Serological and histological diagnosis of primary biliary cirrhosis

76506_7527_full.pdf

J.clin.Path.(1966),19,527

Serologicalandhistologicaldiagnosisofprimarybiliarycirrhosis

R.B.GOUDIE1,R.N.M.MACSWEEN,ANDD.M.GOLDBERG

FromtheWesternInfirmary,Glasgow

SYNOPSISAsimpleimmunofluorescencetestforantibodytoamitochondrialantigenpresentin

manytissuesisareliablemethodofdistinguishingmostcasesofprimarybiliarycirrhosisfromjaundiceduetoextrahepaticbiliarytractobstruction.Of30casesdiagnosedasprimarybiliarycirrhosis,26hadantimitochondrialantibodywhereasnoneof77caseswithjaundiceduetoextra-hepaticbileductobstructionshowedthisserologicalabnormality.Theantibodywasalsofoundintheserumofthreeof42patientswhohadotherformsofcirrhosisandintwoof266patientswith

noevidenceofliverdisease.Clinical,biochemical,andserologicalfindingsfavourtheviewthatprimarybiliarycirrhosisisarealentitywhich,inourpresentstateofknowledge,cannotbedefinedclearlybyanysinglemethodofinvestigation.Inparticular,thelivermayshowavarietyofhistologicalappearanceswhich,interpretedwithoutregardtotheotherfeaturesofthecase,mayleadtoerrorsindiagnosis.

Primarybiliarycirrhosisisarareconditionofun-knowncausecharacterizedclinicallybypruritus,hepatomegaly,andjaundiceofobstructivetypeintheabsenceofextrahepaticbiliarytractobstructionandwithoutpainorfever.Pathologicallythebestknownfeaturesareafine'monolobular'cirrhosisoftheliver,pericholangitisaffectingthemedium-gizedbileductsintheportaltracts,anddisappearanceofmanyofthesmallportalbileducts;intheearlystagestruecirrhosismaynothavedeveloped,andinthelatestagestheappearancesofthelivermaybeindi-stinguishablefromthoseofportalcirrhosis(Sherlock,1963).In1965,Walker,Doniach,Roitt,andSherlockreportedthatantimitochondrialantibodywasdemonstrablebymeansofanindirectimmuno-fluorescencetechniqueintheserumofallof32patientswithprimarybiliarycirrhosis.Theantibodywasnotfoundinpatientswithextrahepaticbiliarytractobstruction.Thesefindingssupporttheconceptthatprimarybiliarycirrhosisasdescribedaboveisanosologicalentity,andtheymayshedsomelightonthecausationofthisobscurecondition;theyalsoprovidethebasisforasimpletesttodifferentiateprimarybiliarycirrhosisfromotherformsofobstructivejaundice,andsomaypreventundesirablesurgicalexplorationoftheextrahepaticbiliary

ISupportedbyagrantfromtheScottishHomeandHealthDepartment

Receivedforpublication13July1966passages.WehaverepeatedthestudiesofWalkeretal.(1965)andinthispaperreportpartialconfirma-tionoftheirfindings.Inadditionwedrawattentiontothefrequencyofmisleadingreportsonliverbiopsymaterialinprimarybiliarycirrhosisandemphasizesomeofthedifficultiesinmakinganaccuratehistologicaldiagnosis.

CASESSTUDIED

Fluorescentantibodytestswerecarriedoutontheserumof204patientswithjaundiceduetoavarietyofcauses,onfivepatientswithprimarybiliarycirrhosiswhodidnothavejaundice,andon266patientswithoutjaundiceorotherevidenceofliverdisease.

PRIMARYBILIARYCIRRHOSIS(30CASES)Inviewofthedifficultyindefiningrigiddiagnosticcriteria,themainclinicalandbiochemicalfeaturesofthe30casesaregiveninTablesIandII.In24casesthediagnosisofprimarybiliarycirrhosishadbeenmadebeforeserologicaltesting.Thediagnosisofthreepatientsthoughttohaveportalcirrhosis(cases2,24,and25)andofthreepatientsthoughttohavelupoidhepatitis(cases1,9,and26)waslaterchangedafterserologicaltestingandreviewoftheclinicalandbiochemicalfindings.

527copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

R.B.Goudie,R.N.M.Macsween,andD.M.Goldberg

TABLEI

CLINICALDETAILSOFPATIENTSWITHPRIMARYBILIARYCIRRHOSISORMITOCHONDRIALANTIBODY

AgeSexDurationPruritusXanthomaTitreofAntibodyTitreofAntibodyComplement-fixationImmunofluorescenceTestEvidenceagainstExtrahepaticBiliaryObstruction

PrimaryBiliaryCirrhosis154F267F354F

4S6789101112131415161718192067

64433745605143555967494272385644F

FFFFFMFFFFFFFFFF5yr.3yr.10yr.

