Pharmaceutical Engineering - July August 2008

July/August 2008

Volume 28, Number 4

Theme: Process Development

Table of Contents

Articles

1. Increased Process Understanding Through Monitoring and Scale-Down Models:

Case Study of a Cell Culture Harvest Fluid Titration and Filtration Process

by Kurt Yanagimachi, Corey Dodge, and Marisa Hewitt

This article presents a case study demonstrating the use of detailed process monitoring and

scale-down modeling in determining the root causes of yield losses in the manufacture of a

therapeutic enzyme and identifying alternative technologies to improve the manufacturing

process.

2. Case Study: Parenteral Facility Upgrade Project with Fill Line Install

by Keith Weseli and Michael DiGiovanni

This article presents a case study illustrating project management and commissioning and

qualification processes that allowed for accelerated completion of a renovation project.

3. Integrating Industrial Engineering and Lean Techniques at a Contract

Pharmaceutical Manufacturer

by Valerie Maier-Speredelozzi, Cyrus Agarabi, Thomas Needham, and Sirine A. Saleem

This article presents the application of industrial engineering and lean techniques to a

contract pharmaceutical manufacturing facility.

4. Retrofitting CIP into API Plants

by Nigel A. Fletcher

This article describes the physical modifications and additions retrofitted into existing plants

to incorporate CIP technology and some of the techniques that can be used to ‘stretch’ the

existing CIP systems for best effect.

5. Solving the Terminology Conundrum

by Robert Adamson, Nuala Calnan, Robert E. Chew, and Steven J. Wisniewski

This article discusses the terms “commissioning,” “qualification,” and “verification.” Do the

terms refer to the same or different ideas? How should the pharmaceutical and

biotechnology industries use these terms in a consistent and meaningful way? This article

provides a compilation of how these terms are used in regulations and by various

industries, and provides a proposal for clear definitions to be used as ISPE updates and

creates Baseline® Guides.

Departments

6. ISPE Update

7. Classified Advertising

Online Exclusive Articles

8. Your Questions, Your Answers

A sampling of interactive discussions taking place online among colleagues through ISPE’s

Communities of Practice.

9. Global Regulatory News

JULY/AUGUST 2008 PHARMACEUTICAL ENGINEERING 1

Process Monitoring

©Copyright ISPE 2008

This articlepresents a casestudydemonstratingthe use ofdetailed processmonitoring andscale-downmodeling indetermining theroot causes ofyield losses inthe manufactureof a therapeuticenzyme andidentifyingalternativetechnologies toimprove themanufacturingprocess.

Increased Process UnderstandingThrough Monitoring and Scale-DownModels: Case Study of a Cell CultureHarvest Fluid Titration and FiltrationProcess

by Kurt Yanagimachi, Corey Dodge, and Marisa Hewitt

Introduction

The identification of ways to improve amanufacturing process often requireunderstanding of the process beyondthe “black-box” treatment, in which only

the basic metrics such as step yield and product

quality are determined. A higher level of pro-cess understanding can be accomplished byclosely monitoring key process attributes, ide-ally in real-time, and performing an in-depthstudy of the process using scale-down models.The combination of these two approaches also

Figure 1. rhXmanufacturing processand pH-Adjust processflow diagram.

Reprinted from

PHARMACEUTICAL ENGINEERING®

The Official Magazine of ISPE

July/August 2008, Vol. 28 No. 4

2 PHARMACEUTICAL ENGINEERING JULY/AUGUST 2008

Process Monitoring


can allow the identification of discrepancies between scalesand can potentially highlight process improvements thatmay not be obvious with study at only one particular scale.The following case study is an example of how this combina-tion approach lead to the identification of the root cause forprocess-scale yield loss and highlighted alternative technolo-gies that would allow for the recovery of this product uponimplementation.

An important step in the manufacture of a recombinantenzyme biopharmaceutical “X” (rhX) is the titration andsubsequent filtration of cell culture Concentrated HarvestedFluid (CHF) prior to the first chromatographic step. Thistitration results in significant precipitation of Host CellProteins (HCP) and the loss of a small, but significant amountof rhX that becomes entrained in the precipitate. Figure 1shows the position of the pH adjust unit operation within theproduction process of rhX, as well as a flow diagram of thatparticular step. The process is summarized as follows:

• A pool of concentrated harvest fluid is pumped into the pHadjust tank by way of a heat exchanger.

• Low pH titrant is added via a pH controller until the CHFpool reaches the target pH. During titration, isoelectricprecipitate consisting mainly of HCP is formed and mustbe filtered.

• Titrated CHF is pumped through the filter train to removeprecipitate. The filter train consists of a lenticular depthfilter, followed by a 0.2 µm (5.1 mil) membrane filter,followed by a second 0.2 µm sterile bag filter. The filteredproduct is stored in a sterile collection bag until ready toload onto the chromatography column.

• Chase buffer is added to the pH adjust tank and is pumpedthrough the filter train into the collection bag. The filterhousings are then purged with air to collect as much of theresidual product as possible.

The yield of this unit operation at the bench scale duringprocess development was in the range of 85-90%. However,upon scale-up to the cGMP manufacturing facility, the yieldhad been considerably lower, averaging 77%. Also, filtrationclogging occasionally occurred with one of the filters (labeled“Membrane Filter A” in Figure 1) at manufacturing scalewhich was not observed during process development andwhich continues to elude reproduction at the bench scale.Some possible reasons for these discrepancies could include:

• Differences in mixing or mass transfer characteristics atsmall and large scale that lead to differences in yield on thetitration step.

• Differences in filtration due to filter geometry or filterdesign impact. The manufacturing filters are of a lens-shaped lenticular cell design, allowing more filter surfacearea to be condensed into a smaller volume. Conversely,the filters available for small-scale studies are flat, circu-lar discs contained in a small capsule or housing.

• Differences in filtration due to filterability of the titratedCHF. If mixing characteristics are different at small vs.

large scale, this could lead to differences in the sizedistribution of the precipitate particles formed.

ObjectivesThe primary objective for this work was to increase thefundamental process understanding, and if possible, to for-mulate a plan of action for increasing yield and/or reducing/eliminating the occurrence of filter clogging. The plan for thisstudy was to:

1. Follow two batches of the pH-adjust process in the manu-facturing plant and conduct extensive sampling and real-time monitoring to quantify each source of yield loss.

2. Develop a representative scale-down model of the pHadjust operation.

3. Utilize the scale-down model to determine the root causesof the drop in yield and occasional membrane filter clog-ging incidents.

4. Evaluate process alternatives that could possibly allevi-ate the aforementioned problems.

Monitoring of Manufacturing Batches and Scale-Down ModelingBefore evaluating any potential process alternatives with thehopes of improving yield and filtration performance, under-standing of the process in its current state was enhanced.According to current standard operating procedure, samplesare taken only of the starting material (CHF) and of the finalsterile-filtered column load. In order to determine exactlywhere in the process rhX is being lost, two batches of the pH-adjust process in the manufacturing plant were followed.Samples of each pool of CHF were taken before titration,during titration at regular pH intervals, and after the finalpH setpoint was reached (refer to Figure 1 for sample loca-tions). Samples also were taken from several points in thefiltration train throughout the filtration of each pool of pH-adjusted CHF and during the buffer chase. Each sample wasassayed for rhX activity, total protein, turbidity, and in somecases, particle size. In addition to the detailed sampling, anon-line back-scattering turbidity probe was installed in thepH adjust tank to monitor turbidity during titration and anin-line forward-scattering turbidity probe was installed inbetween the depth filter and the first 0.2 µm membrane filterin order to monitor depth filter breakthrough in real time.

Power input/unit volumeImpeller type (A310 hydrofoil)Turbulence (as determined by the Reynolds Number)Acid addition siteAcid titrant addition rate per volume CHFFinal pHTemperatureFlux (CHF flow rate/unit filter area)Load (Total volume CHF/unit filter area)Volume chase buffer/unit filter areaFilter media types

Titr

atio

nFi

ltrat

ion

Table A. Summary of scaling variables and conditions maintainedfor titration and filtration.


Process Monitoring


Scale-Down ExperimentsSmall-scale pH adjustments and filtrations were conductedwith the goal of reproducing the performance of the manufac-turing-scale pH adjustments and filtrations that were previ-ously followed. Samples of each CHF pool prior to titrationwere taken to use as starting material for the small-scale pHadjustments in order to rule out lot-to-lot variability of CHFas source of discrepancy between small and large-scale data.Table A details the scaling variables used to scale down thetitration and filtration steps.

Literature references were consulted that highlighted theimportance of mixing in processes involving reaction, aggre-gation, and precipitation.1,2 In the majority of these cases, themost important variable in scaling down a mixing step is thepower input per unit volume (P/V), which determines theaverage shear rate and micromixing time scale. P/V iscalculated as follows:3,4

P Po N3 D5ρ___ = ____________V V

Where:N = Impeller rotation rateD = Impeller diameterρ = Liquid densityV = Liquid volumePo = Dimensionless power number.

The P/V was matched as close as possible to the process-scaleP/V calculated for each individual batch, while maintaininggeometric similarity if possible (see Figure 2 for photos of

small and large scale titration equipment). The titrant addi-tion rate per volume of concentrated harvest fluid and theposition of the acid addition site relative to the impeller alsowere kept the same between the small and large scale.Turbidity was monitored on-line using the same back-scat-tering turbidity probe used with the full-scale system. Sampleswere taken at prescribed pH values and later assayed fortotal protein, rhX activity, and in some cases, particle size.

For filtration scale-down, the filter surface area ratiobetween the depth filter and downstream membrane filterwas maintained.5 A small-scale filterability system (see Fig-ure 2 for photos of small and large scale filtration equipment)was used to measure the in-line system pressures, volume offiltrate recovered, and instantaneous flow rate. The forward-scattering in-line turbidity meter was installed in between

Figure 2. Small-scale and manufacturing equipment used fortitration and filtration.

Figure 3. Comparison of the pH adjustment at different scales, (a)Averaged titration and filtration yields for each step at Scale-down(SD) and Manufacturing scale (MFG), (b) Turbidity profile of twotitrations at both scales, and (c) Particle size distribution ofprecipitates formed at each scale.


Process Monitoring


the depth filter capsules and membrane filter disk housing.The amount of each pool to be filtered was determined byusing throughput/unit depth filter area as a scaling factor.Flow rate/unit area also was conserved between large andsmall scale. Samples were taken in between the depth filterand first 0.2 µm membrane filter, as well as between the 0.2µm filters. Following filtration of the appropriate amount ofeach titrated pool, the scaled-down amount of chase bufferwas pumped through the filtration train.

Scale-Down ResultsFigure 3(a) displays the results of the rhX titration andfiltration yield comparison for the average of two batches runwith the Scale-Down (SD) model and in Manufacturing (MFG).It is evident from these graphs that the SD model provides anaccurate method for determining titration yield. Figure 3(b)shows the turbidity profile evolution throughout the course oftitration for one particular CHF pool at both scales. Simi-larly, there was close agreement between scales when mea-suring the enzyme activity and total protein during titration,adding further evidence to the accuracy of the scale-downtitration model. However, filtration yield is higher at thesmall scale. This in turn results in the overall discrepancybetween the small scale and large scale processes.

Figure 3(c) shows the particle size distribution of the pre-cipitate formed during one titration of the same starting poolof CHF at manufacturing scale and small-scale. The meanparticle size is slightly lower for manufacturing scale. Thismay be due to the higher impeller tip speed, and thus, highermaximum shear rate at full-scale. It also could be a result ofimmediate analysis of the small-scale samples, whereas thefull-scale sample was analyzed two days later after storage at4°C (39°F). In both cases, the majority of particles formed areabove the 6 to 15 µm range in diameter (152 to 381 mil), thenominal rating of the particular depth filter used.

The onset of the occasional clogging of the membrane filterdirectly following the depth filter also was investigated bymonitoring the turbidity of the depth filtrate. While the firstbatch was completed without any filtration difficulties, the

membrane filter did clog during the second batch, resulting inthe need to replace the clogged filter cartridges before complet-ing the batch. However, when this second batch was “repro-duced” at the small scale with identical filter fluxes and loadsand identical starting material, no filtration problems wereobserved. Figure 4 displays the turbidity profiles of the filtratefrom the depth filter for the two manufacturing batches.Turbidity of the filtrates at small-scale was undetectable.

One theory for this discrepancy is that the process-scaledepth filter is not performing to the same degree as its small-scale counterpart, resulting in considerable breakthrough ofparticles, which then are retained in the membrane filter. Tofurther investigate how the depth filter cartridge could con-tribute to the difference in yield seen across the scales, one ofthe used process-scale depth filter cartridges was retrievedand an “autopsy” was performed.

Depth Filter AutopsyFigure 5(a) shows a picture of the depth filter cartridge (lens-style lenticular) after completion of the first batch. Uponinspection, it became evident that the filter was not perform-ing as it should with a significant amount of area around thecenter of each cell remaining free of precipitate cake build-up.

Because of the lens-shape of the lenticular cells, the gapbetween cells is wider at the perimeter of the cartridge thanin the center region. Apparently, the gap spacing close to thecenter region was not sufficient to allow passage of theprecipitate for cake formation. Thus, the surface area at thecenter of the cartridge is largely underutilized. It is evidenthow such a cartridge design could impact filtration, as theeffective flux and load would be higher at process scale thanwhat was utilized at bench scale. Either of these effects bythemselves could result in the reduced adsorptive capacity ofthe depth filter, and so the cumulative effect could explain thepoorer depth filter performance at process scale. By cuttingsections of defined dimension (2" × 2" or 5 cm × 5 cm) fromdifferent areas on the used cartridge and extracting andassaying the residual rhX, the concentration of rhX wasdetermined in the residual liquid contained within the filter

Figure 4. In-line turbidity of filtrate from depth filter, (a) Batch 1, (b) Batch 2.


Process Monitoring


at four different areas from the filter cartridge that werecaked to different extents. Figure 5(b) shows the results ofthis assay. In this graph, the concentrations of rhX have beennormalized to the final concentration measured in the chasebuffer as it left the depth filter. Ideally, the residual rhXconcentration in the filter would be equal to the final concen-tration of chase buffer, giving a normalized value of 1. Thereis a significant amount of product left behind in the filter andheterogeneity with respect to position within the filter car-tridge. The amount of product left behind on the filter at anygiven position correlated directly to the relative extent ofcaking above that piece of media.

It is likely that the exclusion of precipitate cake at thecenter-most depth filter media creates a path-of-least-resis-tance for the chase buffer, which then flows straight throughthe center region of the cartridge without sweeping residualsoluble rhX from the precipitate cake or liquid holdup con-tained within these covered regions of the depth filter media.In Table B, the impact of this path-of-least resistance phe-nomenon and of other sources was determined on the rhXyield for the filtration step of the process. It was assumed thatthe 75% of the total filter surface area was covered with cakeand 80% of the volume of the wet cake was liquid. All of thesesources of product loss would only be encountered in the full-scale process, and nearly all of the difference between thesmall-scale and process-scale filtration yields were accounted

for by the chase buffer path-of-least-resistance phenomenon(7% loss).

Evaluation of Filtration ProcessModifications

The next objective was to explore several potential modifica-tions to the process with the goal of recovering that lost 7%and bringing the pH adjust yield more in line with the small-scale yield and reduce the incidence of membrane filterclogging. The modifications investigated were:

1. Replace the depth filter cartridge with a cartridge designthat allows more uniform cake distribution across itsentire surface area.

2. Replace the existing dead-end filtration process with amicrofiltration/diafiltration process.

Alternative Depth Filter Cartridge DesignsTwo alternative depth filter cartridge designs were studied:a lenticular design in which each cell is enclosed in a rigidpolypropylene cage (Figure 6) and a fully encapsulated non-lenticular design. The caged lenticular design allows uniformspacing between filter cells eliminating the tapered spacingbetween the lens-style cells. The fully encapsulated designeliminates the need for a central filtrate core altogether andthus, also has more desirable flow properties. While candi-date depth filter grades using the encapsulated design weretested and identified, the cage-style lenticular design waseventually chosen because it could utilize the same depthfilter media and be used in the existing filter housings. Thesewere distinct advantages when considering implementing aprocess change to an approved commercial process. Thiscage-style cartridge was evaluated at the pilot scale, main-taining filter flux and load consistent with the process scale.

Cage-Style Pilot-Scale ResultsA filter autopsy was performed on the spent filter cartridge

Contributors to Product Loss Batch 1 Batch 2Filter Sheet (Retained Liquid) 4.1% 5.0%Precipitate Cake (Retained Liquid) 2.3% 2.2%Filter Housing Venting 2.7% 2.8%Filter Cartridge Change-Out 0.0% 3.0%

Sum of Losses 9.1% 12.9%Overall Yield Loss Determined by Mass Balance 10.7% 13.7%Difference Between Yield and Volumetric Loss 1.6% 0.7%Quantities

Table B. Filtration Yield Analysis.

Figure 5. (a) Photo of spent depth filter showing heterogeneity of precipitate cake distribution in between lenticular cells. (b) NormalizedrhX residual concentration at various spots in the filter cartridge: A=Bottom cell, center region, B=Bottom cell, side region, C=Top cell,center region, D=Top cell, side region.


Process Monitoring


Figure 7. Microfiltration/diafiltration system setup.

using the same method previously established. Figure 6shows a photo of one of the cells with the polypropylene cagepartially removed. Upon opening the filter cartridge, it wasevident that the filter media was more evenly utilized (thebare spots are places where the filter cake stuck to the plasticmesh cage that was cut away). Samples were taken fromvarious spots in the filter cartridge and the residual rhX wasmeasured as previously. Again, the residual rhX concentra-tion was normalized to the final concentration of chase bufferleaving the filter.

These results support the theory for how the previous depthfilter’s design contributed to yield loss. The overall productyield for this trial was 94% (compared to an average of 88% forthe two manufacturing batches followed). The improved de-sign of this cartridge appears to allow a more uniform build upof the cake on the filter sheets that, in turn, prevents theformation of a path-of-least-resistance, allowing for a moreeffective chase buffer flow distribution and recovery of residualrhX within the cake and filter sheets. Depth filter break-

through also was not observed during the filtration, and therewas no pressure build-up on the 0.2 µm filter capsule even with40% less relative surface area. This is a good indication that theuse of this cage-style cartridge at the process scale shouldresult in fewer membrane filter clogging incidents.

Microfiltration/DiafiltrationThree trials of a hollow fiber microfiltration/diafiltration unitoperation were performed at bench-scale using CHF suppliedfrom manufacturing. A diagram of the TFF system is dis-played in Figure 7. The process can be summarized as follows:

• CHF is titrated per SOP to the target pH.• The titrated CHF was filtered by hollow fiber TFF until

approximately 66% of the liquid volume had been collectedas permeate, resulting in a 3X concentration of particu-lates. 0.65 µm (15 mil) hollow fiber membranes with aneffective membrane area of 460 cm2 (71.3 in2) and aflowpath length of 30 cm (11.8 in) were used.

Figure 6. Used depth filter cartridge cake distribution (a) and residual rhX measurements from filter autopsy; A=Bottom pad, B=Centerpad, C=Top pad (b).


Process Monitoring


Figure 8. Particle size distributions of the initial titrated CHF (feed) and intermediate retentate following microfiltration.

• Chase buffer was used as a diafiltration buffer. Diafiltrationbuffer was added to the retentate at the same rate at whichpermeate was removed. In total, 5 diavolumes of bufferwere used.

• The diafiltration rate and recycle rate were controlled tomaintain the transmembrane pressure within vendorspecification.

MF/DF ResultsThe MF/DF filtration process resulted in near completerecovery of rhX for all three trials with an average measuredyield for the three trials of 115%. The >100% result could bedue to sampling if the mixture was not uniformly mixed, or itcould be due to the resolubilization of the rhX which wasinitially entrained in the precipitate following titration. Theyield of HCP impurities also was higher (87% for MF/DF asopposed to 83% for bench-scale depth filtration), indicatingthat some HCP may be subject to the same resolubilizationphenomenon, or this may be a consequence of the removal ofthe adsorptive capacity of the depth filter media, which ispositively charged.

Particle size analysis of the initial pH adjusted CHF (feed)and intermediate retentate was performed and Figure 8shows the results of this analysis. It is apparent from thechanges in particle size that the shear force encounteredduring TFF is of great enough magnitude that the precipitateis broken up into smaller particles. A considerable number of

these particles are smaller than the 0.65 µm nominal ratingof the hollow fiber filter, which could increase the particleburden on the 0.2 µm membrane filter. Hence, there is thepotential for full recovery of the rhX with MF/DF, but theimpact on the downstream processing needs to be evaluatedin light of the higher HCP recovery.

ConclusionsThrough a combination of scale-down analysis and processmonitoring, the lens-style lenticular depth filter cartridgedesign was identified as the main contributor to the addi-tional yield loss seen at process-scale. A new lenticular cage-style filter design was evaluated at pilot scale and allowed formore uniform accessibility of the filter media. This in turn ledto a significantly lower level of rhX left behind in the filterholdup, and thus, a significantly higher rhX recovery. Chang-ing to the new design should result in improvements in yieldand a reduction of clogging incidents. The microfiltration/diafiltration process is promising based on the product yieldobtained in initial bench-scale experiments, but the impacton downstream purification operations needs to be deter-mined before proceeding any further. Table C summarizesthe key findings and recommendations of this study.

AbbreviationsCHF Concentrated Harvest Fluid

Filter Technology rhX yield Other Benefits Potential Issues with Adoption DecisionLens-Style Depth Filter Cartridges 88% Status quo N/A N/A(2 process-scale batches)Cage-Style Depth Filter Cartridges 94% Potential reduction in 0.2 μm Minimal – Same depth filter media as current Submitted as a process(one pilot-scale trial) membrane filter clogging due process. Cost increase is not significant change.

to more efficient utilization of compared to yield enhancement. Lowerdepth filter surface area. extractables than current filter.

Microfiltration/diafiltration 115% TFF could result in more Unknown impact on 0.2 μm membrane Further studies needed to(Three bench-scale trials) consistent, reliable filter filtration due to small particle generation. assess impact on product

performance, fewer Unknown impact on product quality. Lower quality and onoperational issues. HCP removal and increased volume could downstream purification

impact capture chromatography. processes.

Table C. Process modification evaluation summary


Process Monitoring


rhX Recombinant enzyme biopharmaceutical “X”, oneof BioMarin’s approved products

HCP Host Cell ProteinP/V Mixing Power per unit Volume

References1. Ducoste, J.J. Clark, M.M., and Weetman, R.J., “Turbu-

lence in Flocculators: Effects of Tank Size and ImpellerType,” AIChE Jorunal, 43 (2), 329-338, 1997.

2. Kim, Woon-Soo, Hirasawa, I., and Kim, Woo-Sik, “AgingCharacteristics of Protein Precipitates Produced by Poly-electrolyte Precipitation in Turbulently Agitated Reac-tor,” Chemical Engineering Science, 57, 4077-4085, 2002.

3. Oldshue, J.Y., Herbst, N.R, and Post, T.A., A Guide toFluid Mixing. p. 13. Lightnin Inc., 1992.

4. Weetman, R.J., and Oldshue, J.Y., “Comparison of MassTransfer Characteristics of Radial and Axial Flow Impel-lers,” Proceedings of the 6th European Conference on Mix-ing, Pavia, Italy, May 1998.

5. Rathore, A.S., and Wang, A., “Optimization, Scale-up, andValidation Issues in Filtration of Biopharmaceuticals,Part 1,” BioPharm International, 17 (8), 50-58. 2004.

AcknowledgementsWe would like to thank our colleagues in the ManufacturingSciences Group and Purification Process Development atBioMarin Pharmaceutical for their valuable insights. Wealso would thank the UC Davis Biotechnology Program forfacilitating Mr. Dodge’s internship.

About the AuthorsKurt Yanagimachi obtained his BS fromthe University of Washington and PhD fromMassachusetts Institute of Technology, bothin chemical engineering. He began his careerin 2000 with Bristol-Myers Squibb, as a re-search investigator in the Engineering Tech-nology Department where he developed re-covery and chromatographic processes for

small-molecule clinical materials. In 2002-2003, hetransitioned to biologics processes at Onyx Pharmaceuticalswhere he worked as an associate scientist in the Manufactur-

ing Sciences Department developing oncolytic virus purifica-tion processes. Currently, he is with BioMarin Pharmaceuti-cal Inc. as a Senior Process Scientist in the ManufacturingSciences Department evaluating and implementing processimprovements and new technologies to existing commercialbiopharmaceutical manufacturing processes. He can bereached by telephone: +1-415-506-6360 or by e-mail:[email protected].

Marisa Hewitt graduated with a BS inchemical engineering from Florida StateUniversity in 2003. Her career began at EliLilly and Company in plant engineering,where she delivered capital projects. In hernext role in Manufacturing Science and Tech-nology, she served as the technical represen-tative for a commercial antibiotic purifica-

tion process and initiated deviation and variability reductionprojects. In 2007, Hewitt joined BioMarin PharmaceuticalInc. as a Process Engineer in Manufacturing Sciences, whereshe provides technical support for biopharmaceutical manu-facturing processes and evaluates equipment and processchanges to improve GMP manufacturing. She can be con-tacted by telephone: +1-415-506-6679 or by e-mail:[email protected].

BioMarin Pharmaceutical Inc., 105 Digital Dr., Novato,California 94949, USA.

Corey Dodge is a PhD candidate in thechemical engineering program at the Uni-versity of California, Davis. He earned hisBS in chemical engineering at Texas A&MUniversity before beginning his graduatestudies. In 2006, he completed a six-monthinternship at BioMarin Pharmaceutical aspart of the curriculum of the cross-disciplin-

ary UC Davis Designated Emphasis in Biotechnology Pro-gram.

University of California, Chemical Engineering and Mate-rials Science Department, 1 Shields Ave., Davis, California95616, USA.


Commissioning and Qualification


This articlepresents a casestudy illustratingprojectmanagementandcommissioningand qualificationprocesses thatallowed foracceleratedcompletion of arenovationproject.

Case Study: Parenteral FacilityUpgrade Project with Fill Line Install

by Keith Weseli and Michael DiGiovanni

Project Summary

As part of an overall parenteral facilityupgrade, which spanned a number ofyears, this case study focuses on thestart-up, commissioning, and qualifi-

cation of a new aseptic filling line for thisoperational facility.

