TECHNOLOGY APPLICATION

Application of Total Organic Carbon Analysis to Cleaning Validation

K. M. JENKENS* , A. J. VANDERWIELEN +, J. A. ARMSTRONG,L. M. LEONARD, G. P. MURPHY, and N. A. PIROS,

Quality Control, The Upjohn Company, Kalamazoo, Michigan

ABSTRACT: Cleaning Validation is the process of assuring that cleaning procedures effectively remove residue From manufacturing equipment /facilities below a predetermined level. This is necessary to assure the quality of Future products using the equipment, to prevent cross- contamination and as a GMP requirement. Currently, Cleaning validation sample are measured using HPLC or spectrophtometric Method of analysis which are Often time consuming and subject to a number of interference. Total Organic Carbon (TOC) analysis is a new Method which has previously only been applied to measurement of carbon residues on production surfaces for Biopharmaceuticals. We have applied the Toc analysis method to a number of traditional pharmaceuticals Products including antibiotics, steroids, and antinaiseant in addition to biopharmaceuticals. The method offers Extremely low detection capability (ppm and ppb), rapid sample analysis time and therefore quick turn- around of Production equipment and facilities. TOC analysis is also applicable to on- line analysis. The method allows the Measurement of extraneous materials such as process intermediates, cleaning agents, and protein materials notPossible by other methods.

Introduction area such as nuclear power plant , semiconductor Production and for pharmaceutical grade water for Cleaning validation has been an area of increased injection . regulatory and industry scrutiny. This attention is war- In 1990 the use of TOC for pharmaceuticals in ranted given the increased use of multipurpose equip - regard to cleaning validation was first reported(1). ment and an overall concern for quality of product. For since that publication the use of TOCfor biopharmaceu- both industry and FDA inspectors the concern is the ticals has been widely used. The same characteristics same_______assurance that equipment is clean and that that make TOC applicable for biopharmaceuticals are product quality and safety are maintained. Cleaning Appropriate for traditoional pharmaceuticals .TOC has validation provides such assurance. One of the key low level detection(ppb), rapid analysis time, is low cost components of any cleaning validation strategy are the as compared to other methods, and can detect all analytical methods use measure Residuals. The carbon – based residuals. overall goal is to assure that the product does not become adulterated from any source whatever ; interme-

diates , excipients and cleaning agents must therefore be Theory__ TOC considered . In most cases , however, the active drug ingredients are still the overriding consideration . TOC analysis involves the oxidation of carbon and the The measurement of total organic carbon (TOC) detection of the resulting carbon dioxide produced from content of contamination by organic compound is a new oxidation. This oxidation can be induced by a few method applied to cleaning validation . The method has common method including photocatalytic oxidation (2), previously been used for analysis of high purity water in chemical oxidation (peroxdisulfate or oxygen) (3) and High temperature combustion (4) . The work presented In this paper was accomplished by use of the shimadzu __________________________ TOC 5000 analyzer. Received December 20, 1994. accepted for publication May 22, 1995 The Shimadzu TOC -5000 (Shimadzu instruments,

Presented at the PDA Annual meeting , Philadelphia, PA November Columbia Maryland ) is an elemental analyzer capabel 1994 of independent measurement of TC (Total Carbon) , Author to whom correspondence should be addressed: Western TOC (Total Organic Carbon) and IC (Inorganic Car-

Michigan University, Kalamazoo, MI 49008 -3842. bon ) in water. The Shimadzu instruments measures * Author to whom correspondence should be addressed: Sterile Manu- organic specices by combustion to produce CO2. An Fecturing , The Upjohn Company Kalamazoo, MI,49001. NDIR(non dispersive infrared )detector is then utilized Chemistry Department , Western Michigan University, MI , 49008 to measure the resulting CO2 produce from combus- Chemistry Department , Illonois University, Normal, IL, 61761. tion. This combustion takes place at high temperature $ R. Baffi, et. Al.,1991 . A Total Organic Carbon Analysis for Vidating Cleaning Between products in Biopharmaceutical Manu- Fcturing Journal of Parenteral Science & Technology, Volume 45 (1), Pages 13-19 PDA Journal of Pharmaceutical Science & Technology

And produces CO2 by the reaction described below. Measurement of TC (Tital Carbon)