3yr.1yr.4yr.3yr.1yr.6yr.10yr.6mth.6mth.3yr.9mth.7yr.2yr.3yr.1yr.1yr.14yr.

2150F2yr.2241F9yr.

23

242526272850

4651545165F

FFFFF11yr.5yr.3yr.5yr.2yr.2yr.

2958F6yr.3056F2yr.

PortalandPostnecroticCirrhosis3148F3yr.3256F8yr.3339F15yr.

Controls3467M-3548F-

PORTALANDPOSTNECROTICCIRRHOSIS(42CASES)Histologicalevidenceofportalorpostnecroticcirrhosiswasobtainedin29cases;theremaining13subjectshad(1)aclinicalhistoryofchronicliverdiseasewithrecurrentjaundiceofatleastoneyear'sduration;(2)portalhypertension,ascites,orepisodesofhepaticcoma;(3)abnormalserumturbidityandflocculationtests;(4)serumalkalinephosphatasebelow35King-Armstrong(K.A.)units/100ml.;(5)serumtransaminaseactivitiesconsistentwithcirrhosisasdescribedbyWr6blewski(1958).

EXTRAHEPATICBILLARYTRACTOBSTRUCTION(72

CASES)ANDSECONDARYBILIARYCnuRosis(5CASES)Fifty-eightpatientshadoperativeornecropsyevidenceofextrahepaticbiliarytractobstructiondueeithertostoneortumour.Infivetherewashisto-logicalevidenceofsecondarybiliarycirrhosis.Thebiliarytractwasnotexploredin19whohadclinicalandbiochemicalevidenceofobstructivejaundice;ofthese,13wereknowntohavecarcinoma,andtheremainingsixhadgallstonesoranon-functioninggallbladderdemonstratedradiologically.

INFECTIVEHEPATmS(25CASES)Allhadtypicalfeaturesofthedisease,includingaprodromalillnessandanacuteattackofjaundicelastingnotmorethansixweeks.In20theserumglutamicpyruvatetrans-

aminase(S.G.P.T.)wasatleast1,000units/ml.;intheotherfive,whowereallunder23yearsold,thetransaminaselevelsestimatedlateintheillnesswere

notdiagnostic. DRUG-INDUCEDJAUNDICE(10CASES)Thisconsistedoftransientjaundicefollowingadministrationof528 Case No. + + ++ ±+ +

±++

+Xanthela

Xanthelasma

+Xanthelasma+ + +

XantheLasma2,048

2,048256

256256128512256

1282,0481,024>4,0961,0001,0001,0002565125121,000

2,048500

500
4,096 <8 16 <8<82,048

2,0482,048

1281,0241285121281,024128>8,1921285122,048128>8,192512512512512

1,0242,048

128

1281,024256Neg.Neg.

Neg. Neg. 128

2,048512NeedlebiopsyNecropsyLaparotomy,cholangiogramNecropsyNeedlebiopsyLaparotomyLaparotomyNeedlebiopsyNeedlebiopsyLaparotomyNeedlebiopsy

NeedlebiopsyLaparotomyNeedlebiopsyNeedlebiopsyLaparotomyCholangiogramLaparotomyLaparotomy,cholangiogramCholangiogramLaparotomy,cholangiogram

NecropsyNeedlebiopsyNeedlebiopsyLaparotomyCholangiogramLaparotomy,cholangiogramLaparotomy

Necropsy

Laparotomy+

64

2,048>4,096

500

12832512copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

529

TABLEII

BIOCHEMICALFINDINGSINPATIENTSWITHPRIMARYBILIARYCIRRHOSISORMITOCHONDRIALANTIBODY1

CaseNo.SerumBilirubinSerumAlkalineThymolTurbiditySGOTSGPTSerum(mg./100ml.)Phosphatase(units)(units/ml.)(units/ml.)Cholesterol(KA.units/100ml.)(mg./100ml.)

PrimaryBiliaryCirrhosis13-825-438-1413-453-067-6720-082-891-6105-71150-02122-5131-7140-9155-2160-71715-6180-71924-0201-1212-82226-02322-8246-2250-8263-12711-22813-62916-0301-6

PortalandPostnecroticCirrhosis3132-8322-2332-3

Controls3435040-5120

62629669922208085701001154386707680971541085020890902746100140136105

20 3035
16 1218

111591248594141710106244611912251012131423326

13 S8 4125

1178670381402603026196180138984960361041152008072982151031321961681044146

496
28124
14 1578

6741404660220602511220265501458427418625015023418015570568224822025022

402
1245
12 13233

4106841701301,1901,4352503555409361,6943703232853604805051,6504153006751,6504732402646308851,250347

157
195
165
310
310
'Maximumrecordedvalues.'Isolatedfinding. drugsknowntobehepatotoxic:para-aminosali-

cylicacid(4cases),chlorpromazine(3),imipramine(1),phenindione(1),andalcohol(1).Ineight,liverbiopsyshowedappropriatehistologicalchanges.

