assessment - a bmj journal

Occupational and Environmental Medicine 1994;51:371-380

A study of chromium induced allergic contactdermatitis with 54 volunteers: implications forenvironmental risk assessment

J Nethercott, D Paustenbach, R Adams, J Fowler, J Marks, C Morton, J Taylor,S Horowitz, B Finley

Chem Risk Division,McLaren/HartEnvironmentalEngineering,Alameda, CA, USAD PaustenbachS HorowitzB FinleyJohns HopkinsUniversity, Baltimore,MD, USAJ NethercottStanford University,Palo Alto, California,USAR AdamsUniversity ofLouisville, Louisville,Kentucky, USAJ FowlerPennsylvania StateUniversity, Hershey,Pennsylvania, USAJ MarksUniversity ofBritishColumbia, Vancouver,CanadaC MortonCleveland ClinicFoundation,Cleveland, Ohio, USAJ TaylorCorrespondence to:Dr D J Paustenbach,Chem Risk Division,McLaren/HartEnvironmental Engineering,1135 Atlantic Avenue,Alameda, CA, USA.

Accepted for publication1 March 1994

AbstractOver the past 60 years, dose-responsepatch test studies by various methodshave been conducted in an attempt toidentify the mi imum elicitation thresh-old (MET) concentration of hexavalentchromium (Cr(VI)) that produces anallergic response in Cr(VI) sensitive sub-jects. These data are not adequate, how-ever, to provide an accurate estimate ofthe MET because of the variability in thepatch testing techniques and the variabil-ity in diagnostic criteria used. Further-more, the data were not reported interms of mass of allergen per surfacearea of skin (mg Cricm2-skin), which isnecessary for conducting occupational orenvironmental health risk assessments.Thus the purpose of this study was todetermine the MET (mg allergenlcm2) forCr(VI) and trivalent chromium (Cr(III))by patch testing techniques. A patch testmethod that delivers a controlled amountof allergen per surface area of skin wasused. A group of 54 Cr(VI) sensitised vol-unteers were patch tested with serialdilutions of Cr(VI) and Cr(III) to deter-mine the cumulative response rate at sev-eral concentrations. The results indicatethat the 10% MET for Cr(VI) based onthe cumulative response was 0-089 pgCr(VI)/cm2-skin. Only one of the 54volunteers may have responded to 33,ugCr(III)/cm2-skin, otherwise Cr(III) wasunable to produce allergic contact der-matitis in these highly sensitive volun-teers. Two supplemental studies werealso conducted to assess whether the sur-face area of the patch and the concentra-tion of Cr(VI) in the patch (related topatch thickness) were likely to influencethe results. The data from these studieswere used to assess the risk of developingallergic contact dermatitis due to contactwith Cr(VI) and Cr(III) in soil. The find-ings indicated that soil concentrations atleast as high as 450 ppm Cr(VI) and165 000 ppm Cr(Ill) should not pose anallergic contact dermatitis hazard for atleast 99*99%/o of the people in the commu-nity who might be exposed.

(Occup Environ Med 1994;51:371-380)

Hexavalent chromium Cr(VI) is one of themost common dermal sensitisers in theoccupational setting and accounts for about

5% of clinically reported cases of allergic con-tact dermatitis in the United States.' Cr(VI)related allergic contact dermatitis has beenreported in chromium plating workers,23 litho-graphers,4 diesel repair shop workers,5 andleather workers.6 Additionally, Cr(VI) inhousehold products such as bleaches anddetergents, cosmetics, and shoe polish hasbeen cited as a potential cause of allergic con-tact dermatitis.' I In many areas, there can beappreciable human exposure to chromite oreprocessing residue (COPR) and mine tailingswhen either is mixed with soil.8

Cr(VI) induced allergic contact dermatitisis known as a type IV, delayed, or cell medi-ated allergic reaction.' 29 The localised bio-logical response of allergic contact dermatitisis similar to a "poison oak" hypersensitivereaction, and elicits the standard symptoms oferythema, oedema, and small vesicles.' 9 10Type IV allergic dermatitis reactions are mostoften not life threatening and their effect isgenerally limited to the skin. Sensitisation is athreshold response where single or repeatedexposure to low doses of an allergen may notproduce an allergic response if the thresholddose is not reached.8 Epidermal contact withhigh concentrations of Cr(VI) can also pro-duce irritant contact dermatitis, which is aheterogeneous symptom wherein a chemicalinduces a non-immunological dermatitis.""-'A cutaneous irritant causes direct damage tothe skin without prior sensitisation.' 12

It has been known since 1925 that dermalcontact with Cr(VI) can elicit allergic der-matoses.'4 Due to the strong sensitisingpotential of Cr(VI) and the desire to reducethe incidence of allergies in workers, severalstudies involving patch testing have been con-ducted to determine the "minimum elicitationthreshold" (MET) of Cr(VI) in sensitised per-sons."52" In those studies, people known to beCr(VI) sensitive were tested with patches con-taining serial dilutions of Cr(VI) (usually aspotassium dichromate (K2Cr2O,)) in petro-leum jelly, water, and/or acid-glycine.

Dose-response studies of dermal sensitisersconducted before 1988 were much lesscomplex than those used today. For example,much of the existing patch test data forCr(VI) were collected before the improvedand standardised diagnostic criteria developedby the North American Contact DermatitisGroup (NACDG)22 and the InternationalContact Dermatitis Research Group(ICDRG).2' One deficiency in prior studies isbelieved to be that some irritant reactionswere scored as an allergic response. Early

371

on January 1, 2022 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/

Nethercott, Paustenbach, Adams, Fowler, Marks, Morton, Taylor, Horowitz, Finley

reports of Cr(VI) patch test data often failedto disclose information regarding the diagnos-tic criteria to determine allergy, duration ofapplication, and the analytical methods usedto validate the chromium concentration andvalency state. Also, it is known that the patchpreparation methods were inconsistent andthat interpatch variability of the amount ofCr(VI) applied could be as high as an order ofmagnitude.24 In some of these previous stud-ies, 1-21 a patch test was simply a gauze padthat was immersed into the aqueous Cr(VI)solution before application, an open area ofskin covered with tape, or a measured amountof Cr(VI)-petroleum mixture loaded intoFinn chambers. At least two recent papershave attempted to identify a MET for Cr(VI)-induced allergic contact dermatitis from thesestudies. Based on a statistical analysis of patchtest data, it has been suggested that 10 ppmCr(VI) was the MET at the 10% responselevel.25 26 Paustenbach et a127 examined thedata from the same studies and suggested thatthe 10% MET should be about 50 ppmCr(VI) in solution.