Reaction of Catalytic Combustion The TC combustion tube is filled with TC catalyst and Heated to 68oC. Carrier (purified air) is supplied (Oxidation of Organic Compound) into this tube after the flow rate is adjusted to 150CaHbNcO + O2 - aCO2 __ +b/2H2O +cNO ml/ min by pressure and mass flow controllers and Mistened by a humidifier. When the sample is introduced in to the TC combus- (Decomposition of Carbonate) tion tube via yhe sample injector (ASI- 5000), the TC MeCO3 __ CO2 +MeO component in the sample (comprising TOC and IC ) is (Decomposition of Bicarbonate) combusted /oxidized to from CO2.The carrier gas with The combustion product (CO2 ) from the combustion MeHCO __ CO2 + 1/2Me2O +1/2H2O flow through the IC reaction vassel and is cooled and dried by a dehumidifier. It then passes through aSharp peak heights are obtained by instantaneously halogen scrubber into the sample cell of the non-converting all carbon compound toCO2 as described. dispersive infrared detector ( NDIR ), where CO2 is the TOC- 5000 uses paek area as the actual measure- detected.The NDIR outputs a detection signal ( analogment method. signal ) wich generates a peak, the area of wich is Calculated by a data processor. The peak is proportional to the TC concentration inTOC/ TC/IC/ NPOC the sample . If a calibration curve expressing the rel- The Shimadzu TOC- 5000 instrument(fig. 1) IS ca- ationship between peak area and the TC concentrationpable of performing total organic carbon analysis (TOC), is obtained using a TC standard solution, the TC con-total carbon analysis ( TC ), inorganic carbon analysis entration of the sample may be determined. This is an( IC) and non –purgable carbon analysis ( NPOC). External standard method of calibreation. The standardeach method is described further and are related by the curve used can be either polygonal or a liner least following sample equation. Squares fit.

TC = TOC +IC

NPOC = TOC + purgables

Figure1____ Schematic of the TOC 5000 instrument. The Schematic was reproductive from the instrument operating procedure with the Permission of the Shimadzu corporation.

Vol. 50, NO. 1 / January __ February 1996Measurement of IC ( Inorganic Carbon )

TABLE I Analysis Method ___ Common Uses The sample is introduced via sample injector (ASI- 5000)Into the IC reaction vessel ( containing IC reagent-Phosphoric acid ), through which carrier gas is following formtiny bubbles. Only the IC component in thesample is the decomposed to from CO2 which is detected upon reaching the NDIR. The concentration of IC is HPLC X X Xdetermined in the same manner as that of the TC except TLC X Xthat an IC calibration curve is used to determine sample spectrophoto-concenteration.Carbon in the form of carbonate and metric X X XHydrogen carbonate may be measured as TOC X X X X

ELISA X electropho-

Resis X PH X Conductivity X Measurement of TOC ( Total Organic Carbon) Gravimetric X X

The TOC concentration may be determined by sub-Tracing the IC concentration , obtained as described Above under measurement of IC ( Inorganic Carbon ) fromThe TC concentration as describedunder measurement Of TC ( Total Carbon ). This subtraction is performed by TABLE IIThe TOC-5000 and reported as the toc value. Analysis Method _______ Advantages/ Disadvantages

Method Advantages Disadvantages

Measurement of NPOC ( non- purgable Organic Carbon The TOC concentration may be determined directly HPLC highly specific Long analysis timeUsing another procedure. In this case , the sample is moderate to high expensiveacidified beforehand and then spraged automatically sensitivity with purified gas to remove the entire IC component. TLC QuantitativeThe is then analyzed to obtained the TC concentro- Moderate to high visual endpointTion as described under the section Measurement of TC Senstivity Long sample prep(Total Carbon). This method is refrred to as the NPOC Fairly inexpensive Method. Spectrophoto- Moderate to high Not quantitative NPOC specifically refers to non- volatile organic car- metric specificity bon which is not eliminated by ovaporation during the High sensitivity sparging process. Volatile organic compound s such as methodorganic solventwhich are not soluble in water TOC Broad spectrum Non- specificare eliminated from a sample at room temperature via Low level detec- Aqueous solublesparging . The organic carbon which is evaporated during tion samples onlysparging is referred to as POC ( purgable Organic on- line capability Carbon ). There is however an increase in analysis time Rapid sample When performing NPOC on sample since the purge turn- around Time with gas ( high purity air ) is 2 minutes or more minimal sample Prep ELISA specific for bio- Long sample turn Pharmaceutical around