HAEMOLYTICJAUNDICE(4CASES)Thepatientshadtypicalhaematologicalandbiochemicalfeatures.

MISCELLANEOUSFORMSOFJAUNDICE(21CASES)Theseincludedrecurrentintrahepaticcholestasis(1case),primaryshunthyper-bilirubinaemia(1),Gilbert'sdisease(1),jaundiceassociatedwithperniciousanaemia(1),cirrhosiscomplicatingulcerativecolitis(1),jaundiceassociatedwithmusculardystrophy(1),jaundiceduetochronicvenouscongestionoftheliver(4),lupoidhepatitis(4),andsevencasesinwhich

adefinitivediagnosishasnotyetbeenmade.

PATIENTSWITHOUTJAUNDICEGeneralhospitalpatientsincluded183patientswithavarietyofdiseasesroutinelyadmittedtoageneralmedicalunit,andoneselectedpatient(case34)notedtogivegranularstainingofthyroidepithelium(seebelow)inatestforthyroidmicrosomalantibody.In20patientswithseverechronicthyroiditisthediagnosis

wasconfirmedbyaprecipitinreactionwiththyro-globulin.Nineteenpatientswithperniciousanaemia,22withrheumatoidarthritis,18withsystemiclupuserythematosus,andthreewithidiopathichyper-lipaemiaandxanthomatosiswerealsotested.

METHODS

IMMUNOFLUORESCENCETECHNIQUEBlocksofrenalcortexfromfreshlykilledratswerefrozenontometalchuckswithCO2snow,and6,usectionswerecutinacryostat.Serumtobetestedwasdiluted1in8innormalsalinebufferedwithveronal(pH72)andappliedtotheunfixedSerologicalandhistologicaldiagnosisofprimarybiliarycirrhosiscopyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

R.B.Goudie,R.N.M.Macsween,andD.M.Goldberg

sectionsfor30minutesat20°C.Afterwashingwithbufferedsalinefor10minutesfluorescein-conjugatedantihumanglobulinserum(BurroughsWelcomeandCompany)wasappliedfor30minutesandafterafinal10-minutewashinbufferedsalinethesectionsweremountedinbufferedglycerolandexaminedwithaGillettandSibertconferencemicroscopeusingbluelight.Toreducenon-specificfluorescentstaining,thefluorescein-conjugatedantihumanglobulinserumhadbeenprevious-lyabsorbedwithcellfragments(obtainedbycentrifuga-tionat35,000gforonehourat4°C.)ofratkidneyhomogenizedinbufferedsaline.Serawhichgavepositiveresultswhendiluted1in8weretitratedinfourfolddilutionsinparallelwithsimilardilutionsofnormalcontrolserum.Alltestswerereadindependentlybytwoobserverswhowereunawareofthenatureoftheserabeingtested.

COMPLEMENT-FIXATIONTESTSThesewereperformedonallbutthreeoftheseragivingimmunofluorescentstainingofratkidneyepithelium.Guinea-pigcomplementpreservedinRichardson'sfluid,sensitizedsheepredcells.standardizedbythemethodofOsler,Strauss,andMayer(1952)andcomplement-fixationbuffer(Oxoid)wereused.Theantigenwaspreparedbyhomogenizingratliverin0-25M-sucroseat4°C.withateflonpestle;afterremov-ingnucleiandcelldebrisbycentrifugationat1,500gfor15minutes,thecrudeantigenwasmadebysuspendingincomplement-fixationbufferthedepositobtainedbycentrifugationat35,000gfor60minutesat4°C.Testswereperformedinperspextraysusing0-1ml.ofdoublingdilutionsofseruminactivatedbyheatingat56°C.for30minutes,0-2ml.ofcomplementcontaining5HD50,and0-1ml.ofasuitabledilutionofantigen(usuallythebuttonobtainedfrom1g.ofliverin40ml.ofbuffer).Afteronehourat37°C.withfrequentshaking,0-1ml.of2%sensitizedsheepcellswasaddedandthetraysincubatedforafurtherhourat37°C.,thenleftover-night4°C.atbeforereading.Anticomplementaryactivityofserumwasassessedbytestingserumdilutionsinthepresenceof3HD50ofcomplementwith0-1ml.ofbufferinsteadofantigen;50%haemolysisjudgedvisuallywastakenastheendpoint.