Certainly, the most important deficiency inthese studies with respect to their usefulnessin environmental risk assessment is that thedata were not reported in terms of mass ofallergen per unit area (for example, mg/cm2).As has been discussed elsewhere, by contrastwith using patch tests to identify thosechemicals to which a person is sensitised (aqualitative decision), in order to performhealth risk assessments patch testing datamust be presented in terms ofmg of chemicalper skin area.24 28-31 For example, when deter-mining acceptable surface concentrations fortoxicants on walls, process equipment, or insoils, uptake per unit area of skin (for exam-ple, mg/cm2-skin) is needed.32 This is impor-tant because at least one regulatory agencyhas suggested that allergic contact dermatitisis a proper health endpoint for regulating levelsof Cr(VI) in soils.25 Because the existing datado not seem to provide an accurate estimateof the Cr(VI) MET and the data are not ade-quate to quantitatively predict the health risksrelated to Cr(VI) in contaminated media, weconducted a carefully controlled study thatused dose per unit area as dosimetric index.

This paper describes the results of a patchtest study designed to determine the areabased MET (ug Cr/cm2 skin) of solubilisedCr(VI) and Cr(III) that will elicit allergic con-tact dermatitis in chromium sensitised sub-jects. Under the direction of several membersof the NACDG, a group of 54 volunteersknown to be sensitised to Cr(VI) were patchtested with serial dilutions of Cr(VI) andCr(III). To reduce the variability inherent in

Table 1 Chromium (Cr(VI) and Cr(III)) concentrations used in the patch testing study

Potassium dichromate Concentration per unit Chromium trichloride Concentration per unit(pg K2Cr207/cm2) area (,pg Cr(VI)/cm2) (pg CrC13/cm2) area (pigCr(III)/cm')12-5 (diagnostic) 4-4 100 33-02-5 0-88 20 6-60-5 0-18 10 3-30-25 0-088 2 0-660-05 0-018

earlier patch preparation methods, "TRUE-Test" patches were specifically manufacturedfor use in this study.

Materials and methodsIn general, as the water solubility of aCr(VI) salt increases, its ability to penetratethe skin barrier and elicit allergic contact der-matitis increases."3 Due to its high degree ofwater solubility, K2Cr2O7 is one of the mostpenetrating and therefore most potentiallyreactive Cr(VI) species.834 Potassium dichro-mate is currently used by members of theNACDG and ICDRG as the standard Cr(VI)patch test agent for diagnostic purposes.2' 35 36Accordingly, for the purposes of this study,K2Cr2O, was chosen as the test Cr(VI) com-pound. Trivalent Cr compounds are not rou-tinely used for clinical patch testing becauseCr(III) is considered to have little or no allergicpotential.37-39 This is due, in part, to the lowwater solubility of most Cr(III) compoundsand the long history of uneventful occupa-tional exposure." Chromium trichloride(CrCl3) was used in this study because it isone of the most water soluble Cr(III)species.'4 Chromium trichloride has also beenused in some earlier patch testing studies.41 42TRUE-Test (thin layer rapid use epicuta-

neous test) gel matrix patches were manufac-tured by Kabi Pharmacia Research CenterAS, Inc, Hillerod, Denmark. Table 1 presentsthe patch concentrations of K2Cr2O7 andCrCl3 used in this study. Patches containingK2Cr2O7 were prepared by mixing K2Cr2O7(purity 98.5-101.5%) with a wet hydrox-ypropyl cellulose gel to the specified concen-trations. The patches containing CrCl3 wereprepared by mixing CrCl3 with a wet polyvi-done gel to the appropriate concentrations.These gels have little or no sensitising poten-tial.24 36 43 The allergens were mixed to a speci-fied mass of Cr(VI) or Cr(III) per unit area,printed on a sheet of polyester, and dried to athin film. These coated water impermeablesheets were then cut into square patches of0-81 cm2, mounted on a piece of adhesive,non-allergenic tape, and packaged in an air-tight and light impermeable envelope. TheK2Cr2O7 test patches were also sealed withdesiccant paper to prevent adsorption ofmoisture.The TRUE-Test patches are specifically

designed to hydrate by perspiration whentaped to the skin under occlusion. The driedfilm is hydrated into a gel thickness of 50 to70 gm from which the allergen migrates to theskin. The hydrated gel, occluded by adhesivebacking and plastic, ensures maximal contactwith the skin, thus enabling high allergenbioavailability." With this approach, the aller-gen is evenly distributed over the test area andthe quantitative dose of allergen challenge isaccurately controlled.24 30 This provides asignificant advantage over other current tech-niques, such as the Finn chamber, in whichthe mass of allergen loaded on to the skin mayvary by up to an order of magnitude frompatch to patch.24

372


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/

A study ofchromium induced allergic contact dermatitis wih 54 volunteers: implications for environmental risk assessment

ANALYTICAL VALIDATION OF PATCHCONCENTRATIONSAnalytical confirmation of the patch concen-trations was conducted before performing thestudy to ensure that each patch contained theappropriate valency and concentration of Cr.The concentrations of Cr(VI) in the patchwere determined with Modified NIOSHMethod 760045 as the extraction procedureand inductively coupled-visible absorptionspectrometry (IC-VAS)46 to measure the solu-bilised Cr(VI). Modified NIOSH Method7600 is essentially a modified version of EPAMethod 3060,47 with slight differences inpreparation and buffer solutions. A singlepatch was immersed in 25 ml of extractionsolution (1-7 g of 0-2 M sodium bicarbonatein 1-0 litre distilled water). The polyethylenebottle was capped and ultrasonicated forabout three hours, at which time the bottleswere removed and placed in a rotating shakerfor 48 hours at room temperature. Theextractant was then injected into an ion chro-matograph where the Cr(VI) was eluted withan ammonium sulphate and ammoniumhydroxide eluent mixture. After elution fromthe column, the Cr(VI) was reacted with 1,5-diphenylcarbohydrazide (DPC) and the peakarea absorbance of the coloured DPC-Cr(VI)complex was photometrically measured at520 nm. The CrCl3 patches were alsoanalysed by the modified NIOSH Method7600/IC-VAS to ensure that Cr(VI) was notpresent in the Cr(III) patches.The concentrations of Cr(III) in the

patches were determined by EPA Method305048 for extraction and EPA Method 601049for analysis. A modified EPA Method 3050was used to extract the Cr(III) patches.Specifically, single patches were added to 1-5ml of 0 1 M sodium hydroxide and heated to85'C for one hour, followed by the additionof 300,ul 30% hydrogen peroxide and heat foranother hour. Two millilitres of 0 1 Mhydrochloric acid were added to the digestantafter heating and the solution was cooled toroom temperature. Another 2-2 ml 0-1 Mhydrochloric acid was then added to the solu-tion and the volume was adjusted to 25 mlwith purified water. The patch digestantswere then analysed for total Cr (as CrCl,) byEPA Method 6010, or Inductively CoupledPlasma (ICP). The EPA Method 6010 mea-sures element emitted light by optical spec-trometry at 267-7 nm. After the CrCl, isdigested, the ICP column is calibrated withstock solutions of 2 ml (1: 1) HNO and 10 ml(1: 1) HCL and diluted to 100 ml with type IIwater. The digestant is directly injected intothe column (about 2 ml) through the torchand a direct reading is taken from theinstrument.