Cleaning Drug Agent Biopharma- Method Residue Excipient Residue ceutical

Very sensitive very expensive Labor intensive

Problems withAdvantages for cleaning Validation TOC analysis offers some distinct advantages forCleaining validation over other traditional method listed denatured pro-In Table I. For each of these method s the common uses teinAre listed in terms of residuals they detect. Of all the Electrophoresis Specific for bio- Very expensiveMethods TOC has the broadest range of uses since it can pharmaceutical Labor intensivedetect any carbon- based residuals. Table II list some of Moderate sensitive Problems with the distinct advantages and disadvantages of the other denatured pro-methods used for cleaning validation. The most signify- teincan differences among the methods are in the areas of Long sample turncost and detection capabilities.HPLC has the capability aroundfor the lowest level of detection,but is also quite costly. PH Rapid Non- specificSpectrometric methods ( UV ) and thin- layer chromatog Inexpensive Water soluble only raphy (TLC ) are lower cost methods but at the expensive On- line capability Only useful for of detection capability. Of the methods applied to Cleaning agents biopharmaceutical, ELISA is a high cost, time- consum- Limited sensitivity ing method. The electrophoresis method, polyacryl- Visual Detection Immediate result Not quantitaveamide gel electrophoresis (PAGE) is an extremely Good for genral Subjectivesensitive assay method. This method is primarily used Screening

PDA Journal of Pharmaceutical Science & Technology

Background Material Description Particales ( PPM ) % Recovery

Quartz Wool+ Spun quartz Excessive 0.00 91.72 Absorbond TM * Hydroentagled poly- 30million particales/m2 0.00 29.01 ester Cotton Prewashed in methanol High 9.71 66.29 And chloroform Foam cubes* 5/ 16” foam cubes Low 61.9 __Metrigard TM* Ulterafine glass fiber High 0.00 __ Acrylic binder Clean cotton TM * sealed edge cotton 146 million particales/m2 5.50 48.83 BlendSterile Wipe HS II TM * Nonwoven cellulose/ 145 million particales/m2 7.42 74.43 Polyester blendAllpha Wipe TM * 100% continuous fila- 30 millio particales /m2 0.0 72.89 Ment double knit PolyesterSterile WipeLP TM * 100% continuous fila- 30 million particales /m2 0.0 68.95 Ment polyester Sterile Wipe LP 10 TM* 100 % continuous fila- 5 million particales/m2 0.0 82.21 Ment polyester with

Thermal order sealSterile Wipe HS TM * Nonwoven cellulose and 165 million particales/m2 17.64 ___ Polypropylene com- PositeBeta Wipe TM* 70% polypropyl- 165 million particales/m2 10.78 ____ Ene / 30% cellulose

* Material obtained from the Texwipe Company; particale count is from Rexwipe literature.** Material obtained from Gelman sciences.+ Material obtained from Shimadzu through Giangarlo scientficCompany, Inc.

for the separation of protein based upon molecular of assay time , cost and downtime. TOC provides anweight and can detect both variant and contaminating immediate answer for validationand for on going moni-proteins, yet identification is somewhat ambiguous . TOC torning at a lower cost. It provides assurance that theprovides a method which is relatively inexpensive.rapid surface of rinsate monitored is clean to a low residual in analysis time, and has low level detection ( ppm __ ppb ). Level.Since it detect all carbon based residues it can detectDrug, excipient and cleaning agent residues. Given the Assay developmentrapid analysis time (3__ 5 ) and the availability ofcompact instrumentation it can be applicable to on- line Swab Challengeanalysis. As noted in the resent FDA guide to inspection of There are some limitation of the TOC method. Only cleaning process, “There are two genral types ofWater soluble compounds can be analyzed by the method. sampling that have been found acceptable. The mostIt is an Nonspecific method; however, that may not be a desirable is the direct method of sampling the surface ofSerious limitation since most of the result are negative the equipment” (5 ). It was therefore desireable to de-Due to the fact that the cleaning process is optimized and velop of swabbing technique for TOC analysis. TheThe surfaces are clean. In these cases a specific method swabbing technique involvesthe use of a swabbing Such as HPLC , TLC or any of the other specific assay material, often saturatedwith a solvent, and physicallyMethods are an unnecessary and added expense in terms sampling the equipment or facility surface. In the case of

TABLE IV Sample Fillter Materials __ Background

Millipore Millipore Millipore Sterile Gelman Gelman Gelman

Millex LCR Millex- HV Millex- HA Acordisc GHP Acordisc Filter 0.05um 0.45um 0.45um 0.45um Acirdisc CR

Composition Hydrophilic Durapore PVDF Cellulose PVDF Hydrpphilic PTFE PTFE Polypropylene Background 0.0* 3.2 0.0* 184.9 11.5 0.0 ( ppm carbon )% Recovery 106.9 101.1 100.2