RESULTS

Figure1showspositiveandnegativeimmuno-fluorescencetestsonratkidney.Inthepositivetestwithserumfromapatientwithprimarybiliarycirrhosisthecytoplasmoftherenaltubularepithelialcellsshowsbrilliantfluorescenceandtheglomeruliappearasdarkareasinthesection;withnormalserumthereisfaintgeneralstainingofalltherenalstructures,sometimeswithbrighterstainingoftheluminalmarginsofthetubularcellsandofcastswithinthetubules.Thesetestshavebeenperformedwithseradiluted1in8,sinceundilutednormalserumcausesconsiderableimmunofluorescentstain-ingwhichisdifficulttodistinguishfromatrueposi-FIG.1.Frozensectionofratkidneystainedbyindirectimmunofluorescence.(a)Withprimarybiliarycirrhosisserum.Thereisbrilliantgranularfluorescence(white)ofthecytoplasmofrenaltubularepithelium.Thetwolargedarkspacesareunstainedglomeruli.(b)Withnormalserum.Thereisdiffusefaintnon-specificstainingoftubulesandglomeruli(toprightandbottomlzft).530

orcopyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

Serologicalandhistologicaldiagnosisofprimarybiliarycirrhosis -a4s *.,.>> bb__

O*1x^.l

=3 _h!-_aF- |9E_:s..S.'Es _~~~~~~~%oh".:ii/£om-.*~

FIG.2aFIG.2bFIG.2c

FIG.2.Frozensectionofhumanthyroidstainedbyindirectimmunofluorescence.(a)Withprimarybiliarycirrhosisserum.Thereisdistinctlygranularstainingoftheepitheliumliningthefollicles.(b)Withthyroiditisserum.Theepithelialcytoplasmisuniformlystained.(c)Withnormalserum.Theepitheliumisvirtuallyunstainedapartfromveryweaknuclearfluorescence.

tivereaction.(Incase23,duetounusuallystrongnon-specificstainingatadilutionof1in8,anobviouslypositiveresultwasobtainedonlyafterdilutingtheserum1in32.)Serareactingpositivelywithratkidneywerealsoshowntogivepositiveresultswithsimilarlypre-paredsectionsofratliver,brainandadrenal,humangastricmucosa,adrenalcortexandmedulla,liver,andthyroid.AsnotedbyWalkeretal.(1965),thestainingofhumangastricmucosacloselyresemblesthatseenwiththemajorityofperniciousanaemiasera,whereasthestainingofthyroidepitheliumhasamuchcoarsergranularappearance(Fig.2)thanthatobtainedwiththethyroid-specificmicrosomalantibodyfoundinthyroiditisserum.DetailsofantibodytitresinpatientswithpositiveimmunofluorescencetestsareshowninTableIandtheresultsofallimmunofluorescencetestsaresum-marizedinTableIII.Antibodywasdemonstratedin26of30(87%)caseswithprimarybiliarycirrhosis;inthepositivecasestitresobtainedbyimmunofluor-escencewerenotlessthan1/128andmostofthetitresweremuchhigher.Antibodywasfoundinthreepatients(7°%)withportalorpostnecroticcirrhosis.Case31hadpostnecroticcirrhosisdemonstratedatnecropsythreeyearsafteraclinicallytypicalattackofinfectivehepatitis,case32hadhistologicalevidenceofportalcirrhosis,andcase33,withhepato-splenomegaly,bledfromoesophagealvarices15yearsafteranattackofjaundice.Theseraofcases34and35gaveagranularstainingpatternofthyroidepitheliumduringtestsforthyroidmicrosomalanti-TABLEIII

IMMUNOFLUORESCENCEWIrHRATKIDNEYINJAUNDICED

PATIENTSANDCONTROLS

No.TestedNo.Positive

JaundicedPatientsPrimarybiliarycirrhosisPortalandpostnecroticcirrhosisExtrahepaticbiliarytractobstructionSecondarybiliarycirrhosisInfectivehepatitisDrug-inducedjaundiceHaemolyticjaundiceMiscellaneous

PatientswithoutjaundiceGeneralhospitalpatientsChronicthyroiditisPerniciousanaemiaRheumatoidarthritisSystemiclupuserythematosusIdiopathichyperlipaemiaandxanthomatosis30

427252510421

184

2019221826(87%)3(7%)000000

1(05%)001(<5%)0

0

body,andweresubsequentlytestedwithratkidney.Case34hadosteoarthritis,chronicbronchitis,andanE.S.R.of48mm.inonehour,andcase35hadclassicalrheumatoidarthritis.Bothhaveslightlyraisedserumcholesterollevels.Thesetwospeciallyselectedcaseshave beenincludedintheappropriatecontrolgroupsinTableIII,andgiveanexaggeratedincidenceofpositivetestsingeneralhospitalpatients(05%)andincasesofrheumatoidarthritis(5%).Positivetestswerenotfoundintheotherjaundicedandcontrolgroupsstudied.Figure3illustratespositiveandnegativecomple-ment-fixationreactionsusingthe35,000gsediment531copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

R.B.Goudie,R.N.M.Macsween,andD.M.Goldberg

*m-.... e.......... +. e'^^< >e::A.,zIAte ,;......R *:.*.4*.t4:M.;';*~~~~~~~~~~~96.::aa- 0...