Representative patches of each Cr concen-tration were analysed in duplicate to evaluateinterpatch variability. As part of the confirma-tion analyses, matrix spikes and matrix spikeduplicates (MS/MSDs) were analysed byspiking selected patches with 10ul ofKCrO,at a concentration 10 times that of the patchconcentration to assess recovery. The

KCrO, patches analysed by modifiedNIOSH 7600/IC-VAS were within 89-101%of the theoretical Cr(VI) concentrations. Also,the MS/MSDs averaged 121% and 111-5%for the 44pug Cr(VI)/cm2 and 0-018,ugCr(VI)/cm2 patches with reported relative per-centage differences (RPDs) of 2% for both.Therefore, the patch concentrations werewithin Contract Laboratory Procedure (CLP)criteria for acceptable variability (± 25%).Additionally, the K2Cr2,7 patches analysedfor total Cr using EPA Method 6010 werewithin 88-97% of the theoretical total Crconcentrations. The MS/MSDs measured forthe 4-4,ug Cr(VI)/cm2 patch were both 106%.We analysed the data by SYSTAT version 5-0by the paired t test with a 95% confidencelevel and the data sets were not statisticallydifferent. Hence, the Cr(VI) concentrationsmeasured were considered to be stable in thehexavalent form and in the appropriateamounts.The concentrations of total Cr in the

Cr(III) patches ranged from 89%-105% ofthe theoretical concentrations. TheMS/MSDs measured for the 33 ugCr(III)/cm2 concentration were 90% and80% respectively. The 0 33,ug Cr(III)/cm2MS/MSD concentrations were 110% and113%, respectively, with reported RPDs of3% for both. The Cr(III) patches were alsoanalysed for Cr(VI) by the modified NIOSH7600/IC-VAS method and none was found.In summary, these analyses were reproducibleand within the CLP acceptable range. Blankpatches were also analysed for Cr(VI) andCr(III). No Cr was detected on either patchesusing reported limits of 0-00025 mg/cm2 forCr(III) and 0-000015 mg/cm2 for Cr(VI).Representative patches were reanalysed aboutevery three months, and, in every instance,the appropriate valency states and concentra-tions did not change (less than 5%) through-out the duration of the study (about oneyear).

STUDY DESIGNPatch concentrationsThe Cr(VI) patch concentrations used in thisstudy were chosen based on our best esti-mates of a range that would provide amaximal (100%) response at the highest con-centration and a minimal response (< 10%) atthe lowest concentration. This is similar to thedesign of other patch testing studies.28 Basedon the theoretical underpinnings of the dose-response relation50 our review of the literaturesuggested that an approximate 250-fold rangeof Cr(VI) concentrations from 0-018 to44 ,ug Cr(VI)/cm2 would yield useful results.Similarly, for Cr(III), a 50-fold range of con-centrations (0-66-33 ug Cr(III)/cm2) waschosen based on previously published data.5'

Patch testing strategyThe patch testing study was designed to occurin three rounds. In the first round, all subjectswere tested with a diagnostic Cr(VI) patch(444ug Cr(VI)/cm2) to confirm that all volun-teers in the study were allergic to Cr(VI).

373


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/


Those who were confirmed as sensitised inround one were then tested in round two withthe two lowest Cr(VI) concentrations and allfour Cr(III) concentrations. If the subjectsresponded to both of the Cr(VI) patches oronly to the higher concentration (0-088 pgCr(VI)/cm2), a threshold was considered tohave been identified and the subjects did notcomplete round three. If no responseoccurred at either of the two low Cr(VI) con-centrations, then the subjects were tested inround three with the two higher Cr(VI)-concentrations. For each subject, the lowestCr(VI) concentration at which a positiveresponse occurred was considered to be theMET for that subject.The rationale for testing with low Cr(VI)

concentrations initially, followed by higherconcentrations if necessary, was to minimisethe incidence of "false positives" and "excitedskin syndrome" that can occur when multiplepatches are applied to the subject's skin in asingle dosing.52 54 Indeed, many of the earlierreported positive responses to low Cr(VI)concentrations almost certainly were false-positives resulting from multiple testing.852-55Patch testing with Cr(III) was performed inone round of testing (round two) as theCr(III) patches were not expected to elicit aresponse in most of the subjects. All testpatches were supplied to the physicians undercode to double blind the study. The purposeof blinding was to achieve unbiased readingsof the test results by the study physicians.Only one physician evaluated each patient.The physicians received approval from theirrespective human use committees as appro-priate.

Population sizeBefore starting the study, an analysis was per-formed to estimate the total number of sub-jects required to achieve acceptable statisticalpower. Our goal was to identify a lowerthreshold at which no more than 10% of theCr(VI) sensitive subjects would respond with90% confidence. A figure of 10% was usedsince a number of dermatologists indicatedthat it would be very difficult to accuratelyidentify through testing a 1%-5% valuebecause 10% is probably very close to thethreshold dose for all people (including themost sensitised). Estimating the number ofsubjects needed for the study required anassumption about the number of respondersat a particular test concentration. Based onresults of previous reports56 we concluded thatabout 50-80 subjects would be required tomeet these criteria.

It was anticipated that a number of thesubjects may have actually had irritant,atopic, or "excited skin syndrome" reactions,

Table 2 Description of Cr(VI) sensitised volunteers who participated in the study

No No whooriginally patch qualified and Range of ages Average age

Sex Tested paricipated (y) (y (SD))

Men 78 39 24-74 45-6 (12 6)Women 24 15 25-59 39-6 (10-3)

rather than a true allergic contact dermatitisresponse during previous testing and thatabout 80% of the initial subjects would, onround one testing, show a strong positiveresponse to Cr(VI). Of these, it was estimatedthat at least 75% of sensitised persons wouldconsent to participate in the proposed studiesand complete the required testing rounds.The drop out rate for these subjects was notknown, but was not expected to exceed 10%-20%. In summary, to have a study populationof no less than 50 subjects, it was determinedthat we needed to find about 100 Cr(VI)sensitised volunteers.

PARTICIPATING PHYSICIANS AND STUDYPOPULATIONParticipating physiciansSix practising dermatologists conducted theclinical aspects of the study (Dr RobertAdams, Dr Joseph Fowler, Dr James Marks,Dr Charles Morton, Dr James Nethercott,and Dr James Taylor). They also participatedin the study design.

Study populationMore than 6000 patient files from variousdermatologists (who specialised in patch test-ing) were examined before 100 possible vol-unteers were identified. Eventually, a group of113 potential subjects were found by the par-ticipating physicians, of which 102 eventuallytook part in the study (11 subjects droppedout due to personal reasons). All werebelieved to be Cr(VI) sensitised based primar-ily on previous clinical patch tests performedby these physicians. As presented in table 2,this initial study population consisted of 78men (76%) and 24 women (24%). All wereover 18 years of age. Persons taking immuno-suppressive or steroidal medications and preg-nant women were excluded from the study.Subjects with eczema at the scheduled time oftesting were not tested until the dermatitissubsided and it was requested that topicalsteroids not be used for two weeks beforetesting. All volunteers provided their doctorswith written consent to participate in thestudy.