Background was noted below the detection limit of 1.4 ppm TABLE V Swab / Rinse Recovery

Drug Drug Swab % Recovery Category Formulation Type/ indication Agent Swab/ Rinse

Biopharmaceuticals Sterile Recombinant protein Water 68 79 Sterile Immunosuppressant Tween 80/ Water 52 83Antinauseant Dry product __ Tablet Motion sickness Water 83 103Antibiotics Dry product __ Tablet Acidified Water 83 102 Sterile solution Anaeribic bacteria semi- Water 68 94 synthetic antibiotic Dry product __ powder Broad spectrum antibiotic Water 84 103 Fluid Sterile solution Gram- positive antibiotic Water 50 98 Dry product __ tablet Steroids Sterile solution Steriodal research com- Acidified water 73 90 Pound Sterile solution Glucocorticiod Water 68 89

TOC only water or agents with low carbon background work. The alphawipe ( Texwipe Cat. No. TX 1004 ) wasCan be used for swabbing. These agents can include used as the swab material since it was less costly than the Acids, bases, low carbon surfactants and wetting agents Sterile wipe and sterile wipe LP10which is the same(Tween 80 ). In adition, theswab it self must be of such a material only sterilized. The strile Wipe LP10 has aMaterial that it does not contribute significant back- manufacture-rated partical level less then the other twoGround. Both of these criteria present a challenge. A Swab only due to the thermal border seal. This was notNumber of swabbing material were evaluated for use significant since all swab material would be cut intoWith TOC. In all cases HPLC grade water was used as a smaller size, thereby eliminating any advantage of aSolvent choice for the analysis. Table III lists material sealed border. Alpha Wipe swab are prepared by cit-Evaluated as possible swab, their composition, rated ting sheet of the material into 2cm by 2cm squarePartical count information and the background level of which are then used as swab. NO pretreatment of theCarbon determined for two swab in 10.0ml of high material is necessary prior to use.Purity water. In adition, the recovery of each swab was Tasted using spiked sample of a water soluble drug( methscopolamine bromide ). This evaluation demon- Sample FiltrationStrated that quartz wool at the highest recovery with no Since large insoluble particalesfrom swabbing surfacesDetectable background. However, the quartz wool mate- could be detected by TOC and cause mechanical prob-Rial left a residue of particulates on swabbed surfaces lems in the sample injection system, sample filtrationAns was therefore determined unacceptable. Was necessary. Various filters from Millipore ( Millipore The Alpha Wipe sterile Wipe, and sterile Wipe LP 10 Crop., Bedford, MA ) and Gelman ( Gelman Sci., AnnSwab (The Texwipe Company, Upper saddle River, Arbor, MI ) were tasted background. The resultsNJ ) demonstrated the next best recovery with no signify- ( Table IV ) demonstrated that only the Millipore LCR,Cant background. Since these swabs are a knitted polyes- Millipore HA, and Gelman CR filters had backgroundTer material there was no residue detected on the reading of 0.0 ppm carbon. For each of these filters

Swabbed area at the ppm level. This material was chosen studies were conducted to confirm that absorption of theAs the swabbing materialfor all further development drug did not occur on the filter material. The Millipore

TABLE VI Swab / Rinse Recovery Over Range

Drug Drug Swab/ Rinse Swab / Rinse Swab / Rinse Swab/ Rinse Category Type/ Indication Agent Slope Intercept Correlation Coef.

Biopharmaceuticales Recombinant protein Water 0.884 1.004 __ 5.199 0.570 0.996 0.996 Immunosuppressant Tween 80/water 0.585 1.048 5.158 22.609 0.996 0.998Antinauseant Motion sickness Water 0.758 1.119 3.882 5.138 0.999 0.999Antibiotics Acidified water 0.834 1.031 3.454 _ 2.552 0.999 0.999 Anaerobic bacteria Water 0.668 0.988 _3.473 _6.705 0.999 0.999 Semisynthitic antibi- Otic Broad spectrum- antibi- Water 1.318 0.991 12.208 0.558 0.999 0.999 Otic Gram- pissitive antibiotic Water 0.614 1.009 _ 0.569 _9.029 0.993 0.999Steroids Steroidal research com- Acidified Water 0.765 0.981 3.122 _1.557 0.999 0.999 Pound Glucocorticoid Water 0.769 0.962 _7.230 1.112 0.999 0.999