The location for the new aseptic filling lineis in an area of the plant where two asepticfilling lines already existed. These lines con-nect to multiple freeze dryers and capping linesand supported the filling, freeze drying, andcapping of multiple products. The majority ofproducts on these lines are dried, but there aresome liquid presentations. Multiple fill vol-umes and stopper/vial combinations were ac-commodated. The long term goal for the newfilling line is to replace both existing fillinglines.

The first phase called for the replacement ofone filling line and the implementation of anew process designed to reduce the number ofaseptic connections. The existing filling line

used 14 different vial/stopper combinationsranging from a 3 ml vial to a 50 ml vial andrequired line speeds ranging from 80 to 300vials per minute. This line was the only linefully validated to fill a number of products forthe facility and the corporation. Therefore, suc-cessful design, commissioning, qualification,and start-up were critical. The second existingline is scheduled to be decommissioned 12 to 18months following the successful installation ofthe new filling line.

One major goal of this project was to opti-mize the qualification documentation by focus-ing on key regulatory requirements rather thana combination of regulatory and operationalrequirements. While this focus significantlyreduced the number of requirements tested aspart of the qualification effort, it did not effec-tively identify key operational performancemetrics early in the project’s development phase.However, the Commissioning and Qualifica-tion (C&Q) team was able to identify and inte-grate these key operational requirements into

the C&Q process and used them toeffectively improve the performanceof the system prior to installing itinto the production facility.

The new process for connectingthe product fill tanks to the filler toreduce aseptic connections pre-sented additional challenges. Anumber of new equipment itemswere introduced to operators andthe level of automated control ofthe fill line was increased. Thesechanges brought about additionalaseptic equipment preparation chal-lenges and added an additionallayer of commissioning, qualifica-tion, and validation activities tothe project.

Activity At Off-Site At ProductionTesting Facility Facility

Requirements X Design Documents X Commissioning Testing X X*(Receipt, installation checks,throughput, startup andperformance tests)Installation Qualification X X*Operational Qualification X X*Performance Qualification XCleaning Validation XProcess Validation X*All documentation and testing activities necessary to qualify the systemand verify acceptable performance were completed at the off-site testingfacility. After installing the unit in the final location, minimal IQ documentswere re-executed to insure all equipment was accounted for and installedproperly. Additionally, some OQ and Comm testing was re-executed to verifyperformance in the final location was as consistent with what was seen inthe test facility.

Table A.Commissioning,qualification, andvalidation activitylocations.

Reprinted from







Installation for the new line was accomplished during afull maintenance shutdown of the production facility. How-ever, the location for the new line was in a common asepticfreeze dry area which was required to return to production asquickly as possible. Initial start-up, commissioning, andqualification through Operational Qualification (OQ) of thenew line were accomplished during the maintenance shut-down. However, Performance Qualification (PQ) and ProcessValidation (PV) activities on the new line had to be accom-plished, while aseptic manufacturing operations were occur-ring on the line that was not removed during the shutdown.PQ utilized the Manufacturing Execution System (MES)with the new equipment filling water. PV was accomplishedin aseptic conditions filling actual drug substance. Trouble-shooting and testing a line in an operational aseptic environ-ment proved to be extremely difficult. Testing activities couldnot disrupt production demands during this period. Opera-tional resources were challenged every day to continue pro-duction, while trying to learn and support the qualificationand validation of the new line. Refer to Table A for a break-down of commissioning, qualification, and validation activi-ties that were accomplished in each location.

ObjectiveDue to the criticality of the project, the filling line was firstinstalled in a former aseptic production plant located off-site.The objective was to reduce the potential for significant start-up delays by completing as much design, installation, start-up, and commissioning activity as possible in this off-sitetesting facility. The construction downtime, which involved amuch larger overall scope, was to last eight weeks. This wouldallow time for construction and additional commissioningand qualification activities in the actual facility during theshutdown period. Intensive static and dynamic airflow pat-tern testing also were executed in the actual facility to prove

that airflows throughout the actual production facility werenot adversely affected by the installation of the new line. Theproduction area and the remaining fill line, were then re-turned to service in an aseptic state. Once aseptic, media fillswere performed on the existing line and production beganagain on this existing line. The new line then progressedthrough a series of PQ tests, media fills, and PV. The schedulecalled for the new line to begin making marketable medicineapproximately eight weeks after the facility was returned toaseptic conditions. With a full maintenance shutdown takingplace at the same time as the construction activities, multiplecommissioning and qualification activities to qualify newequipment installations, airflow pattern testing, and thestart-up of the aseptic environment, it was critical that theproject develop and maintain one integrated schedule withproper sequencing and interdependencies.

Testing StrategyPlanningThe principles in the ISPE Baseline® Guide to Commission-ing and Qualification were followed in planning, testing, andverification activities for this renovation project. Anoverarching project validation plan was developed, whichprovided the outline for all commissioning, qualification, andvalidation strategies. Test plans were developed for systemsor groups of systems, which contained the detailed approachesto commissioning, qualification, and computer systems vali-dation. The test plans dictated that all requirements weretested and only those associated with higher risk, compliancerequirements, and could affect product quality, were testedduring qualification. The test plans also laid out which testswould occur in the offsite testing facility and which wouldoccur in the production facility. Each test plan was summa-rized following testing. For equipment that was tested in theoff-site facility, separate test summary reports were gener-

Figure 1. Test case cover sheet.




ated and approved—one for the off-site testing and one fortesting in the production facility.

Planning documents were approved by affected functionalgroups’ leadership with final approval from Quality Assur-ance. User Requirement documents were developed and theverification that the design met the requirements was accom-plished via Design Qualification (or Design Review for indi-rect impact systems). Planning, requirements definition, andverification of design adequacy took standard approaches.The remaining subsections discuss some unique aspects ofthis project.

Benefits of Off-Site TestingThe project team had a number of key challenges it needed toovercome in order to deliver the new fill line successfully. Theuse of an off-site testing facility was a critical piece inovercoming these challenges, and provided significant ben-efit to the facility.

The new fill line was being installed in an existing,validated, and fully operational aseptic filling operation.While there was some ability to build inventories of keyproducts to allow for an extended production outage, due tostrong demand for these products and capacity constraints inthe existing facility, an eight week shutdown was all thatcould be accommodated. With a significant construction scope,

including architectural modifications, installation of a newfilling line, installation of multiple aseptic filling hoods, andother items, it was necessary to limit the amount of start-upproblems to ensure the project could be completed in eightweeks. It also was important that once the facility had beenrestarted, issues that needed to be addressed on the new linecould not affect production on the sister line in the sameaseptic area. These issues drove the need for an off-sitetesting facility to ensure the unit was operating properly, wasvalidatable, and could be operated effectively by the opera-tions personnel once it was installed.

From a cost perspective, the use of the off-site testingfacility added very few additional costs to the project, butprovided significant savings. There was no additional cost forthe off-site facility since it was already owned and operatedby the organization installing the fill line. However, therewas some additional construction costs associated with themoving and final installation of the fill line since the systemwas fully assembled and operational in the test facility. Thiswas the only redundant activity for the engineering andconstruction portion of the project. Additionally, there wasvery little duplication of validation documentation from onesite to the other through the effective use of commissioningtesting, and the focusing of qualification efforts on key pro-cess requirements and not on engineering requirements. The

Figure 2. Dosing performance tracking tool.




use of test cases helped to greatly reduce the costs associatedwith these redundancies by allowing the team to minimizethe number of test cases needed by establishing a format thatallowed for both commissioning and qualification testing ofmultiple variables in one test case, by focusing qualificationtesting around critical process parameters, providing anefficient documentation process for commissioning testing,which allowed for trouble shooting and re-execution of testcases without revisions. The real savings for the overallproject came from a significantly reduced inventory buildprior to the outage, the ability to more efficiently identify andresolve problems in a non-aseptic environment, and theability to transition to a full production mode sooner, elimi-nating over-time, outsourcing of products, operational ineffi-ciencies, and the need for changes after completion of theproject thus delay final validation of the new filling line.These larger savings off-set the minor additional costs asso-ciated with the off-site testing facility and also made theproject possible.

The Use of Test Cases Instead of ProtocolsThe ISPE Baseline Guide® to Commissioning and Qualifica-tion defines commissioning as “a well planned, documented,and managed engineering approach to the start-up andturnover of facilities, systems, and equipment to the End-User that results in a safe and functional environment thatmeets established design requirements and stakeholder ex-pectations.” Performance testing is a key aspect of the com-missioning process. The typical way to document this testingis in a protocol which has a series of test cases and executioninstructions. This protocol can be a very lengthy document toensure that all of the system’s functional requirements andspecifications are tested. This can lead to an arduous ap-proval process as differences of opinion regarding one specifictest case can impede the approval of the entire protocol,resulting in schedule delays. Another potential shortcomingof this approach is that if a testing gap is discovered (i.e., arequirement was not tested) the testing protocol must berevised or addenda must be created depending on the processestablished by the project team.

To avoid some of these drawbacks, this facility upgradeproject implemented a process whereby test cases were writ-ten and approved individually rather than within the contextof a testing protocol. This sped up the document developmentprocess significantly. Another advantage of this process wasthat test cases could be run multiple times without thecreation of additional documentation. This was crucial to theproject given the diversity of products filled by the line. Thesame test case could be executed for each of the nine productrecipes and their associated vial/stopper combinations with-out assembling a testing protocol that would have beenhundreds of pages in length. Recipe, vial type, stopper type,etc. were documented by simply checking a series of boxes onthe first page of the test script - Figure 1. Figure 1 is the firstpage of a test case that was approved prior to execution andpost-execution reviewed and approved (not shown). The testcases aided the project team in determining which recipes

would be the most challenging from whatever performanceaspect the team wanted to examine. For example, the teamcould run the same test case on Recipe X and Y and analyzethe percentage of vials displaying missing stoppers by simplycomparing the test cases. The results were then examinedwithout the burden of generating new testing documents orwriting a summary report. The test cases also could be usedto document testing at the test facility and in the mainproduction facility, further reducing the amount of newdocumentation that would need to be created. Moreover, theuse of test cases facilitated assessing the impact of changeswithout new testing documentation. If a physical change wasmade to the system, test cases for the same recipe executedbefore and after the change could be directly compared. Re-execution of failed test cases or re-execution to verify perfor-mance after a physical change also was simplified throughthis process.

The use of test cases also facilitated the testing of theinline filtration system, which was the system implementedto significantly reduce the number of aseptic connectionsrequired by the process. This system pressurized a manufac-turing tank, allowing product to be filtered by a dual filtrationassembly into a smaller fill tank located in the fill room. Thecontrol system actively controls fill tank pressure. Besidesreducing the number of aseptic connections required, theinline filtration system was believed to provide the additionalbenefit of tighter dosing control, due to active pressure andlevel control on the smaller filling tank. To prove this theory,the standard filler commissioning test was used. A directcomparison of dosing results using the inline filtration sys-tem and its active pressure control could be made to previousfiller runs completed prior to installation of the new system.The performance testing results confirmed that the newsystem did indeed provide tighter dosing control for the filler.The use of test cases instead of protocols allowed the inlinefiltration system to be tested in parallel with testing of thefiller, thus reducing the time to effectively test the integratedsystem. Since the inline filtration system was not completeduntil several months into the testing period at the off-site testfacility, the time saved by parallel testing of the systems wascrucial.

Using the Off-Site Test FacilityThe owner of the final facility operates and maintains mul-tiple production and development facilities around the world.There are a number of legacy production facilities still inplace, and recently the owner had designated one facility inparticular for development activities. This facility was ap-proximately 30 miles from the final production facility. Theopportunity to conduct commissioning and qualification atan off-site test facility provided some distinct advantages forthe project team. First, being physically removed from theproduction facility allowed the project team to fully concen-trate on the task of commissioning the new filler without thedistractions of the day-to-day production routine. Using theoff-site testing facility gave the project team access to experi-enced personnel who had been involved in similar projects




and had operational knowledge that benefited the projectteam. The facility allowed for the replication of productionconditions for performance testing. Vials were washed andsterilized prior to filling operations so vial handling charac-teristics of clean, sterilized glass were analyzed (if desired,the equipment allowed for manually loading vials directlyonto the pre-fill accumulation table). Stoppers were auto-claved and prepared according to operational procedures soperformance of the stoppering system could be effectivelycompared to production rates. Unidirectional airflow hoodswere turned on so the performance of the check-weigh scaleswould be comparable to that in the Grade A productionenvironment.

To ensure that testing results from the testing facilitycould be considered valid after the equipment was moved intothe production facility, the project team took a detailedinventory of all equipment prior to packing for the move.Since the test-facility and final production facility were onlyabout 30 miles from one another, the team was able to havesignificant oversite and involvement in the movement of theequipment. The team witnessed the movers to ensure thatcare was taken as the equipment was loaded for movement.A commissioning receipt verification was conducted as theequipment arrived at the production facility, and installationchecks were done following equipment placement in thebuilding. These efforts mitigated risks associated with mov-ing the equipment and accepting testing conducted in twodifferent locations as equivalent. Project teams regularlymake such risk-based decisions when deciding to leveragingtesting conducted on a vendor’s factory during FAT.

Testing Beyond the RequirementsThe requirements documents developed for the project wereintentionally focused specifically on regulatory requirementsrather than operational requirements. While this strategygreatly reduced the amount of regulatory documentation andlimited the scope of qualification/validation testing, it left agap with the overall functional requirements which arenormally tied to commissioning activities and operationalexpectations of the system. The commissioning and qualifica-tion test cases were written to align with the User and

Functional Requirements associated with the filler. How-ever, things like the number of alarms, the number of asepticinterventions, and the percentage of vials automaticallyrejected by the control system, which are certainly importantfrom an operational and business perspective, were notspecified. The project team quickly realized that they could“meet established design requirements” while failing to meet“stakeholder expectations.”

The installation, start-up, and validation of this new linewere critical for the plant. The existing filling line was theonly line the company had validated to fill a number of keyproducts. Therefore, it was very important for the projectteam as well as the operations team to accurately understandhow well the machine would perform after installation. Toeffectively communicate the capability of the new filler, theproject team had to develop a tool to document the system’sperformance above and beyond what was requested in thetest scripts. The project team worked with the plant’s opera-tions team to develop acceptance criteria focused around theline’s overall performance. These criteria included missingstopper rates for each run, number of aseptic interventions,total yield for the fill, and others. The team developed aprocess that allowed commissioning test executors to effi-ciently track the issues encountered during the test run. Fourseparate tools were created—one for filler infeed, one fordosing performance, one for stoppering performance, and onefor tray loading performance. The dosing performance tool isshown in Figure 2 for illustrative purposes. This documenta-tion process allowed for the commissioning executors to testin parallel with the formal testing script. Results of commis-sioning were then attached to the test script so this additionalinformation was captured. After several test runs, thesedocuments enabled the team to quickly generate Paretocharts for various issues. With the Pareto charts, the teamwas then able to effectively focus on critical issues, solvethem, and then re-execute runs to ensure changes wereeffective. With a clear set of operational expectations, regula-tory requirements, documented solutions and their effective-ness, management was able to understand in detail the riskswith the new fill line from a performance perspective andmake a sound decision as to whether to move forward with theproject during this shutdown period or wait until the nextavailable shutdown period given the inventory demands ofthe facility.

Solving Technical ProblemsThe Pareto charts became the primary decision tool for theproject team to address technical and performance prob-lems witnessed at the off-site test facility. Using thesecharts, the team addressed several problems in a relativelyshort period of time. Examining the Infeed Pareto chart(Figure 3) illustrates the approach the team took to solvingtechnical issues. A quick glance at the Pareto shows that thetwo most occurring alarms and aseptic interventions were“No vials at infeed starwheel” and “Remove downed bottle.”Through testing experience, the C&Q team knew that thisin fact was one in the same problem. When a “No vials at

Figure 3. Infeed Pareto chart.




infeed starwheel” alarm was received, this was usually dueto a downed bottle jamming the flow of vials from the pre-fillaccumulation table to the filler infeed starwheel. When thisoccurred, operator intervention was required to remove thedowned vial. The infeed transition did have a downed vialescapement, but the jam was often created because the vialsfell downstream of the escapement.

Again from the testing, the C&Q team also knew that thisproblem was primarily related to the 10 ml vial. Qualitativedata demonstrated that more than 50 percent of the occur-rences of “No vials at infeed starwheel” were associated withtesting runs using the 10 ml vial. This vial is tall and has arelatively high center of gravity. If there was any interruptionin the flow of vials through the infeed transition, the vialsupstream of the interruption had a tendency to wobble andfall. The project engineers designed a vial neck guide at thevial escapement that prevented vials from tipping over. Thisguide was installed over the infeed vial guide and grabbedvials so that if there was motion due to a break in the streamof vials, the vials would not fall over. Using the test trackingtools, the team documented a significant reduction in thenumber of “No vials at infeed starwheel” and “Remove downedbottle” alarms following the implementation of this physicalchange to the system. Similar troubleshooting during thetesting phase at the off-site test facility, as well as changemanagement tracking using the repeatable test case execu-tion, also were employed.

Managing ComponentsProcurement is a functional area involved with any capitalproject. One generally assumes that this function will acquirethe components of the process systems. While this was truefor this facility upgrade project, the team also had to managetesting components – specifically coordinating the acquisi-tion and use of vials and stoppers for testing. Since themajority of the testing took place at an off-site test facility, theproject team had to interface with the procurement group toensure testing supplies were available. The procurementteam then had to balance the requests of the project teamwith the needs of the manufacturing facility, which wasramping up production to meet inventory needs prior to theupcoming shutdown. Clearly, procurement had to defer toproduction when there were conflicts with a given vial orstopper.

A logical solution to this situation would simply be to ordermore vials and stoppers from the vendors. However, the leadtime for these items is quite significant – on the scale ofmonths, not weeks. While this was an option, it was not apanacea for the project team and the tight demands of theschedule. Planning for testing needs relatively far in advancebecame a necessity for the project, not a luxury. With onlyfour focused months of testing at the test facility and twomonths of construction and C&Q testing in the new facility,precise planning was a necessity and there was not muchtime to deal with procurement issues.

With the wide variety of vials and stoppers in use onexisting production lines, other performance issues had been

observed and were being addressed by other project groups.These groups often wanted to conduct studies on how poten-tial replacement container closure systems would functionwith the new filler, even though these would not immediatelybe employed in production following the shutdown period.Despite the pressing production needs following the shut-down, the commissioning team had to examine the big pic-ture and allow potential improved vials and stoppers to berun on the new machine. This enabled the characteristics ofthese components to be examined without production down-time, which would be required if these studies were con-ducted once the filler was placed in the manufacturing facil-ity.

Lessons LearnedExecuting the Commissioning TestsSeparating business needs from regulatory/quality require-ments provided a number of key benefits for the project. Itgreatly reduced the amount of regulatory documentationwithout reducing the effectiveness of the regulatory effort.The definition of the operational requirements late in theproject proved to be a major obstacle, and added significantinefficiencies to the project early on. While the team was ableto recover from this deficiency, as a whole, integration ofoperational requirements into the C&Q effort can be donewithout reducing the regulatory optimization effort, andensures a more operable, efficient, and maintainable systemin the future. Another objective of the project was to providebetter integration of the mechanical qualification and com-puter system validation effort. While progress was made inthis area, there was clearly more that could have been done.

The commissioning test cases were generally executed byoperators familiar with the existing filler and procedures.This use of experienced production staff eased the technologyand knowledge transfer process that occurred at the end ofthe project. The knowledge operators brought to the commis-sioning team cannot be overstated. Their expertise in theareas of aseptic technique and documenting aseptic interven-tions was essential for developing robust testing data anddefining operational requirements that were not part of theregulatory effort.

Human ResourcesAlthough there were operators who supported the commis-sioning phase of the project, a full fill team could not be sparedby the production facility to support full-time commissioningtesting given the pre-shutdown production schedule. There-fore, the commissioning contractors often had to assume therole of operators. This required the commissioning contrac-tors to quickly learn the functionality of the machine whichshould be expected of any professional commissioning teamwhile maintaining their objectivity during the testing pro-cess. The ability of the commissioning team to run theequipment enhanced troubleshooting efforts and the resolu-tion of operational issues.

Other critical resources for the project included Equip-ment Engineers and Maintenance Mechanics with a very




solid background in parenteral equipment, maintenance,filling processes, and equipment troubleshooting. Automa-tion engineers with the ability to cross the boundary betweenprocess/equipment issues and instrumentation and codingwere invaluable toward ensuring the process and the systemworked in harmony. This group was very important duringthe troubleshooting phase of the project. Their understand-ing of the machines, ability to perform effective root causeanalysis, and identify effective solutions quickly led to sig-nificant improvements in performance over very short peri-ods of time. Lastly, a responsive Quality organization maderapid turnaround on testing documentation possible.

Coordination with End UserIntegration of a new filling line into an existing facilitycannot be done without the commitment and support of thefuture system owners. Their understanding of the day-to-daybusiness of making medicine in the particular facility wherethis system was installed is invaluable. They are the bestequipped to articulate issues, provide clear direction aroundpriorities. They also understand the operational and regula-tory limitations and improvements of the system, and theireffect on the overall operability of the area.

About the AuthorsMichael DiGiovanni is a Project Engineerfor Eli Lilly & Company. He was responsiblefor the delivery of the filling line described inthis article. He has previous experience as aUtilities and HVAC engineering team lead.He obtained a BA in economics in 1992 fromthe University of Pittsburgh and a BSc inmechanical engineering from the University

of Pittsburgh. He is a member of ISPE. He can be contactedby telephone: +1-317-651-9147 or by e-mail: [email protected].

Eli Lilly & Co., Lilly Corporate Center, Indianapolis,Indiana 46285, USA.

Keith Weseli is a Project Manager for Com-missioning Agents, Inc. He earned a BS inmathematics from Vanderbilt University in1997 and a MBA from the University ofRhode Island in 2005. He has been certifiedas a Nuclear Engineering Officer by the USNavy and served as a submarine officer forseven years. He has been involved with the

commissioning, qualification, and validation of manufactur-ing equipment associated with several capital projects forlarge pharmaceutical and biotech companies and is a mem-ber of the ISPE Commissioning and Qualification Commu-nity of Practice. He can be contacted by telephone: +1-317-289-2804 or by e-mail: [email protected].

Commissioning Agents, Inc., 1515 N. Girls School Rd.,Indianapolis, Indiana 46214, USA.


Lean Manufacturing Techniques


This articlepresents theapplication ofindustrialengineering andlean techniquesto a contractpharmaceuticalmanufacturingfacility.

Integrating Industrial Engineering andLean Techniques at a ContractPharmaceutical Manufacturer

by Valerie Maier-Speredelozzi, Cyrus Agarabi,Thomas Needham, and Sirine A. Saleem

Figure 1. Simulationoverview.15

Introduction

Current Good Manufacturing Practices(cGMPs) were developed to ensure qual-ity pharmaceutical products and keepthe public safe. The regulatory ap-

proach by the US Food and Drug Administra-tion (FDA), combined with an environmentthat did not encourage manufacturing innova-tion, resulted in a pharmaceutical industrythat did not keep pace with technological evo-lution, and ultimately had a restrictive effecton daily operations and process improvementsfor pharmaceutical manufacturers. Companieswere apprehensive to be the first to initiatemajor changes in their production environ-ments, without knowing how regulators wouldrespond. In response, the FDA released twodocuments to encourage more innovation inpharmaceutical manufacturing: Guidances forIndustry: PAT – A Framework for InnovativePharmaceutical Development, Manufacturing,and Quality Assurance1 and PharmaceuticalcGMPs for the 21st Century-A Risk-Based Ap-proach.2 These guidances have excited many inthe pharmaceutical industry who realize thepotential to continuously improve processes,as occurs in other manufacturing industries.Consequently, the focus changed from producttesting and release to understanding the prod-uct, the manufacturing process, and opera-tions.3,4

Other industries have successfully devel-oped innovative approaches to continuouslyimprove and remain competitive, so it is impor-tant to learn from their successes and failures.A facility producing pharmaceutical productsunder cGMPs was evaluated to gain a “baseline”understanding of current manufacturing prac-tices. Critical and problematic areas were iden-tified as well as potential opportunities to in-corporate external industry practices to im-prove the manufacturing process with a focuson the Toyota Production System®.

There is considerable opportunity to investi-gate the implementation of current practicesand knowledge found outside the pharmaceuti-cal industry for incorporation into novel pro-cesses in line with the FDA’s cGMP regulationsand PAT guidance.5 The concept of incorporat-ing external industries’ practices has been re-cently proposed.6 However, research focused onthese principles and their effects on manufac-turing operations and pharmaceutical productdevelopment has not been explored. It is piv-otal to approach improvements to pharmaceu-tical manufacturing processes, while still com-plying with cGMP and FDA regulations, as wellas other regulatory agencies around the globe.

BackgroundThis section provides an overview of current

Reprinted from







practices and tools from outside the pharmaceutical industrythat were investigated for a contract pharmaceutical com-pany (XYZ Pharma) to improve its current manufacturingsystem. An extensive overview of these principles and toolscan be found in books on lean manufacturing.7,8,9,10

Lean Manufacturing and its ToolsToyota was a leading developer in the use of lean manufactur-ing, which has been widely adopted with applicability to anyindustry as the “Toyota Production System®.” Production isconsidered lean if it is accomplished with minimal waste,11

and if it utilizes far fewer resources, such as worker effort,production, storage space, or equipment investments whilestriving to achieve defect free processing.12 Taiichi Ohno ofToyota®7 identified seven forms of waste: defects, waiting,motion, over-processing, over-production, inventory, and in-efficiency. Several lean techniques can be used in the pursuitof achieving zero waste such as the following:

• Autonomation (or Jidoka) has two distinct meanings: 1) Achange from a manual process to a machine process. 2)Automatic control of defects or automation with a humanmind.8

• Error-Proofing: This technique places various checkingdevices on equipment and tools to remove the potential forerror and ultimately the creation of defects.

