction:; start-up component supply ox introduction; initial fi weight or volume adjustment; periodic component replenishment; periodic fill weight or volume checking and verification; fil weight or volume adjustments environmental monitoring: ‘operator breaks and meals; operator shift changes: ‘product sempling; fier integrity testing; product container replacement; any other interventional activity which isan integral part of. the process. Noncoutine interventions are activities that are predamticartly- tablishing tighter acceptable quality levels for containers seals and other parts that must be assembled: and ensuring better contro! over component preparation to provide greater oper- tional comsistency. Process aetomation. Horan intervention alo can be reduced ‘by automating processes. Framples include robotic sampling fr fill weights, servo-adjustable fill volumes, automated te- moval of downed containers, and automated stopper seal in- tegrity testing. ‘Another way to reduce the number of interventions i to en- ‘corurage operators to examine the necessity of an intervention before it. For exarnplea vial that has fallen over on. «a turntable should be left on its side until t presents a problem feeding other containers. If no problem occurs, the fallen vial ‘wil sill be om the turntable atthe completion of the ill, thereby climinating the intervention entirely. Looking at the fst of interventions provided above, it might appear that a firm need only address those items. That ap- ‘proach may be insufficient, however. The preferred approach corrective and may not be a part of every batch. Although in. ite asserbble the operating personnel and have each operator iteory, nonrentine interventions may not be necessary deg the aseptic process, in practice such interventions are ast always required to correct some anornaly- Sonne common routine interventions imolve: stopper misfeeds or camping; fallen, broken, or jammed containers; any other ine malfunction requiring mamual correction ‘The perfect intervention In. aseptic processing, the perfect intervention isthe one that is not required. The fewer the interventions, the lower the Hikei- |rond of contamination. Reducing the number of interventions should be the goal in every aseptic process. Operators should ‘trive fortis goal throughout the operational feof the process. Interventions can be eliminated by at least three major means: Precessand procedaral design. Process and procedural design - “ements tat can reduce interventions include chminating,inter- ‘ventions by performing cean-in-place and steriize-in-place pro- cedures forthe filing assembly; removing samples afr process materials have been transferred; eliminating unnecessary sam- pling steps and using the pressure-hokd method for fier integrity verification to obviate the need fora downstream connection. Inaproving component quality. Interventions also can be elim- inated by improving component quality. Examples inchide es- 50 Pmmment ary wv 205 ‘develop a Bist of routine and nonroutine interventions. For {tis Wits inckipe sscpic operations, interventions may vary ‘nie from fill Fine to another, even if both are filing similar ‘products and containers. Variations in product type may add activities that are specific to individual situations. The initial ‘0a oF this activity i to compile a Fist ofall interventions for precttan pccepbl anno isk. Thenhseach inter- ‘vention should be discussed im detail anda preferred means for ‘executing the intervention should be identified —although two (or more operators may have identifi the same intervention, they may not perform it in an identical or even fully acceptable ‘manner. A cornpetent microbiologist familar with aseptic tech nique should participate in this process. Once agreed upon, cach intervention (whether routine and nonroutine) should be established within 2 single standard operating procedure for cach fill line, process, or prochct type. This SOP should be ap- plied to both process simalations and routine operations. ‘The procedures must include sufficient detail to eliminate ambiguity. For example, because many interventions require the removal of containers or components from the line the ex- tent of removal must be exact. For instance, an SOP might in- dicate that when clearing a stopper jar, the operator should remore three containers on cither side of the container under the track where the jam occurred. Whenever possible, the mim- bers of components to be rerooved during each intervention should be identified and that number should be consistent for simulation and production. In some instances, it may be preferable to define the extent of removal by location rather than number. The procedure might state, “remove all open containers between the lat fill wow, pharmteeh.comDATA AND REVIEW head and