J. Soc. Cosmet. Chem., 24, 135-146 (February 2, 1973)

Animal,Human,and Microbiological SafetyTestingof Cosmetic Products M. J. THOMAS, Ph.D., and P. A. MAJORS,M.S.* PresentedOctober27, 1971,beforethe CaliforniaChapter

Synopsis-Thethreeprincipalareasof TESTING pertinentto COSMETICS are discussed.

These include SAFETY determination in ANIMALS, safety and efficacystudiesin HUMANS, and MICROBIOLOGICAL STUDIES for safetyand productstability.Products must be shownsafeto animalsprior to use on humans.Most animal studiesinclude, for the most part, acute and subacutestudies.The former are done in accordancewith

the FederalHazardousSubstances Act (FHSA) requirements. There are a numberof patchtestsapplicableto humansafetytestingof products.The modifiedDraize techniqueappearsto be the preferredprocedure.It servesas a good predictorof sensitizing potentialand suppliesinformationon the primaryirritant characteristicsand cmnulativeprimaryirritantpropertiesof the test materials.A test program is also presentedfor determinationof adequacyof PRESERVATIVES in new and current cosmetic products.

INTRODUCTION

The increasing emphasis placedby our Government on the safetyof foods, drugs,andsimilarproducts hasbeenexpanded intothe areaof cosmetic productsin the last few years.Thosein the independentresearchand testinglab-

oratoryfieldsare verymuchawareof thischangebecausetheyhavebeenan integralpart of the developingprogramand have seenthe variousregulations evolve.Just as the regulatoryrequirementsfor safetyin food additivesand drug productsbecamemoredefinitiveas time passed,so will the regulatory requirementsfor cosmeticproductsbe more specificand morepronouncedas time goeson.

The objectivehere is to summarizethe presenttrendsand methodsused in testingfor the safetyand efficacyof cosmeticproducts.Recent action of the *Hill Top Research, Inc., Miamiville, Ohio 45147. 135

136

JOURNAL

OF THE

SOCIETY

OF COSMETIC

CHEMISTS

FDA relative to voluntary disclosureof productingredientswill not be discussedin detail. Accompanyingthe publicationof the ingredientdisclosure regulationswas the unexpectedchangein the classificationof 13 cosmetic product categoriesas also being drugs. This reclassification does not necessarily mean that additionaltestswill be required. Many cosmeticmanufacturers have productsafety standardswhich exceedthe standardswhich apply to new drug approvals. The one area of testing which will be affectedis the evaluationof these productson humans.Before this can be done, proper investigationalforms mustbe filed xviththe Food and Drug Administration.This requiressubmissionof detailedprotocolsfor review, and review of theseprotocolsby peer committees.

The three ,principalaras of testingpertinentto the cosmeticfield include (a) safetydeterminationsin animals,(b) safetyand efficacystudiesin human panelists,and (c) safetyand productstabilityfrom the microbiological standpoint. When a productis submittedfor humanpanelisttesting,it is mandatorythat sufficientanimaltoxicological data existto showthat the product is relativelynontoxicand will not induceexcessive reactionsin the testingon humans.After the producthasbeenshownto be safein animaland human panelisttesting,it is readyfor the market.However,there are still microbiologicalcontamination factorswhich shouldbe considered. During the manufacturingprocessthe product may be inadvertentlyexposedto unsanitary conditions in the plant or aroundequipment,or to contaminatedair entering

the plant, or to obiectionable microorganisms in the basic raw materials. ANIMAL

TOXICOLOGY

It is usuallyrecommended that the animaltoxicologyprogramsfollow the protocols specified by the FederalHazardousSubstances Act (FHSA) and its subsequent Regulations, publishedin the FederalRegister(1), and thoseof additionalstudieswhich are frequentlyrequired.