• Just-In-Time (JIT) systems: This technique requires aholistic approach to ensure accurate production, ordering,and stock quantities to ensure that the right parts neededin an assembly are available at the exact time they areneeded and only in the amount needed.7

• Kanban (sign board): This technique is a communicationtool to convey information about picking up or receivingthe production order.7

• Kaizen (good change): This technique is defined as con-tinuous improvement.

• “5S:” This technique is used during a Kaizen event toreduce hidden wastes in the plant through a cleanupactivity.

• System Configurations: This technique is essential toidentify a system’s bottlenecks and focus on improvingthem in order to realize the full potential of the system.

These various lean tools were each considered for possibleapplication at XYZ Pharma, following the preliminary appli-cation of another lean tool, Value Stream Mapping.

Value Stream Mapping® (VSM)The first step in achieving lean manufacturing is to under-stand the current system by applying VSM. This processinvolves recording all activities involved with manufacturinga product from raw materials to finished goods. A VSM iscreated to evaluate total efficiency instead of individualefficiencies. The map is comprised of three basic elements: 1.Product flow, 2. Information flow, and 3. Material flow.9

Furthermore, it includes two classes of work: work that addsvalue to the product and non-value added work (waste). By

addressing both types of work, a strategy can be devised toimplement lean tools to minimize waste and create a longterm vision. Eventually, this can be shown through a futurestate map which would employ the improvements and theirestimated effects on the system.9 It should be noted that tocreate an accurate model of the three flows in the map, theremust be accurate data collected from the production floor.This is accomplished by using “Gemba” which is defined asthe “actual place” and is the first step used by Toyota® whensolving a problem.10 The concept reinforces the need forfirsthand knowledge and challenges conventional manage-ment methods of system reports and computer analysis.

The created VSM might be the foundation of a simulationmodel that enables the monitoring of Work In Progress(WIP), production lead time, cycle time, changeover time,efficiency, etc.13 As depicted in Figure 1, simulation will allowvisualization and incorporate detailed information about thesystem while closely conforming to the individual aspects.14

This will facilitate experimentation with theoretical sce-narios to identify problematic areas and potential failures.

Additionally, using other methods such as Quality byDesign (QbD) and Design of Experiments (DOE) may help indetermining the critical factors and interactions identifiedfrom a VSM or simulation model results that achieved statis-tical significance. Such significant factors will become thefocus of the future state model.

Company and Process BackgroundThe pharmaceutical company in this study is referred to asXYZ Pharma and primarily serves as an OTC contract manu-facturer for more than 100 products. They are housed in a25,000 sq.ft. production facility with a detached 75,000 sq.ft.warehouse building nearby. There are approximately 55 per-manent and temporary employees working three overlappingshifts (5 am to 2 pm, 7:30 am to 4 pm, and 2 pm to 10:30 pm).There are between 20 and 30 operators during the shifts to runthe filling and packaging machines. Only a select group ofoperators have been trained on the newest filling machine, andtheir operating times are limited to the second shift. The 12person Quality Control (QC) department functions to ensurethe quality of all incoming shipments, all cGMP requirementsduring production, and the quality specifications of all out-bound products, including maintaining documentation and a“quarantine” area. One driver is responsible for operating XYZPharma’s truck, which travels between the main warehousebuilding and the production facility approximately every twohours with raw materials, finished goods and recyclables.

This study will evaluate the manufacturing of a poloxamerbased topical gel, referred to as Product X, produced in three30,000 tube batch sizes for one 90,000 tube job. This productis similar in formulation and manufacturing processes (Fig-ure 2) to many other products manufactured at XYZ Pharmaand is representative of the system. Raw materials undergoa 23 step proprietary process to formulate the bulk drug.During production, the QC inspector periodically checks onthe manufacturing specialist to verify his work. The formula-tion process is completed when the product has been entirely




transferred into a stainless steel storage vat through amilling process. A mechanic is responsible for the setup of thefilling equipment and must prime the machine with productand make adjustments to ensure accurate filling withinspecifications. The company currently has two different typesof tube filling machines: three older machines and one new,which have both been used to manufacture Product X. Themajority of employees are trained on the older machines,which can be run on all three shifts with only minor adjust-ments after the initial setup. The new system is only run bya select number of operators and generally runs during the7:30 am to 4:00 pm second shift. The older machines arecapable of producing 35 tubes per minute of Product X usingtwo to three operators, while the new system produces 70tubes per minute using between two and four operators. If a

filled tube passes the inspection criteria, it is put into a plasticstorage tote on a pallet. Each pallet holds up to 20 plastictotes, and the company owns 52 totes. If the totes are filledbefore the next process (cartoning) is running, then operatorsmust assemble boxes and fill those instead of totes. A me-chanic must setup and adjust the cartoner to accommodatethe size and feed rate of the cartons, and the subsequent tapemachine. Cartoning occurs at a load rate of 78 to 82 cartonsper minute and can operate through all three shifts with sixto eight operators. Finished goods are placed in the produc-tion holding area under quarantine to await final assayresults and reconciliation of quantities. The product may bemoved via an internal shipping truck to the warehousefacility to await final release of the lot from quarantine toallow shipping to the distributor.

Figure 3. Current state VSM.

Figure 2. Overview of manufacturing of Product X.




Figure 4. Current state simulation layout.

Current State VSMA current state Value Stream Map (VSM) of the system forProduct X at XYZ Pharma is shown in Figure 3. Data for thisVSM was collected through time studies and retrieved fromthe master batch records of four recent production runs ofProduct X. The suppliers are located at the top left, and theXYZ Pharma customers at the top right. Information flowsacross the top of the diagram from right to left, and downwardin the form of weekly schedules for each department. Mate-rials and product flow across the diagram from left to rightthrough the various processing operations. Each major op-eration has a “process box” underneath with metrics that helpto quantify the time and resources required for that produc-tion stage. Uptime percentage was calculated by dividing thevalue added time (of the batch) by total production time, bothvalue and non-value added. Non-value added time for a singleoperation was comprised of breaks, lunch, machine break-downs, and setup/change over times. Along the bottom of theVSM, a “ladder diagram” shows value adding processing time(high points) and non-value adding time (low points) of theentire production process, which includes batching delays,time spent in inventory, which is represented by triangles onthe diagram, waiting, or other wastes. The percent of valueadding time for the entire process is 5.9% which is actuallyvery typical for many “current state” non-lean facilities.

The current state VSM gives the Cycle Time (CT) of anindividual tube through processing, except during formula-tion where the entire batch is undergoing value added workat the same time. The map shows that the greatest wastes oftime are the buildup of WIP in the system before the fillingline and after the cartoning process, where finished goodspallets are stored until the entire batch is complete and readyfor shipment. The lead time of the system was calculated tobe 7925 min and 50 sec. When divided by the total amount ofoperational hours from 5:00 am to 10:30 pm (17.5 hours), thelead time is roughly 7.5 days. The current state is based on afive day work week with Saturday and Sunday as days off.

This translates into a “real world” lead time of 9.5 calendardays for one batch. One order for this product generallyconsists of three batches requiring minimal inter-batchchanges once the initial set-ups have occurred, resulting indecreased holding times to avoid starving down processmachines. Therefore, the second and third batches are pro-duced faster, and the entire three batch process requiresapproximately 11 days of processing or 15 calendar days.

Simulation for the Current VSMThe simulation utilizes collected data and compiled masterbatch records to create a model which depicts the currentstate of the system. All processing times, waiting times, andpersonnel assignments have been created through observa-tion and verified through company documentation. A portionof the facility layout and the simulation locations which havebeen built are shown in Figure 4. A Computer Aided Drafting(CAD) model of the facility layout, including accurate dis-tances between departments, was developed. This layout wasthen imported into the commercial simulation software pack-age as a backdrop for the simulation model. Travel times foremployees moving product between processing operationswere entered into the model, and a visual representation ofthe movement of product through the facility was available asthe simulation model ran. The flow of materials follows thecurrent VSM in Figure 3 as discussed earlier.

The results of the current state simulation help to gain anunderstanding of the steady state of the system as shown inTable A. The simulation runs overnight and on the weekendseven when the virtual equipment is not operating to realisti-cally represent the current state in the facility, which does notoperate on the weekends. Thus, a batch that is in productionon Friday may not be filled and cartoned until Monday,increasing the lead time. The simulation was run for 999replications, the maximum allowed by the commercial soft-ware package, and the calculated average throughput timethrough the system was found to be 370.92 hours or 15.45




days. In order to ensure that the simulation adequatelyrepresented real world conditions, the average time perbatch, 123.88 hours, is multiplied by three to give a totalorder time of 371.65 hours. The average time from thesimulations was analyzed to be 99.8% similar to that from themaster batch records.

The largest periods of non-value added time are when theoperation waits to batch. The time that formulated bulk drugproduct spends in the transfer vat waiting to be brought to thefilling line is very long. Also the time spent waiting for thebatch of wrapped and palletized finished goods is considerable.Due to the infinite capacity of the production warehouse areain the model, the utilization percent is not calculated. This waspurposely avoided to prevent blockage of incoming, outgoing,and stored materials, which would require a separate in-depthmaterial handling study out of this project’s scope.

Furthermore, Table A shows that the two highest utiliza-tion percentages of equipment or inventory transfer andholding locations are the vat holding area and the new fillingmachine, respectively. These areas appear to be bottlenecksin the current system and should be viewed as areas forimprovement for the future state by introducing parallelmachining capabilities. This is possible because the companyhad actually replaced a slower old filling machine with afaster new filling machine. Both machines were alreadyvalidated, so utilizing them simultaneously to fill one batchof Product X in less time would be acceptable within FDAguidelines and would not require a new investment. Whilethe production warehouse area does not have a calculatedutilization percent, the average time per entry is the secondhighest and also should be considered as a potential area forimprovement.

Areas for ImprovementsAfter developing the current VSM and running the simula-tion, critical and problematic areas were identified as well aspotential opportunities to incorporate lean tools to improvethe manufacturing process. These areas for improvementsinclude:

• shrinking traditional batch sizes to create a more semi-continuous production system

• improving efficiency of “milk run” truck deliveries• reducing inventories by decreasing storage vessel batch

sizes and decreasing WIP buffers• cleaning and organizing facilities by applying the 5S

techniques• improving equipment by using visual controls and Single

Minute Exchange of Dies techniques to decrease setuptimes

• incorporating automation that would result in improvedproduct quality and fewer operators needed on machines

• adding parallel machines at the equipment bottleneck• cross training personnel so that most operators would be

trained on multiple pieces of equipment, allowing them tooperate during any shift and rotate when needed

The Future State SimulationTo develop the future state model, a Design of Experiments(DOE) methodology was used to systematically vary the stateof certain factors or operational parameters in the simula-tion. After analyzing the areas for improvement, three criti-cal factors were identified:

Factor A – parallel manufacturing by reinstating use of an oldfilling machineFactor B – cross training of personnelFactor C – changes in move batching rules during production

Some of the other lean improvements for the facility thatwere considered above were evaluated and suggested to thecompany, but the implications and results of these changeswould be more difficult to model in a simulation environment.For the three selected factors, the simulation model could bealtered by adding machine resources, changing the rules thatgovern which virtual operators can operate which equip-ment, and changing the size of move batches. It is importantto note that the validated batch size of Product X, constitutinga full production run on a master batch record for FDAdocumentation purposes, has not been changed. What ischanged with Factor C is that the quantities of work-in-process that can be moved between operations are reduced.For instance, the first tube of Product X that is filled no longerneeds to wait for the last tube in the batch to be completed

FactorsScenario # A B C1 0 0 02 0 0 13 0 1 04 0 1 15 1 0 06 1 0 17 1 1 08 1 1 1

Table B. 23 factorial design of three factors selected forexperimentation.

Location Name Avg Time per % UtilizationEntry (MIN)

Inventory Transfer and HoldingProduction Warehouse Area 3081.99 0Formulation Holding Area 934.16 12.79Storage Vat Load 977.77 13.39Vat Holding Area 7097.92 31.64New Filling Machine WIP Holding 54.86 0.64Cartoning Machine WIP Feed 353.97 4.06Cartoning Machine FGI 203.02 14.32EquipmentKettle 1397.93 19.14New Filling Machine 1816.03 24.58Cartoning Machine 142.85 16.49Pallet Wrapper 27.14 2.26Total System Throughput Time 370.92 hrs 15.45 days

Table A. Results of the current state simulation.




Location Name 1 2 3 4 5 6 7 8Inventory Transfer and HoldingProduction Warehouse Area 0 0 0 0 0 0 0 0Formulation Holding Area 13 42 17 69 12 36 16 54Storage Vat Load 13 67 18 30 12 16 17 17Vat Holding Area 32 0 26 0 36 0 29 0New Filling Machine WIP Holding 1 0 1 0 3 0 4 0Cartoning Machine WIP feed 4 0 7 1 12 5 12 7Cartoning Machine FGI 14 32 13 25 14 37 12 40EquipmentKettle 19 86 25 91 18 52 25 74Hopper 0 70 0 28 16 13 13 23New Filling Machine 25 74 18 36 25 45 19 26Old Filling Machine 0 0 0 0 23 37 19 26Cartoning Machine 16 26 17 25 18 37 18 31Pallet Wrapper 2 3 3 7 2 8 4 45

Table E. Utilization (%) of locations.

before it is moved to the cartoning operation.Subsequently, a 23 factorial design of the critical three

factors was selected for evaluation in a simulation as shownin Table B where 0 refers to the current state and 1 refers tothe suggested improvement.

The impact of the three critical factors on the averagethroughput times and average time per entry is shown inTable C and Table D, respectively. Table C summarizes theaverage throughput times of 999 simulation replications foreach of the eight scenarios. Scenarios 4 and 8 show thegreatest improvement over the original current state through-put time (Scenario 1). Scenario 8 is a future state with allthree factors changed, which is comprised of parallel machin-ing, cross training, and changing the batch rules. Scenario 4is the same as scenario 8, except it does not use parallelmachining methods.

Table D compares all of the simulation scenarios and

shows the average time per entry, which gives the averagetime (in minutes) spent at a given location by each componenttraveling through the system. This is helpful in evaluatingthe effects of the significant rule changes in the variousscenarios on processing and holding times.

The changing of the batching rules to a more continuousapproach resulted in considerable time savings. For example,the average time per entry (min) that a box of finished goodsinventory spends waiting after completion of operations atthe cartoning machine using Scenario 4 was approximately50% of the time as the current state, Scenario 1.

The utilization (%) of locations resulting from the simula-tions is given in Table E. This is useful to identify possiblenew bottlenecks created in the system after changes havebeen made. Many manufacturers strive to reach high utiliza-tion rates for expensive machinery and research has shownthat this can result in a large buildup of WIP in front of the

Location Name 1 2 3 4 5 6 7 8Inventory Transfer and HoldingProduction Warehouse Area 3082 598 2242 164 3061 629 2198 136Formulation Holding Area 934 1584 943 1458 939 1076 939 1140Storage Vat Load 978 631 986 154 979 119 980 87Vat Holding Area 7098 0 4537 0 8912 0 5021 0New Filling Machine WIP Holding 55 0 97 0 304 33 235 1Cartoning Machine WIP Feed 354 11 493 33 1127 207 785 171Cartoning Machine FGI 203 226 132 99 222 235 126 171EquipmentKettle 1398 3266 1411 1909 1404 1560 1404 1576Hopper 0 658 0 147 1262 385 734 112New Filling Machine 1816 33 1021 9 46 31 22 9Old Filling Machine 0 0 0 0 44 27 32 20Cartoning Machine 143 117 112 63 172 142 118 75Pallet Wrapper 27 22 31 25 26 40 41 169

Table D. Average time per entry (minutes) of all simulation scenarios.

Scenario 1 2 3 4 5 6 7 8Avg. Throughput (Hr) 371 189 284 105 406 161 290 110

Table C. Rounded average throughput times (Hr) of all simulation scenarios.




machine to avoid starvation.16 This buildup of WIP hasnegative effects on the system as can be seen from thebatching used at XYZ Pharma. It is interesting to note thatincreasing the utilization of holding and transfer areas is notthe goal of lean manufacturing and implies that raw materi-als or WIP is occupying holding areas, which adds to non-value added time.

Scenario 4 provides the lowest throughput time (Hr) out ofthe eight tested scenarios. Following the continuous im-provement philosophy, the next phase of improvements wouldaddress the new locations which have subsequently becomebottlenecks in the system. The addition of cross training anda change in the batching rules has shifted the bottleneck tothe formulation step. The utilization of the kettle has in-creased from 19% to 91% between Scenarios 1 and 4, whichstrongly suggests further improvements to setups, cleaningand removal of other non-value added operations. If minimi-zation of non-value added time does not relieve the formula-tion bottleneck, parallel machining should be investigated.Within the inventory transfer and holding areas, the utiliza-tion of the formulation holding area experiences a largeincrease from 13% in the current state (Scenario 1) to 69% inthe future state (Scenario 4). Further improvements wouldconsist of a more precise JIT system, which would bring the

correct amount of materials for formulation at the time thatthe manufacturing operator requires them. Scenario 4 im-proves upon the current system by spreading out the arrivalof the raw materials, but inventory is still held in the formu-lation holding area. If this JIT system was instated, thecurrent formulation holding area could be converted into aprocessing area or could be used for other purposes.

Statistical Analysis of ResultsTo determine whether the three critical factors and theirinteractions are significant, a full factorial design was ana-lyzed. The coefficient of determination, or adjusted R2 value,was calculated to be 0.89, indicating that 89% of variabilityin the data can be accounted for by the model. Also, thecalculated P value is less than 0.05; so therefore, the hypoth-esis that this model is adequate has less than a 5% chance ofbeing rejected.

The Analysis of Variance (ANOVA) statistical results inTable F show that all three main factors, parallel manufac-turing, cross training personnel, and a change in batching aresignificant. The interaction between each of these factors isalso examined by this statistical method. The interactionbetween parallel manufacturing and changing batching rulesis significant, as is the interaction between cross training

Figure 5. Future State Value Stream Map based on Scenario 4.




Source Nparm DF Sum of Squares F Ratio Prob > FParallel MFG 1 1 40767 28.7772 <0.0001Cross Train 1 1 14217695 10036.16 0.0000Batch 1 1 77386070 54626.21 0.0000Parallel MFG*Cross Train 1 1 1147 0.8099 0.3682Parallel MFG*Batch 1 1 514934 363.4881 <0.0001Cross Train*Batch 1 1 582003 410.8311 <0.0001

Table F. ANOVA table results of the three critical factors.

personnel and changing the batching rules. The interactionbetween parallel manufacturing and cross training is not asignificant interaction. The nonsignificant interaction is mostlikely due to the overpowering effect of changing the batchingrules.

Future State VSMThe envisioned future state VSM was created as illustratedin Figure 5 based on results found from Scenario 4. Thisscenario yields the lowest simulation throughput time, whichis the primary goal for a contract manufacturer. It is wellestablished that in contract manufacturing, overproductionis not a concern because XYZ Pharma only produces whattheir customers have ordered. This scenario utilizes crosstraining and change in batching rules to decrease lead times.One difference from the current state with regard to crosstraining is the increase in shifts that are available for the newfilling machine to operate, due to a greater number of opera-tors able to run the machine. The most significant factor thatwas found through the statistical analysis was the batchingrule changes. These changes affect the entire process fromarrival of raw materials to departure of finished goods inven-tories. Arrival of raw ingredients are limited to quantitiesrequired for formulation of bulk drug at that time and arestored directly outside of the manufacturing areas. Thereplacement of the transfer vat with a smaller, more flexible,and mobile drum has decreased waiting times during trans-fers and setups, subsequently decreasing the cycle time byapproximately 90 minutes. Another benefit of earlier bulkdrug substance arrivals are the completion of final adjust-ments to the filling equipment sooner. Waiting times alsohave been decreased because WIP is no longer waiting tobatch prior to movement. Bulk drug in drums, WIP totes, andfinished goods pallets are all moved individually, and requirefewer quantities to be moved. The cumulative effects of thesechanges result in a lead time of approximately 2130 minutes,which is about a 75% reduction of time. It is important to notethat this future state value stream map indicates the time toproduce the first pallet of finished goods inventory. In thefuture state, the company may not have to batch the finishedgoods prior to shipping. Therefore, this estimate is useful todetermine how quickly finished goods would be ready to beginshipping if there was flexibility with the customers to receivegoods in more frequent smaller delivery amounts, while stillcomplying with regulations, and if it did not increase trans-portation costs. The classic Economic Order Quantity modelcontinues to apply with a trade-off between holding inven-tory, while large batches are completed versus paying order

set-up costs for more frequent, smaller batches. For FDAvalidation purposes, the defined batch size has not changed,but movement within the facility is allowed in smaller trans-fer vessels and in smaller quantities of tubes and cartons.

Recommendations were made to the company, PharmaXYZ, based on the results of the Value Stream Mapping andsimulation activities. The company, which is too small toemploy industrial or manufacturing engineers of their own,benefited from seeing models of both the current state and anenvisioned future state of their operations, which employsvarious potential lean techniques. The project was completedas part of a graduate student thesis project, and a teamproject for a class, and, had no cost to the company. To date,they have not specifically implemented the described recom-mended changes, as their focus is on daily production andregulatory compliance, as opposed to process improvement.

ConclusionsThe focus on product quality is extremely high in the pharma-ceutical industry to avoid potentially fatal and costly defects.With the advent of pharmaceutical quality systems, theindustry is moving away from end product testing and towardin process testing, which has been used for many years byother industries. Some pharmaceutical production facilitieshave a “Job Shop” layout, where all formulation equipment isgrouped near each other and products are transported inlarge vats for filling and on palletized totes of tubes for otherwork in process. A redesigned layout would place all of theequipment needed for a particular product in close proximityto each other in order to achieve a more continuous flowthrough the facility.

Lean manufacturing techniques should be explored in thepharmaceutical industry to improve current systems. Leantechniques also should be utilized early in the development ofnew systems. This case study represents a number of possibleopportunities for specific areas of improvement as well assuggesting an overall change in the manufacturing mindset.The pharmaceutical industry can learn a great deal fromoutside industries, such as using industrial engineering andlean techniques to enhance competitiveness and thereby helpto ensure solvency.

References1. Food and Drug Administration, “Guidance for Industry:

PAT-A Framework for Innovative Pharmaceutical Devel-opment, Manufacturing, and Quality Assurance.” http://www.fda.gov/cder/OPS/PAT.htm accessed on 12 October2006, September 2004a.




2. Food and Drug Administration, “Pharmaceutical cGMPsfor the 21st Century-A Risk-Based Approach,” http://www.fda.gov/cder/gmp/gmp2004/GMP_finalreport2004.htm accessed on 12 October 2006, September 2004b.

3. Langberg, H.C., Ruud, H. J., “Implementing PAT Step ByStep as a Process Optimization Tool,” PharmaceuticalEngineering, Vol. 25, No. 3, 2005.

4. Van Liedekerke, B., Maes, I., “Pharmaceutical Manufac-turing: Linking Vision and Decision-Making to Achieve aRoadmap Toward cGMPs for the 21st Century,” Pharma-ceutical Engineering, Vol. 27, No. 4, 2007.

5. Ganguly, J., Vogel, G., “Process Analytical Technology(PAT) and Scalable Automation for Bioprocess Controland Monitoring – A Case Study,” Pharmaceutical Engi-neering, Vol. 26, No. 1, 2006.

6. Crosby, T., “Designing For the Future of ContinuousProcessing,” Pharmaceutical Processing Online, Jan. 1,2006, http://www.pharmpro.com/.

7. Ohno, T., “Toyota Production System: Beyond Large-Scale Production,” Productivity Press, 1988.

8. Monden, Y., “Toyota Production System: An IntegratedApproach to Just-In-Time,” Norcross, Georgia: Instituteof Industrial Engineers, Third edition. 1998.

9. Allen, J., Robinson, C., Stewart, D., “Lean Manufactur-ing: A Plant Floor Guide,” Dearborn, Michigan: Society ofManufacturing Engineers, 2001.

10. Liker, J., “The Toyota Way: 14 Management Principlesfrom the World’s Greatest Manufacturer,” McGraw HillPublishing, 2004.

11. Narasimhan, R., Swink, M., Kim, S., “Disentangling Lean-ness and Agility: An Empirical Investigation,” Journal ofOperations Management, Vol. 24, 2006, pp. 440-457.

12. Womack, J., Jones, D., Roos, D., “The Machine ThatChanged the World,” Harper Perennial Publishing, 1990.

13. Abdulamek, F., Rajgopal, J., “Analyzing The Benefits ofLean Manufacturing and Value Stream Mapping ViaSimulation: A Process Sector Case Study,” InternationalJournal of Production Economics, Vol. 107, 2007, pp. 223-236.

14. Askin, R., Standridge, C., “Modeling and Analysis of Manu-facturing Systems,” John Wiley and Sons, Inc., 1993.

15. Sanchez, S., Moeeni, F., Sanchez, P., “So Many Factors, SoLittle Time... Simulation Experiments in the FrequencyDomain,” International Journal of Production Econom-ics, Vol. 103, 2006, pp. 149-165.

16. Li, N., Zhang, M., Deng, S., Lee, Z., Zhang, L., Zheng, L.,“Single-Station Performance Evaluation and Improve-ment in Semiconductor Manufacturing: A Graphical Ap-proach,” International Journal of Production Economics,Vol. 107, No. 2, 2007, pp. 397-403.

About the AuthorsDr. Valerie Maier-Speredelozzi is an As-sistant Professor in the Industrial and Sys-tems Engineering Department at the Uni-versity of Rhode Island. She received herPhD in 2003, and Masters degrees in bothmechanical engineering and industrial andoperations engineering in 2001, all from theUniversity of Michigan. Her Bachelor of

Mechanical Engineering degree is from Georgia Institute ofTechnology. She has worked in industry during undergradu-ate and graduate school with Georgia Power Company, VolvoPenta of the Americas, and General Motors, and completednumerous other projects with industry. Her research inter-ests include flexibility metrics, lean systems, and manufac-turing system design across multiple industries. She hassupervised students in projects applying lean manufacturingprinciples to diverse industries including hospitals, banks,restaurants, offices, and pharmaceutical manufacturing. Shecan be contacted by e-mail: [email protected].

University of Rhode Island, Industrial and Systems Engi-neering, 103 Gilbreth Hall, 2 East Alumni Ave., Kingston,Rhode Island 02881, USA.