the stopper machine” rather than, “remove the six open containers between the stopper machine and the lst ill head” If the conweyor happens not tobe full be- tween those locations, requiring the re moval of six vials may cruse confusion and lead to inconsistent execution of the task. ‘Training personnel to execute the in- terventions sof paramount importance. Fach person vo is expected to perform 60 Parmar chai sic 205 an intervention in routine operation must become proficient in its execution. The preparation of a video depicting the proper methods for each intervention is highly recommended because it can be ex- tremely difficult to convey in words the proper method for performing an inter vention. Ifthe intervention methods are identical for all operators, questions re- garding operator pasticipation will be re- duced. Thus, the intervention procedure becomes qualified regardless of which op- erator performs it. If firm has many op- erators, this alignment of procedures can help simplify the process simulation pro- gram. When new operators are introduced into the operation, the approved interven- tion procedures become a major focusof their tak-specific training. Process simulation of routine and — noaroutine interventions | Because routine interventions ae integral | and necessary parts of every aseptic | process, their execution during process | simulation should occur atthe same fre- quency asin an ordinary aseptic produc tion process Fortunately, thisis relatively | easy to accomplish. The setup of the Hine for the process simulation should follow ‘essentially identical procedures o those used for production. The only differences _might be using ar instead of nitrogen for blanketing or purging (to enhance recov- ‘xy of potential microbial contaminants) | and adapting in-line polishing filters to | maintain flow rates The remaining rou- | tine interventions are cther prescheduled be performed at the same frequency in both, thus ensuring thatthe simulation is Ja valid representation of the routine | process. | Nouroutine interventions vary substan- | tally from routine interventions. Noarou- | tine interventions occur randomly during | the proces in response to faults. The fre- quency with which they occur may vary substantially 2s a result of factors outside ‘our knowledge or bility to control. To en- sare their correct execution during rou- ‘ine operations these interventions must be included in proces simulations at are- alistic frequency level. ‘Nonroutine interventions should not | be optional in simulations. If operators | do not practice nonroutine interventions during simulations, the operators will be "unable to perform those interventions ‘The suggested means for integrating, nonroutine interventions into a process ww phacmtech.comDaTA AND REVIEW simulation isto schedule them as if they were integral to the process, al approxi mately the same frequency with which they occur during normal operations. ‘The operators should perform the non- routine intervention following the ap- proved procedure as closely as possible. “The media fil observer (whose presence is strongly recommended during every Process simulation) must ensure that — 62 Permit chang x20 205 nonroutine interventions are executed correctly. Routine and nenrowtine interventions in routine operation | Interventions that occur during routine ‘operations should be performed using the methods that have been established as | SOPsand practiced during process sima- [lation studies. Occasionally, however, something may occur during an aseptic process that requires an intervention that was never considered or that was consid- cred to present a unacceptable level of rik. "The process should he terminated at that point. Unless an intervention has been fully integrated into both the process sim tiation and routine operational setting, ino assurance exists that it can be per formed without introducing contamina- Bon, No intervention should he permit- junkessit has been satisfactorily accom- ed in atleast one media fil. If the | seed to perform an unapproved interven- ton occurs with increasing frequency the | firm should take corrective measures to | alleviate or eliminate the situation. If ap- propriate, the firm can add interventions to its approved lst, provided those inter- ‘ventions are incorporated into the process ) simulation program. | 18's important to document in the | | batch records ll nonroutine interventions | that are performed during both process | simulation and routine filling. The docu sentation should include the time the in tervention was performed and itsidenti- | ion frequency allows the firm to develop | its process simulation program to support | the more-cormmon nonroutine interven- | sions in every process simulation and rel- | egate those that are less frequent to 2 