1-i0:JOY

aa**wr-t.: i** ****A3).

HIG.3.Complement-fixationtestwithserafromprimarybiliarycirrhosis.Thecontrolrowscontainserum,3HD50ofcomplementandnoantigen(testforserumanticomplementaryactivity).Thetestrowscontainserum,5HD50ofcomple-

mentandantigenpreparedfromratliver.Tests21and22arenegative,and23and24arepositive(titre256).

ofratliverhomogenateasantigen.DetailsoftitresobtainedareshowninTableI.Positivecomplement-fixationreactionswerefoundinthosepatientswhose

serumgaveimmunofluorescentstainingofratkidneyepithelium.Thecorrelationcoefficient(r)betweenlog2immunofluorescenceandcomplement-fixationtitresis072(t=-804,p<0001).

HISTOLOGICALSTUDIES

Thefollowingdescriptionisbasedontheappearanceofparaffinsectionsstainedbyhaematoxylinandeosinoftheeightwedgebiopsiesofliverobtainedatlaparotomy(cases3,7,10,14,17,20, 22,and28)and

onthefournecropsycases(cases2,4,17,and23)whichwereavailabletousforstudy.

EVIDENCEOFCIRRHOSISWell-developedmono-lobularcirrhosiswithfibrousbandsregularlyextend-ingroundliverlobulestolinkadjacentportaltractswasseeninoneofthewedgebiopsiesandintwoofthefournecropsycases,whilemonolobularcirrhosiswithlinkingofonlysomeofthe portaltractswas

seeninthreewedgebiopsies.Inthreebiopsiesthere werefibrousbandsextendingoutfromtheportal

tracts(portaltractfibrosis)butcontinuitybetweenadjacenttractswasnotestablished.Thecirrhosiswasofdistinctlymultilobulartypeinthreecases,allofthreeyears'durationorlonger.

CHOLESTASISCholestasis,asshownbybilethrombiand/orintracellulargranulesofbilepigment,was presentinonlyfivewedgebiopsiesbutwasacon-sistentfeatureinthenecropsymaterial.

CELLULARINFILTRATIONOFPORTALTRACTSThis

waspresentineverycase,andconsistedmainlyoflymphocyteswithavariablenumberofplasmacells.Theinfiltratewasdiffuse,frequentlywithnodulesofdenselyaggregatedcells.Theinfiltratingcells,r

532copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

Serologicalandhistologicaldiagnosisofprimarybiliarycirrhosis

togetherwithfibroblasts,frequentlyinvadedthelimitingplateoflivercellsroundtheportaltractsgivinganerodedappearance.Afeweosinophilpoly-morphswerepresentintheportalinfiltrateinmostcases,andinathirdofthecasessmallnumbersofneutrophilpolymorphswerenoted.

SMALLINTRAHEPATICBILEDUCTSThoselessthan50,uindiameterweremarkedlyreducedinnumberin10ofour12cases(Fig.4).Infourofthesetherewasfocalproliferationoftubularstructuressimulat-ingbileductproliferationandcomposedofflateosinophiliccellslyingnearthehepaticlobulesandpossiblyderivedtherefrom.Markedproliferationofbileductswasobservedincase23,whichwasof11years'duration(Fig.5).

MEDIUMSIZEDINTRAHEPATICBILEDUCTSThosebetween50and100p.indiametershowedperi-cholangitisinsixofthesevenwedgebiopsies.Thiswascharacterizedbyaccumulationoflymphocytesandplasmacellsaroundthesestructuresandbyinfiltrationofthebileductepitheliumbylympho-cytes,withswelling,eosinophilia,andheaping-upoftheepitheliumandabsenceofsmallbranches(Figs.6and7).Thisfeaturewasidentifiedinonlyoneofthefournecropsycasespossiblyduetodifficultiesim-posedbypost-mortemautolysis.

LIVERCELLDEGENERATIONLivercelldegenerationoffeatherytypewascommon,especiallyinareasshowingerosionofthelimitingplate.Areasoffocallivercellnecrosiswithinthelobules,andusuallysurroundedbypolymorphs,wereseeninsixofthewedgebiopsies.Thefrequencyofthischangeinthenecropsycasesisuncertainbecauseofpost-mortemautolysis.Mildfattychangewasseeninhalfthetotalcases.