Cyrus Agarabi is currently a PhD Scholarin the Department of Biomedical and Phar-maceutical Sciences, College of Pharmacy, atUniversity of Rhode Island. He earned aDoctor of Pharmacy (Pharm. D) degree in2005. Moreover, he has completed both a MSin pharmaceutical sciences and a MSc inindustrial engineering from the same uni-

versity. He is the recipient of the Rainville Leadership GroupAward, URI Centennial Scholarship, Daniel P. Tsao Scholar-ship, and Pharmacy Alumni Scholarship, to name a few. Hisresearch interests are focused on improvements of currentpharmaceutical manufacturing systems utilizing principlesof lean manufacturing and formulation development of anovel multiparticulate extended release pharmaceutical de-livery platform. He can be contacted by e-mail:[email protected].

Dr. Thomas E. Needham is a Professor inthe Department of Applied PharmaceuticalSciences at the University of Rhode Island,College of Pharmacy and has more than 35years of experience in the pharmaceuticalindustry and academia. He has extensiveexperience in the development of drug deliv-ery product systems that incorporate many

different types of conventional and biotechnology drugs fororal, parenteral, intranasal, and transdermal administra-tion. Needham has a wide variety of management experiencethat includes responsibility for product development, scaleup,regulatory documentation, manufacturing specifications, andtechnical support for manufacturing and quality control. Hehas developed, directed, and presented GMP training pro-




grams for the US FDA and a number of pharmaceuticalcompanies. He has served as a consultant and expert witnessfor both the FDA and a number of US and internationalpharmaceutical companies in the areas of GMPs and productdevelopment. Needham earned his BS in pharmacy and MSand PhD in pharmaceutics from the University of RhodeIsland. He has more than 70 publications and five patents. Hecan be contacted by e-mail: [email protected].

University of Rhode Island, College of Pharmacy, Bio-medical and Pharmaceutical Sciences, Kingston, Rhode Is-land 02881 USA.

Sirine A. Saleem graduated with a BSc inindustrial engineering from University ofJordan in 1997, where she implemented vali-dation guidelines and employed quality con-trol and Statistical Process Control (SPC)methodologies to identify improvements fora manufacturing process in a pharmaceuti-cal facility. After that, she received her MSc

degree in mechanical engineering from University of SouthCarolina where her research focused on environmentallyconscious manufacturing. Currently, she is pursuing a PhDin industrial and systems engineering at the University ofRhode Island. She can be contacted by e-mail: [email protected].

University of Rhode Island, Industrial and Systems Engi-neering, 103 Gilbreth Hall, 2 East Alumni Ave., Kingston,Rhode Island 02881, USA.


Retrofitting CIP


This articledescribes thephysicalmodificationsand additionsretrofitted intoexisting plantsto incorporateCIP technologyand some of thetechniques thatcan be used to‘stretch’ theexisting CIPsystems for besteffect.

Retrofitting CIP into API Plants

by Nigel A. Fletcher

Introduction

With the ever increasing demand forimproved quality, the API indus-try has turned its attention to clean-ing. Every aspect of cleaning is

being considered from the use, or not, of deter-gents, to the methods used and how these affectthe final quality of the API product. This is nota problem for new plant designs where clean-ing and Cleaning-In-Place (CIP) can be inte-grated into the design, but it is a differentmatter for API facilities already in production.

Five to 10 years ago CIP was simply some-thing that only really happened in plants pro-ducing high quality products or those thatproduced sterile, aseptic, or special products.Many plants appear to have paid scant regardto cleaning, restricting cleaning to boiling outthe reactors, and cursorily, spraying centri-fuges and dryers to remove ‘gross’ contamina-tion. Little attention was paid to out-of-the-way places in nozzles, valve bodies, or vent

lines etc. because the strict quality require-ments in place now were not in place then. Thiswas possibly because the APIs produced tendedto be less potent and could tolerate a smallamount of contamination from another productwithout there being a serious risk to the enduser.

Although there are simple solutions to thecleaning problem, the most difficult part isretrofitting CIP into a plant not originally de-signed to be cleaned. This particularly appliesto the integration of CIP as part of plant retro-fits, turnaround optimizations, or plant exten-sions. Let us be clear at this point that CIP isbeing added to plants not only to improve prod-uct quality, but also in the case of multipurposeplants, to reduce the time between productcampaigns. Thus, CIP retrofits are being seenas a way of improving plant productivity (pay-ing for itself) as well as satisfying improvedquality requirements. This article will providesome examples of potential solutions that could

be used to improve cleaning. Many‘solutions’ in this article are spe-cific to a particular plant or instal-lation, but are intended to illus-trate cleaning/CIP can be retrofit-ted even in older plants. In theexamples used, operator time hasbeen reduced and plant turn-around times have been improved.Time savings range from a fewhours to as much as several shifts.The precise time saving is, ofcourse, dependent on the extent ofthe plant modification, when it ismodified and how much automa-tion has been included in thechange. Many of the changes out-lined in this article were installedpiecemeal to fit with the manufac-turers’ budgets and manufactur-ing timetables.

Things have changed in the

Figure 1. Spray patternin a typical API reactor.

Reprinted from





Retrofitting CIP


past five to 10 years in quality terms and there is a signifi-cantly greater requirement for plants to be demonstrablyclean and prepared for processing the next product. Whilethis QA requirement is perfectly reasonable, it has led to asituation where new plants have to be designed and con-structed to integrate CIP into their operation, but olderplants have been left behind. As these plants often have aconsiderable residual life before they are decommissioned,the operating company is left with the situation of having toimplement a cleaning regime in a plant that was not designedfor it. It is at this point that consultants get called in to ‘solve’the situation as they are specialists and the operating com-pany personnel are unlikely to have the necessary time norexpertise to perform the analysis and design required.

Having identified the problem, the client usually asks ifthe recommended solutions also can improve/reduce theexposure to the operator, reduce the turnaround time be-tween product campaigns, reduce solvent usage, reduce wastedisposal costs, and reduce environmental impact from clean-ing operations/wastes. Ideally, the consultant’s proposalswill achieve all of these, but their main purpose must be toimprove product quality.

Where the Problems are FoundSince there are many areas in an API plant that needcleaning and the potential solutions would number in thethousands, this article will focus on a few examples in anumber of key areas in an API plant, including:

• reactor heads and some of the associated overheads• Nütsche-type filters• centrifuges• dryers

Other considerations are the protocols used for the cleaningand the ‘solvents’ used. These critical areas are often over-looked or, worse, treated as unimportant. Undervaluing ofthe cleaning protocol often manifests itself in ‘one protocolcleans everything’ or ‘if it is wetted by the CIP, then it will becleaned.’ The same applies to the choice of cleaning solventwhere ‘one solvent does all the cleaning,’ i.e., is the ‘universal’solvent. These inappropriate attitudes are all too prevalentand the industry needs education in this area; however, thissubject will not be addressed in this article.

API ReactorsReactors in API plants have a relatively common design,whether fabricated in glass-lined carbon steel or stainlesssteel, and generally consist of a vertical shell, closed top andbottom with dished ends. Nozzles on the top head provide themajority of the access to the vessel. These nozzles are closedby valves, agitators, baffles, or dip-pipes, and are often citedas the areas where the client company has most problemswith residues. Other ‘dirty’ areas are crusty product ringspart way up the vessel sidewall and underside of the agitatorblades.

All of these areas can be cleaned, but there are two

principal approaches that can be adopted. The first is themost simple – open the manway and lower a high pressurerotary spray unit into the vessel. These provide thin jets thatimpact on the walls in a defined pattern that progresses overthe whole interior surface. Cleaning is achieved mainly by‘impact cleaning.’ This is very effective and many suppliers ofthese types of units will show photographs highlighting whathas been achieved by their units. This is a perfectly accept-able solution with many benefits, but there are two reasonswhy many companies will not adopt this solution. The first isthat these systems use water and there is a concern that thedirt will not be properly dissolved and removed especiallyfrom ‘difficult’ areas deep in the recesses of nozzles. Secondly,the vessel to be cleaned needs to be opened to allow theinsertion of the (bulky) spray unit. If the product to be cleanedis highly potent or toxic, then there is a serious risk that theoperators could be exposed/contaminated. These objectionscan be overcome, but the cost and the operational inconve-nience can be high and so the integration of these systems islow.

The second technique is to permanently install sprayballsor nozzles to allow CIP. This technique is commonly used, butcan fail easily. The first difficulty to overcome is that there isoften only one nozzle that can be adapted to install a spraydevice. This is doomed to failure as a simple pattern analysisshows - Figure 1. Objects such as the agitator shaft impedethe spray and the vessel is left dirty in the ‘shadow’ area. Thedark areas in the Figure show the untouched and uncleanedareas. A point also worth noting from the pattern is that if thethrow of the spray device is inadequate as shown by the 1.0mand 1.5m dotted lines, it is quite clear that parts of the vesselwill not be cleaned adequately.

If only one nozzle is available, the installation of a secondspray device is possible (Figure 2) by means of an internally

Figure 2. Second sprayball installation in a reactor with only oneaccess nozzle for CIP sprays.


Retrofitting CIP


mounted angled supply line from the first sprayball’s dip-pipe. This adds a second spray on the opposite side of thevessel so effectively eliminating the shadow area of theagitator shaft and other obstructions.

This is not a perfect solution because the installation ofpiping inside this type of vessel is not ideal and finding asecond nozzle would be a better solution. This can involve adifficult analysis of the vessel processes and may requiresome re-piping of the top head of the vessel.

However, it is rare that the simple installation of twosprayballs will clean the top head nozzles as Figure 3 suggests.Although the shadow area in the nozzle, shown in Figure 3, issmall, it is difficult to clean. There used to be a mistaken beliefthat ‘rebound’ washing would occur in these areas. This ideameant that the spray hitting one side of the nozzle wouldbounce off the nozzle wall and hit the side in shadow. Thus, theshadow areas would be wetted, and so in time, would be

cleaned. While rebound washing can occur, it only happenswhere the jets or sprays are moving at high velocity.

High velocity jets usually only occur locally to the sprayballso by the time they reach the deeper recesses of a nozzle, theysimply do not have the energy to rebound according to thetheory. Thus, rebound washing does not generally work. So ifthis method is not feasible, then this means that an alterna-tive must be found. Washing the deep parts of a nozzle is, ofcourse, trivial if the nozzle can be directly flushed by solventfrom a bulk system or can be flushed by condensate from acondenser. Where this is not possible, the use of a spray ring(Figure 5) or a standard instrument tee (Figure 4) should beconsidered.

Both of these installed at the top of the nozzle allowcleaning fluid to be introduced to produce the desired clean-ing effect. The spray ring, supplied by a number of manufac-turers, can be used to flush nozzles that have dip-pipes orbaffles.

Putting all the above ideas together achieves very goodcleaning of the top head and nozzles of a reactor. However,this does not address other reactor cleaning problems, suchas a crusted ring of product on the side wall or deposits on theunderside of the agitator blades. In these cases, the use ofhigh pressure washers or traditional boil-out techniques areprobably the most effective short-term solutions. If the crustedmaterials are sufficiently soluble, the solvent running downthe walls from the sprays in the top head may slowly removethe crusted ring. At this point, it is worth commenting thatthe choice of cleaning solvent is critical for these moredifficult cleaning duties and an area where many fail. De-tailed analysis of the cleaning problems is required at thispoint. This is not covered in this article.

Whatever solution has been adopted, the spent CIP solu-tion flows out of the base of the vessel through the bottomoutlet valve. The supply of cleaning solution from the spraysmust be such that flooding in the bottom of the vessel isavoided. If flooding does occur, this can give rise to re-deposition of dislodged product and a cleaning failure.

This bottom outlet valve is another area where contamina-tion occurs and the selection of the right valve can improvethis situation. For example, a ball valve can be purchasedwith side ports to allow flushing of the body cavity. If the valvetype is unchangeable, repeatedly opening and closing thevalve during the cleaning sequence may help. If this tech-nique does not work and no alternative valve can be installed,dismantling and manual cleaning may be necessary.

Nütsche Filter CleaningA review of Nütsche-type filters reveals that they closelyresemble vessels. The Nütsche filter and filter-dryer aresimilar machines in many respects and suggestions in thissection apply to both types. However, the dryer has the addedcapability of heating cleaning solutions. The top head is verysimilar to that of a vessel with numerous nozzles, dip-pipes,and an agitator. Thus, the solutions that serve to cleanvessels also serve to clean the top heads of filters and filter-dryers. However, there are exceptions including which is the

Figure 4. ‘Flushable’ nozzle which includes an instrument teewhich can be connected to the cleaning system.

Figure 3. Nozzle shadowing resulting from the spray not fullypenetrating into the nozzle.


Retrofitting CIP


dust dome or filter that serves to remove dust fines from thevented gas during depressurization or in the case of a filter-dryer, the dry product fines from the drying operation.

The cleaning of this filter is not trivial and the author isunaware of a guaranteed way of achieving complete cleaningof the filter elements. The author is aware of customers’ trialsthat are examining the use of sinter elements and how theymay be cleaned in place. The author understands that someresults have been validated, but many pharmaceutical manu-facturers still have concerns about the cleanability of theseelements. If we accept this limitation, we need to considerwhat we can achieve. This may be gross cleaning of theelements and wetting them in situ (so called Wet-In-Place orWIP) so that they can be removed manually for disposal orwashing externally. It must be remembered that there aretwo sides to the filters – the so-called dirty and clean sides.

For cleaning purposes, both these sides must be wettedalthough the clean side is not so critical in this respect. Figure6 shows the access door to the filter dome, normally used tochange the filter elements. The door/cover has been modifiedto include a sprayball (see tri-clamp connection). This sprayprovides a 180° pattern spray that will wet a reasonableproportion of the elements, but not all of them. Ideally, thisunit will have an internal pipe which holds the sprayball andpossibly a second sprayball inside the ring of filter elements.However, this configuration is difficult to arrange because thepiping and sprays must not interfere with the filter elements.Another spray needs to be installed on the opposite side of thehousing to improve the overall wetting of the elements. It isdifficult to be precise as the number of elements varies accord-ing to duty, size, and make of the Nütsche filter.

An important point to note is that the modification toinstall the sprayball is restricted to a part (the access door inthis case) that can easily be removed from the machine formodification in the engineering workshop. If the filter is afilter-dryer, it is also possible to use the explosion relief hatch(if installed) as a suitable nozzle that can be modified toprovide sprays into the filter dome. A similar modificationcan be made to the top head of the filter dome on the clean side

where a 180° pattern sprayball can be installed to spray thetop end of the filter candles. This gets the internal surfaces ofthe filter candles wet as well as the top head of the filter dome.Thus, spraying both sides of the candles means that they aresafer to remove for manual cleaning or disposal.

There is an important point to be noted in relation to the twomodifications described above. Where possible, all CIP retro-fits should be carried out either by insertion into existingnozzles or by modification of parts that can be removed fromthe plant to a suitable workshop. In the case of the filter dome,this means the filter access door/hatch and the top head, whichis usually flanged and bolted to the body. Where a nozzle needsto be modified – then whatever its function – it can often bemodified to include an extra connection for CIP. This is wherethe experience of the consultant can be useful to recommend asolution that needs unusual or lateral thinking.

Peeler CentrifugesIf the Nütsche filter is not the main means of isolating theproduct from its mother liquor, a centrifuge is used. Thesecome in a variety of types, but the two principal typesencountered are the inverting bag or peeler centrifuges. Thisarticle will consider the latter, peeler, type. Retrofitting CIPinto and around the inverting bag centrifuge will be consid-ered in another article.

The peeler centrifuge is very complex internally. There isthe feed pipe, the solids peeler mechanism, solids levelmeasurement device, wash fluid inlet, splash guard(s), andbasket assembly. If the machine is reasonably modern, therealso may be some internal spray devices for CIP/cleaningalthough these may not be as useful as they appear. All ofthese items are housed in a large outer housing, which wrapsaround the basket and out of the bottom of which passes themother liquor to an external collection system. The housingis closed by a full face door, which acts as the mounting pointfor the peeler mechanism, feed pipe etc. and has a largediameter seal with the main housing.

This main seal is the source of many problems. There arefew centrifuges where this seal does not leak, if not regularly,then frequently enough to be a source of irritation to theoperating staff. This leakage is a problem for cleaning as themost effective CIP mechanism for this type of machine is topartially flood the machine and turn the basket round slowly.This ‘washing machine’ action is very effective for cleaningthe basket and filter media although not always successfulfor the solids outlet chute. If it can be engineered and willtolerate it, without leaking, then reverse flooding the solidsoutlet chute can be very effective. If flooding is used , a shortinternal spray may be needed to reduce the risk of a residual‘ring’ at the surface of the cleaning ‘pool.’ This final wash-down also will remove re-deposited solids after the floodsolution has been drained away. The list of difficult areas towash in the peeler centrifuge can be summarized as follows:

• behind the basket• clamping bars for the filtration media• solids discharge chute

Figure 5. Proprietary spray ring to be installed inside the bolt circle ofthe nozzle and connected to the cleaning system by a short branch.


Retrofitting CIP


• around mounting screws for internals• behind cover plates, e.g., end of the peeler arm• inlet and outlet nozzles on the main housing

Flooding the machine can deal with some (but not all) of thesequite effectively, and if the machine cannot be flooded due toa poor door seal, other measures need to be considered.

If there is an inert gas nozzle on the back of the housing,this can be adapted to incorporate a low profile spray nozzlewith 180° spray pattern, which can provide a reasonablewash behind the basket. If this is not available, the vapor venton the top of the housing can be adapted in the same way witha low profile, 180° spray pattern, spray nozzle. This does nottarget the back of the basket in the same way, and results canbe mixed, but can be improved if the basket is rotated slowly,simultaneously with spraying. Recirculating the cleaningfluid helps combined with spraying for 15 to 20 minutes. Donot continue with a single CIP operation for more than 30minutes continuously because it rarely increases the qualityof the end result. These methods cover washing behind thebasket, but do not address cleaning areas inside the basket.

If the machine has been fitted with a spray arm forinjecting wash fluid during operation, it can be used toprovide a very satisfactory method of introducing cleaningfluid to the basket. An example is shown in Figure 8. This canbe used to wash the media clamping bars, but it should bepointed out that this system does not usually operate at a

sufficiently high pressure and so this may require a pressureboost from either a pump or increased pressure in the sourcevessel. Choice of cleaning fluid is critical as the cake washsprays will not provide impact cleaning jets so the cleaningfluid must be able to dissolve the product.

In the example shown in Figure 8, the centrifuge has beenfitted with various spray devices. One above the slurry feed forgeneral washing of the internal area of the basket and a secondone immediately behind the peeler arm (seen in the top left ofthe photograph) and these provide for fairly good cleaning ofthe internals of the machine. However, they do not cleanmounting screws or studs or behind cover plates. A reportedproblem is the ingress of product behind the cover plate at theend of the peeler arm. Commonly, these plates are mountedwith a metal to metal seal which is not always a good seal. Theresult is that a brown, sticky residue can be found behind theplate when the maintenance team carry out work on themechanism. This leakage is caused by distortion or damage ofthe cover plate. The solution is to retrofit a very thin (0.25-0.5mm), full-face PTFE gasket under the cover plate with its edgeflush with the outer face of the peeler arm and cover plate.When the cover plate is re-bolted, then the gasket performs asany other gasket and seals the gap effectively. This eliminatesthe cleaning problem permanently.

A similar problem is seen around the mounting screws forthe peeler knife where product is forced due to the productpressure during the peeling operation. While the fitting of aPTFE gasket may solve this problem, it is likely to return andso a simpler arrangement is to try to space the knife awayfrom the support bracket. This larger gap/space is a lot easierto wash out than the narrow one where the knife is bolteddirectly to the support.

Figure 6. Nütsche dust filter showing the access cover fitted witha cleaning nozzle.

Figure 7. Typical peeler type centrifuge.


Retrofitting CIP


Figure 9. Spray installation in elbow.

Figure 10. Bubble spray unit.

Centrate Liquor Outlet and PipingThe inlet and outlet (mother liquor) nozzles need to beconsidered. Generally, the inlet nozzle is not too much of aproblem as the slurry is fed into the machine at a reasonablevelocity; so, this nozzle does not get particularly contami-nated and can be cleaned by pumping cleaning solutionthrough it at high velocity with solution recirculation ifpossible. The outlet or mother liquor nozzle is rather differ-ent. The flow at the start of the centrifuging operation is high,but when the main dewatering phase finishes, the liquorspass through this nozzle as a spray or at low flowrate. Thisleaves the nozzle and downstream mother liquor systemcontaminated. Normally, this is not a problem as it is rarethat the mother liquor is the desired product and the motherliquor system is usually beyond the GMP boundary. How-ever, if the product is the mother liquor or there is a cloth/media failure and the batch needs recovery, the cleanliness ofthe mother liquor system becomes important. The idealsolution is to flood the base of the centrifuge together with themother liquor system. However, this is not always possibleand an alternative solution must be found. Normally, themother liquor outlet has an elbow on it pointing down forgravity drainage of the liquors or, occasionally, the elbow ishorizontal. In either case, the solution is to replace the elbowwith a tee. This is an opportunity to create a new access pointonto the system into which a spray device may be fitted. Theonly disadvantage with this method is that the tee where the

spray unit is fitted becomes a ‘dead area’ or an area forcontamination and must be cleaned. Alternatively, a shortspray ‘bubble’ can be fitted on the end of the elbow whichvirtually eliminates the dead area. These alternatives areshown in the illustrations in Figures 9 and 10.

The spray bubble can be mounted on a short reach tri-clamp nozzle ‘stabbed’ into the line on the corner of the elbow.Its spray pattern can be adjusted to jet into the centrifugeoutlet and also simultaneously down the mother liquor pip-ing.

Vertical and Horizontal Axis DryersThe final item of equipment to consider is the dryer, which isavailable in a number of different formats. Typically, thereare two types in most pharmaceutical plants: the vertical axisconical dryer and the horizontal axis paddle dryer. It canreadily be seen that the conical dryer has many similaritieswith the reactor and the solutions for cleaning these ma-chines are broadly the same as for the reactor. Flood washingis a common method, but suffers the problem of a largevolume of contaminated waste solvent. A considerably bettermethod is to use a tank washer lowered into the body througha suitable nozzle and to recirculate the wash solution.

This type of machine is not difficult to wash compared tothe horizontal axis paddle dryer. This is because the bladeson a paddle dryer are usually arranged to sweep nearly 100%of the internal surface of the dryer body. This is a problembecause there is no point to introduce a spray device to

Figure 8. Typical spray installation in centrifuge body as shown bythe three arrows.


Retrofitting CIP


intrude inside the body. This means that you cannot use themost effective method of cleaning with the paddles turning.If the paddles are static, there is a concern that areas may bemissed due to ‘shadowing’ and not cleaned properly. A com-mon solution to this situation is to flood the body and agitatethe solution using the paddles. This is effective, but uses alarge amount of solution and can take a considerable amountof time if the method uses total flooding. The target in sucha situation is to reduce both time and solvent usage.

If the idea of flooding is disregarded, two possibilities canbe examined: static or dynamic washing. In the case of staticwashing, the paddles are brought to a standstill at a knownposition and a sprayball inserted into the body through thetop manway or a nozzle on the top of the machine. In this case,the spray is operated, removed, the paddles rotated 90°, thespray reinserted, the spray operated a second time, and thissequence repeated until the desired end result achieved.Washing solutions can be re-circulated or reused until ex-hausted so that the total volume of solution is kept low. There

are clear risks with this approach, for example, operating thepaddles without having withdrawn the spray device. Thesecan be reduced by careful operating practice and having aPTFE dip-pipe and sprayball, which will shear off if struck bythe paddles without damaging the dryer. Alternatively, thesprayball can be the type which extends from a housingmounted on the nozzle when there is sufficient pressure ofcleaning solution. An example is shown in Figure 11. Thistype usually rotates and provides good coverage of theinternals. The same sequence of withdraw, rotate paddles,insert and wash, can be used with this type, but it requiresless operator input and can be automated.

While static washing techniques have many benefits, thealternative dynamic washing technique is usually more effec-tive. This is because the cleaning fluid can often contact allthe internal parts of the machine compared to the shadowingof the spray in the static situation. But as stated above, thisis more difficult as the opportunities for this in the paddledryer are so limited.

Only if the body has nozzles at each end that can beadapted is there a realistic opportunity of undertaking dy-namic cleaning. It should be noted that a spray mounted inthe end of a cylinder can only spray about half the cylinderand opposite end as it is operating from inside a nozzle. Thesimple diagram in Figure 12 illustrates this with the darkareas showing unwashed sections. Thus, two such sprayballswould be required to cover the whole of the interior of thedryer.

The throw and spray angle of the spray devices need to beselected carefully. The illustration omits the paddles andthese would interfere with the spray patterns from the twospray devices. However, as they are rotating during thewashing cycle, the interference is significantly reduced for inone position the paddles obstruct the spray, but a few degreesof rotation later they are less of an obstruction. This aspect ofthe cleaning has to be checked by spray pattern analysis forthe paddles in a number of positions. This method does relyon the availability of nozzles or manways. Even a manwaypositioned low down on the body can be used, providing themanway plug can be modified and the cleaning operation isperformed with the dryer body being kept fully drained so thespray does not become submerged.

As with the other items of equipment covered in thisarticle, there are many other areas of dryers that needcleaning and they need to be considered separately. These

Figure 11. Extending spray unit which extends when the cleaningsolution pressure is high enough.

Figure 12. Spray coverage in the cylindrical body of a dryer.


Retrofitting CIP


include the solids outlet valve, solids outlet chute, the shaftseal area inside the body of the machine, filter dome, andother nozzles on the body.