PDA Journal of pharmaceutical science & Technology TABLE VII Limit of Drug Rinse Detection Drug Category Type / Indication Agent (ppm range)

Biopharmaceuticals Recombinant protein Water 14 Immunosuppresant Tween 80/Water 6Antinauseant Motion sickness Water 4 5Antibiotics Anaerobic bacteria semisynthetic Water 8 Antibiotic Broad spectrum-antibiotic Water 1 Gram positive-antibiotic Water 2Steroids Steroidal research compound Acidified Water 7 Glucocorticoid Water 1

LCR and HA and Gelman CR filters all performed used to show that contaminants can be recoverd fromEqually well in this regard . In the case of protein the the equipment surfaces and at what level, ie; 50 %Gelman CR and GHP Acrodisc filers provided low recovery 90%.” This reinforces importance of deter-Background without absorption of the protein material. Mining the actual swab and rinse recovery of the Method. These recovery valuesare then used in the finalSwab recovery calculation. Six 1.0ml- sample of various drugat 225 mcg/ ml Rinse RecoveryWere spiked on stainless steel surfaces and allowed toDry. These sample were spiked onto an area approxi- Six 1.0ml sample of various drug at 225 mcg / mlMately225 cm2. swabbing was conducted using Alpha- were spiked intostainless steel beaker and allowed toWipes at the swab wetted with the slected swabbing dry. 20 ml of the solution used for swabbing was used to Agents. Two separate swab held with a stainless steel rinse the beaker. This solution was collected and Forceps were used to swab the 225 cm2 surface both filtered prior to analysis. Rinse recovery data for various Vertically and horizontly. Drug compounds are listed in Table V. Although rinse The swab were placed in acid washed vails (Eagle recovery values from 79__103%, recovery wasPicher, Miami, Oklahoma,) and 10 ml; of diluent (HPLC generally in the 90 % range of better.Grade water ) was added. After extraction by shaking for These recovery values are considered a worst caseTen minutes the solution was passed through a filter since the sample are allowed to dryon the surfaces. The( Millipore , Gilman, CR or GHP Acordisc )to remove Swab and rinse recovery values are used in the finalAny particulates. Swabsample were then analyzed by Calculation for sample taken from production surfaces.The total carbon analysis method described. The average Recovery from stainless steelsurfaces for the drugs Linearity of detector responseTasted and the swabbing agent used are listed in Table V.Swab recovery were greater than 50% in all cases, Linearity of detector response was measured forWhich was considered acceptable. Potassium hydrogen phthalate standards diluted from a 1000 ppm stock solution. Potassium hydrogen phthalate C8H5O4K ( Sigma, St. Louis, Mo. )provides a pure carbonDetermination of Swab and Rinse Recovery Source which was used a standard. The range of The FDA guide to inspectionfor cleaning validation detection measuredfor Linearity wasfrom 20 % to 400 %Notes that “The firm should challenge the analytical of the 1.0mcg/cm2 level for each drug tasted in the ppmMethod in combination with the sampling method ( S) range. The detector responsefor a typical analysis was

TABLE VIII Comparison to HPLC

HPLC Drug TOC Average AverageDrug Category Type/ Indication Recovery RSD Recovery RSD

Antinauseant Motion sickness 83.3 4.2 55 3.1Antibiotics Anaerobic bacteria 68 14 39 7.5 Semisynthtic antibi- Otic

Gram positive – antibiotic 50 9.9 86 6.4Steroids Steroidal research com- Pound Glucocorticoid 71.6 8 72.4 7

TABLE IX Comparison to UV Total protein methods

TOC UV TotalDrug Drug Average protienCategory Type Indication Surface Recovery RSD Average Recovery RSD

Biopharmaceutical Immunosuppreseant Glass 51 19.5 21 32 Stainless steel 52 19 16 30 Recombinant protein Stainless steel 68 10.8 55 5.7