Patient questionnaireBefore initial testing, each patient filled out amedical and occupational history question-naire. Each questionnaire was screened by aqualified person to insure proper completion.The medical history discussed in the formincluded incidence, type, and duration of pastpresent dermatitis and other known allergiesincluding asthma and any other skin problemsor sensitivities. The questionnaire also askedfor history of jobs held and correspondingduration, as well as any known or potentialexposure to Cr in the workplace. Use of overthe counter medications and vitamins wasalso recorded. The purpose of collecting thisinformation was to assess whether previousexposure to Cr compounds was a significantfactor in individual METs.

Information on allergic and atopic dermati-tis was available for all of the 102 subjects.

374


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/


Table 3 Summary ofoccupations of the 54 Cr(IT>)sensitised volunteers who participated in the study

Occupation No ofpeople

Accounting 2Art dealer 1Brick mason* 3Carpenter 2Carpet layer 1Cement mason/finisher* 4Truck driver (concrete)* 1Core driller* 1Electrician 1Engineer 2Gardener 1Horse trainer 1Insulation installer 1Lab technician 2Mechanic 1Medical assistant 1Photoengraver* 3Plumber (retired) 1Print estimator 1Production manager 2Publisher 1Real estate associate 1Retired 5Saw cutter 1Secretary 5Self employed 1Student 1Teacher educator 3Tyre changer 1Transit operator 1Warehouse worker 1Unemployed 1

*Construction related jobs.

Present or past atopic dermatitis was presentin 15% of those who completed the study(eight people). This is in keeping with previ-ous data relating to the prevalence of allergiccontact dermatitis. Only 22% of the volun-teers worked in the construction industry or arelated field and a significant number of par-ticipants had no known previous occupationalexposure to Cr(VI) (table 3).

PATCH TESTING PROCEDUREIn each round of testing, all patches wereapplied to the upper sides of the back at 7 cmapart and fixed with Scanpor or paper tape fortotal occlusion with each patient serving as hisor her own control. The patches remained inplace for 48 hours, at which time they wereremoved and readings were taken then and at96 hours. Test sites with positive reactionsafter the 96 hour exposure duration were pho-tographed. For each patch in each round, thephysicians recorded one of the followingresponses:

1 = Weak (no vesicular) reaction: ery-thema, infiltration, papules (+)

2 = Strong (oedematous or vesicular) reac-tion (+ +)

3 = Extreme (spreading, bullous, ulcera-tive) reaction (+ +)

4 = Doubtful reaction, macular erythemaonly (?)

5 = Irritant morphology6 = Negative reaction (-)7 = Not testedFor the purposes of this study, any degree

of positive response (1-3), including a veryweak response, was considered to be positive.This is consistent with current NACDG crite-ria for patch test interpretation.35Round one testing took place between June

and September 1992. The first round con-

sisted of testing subjects with a TRUE-Testdiagnostic patch concentration of 44 ,igCr(VI)/cm2 and a control patch (containingonly hydroxypropyl cellulose) to verify sensiti-sation to Cr(VI). The patches were removedafter 48 hours and an initial reading of the testsite was recorded. The final reading, alongwith photographs of the test sites, was taken96 hours after patch application. Those whodeveloped an allergic response were consid-ered to be Cr(VI) sensitised and qualified forround two testing of the study. Three weeksafter round one was completed, those withdefinite positive reactions were tested inround two.Round two testing took place between July

and October 1992. The second test roundconsisted of patch testing volunteers with0 018 and 0-088 pug Cr(VI)/cm2 and all fourCr(III) concentrations (0-66, 3.3, 6&6, and 33jug Cr(III)/cm2). Control patches consisting ofblank hydroxypropyl cellulose and polyvidonewere also tested. The patches were removedafter 48 hours and an initial reading of the testsite was recorded. The final reading, alongwith photographs of the test sites, was taken96 hours after patch application.Round three testing took place between

August and December 1992. Those volun-teers showing responses to both 0-01 8 and0.088 /ug Cr(VI)/cm2 or to only the 0-088 jugCr(VI)/cm2 patch concentrations were nottested in round three. Those who did notrespond to either of the two lowest Cr(VI)concentrations were tested in the third roundwith the two highest Cr(VI) concentrations(0 18 and 0-88 pug-Cr(VI)/cm2). Any who hadanomalous results in round two (positive reac-tion to 0-018 Cr(VI)/cm2 but not to 0 088 jugCr(VI)/cm2) were retested in round three withall four Cr(VI) concentrations. The roundthree patches were removed after 48 hours ofapplication and an initial reading of the testsite was recorded. The final reading, alongwith photographs of the test sites, was taken96 hours after patch application. All volun-teers tested in round three were given a threeweek interval between round two and roundthree testing.

Patch test resultsRound oneOf the 102 people who were initially selectedfor the study, only 54 responded positively(+, + +, or + + +) to the diagnostic Cr(VI)patch (4.4 jug Cr(VI)/cm2). This response rate(47%) was far less than the projected estimateof 80%, but still yielded enough people (54)to provide statistically significant results. The54 subjects who proceeded to round two ofthe study consisted of 39 men (72%) and 15women (28%). The men ranged in age from25-74 (mean 47 9) years; the women from25-59 (mean 41-2) years.

Round twoIn round two, four of 54 subjects (7%)responded positively to 0-088 jug Cr(VI)/cm2but not to 0-018 jug Cr(VI)/cm2. Accordingly, a

375


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/

Nethercott, Paustenbach, Adams, Fowler, Marks, Morton, Taylor, Horowitz, Finleyy

Table 4 Cumulative dernal response of 54 Cr(VI)sensitised volunteers to variousconcentrations ofCr(VI)

Minimum elicitation thresholdCr(VI) (pug/cm2) response (%) Cumulative response (%)

0-018 1/54 (2) 1/54 (2)0-088 4/54 (7) 5/54 (9)0-18 5/54 (9) 10/54 (19)0-88 22/54 (41) 32/54 (59)4-4 22/54 (41) 54/54 (100)

Table S Cumulative response of54 Cr(VI)sensitisedvolunteers to various concentrations of Cr(III)

Cr(III) concentration (#g/2cm) Cumulative response (%/6)

0-66 0/54 (0)3-3 0/54 (0)6-6 0/54 (0)33 1/54 (0)*

*This person, on retest, did not respond to this concentration ofCr(III).

MET of 0-088 pg Cr(VI)/cm2 was recordedfor these four people and they were notrequired to complete round three. Only one of54 (2%) responded positively to both 0-018pg-Cr(VI)/cm2 and 0 088 pg Cr(VI)/cm2.Accordingly, a MET of 0 018 pg Cr(VI)/cm2was recorded for this person and he was notrequired to complete round three. Forty nineof the 54 subjects did not respond to eitherconcentration and, therefore, proceeded toround three. One person had a weak responseat the highest Cr(Ill) concentration tested (33pg Cr(III)/cm2). On retesting to confirm theweak response, this person failed to elicit apositive response to 33,ug Cr(lIj)/cm2. Theremaining 53 volunteers also failed to respondpositively to any of the Cr(Ill) concentrations.