Other Equipment and CIP NozzlesThere are many different types of spray nozzles with somenozzles providing uniform mists and others providing highlydirected jets. The consultant has to choose which type bestsuits each particular installation. For example, a spray headwith a solid cone spray with narrow included angle may be thecorrect choice to clean an isolated side nozzle. Each type willachieve a certain degree of cleaning so it is important thatwhen considering a particular installation that the nozzle ischosen to give the correct coverage, the right ‘direction’, theright fluid velocity if impact cleaning is required, and theright physical size to fit the nozzle or chosen location. Ifimpact cleaning is required, many spray nozzles need to belocated well within two meters of the ‘target.’ Beyond thispoint, spray nozzles lose much of their energy and becomeunsuitable for impact cleaning. If the ‘target’ is more than 1.5to 2.0 meters from the nearest spray, then as stated above, itmay be more appropriate to choose a rotary or powered jetunit which is ideal for this type of duty. Sprayball or spraynozzle selection may take some time to find, but the resultswill be better if the correct nozzle is used. Where this is notachievable, it is possible to work directly with sprayballmanufacturers to achieve the precise flows and spray pat-terns to solve specific cleaning problems.

The selection of spray device becomes more critical asother items of equipment are considered. Many, includingthose considered in this article, have either complex struc-tures or deep, difficult to clean areas and specialist advice isusually required to ensure a successful solution to the clean-ing problem. Gasket areas are frequently a problem, includ-ing body gaskets for vessels or equipment or nozzle gaskets.The gasket is rarely a perfect fit in the flange and there isoften a recess or pair of crevices (one either side of the gasket),which can become contaminated with ‘product.’ Gasket areashave to be very specifically addressed and the solution willdepend on gasket location and the cleaning technique used.

For gaskets in a horizontal plane, the risk is usuallylower than vertically orientated gaskets. With a verticalgasket the bottom position allows settling of contaminationinto the crevices and the removal of this contamination isvery difficult. Critical to this cleaning is the selection of thecorrect cleaning agent to dissolve the ‘dirt’; frequent refresh-ing of the solution in contact with the dirt so that themaximum dissolution driving force is available; good agita-tion of the solution during cleaning, and if possible, the useof high velocity jets for cleaning to give a scouring action.The use of the same techniques is also applicable to horizon-tal gaskets, but the use of burst washing is advantageous asthis allows contaminated cleaning solution to drain awaybefore new solution is introduced. These techniques doachieve most of the cleaning required except where theproduct is highly potent. In this case, dismantling may bethe only method to ensure removal of the product. The

author has noted that process transfer piping gaskets arevirtually never considered a problem although these gas-kets are also in contact with slurries and product solutions.Transfer piping also includes sample points, drains, instru-ment branches, and various other crevice areas. These areoften ignored as they are not visible, but they can be as mucha source of problems as any equipment gasket.

ConclusionsRetrofitting of CIP is most successful when simple solutionsare adopted with relatively little engineering work on theplant. Most retrofits use existing nozzles or access points intothe process, or where this has not been possible, then simple‘stab-in’ connections. If the CIP retrofit is too complex, tooexpensive, or requires a long shut-down period, there will belittle incentive to perform the work as it will be perceived tobe too difficult. Retrofits should be designed so that they canbe performed during a campaign turnaround, during a shortshut-down, or as a last resort, during the annual plant shut-down. In other words, the easier it is to fit the modification,the more likely it is to be performed and the greater thechance of the client/operating company realizing the desiredCIP goals. Successful CIP retrofits are effective and pay backall of their costs in improved quality and shortened turn-around time.

As mentioned earlier, the protocol used for cleaning also isimportant. For example, washing continuously may not be aseffective as burst washing. The protocol may or may not usedetergent, acid or alkaline wash solutions, and these have tobe considered when retrofitting CIP or cleaning systems intothe plant to ensure they can be removed at the finish of thecleaning sequence. The selection of the correct ‘solvent’ is,similarly critical, and time may have to be spent reviewingand considering alternatives to achieve the best results. Thisactivity is not a trivial one and requires experience and astructured approach.

In summary, when retrofitting CIP to an existing plant,the consultant has to consider every aspect of the cleaningprocess, including the plant equipment, the operational re-quirements, PED (or equivalent pressure equipment codes),and HSE issues. Thus, retrofitting CIP must start with thedesired end result or target and accommodate the physicallimitations of the plant and result in the potential changesthat need to be incorporated into the cleaning protocol. Thekey to a successful result is not to be too narrow in thinkingabout physical solutions. Problems have to be analyzedsystematically and the resultant solutions kept as simple aspossible.

About the AuthorNigel Fletcher joined Foster Wheeler in1975 and has worked on pharmaceuticalprojects for more than 25 years. He hasextensive experience in very large and verysmall specialized API and aseptic facilitiesdesign. He has been Process Manager for alarge number of projects, including multi-


Retrofitting CIP


purpose and dedicated API facilities, pilot plant API units,and a multi-product primary facility featuring an extremelyconfigurable arrangement of equipment. Nigel also has beenTechnical Manager for projects, including a bulk sterile andbulk fermentation facilities. He has specialist experience inCIP and specialist cleaning, high containment systems, andmaterials handling. He consults on these and many otheraspects of pharmaceutical manufacturing. He has experiencefor anti-ulcer, cytotoxic, anti-depressant, steroids, antibiot-ics, parenteral, and oral drugs together with upstream anddownstream processing biotechnology facilities. Fletchergraduated from the Imperial College of Science and Technol-ogy, London, and holds a BSc (Hons) degree in chemicalengineering. He is a Fellow of the Institution of ChemicalEngineers and is a member of the International Society forPharmaceutical Engineering. He can be reached by tele-phone: +1-44-118-913-2184 or by e-mail: [email protected].

Foster Wheeler Energy Limited, Shinfield Park, Reading,Berkshire RG2 9FW, United Kingdom.


Commissioning, Qualification, and Verification


Solving the Terminology Conundrum

by Robert Adamson, Nuala Calnan, Robert E. Chew,and Steven J. Wisniewski

Introduction

In today’s biopharma and pharmaceuticalindustries, three related, but distinct termsare in common use: commissioning, qualification, and verification. Inconsistent in-

terpretation and application of these termsleads to misunderstandings and inefficiencieson the part of vendors, service providers, andmanufacturing personnel from company to com-pany. This article, through a review of theindustry definitions and associated practices,is intended to stimulate discussion on resolv-ing this terminology conundrum and providekey input to pending publications of ISPEBaseline® Guides.

In 2001, ISPE issued the Baseline® GuideVolume 5: Commissioning and Qualification,that provided definitions for two of these terms:Commissioning and Qualification. In 2007,ASTM E2500-07: A Standard Guide for theSpecification, Design, and Verification of Phar-maceutical and Biopharmaceutical Manufac-turing Systems and Equipment was issued.This standard introduced the term “verifica-tion” as a new term for demonstrating suitabil-ity and fitness for intended purpose, in place ofthe terms commissioning and qualification.

The terms “verification” and “commission-ing” are used in many industries and have afairly consistent meaning. The term “qualifica-tion” has been used by the regulated pharma-ceutical and biotech industries, and can befound in EU regulations, as well as US, EU,and ICH guidance documents. Do these termsmean the same thing (more or less) or do theyconvey three different necessary and uniquemeanings?

This article is divided into two parts:

1. definitions and use of the terms found inpublished regulatory and guidance docu-ments

2. analysis of the terms in light of currentpractices

The authors invite readers to respond to thisdiscussion, either through the ISPE Commis-sioning and Qualification Community of Prac-tice (C&Q COP) discussion board, or via directcommunication. Such input will be consideredwhen any related updates to the Baseline®

Guides are undertaken.

Part I – Definitions and CitationsQualificationThe term qualification, while not specificallyfound in US GMP regulations, is found in EUregulations, ICH Q7A, and ICH Q9, as well asWHO and other country regulations and guid-ance documents.

US – FDAThe US GMPs do not explicitly mention theterm qualification – in that there is no specificregulatory requirement to produce documentslabeled installation, operation, or performancequalification. However, there are clear expec-tations of a process that demonstrates fitnessfor intended use and assures proper perfor-mance.

US GMPs require that:

• Facilities be “suitable... to facilitate cleaning,maintenance, and proper operation.”

• Equipment is to “be of appropriate design...to facilitate operations for its intended use.”(21 CFR 211.42, 211.63, 606.40, 606.60,820.40, 820.60).

• Automated systems are required to be “checkedaccording to a written program designed toassure proper performance” (211.68).

The medical device regulations (21 CFR 820)require that: “computer software programs shallbe validated by adequate and documented test-ing” (820.61).

21 CFR Part 11 requires [for those systems to

This articlediscusses theterms“commissioning,”“qualification,”and“verification.”Do the termsrefer to thesame ordifferent ideas?How should thepharmaceuticalandbiotechnologyindustries usethese terms in aconsistent andmeaningfulway? Thisarticle providesa compilation ofhow theseterms are usedin regulationsand by variousindustries, andprovides aproposal forclear definitionsto be used asISPE updatesand createsBaseline®Guides.

Reprinted from







which Part 11 applies]: “Validation ofsystems to ensure accuracy, reliability,consistent intended performance, andthe ability to discern invalid or alteredrecord.”

The 1987 FDA guidance on processvalidation first introduced the termqualification in these terms:

Installation qualification studies estab-lish confidence that the process equip-ment and ancillary systems are capableof consistently operating within estab-lished limits and tolerances. After pro-cess equipment is designed or selected, itshould be evaluated and tested to verifythat it is capable of operating satisfacto-rily within the operating limits requiredby the process. This phase of validationincludes examination of equipment de-sign; determination of calibration, main-tenance, and adjustment requirements;and identifying critical equipment fea-tures that could affect the process andproduct. Information obtained from thesestudies should be used to establish writ-ten procedures covering equipment cali-bration, maintenance, monitoring, andcontrol. In assessing the suitabilityof a given piece of equipment [em-phasis added], it is usually insufficientto rely solely upon the representations ofthe equipment supplier, or upon experi-ence in producing some other product.Sound theoretical and practical engi-neering principles and considerationsare a first step in the assessment.

The Food and Drug Administration’s(FDA’s) current thinking on the topic ofActive Substances Used as StartingMaterials is represented by the ICHQ7A guidance, which includes refer-ences to Qualification.

EU – EMEAEU Volume 4: EU Guidelines to GoodManufacturing Practice MedicinalProducts for Human and VeterinaryUse, Annex 15 (Qualification and Vali-dation), while specifically referencingboth qualification and validation, fur-ther outlines in its lead Principle State-ment that:

quirements for Active Substancesused as Starting Materials, whichalso includes specific references toqualification activities.

ICH Q9 has recently been adopted bythe EU as part of its Vol 4 GMPs asAnnex 20.

ICH Harmonized TripartiteGuidelinesThe ICH International guidance docu-ments contain additional references toqualification. ICH Q7A, GMPs for Ac-tive Pharmaceutical Ingredients statesthat:

“Before initiating process validationactivities, appropriate qualification ofcritical equipment and systems shouldbe completed. Qualification is usuallycarried out by conducting the followingactivities, individually or combined:

• Design Qualification (DQ): docu-mented verification that the proposeddesign of the facilities, equipment, orsystems is suitable for the intendeduse.

• Installation Qualification (IQ): docu-mented verification that the equip-ment or systems, as installed or modi-fied.

• Operational Qualification (OQ):documented verification that theequipment or systems, as installedor modified, perform as intendedthroughout the anticipated operat-ing ranges.

• Performance Qualification (PQ):documented verification that theequipment and ancillary systems, asconnected together, can perform ef-fectively and reproducibly based onthe approved process method andspecifications.

The recent ICH Q9, Quality Risk Man-agement, includes an appendix of ap-plications of quality risk management;Appendix II.4 describes how to usequality risk management for facilities,equipment, and utilities, including:

“...manufacturers identify what vali-dation work is needed to prove controlof the critical aspects of their particularoperations... A risk assessment approachshould be used to determine the scopeand extent of validation.”

The Annex goes on to describe the fol-lowing validation and qualificationactivities as:

• The first element of the validation...could be design qualification.

• Installation qualification should beperformed on new or modified facili-ties, systems, and equipment.

• Operational qualification shouldfollow installation qualification.

• Performance qualification shouldfollow successful completion of in-stallation qualification and opera-tional qualification.

The annex includes specifics regardingthe content and execution of qualifica-tion work. Content requirements in-clude the items typically found in anIQ, OQ, or PQ protocol, such as instal-lation verification, collection of equip-ment manuals, calibration, materialsof construction, testing across operat-ing ranges, etc. Execution requirementsinclude:

• Written protocol specifying criticalsteps and acceptance criteria.

• Protocol reviewed and approved(does not specify by whom).

• A report written summarizing re-sults, including recommendingchanges necessary to correct defi-ciencies, and documenting changeswith appropriate justification.

• Formal release to the next step inqualification or validation as a writ-ten authorization (does not specifyby whom).

ICH Q7A has been incorporated intothe EU GMPs as Part II: Basic Re-




“...to determine the scope and extentof qualification of facilities, buildings,and production equipment...”

ISPE Baseline® GuideThe 2001 Commissioning and Qualifi-cation Baseline® Guide defines IQ, OQ,and PQ in similar terms:

• Installation Qualification: the docu-mented verification that all aspectsof a facility, utility or equipmentthat can affect product quality ad-here to approved specifications (e.g.,construction, materials) and are cor-rectly installed.

• Operational Qualification: the docu-mented verification that all aspectsof a facility, utility, or equipmentthat can affect product quality oper-ate as intended throughout all an-ticipated ranges.

• Performance Qualification: the docu-mented verification that all aspectsof a facility, utility, or equipmentthat can affect product quality per-form as intended meeting predeter-mined acceptance criteria.

World Health Organization(WHO)World Health Organization (WHO)Guidance on Validation defines Quali-fication as “Action of proving and docu-menting that any premises, systems,and equipment are properly installedand/or work correctly and lead to theexpected results.”

CommissioningEU – EMEAEU GMPs Annex 11, ComputerisedSystems positions commissioning as acomponent of computer validation:

[The computer validation life] “cycleincludes the stages of planning, specify-ing, programming, testing, commission-

ing, documentation, operation, moni-toring and modifying.”

ISPE Baseline® GuideThe 2001 Commissioning and Qualifi-cation Baseline® Guide defines Com-missioning as “A well planned, docu-mented, and managed engineering ap-proach to the start-up and turnover offacilities, systems, and equipment tothe end-user that results in a safe andfunctional environment that meets es-tablished design requirements andstakeholder expectations.”

The material that follows this defi-nition positions commissioning as aprocess that includes inspections, op-erational testing, and performance test-ing.

Commissioning as defined by non-drug industries:

• Building commissioning providesdocumented confirmation that build-ing systems function according tocriteria set forth in the project docu-ments to satisfy the owner’s opera-tional needs (Building Commission-ing Association).

• Commissioning means to verify thatthe building’s energy related systemsare installed, calibrated and per-form according to the owner’s projectrequirements, basis of design, andconstruction documents (LEED re-quirements).

• Building commissioning is the pro-cess of ensuring that building sys-tems and equipment are designed,installed, tested, and capable of be-ing operated and maintained accord-ing to the owner’s operational needs(US Department of Energy).

• Process of ensuring that new build-ings and their systems perform asdesigned (Oak Ridge National Labo-ratory).

VerificationUS – FDA21 CFR Part 820 (U.S. medical devicequality system regulations) definesVerification to mean: “confirmation byexamination and provision of objectiveevidence that specified requirementshave been fulfilled.” This definition maybe contrasted with the Part 820 defini-tion of Validation, “confirmation by ex-amination and provision of objectiveevidence that the particular require-ments for a specific intended use can beconsistently fulfilled.”

EU – EMEAEU Volume 4: EU Guidelines to GoodManufacturing Practice MedicinalProducts for Human and VeterinaryUse, Annex 15 (Qualification and Vali-dation), Glossary, includes the samedefinitions for DQ, IQ, OQ, and PQ asoriginated in the ICH Q7A document,which defines these activities in termsof a “Documented Verification.”

ASTM E2500 defines Verification as:“A systematic approach to verify thatmanufacturing systems, acting singlyor in combination, are fit for intendeduse, have been properly installed, andare operating correctly. This is an um-brella term that encompasses all typesof approaches to assuring systems arefit for use such as qualification, com-missioning and qualification, verifica-tion, system validation, or other.”

According to ISO 9000:2000 Verifica-tion is defined as the: “Confirmation,through the provision of objective evi-dence, that specified requirements havebeen fulfilled.” Objective evidence isdefined as “data supporting the exist-ence or verity of something.”

IEEE Standard 1012-2004, Standardfor Independent Verification and Vali-dation, defines Verification as: “Pro-cess for determining whether the soft-

“We leave it to industry to debate these proposals; it is important that we achieve aconsistent understanding and application of these terms. Once the debate is complete, it is

for ISPE to incorporate the results into upcoming Baseline® Guides.”




ware products of an activity fulfill therequirements or conditions imposed onthem in the previous activities.”

Of note, the definition of Validation inIEEE standard 1012-2004 is: Valida-tion – process for determining whetherthe requirements and the final as-builtsystem or software product fulfills itsspecific intended use.

Part II – Analysis in Light ofCurrent Practices

The question is, do these three terms –verification, commissioning, qualifica-tion – describe the same or differentthings?

The simplest term to analyze is verifi-cation. For the most part, the defini-tions of verification are consistent (asfound in 21 CFR 820, ISO 9000, IEEE1012-2004, and other sources). Thesedefinitions focus on the idea of “con-firming, through objective evidence,that a specified requirement has beenmet (fulfilled).” ASTM E2500 definesVerification using the same base word:“to verify.” The standard assigns abroader mission for verification, “a sys-tematic approach to verify... systemsand equipment are fit for intended use,properly installed, operating correctly...an umbrella term.”

The term commissioning is more com-plex – different organizations in ourindustry assign different meanings tocommissioning. Some view it as thework that is necessary to make a pieceof equipment ready to start, i.e., thepre-functional inspections and checks(sometimes referred to as pre-commis-sioning). Other organizations are morealigned with the 2001 Commissioningand Qualification Baseline® Guide defi-nition, which positions commissioningas a project lifecycle activity that con-sists of a planned, managed, and docu-mented approach to bringing equip-ment or systems to a full operationalstate, and demonstrating conformancewith specifications and user require-ments. Depending on system complex-ity, the start-up, setting to work, regu-lation and adjustments, cycle develop-

ment, and related work can be signifi-cant, not to mention the actual inspec-tions and testing activities. Using thisidea of commissioning means it mayinclude a number of diverse activitiesrequiring significant planning and co-ordination. Other industries definecommissioning in terms that empha-size the performance testing of a sys-tem or group of systems against end-user requirements.

Finally, qualification, as shown above,is specifically mentioned in EU regula-tions as well as ICH Q7A and ICH Q9.Although the word Qualification is notexplicitly mentioned in US GMP regu-lations, the concept of equipment andfacilities being suitable for their in-tended use is clearly referenced. Fur-thermore, US GMPs do contain a re-quirement to validate certain automa-tion systems, and everyone recognizesthat the typical current industry prac-tice is to include installation, opera-tion, and performance qualification.

How do we reconcile this TerminologyConundrum? Are we to adopt the stancethat if one uses the term “verification,”that this implies a science- and risk-based approach as defined by ASTME2500, whereas use of the terms “com-missioning” and “qualification” impliesa more traditional approach not basedon science and risk? Or do these threeterms describe three different ideas orprocesses, each of which can have auseful place in our approach to deliver-ing equipment, systems, and automa-tion that are suitable for their in-tended use?

1. Irrespective of an organization’sregulatory compliance strategy ofusing either a program labeled “Veri-fication” or “Qualification,” facili-ties and equipment will still need tobe commissioned as defined above.Therefore, a well planned, managed,and documented effort to start-upand place into service a system,equipment, or combination thereof,including automation, will need tobe undertaken – commissioning.

This phase includes safe start-up,setting to work, regulation and ad-justment, cycle development, etc.,which contribute to achieving a fulloperational state.

2. A significant amount of valuableverification work may occur duringthis commissioning process, e.g.,physical inspections, documentationreviews, operational testing, andperformance testing. Retention ofthe term commissioning for this com-plex process of placing equipmentinto operation may therefore be ap-propriate, and for this term to ex-tend to and include, the verificationwork that may occur at this time.

3. Assignment of the term verificationto the act of confirming, throughobjective evidence, that a particularspecification has been met is appro-priate, given the common under-standing of the meaning of this termand its use by the medical deviceregulations, ICH guidance, etc. Thisconfirmation can take many forms:physical inspection, operationaltesting, performance testing, as wellas other methods such as review ofa material certification document,software code inspection for con-formance to programming stan-dards, etc.

a. This verification could occur atany point in the overall lifecycleof, design, fabrication, installa-tion, pre-start-up, start-up, orinitial operation of the overallsystem or process.

b. This verification should occur atthe most appropriate point in theoverall lifecycle – as defined andjustified though the Quality RiskManagement (QRM) process.

c. This verification work may occurduring factory acceptance, siteacceptance testing, installation,or formal commissioning phasesof the project.

d. This verification work is per-formed under Good Engineering




Practice (GEP), and executed byappropriate Subject Matter Ex-perts (SME).

4. A common requirement of all of theregulatory references above is thatfacilities, equipment systems, andassociated automation are docu-mented and authorized as suitablefor the intended use. The determina-tion that systems are suitable fortheir intended use present a diffi-culty in ensuring that there is aclear understanding of what suit-ability means. Suitability for usecan be defined in many ways, andthere may be different possible de-sign solutions, which will achieve adesired result. We strongly recom-mend that suitability for use is notequivalent to meeting a particularengineering design specification.Instead, we propose that suitabilityfor use be defined in terms of abilityto meet product and process require-ments necessary to manufacture aquality product, and ability to pro-vide sufficient control of risks to thepatient (this is what ASTM E2500has as its approach). Suitability foruse is therefore linked to:

a. A specific manufacturing processand product (or class of prod-ucts).

b. It is based on knowledge of theprocess and an analysis of risk tothe patient.

Qualification should mean that equip-ment has been found to be suitable forits intended use, based on the designcriteria (process requirements orequivalent) and the verification workthat was performed throughout thedelivery process, in particular includ-ing that which occurred during thecommissioning phase. Qualified nolonger means the completion of an IQ/OQ/PQ protocol as traditionally for-mulated – leveraged or otherwise, butis instead a state or condition of certi-fied suitability for use. Graphically,these three terms relate as illustratedin Figure 1.

The question is, can we adopt this

use of the word qualification withoutinvoking the non-value added prac-tices of the past? Can people get pastthe habit of creating separate IQ, OQ,and PQ protocols, and instead adoptthe idea that qualification is a “state”achieved as shown above? Or shouldwe adopt a different definition? Or dropthe use of the term altogether (as ASTMhas done), and leave it to the operatingcompany to explain how their programnonetheless meets the intent of EUand other global regulations?

For those who feel the need to havesome form of qualification documen-tation, the determination that equip-ment is suitable for its intended use

could be equivalent to either the Ac-ceptance and Release phase describedin ASTM E2500 or to the Qualifica-tion Summary Report phase currentlyundertaken in many traditional com-pliance programs, as illustrated inFigure 2.

Both these representations and therelationship of the terminology meetthe intent of all regulations for demon-strating Suitability for Use and do notpresent non-compliance concerns withinthe ICH or EU regulated regions. De-sign qualification also can fit into thisscheme should that be desired. There-fore, the idea that suitability for use canbe determined based on patient risk

Figure 2. Verification, commissioning, and qualification as distinct steps. Note that in thismodel, there is no extra qualification-related field work or documentation when compared tothe ASTM E2500 process. It is simply a repackaging of the acceptance and release phasefor those organizations that require a document labeled “qualification protocol/ report.”

Figure 1. Relationships between the concept of verification, commissioning, andqualification.




The authors invite readers torespond to this discussion,

either throughthe ISPE Commissioning

and Qualification Communityof Practice (C&Q COP)

discussion board throughwww.ISPE.org/cops

or via direct communication.

and process requirements is wellgrounded in EU regulations and ICHdocuments, and is supported by US regu-lations and guidance documents.

We leave it to industry to debatethese proposals; it is important that weachieve a consistent understanding andapplication of these terms. Once thedebate is complete, it is for ISPE toincorporate the results into upcomingBaseline® Guides.

About the AuthorsJohn Robert (Bob)Adamson is Managerof PharmaceuticalCompliance and Vali-dation with FosterWheeler based inReading, UK. He has38 years of experience

in the pharmaceutical industry includ-ing positions with Foster Wheeler,Beecham, and Glaxo (now GSK), andexperience in R&D, production man-agement for APIs, and sterile manu-facturing. He has been responsible forthe start-up of new facilities and anumber of major refurbishments. He isa chartered chemist and chartered en-gineer. His experience includes com-pliance, regulatory topics including lat-est risk-based approaches, qualifica-tion and commissioning of: oral andsterile dosage forms, biotechnology andAPI facilities to meet EU, FDA, andMHLW requirements. He worked witha wide variety of clients and on occa-sions has represented them at FDAmeetings. Bob is Co-chair of the ISPECOP for C&Q, is actively involved inhis local affiliate, has written and pre-sented a number of papers, and con-tributed to a number of books. He waselected U.K. Fellow of the Year in 2003.He can be contacted by telephone: 44-1189133041 or by e-mail: [email protected].

Foster Wheeler Energy Ltd.,Shinfield Park, Reading, Berks RG29FW, United Kingdom.

Nuala Calnan is aPrincipal Consultantwith Project Manage-ment Group, Irelandand has more than 17years experience in thepharmaceutical in-dustry, with experi-

ence across the secondary pharmaceu-tical sector and the biopharmaceuticalsector. She is the current a member ofthe ISPE International Board of Direc-tors, Chair of the International Execu-tive Education Committee, a memberof the Steering Committee for the Com-missioning and Qualification Commu-nity of Practice (C&Q COP) and a mem-ber of the UK and Ireland Process Ana-lytical Technology (PAT) COP. She isactively involved in responding to therecent initiatives within the regula-tory environment and was a member ofthe Author Task Team which producedthe recent ASTM International Stan-dard on Commissioning and Qualifica-tion and is also a member of the teamcurrently updating the ISPE BaselineGuide for C&Q. She graduated in 1991with a Bachelor of Science in engineer-ing (BSc Eng) degree and achieved herMBA in 2002. She can be contacted bytelephone: +353-1-4040700 or by e-mail: [email protected].

PM Group, Kilakee House, BelgardSquare, Tallagh, Dublin 24, Ireland.