Theseprogramswfil require one or moretestsaccordingto the type of productinvolved.Theseare: 1. Acute oral toxicity-rats 2. Acute eye irritation-rabbits 3. Primaryskinirritation and corrosivity-rabbits 4. Acute dermaltoxicity-rabbits 5. Acute inhalationtoxicity-rats 6. Subacutedermal (21-day) toxicity-rabbits 7. Landsteinerskinsensitizationin guineapigs Acute Oral Toxicity

These studiesare conductedwith six groupsof five male albino rats (30

rats).A singleoraldoseof the testmaterialis administered at graduated dos-

SAFETY

TESTING

OF COSMETICS

137

agelevels.The ratsare observed for grosssignsof systemic toxicity,pharmacological effects, andmortalityat frequentintervals duringthedayof dosage, andat leastoncedailythereafterfor 14 days.Grossautopsies are performed on any animalsxvhiehdie. Acute Eye Irritation

Sixalbinorabbitsareusedin thesetests.One-tenthmilliliterof a liquidtest material,or 100 mg of solidsor pastes,is placedin one eye of eachof the rabbits.The oppositeeye is untreatedand servesas a control.Readingsfor eyeirritationare madeat 24, 48, and 72 hoursfollowingapplication.The scoring scalefor recordingeye irritationis that of Draize (2). Primary SkinIrritation and Corrosivity

The test material is applied under a 1-in.2 gauzepatch to both an intact and an abraded skin area on each of six albino rabbits, whose trunks have

beenclippedfree of fur. Four patchescan be applied,allowingfor two samplesto be testedon eachanimal. Acute Derreal Toxicity

This studyusesfour groupsof four albinorabbits.A singlederrealapplicationof the testmaterialis madeto eachgroupat graduateddosagelevels. The skinof two rabbitsper groupis abraded;the skinof the two remaining rabbitsper groupremainsintact.The test materialis appliedunder a rubber dental dam and remains in contact with the skin for 24 hours while the animals are restrained in stocks.

Acute Inhalation Toxicity

Ten male albinorats are exposedin an inhalationchamberfor onehour to an atmosphericconcentration .of20,000ppm of gas.orvapor or 200 mg/1. of mist or dust.The rats are observedduringthe exposureperiod and daily for 14 daysfor grosssignsof systemictoxicity.Grossautopsiesare performedon rats which die.

Standard criteria are used to evaluate the results of 'these various animal studies as shown in Table I.

SubacuteDerreal (21-Day) Toxicity

Thisstudyevaluates the percutaneous absorption andirritativepotentialof the testmaterialfollowingsubacutederrealapplicationto rabbits.The testis not specifiedby the regulationbut is frequentlycarriedout. Eighteenyoung adult albino rabbits are divided randomlyinto three groupsof six rabbits each.

Group1 is the controlgroupand receivesrepeatedskinapplicationsof dis-

tilledwater.Group2 receives repeatedskinapplications of the testmaterial

JOURNAL OF TIlE SOCIETY OF COSMETIC CHEMISTS Table

I

Evaluation of Toxicity in Animal Studies Route of Administration Oral

Derreal Inhalation

Highly Toxic

Toxic

Nontoxic

LD5o of 50 mg/kg or less LD,o >50 mg/kg but LD•,o> g/kg not >5 g/kg LD5oof 200 mg/kg or less LD,o >200 mg/kg but LD•o >2 g/kg not >2 g/kg LD5o of 200 ppm of gas or LD5o >200 ppm of gas LD,o >20,000 ppm vapor (2 rag/1. for mist or or vapor (2 mg/]. for of gas or vapor rag/1. of dust) or less during one- mist or dust) but not (>200 hour exposure more than 20,000 ppm mist or dust) (200 rag/1. of mist or dust)

at a low dosagelevel. Group3 receivesrepeatedskin applications of the test material at a high dosagelevel ( 10 timesthat of Group 2).

Three rabbitsin eachgroupare abradedat the beginningof the first, second, and third weeksof the study.The skin of the remainingrabbitsis left intact. The controland test materialsare applied oncedaily on a 5-day per weekbasisovera 3-weekperiod,for a total of 15 applications. Eachdaily ex-

posureperiodis for 6 to 8 hours.Followingthe 15 applications, the rabbits areobserved for 2 weeks,thensacrificed, andgrossnecropsies areperformed. During the studyeachanimalis weighedweeklyand observed daily for grosssignsof derrealirritationandsystemic toxicity.A completebloodcount, hemoglobindetermination, and urinalysisare performedinitially and at terminationof the study.At necropsy, representative tissues from eachcontrol andtestanimalare preservedin 10%formaldehyde.

Microscopic examination of tissues is madeon eachcontrolandhigh-level rabbit,includingthe skin,heart,liver,kidney,spleen,adrenal,stomach, small intestine,gonad,and bonemarrow.From the low-levelgroup,sectionsof skin,liver, and kidneyare examined. All remainingtissues from all animals are held for the possibilityof further micrscopieexaminations.