Round threeIn round three, one of the 49 subjectsresponded to both 0-88 and 0-18 pgCr(VI)/cm2. Accordingly, the MET for thisperson was recorded as 0-18/pg Cr(VI)/cm2.Also, 22 subjects responded to 0-88 pg

Cumulative response of54Cr(V7) sensitisedvolunteers to test patchescontaining variousconcentrations of Cr(VI).

100 r

a)enc

0.C.U)

0co

-s

L-

75

50~

25

o0.01

U

* Response results* Extrapolation

0

i

II

..-

I:0.1 1

Concentration of Cr (VI) intest patches (pLg/cm2)

Cr(VI)/cm2, but not 0 18pg Cr(II)/cm2, andtherefore 0-88 jpg Cr(VI)/cm2 was recorded asthe MET for those subjects. Twenty two ofthe volunteers tested in round three had noresponse to either concentration. Becausethese volunteers did not respond to any patchconcentration less than the diagnostic patch,the MET recorded for them was 44 pigCr(VI)/cm2. Tables 4 and 5 present the METand cumulative frequency of response forCr(VI) and Cr(III). The figure presents thedose-response curve.

STATISTICAL ANALYSIS OF RESULTSThe data were analysed by a computer mod-elled data technique.58 A truncated log normaldistribution was fitted to the data with maxi-mum likelihood methods, which is a tech-nique for choosing parameters of a selecteddistribution of the observed data such that theprobability of response of the observed data ismaximised.59 The fit to the data was excellentas confirmed by thex2 goodness of fit test,which gave a p value of >0 05. The 10%cumulative response MET for Cr(VI) was0-089 pg/cm2. The log normal distribution isconventionally used in the analysis of bioassaydata and was used here.60 Because a reactionat the maximum tested concentration was acriterion for inclusion in the study, the distrib-ution of response concentrations is truncated.Thus the truncated log normal distributionwas fitted to the dataset with the highest trun-cation point being the highest Cr(VI) concen-tration tested (44 pig Cr(VI)/cm2).

IMPORTANCE OF PATCH CONCENTRATIONRecently, some dermatologists have con-cluded that the mass of allergen per unit areaof skin is the correct measure of dermal dose,not the applied patch concentration ofallergen.29 To assess the influence of the con-centration in the patch test material v the doseapplied in mass per unit area on the potentialto elicit a response, two different sets ofpatches were prepared. The first set wasprepared with the gel material normally usedto manufacture the 0-88 ug Cr(VI)/cm2 patch.A second patch was constructed from thesame gel but it was only 1/7 as thick (the thinpatch). Hence, the thin patches contained0-13 pg Cr(VI)/cm2, but were identical to the0 88 pg Cr(VI)/cm2 patch with respect tomass of Cr(VI) per mass of patch (in this case,175 ppm).These patches were then evaluated with

nine volunteers previously shown to have aMET of 0-88 pg Cr(VI)/cm2. The results(table 6) indicate that six of the nine subjectsshowed a response deemed to be morphologi-cally positive at 0-88 pg Cr(VI)/cm2 but noneshowed such a response at 0-13 pg Cr(VI)/cm2.Thus although both patches were identical ona mass of Cr(VI)/mass of patch basis (175"ppm" Cr(VI)) the elicited response was dif-ferent indicating that mass per unit arrea is amore appropriate measure of dose. Hence, aspredicted by Upadhye and Maibach,29 themass of allergen per unit area of skin, and notthe mass of Cr(VI)/mass of patch (concentra-

376


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/

A study of chromium induced allergic contact dermatitis wih 54 volunteers: implications for environmental risk assessment

Table 6 Response ofvolunteers to patches containingidentical PPM concentrations of Cr(VI) (mass ofCr(VI)/mass ofpatch) but differing mass loading ofCr(VI)1cm2

Patch concentration

Subject 0 13 pig Cr(VI)/cm2 0-88 pig Cr(VI)/cm2

1 Negative Negative2 Doubtful Positive3 Negative Positive4 Doubtful Positive5 Negative Positive6 Negative Positive7 Negative Positive8 Negative Negative9 Negative Negative

tion in ppm) is the critical factor to quantita-tively characterise the dermal hazard.

IMPORTANCE OF PATCH SURFACE AREAIn the past, it has been suggested that if thesurface area of the standard patch was signifi-cantly increased, then a sub-MET dose per

unit area might produce a response becausethe total systemic uptake of the chemicalwould be greater. To assess whether patchesof a larger surface area might elicit a responseat concentrations lower than the METs identi-fied in this study, another study was con-

ducted. Four volunteers who had previouslybeen shown to have a MET of 0-88 ,ugCr(VI)/cm2 were patch tested with five 0-18,ug Cr(VI)/cm2 patches placed side by side.Hence, the total amount of allergen requiredto elicit allergic contact dermatitis in a stan-dard patch was applied, but over a larger sur-

face area. The results showed that sub-METconcentrations of Cr(VI) applied over a largerskin surface area did not elicit the positiveresponses seen when the MET concentrationwas applied in the standard patch. Theseresults support the belief that to produceallergic contact dermatitis, the concentrationof allergen delivered to the target organ (theLangerhans cell) in a given area of skin is thecorrect dosimetric index for evaluating thedermal allergic contact dermatitis hazard.

DiscussionThe results of this study corroborate earlierreports that Cr(VI) sensitive persons respondto serial dilutions of Cr(VI) in a fairly linearmanner at moderate concentrations.'1'8 20 21Interestingly, almost half (22/54) of theCr(VI) sensitised volunteers in this study didnot respond to Cr(VI) concentrations lessthan the diagnostic concentration of 44 jugCr(VI)/cm2. The diagnostic patch contains a

fairly high Cr(VI) concentration that isunlikely to be found in the environment andmost workplaces.8Many of the subjects who failed to respond

in round one had been thought by the physi-cians to be Cr(VI) sensitive due to earlierinstances of positive responses to 0-5%K2Cr207 in petroleum jelly or to 0-25%K2Cr2O, in TRUE-Test matrix. As discussedearlier, 0-5% K2Cr20 is no longer commonlyused clinically in the United States due to thehigh rate of irritant (non-allergic) responses,1 43

although it is still used in Europe. Manydermatologists in North America believe that

use of 0-25% K2CrO, to identify sensitisa-tion, as was done in this study, minimises oreliminates the incidence of irritant reactions.36Therefore, it is likely that many of those vol-unteers who initially reacted to 0 5% K2Cr2O,exhibited an irritant (false positive) reaction inearlier testing and that the 0-25% K2Cr2O,patch used in this study failed to produce theirritant response. Alternatively, it is possiblethat some of the subjects who failed torespond in round one truly are Cr(VI) sensi-tive but have a threshold that is greater than0-25% KCrO,. If this is true, then the resultsof our study could be considered "worst case"as only the hypersensitive volunteers wereused to represent the sensitised population.The results of this study suggest that

the 10% MET for Cr(VI) induced allergiccontact dermatitis is approximately 0-089 pug-Cr(VI)/cm2. No significant correlations be-tween age, sex, or occupation were found atany of the tested concentrations.