Robert E. Chew, PE,is President of Com-missioning Agents,Inc. He initiated theeffort to develop thisstandard, and was in-volved in the ASTME2500 writing effort

throughout its three year process. Chewhas a BS in Chemical Engineering andis a registered Professional Engineer.He can be contacted by telephone: +1-317-710-1530 or by e-mail: [email protected].

Commissioning Agents, Inc., 1515N. Girls School Rd., Indianapolis, Indi-ana 46214, USA.

Steven J. Wisniew-ski is Senior Associ-ate and Director ofCompliance for Inte-grated Project Services(IPS), a full-service en-gineering firm special-izing in the delivery of

technical complex projects, and thatoffers complete design/build, commis-sioning, validation, and FDA compli-ance services for the pharmaceutical,biotech, health care, and specialtymanufacturing industries. Wisniewskihas more then 30 years experience inthe pharmaceutical, biotech, and de-vice industries. Prior to joining IPS, hewas senior consultant for drug anddevice associates and has served insenior management roles at SterlingWinthrop and Bausch & Lomb. He hascompleted a wide variety of pharma-ceutical manufacturing, filling, andcritical support operations to majorR&D laboratories, facilities, and up-grades. He holds a BSME fromRensselaer Polytechnic Institute, is aMember of PDA, and an active Mem-ber of ISPE. He has served on the ISPEboard of directors since 1982, and waschairman of the board in 1991. Cur-rently, Wisniewski is chairman of theISPE Community of Practice for Com-missioning and Qualification, is on theISPE task teams that developed theASTM E2500 Verification Standard,and is working on the revised C&QBaseline® Guide. He can be contactedby telephone: +1-610-828-4090 or by e-mail: [email protected].

IPS, 2001 Joshua Rd., Lafayette Hill,Pennsylvania 19444, USA.


ISPE Update


Concludes on page 7.

French Medicines Agency Voices Support for ISPE’s Risk-MaPPGuide Principlesby Rochelle Runas, ISPE Technical Writer

ICH Officials Announce Adoption of Q10 atISPE’s Washington Conferenceby Rochelle Runas, ISPE Technical Writer

The final draft of the much anticipated ICH GuidelineQ10 (Pharmaceutical Quality Systems) has been adoptedby the ICH Steering Committee, ICH officials announced

5 June at the ISPE 2008 Washington Conference: Engineer-ing Regulatory Compliance, Washington, D.C., USA.

During the ISPE educational session, “Regulatory Per-spectives on Hot Topics, Regulatory Trends, and Observa-tions,” members of the ICH Q10 Expert Working Group,including Joe Famulare, Deputy Director, Center for DrugEvaluation and Research (CEDER) Office of Compliance, USFDA; Moheb Nasr, Office of New Drug Quality Assessment(ONDQA), CEDER, US FDA; and Robert Baum, ExecutiveDirector, Pfizer, Inc., delivered the news live via teleconfer-ence from the ICH Steering Committee Meeting in Portland,Oregon, USA.

“We have successfully reached Step 4,” Famulare told 70ISPE education delegates. “There is a consensus, we’ve signedoff on it, and it is ready for publication.”

Famulare, Nasr, and Baum gave the latest update fromthe meeting, including highlights of the new document. Q10includes a revised section on transfer of ownership of prod-ucts to include additional information addressing quality.

The document also includes a robust section on outsourcingwith the key message that ultimate responsibility falls on themanufacturer.

Q10 describes a model for an effective quality manage-ment system for the pharmaceutical industry that can beimplemented throughout the different stages of a productlifecycle. Implementation should facilitate innovation andcontinual improvement and strengthen the link betweenpharmaceutical development and manufacturing activities.

Having reached Step 4, Q10 moves immediately to thefinal step (Step 5: Implementation) of the process, which isregulatory implementation. This step is carried out accordingto the same national/regional procedures that apply to otherregional regulatory guidelines and requirements, in the EU,Japan, and the US.

ISPE’s Product Quality Lifecycle Implementation (PQLI)initiative is helping industry define areas where they canprovide the technical framework for implementation of QbDin regulatory submissions, and turn the ICH Guidelines Q8,Q9, and Q10 into a cross-functional and practical reality.

More information on Q10 is expected to be posted soon onthe ICH Web site, www.ich.org.

Representatives from the French Medicines Agency(AFSSAPS) have voiced support for risk-management prin-ciples that underpin ISPE’s upcoming ISPE Baseline® GuideVolume 10: Risk-Based Manufacture of Pharmaceutical Prod-ucts (Risk-MaPP).

“AFSSAPS is in favor of a risk-management method,” saidVincent Gazin, Head of the Clinical Toxicology Unit,AFSSAPS, at the ISPE 2008 Washington Conference: Engi-neering Regulatory Compliance held June 2 – 5 in Washing-ton, D.C., USA. “We agree to have a scientific discussion morethan an interpretation of regulatory text.”

The need for dedicated facilities for the manufacture ofcertain classes of high hazard compounds has been thesubject of much debate in recent years. The rationale forseparating certain compounds has not always been clear andregulators in the US and Europe are working on revisions toparts of their Good Manufacturing Practice (GMP) guidelinesaddressing this issue.

The ISPE Risk-MaPP Baseline® Guide shows how the

rational use of a science-based risk assessment process can beused to assess compounds, on a case-by-case basis, to supportmanufacturing strategies that allow for the use of multi-product facilities.

In January the EMEA released a “State of the Status of theRevision of Chapter 5 of the GMP Guide Concerning “Dedi-cated Facilities,” indicating that the EMEA will provide a listof products that mandates “dedicated facilities.” The develop-ment of this list is pending input from toxicological/pharma-cological experts, including Gazin.

The EMEA anticipates that a text will be submitted to theEuropean Commission at the end of 2008 or the beginning of2009 for public consultation.

At the ISPE educational session, “Risk-MaPP: Applica-tion of the new ISPE Baseline® Guide Volume 10: Risk-BasedManufacture of Pharmaceutical Products (Risk-MaPP),” Gazinand Nicolas Chauviere-Courcol, Mechanical Engineer, gavea presentation on how risk management principles are ap-plied in the toxicological unit of AFSSAPS.

Reprinted from





ISPE Update


JPI Publishes Ground-breaking Scientific Papers onReshaping Pharmaceutical Quality

The June 2008 issue of the Journal of PharmaceuticalInnovation (JPI) has published the first scientific pa-pers outlining the progress made on ISPE’s Product

Quality Lifecycle Implementation (PQLI) initiative.Written by subject matter experts representing the global

pharmaceutical manufacturing industry, these papers presentpreliminary practical scientific and technological approachesto implementing ICH documents that address Pharmaceuti-cal Development (Q8 and Q8(R)), Quality Risk Management(Q9), and Pharmaceutical Quality Systems (Q10).

The June issue is available in print and with Open Accesson SpringerLink (available at http://www.springer.com/jour-nal/12247) with the possibility to comment.

The Product Quality Lifecycle Implementation (PQLI)initiative was launched by ISPE in June 2007 to help indus-try find practical technical solutions to the challenges ofimplementing guidelines put forth by the ICH. The first threeTask Teams formed focused on Criticality, Design Space andControl Strategy, and how these areas are linked; a LegacyProducts Task team has also been formed as the fourthtopical area.

Through PQLI, ISPE isproviding technical frame-works to facilitate theimplementation of Q8, Q9,and Q10 for new productsand processes, as well asfor existing approved prod-ucts which could benefit.PQLI will provide betterunderstanding of Qualityby Design (QbD) applied tonew products and pro-cesses, and is developingcross-functional tools val-ued by both the Industryand Regulatory Authorities worldwide.

PQLI is projected to be at least a five-year initiative thathas started with highly interactive fact-gathering sessionsheld in the USA and Europe. Working groups will continue tocollect and process information for distribution as whitepapers, articles to be published in ISPE’s Journal of Pharma-

Continued on page 3.

Employers and Industry ProfessionalsRealize Value of the CPIPSM

As acceptance for the Certified Pharmaceutical Industry ProfessionalSM (CPIPSM)credential gains momentum worldwide, industry professionals and employers

remark on its significance as a powerful tool for professional development and topjob performance.

“We intend to strategically use the CPIP credential now and in the future toqualify our team and support their on-going professional development,” saidDonovan Wearne, CEO SeerPharma Pty., Ltd.

“Our company has a dual career ladder, allowing technical staff to advance tolevels that were once only open to individuals on a management track. I challengedthose aspiring to Senior Principal Consultant levels to pursue the CPIP credentialas a sign of their commitment to being recognized by our industry as a professionalwith a high proficiency level,” said Ken Ewan, Director, Corporate Engineering,Amgen. “The CPIP program’s seven knowledge areas allow our managers to identifythe focus each technical staff member needs to advance their careers.”

“Our industry benefits from employees certified in diverse knowledge, and withthe ability to apply this knowledge across all segments of our industry,” said AliAfnan, PhD, U.S. FDA. “In addition, it allows employers to be able to recognize topperformers, attain better product quality, industry-wide recognition, and commit-ment to innovation. Certified employees will become more valuable as team leaders,develop keener awareness, and perform their job more efficiently.”

Visit www.ISPE-PCC.org for up-to-the-minute CPIP news, exam information,test dates, and more.


ISPE Update


ISPE Singapore Conference and Region’s IndustryRemains Strong

JPI Publishes Ground-breaking Scientific Papers...Continued from page 2.

ceutical Innovation and Pharmaceutical Engineering Maga-zine, leading to detailed technical documents, and trainingprograms that will be produced by ISPE for the industryworldwide.

With the publication of these articles, the ISPE PQLI TaskTeams are seeking additional feedback prior to developingtheir respective positions into technical documents.

The Criticality article describes a mechanism for catego-rizing and delineating criticality for quality attributes, vari-ables, material attributes and process parameters in accor-

dance with a risk-based approach reflective of QbD principlesarticulated in ICH Q8R. The article introduces the adoptionof a Criticality Analysis Decision Tree to categorize criticalityrelative to a variable’s impact to quality and delineate levelsof criticality with respect to relative risk.

Design Space discussions considered the linkage of thepatient experience with product quality. It also focused onhow risk assessment methodologies integrate with processdesign principles, provided perspective on selection of mecha-nistic versus empirical approaches, and clarified how they

More than 300 industry professionals from the Singaporeregion and beyond attended the 8th edition of the ISPE

Singapore Conference 1 – 3 June at SUNTEC, Singapore.“Enhancing Regional Pharmaceutical Manufacturing

Excellence,” co-organized by ISPE and Reed Exhibitions,addressed the latest regulatory, technological, and practi-cal issues facing both multi-national and regional phar-maceutical manufacturers in API, secondary, and biotechmanufacturing, through workshops and various sessions.

More than 35 international regional speakers drawnfrom the US FDA, WHO, Singapore Health Sciences Au-thority (HSA), and the pharmaceutical industry sharedtheir insights and views on various issues.

In addition, many of the delegates had the opportunityto visit international pharmaceutical manufacturing fa-cilities based in Singapore. The Interphex Asia 2008exhibition was also held 2 – 3 at the same venue, inconjunction with the conference.

The strong attendance at the ISPE Singapore Confer-ence is indicative of the speed of pharmaceutical produc-

tion in Singapore. Despite competition from emergingcheaper manufacturing facilities in markets such as Chinaand India, multinational corporations are already invest-ing some $1.3 billion US dollars in plants in Singapore,and pharmaceutical manufacturing output doubled lastyear, according to a Channel News Asia report.

The pharmaceutical manufacturing industry inSingapore is expected to stay strong for the next five years,according to Gus Abdallah, Past President, ISPE SingaporeAffiliate. “Over the next five years, I would say you arelooking at a similar in-crease, just with the num-ber of pharmaceutical com-panies coming in here, andspeaking with pharmaceu-tical companies, they re-ally intend to push the out-put from Singapore, so Isee a very health increase,”Abdallah said in an inter-view with Channel NewsAsia.

A good turnout of delegates at the conference.

Industry professionals engaged in a workshop session duringthe ISPE Singapore Conference.

Gus Abdallah, Past President,ISPE Singapore Affiliate, onChannel News Asia.



ISPE Update


JPI Publishes Ground-breaking Scientific Papers...Continued from page 3.

may be applied to legacy products, and biotech products. Theteam also discussed a number of useful methods for depictingdesign space. The team recognizes that organizations maychoose different, scientifically defensible means to arrive atdesign space.

The Control Strategy team has proposed a Model processto enable a clear logic to be used on how a Control Strategydifferentiates between patient and business requirements,as well as showing the linkage from Critical Quality At-tributes, e.g. via Critical Process Parameters, to individualcontrols such as analytical, PAT, engineering, procedural orother controls. The Model illustrates how the Control Strat-egy embraces ICH requirements (product and systems). Itwill also provide a discussion bridge between disciplines suchas development scientists and controls engineers.

The Legacy Products team has started work and willproduce a paper later in 2008 in JPI. The team is consideringhow to derive business benefits by reviewing knowledgeabout a product and/or process and proposing opportunitiesfor flexibility in a post approval regulatory application for anapproved product. A suggested workflow process will beproduced and supported by case studies.

“The publication of these papers is a milestone event as it

will bring together an industry view of a risk- and science-based design approach for pharmaceuticals,” said James C.Spavins, Vice President, Global CMC, Pfizer. “The use of risk-based analyses to determine design constraints and thendetermine appropriate controls is a foundational process forthe advancement of science and technology - it is time forpharmaceutical professionals to have an aligned view.”

The Journal of Pharmaceutical Innovation (JPI) is aninternational, multidisciplinary peer-reviewed scientific jour-nal dedicated to publishing high quality papers emphasizinginnovative research and applied technologies within the phar-maceutical and biotechnology industries. JPI’s goal is to be thepremier communication vehicle for the critical body of knowl-edge that is needed for scientific evolution and technicalinnovation. The journal brings together in a single source themost exciting work from a variety of fields - from R&D tomarket. JPI publishes Perspectives, Case Studies, ResearchLetters, Research Articles, and Reviews in the following cat-egories: materials science; process design, optimization, auto-mation, and control; product design; facilities; informationmanagement; regulatory policy and strategy; supply chaindevelopments; and education and professional development.JPI is published by ISPE in collaboration with Springer.

ISPE Manchester Conference to Focus on Product andProcess Quality

The ISPE Manchester Conference on Product and Pro-cess Quality will be held 15 – 18 September at TheLowry Hotel, Manchester, UK. There will be six semi-

nars on:

• Innovations in OSD Processing• Barrier Isolation Technology• Science- and Risk-based Approach for C&Q: Application of

the New Baseline® Guide: Installation and Verification inSupport of ASTM E2500

• Applying GAMP® 5 Risk-based Approaches in Practice• Investigational Products – Delivering Quality by Design• PAT Data Management – Impact on Business Processes

Used for Improving Product and Process Quality

The following are descriptions of each seminar.

Innovations in OSD ProcessingLearn about new technologies in OSD manufacturing, inno-vative Quality by Design approaches to product and processdesign, innovation in other processing industries. The two-day seminar will feature sessions on:

• Nano-chemical Approaches to Controlled and TargetedDrug Delivery

• Lyopan – A Lyophilisation Technology for producing Fast-melting Tablets

• Continuous Processing of OSD• Scientific Approach to Process Design• DEM Modelling• Real Time Statistical Process Control• QbD and ISPE OSD Baseline® Guide

There will also be case studies on design Space, real-timerelease, and real time performance management in the pro-cess industry.

Barrier Isolation TechnologyThrough technology updates, case studies, discussion groupsand industry comment, this seminar will present the latestdevelopments in barrier isolation technology. Vendors willbriefly introduce the latest and most innovative technologies,while case studies will focus on the implementation of re-cently developed isolators.



ISPE Update


ISPE Manchester Conference to Focus on Product and Process QualityContinued from page 4.

Sessions will include sterile transfer, electron beam tech-nology, clean-in-place applications, and a range of topics suchas the influence of humidity on concentration and the decon-tamination trusses.

Science- and Risk-based Approach for C&Q:Application of the New Baseline® Guide: Installationand Verification in Support of ASTM E2500This seminar offers an exclusive opportunity to receive anoverview of the Baseline® Pharmaceutical Engineering Guidefor Installation and Verification – An Implementation Guidein Support of Science and Risk-based approaches for C&Q andIn Support of ASTM E2500, which is currently in development.

The new guide will coexist with the current C&Q BaselineGuide and provides guidance on the implementation of risk-based approaches and verification of a system under theASTM Standard E2500 – Standard for Specification, Design,and Verification of Pharmaceutical and BiopharmaceuticalManufacturing Systems and Equipment. The seminar alsofocuses on how the new Guide incorporates concepts fromICH Q8 and Q9 guidance, and provides procedures to improvedelivery of regulated manufacturing capacity.

This is a topic of great interest to the industry as the ASTME2500 standard was published in 2007 and delegates will beinterested to understand more about ISPE’s response. It isalso of major significance due to the ongoing debate as towhether the previous guide and methodology are to be fol-lowed or whether the industry should move to this newparadigm.

Applying GAMP® 5 Risk-based Approaches in PracticeThis seminar will provide delegates with current thinkingand examples on how the risk-based approaches described inGAMP 5 may be applied in practice. The sessions will demon-strate that the specification, design, implementation, opera-tion, and subsequent retirement of a computerized systemrequire careful planning and organization within a struc-tured framework. If managed correctly, this not only ensurescompliance with regulatory requirements but also enablestechnological advance and encourages innovation.

The seminar will offer the first opportunities for theindustry to present and discuss the impact of GAMP 5 andhow it is being put into practice. It will also cover two piecesof legislation that are likely to appear in the near future – EUAnnex 11 and the Revised 21 CFR Part 11.

Investigational Products – Delivering Quality by DesignThis seminar will cover the future of clinical supplies, addressefficiencies and improvements in labelling processes, andinclude a regulatory update workshop on the proposed changesto Annex 13 (a key document for those in clinical supplies).

Supporting those working with investigational products(IP) and clinical trials, this seminar will work with delegatesto understand clinical trials regulations in Europe, anddevelop better ways of working in line with these regulations.Using case studies and real examples, the focus will be onsharing experiences from the wide range of companies in-volved in the manufacture, packaging and distribution ofinvestigational medicinal products. Through networkingevents, interactive workshops, and seminar presentationsled by key opinion leaders, including those from within ourindustry and clinical/hospital environments, this seminarprovides a unique and valuable forum to challenge existingpreconceptions, explore alternative approaches and to share“best practice” ideas.

PAT Data Management – Impact on BusinessProcesses Used for Improving Product and ProcessQualityOne of the challenges resulting from the PAT initiative iswhat to do with all of the additional data that is beinggenerated and how to use these data for quality and otherbusiness decisions. This session looks at the progress thathas been made in supporting the manufacturing processesutilizing PAT generated data.

It concentrates on the business challenges in implement-ing PAT systems to design, analyze, and control manufactur-ing operations to improve the processability and productquality. Case studies will help delegates understand howcompanies have utilized data management throughout theirdevelopment and manufacturing environment as well asdiscussing the challenges of integrating these data in thebusiness processes.

A full day will be devoted to PAT data management. A half-day will consider how PAT can be applied to make betterbusiness decisions. A further half-day will involve an interac-tive European Community of Practice (COP) meeting, whichwill give participants the opportunity to discuss PAT imple-mentation with equally minded scientists. It will also givethem valuable information which they can take back to theirlocal COP to work on PAT issues in more detail.

For more detailed information,visit www.ISPE.org/manchesterconference.


ISPE Update


Mark Your Calendar with these ISPE EventsAugust 20086 Greater Los Angeles Area Chapter, Social Regulatory Aspects of Clinical Research and Overview of FDA Good Clinical Practice

Guidelines, David Geffen School of Medicine, Irvine, California, USA8 Japan Affiliate, 17th SAM and GMP Meeting, Yamaguchi, Japan8 Puerto Rico Chapter, Site Tour and Training on “How to Conduct an Effective Investigation (CAPA),” Guayama, Puerto Rico,

USA12 – 14 Brazil Affiliate, GAMP® Forum Three-Day Event, Mercure Apartments, Sao Paulo, Brazil18 – 19 Argentina Affiliate, Course II: Water for Pharmaceutical Use and ISPE Baseline® Guide and Regulations, Laboratorios Rontag

Auditorium, Buenos Aires, Argentina21 Puerto Rico Chapter, Technology Showcase, San Juan, Puerto Rico, USA21 San Diego Chapter, Vendor Night, Hilton La Jolla Torrey Pines, La Jolla, California, USA22 San Diego Chapter, Golf Tournament, Twin Oaks Golf Course, San Marcos, California, USA27 Nordic Affiliate, Conference on Cleaning, Helsinki, Finland28 DACH Affiliate, Workshop on “New Technologies in Manufacturing Effervescent Tablets” and site visit at Hermes, Wolfsberg,

Austria28 – 30 INTERPHEX India 2008, HITEX Exhibition Centre, Hyderabad, India21 San Diego Chapter, Vendor Night, Theme: Football Tailgate Party, Hilton La Jolla Torrey Pines, La Jolla, California, USA22 San Diego Chapter, 11th Annual Golf Tournament, Twin Oaks Golf Course, San Marcos, California, USA

September 20082 UK Affiliate – Central Region, Visit to Superconducting Magnets Facility at Siemens Medical Amysham, Oxford, United Kingdom4 Puerto Rico Chapter, Full-Day Biotechnology Program, Guaynabo, Puerto Rico, USA9 San Diego Chapter, Padres versus Dodgers Game, San Diego, California, USA9 San Francisco/Bay Area Chapter, Commuter Conference on “Alternative Delivery and Contracting Methods – IPD, DB, DBOM,”

UCSF Mission Bay, San Francisco, California, USA11 Ireland Affiliate, UCB/Schwarz Plant Tour and Golf Outing, Shannon, Ireland12 DACH Affiliate, Workshop at Bosch on “Containment/Asept. Filling of Liquid Products,” Crailsheim, Germany15 – 18 ISPE Conference on Product and Process Quality, The Lowry Hotel, Manchester, United Kingdom16 Boston Area Chapter, Six Sigma Seminar, Genzyme Corporate Center, Cambridge, Massachusetts, USA16 Chesapeake Bay Area, Annual Golf Tournament, Whiskey Creek Golf Club, Ijamsville, Maryland, USA16 – 17 Great Lakes Chapter, Vendor Show and Education Program, Indianapolis, Indiana, USA18 Brazil Affiliate, One-Day Event on “HVAC in a Pharmaceutical Industry,” Mercure Apartments, Sao Paulo, Brazil18 Pacific Northwest Chapter, Vendor Night, Bellevue, Washington, USA19 Pacific Northwest Chapter, Golf Tournament, Echo Falls Golf Course, Echo Falls, Washington, USA19 Rocky Mountain Chapter, Annual Golf Tournament, Indian Peaks Golf Course, Lafeyette, Colorado, USA22 Argentina Affiliate, Workshop with Topics on “Plant Design, Construction, New Alternatives Manufacturing Facilities and

Regulations, Laboratorios Rontag Auditorium, Buenos Aires, Argentina22 – 25 ISPE New Jersey Classroom Training, Holiday Inn Somerset, Somerset, New Jersey, USA24 – 25 INTERPHEX Canada, including educational programming by the ISPE Central Canada Chapter, Palais des congres de Montreal,

Montreal, Quebec, Canada24 – 25 Spain Affiliate, Project Management Conference, Spain25 Italy Affiliate, Pharmaceutical Management Forum, Florence, Italy25 San Diego Chapter, Dinner Meeting – “Amylin Ohio Case Study,” San Diego, California, USA30 Nordic Affiliate, Conference on Project Management, Helsingborg, Sweden

October 20082 Greater Los Angeles Area Chapter, Golf Tournament, Strawberry Farms Golf Club, Irvine, California, USA6 – 9 ISPE Milan Classroom Training, Crowne Plaza Milan Linate, Milan, Italy7 – 8 Nordic Affiliate, Event on EuPAT3, Stockholm, Sweden8 UK Affiliate – Northwest Affiliate, Joint ISPE/IChemE Day Seminar on Sustainability, Science and Industry Museum,

Manchester, United Kingdom8 Boston Area Chapter, Annual Product Show, Gillette Stadium Clubhouse, Foxboro, Massachusetts, USA10 Puerto Rico Chapter, Site Tour and Training “Maintenance and Reliability – Predictive Maintenance,” Puerto Rico, USA10 – 11 ISPE China Conference, Organized in conjunction with the 13th China International Pharmaceutical Industry Exhibition (China

Pharm), Beijing, China16 Ireland Affiliate, Workshop/Seminar on “Regulatory Environment,” Crowne Plaza Dublin Airport Hotel, Dublin, Ireland16 South Central Chapter, Plant Tour and Dinner, Austin, Texas, USA16 San Diego Chapter, Full Day Risk Validation Class, Biogen Idec, San Diego, California, USA16 – 17 2008 Istanbul Classroom Training, Sheraton Istanbul Maslak Hotel, Istanbul, Turkey22 Italy Affiliate, Event Topic: ISPE Maintenance Baseline® Guide, Verona, Italy22 Nordic Affiliate, Conference on GAMP®, Oslo, Norway23 DACH Affiliate, Workshop at Bosch on “New Technologies for Powder Filling,” Waiblingen, Germany26 – 29 2008 ISPE Annual Meeting, Boca Raton Resort, Boca Raton, Florida, USA

Dates and Topics are subject to change.


ISPE Update


The BTEC Experienceby Jeff Odum, ISPE North American Education Advisor

The week of May 12th was a historical one for ISPE. Thatweek, ISPE and North Carolina State University held a

first-of-its-kind public training event for the biotech indus-try. Three courses were offered by ISPE and BTEC instruc-tors at the Golden Leaf Biomanufacturing Training andEducation Center (BTEC).

The BTEC training program offered the coursesBiopharmaceutical Manufacturing Facilities, Process Vali-dation in Biotechnology Manufacturing, and Getting theMost from your Bioreactor. The program included lecturesand hands-on activities at a state-of-the-art cGMP pilot plantfacility, the first commercial-scale bioprocess training centerin the United States. The intimate class sizes providedindividuals with an opportunity to share best practices andbenchmark their efforts against peers.