GRANULOMATALocalizedaggregatesofendothe-lioidcellsresemblingsarcoidfollicleswerefoundintheportaltractsandinthehepaticparenchymaintwoofthewedgebiopsies(Figs.8and9).IntwootherwedgebiopsieslessclearlydemarcatedclustersofendothelioidcellsasdescribedbyGrevilleWilliams

b~~~~~~~~~~~~~~fiA-L'It At'Lkq4~FIG.4.Portaltractinprimaiy~~biliArycrhssMhwigfboi,fcanifsymhctcadpasaclnta

don,4erosionaothecliiingpiateofiliverycelloss,andoabsnceofibroileducts.aneathfueliverhobulesarepafwsmaclltubultar

structures,probablyderivedfromlivercells(haemzatoxylinandeosinx135).533copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

R.B.Goudie,R.N.M.Macsween,andD.M.Goldberg

FIG.6 FIG.5

FIG.5.Proliferationofsnmallbileductsinfibroustissueinprimarybiliarycirrhosisoflongduration(haematoxylinandeosinx135).

FIG.6.Mediumsizedbileductinprimarybiliarycirrhosis.Thereismassivelymphocyticandplasmacellinfiltrationoftheportaltractandthesecellshaveinvadedthebileduct.Thereisepithelialdegenerationintheleftlowerpartoftheduct.Notetheabsenceofsmallbranches(haematoxylinandeosinx150).

FIG.7.Mediumsizedbileductinprimarybiliarycirrhosis.Theheaped-upepitheliumisinfiltratedbylymphocytesandplasmacellsasistheperiductaltissue.Thereisanill-definedclusterofendothelioidcellsatthetoprighthandcorner(Massonx370).

wt..,P%*ft&I..%qmb.a'lbNo0OF.,*0404-,t.ir-lbAl.t0

FIG.7534copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

Serologicalandhistologicaldiagnosisofprimarybiliarycirrhosis .FIGis.9 FIG.9 FIG.8.Primarybiliarycirrhosis.Sarcoid-likefolliclecomposedofendothelioidandsmallgiantcellsinaportal tract(haematoxylinandeosinx205). FIG.9.Primarybiliarycirrhosis.Sarcoid-likefolliclewithinliverlobule(haematoxylinandeosinx205).

FIG.10.Primarybiliarycirrhosis.NecroticfocuswithLanghansgiantcellandendothelioidcells,resemblingtuberculousfollicle(haematoxvlinandeosinx205).

FIG.10FIG.8535copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

R.B.Goudie,R.N.M.Macsween,andD.M.Goldberg

(1965)wereseenintheportaltracts(seeFig.7).InonenecropsycasenecroticfollicleswithLanghanstypegiantcellswerenoted(Fig.10);notuberclebacilliweredemonstratedinsectionsfromthisliver,whichwasunfortunatelytheonlyorganavailableforexamination.