An evaluation whichfrequentlyis carriedoutin guineapigsis for detection of allergenic potentialof testmaterials. This procedure shouldbe applied solelyin the evaluationof allergenicpotentialof proposed productcomponents.The evaluation of proposed marketformulations isusuallyrestricted to testson humanpanelists.

In reviewingthe applications of the guineapig testfor sensitization, the generalopinionis that the productionof positivereactionsin guinea pigs shouldexcludethe material from further considerationfor usein preparations

whichwill be topicallyappliedto the skin,or it shouldindicatecautionin subsequent testson humanswhichmayfollowin orderto determineconcen-

SAFETY

TESTING

OF

COSMETICS

139

trationswhich can be e•nployedwithout inducing sensitizationby finished formulations.

Experiencesof someinvestigators, however,have shownsomefalse positive results;that is, sensitization seeminglyinducedin the guineapig that was not confirmedby subsequent testsin humans.The possibilityof suchdifferences in opinionsis readily apparentwhen oneconsidersthe variationsin procedure which have been employedin the testson guineapigs and on humans.There are asmany,probablymore,variationsin the predictivepatch test on humans asthereare in the testsemployingguineapigs. Recentreviewsof the importanceof predictivetestingand of the use of guinea pigs and humansin predictive testing include one by Griffith and Buehler (3). In this review, someof the findingsof Buehlerwere analyzed. The studieswere carriedout to determinethe relativesensitivityof variations in proceduresutilizingguineapigs. The techniquescomparedwere the intradermalinjectionmethodof Landsteinerand Jacobs(4) and a new procedurewhich employeda closedpatch. In thisseriesof studies,it was determinedthat the closed-patch proceduredetectedsensitizing potentialof eightmaterialsknownto be contactallergens, while onlytwo of the eightweredetectedby the intradermalprocedure.The applicationscheduleand test materialconcentration were the samein both series.This differencein sensitivitywas confirmedby additionalstudiesin which sensitizationof 100% of the animalswas obtainedby the patch procedure with concentrations of dinitroehlorobenzene (DNCB) which failed to inducesensitization of any animalstestedby the intradermaltechnique. The value of the intradermaltechniquecannotbe questioned,but it has

limitations.With manymaterials,the concentrations whichcanbe employed withoutproducingsevereprimaryirritation,includingnecrosis, are limitedto levelssomewhatbelowthe practicaluse level of the material.An equally importantfactoris the dittleultyin interpretingthe reactions to the intraderreal injections.Unfortunately,sensitizationreactionsin guinea pigs are not

characterized by the typicalgrossly visiblepapularor vesieularreactionscommonlyobservedin humans.Judgmentmustbe madeon the basisof the extent of erythematous or mfi•imaledematous reactions andcomparing the challenge reactionsto similar but less extensivereactionsto initial injections.When

properlyapplied,however,it is a valuabletoolfor the detectionof strong sensitizers.

HUMAN CLINICAL STUDIES

After the animaltoxicologicalstudiesare completedand the producthas been shown to be safe for human clinical studies,the latter are then initiated.

Clinical studiesperformedwith cosmetics on humansare referredto by our laboratoriesas proprietaryclinicalstudiesto distinguishthem from clinical studieswhich are carried out on ethical drugs.

140

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Patch Tests

Two basicprocedureshave been mostfrequently employedand numerous variationsof thesehave been used.One of the basicprocedures,the Schwartz and Peck (5), has limited value sinceit will detect only very strongsensitiz-

ers.Presentconsideration of the humanpatchtestwill primarilybe limited to the repeatedinsult procedureand its variousmodifications.Two variations were originallypresented-oneby Shelanskiand Shelanski(6) and the other by Draize (2). Both of fl•eseproceduresemployrepeatedapplicationslnade over a period of 4 or 5 weeks.They differ in that 10 sensitizingapplications are madein the Draize procedureand 15 are made in the procedureof Shelanski.They alsodifferin that challengeapplications are madeto the original testsitesonly in the Draize procedureand are madeto the originaltestsite andto previouslyunpatchedsitesin the procedureof Shelanski. A shortreview of theseand otherproceduresemployinghumanpanelists hasbeenpresentedby Kligman (7), who suggested the maximizationprocedure as an alternative.Variousmethodsfor achievingthe objectiveof Kligman'smaximization procedurehavebeenemployed.We useda procedurein whichthe skinis abradedprior to sampleapplication.The testmaterialsare appliedto abradedsitesunderpatcheswhich are appliedto the upper arms.