Linear over the range as determined by peak area value of the intercept plus three times the estimatedResponse. The correlation for the detector response was standard daviation of the slop ( 6 ) . Table VII lists theAlways greater than 0.95 with an intercept that was not determined detection capability is only limited by the Significantly different from zero. All standard curves are tasted. These detection level were determined only forPrepared using potassium hydrogen phthalate and the ppm range of the instrument. In the ppb range of theSample are analyzed against these stored curves. Check instrument the detection capability is only limited by theSolution over the range of testing are used to ensure background of the process water used in that environ-Accuracy of the curves. Ment . HPLC grade water used in the lab has a carbon Background of 120__ 200ppb. The background of produc-Recovery Over Range tion- purified rinse water will vary with the source and Treatment . Analysis of production equipment rinses inRecovery over the range of 20% to 400% of various the ppb range should take account the backgroundDrugs at 1.0 mcg/m2 was determined for both the carbon level of the water used for rinse.Swabbing and rinse method.For the swab method 1.0ml Sample at 20% , 50%, 100%, 150%and 400% of the 1.0Mcg/cm2 levelwere spiked onto 15cm by 15 cm areas on Comparison to Other MethodsStainless steel sheet and allowed to dry. High purity Water was used as a swabbing agent and Alpha Wipes Comparison of the TOC method to existing HPLCWere used as the swab. These result demonstrate that Methods for each drug was conducted to assure equiva-the recovery was linear for all drugs tasted. Lency and controlled tested. Sample of drugs (1.0 ) ml For the rinse method 2.0ml sampleat 20% , 50% , spiked onto 225cm2areas of stainless steel at a concen-100%, 150%. And 400%of the 1.0mcg/cm2levelwere tration of225mcg/ml and allowed to dry.These areasSpiked into stainless steel beakers and allowed to dry.A were swabbed and separate sample were analyzed by20.0ml rinse solution of HPLC grade water was added both the HPLC and TOC method. Table VIII lists theTo each beaker for recovery. The solution was agitated average recovery and RSD for six seprate sites ana-And collected for analysis. The analysis of the rinse lyzed by each method. In genral, the result were eitherSolution demonstrated that the recovery was linear comparable or the TOC method had greater recovery( Table VI ). than the JPLC method.

In the case of proteins the TOC method was comp- Pared to spectrophotometric total proteins method.Limit of Detection These total proteins methods both used the lowery method Due to the background contributed by the Alpha Wipe for total proteins determination. The lowry methodSwabs, the limit of detection for swabbing material. The swab suffers from interference due to surfactants and otherGround contributed by the swabbing material. The swab organic compounds (7 ). Table IX list the everage recover-Background ranges typically from 0__15ppm . The swab ery and RSD for each method. In all cases the TOCBackground value from each run is used to subtract this method resulted in higher average recovery. One signify-Background from all swab sample. Thereis no pretreat Ment of the swab . The swab are handaled with powder –Free class 100 clean rooms latex gloves (QRP, Tucson, TABLE XAZ ) and acid washed stainless steel foreceps. The gloves Production Monotoring Environmental BackgroundAre powder – free , have low particulates and thereby do _________________________________________________Not contribute to the background for the swabbing ResultsAgent. Surface area ppm/225 cm2 _________________________________________________

To determine the rinse limit of detection , solution Sterile filling pump 0Were prepared at 10% , 70% and 200% of the 100% level Sterile holding tank 2( 225 mcg drug ), spiked into 1.2 liter stainless steel Sterile trays 3.8Beakers and allowed to dry. The beaker were rinsed Blender interior 0With 10.0 ml rinse solution (water ) and analyzed for Blender lid 4Total carbon. The results were graphed and the intercept Loading chute 0Calculated. The detection limit is defined as the absolute Blending Rm Floor 60 Module Floor 177 ___________________________________________________ PDA Journal of pharmaceutical science& Technology

TABLE XI TABLE XIII Freez Dryer Trays Sterile Manufecturing Tank __ protein_______________________________________________________ __________________________________________________Equipment Antibiotic Steroid Equipment Run 1 Run 2

Part Result (ppm ) Result ( ppm) Part Result (ppm) Result (ppm)_______________________________________________________ __________________________________________________Tray 1 interior 2 9 Tank interior 4 10Tray 1 lid 5 3 Tank lid <1 4Tray 2 interior <1 3 Tank discharge port <1 6Tray 2 lid <1 <1 Qs rod 13 7Tray 3 interior 3 <1 ___________________________________________________