DISCUSSION

Ourfindingthat87%ofpatientsdiagnosedascasesofprimarybiliarycirrhosisandasmallproportionofpatientswithothertypesofcirrhosishavecirculatingantibodywhichisreadilydemonstrablebygranularimmunofluorescentstainingofratkidneyepitheliumlargelyconfirmstheobservationsofWalkeretal.(1965).Complementfixationbetweenprimarybiliarycirrhosisseraandtissueextractshaspreviouslybeendescribed(Gajdusek,1958;Deicher,Holman,andKunkel,1960;Walkeretal.,1965)andourdemon-strationofasignificantcorrelationbetweencomple-ment-fixationtitreswithratliverextractandimmunofluorescentstainingofratkidneysuggeststhatbothmethodsmaybroadlyreflectthesameantigen-antibodysystems.Theoccasionalfindingofdisparitybetweentitresbythetwomethodsmaybeaconsequenceoftheindefiniteendpointobservedwithtitrations,especiallybytheimmunofluorescencetechnique,ormayindicatesomevariationinthespecificityoftheantibodyinindividualpatients.Beyondconfirmingabsenceoforganandspeciesspecificity,wehavenothingtoaddconcerningthenatureoftheantigenconcernedinthesereactions.FollowingthestudiesofAhrens,Payne,Kunkel,Eisenmenger,andBlondheim(1950)primarybiliarycirrhosishasbeenestablishedasaclinicalentitythoughitisdifficulttodefineitprecisely.Probablythelackofapreciseclinicaldefinitionaccountsforourfindingofantibodyinonly870%ofcasesofprimarybiliarycirrhosisincontrasttothe10000reportedbyWalkeretal.(1965).Oursero-negativecasesareindistinguishablefromthosewithpositiveantibodytestsontheclinical,biochemical(TablesIand11),andhistologicalevidenceavailabletous.Itremainstobeseenwhetherthesero-negativecasesarefundamentallydifferentfromthosewithanti-body,oraresufferingfromessentiallythesamediseaseinwhichantibodyformationismerelyacommonepiphenomenon.Thefindingofantibodyinoccasionalcasesofportalorpostnecroticcirrhosismayindicatethatthesecasesareatypicalvariantsofprimarybiliarycirrhosis(e.g.,case33whichhasarelativelyhighserumalkalinephosphataseof35K.A.units/100ml.)ormaybetakenasfurtherevidenceagainstthefundamentalimportanceofantibodyformationinthepathogenesisofprimarybiliarycirrhosis.Experimentalfindingsfavourthelatterviewsinceliverdamagebycarbontetrachloridecausesanti-mitochrondrialantibodyformationinrats,butnotthedevelopmentofbiliarycirrhosis(Weir,1963);andindiseasesinwhichautoimmunityisbelievedtohaveapathogenicrole,e.g.,experimentalallergicencephalomyelitisorthyroiditis,sensitizedlympho-cytes(delayedhypersensitivity)andnotcirculatingantibodyarethoughttobetheharmfulagents(Paterson,1960;Felix-DaviesandWaksman,1961).Whetherornotantibodyisofimportanceinthepathogenesisofprimarybiliarycirrhosis,itisclearthattheimmunologicaltestisanimportantnewparameterinthestudyofhumanliverdisease.Itstronglysupportstheviewthat,byandlarge,primarybiliarycirrhosisasdescribedinclinicalandbio-chemicaltermsisarealentity.Furthermoreitprovidesausefulconfirmatorytestfordistinguishingthemajorityofcasesfromconditions,suchasextra-hepaticbiliarytractobstructioninwhichsurgicalexplorationiscalledfor,andfromdrug-inducedcholestaticjaundice.Thebiochemicalfeaturesobservedinthepresentsubjectsconformtothepatternusuallyassociatedwithprimarybiliarycirrhosis(Ahrensetal.,1950;Sherlock,1959).Thisincludesdisproportionatelyhighlevelsofserumalkalinephosphataserelativetothefrequentlymilddegreeofjaundice.Theveryhighbilirubinvaluesrecordedforcases4and11weresporadicorterminalphenomena,whereasthealka-linephosphatasewasneververymuchlowerthanthehighestobservedlevelsdocumentedinTableII.Thymolturbiditywasusuallyabnormalinestablishedcases,thoughnotinfrequentlynormalonfirstpresentation.Ingeneral,valuesforthistestwerehigherthanthoseencounteredinextrahepaticobstruction,thoughitwasuncommontofindsuchdramaticelevationsasareseeninahighproportionofcaseswithportalcirrhosis.Ontheotherhand,transaminaseactivitieswereusuallyhigherthanthosemetwithinportalcirrhosiswithcomparablebili-rubinaemia.Unlikethepatternmostoftenen-counteredinnon-malignantextrahepaticobstruction,activityofserumglutamicoxalacetictransaminase(SGOT)wasfrequentlyhigherthanthatofSGPTinthesamespecimen.Inmanycasestherewasaslightormoderateelevationinserumglobulin,butcom-paredwithtypicalcasesofportalcirrhosis,thealbuminwasonlyslightlydepressed.Theserumcholesterollevelwasraisedin25casesand,wheremeasured,serumphospholipidswerealsoelevated.Theseabnormalitiesneednotbeassociatedwithdermalxanthomataandshouldalwaysbesoughtinpatientswithlong-standingjaundicesincetheyaremostusefulindicesfordistinguishingprimarybiliaryfromotherformsofcirrhosis.536copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

Serologicalandhistologicaldiagnosisofprimarybiliarycirrhosis

Thepathologicalchangesintheliverinprimarybiliarycirrhosisseemtobelesswidelyknownthantheclinicalfeatures.Inthecasesdescribedinthispaperseveralincorrecthistologicaldiagnoseswereinitiallymadewhichmisledtheclinicians;theseincludedportalcirrhosis,cryptogeniccirrhosis,continuinghepatitis,lupoidhepatitis,andcholan-gitis.Theseerrorsundoubtedlyarisefromun-familiaritywiththisratheruncommondisease,fromattempteddiagnosisonbiopsiestoosmalltorevealthecharacteristicchangesinthemedium-sizedandsmallbileducts,fromvariationinliverhistologydependingonthedurationofthedisease,andfromtheconflictingaccountsofthediseaseinthelitera-ture.WecanconfirmSherlock'sstatements(1963)thatneedlebiopsyisusefulmainlyfortheexclusionofextrahepaticbiliarytractobstructioninwhichthereismarkedproliferationoftheintrahepaticbileducts(Fig.11),andthatinsteadofthewell-knownmonolobularcirrhosistheremaybemerelyportaltractfibrosisorfrankportalcirrhosisinpatientswithprimarybiliarycirrhosis.ThestatementbyCameronandHou(1962)thatbilethrombiareinvariablypresentisapparentlybasedonnecropsymaterial;asemphasizedbyPopper,Rubin,andSchaffner(1962)theymaybeabsent,particularlyintheearlystageofthedisease.Undoubtedlythemostcharacteristichistologicallesionofprimarybiliarycirrhosisisthereductioninnumberofthesmallintrahepaticbileducts,butfocalproliferationofstructuresresemblingverysmallbileductsisnotuncommon,andinthelatestages,whenportalcirrhosisisestablished,bileductproliferationmaybewidespread.Theotherfindingsfavouringapositivediagnosisofprimarybiliarycirrhosisisthepericholangitisofmedium-sizedintrahepaticbileductsdescribedbyGrevilleWilliams(1965),thoughthislesionisoftendifficulttofindandissometimesabsent.Thepresenceofgranulomataresemblingsarcoidfollicleshasprevious-lvbeennotedbyRubin(1963),andScheuer(1966)hasdrawnourattentiontohisfindingoftheselesionsinaboutonethirdofcasesofprimarybiliarycirrhosis.