Several estimates of the prevalence ofCr(VI) related allergic contact dermatitis inthe general population have been published inthe past three years. Hostynek and Maibachestimated that less than 1% of the generalpopulation is Cr(VI) sensitive based on theirreview of several years of clinical data.7Paustenbach et a18 estimated that about 07%of the general population is Cr(VI) sensitive.Finally, in a review of clinical data collectedfrom several hundred human volunteers,Nethercott57 estimated that "a 06% preva-lence is likely a reasonable reflection of theprevalence of Cr(VI) sensitivity in the generalpopulation, although false positive irritantresponses to 05% K2Cr,07 in petroleum jellymay explain even some of these responses."In all three analyses, if one excludes thosewho have had more than 10 years of experi-ence working with wet cement, the prevalenceof Cr(VI) sensitisation in the general popula-tion is probably no greater than 0 1%.Therefore, if Cr(VI) sensitised subjects com-prise less than 0-1% of the general population,an allergic contact dermatitis based soil cleanup criterion that is protective of 10% of theCr(VI) sensitised population will probably beprotective of about 99 99% of the generalpopulation.When attempting to use these data to iden-

tify safe concentrations of Cr(VI) in environ-mental media (soil, dust, etc), it is necessaryto consider whether children are more or lessprone to Cr(VI) sensitisation than adults. Ingeneral, most authorities believe that sensitivityto contact allergens increases with age, dueprimarily to increased exposure.626' Childrenunder the age of three react less often and lessintensely than older children, whereas chil-dren older than three react to contact aller-gens in a manner similar to adults.6465 Instudies of patients in paediatric dermatologyclinics, Schachner et a166 diagnosed contactdermatitis in only 1 2% of children testedwith KCr,07. Romaguera et a167 reportedpositive reactions in 11 of 1023 childrentested (1.1%) and Tunnessen68 reported asimilar prevalence rate of 2-4%. In a study of

377


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/

Nethercotr, Paustenbach, Adams, Fowler, Marks, Morton, Taylor, Horowitz, Finley

Soil concentration (mg-allergen/kg-soil) =

422 paediatric patients, Angelini andMeneghini69 concluded that: (a) sensitisationcan occur in children as well as adults, but isextremely rare in the first few years of life; (b)the incidence of allergic contact dermatitisincreases with age and escalates significantlyfrom the age of 12 years, primarily because ofincreased exposure to environmental sensitis-ers and, perhaps, a fully matured immune sys-tem; and (c) the patch testing techniqueemployed in adults, with the same substancesand concentrations, is also well suited fordiagnostic purposes in children. Thus theMETs identified in our study can be consid-ered applicable to children as well as adults.

It is noteworthy that there is some uncer-tainty associated with reading a patch testresponse as an allergic v an irritant reaction.Nethercott57 has stated that the incidence of"misread" test sites can be as high as 20%.Because the patch test reactions in this studywere interpreted by members of the NACDGthis rate should be much less as the interpre-tation of patch tests is a routine part of theirpractice. Because it is more common to mis-read irritant reactions as allergic than discountan allergic reaction as an irritant one,43 theMET estimated from this study can be con-

sidered conservative (no less than predictedhere).

CALCULATING ACCEPTABLE SOILCONCENTRATIONS FOR Cr(VI)As described by Horowitz and Finley,3" an

acceptable concentration of allergen in soilcan be derived as above: Where MET = theminimum elicitation threshold determinedfrom patch test data; CF = conversion factor;SA = soil adherence factor; BVA = bioavail-ability.

If the leachability (bioavailability) of theallergen from soil to skin is less than 100%,then the soil concentration must be adjustedupwards accordingly.The 10% MET for Cr(VI) in sensitised

people, as identified in the patch testingstudy, was 0-089 pg Cr(VI)/cm2 skin. Asdiscussed previously, a threshold derived froma 48 hour, occluded patch test study is highlyconservative (health protective) with respectto realistic environmental exposures. Accord-ingly, to remain consistent with USEPA's"reasonable maximal exposure" philosophy ofrisk assessment (wherein average and upperbound values are combined to give a "reason-able" estimate, rather than simply compound-ing worst case assumptions), it is appropriateto use an average value to represent soiladherence. The USEPA's suggested averagevalue of 0-20 mg/soil cm2-skin was used forthis analysis.With these values for MET and soil

fmg-allergen\ (106 mg-soil\MET t cm _skin ) x CF kg-soil )

lmg-soil \SA _cm-skin_ x BVA (unitless)

adherence factor, it can be estimated that aconcentration of 445 mg Cr(VI)/kg soil (445ppm) should be acceptable. Hence, even if itwere conservatively assumed that all Cr(VI) inchromite ore processing residue adhering toskin was able to leach into skin moisture andcross the skin barrier, a chromite ore process-ing residue concentration of 445 ppm Cr(VI)should protect the vast majority of the Cr(VI)-sensitised population from the allergic contactdermatitis hazard. As noted by Horowitz andFinley70 however, extraction of chromite oreprocessing residue containing up to 1240 mgCr(VI)/kg soil failed to leach more than 0 1%of the Cr(VI). Hence, chromite processingresidue concentrations much higher than445 ppm Cr(VI) should also be health pro-tective for even those people who are veryallergic to chromium. As soil cleanup concen-trations of about 400 ppm Cr(VI) or greaterare considered to be health protective for themore traditional exposure pathways (soilingestion and particulate inhalation), it seemsthat elicitation of allergic contact dermatitison contact with chromite ore processingresidue could occur only at Cr(VI) concentra-tions much higher than the soil concentra-tions that may pose other hazards.

CALCULATING ACCEPTABLE SOILCONCENTRATIONS FOR cr(m)For the purpose of calculating a Cr(III)concentration in soil, if one conservativelyassumes a 10% MET at the maximum testeddose used in this study (33 jug Cr(III)/cm2-skin), the concentration that should not posean allergic contract dermatitis hazard is165 000 mg Cr(IIII)/kg-soil (assuming a soiladherence factor of 0-20 mg-soil/cm2-skin and100% leachability. Adjustment for less than100% leachability would yield a chromite oreprocessing residue concentration of high hun-dreds of thousands of ppm. In short, it isimplausible that soil contaminated withCr(III) could pose a dermal hazard.

ConclusionsThis patch testing study was designed todetermine a MET for Cr(VI) induced allergiccontact dermatitis on a mg Cr(VI)/cm2 skinbasis so that the results could be used toperform occupational and environmental riskassessment. The results suggest that the 10%MET for Cr(VI) induced allergic contact der-matitis is about 0-089 pg/cm2. It was alsoshown that the concentration (in ppm) in thepatch was not as relevant a dosimetric indexas dose per area. With quantitative risk assess-ment principles, these patch test data wereused to identify safe concentrations of Cr(VI)in soil.The regulatory concern about Cr(VI) in

378


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/


soil was born of the belief that because Cr(VI)sensitisation can occur in the workplace, envi-ronmental exposure to Cr(VI) in a soil matrixmight also elicit allergic contact dermatitis.Initial estimates of the likely safe concentra-tions of Cr(VI) in soil were based on patchtesting data from the 1950s and 1960s51; how-ever, as stated previously, there is a highdegree of uncertainty associated with thosedata. Furthermore, the results of those studieswere not reported in terms of mg Cr(VI)/cm2skin.The results of this patch test study were

used to develop soil clean up levels whichshould prevent the elicitation of allergic con-tact dermatitis even in sensitised persons.'1The results also indicate that concentrationsof Cr(VI) in soil due to chromite ore process-ing residue should not produce allergic con-tact dermatitis, even in Cr(VI) sensitisedpersons, below about 450 ppm. Our resultsindicate that traditional patch testing, which isused routinely to diagnose illness, can bemodified so that the data can then be used toresolve issues in environmental medicine andquantitative risk assessment.