Reactions to the testapplications arescoredon eachWednesday, Friday,and Mondayfollowingapplications on the preceding Monday,Wednesday, and Friday.The patchesare removedby the panelists 9,4hoursafterapplication. Nineapplications aremadein 3 successive weeks,andthechallenge application is madeon Mondayof the sixthweek.The challengeapplications are madeto previouslyunpatchedsites. Abrasions of the testsitesare madejustpriorto sampleapplications on test

days1, 4, 7, and11, or preferably, priorto eachapplication. Thismorefrequentschedule of abradingcanbe donewhenup to fourmaterials areevaluatedon abradedskinonly.The abrasions are in the form of two concentric circleswith diametersof % in. and "%in. and are inflictedby a speciallycon-

structed apparatus. The skinareasarecleansed with isopropyl alcoholjust priorto abrading. A separate sterileabraderisusedfor eachpanelist. All adhesive patches andabsorbent padsappliedoverabraded areasaresterilized byexposure to ethylene oxidepriortouse.Theusualgrading scaleisutilized. Anotherprocedure employed is to stripthe skinby repeatedapplication andremovalof adhesive tapefromthetestsitespriorto sampleapplications. All of theseprocedures havevaluein certainapplications. The onechosen should providetherequired degree of sensitivity aswellasdatafromwhich valid decisions can be derived.

A variation of therepeated insultpatchtesthasprovedto be themostvaluableprocedure available. Mostof the criticisms of the basicprocedure shouldhavebeencriticism of thewayin whichtheprocedurc •vascarriedout.

Products arefrequently formulated onthepremise thatthiscanbe safely

SAFETY TESTING

OF COSMETICS

141

doneprovidedthat all components havea significant historyof safeuse.This is trueonlyif thereis nosignificant deviation frmnprovenuseof all components,bothasto typeof productandingredientconcentration. It mustalsobe recognizedthat there alwaysis the possibilityof synergisticactionbetween components.

The importanceof evaluatingthe productin its finishedform is e•nphasizedby observations of manyevaluators. Oneof our •noststrikingpersonal experiences was with triehlorosalieylanilide (TCSA). At the ti•ne of its intro-

ductionand prmnotion as an antibacterial agentfor topicalapplications, it was not recognizedthat it was a potent contactas well as a photoallergen. We evaluateda cleansing lotioncontaining1% TCSA in a repeatedinsult patchtest.About5% of the panelistsdevelopedstrongcontactsensitization to the .product. It wasestablished, in subsequent reehallenges, that TCSA was the sensitizingagent.A surprisingresultof thesereehallenges wasthat the responseto TCSA was evokedwhen dispersions were preparedin water by means of a low concentrationof a surfaetant,but was not evoked frmn these samereactivepanelistswhen they were challengedwith solutionsof TCSA in

diethylphthalate. The originaltestsfor sensitizing potentialwere carriedout on solutionsin thissolvent.We frequentlyare proneto assrune that the preferredsolventis onein whichthe •naterialunderinvestigation is highlysoluble. The above instance,as well as otherswhich have been encountered,indi-

catesthe necessityof any candidateproduct additivebeing evaluatedin finishedformulationsor in producttypessi•nilarto finishedformulations. The failure to detectthe sensitization potentialof TCSA in initial testswas not due to the inabilityof the procedureto detectthis property.It wasdue to the failureto insultthe skinwith preparations havingthe properskinpenetration characteristics.