Theinfiltrationofportaltractsbylymphocytesandplasmacells,theproliferationoffibrousconnec-tivetissue,andthealterationofepithelialstructuresfoundintheliverinprimarybiliarycirrhosismayberegardedasanalogoustothelesionsofautoimmunethyroiditisandgastritis.Localizedgranulomaformationisnot,however,afeatureofthelesionsassociatedwithorgan-specificautoimmunity.Granu-lomatacomposedofendothelioidcellsaresometimesfoundinthelesionsofulcerativecolitis(Ackerman,1959)andofsystemiclupuserythematosus(Teilum,1946)andonmorphologicalgroundstheremightbeFIG.11.Secondarybiliarycirrhosisshowingfloridproliferationofsmallbileducts(haematoxylinandeosinx135).

reasontosuspectthatprimarybiliarycirrhosishasapeculiarityofimmunologicalreactivityincommonwiththesediseases.

Wewishtorecordourthankstothefollowingphysicianswhosupplieduswithsera:Drs.H.T.Howat,W.J.Cunliffe,andJ.P.O'Loughlin,ManchesterRoyalInfirmary;Dr.G.P.Lewis,StobhillHospital,Glasgow;Drs.A.G.MelroseandR.Hume,SouthernGeneralHospital,Glasgow;Drs.A.LyellandJ.A.Boyle,RoyalInfirmary,Glasgow.Dr.I.RoittoftheMiddlesexHospital,London,supplieduswithouroriginalpositivecontrolsera.Mr.E.MacDonaldandMissM.Barrgavevaluabletechnicalassistance.

REFERENCES

Ackerman,L.V.(1959).SurgicalPathology,2nded.,p.383.Mosby,St.Louis.Ahrens,E.H.,Jr.,Payne,M.A.,Kunkel,H.G.,Eisenmenger,W.J.,andBlondheim,S.H.(1950).Medicine(Baltimore),29,299.Cameron,R.,andHou,P.C.(1962).BiliaryCirrhosis,p.149.OliverandBoyd,EdinburghandLondon.Deicher,H.R.G.,Holman,H.R.,andKunkel,H.G.(1960).Arthr.andRheum.,3,1.537copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from

R.B.Goudie,R.N.M.Macsweein,andD.M.Goldberg

Felix-Davies,D.,andWaksman,B.H.(1961).Ibid.,4,416.Gajdusek,D.C.(1958).Arch.intern.Med.,101,9.Osler,A.G.,Strauss,J.H.,andMayer,M.M.(1952).Anmer.J.Syph.,36,140.Paterson,P.Y.(1960).J.exp.Med.,111,119.Popper,H.,Rubin,E.,andSchaffner,F.(1962).Amer.J.Med.,33,807.Rubin,E.(1963).Gastroenterology,45,400.Scheuer,P.J.(1966).Personalcommunication.Sherlock,S.(1959).Gastroenterology,37,574.Sherlock,S.(1963).DiseasesoftheLiverandBiliarySystem,3rded.,

p.415.Blackwell,Oxford.Teilum,G.(1946).Actamed.scand.,123,126.Walker,J.G.,Doniach,D.,Roitt,1.M.,andSherlock,S.(1965).Lancet,1,827.Weir,D.M.(1963).Immunology,6,581.Williams,G.E.G.(1965).J.Path.Bact.,89,23.Wr6blewski,F.(1958).Advanc.clin.Chem.,1,314.538copyright. on August 15, 2023 by guest. Protected byhttp://jcp.bmj.com/J Clin Pathol: first published as 10.1136/jcp.19.6.527 on 1 November 196

6. Downloaded from