I had the pleasure of being teamed with Dr. MichaelFlickinger, the Director of BTEC, to deliver a course onBiopharmaceutical Manufacturing Facility Design, usingthe BTEC as a “working model” for implementation of manyof the design principles of the Biopharmaceutical Manufac-turing Facilities Baseline® Guide. The focus of the course was

simple: allow a small group of experienced industry profes-sionals the chance to test their knowledge via a unique designproblem.

A focus of the course was to use the BiopharmaceuticalManufacturing Facilities Baseline® Guide as a text for studyof how to address design issues faced by many companies. TheBTEC served as a great teaching tool, one that gave studentsthe opportunity to “roll up their sleeves” (as best you canwhen gowned up) and dive into the internal workings of amanufacturing operation. The students were divided intotwo teams and given a chance to redesign the facility based ona unique case study scenario.

And to make it interesting, a distinguished panel of fiveindustry professionals served as a selection jury for choosingthe best approach to meet the design criteria. My thanks toGlen Williams (Biolex), Ed George (Wyeth), Steve Errico(Eisai), John Wagner (Merck), and Mitch Lower (BiogenIDEC) for their participation and insights.

The experience was great. Having a working, commercial-scale teaching facility at your disposal allows you to do somany more things than you could do simply using a lectureformat. The students dove into the design problem head firstand did a fantastic job in not only coming together as a team,but in developing two sound design solutions for the caseproblem.

I’d like to extend my sincere thanks to BTEC and all of theinstructors for their expertise and dedication to this success-ful, ground-breaking training event.

BTEC training participants ready themselves for their hands-oncoursework.

...Risk-MaPP Guide PrinciplesContinued from page 1.

While AFSSAPS has not yet officially reviewed the Risk-MaPP Guide, Gazin’s comments during his presentationindicated his support for the Guide’s principles.

Gazin commented on the following text taken from theRisk-MaPP draft September 2007:

Pharmacological and toxicological descriptions (dose-response, no-observed-adverse-effect level (NOAEL) andADI) should be used to assess compounds instead ofhazard labels. Terms such as potent, cytotoxic, cyto-static, and other product class definitions tend to inducean emotional response that might imply that thesecompounds are always difficult to handle and requirethe highest level of control.

“Emotional responses should be rationalized,” said Gazin.“Toxicity depends on the quantity as well as on the durationand the route of exposure. NOAELs and usual toxicologicalreference values should be integrated in a risk managementprogram. But the choice of NOAELs and toxicological refer-ence values (such as ADI) could be different among toxicolo-gists.”

“The objective is to have harmonization between asses-sors,” said Gazin. “When we are talking about scientific data,it is easier to have harmonization.”


Classified Advertising


Architects, Engineers – Constructors

CRB Consulting Engineers, 7410 N.W.Tiffany Springs Pkwy., Suite 100, KansasCity, MO 64153. (816) 880-9800. See ourad in this issue.

EI Associates, 8 Ridgedale Ave., CedarKnolls, NJ 07927. (973) 775-7777. Seeour ad in this issue.

IPS – Integrated Project Services, 2001Joshua Rd., Lafayette Hill, PA 19444.(610) 828-4090. See our ad in this issue.

Parsons, 150 Federal St., Boston, MA02110. (617)-946-9400. See our ad inthis issue.

Bioreactors/Fermenters

Cleanroom Products/Services

AdvanceTec, 485 Southlake Blvd.,Southport Corporate Center, Richmond,VA 23236. (804) 378-1550. See our ad inthis issue.

AES Clean Technology, 422 Stump Rd.,Montgomeryville, PA 18936. (215) 393-6810. See our ad in this issue.

Dagard USA Corp., 1251 Avenue of theAmericas, 34th Floor, New York, NY10020. (212) 583-4928. See our ad in thisissue.

Employment Search Firms

Jim Crumpley & Associates, 1200 E.Woodhurst Dr., Bldg. B-400, Springfield,MO 65804. (417) 882-7555. See our ad inthis issue.

Filtration Products

MKS Instruments, 5330 Sterling Dr.,Boulder, CO 80301. (800) 345-1967. Seeour ad in this issue.

Siemens Water Technologies, 10Technology Dr., Lowell, MA 01851. (978)934-9349. See our ad in this issue.

Instrumentation

Hach Ultra Analytics, 5600 LindberghDr., Loveland, CO 80539. (970) 663-1377. See our ad in this issue.

Label Removal Equipment

Hurst Corp., Box 737, Devon, PA 19333.(610) 687-2404. See our ad in this issue.

Passivation andContract Cleaning Services

Active Chemical Corp., 4520 Old LincolnHwy., Oakford, PA 19053. (215) 676-1111. See our ad in this issue.

Astro Pak Corp., 270 E. Baker St., Suite100, Costa Mesa, CA 92626. (800) 743-5444. See our ad in this issue.

Passivation andContract Cleaning Services (cont.)

Cal-Chem Corp., 2102 Merced Ave., SouthEl Monte, CA 91733. (800) 444-6786.See our ad in this issue.

Spray Dryers

GEA Niro Pharma Systems, 9165Rumsey Rd., Columbia, MD 21045. Seeour ad in this issue.

Heinen Drying Inc., 1504 Grundy’s Ln.,Bristol, PA 19007. (215) 788-8196. Seeour ad in this issue.

Sterile Products Manufacturing

Sterilization Systems

Environmental Tectonics Corp., 125James Way, Southampton, PA 18966.(215) 957-9333. See our ad in this issue.

Tanks/Vessels

Eagle Stainless, 816 Nina Way,Warminster, PA 18974. (215) 957-9333.See our ad in this issue.

Used Machinery

Validation Services

Commissioning Agents, Inc., 1515 N.Girls School Rd., Indianapolis, IN 46214.(317) 710-1530. See our ad in this issue.

ProPharma Group, 10975 Benson Dr.,Suite 330, Overland Park, KS 66210;5235 Westview Dr., Suite 100, Frederick,MD 21703. (888) 242-0559. See our ad inthis issue.

Valves

Gemu GmbH & Co., Fritz-Mueller-Str. 6-8, D-74653 Ingelfingen, Germany. +497940123-0. See our ad in this issue.

Water Treatment

Siemens Water Technologies, 10Technology Dr., Lowell, MA 01851. (978)934-9349. See our ad in this issue.

Veolia Water Solutions & Technologies,Global Headquarters, L’Aquarène – 1,place Montgolfier, 94417 Saint-MauriceCedex, France, www.pharma.veoliawaterst.com, Email: [email protected]. See our ad in this issue.

Reprinted from




ISPE's Communities of Practice Q&A

JULY/AUGUST 2008 PHARMACEUTICAL ENGINEERING On-Line Exclusive 1www.ispe.org/PE_Online_Exclusive

©Cop

yrig

ht IS

PE 2

008

A sampling of interactive discussions taking place online amongcolleagues through ISPE’s Communities of Practice.

Your Questions, Your Answers

Exclusive On-Line Article




Recognizing the challenge of the phar-maceutical industry to achieve con-tinuous improvement and productiv-ity, ISPE has implemented Communi-

ties of Practice (COP) as an online forum toenable colleagues in specific disciplines to rap-idly exchange information and solutions toeveryday problems.

Through professional networking and peercollaboration, ISPE COPs produce discipline-specific content that deepens members’ knowl-edge and expertise, increases quality and con-tinuous improvement in the industry, andachieves ISPE’s core purpose of leading globalinnovation.

Currently, there are 17 active COPs in areasof pharmaceutical expertise such as ActivePharmaceutical Ingredients (API); Biotechnol-ogy (Biotech); Commissioning and Qualifica-tion (C&Q); Containment; Critical Utilities(CU); Disposables; Engineering StandardsBenchmarking; Good Automated Manufactur-ing Practice (GAMP®); Good Control Labora-tory Practices (GCLP); Heating, Ventilation,and Air Conditioning (HVAC); InvestigationalProducts (IP); Packaging; Process AnalyticalTechnology (PAT); Process/Product Develop-ment (PPD); Project Management (PM); Ster-ile Products Processing (SPP); and SustainableFacilities. Initiatives for the COP in the field ofOral Solid Dosage is in development.

By engaging in ISPE COPs and becomingactive in their communities, participants havethe ability to connect with like-minded profes-sionals through an interactive online commu-nity that offers global networking opportuni-ties and access to a discipline-specific body ofknowledge. As the following pages of this ar-ticle demonstrate, Members of the diverse COPsare actively participating in a continuous dis-cussion of ideas to solve everyday problems.They are frequently using COPs to validateindustrial measures, reference guidelines, con-vey best practices, develop technical documents,

and keep abreast of current issues and impor-tant trends and developments of the commu-nity discipline.

Joining one or multiple ISPE COPs is thefirst step to connecting with peers and col-leagues around the world. With the acceleratedgrowth of pharmaceutical domains such as bio-technology and generics, the demand for leanmanufacturing methodologies, and the extraor-dinary costs in producing drugs, COPs repre-sent an effective outlet for the dissemination ofvaluable content produced by and for ISPECOP Members. ISPE COPs continue to be aprogressive and efficient response tool in thecompetitive environment of pharmaceuticalengineering.

You do not have to be an ISPE member tojoin and participate in ISPE COPs; however,ISPE Members gain full access to allfunctionalities of the COPs.

To join and become active in your ISPECommunities of Practice, visit www.ISPE.org/cops.

Active PharmaceuticalIngredients (API) COP

Validated Status of a ManufacturingProcess

Q Typically, when a process is validated,the passrate for the number of batches

manufactured per year should be high (i.e.,maybe >90%). What should be the passrate forbatches manufactured in order to consider aprocess in a validated status?

A There is more than one area to be consid-ered as part of a periodic validation re-

view to deem the process validated. One ofthem is the product quality review includinganalytical results, process capability, trends,out of specifications, reworks, yield etc.

This data should be reviewed to determine ifa change in the validation status is indicated by


2 PHARMACEUTICAL ENGINEERING On-Line Exclusive JULY/AUGUST 2008 www.ispe.org/PE_Online_Exclusive

©Copyright IS

PE 2008

trends in any of the above. In addition,the trended analytical data in the Prod-uct review should be compared withthe analytical data from the most re-cent PQ exercise. If a statistically sig-nificant shift is seen (even if the resultremains within specification) then aninvestigation into the probable rootcause must be undertaken in accor-dance with the Investigation of NonConformances and Deviations proce-dure. If the investigation determinesthat the process can no longer be con-sidered validated, then a revalidationexercise should be executed.

A A batch failure rate of 10% wouldsuggest that the process was out

of control. It is likely that a trend couldbe found in the 10% of failures: all lateat night, all just after start-up, etc.That trend would be the part of theprocess that is out of control.

A A 10% failure rate would indeedbe very high. The question is “Is

there an acceptable failure rate or con-versely an acceptable passrate” thatwould deem the process to be in avalidated status. In a nutshell, is therea quantifiable number (in terms of %passrate) that the community recog-nized to say that the process is vali-dated or out of control?

A As we all know, some processesare more robust than others. A

robust process may be out of controlwhen <1% failures occur. A new pro-cess for an innovative drug just beinglaunched may still need optimizationso a much higher failure rate may beseen in Year One, but the companywould want to improve upon that an-nually. Perhaps the Six-Sigma ap-proach to failure would be a usefultarget for processes.

Biotechnology(Biotech) COP

Emergency Showers in ClassA/B Area

Q I am involved in a discussionwith our site safety personnel

about installing an emergency showerand eyewash in Class A/B (Class 100)

areas in a pilot plant facility. My un-derstanding is that the presence of theshower/eyewash would increase therisk of microbial failures in the envi-ronmental monitoring of the area. Noone I have talked to has seen any show-ers/eyewashes in these areas in otherfacilities. Under what circumstancesare there showers/eyewashes in ClassA/B (Class 100) areas? Can someoneprovide me with references from regu-latory guidelines on this issue?

A Grade A/Class 100 areas are typi-cally kept small due to the high

capital and operating costs, and thechallenges of maintaining that level ofclassification. I am not aware of any-one installing an emergency shower oreye wash in this type of area. Grade A/Class 100 areas usually have a sur-rounding or background area of GradeB or Class 10,000. Emergency showersand eye washes have been installed inGrade B/Class 10,000 when requiredfor operator safety. You may be able toinstall these safety systems just out-side of the Grade B/Class 10,000 area,but they must be readily accessible inan emergency situation. ANSI Stan-dard Z358.1-2004 provides guidelinesfor safety equipment fixtures and in-stallation practices. I would first con-firm that the operations pose a safetyhazard that requires installation ofthese safety systems.

Commissioning andQualification (C&Q) COP

Cleanroom vs. EnvironmentallyControlled Room

Q This may not be the correct COPfor this topic, but has anyone else

been asked to stop using the termcleanroom and instead use the phraseEnvironmentally Controlled Room? Inmy case, I am referring to ISO 7. Upuntil the beginning of this year, weused the term cleanroom as defined inISO 14644-1. Just recently we havebeen asked to change to environmen-tally controlled rooms, but I have yet tofind the rationale for the change.

A Actually, we use the term“cleanroom” environments up to

and including ISO 8.

A I will try to explain what I under-stand about these two terms. I

was also asked to change the term andhere is the explanation:

ISO 14644-1 and 2 refer to the par-ticle size and amount of it in the air orspace while the controlled environmentgoes beyond; we have to continuouslymonitor the microbial growth and rout-ing of the cleaning agents, but also wehave to monitor the temperature, hu-midity, differential pressure, and airflows to keep the environment undercontrol and avoid cross contamination.If you read USP chapter 1116 Micro-biological Evaluation of Clean Roomand Other Controlled Environments,there is an explanation of these twoterms. The difference of Clean Roomand Environmentally controlled roomsreside in the amount of controls you setfor that particular area. You may havea room that is Class ISO 8 where youdon’t have to control cleaning as youneed to control it in an ISO Class 8room where you have a critical step ofyour process. I will say that the differ-ence resides in the controls that youneed to establish due to your processsteps.

Containment COP

Segregated vs. Stand AlonePotent Facility

Q Does anyone know of any currentor proposed regulations in North

America or EU which require/proposea stand alone facility for manufactureof GMP potent and/or cytotoxic prod-ucts (oral solid dosage forms)? At onepoint the FDA was talking about thisand I hear rumors from my EU friendsfrom time to time.

A First, what is the definition of po-tent? Industry cannot agree and

the regulators do not have a definition.The current trend is to use risk

assessment to determine the need forstand alone/segregated facilities. In-dustry actually needs to come up witha common definition for segregated andstand alone facilities!

I would suggest you attend the 2 – 5



©Cop

yrig

ht IS

PE 2

008

June 2008 Containment TechnologiesForum in DC as Edwin Melendez, theFDA “expert” in “potent” compounds,will be presenting the FDA’s views onrisk-based approaches. Last yearEdwin answered many questions in acasual Q&A at the end of the session.He clarified many points, like dedi-cated facility does not necessarily meana separate facility to the FDA. I believesome of that Q&A is available on thissite.

If you look in the Community News/Community Files portion of this siteyou will see that the EMEA are cur-rently evaluating which compounds tolist as requiring dedicated facilities.The issue is on hold until Toxicologicalstudies are performed. During the JuneRisk-MaPP session in DC, VincentGazin, Toxicologist from AFSSAPS, willbe presenting on the status of thesestudies... come hear first hand what ishappening and you may even be able tosway his opinion!

Hope to see you in DC; I really thinkyou will get more than this questionanswered!

A I don’t recall seeing you at theDC conference, so I wanted to let

you know that ISPE recorded all thesessions and is selling all the proceed-ings (the CD will have a sound file anda PDF file of the presentations for all10 sessions), so if you are interested,please contact ISPE.

For your information, the next con-tainment technology session will be 2 –5 March 2009 in Tampa.

Critical Utilities (CU) COP

Hydrocarbon in Compress Air

Q We are looking for a method toknow the hydrocarbon content in

compress air. Do you have any refer-ence?

A There are tubes that you hook upto your compressed air. Once you

run a certain quantity through thetube you can tell the concentration ofhydrocarbon by the color change. Onesupplier is Draeger and they have aquality article I came across that mightbe useful. http://www.draeger.com/ST/

internet/pdf/CS/en/DraegerReview/DR94/DR94_article_4.pdf

A Are you looking for levels in mg/m3 or in ppm? The Draeger tubes

will give the levels in mg/m3 and thenmust be converted to ppm if required.The problem with Draeger tubes isthat it is hard to get an accurate con-centration level. Also, you must knowthe type of oil you are testing for. Thetype can have a direct effect on detect-ability of the Draeger tube.

A Which ISO 8573 Class shouldthe compress air comply for a use

in pharmaceutical facilities (sterile andnon sterile areas) (maximum oil con-tent).

A ISO 8573 part 1 identifies classesof compressed air with allowable

levels of contaminants. Some wouldsuggest a Class 2 or better for hydro-carbons. As far as testing for hydrocar-bons, ISO 8573 parts 2 and 5 are thetesting methods.

At the Annual Meeting at ISPE whenI was presenting on this, one personsuggested that if a sterile filter is 0.2micron, why should the filtration needto be any greater than this? So thisappears to be the standard his com-pany was using. Most coalescing filterswill offer a 0.01 mg/m3 rating basis70°F temp and clean. Some would sayto minimize risk; you should have themost stringent Class, which is Classzero, so this is up to your risk tolerance.Oil Free compressors do not add hydro-carbons to the air and are the leastrisk, while oil injected compressors needclean up equipment. Intake is anotherconcern. Make sure your compressor isnot taking in hydrocarbons as even oilfree compressors put out what theytake in (concerns would be powder lu-bricants used in the facility that getairborne or intakes near loading dockswhich will absorb exhaust fumes).

Disposables COP

Retrofitting Disposables

Q Please share your experiences/opinions during retrofitting

disposables in existing SS facilities.

What are the challenges involved in it?

A The most challenging issue iscreating appropriate interfaces

from single use items to existing stain-less items, especially for aseptic sys-tems, if that is your context. There areseveral ways of doing it. We commonlyset up the stainless systems with theinterfaces ready to go, called “pigtails,”to which we can perform tubing weldsor use an aseptic single use connector.We also use different single use SIPconnectors at standard SIP ports. De-pending on the room classification andbioburden requirements of the processone can simply connect using a plastictriclamp connector on the disposableitem.

Engineering StandardsBenchmarking COP

Global Approach

Q How does your company addressthe design and construction of

facilities globally when facing the dif-ferent international regulations?

a. Controlled environments (i.e. par-ticulate control, viable & non viablemonitoring, temperature and hu-midity controls)?

b. Environmental, health and safetyissues (i.e. local codes vs. corporaterequirements)?

Respond to this question by joiningthis newly developed COP. Questionsand answers are welcome under thesection “Community Discussions,” onthe Engineering Standards Bench-marking COP site.

GAMP® COP

Infrastructure Qualification

Q Some years ago there was a lot ofdiscussion about infrastructure

qualification. An ISPE GAMP® GoodPractice Guide (“IT Infrastructure Con-trol and Compliance”) was published,giving great guidance on how to achievea qualified platform.

Since then I haven’t heard muchabout infrastructure qualification. Iwas just wondering if any of you have:



©Copyright IS

PE 2008

• experienced that infrastructurequalification is a hot topic in yourcorporation/among your customers

• heard of inspections where infra-structure qualification was/becamean issue

• experience implementing the GoodPractice Guide and wants to sharesome of your key learning on thistopic

A Since the IT Infrastructure SIGwas reformed and following the

publication of the Good Practice Guide,the approach to IT Infrastructure quali-fication has become more pragmatic.

While this is certainly still an issuethat needs to be addressed, most regu-latory agencies see this as a relativelylow risk area and now recognize that if(to paraphrase GAMP® 5) computer-ized systems can not represent a higherrisk, that the associated processes andproducts, the associated IT infrastruc-ture is even less of a risk.

Some of the key regulatory inspec-tions in this area (1999 – 2001) whichare often quoted focused principally oninfrastructure.

Regulatory agencies generally ac-cept that this was not/is not a usefulapproach. Typically nowadays Infra-structure is only cited as an issue dur-ing inspections if the lack of qualifica-tion/control is reasonably likely to leadto risk to product quality or patientsafety.

For most companies IT Infrastruc-ture is not the hot topic that it wasbecause of this more pragmatic ap-proach to inspections and because mostreputable companies have addressedthe topic (or have started to).

Having said all of that (and to ad-dress your second point) where it doesstill represent a risk to product qualityor patient safety it will be cited duringan inspection and not only by the FDA.There are a number of occasions whereEuropean inspectors have included ITinfrastructure findings as part of theirobservations.

My experience with using the GoodPractice Guide (and I should declarehere that I was a member of the SIG) isthat it provides a sound and pragmaticapproach.

However, some smaller companieshave struggled with some of the ex-amples and guidance given because toa significant extent it was written with‘Big Pharma’ organizations in mindand many smaller companies struggleto apply the concepts when they onlyhave three people in their IT depart-ments. The concepts are however soundand when applied properly the guid-ance can be scaled to organizations ofany size.

Another area where people seem tostruggle is taking a risk-based approachto infrastructure qualification and likeGAMP® 4, the Good Practice Guide isprobably a little light in that area. Thiscan result in a ‘one-size-fits-all’ ap-proach to Infrastructure Qualificationwhich is fine for most areas, but issometimes too much or too little forspecific components.

However, the expanded Risk Man-agement section to GAMP® 5 can beapplied to the principles of Infrastruc-ture Qualification and my experienceis that a risk-based approach can besuccessfully used to scale qualificationand control activities to those areaswhere the risk likelihood is greater(e.g. when using ‘novel’ infrastructureor using infrastructure for a purposenot intended by the manufacturer) orwhere the probability of detection islow (e.g. where there is no resilienceand no in-built diagnostics or perfor-mance monitoring).

Technology moves on and manypeople are struggling with how to applythe principles to technology not ad-dressed in detail in the Good PracticeGuide, i.e., looking towardsvirtualization (e.g. VMWare) and Ser-vice Oriented Architecture/Middle-ware. Although not addressed in detailin the Good Practice Guide the ‘hierar-chical’ layer and component model canbe extended to additional dimensionsand allows a pragmatic approach to bedeveloped for the appropriate qualifica-tion of newer technology not explicitlyaddressed in the Good Practice Guide.

To summarize, the pragmatism in-herent within the Good Practice Guidehas helped to define a sensible ap-proach to Infrastructure Qualificationand has gone a long way to making this

less of a hot topic. Although they needthinking about in some circumstancesthe principles of the Good PracticeGuide are still sound.

A Thank you for this very compre-hensive statement! I am aware

about the concerns of smaller organi-zations complaining that the GPG isonly achievable for large (and rich)companies. I do not really agree withthis concern because the GPG is notprescriptive regarding the number andthe complexity of required activities.

GPG objectives are to define a frame-work and to provide recommendationregarding how to establish compliantIT infrastructures and how to main-tain the controlled state of them. Weshould consider two main drivers forkeeping IT infrastructure under con-trol:

• Impact on the Patient’s health• Impact on the business capability

(which can also impact the Patient’shealth)

Having reliable IT infrastructure forGCP reasons is – in my humble opinion– a not negotiable requirement. Hav-ing reliable IT infrastructure is part ofGood Business Practice and is alsopart of the SOx scope.

The implementation of PAT is oftenvery challenging for the IT infrastruc-ture; various systems (process controlas well as multiple information sys-tems) have to be interconnected andthey must exchange numerous infor-mation in real time. The deployment ofchromatography data systems requiresalso a reliable IT infrastructure, allow-ing laboratory equipments to commu-nicate accurately with servers. Thedeployment of electronic lab-journals,the use of PKI-based electronic signa-tures, request reliable and secure ITinfrastructure.

All these facts plead for maintain-ing with rigor IT infrastructure undercontrol. All these facts are at least justas important as “pure” regulatory re-quirements.

In my experience as CSV and e-compliance auditor, I met once a small



©Cop

yrig

ht IS

PE 2

008

company (around 200 employees) witha very well structured and reliable ITsupport. Everything (every configura-tion item) was documented and main-tained under change control. Each ac-tivity was documented in log-book. Thereason for this exemplary behavior wasthe size of the team providing IT sup-port: 3 people big.

These three persons had to providesupport regarding the network andserver infrastructure for both office aswell as production areas. They had toensure the external connectivity of thecompany (internet, e-mail, web) andthe company worked 24/7.

Because somebody could becomesick during the colleague’s vacation, itwas vital for this IT team to make allrelevant information available to therest of the team in an accurate andconsistent manner: full cooperationwithout any “kingdom!”

The quality of the documentation,the rigor of activities (including changecontrol and business continuity plan-ning) were the main drivers for thisorganization; not the fear of a GxPinspector!!! Even if regulatory require-ments are important in our industry,we have to think again in terms ofcommon sense. The question is notwhat the Quality costs are!

Furthermore, the question is: Howexpensive is it to not have sufficientQuality (to not have reliable IT infra-structure) in place?

Please consider the Good PracticeGuide about IT Infrastructure andGAMP® 5 as a support for doing theright thing at the right time, because itis good for the business!

Good Control LaboratoryPractices (GCLP) COP

Scanning Electron Microscope

Q I’m researching the regulatoryrequirements for Scanning Elec-

tron Microscopes under GLP. So far,I’ve only found text suggesting thatMicroscopes can not be validated ortested for Robustness. Can any oneoffer insights or opinions on the use ofScanning Electron Microscopes in theLab?

A There is an article in the Journalof Measurement Science and

Technology, Vol. 17, pp. 2613-2622,(2006), titled, Metrology on ScanningElectronic Microscope: Theoretical De-velopments and Experimental Valida-tion (though I have not seen that).Youmay refer to this paper for some infor-mation you are seeking.

Heating, Ventilation, andAir Conditioning

(HVAC) COP

HEPA Filter Integrity TestFailure

Q Ceiling HEPA filters are testedfor integrity annually. If filter

fails integrity test, what is done withall product processed from the last fail-ure till to-day. On testing particulatematter the test passed.

A Your question shows the need forcontinuous monitoring. The only

action can be: to sequester the mate-rial produced since the last good read-ing. If the last good reading was threemonths ago then all the product pro-duced since that date is suspect andneeds to be QCed and tested. If most ofthat product has been shipped, thenonly a recall is possible and not prob-able. If the product was passed by QCbefore shipment, then no foul; no dam-age.