Tray 3 lid 1 4________________________________________________________

In the fluid manufacturing area the TOC analysis Method for an antibiotic was used to validate a newCant advantage to the TOC method is that it detect both cleaning process. Table XIV shows the data from thisNative and denatured protein. Analysis which indicates high result for the tank interior And impeller from two separate monitoring. This dataEnvironmental Background of the shaft which was identified as a cleaning problem To determined areas where TOC analysis was feasible TOC analysis has also been applied in the dry prod-It was necessary to determined the background for produc- ucts area. Table XV lists the monitoring result for aTion surfaces. Background sample were taken from feeder used in one of our micrinizing operations. TheEquipment in the sterile operation and dry products result were all 10ppm or less for each 225 cm2 areaManufacturing area. In the sterile area swb sample monitored by swabbing except for one sample siteWere taken from filling pumps, holding tanks, and sterile These result demonstrate that TOC analysis of swabTrays ( Table X ). In the dry products area, samples were samples can be effectively used for cleaning validation ofTaken from the interior of blenders, the lid and the loading equipment in the sterile, fluid,/ ointments and dry prod-Chute. In both area sample were taken from the ucts manufacturing/ production area.Module and facility floor . The equipment surfaces all Had background carbon levels from 0__4 ppm over a 225 Cm2 area. Floor surfaces were in the range from 60__ 177Ppm of carbon over a 225 cm2 area. This data demon- Determinetion Cleaning Agentstrates that TOC would not be optimal for monitoring non product contract surfaces such as the facility; how- The FDA guide to inspection to cleaning validationever, it is appropriate for the equipment. This was not notes the importance of determining residual cleaningviewed as a serious disadvantage since facilities can be agents with their statement “ As with products residues, itvisually inspected and are usually only required to meet is imported and it is expected that the manufacturerless stringent resodue limit. Evaluate the efficiency of cleaning process for the the removel of residues. Detergents are not part of the manufacturing process and are only added to facilitiesProduction Sampling cleaning during the cleaning process” (5). Because of The TOC method developed were used for cleaning these concern residual cleaning agents should be exam-Validation monitorings in various areas of production. In ined in the cleaning process. TOC is one possibleThe sterile operation the TOC analysis for antibiotics analytical method to detect residual cleaning agentsSteroids and proteins were used to monitor sterile since they are carbon based.Operation equipment such as bulk freez dryer trays , The carbon content of cleaning agent was deter-Sterile manufacturing pumps and sterile manufacturing mined by making up 100 ppm solution of each of the Tanks. Tables XI, XII, XIII, XIV, and XV list data for cleaning agents thought to contain carbon , and determin-Each of these equipment surfaces. In all cases a 225 cm2 ing their carbon content using the shimadzu TOC-5000.Area was swabbed using Alpha Wipe and the swab were The TC content of the samples were measured and usedTested by TOC analysis. In all cases the equipment the determine thr present carbon in each of the cleaningMonitored met test in regard to residual monitoring agents tested. The TOC analysis works best under acidic Limits. Condition due to the platinum catalyst in the combus- Tion. In circomastances where the sample run are

TABLE XII TABLE XIV Sterile Manufecturing Pump___ Protien Fluid Manufacturing Tank ___ Antibiotic_____________________________________________________ ________________________________________________Equipment Part Result ( ppm ) Equipment Run 1 Run 2_____________________________________________________ Part Results ( ppm ) Results (ppm)Filter housing 9 ________________________________________________

Filter housing top 3 Tank interior 7 82s.s. pipes 3 Impellor 66 9Rotors 9 Tank discharge port 15 10Rotor housing 6 ______________________________________________________________________________________________________

Vol. 50, No. 1 / January_ February 1996

TABLE VX CIP/COP have lead to further use in the solid dose, Micronizing Feeder ___ Antibiotic fluid / ointments and bulk drug operation._______________________________________________________ TOC has the capability for no-line testing of such Equipment Run 1 Run 2 system. This would provide advantages such as rapid Part Results (ppm ) Result (ppm ) response in term of resultand provides a means to_______________________________________________________ troubleshoot the cleaning process. A stainless steel Mixing vessel was used to simulate a manufacturing Feeder top 0 1 vessel. High purity water recirculated through the vassel Hopper 6 1 was tested to residual carbon using a CSM-5000 continu- Lower feed chamber 9 0 ous sampling module attached to the Shimadzu TOC- Feed screw 2 0 5000. After a period of time during which the recycled Feed screw housing 16 0 water in the vessel was tested, a standard carbon Blade 5 0 solution was spiked into the vessel . This tested was ________________________________________________________ performed by NPOC analysis. The use of on-line TOC Could be for no-line quality control, cleaning processNeutral or slightly acidic injection of 0.2 N HCL are run auditing validation, and/ or troubleshooting the cleanApproximately every 20 samples, but since many of the ing process during method development, on-line TOCCleaning agents have PHs above 7, Hcl injection were would allow eximanition of the point at which the bulkRun more often. The results of this study, including the of the residual is removed by the cleaning processPpm carbon determined and percentage carbon for each thereby providing optimization of the cleaning cycle inPrepared solution, can be seen in table XVI. The major terms of wash and rinse time.Components of each cleaning agents are listed for each A study of different cleaning agents rinse cycle wasTrade name cleaning agent. Most of the cleaning agents conducted using this on-line approach. For three differ-Tested could be detected at the 2_20 ppm carbon range ent rinse cycle the total carbon, chloride ion andFor a 100 ppm solution. The only exception was Delvak, phosphate level was measured by TOC and ion Chroma Krystall and sentol which could only be detected at tography in the rinse water from the process. This dataSolution concentration of 1000 ppm. Those cleaning ( Table XVII ) showed that two short cycle were moreAgents with low carbon percentage may not be suitable effective thanone longer rinse cycle of removing bothCandidates for detection by TOC. Other methods such drug and cleaning agents. This type of testing allows theAs ion chromatography may be more appropriate in such ability to actually study the effectiveness of the cleaningCases. Process and thereby optimize the process.