Jonger interval between simulations. ‘Which units to incubate? All process simulations inciade three pop- ‘lations (see Figure 1) Population 1 rep- ‘resents marketable product and consists of fully acceptable units with integral ‘oniainer-closure and units with cosmetic defects that ordinarily would be removed during post-fill inspection. Population 2 Consists ofimervention units removed by ‘procedure or practice daring the course of the fill Any units that are automatically sorted out ofthe batch by the equipment ‘ya pnarmtacn.comData AND REVIEW Population 1 Represents marketable product plus cosmetic defects from post-fillinspection operations. r cmp Srtgra mi Sloot map arn ap Seg Population 2 ———» Population 3 Examples: Intervention units Change in filing needle Change in filing equipment ‘Stopper jam + Dose adjustments/samples “= Any units on floor + Automatically ejected units rerentative of marketable product should bein- cubated This means that only Population { should be incubated. Units removed from the filling process, either as part ofan intervention during the media ill, Population 2) or aftr the fill is completed (ie, Population 3) should be dis- carded. The majority of these units willbe un- filled, unsealed, or uncrimped, Manually sealing and crimping these units isnot apart of routine operations because its key to introduce con- tamination. Little benefit willbe gained from testing such vials. Detecting contamination in the removed units does not imply contamina- tion ofthe overall aseptic process the presump- tion of contamination that prompted their re- moval in the first place suggests that any ‘Figure 1: The tree populations that occur in all process siraiatis. contamination detected in those units would nat be represen tative of the the remaining units. during processing would 2kso be included in Population 2. Pop"? ~All oontainers in Population | must be included in the statis- ulation 3 consists of any unit with compromised contaiiereclo- tical evaluation of the media fill. No containers can be disre- Senegal tm ‘garded once they have been incubated. This population repre- marketable product and vials with cosmetic defects that should be removed in posting inspection. The cosmetic de- ‘ects are included in Population L because their removal during 6 Pusat tay sa 205 ww.pharmtech.comDATA AND REVIEW spection process, and thus they are poten tially representative of marketed product. Incubating any remaining media that ‘was not filled into containers may be use- ful, but certainly is not required. If that ‘media is heavily contaminated, it might invalidate the entire fll. However, i the residual media is contaminated but the filled units are fee of growth, then the re- salts of the fil must stand. In effect, the 6s Premera ching sc 205 presence of growth in the residual media has ttle relevance tothe media fill, except in those instances where it results in the failure of the media fll, which ordinarily ‘is considered a non-test,requiring thatthe ner for both process simulation and rou: tine operation, the adequacy of the asep- fic process can be demonstrated by suc- cessful media fils. Aligning. the intervention procedures eliminates the test be repeated “uncertainties associated with varying prac- tices and supports the successful execu- Condusion tion. Encubating the units removed du- ‘When intervention practices are detailed in procedures used in an identical man- ing processing provides no additional assurance—these units are always rejects daring either production or process sim- ulation. Incubating nonintegral contain | ers isan exercise in folly. Process simula- tions are assessments of aseptic processing capability and are not definitive determi- nations of sterility assurance. Inferences ‘that not-incubated units evidencing con- tamination are indicative of aseptic process failure represent a biased viewpoint. | nterences | tL Agalloce and B. Gordon, “Current Prac | eebin the Use of Media Fil i the Vida [tion of Aseptic Processing”. Parenteral Si. Teco AU (8), 128-141 (1987). 2 J. Aglloco and J Akzs, "Current Practices in the Validation of Aseptic Process inng—1992" | Parereral Sot Techasl 47 (2), sauppesnest (1993). 3. J Apalloce and J Akers, “Current Pracices ‘the Validstiog Process * ¥ . ‘5. Rood ana Deg Administration, Waring Lt- tert Akan Laborsories, 17 Nevermier 200, wwncagviStearning leer mn77ngple | (accessed 18 February 2005). |) Feodand Drag Administration, Wiring Let | tert Hii Lally & Co. tne, ? March 2001, |, srewcida gonifeiearning letersm52570, | opal accssd 1 February 2005. 7. PDA, PDA Newsletter, February 2002, US Food and Drug Administration, Draft (Coocept Paper oo Sterde Drg Products Pro doc by Ase Proceasing (EDA, Rockilie, DMD, September 20023. PF J. Pharm, Sei. Technol, $6 (3), supplement wirwepharmtech.com