Other

variations

have been encountered

in which

the

mnountsof test •naterial applied were much reduced-as much as tenfoldfrownthe 0.5-g amountsmostinvestigators apply. It is not possibleto considerthe •neritsof the variousmodificationsof the repeatedinsult patch test procedure.This would require lengthy discussions based on extensiveexperiencewith each modification.These •nodifieations have mostfrequentlybeen e•nployedto satisfythe proposeduse or nature of the product.We prefernot to deviatefrom our standardprocedurein which we employclosedpatchesand follow basicallythe procedureof Draize with the challengeprocedureof Shelanski.It is frequently necessaryto employ se•ni-openor openpatchesor to resortto short exposuresachievedby panelistssittingwith the testmaterialon the flexorsurfaceof their forearms.These variationshavebeenrequiredbecauseof the volatileor irritant natureof the product.We haveinducedtypicalsensitization in numerousinstances by all degreesof occlusionemployed. A significantcauseof possibleerroneousresultsis that of eross-sensitization reactions.It is •vell recognizedthat materialswhich in themselvesare not sensitizers canevoketypicalsensitization reactionsfrom individualswho have

142

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

beensensitizedby chemicallysimilarmaterialswhich are sensitizers. For this reasonproductsfrom different clientsshouldnot be combinedon the same panelists.The one exceptionis joint studieswhich are carried out for ingredient suppliersand formulatorsof finishedproducts.In suchstudiesthe proposedadditivemay be si•nultaneously testedin severalformulatedproducts. Regardless of the procedureemployedin testson humanpanelists,minimal

reactions will be frequentlyevokedduringthe serialapplications and at challenge,whichrequirethe applicationof expertiseacquiredby yearsof observationsto make proper judgmentsas to whetherthe materialis a sensitizer or not. It is frequentlynecessary to conductrechallenges severalweeksfollowingthe initial challengein orderto assess the significance of the observed reactions.

The interpretationof test resultsof the repeatedinsult patch test procedureswill notbe discussed hereat lengthsincethisbecomesa rathercomplex endeavor. Thereis markedvariationin the scoringsystems employed.We employ a seeminglycomplexscoringscale(Table II) which has developed throughthe additionof designations asthe needfor thembecameevident. Table

II

Patch Test Reaction Scoring Scale 0

No

irritation

1 Slight erythema 2 Marked erythema 3 Erythema and papules E Erythema and edema

4

Erythema, papules and edema

5

Vesicles

6

Strong spreading reaction

A B C F G

Effect on Superficial Layers of Skin Slight glaze Marked glaze Glazing with cracking and peeling Moist fissures Film of serous exudate

Whenproperlyapplied,therepeated insultpatchtestis a severetestforthe sensitizing potentialof products. Therearemanyproducts whichwill induce sensitization in a low percentage of panelists testedby thisprocedure which can be usedwithoutuntowardreactions.The proposedusemustbe consid-

ered.Forexample, a product whichshows minimal sensitization potential and is to be usedon the feet shouldnotbe marketed,while a productwhichwould

beinfrequently usedundernonocclusive conditions couldprobably besafely marketed.

Someassurance thatproducts foundto be minimalsensitizers canbe safely marketedcanbe obtainedby conducting product-use challenges of panelists

SAFETY

TESTING

OF COSMETICS

143

who have been sensitizedin the patch test procedure.If thesereactive individualsdo not respondto useapplications,it is mostunlikelythat sensitization would be induced by normal use of the product. However, this reasoning cannotbe appliedto productsusedfor whole-bodyapplications.Whole-body exposures by sensitizedpanelistscannotbe safelydone, and limited area applicationsmay be misleading. In additionto the value of the repeatedinsult patch test as a predictorof sensitizingpotential,it furnishesvaluableinformationon the primaryirritant characteristicsand cumulativeprimary irritant propertiesof test materials. The latter propertyis descriptivelyreferredto frequentlyas"skinfatigue." In summaryof our consideration of the evaluationof sensitizingpotential, guineapig sensitizationstudiesare valuablein the detectionof strongsensitizers.The closedpatch procedureof Buehler (3) is much more sensitivethan the intradermalinjectionprocedure.Productsshowingdefinite evidenceof sensitizationin guineapigs shouldbe droppedfrom further considerationunlessthey can be employedin concentrations significantlybelow the minimal sensitizinglevel for guineapigs. Irrespectivcof negativeresultsobtainedin testson guineapigs, test materials with no historyof toleranceby humansshouldbe initially testedin the repeatedinsultpatch test procedurein a pilot studyon 10-12panelists.Fullscalerepeatedinsultpatchtestsshouldbe carriedout on all productsprior to release to market.