1 Marks JG, DeLeo VA. Contact and occupational dermatol-ogy. St Louis: Mosbey Year Book, 1992.

2 Burrows D, ed. Adverse chromate reactions on the skin.In: Chromium metabolism and toxicity. Boca Raton,Florida: CRC Press, 1983:138-58.

3 Burrows D, Adams RM. Metals. In: Adams RM, ed.Occupational skin disease, 2nd ed. Philadelphia: WBSaunders, 1990:349-86.

4 Levin HM, Brunner NJ, Ratner H. Lithographer's der-matitis JAMA 1959;169:566-9.

5 Winston JR, Walsh EN. Chromate dermatitis in railroademployees working with diesel locomotives. JAMA1951;147:1133-4.

6 Morris GE. "Chrome" dermatitis: A study of the chem-istry of shoe leather with particular reference to basicchromic sulfate. Arch Dermatol 1958;78:612-8.

7 Hostynek JJ, Maibach HI. Chromium in US householdbleach. Contact Dermatitis 1988;18:206-9.

8 Paustenbach DJ, Sheehan PJ, Paull JM, Wisser LM,Finley BL. Review of the allergic contact dermatitis haz-ard posed by chromium-contaminated soil: Identifying a"safe" concentration. J Toxicol Environ Health 1992;37:177-207.

9 Adams RM, ed. Allergic contact dermatitis. In:Occupational skin disease. 2nd ed. Philadelphia: WBSaunders, 1990:26-31.

10 Ackerman AB. Histologic diagnosis of inflammatory skin dis-ease. Philadelphia: Lea and Febiger, 1978:223-6.

11 Rook A, Wilkinson DS, Ebling FJ, Champion RH, BurtonJL, eds. Textbook of dermatology. Oxford: BlackwellScientific, 1986:350-450.

12 Jackson EM, Goldmen R, eds. Irritant contact dermatitis.New York: Marcel Dekker Inc, 1990.

13 Iammintausta KH, Maibach HI. Irritant dermatitissyndrome. Immunological Clinics of North America1990;9:435-46.

14 Parkhurst HJ. Dermatosis industrials in a blueprint workerdue to chromium compounds. Arch Dermatol 1925;15:253-6.

15 Pirili V. On the role of chromium and other trace elementsin cement eczema. Acta Derm Venereol 1954;34: 137-43.

16 Zelger J. Zur Khlnik und pathogenses dis chromateekzems. Archives of Clinical and Experimental Dermatology1964;218:499-542.

17 Zelger J, Wachter H. Uber die beziehungen zwischenchromat-und dichromat-allergie. Dermatologica 1966;132:45-50.

18 Burrows D, Calnan CD. Cement dermatitis II. Clinicalaspects. Transactions of St John Hospital DermatologicalSociety 1965;51:27-39.

19 Skog E, Wahlberg JE. Patch testing with potassium dichro-mate in different vehicles. Arch Dermatol 1969;9:697-700.

20 Anderson FE. Cement and oil dermatitis: the part playedby chrome sensitivity. BrJtDermatol 1960;72:108-17.

21 Geiser JD, Jeanneret JP, Delacretaz J. Eczema au ciment etsensibilization au cobalt. Dermatologica 1960;121:1-7.

22 North American Contact Dermatitis Group (NACDG).Preliminary studies of the TRUE-TestTm patch test sys-tem in the United States. J Am Acad Dermatol1984;21:841-3.

23 ICDRG European Standard Series. Contact Dermatitis.1994;11:63-4.

24 Fischer TI, Maibach HI. The thin layer rapid use epicuta-neous test (TRUE-Test.), a new patch test methodwith high accuracy. BrJ Dermatol 1985;112:63-8.

25 New Jersey Department of Environmental Protection andEnergy. Risk assessment of the allergic contact dermatitispotential of hexavalent chromium in contaminated soil-derivation of an acceptable soil concentration. Draft by DrAlan Stem, Division of Science and Research, Trenton,New Jersey NJDEPE, 9 June 1992.

26 Stem AH, Bagdon RE, Hazen RE, Marzulli FN. Riskassessment of the allergic dermatitis potential of environ-mental exposure to hexavalent chromium. Jf ToxicolEnviron Health 1993;40:613-41.

27 Paustenbach DJ, Jernigan JD, Bass R, Kalmes R, Scott P. Aproposed approach to regulatory contaminated soil:Identify safe concentrations for seven of the most fre-quently encountered exposure scenarios. Reg ToxicolPharmacol 1992;16:21-56.

28 Andersen KE, Linden C, Hansen J, Volund A. The degreeof nickel allergy and the enhancing effect of multiplereactions evaluated by dose-response testing in allergicpatients. First Congress of the European Society of ContactDermatitis, Amsterdam, October 1992.

29 Upadhye MR, Maibach HI. Influence of area of applica-tion of allergen on sensitization in contact dermatitis.Contact Dermatitis 1992;27:281-6.

30 Fischer T, Maibach HI. Amount of nickel applied with astandard patch test. Contact Dermatitis 1984;1l:285-7.

31 Horowitz SB, Finley BL. Setting health protective soilconcentrations for dermal contact allergens: a proposedmethodology. Regul Toxicol Pharm 1994;19:31-47.

32 Leung HW, Paustenbach DJ. Techniques for estimatingdermal absorption of chemicals due to occupational andenvironmental exposure. Applied Occupational andEnvironmental Hygiene 1994;9: 187-97.

33 Gammelgaard B, Fullerton A, Avnstorp C, Menne T.Permeations of chromium salts through human skin invitro. Contact Dermatitis 1992;27:302-10.

34 Agency for Toxic Substances and Disease Registry.Toxicological profile for chromium. ATSDR/TP-88/10.Washington DC: ATSDR; US Public Health Service,1989.

35 Rietschel RL, Marks JG, Adams RM, et al. Preliminarystudies of the TRUE patch test system in the UnitedStates. JAm Acad Derm 1989;21:841-3.

36 Storrs FJ, Rosenthal LE, Adams RM, Clendenning W,Emmett EA, Fisher AA, et al. Prevalence and relevanceof allergic reactions in patients patch tested in NorthAmerica-1984 to 1985. J Am Acad Dermatol 1989;20:1038-44.