The issue is the following: You needreal-time, near-real time, or high fre-quency monitoring to help determinethe health of the environment at alltimes. High frequency monitoring canhelp determine issues long before speci-fication excursions.

A I recommend to look into the re-lease product SOP and QA/QC

batch criteria or document related tothat, there should be a contingencyplan for this kind of event, if there isnone then the Quality system needs tobe reviewed and a plan needs to be inplace (e.g. release criteria and criticalparameters to release product SOPneeds to be in place). If the HEPA filterintegrity test is a criteria to releaseproduct, then you should follow therelease product SOP. However if it is,

consider just an additional test, then asystem has to be implemented to avoidthis situation to happen again.

The situation that I see in here isthat the HEPA filter integrity test SOPwas not followed or was followed butthe SOP is not clear on how to report tothe different departments that the testdidn’t pass and that all production ac-tivities needed to be cancelled due tothe test failure. First the DeviationReport needs to be addressed then thequality system needs to be evaluatedand let QA/QC analyze the data if PAR-TICLE COUNT AND THE MICRO-BIAL LEVEL ARE OK DURING YOURPRODUCTION TIMEFRAME THENYOUR PRODUCT IS OK and your de-viation will address all of these inves-tigational facts.

The main function of the HEPA isfiltration and microbial level control.So, if those two parameters are withinthe acceptance limit your deviationreport must indicate that.

A You say the room particle countswere OK, I assume the particle

counts under the Grade A hoods wereOK. So there’s likely no risk to product.Here’s why:

A pinhole leak in a terminal filterwill NOT cause a measurable rise inroom particles, as the activities of thepeople in the room can cause morevariability in room air particles thanthe few particles per minute leakingthru a pinhole. With a pinhole, youMIGHT see a rise in AT REST particlecounts, but likely not enough to causeanyone to notice. Did the at-rest countsincrease?

If the terminal HEPA filter is theSECOND HEPA filter (i.e., there’s aprimary HEPA filter in the air han-dler) there could be essentially NOparticles passing thru the pinhole.

We have contended that it’s of novalue to pinhole (integrity) test ceilingterminal filters for Grade C and Brooms, as the presence of a pinhole hasnegligible effect on the room’s airbornecounts. You’d need a good sized hole,especially if the HEPA is downstreamof a primary HEPA. However, it is ofutmost importance to integrity test theHEPA filters in a Grade A hood, as the



©Copyright IS

PE 2008

pinhole MAY be directly over a criticalsite.

Bottom line – were the particlecounts near product exposure still “nor-mal?”

A In your last post you mention:“We have contended that it’s of

no value to pinhole (integrity) test ceil-ing terminal filters for Grade C and Brooms.” Is this documented somewhereand is it generally accepted? If no in-tegrity test is performed, how is filterreplacement scheduled? By pressuredrop?

A Unfortunately, there has beenno concession by the regulators

regarding integrity testing vs. overallefficiency testing of terminal (ceiling)filters. The key factor is bacteria reten-tion, a pinhole in a filter will pass a bitmore bacteria than one that has beenintegrity scanned to 99.99% (which alsopasses some bacteria, less than 0.01%of them). (But remember, even morebacteria come from the operators, sothe few bacteria passed may not changeroom counts).

However, there is a lesser expecta-tion for testing rigor and even testingfrequency for Grade C rooms, and Ithink we even say that in the latestdraft of the ISPE Sterile Manufactur-ing Facilities Baseline® Guide, which Idon’t have nearby right now (and ithasn’t been through final FDA review).There are a number of “reality checks”that we need to run by FDA, that gobeyond the content of the SterileBaseline® Guide, and this is one ofthem. Perhaps the fastest way to getan answer is to pose a question at theJune 2008 Washington Conferenceduring the FDA Q&A session.

Regardless of the results of integ-rity tests, I’d suggest replacing TER-MINAL filters (and filters in air han-dlers) based on pressure drop. You canusually repair terminal and AHU fil-ters (to a predetermined limit, no morethan a few percent of face area) andpressure drop will go up a little (due toreduced face area and thus higher ve-locity). Once the DP is about twice theoriginal DP the filter is full. Check theactual DP factor with the filter manu-

facturer. Dirty HEPA filters filter bet-ter than clean ones, but they use en-ergy. Often replacement frequency willbe driven by life cycle cost analysis.

For filters in Grade A (Unidirec-tional flow) hoods, the above discus-sion does not apply. Those filters arepart of the process equipment, andpinholes and repairs can be a majorproblem.

A The nature of the HEPA failureneeds to be understood in rela-

tion to the manufacturing activitiescarried out in the area. Are the ceilingHEPAs in the Grade B area (EU areaclassification) or the Grade A area. Ihave seen over the years a small num-bers of HEPA failures which have notimpacted upon the routine environ-mental data for both viable and non-viable particles for the relevant area.All data was well within the EU speci-fications even when a Grade A areaHEPA was involved in the failure. Youshould also consider when the lastmedia fill was done in the area inrelation to the date of the HEPA fail-ure. The HEPA failure needs to beaddressed as part of the deviation sys-tem and included in the investigationwill be the potential impact on productfilled since the last test passed. It ishighly unlikely that a HEPA failurewill lead to the area classification fail-ing the environmental standards re-quired for the product manufactureand hence there is no risk to the prod-ucts previously manufactured since thelast pass result. The only justificationfor a potential recall of products wouldbe if the failure involved a LAF HEPAor tunnel HEPA and the failure was agross one which did compromise theenvironmental standards of the area.One that comes to mind would be themajor damage of the filter media, i.e., ahole!

Investigational Products(IP) COP

Certificates of Analysis

Q I am attempting to assess theneed for C of A’s with every ship-

ment of drug to centres and distribu-tion sites alike. Does anyone have a

comprehensive list of custom versusregulatory requirements?

A Good question! In my personalopinion, it is more a matter of hab-

its. Site pharmacists were used to re-ceiving CoA, so documents describingthe “chemical and biologics” attributesof drugs, now, after the Directive/QPimplementation, the majority of re-lease certificates have the format of“statements” or “disclaimers” and thisis generating questions and concerns;that’s why the request to also have thedrug CoAs.

A Our experience is that we supplya “Quality Statement” with ship-

ments in the EU. In the US, we onlysupply the statement when requested.We do not generate a CofA for pack-aged supplies.

Packaging COP

Packaging DefectsClassification

Q I am working to classify the Pack-aging defects as Critical, major

and minor in aluminum, PVC,Polietilen bottles, individual boxes,mainly for inspection. Anyone has donesome similar work and has some clas-sification of defects or any orientation?

A First, most all companies use theMIl-STD-105 to determine sam-

pling plans. This is a valuable tool touse. As far as classifying defects it isreally a quality issue. If a package hasa defect that would be a detriment tothe product (Hole in blister, defectivechildproof cap, missing print, or illeg-ible print) these would normally bemajor defects. Minor defects are usu-ally cosmetic defects that would notaffect the product. These could bedented boxes, blisters, smeared print(still readable) or other defects causedby poor material or equipment caused.It really depends on your quality stan-dards. Some companies consider all ofthe above unacceptable and would re-ject them. The major defects wouldmost likely call for a 100% inspection ofthe lot but the minor defects would notcall for this. Hope this helps.



©Cop

yrig

ht IS

PE 2

008

A Thank you for your comments,actually we are using the mil-std

105 (ANZI) as sampling plan, however,what I want to do is quantify the de-fects and level found on each inspec-tion, I mean define an AQL for critical,major and minor defects. In your expe-rience, what are the most common AQLdefined for a critical, major or minordefects?

A Well again it depends on the de-fect. The AQL is defined as 1.0%

for critical defects and 2.5% for minordefects. Major might be in the 1.5%area. For six sigma calculations, thecommon is 3.4 defects per (one) millionopportunities (DPMO) so this is muchtighter. This is a goal some companiesare using. I have not done this forawhile but you really should set thesegoals with the Quality department. A1.0% AQL might not be acceptable fora wrong lot number/bar code or otherproduct quality defect. A wrong lotnumber should never be found and if itis then the entire batch should bestopped and 100% re-inspected. Awrong tablet count might be OK if itstill meets label claims. Some cos. fill alittle over in a tablet bottle. A missingprint is still critical but might not war-rant a 100% inspection unless youwould find a second one. Hope thishelps.

A One thing that I have done in thepast is take the individual com-

ponents that come together that makethe final package (i.e., bottle, closure,label, carton, case, etc), and considerwhat could actually go wrong with eachone of those components during pack-aging. Incoming inspection had its owninspection process which (right orwrong) we used as justification not toinspect for those defects as part of thepackaging AQL.

Take for instance label. Defects thatcould occur to the label include: Miss-ing (i.e., not on the bottle), damage thatcauses information not to be legible,presence of material behind label (sepa-rating material of known and unknownorigin), wrinkled, skewed placement,air pockets behind, damage that pre-sents a poor appearance, etc.

Once we identified all these poten-tial defects, we then considered whatthe impact of each of those defectswould be to the consumers (safety, use,relations with), and to regulatory bod-ies (meets or doesn’t meet regulations).Each one of these defects was thenclassified in these categories to deter-mine which level it would fit under.This developed into the critical, majorand minor defects that we used withthe AQL levels for inspection.

A I would agree that you have tounderstand the interaction of

each component in all your packs toensure that you have the correct classi-fication. It can be too easy to miss-classify and either risk failure or in-crease unit costs.

Process AnalyticalTechnology (PAT) COP

Analysis of Data

Q The major challenge involved inany analysis is generation, inte-

gration and organization of data. Usu-ally data are stored in big warehousesbut rarely retrieved. This is the com-mon situation in (bio)-pharmaceuticalindustry. Multivariate methods areideal for analysis. Any other methodsavailable for Analysis of Variables(ANOVA)? What is the Industry prac-tice?

A Have you tried “parallel coordi-nates.” I have used XMDV tool

(you can go to their web pages and seethe application). It is a very useful toolto evaluate thousands of data!

A We have used the product fromCurvaceous Software for the last

five or so years and found it excellent,especially for explaining complex in-teractions to plant personnel. The factthat no scaling or pre-processing ofdata is required and parameters can beshown at their native scale makes itmuch easier to communicate than amultivariate PCA model.

A You might want to consult an ar-ticle I wrote on Exponentially

Weighted Process Statistics and SPC

published in Pharmaceutical Engineer-ing in March/April 2007. This is anexcellent article on data especially withhigh frequency.

A In regards to multivariate dataanalysis, especially for spectro-

scopic based data, I have always foundThe Unscrambler, by Camo as beingvery useful.

A Another excellent data miningand modeling set of tools comes

from Pavilion Technologies, who wasrecently purchased by Rockwell Auto-mation. They have solid data evalua-tion tools that incorporate multivari-ate analysis and modeling. I wouldcheck out their tools as well. Very effec-tive.

Process/ProductDevelopment (PPD) COP

Product Transfer

Q There are several cases for prod-uct transfer and manufacturing

and development work together likethe transfer from development to manu-facturing (obvious!) or the transfer fromold to a new process system; usuallyproduct transfer from site to site,mainly in the cases of outsourcing,does not include the development team.Is there any real need for that whenprocess does not change? I would sayyes since they are the owners (in someway) of process and product design andany modification should be handled bythem before green light is granted. Anycomments?

A We struggle with who owns whatin transfers all the time. We have

established a system with a MasterTransfer Plan, where each time weapprove a RACI (Responsible, Account-able/Approver, Contributor, Informed).That seems to help. I agree that Devel-opment should be involved to somedegree since they may have intimateknowledge of the process that wouldhelp the team understand special nu-ances in the transfer program.

A This seems to be a topic manystruggle with! I liked the response



©Copyright IS

PE 2008

given in that having the RACI definedcan greatly aid as a communicationtool across a tech transfer team. Thedetails of RACI may also depend on thesize of your organization, complexity ofthe tech transfer project, availabilityof resources, and timeline. Roles mayalso change with the progression of theproject.

Specific to this question, if the de-velopment team has produced a pro-cess that is well characterized, and isDOCUMENTED, there may be a wayto make modifications without includ-ing an intensive commitment from dev,freeing them to work on the next bigthing! Otherwise, subject-matter ex-perts need to be involved when makingchanges, and without the tools to trans-fer the subject-matter expertise, youmay have to commit your dev team tolater phase projects.

A Thanks, good points! Let me putthis question together with ICH

Q8, Q9 and Q10 plus ICH Q8 Annex(draft) and outsourcing activities. TheDesign Space is a multivariate rela-tion involving materials, methods,machines, measures, the environmentand the people, then, when a product ismoved out of the company at least thefollowing is different: environment andpeople. I am just thinking that thecomplexity to manage the transfer couldbe very high unless some basics are

accepted, therefore the question couldbe, what are those basics?

Project Management(PM) COP

POLL Results – Why DoProjects Still Fail

A So the results are in:

1 - 50% of vote – poor planning2 - 25% of vote – lack of control2 - 25% of vote – no management of the

business changes needed to supportthe project

None of you identified the following ascauses:

• inappropriate selection of projectmanager

• no benefits management• lack of teamwork and poor team

culture• no link between the project and the

business

Thanks for voting.

A On larger projects that take morethan 1 year of Pre-Construction

change in Stake Holders, failure to re-visit and confirm expectations can leadto failure. Inadequate time to specifyspeciality equipment, accessories, util-ity requirements, compliance with lo-

cal codes and integration with the ar-chitectural theme can force last minutechanges, add costs and delay the project.It is important to have upper manage-ment support allocating sufficient re-sources to the project with clear under-standing of decision making authority.

A I agree that longer duration pro-jects do have some specific chal-

lenges around both stakeholder man-agement and control of delivery (projectobjectives and associated business ben-efits). Getting key processes in place(like decision making authority) at thestart of the project will reap benefitsparticularly for those projects lasting afew years!

Sterile ProductsProcessing (SPP) COP

Biofilm Removal

Q Could anyone please help me re-garding the method for removal

of biofilm appearing in WFI water tankas well piping? What would be theright chemicals to be used for removalof same?

A Chlorine bleach is harsh but effec-tive. It may not be appropriate in

your system at all. You must ensurethe liquid contact parts, etc. are com-patible with Cl bleach and it is allremoved.

Sustainable Facilities COP

Energy Policy/Carbon ReductionPlans/Aims

Q Can anyone comment on the sta-tus of their company’s corporate

energy policy/plans to reduce carbonemissions or attempt to become carbonneutral and the logistics of this chal-lenge from a cultural and financialviewpoint? Also, is anyone looking atthe viability/possibility of a carbonneutral laboratory/manufacturingplant or even campus/ site?

Respond to this question by joiningthis newly developed COP. Questionsand answers are welcome under thesection “Community Discussions,” onthe Sustainable Facilities COP site.

Have You Accessed ISPE’s Communities of Practice?ISPE’s Communities of Practice (COPs) provide enhanced connectivity through an interactive on-line community. ISPE offers industry professionals 16 different discipline-specific communities to choose from. Each COP provides global networking opportunities and access to a community-specific Body of Knowledge.

Choose one or sixteen - the choice is YOURS!

By accessing ISPE’s enhanced global Communities of Practice, COP members:

or challenge

content relevant to the discipline of the community

community activities

Access your ISPE Communities of Practice at www.ISPE.org/cops

Global Regulatory News

www.ispe.org/PE_Online_Exclusive

©Cop

yrig

ht IS

PE 2

008

JULY/AUGUST 2008 PHARMACEUTICAL ENGINEERING On-Line Exclusive 1

On-Line Exclusive Article




InternationalIn April 2008, The International Con-ference on Harmonisation (ICH) an-nounced on their Web site1 a proposednew tripartite Technical Guidance forActive Pharmaceutical Ingredients.This guidance will harmonise the sci-entific and technical principles for thedescription and justification of the de-velopment and manufacturing processof Drug Substances including bothchemical entities and biotechnologi-cal/biological entities as required forCTD sections S.2.2 to S.2.6. The docu-ment will take into consideration andprovide examples of what might beprovided in accordance with ICH guide-lines on Pharmaceutical Development(Q8), Quality Risk Management (Q9)and Pharmaceutical Quality Systems(Q10).

In April 2008 also, the Pharmaceu-tical Inspection Convention (PIC/S)2

issued a guide to good practices for thepreparation of medicinal products inhealthcare establishments, guide PE010-02. It should be noted that, whereasPIC/S Guide PE 009 applies to indus-trial manufacture of distributed me-dicinal products, the basic require-ments presented in PE 010-02 apply tothe preparation of medicinal productsnormally performed by healthcare es-tablishments for direct supply to pa-tients.

ArgentinaThe Argentinean National Medicines,Food and Medical Technology Admin-istration, ANMAT, has introduced anew regulation3 that defines the re-sponsibilities and procedures that phar-maceutical companies should followwhen recalling a product. The new regu-lation came into effect in March 2008and revises existing guidance on recallmanagement, follow-up and audits.Although the GMP regulations had asection on product recalls, they did notprovide sufficient detail with respectto the recall procedures. The new regu-lation also applies to manufacturers,importers, exporters and distributorsof personal hygiene products, cosmet-ics and perfumes.

AustraliaThe Australian Therapeutic GoodsAdministration (TGA) has issued the16th edition of the Guidance on theGMP clearance of overseas medicinemanufacturers is available via theirTGA Web site.4 The purpose of thisguidance is to provide information tosponsors and manufacturers on theacceptable form of evidence of GMPcompliance for overseas manufactur-ers, and outline how to submit suchevidence to the TGA.

The TGA is also developing a con-solidated reference document detail-ing the Australian regulatory require-ments for medical devices (ARGMD).This document will provide guidanceon all the regulatory requirements formedical devices in Australia and willbe made available on the TGA Website.5 It will replace the existing guid-ance documents and information sheetsfor medical devices. Although due forcompletion by the end of 2008, as eachsection is prepared, it will be releasedfor comment with a four week dead-line. However the existing guidancematerial will remain until it can befully replaced by the ARGMD.

EuropeIn March and April 2008, the Euro-pean Commission DG Enterprise andIndustry released via their Web site6

the following documents for public con-sultation and comment by end October2008:

• Legal proposals for combating thecounterfeiting of medicinal products.The Directorate-General Enterpriseand Industry is consulting all stake-holders and interested parties onkey ideas for amending the regula-tory framework for medicinal prod-ucts in an effort to combat the coun-terfeiting of medicinal products. TheCommission has identified threeareas of regulation of medicinalproducts where it is felt that im-provements to the regulatory frame-work could make a real contributionto protecting against counterfeitmedicinal products. These measuresrelate to:

1. Placing on the market of medici-nal products and related inspec-tions

2. Import, export and transhipmentof medicinal products

3. Placing on the market of activesubstances and related inspec-tions

• A proposed revision of GMP Annex11 on Computerized Systems andrelated changes in GMP. The Annexhas been updated in response to theincreased use of computerised sys-tems and the increased complexityof these systems. Consequentialamendments are also proposed forChapter 4 of the GMP Guide (Docu-mentation).

• A proposed amendment to Part II ofthe GMP Guide to incorporate prin-ciples of Quality Risk Managementas laid down in the ICH guidelineQ9. This also corresponds to similarchanges made to Part I Chapter 1 ofthe Guide and published in Febru-ary 2008. A new section on QualityRisk management is introduced assection 2.2.

• A proposed revision of GMP Annex13 on Investigational MedicinalProducts to clarify certain pointsrelated to reference and retentionsamples, the two-step release pro-cedure for investigational medici-nal products and to the principle ofindependence between productionand quality control functions. Aminor change has been made to sec-tion 3 in order to reinforce the prin-ciple of independence between pro-duction and quality control func-tions in cases where the number ofpersonnel involved is small.Changes are proposed to sections 36and 37 to supplement, for investiga-tional medicinal products, the guid-ance for reference and retentionsamples given in Annex 19. Section44 has been reworded to enhanceunderstanding of the two-step re-lease procedure that applies to in-vestigational medicinal products.



©Copyright IS

PE 2008

2 PHARMACEUTICAL ENGINEERING On-Line Exclusive JULY/AUGUST 2008

In May 2008, DG Enterprise and In-dustry also released an updated ver-sion of CTD Module 1 – AdministrativeInformation and Prescribing Informa-tion of Volume 2B – Presentation andcontent of the dossier. This updatedversion contains a new section 1.10 forinformation relating to paediatrics andincorporates a number of minor correc-tions or updates (e.g. references to newguidance documents).

The European Directorate for theQuality of Medicine & Healthcare(EDQM) has announced via its Website7 that they are now taking ordersfor the European pharmacopoeia 6thEdition 2009 to include Supplements6.3, 6.4, and 6.5. Although Supplement6.3 will become available in June/July2008, a list of its contents is now avail-able from the Web site.

Further, the 14th Edition (March2008) of the guide to the preparation,the use and the quality assurance ofblood components is now available,providing guidelines aimed at ensur-ing the safety, efficacy and quality as-surance of these components. The guidecontains recommendations on bloodcollection, blood components, techni-cal procedures, transfusion practicesand quality systems for blood estab-lishments and is designed for bloodtransfusion services and national regu-lators in Europe and perhaps beyond.

The Committee for Medicinal Prod-ucts for Human Use (CHMP)8 has pub-lished a report from its plenary meet-ing held on 21 to 24 April 2008. Thefollowing relevant guideline11 has beenadopted by the Quality Working Party:

• Guideline on Quality of Combina-tion Herbal Medicinal Products/Tra-ditional Herbal Medicinal Products(EMEA/HMPC/CHMP/CVMP/214869/2006).9 This guideline ad-dresses in detail approaches for iden-tification and quantitative determi-nation of herbal substances and/orherbal preparations in combinationherbal medicinal products.

The Committee on Herbal MedicinalProducts (HMPC)10 had not publishedfull meeting reports for meetings heldin April and May 2008 at the time of

preparing this review. However, it isnoted that the HMPC has adopted theQWP ‘Concept Paper on the develop-ment of a quality guideline on pharma-ceutical development of medicines forpaediatric use’ (EMEA/138931/2008).Upon adoption by the CHMP andPDCO, the concept paper will be pub-lished on the EMEA Web site11

The Paediatric Committee (PDCO)12

has published their monthly meetingreports for the meetings held on 9-11April and 6-8 May 2008. No new rel-evant information was noted.

The Committee for Orphan Medici-nal Products (COMP)13 has publishedtheir monthly meeting reports for themeetings held on 8 April and 13-14May 2008. No new relevant informa-tion was noted.

The Committee for Veterinary Me-dicinal Products (CVMP)14 has pub-lished their Monthly Reports of Appli-cation Procedures, Guidelines and Re-lated Documents for March and April2008. Each includes an accumulativesummary of the opinions issued by theCVMP in the current year and a list ofadopted Guidelines and other publicdocuments.

Noteworthy, the following relevantguidelines were adopted for public con-sultation at their March meeting:15

• Guideline on the requirements forthe replacement of established mas-ter seeds (MS) already used inauthorised immunological veteri-nary medicinal products (IVMPs)(EMEA/CVMP/IWP/105504/2007).

• Guideline on the Quality of Combi-nation Herbal Medicinal Products/Traditional Herbal Medicinal Prod-ucts (EMEA/HMPC/CHMP/CVMP/214869/2006, as above)

FranceThe French Regulatory Authority(AFSSAPS) has via its Webs ite16 is-sued clarification on the question ofcompatibility between medical devicesplaced on the market by different manu-facturers. A manufacturer of medicaldevices may claim compatibility withanother medical device even if themanufacturer of the latter disputes

the use of medical devices of othermanufacturers. The manufacturermust then demonstrate its product’scompatibility with the medical deviceand ensure the continuity of that com-patibility over time in the event thatthe medical device with which it isassociated is upgraded. Guidance onhow this may be achieved is provided.

JordanThe Jordanian Food and Drug Admin-istration (JDFA) has begun re-evalu-ating registration approvals2 grantedto manufacturing sites before July 2000to identify dossiers that were submit-ted with incomplete information relat-ing to the site.

The agency will contact the compa-nies that registered their manufactur-ing sites with incomplete submissions.It will ask them to file a new applica-tion form, together with documenta-tion that provides details about theplant, its location, the products manu-factured on the site and the address ofthe company’s headquarters or mainoffice.

The JFDA also found problems re-lating to specific products which failedto pass quality control testing by theMinistry of Health including problemsregarding the overall quality of prod-ucts manufactured at one particularsite.

Separately, the JFDA has issued acircular asking all pharmaceutical dis-tributors to submit applications to ei-ther renew or delete the registration ofproducts that were originally regis-tered (or re-registered) in 2002. Therenewal will take place in two stagesand the applicant must submit infor-mation about the product andmanufacturer’s.

References1. http://www.ich.org/cache/compo/

276-254-1.html2. h t t p : / / w w w . p i c s c h e m e . o r g /

index.php3. RAJ Pharma May 2008.4. http://www.tga.gov.au/manuf/

gmpsom.htm5. http : / /www.tga .gov .au /new/

new.htm



©Cop

yrig

ht IS

PE 2

008

JULY/AUGUST 2008 PHARMACEUTICAL ENGINEERING On-Line Exclusive 3

6. http://ec.europa.eu/enterprise/pharmaceuticals/pharmacos/new.htm

7. https://www.edqm.eu/store/8. http: / /www.emea.europa.eu/

Press%20Office/chmp.htm9. http://www.emea.europa.eu/pdfs/

human/hmpc/21486906.pdf10. http: / /www.emea.europa.eu/

Press%20Office/hmpc.htm11. http://www.emea.europa.eu/htms/

h u m a n / h u m a n g u i d e l i n e s /quality.htm

12. http: / /www.emea.europa.eu/Press%20Office/pdco.htm

13. http: / /www.emea.europa.eu/Press%20Office/comp.htm

14. t t p : / / w w w . e m e a . e u r o p a . e u /Press%20Office/cvmp.htm

15. http://www.emea.europa.eu/pdfs/vet/press/pr/17496208en.pdf

16. http://agmed.sante.gouv.fr/ang/pdf/compa_tibility_medical_devices.pdf

This information was provided by IanMorland, MRPharmS, PhD, Pharma-ceutical Research Associates (UK).

Date post:	07-Apr-2015
Category:	Documents
Upload:	andyraff
View:	481 times
Download:	4 times