In approximately20 yearsof experiencethe presentauthorshave never encountereda severeincidenceof sensitizingpotentialin full-scale60-200panelisttestswhichwasnot predictedby the pilot study.A pilot repeatedinsult patch test on 10-12 human panelistsis of much more value than guinea pig testsas a screeningtest for finishedproducts. The total number of human panelistswhich shouldbe employedin tests shouldbe governedto someextentby the type and extentof exposureto the product which will be encounteredin use. For materialswhich will be employedin nonocclusive uses,suchasperfumes,hand and face lotions,etc., and will be appliedto limited body areas,completelynegativeresultsin testson 60 panelistshaveprovedto be adequateassurance that suchproductswill be essentiallyinnocuousunder use conditions.Productswhich are used under occlusivetestconditionsshouldbe evaluatedon largernumbersof panelists. It is also desirableto carry out controlleduse studiesof such products. Theseshouldbe precededby a repeatedinsultpatchteston 60 panelistsand shouldbe designedso that one groupof panelistsuse the productsomewhat in excessof recommended use.Challengepatclaesof the test materialshould be appliedat the end of the useperiod. The discussion of patchtestingabovehasbeenrestrictedto testsdesigned primarily for the detectionof sensitizingpotentialof formulationsor ingredients.Many cosmetics produceno primary irritationwhen evaluatedin the

aboveprocedures. In a continuous patch procedurepresentedby Lanman

144

JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

(8), a total of asmanyas8 samplescanbe appliedto the backof eachpanelist.Six of theseare test samplesand two shouldbe referencesamplesof knownirritativepotential. The testpatchis a 1-in?Webrilswatchmoistened with 0.5 ml of liquidor 0.2 ml of semisolid samples. The swatchis covered with Blendermtape. The patchesare removedeachday and the sitesare scoredfor irritation.Freshpatchesare applied immediately.This routineis followedeachdayuntil the firstsignsof erythemaareseenor for a maximum of 21 days.As eachsampleproduces erythemaits application to the affected subjectis discontinued. At the conclusion the IT;o (numberof daysrequired to produceerythemain 50%of the subjects)is calculated for eachsample.By this procedureminimaldifferences in irritant potentialof essentiallyinnocuousformulations canfrequentlybe demonstrated. MICROBIAL STUDIES

The importance of controlof the microbialqualityof formulations needsno emphasis asthisis an areaof whichwe all are acutelyaware.In additionto the healthhazardto consumers from contaminated products,the effectof microbialgrowthon productqualityis alsoan importantfactor.A comprehensivereviewof thisproblemwaspresented by Bruch(9) givingin somedetail the procedures for evaluating microbialqualityof cosmetics andthe testing programs whichshouldbe employedto assurethe manufacturers that their productsare adequatelypreservedand that they are producedunderGMP guidelines. Sinceit is possiblethat preservativesmay act as sensitizers,or irritants to humanskin,both animaltoxicologyand human tests should be conducted be-

fore cosmeticproductsare releasedfor sale. In addition to a recommended testingprogramfor cosmeticproducts,it is highly desirableto monitorthe of the plant manmicrobiological qualityof raxvmaterialsand the cleanliness

ufacturing facilities. Thisis doneby the collection of swabsamples fromvariousareasof the plant,the processing equipmentand containers in orderto evaluatesanitizing procedures, andby sampling raw materials for microbiologicalcontent. It is of primeimportance alsoto carefullyinspectfor other contamination possibfiities duetothegeneral location andlayoutof thebuilding,andforair contaminants originating fromplantsin thenearbyarea.From information obtained by theseprocedures, properactioncanbe takento correct any factorscausingcontamination.

Thetesting program weuseis designed to determine theadequacy of preservatives presentin newandcurrentcosmetic products. Its objectives are: 1. To determine if products in currentproduction aresufficiently biostatic to preventthe growthof microorganisms, and to determine for new

products thelevelsof biostatic agents required to rendersuchproducts suitablybiostatic for extendedperiodsof time. 2. To determine if the products in currentproduction arefree of microorganisms whicharepotentiallyhumanpathogens.

SAFETY

TESTING

OF

COSMETICS

14,5

Therefore,the procedureconsists of two phases: 1. Evaluationof biostaticactivityof the completeformulations. 2. The examinationof productionsamplesfor the presenceof viable organisms.