37 Mali JWH, Van Kooten WJ, van Neer FCJ. Some aspects ofthe behavior of chromium compounds in the skin.J Invest Dermatol 1963;41:111-2.

38 Samitz MH, Katz SA. A study of the chemical reactionbetween chromium and skin. J Invest Dermatol 1964;43:35-43.

39 Polak L. Immunology of chromium. In: Burrows D, ed.Chromium: metabolism and toxicity. Boca Raton, Florida:CRC Press, 1983:51-135.

40 Spruit D, van Neer FCJ. Penetration rate of Cr(III) andCr(VI). Dermatologica 1966;132:179-82.

41 Kligman AM. The identification of contact allergens byhuman assay. III. The maximization test: a procedurefor screening and rating contact sensitizers. Jf InvestDermatol 1966;47:393-409.

42 Fregert S, Rorsman H. Allergy to trivalent chromium.American Medical Association Archives of Dermatology1964;90:4-6.

43 Fischer T, Maibach HI. Easier patch testing with TRUETest.". JAmerAcad Dermatol 1989;20:447-53.

44 Wright RW. Evaluation of contact dermatitis using theTRUE patch test. Journal of the Arkansas Medical Society1991;88:271-2.

45 Molina D, Abell MT. NIOSH manual ofanalytical methods,3rd ed. Method No. 7600. NIOSH publ No 84-100.Cincinnati, Ohio: NIOSH, 1984.

46 California Air Resources Board. Ion chromatograph separa-tion-colorimetry. Modified CARB SOP No MLD 105.Sacramento, California: CARB, 1990.

47 US Environmental Protection Agency. EPA Method 3060.SW-846. 3rd ed. September. Washington, DC: USEPA,1986.

48 US Environmental Protection Agency. EPA Method 3050.SW-846. 3rd ed. September. Washington, DC: USEPA,1986.

49 US Environrmental Protection Agency. EPA Method 6010.SW-846. 3rd ed. September. Washington, DC: USEPA,1986.

50 Klaassen CD, Amdur MO, Doull J, eds. Cassarett andDoull's toxicology: the basic science ofpoisons. 4th ed. NewYork: Macmillan, 1992.

51 Bagdon RE, Hazen RE. Skin permeation and cutaneoushypersensitivity as a basis for making risk assessments ofchromium as a soil contaminant. Environ Health Perspect1991;92:111-9.

52 Maibach HI. The E.S.S excited skin syndrome (alias the"angry back"). In: Ring D, Burg B, eds. New trends inallergy. Berlin: Springer-Verlag, 1981:208-21.

53 Maibach HI, ed. Occupational and industrial dermatology,2nd ed. Chicago: Year Book Medical, 1987: 28-31;190-210.

54 Bruynzeel DP, van Ketel WG, von Blomberg M, ScheperRJ. Angry back or the excited skin syndrome.JAm AcadDermatol 1983;8:392-7.

55 Nethercott JR. Results of routine patch testing of 200

379


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/


patients in Toronto, Canada. Contact Dermatitis 1982;8:389-95.

56 Bagdon RE. Dermal absorption of selected chemicalsunder experimental and human exposure conditions tofacilitate risk assessment and the development of standardsfor soil. Report to the Office of Science and Research,Trenton, New Jersey: New Jersey Department ofEnvironmental Protection & Energy (NJDEPE), 1989.

57 Nethercott JR. Practical problems in the use ofpatch testingin evaluation of patients with contact dermatitis. In:Weston W, ed. Current problems in dermatology, St Louis:Mosby-year Book Inc, 1990:101-23.

58 GENSTAT Reference Manual. GENSTAT 5, release 2.2.Oxford: Clarendon Press, 1982.

59 Rothman KJ. Modern epidemiology. Boston: Little, Brown,and Co 1987.

60 Armitage P, Berry G. Statistical methods in medical research.2nd ed. Oxford: Blackwell Scientific Publications.1987:485-6:495-6.

61 Paustenbach DJ, Sheehan PJ, Lau V, Meyer D. An assess-ment and quantitative uncertainty analysis of the healthrisks to workers exposed to chromium contaminatedsoils. Toxicol Ind Health 1991;7:159-96.

62 Weston WL, Weston JA. Allergic contact dermatitis inchildren. AmJIDis Child 1984;138:932-6.

63 Rudzki E, Rebandel P, Grzywa Z. Patch tests with occupa-tional contactants in nurses, doctors and dentists.Contact Dermatitis 1989;20:247-50.

64 Heskel NS. Contact dermatitis in children. Dermatol Clin1984;2:579-84.

65 Hjorth N. Contact dermatitis in children. Acta DermVenereol (Stockh) 1981;S95:36-9.

66 Schachner I, Ling NS, Press S. A statistical analysis of apediatric dermatology clinic. Pediatr Dermatol 1983;1:157-64.

67 Romaguera C, Alomar A, Camarasa JMG. Contact der-matitis in children. Contact Dermatitis 1985;12:283-4.

68 Tunnessen WW. A survey of skin disorders seen in pedi-atric general and dermatology clinics. Pediatr Dermatol1984;1:219-22.

69 Angelini G, Meneghini CL. Contact dermatitis inchildren. Pediatric Dermatology News 1987;6: 195-204.

70 Finley BL, Horowitz SB. Using human sweat to extractchromium from chromite-ore processing residue: appli-cations to setting health-based cleanup standards.Y Toxicol Environ Health 1993;40:585-99.

Vancouver styleAll manuscripts submitted to Occup EnvironMed should conform to the uniformrequirements for manuscripts submitted tobiomedical journals (known as theVancouver style.)

Occup Environ Med, together with manyother international biomedical journals, hasagreed to accept articles prepared in accor-dance with the Vancouver style. The style(described in full in the BMJ, 24 February1979, p 532) is intended to standardizerequirements for authors.

References should be numbered consec-utively in the order in which they are firstmentioned in the text by Arabic numeralsabove the line on each occasion the refer-ence is cited (Manson' confirmed otherreports . . .). In future references topapers submitted to Occup Environ Med

should include: the names of all authors ifthere are seven or less or, if there are more,the first six followed by et al; the title ofjournal articles or book chapters; the titlesof journals abbreviated according to thestyle of Index Medicus; and the first and finalpage numbers of the article or chapter.Titles not in Index Medicus should be givenin full.

Examples of common forms of refer-ences are:

1 International Steering Committee of Medical Editors,Uniform requirements for manuscripts submitted tobiomedical journals. BrMedJ 1979;1:532-5.

2 Soter NA, Wasserman SI, Austen KF. Cold urticaria:release into the circulation of histamine and eosino-phil chemotactic factor of anaphylaxis during coldchallenge. N EnglJ Med 1976;294:687-90.

3 Weinstein L, Swartz MN. Pathogenic properties ofinvading micro-organisms. In: Sodeman WA Jr,Sodeman WA, eds. Pathologic physiology, mechanismsof disease. Philadelphia: W B Saunders, 1974:457-72.

380


http://oem.bm

j.com/

Occup E


.51.6.371 on 1 June 1994. Dow

nloaded from

http://oem.bmj.com/

assessment - a bmj journal

Documents