On-line CapabilityCIP (cleaning in place) and COP (clean out of place)

System were first introduced in the dairy industry. ConclusionsThese system are used more frequently in our industry. Total Organic Carbon analysis provides a new methodThey provide a consistent level of cleaning. A few years for cleaning validation. The methods offer s low levelAgo they were primarily used in the parentral drug detection, rapid analysis time, and detects all carbonAreas. However, recent advances and decreased costs in

TABLE XVI Cleaning Agent Carbon Content Cleaning Concentration Cleaning Agent (ppm) Percent Agent (ppm) Composition Measured Carbon

Airkem F/ H 98.80 Quaternary ammonia 7.420 7.510Adept 105 Potassium hydroxide 4.363 4.16Alconox 100.2 Phosphates sodium carbonates 12.66 12.64 sulphonatesCIP 100 109 Potassium hydroxide 4.809 4.41Delvak 962 Sodium hydroxide tripolyphos- 14.37 1.49 phate sodium carbonate Grease Ease 104.8 Sodium metasilicate butyl cel- 11.45 10.92 losolveKrystall (immiscible in water) 925 Oil base 1.229 LT 1Resource 102 Sodium hydroxide 2.976 2.92SD-20 114 Sodium phosphate sulfonates 4.984 4.37 etyhoxylated amideSentol 1135 Phosphoric acid 4.625 LT 1Spray Nine 109 Sodium metasilicates 4.704 4.32SDT 99.60 Sodium dichlorotriazene 15.71 15.77 trioneSprex 102.3 Sodium carbonate sodium sili- 10.48 10.24 cates TBQ 104.0 Alkylamine ammonium chlo- 19.32 18.58 ride

PDA journal of pharmaceutical Science & Technology

TABLE XVII References Cleaning Process ___ Rinse Study

____________________________________________________________ 1. R. Baffi, et al; “ A total organic carbon analysis method for validat- Rinse Duration TOC CI Ion Phosphate ing cleaning between products in biopharmaceutical manufectur-

Cycle (minutes) (ppm) (ppb) (ppm) ing, “Journal of parentralScience and Technology,45 (1), 13 _19 ___________________________________________________________ (1991). Run 1 4 50 994 6.7 2. R.W. Mathews, M. Abdullah, and G. K.- C.Low, “ PhotacataltticRun 2 (cycle 1) 2 16 698 2.6 oxidation for total organic carbon analysis,” Analytica Chimica ActaRun 2(cycle 2) 2 6 220 1.4 233, 171__179 (1990)._________________________________________________________________________ 3. ASTM Method D4 129__82, “ Determination of dissolved organic Carbon,” ASTM book of standard, Philadelphia, Pa. (1984)based residual at a moderate costs. The method provide 4. Y. sugimura, Y. Suzuki, “ A high temperature catalytic oxidation

Linear detection over the range studied and acceptable method for the determination of non-volatile dissolved organic

recovery for various pharmaceutical tested. Data from carbon in seawater by direct injection of a liquid sample,” Marine actual production monitoring demonstrate that TOC Chemistry 24, 105__131 (1988).can be used effectively for cleaning validation, In short 5. FDA, “ Guide to inspection of cleaning validation process

TOC method can be developed, validated and utilized (Division of Field investigation, office of regional operarions,

for pharmaceutical cleaning validation. Office of rtegulatory affairs, july 1993).

6. J. K. Tylor, “ Quality assurance of chemical measurements,”Lewis publisher, Chelsea, MI, p. 78__84, 1987

7. Lowry, et al; Journal of bio.Chem; 1951, (193) , p. 265__ 275