Evaluation o1'BiostaticActivity

The efficacyof the preservativein a preparationis determinedin the laboratoryby the additionof pure or mixed culturesof microorganisms to the finishedproduct.The initial microbialpopulationof the inoculatedproductis determinedby plating aliquotsof suitabledilutionsin agar media.Countsof viable microorganisms are redeterminedat selectedtime intervals.The initial populationsin inoculatedproductsshouldbe about5 x 105 per ml or g. Aliquotsof inoculatedproductsare plated at varioustestingintervalsto determinethe microbialpopulation.Variousneutralizersfor the preservative present in the formulationcan be addedto the plating media to insuremaximal recoveryof the challengeorganisms. The inoculumto be usedshouldl'•epreparedby mixing pure culturesof organismsisolatedh'om spoiledor grosslycontaminatedproducts.In the eventthisprocedureis not feasible,pure culturesof organisms which are suspeckas contaminantsof suchproductscan be mixed to prepare the incoulum. The most commonlyencounteredcontaminantsin industrial products are membersof the genusPseudomonas. This organislnis generallythe mostdifficult to controlin the productsto be tested.The inhibitory actionof the products shouldalsobe challengedby suchorganismsas Aspergillusniger, Bacillus subtilis,Staphylococcus aureus,Escherichiacoli, and otherswhich are ca-

pable of causingdiseaseor productspoilage.Samplingintervalsshouldbe such that rapid growth of contaminantsis detected,as well as very slow growth.In somematerialscontaminants will proliferatevery rapidly duringa relativelyshorttime (one week). This rapid increasemay then be followedby a gradualdecreasein numbersof viable organisms to a level below that originally present.In both of thesegrowthpatterns,countsshouldbe performed immediately,after one week, and again one month after inocualtion;and repeatedat 4-monthintervalsfor 12 months.If it is desiredto determinepreservative adequacyonly over a shortperiod of time, terminationof testing at one month should suffice.

A secondchallengewith the inoculummay be indicatedwith someproducts.Productswhoseusagefrom onecontainermay be distributedoverlong periods(severalmonths) would be in this category.The biostaticactivity may diminishafter prolongedstorage.Productswhich are adequatelyprotected at the time of productionmay permit growth of contaminantsintroducedduringextendeduse.If the productis usedup shortlyafter opening, this problemis eliminated.

146

JOURNAL OF THE SOCIETY OF COSMETIC

CHEMISTS

SUMMARY

A shortresumehasbeen presentedof the varioustestsfor productsafety which are commonlycarried out on cosmetics.There is one very important

questionto which an answerhasnot been attempted.After one has carried out all the requiredtests,what assurance is therethat the productwill be safe to use?This is very difficultto predict.A practice,whichshouldbe followed andwhichwill facilitatesuchpredictions, is to alwaysincludein anyseriesof testson new productsa referencesample.This preferablyshouldbe one of your own similarproductswhichhas an acceptablehistoryof freedomfrom complaints. Competitors' samples canalsobe used,but the complaintrecords of suchsampleswill not be available. ( ReceivedApril 14, 1972) REFERENCES

(1) Federal Register,General ServicesAdministration,Washington,D.C., Vol. 9.9, September 17, 1964.

(2) Draize,J. H., in Appraisalof the Safetyof Chemicalsin Foods,Drugsand Cosmetics, Associationof Food and Drug Officialsof the United States,Austin, Tex., 1959. (3) Griffith, J. E., and Buehler,E. V., Experimentalskin sensitizationin the guinea pig and man, Presentedat the 28th Annual Meeting, AmericanAcademyof Dermatology, Bal Harbour, Fla., December9, 1969. (4) Landsteiner,K., and Jacobs,J., Studieson sensitization of animalswith simplechemical compounds,J. Exp. Med., 61, 643-56 (1935). (5) Schwartz,L., and Peek, S. M., The patch test in contactdermatitis,Public Health Rev., 59, 2 (1944).

(6) Shelanski,H. A., and Shelanski,M. V., A new techniqueof human patch tests,Proc. Sci. Sect. Toilet GoodsAss.,19, 46-9 (1953). (7) Kligman, A.M., The identification of eontaet allergens by human assay, J. Invest. DermatoL, 47, 369-74 (1966).

(8) Lanman,B. M., Elvers, W. B., and Howard, C. S., The role of humanpatch testing in a produetdevelopment program,Proc.Joint Conf. Cosmet.Sci., Washington,D.C., 135-45 (1968). (9) Bruch, C. W., Cosmetics;sterility vs. microbial control, Amer. Perrum. Cosmet., 86, 45-50 (April 1971).