clinical associations betweenarthritis and liver diseasejaundice seemed equally effective. attempts...

Annals of the Rheumatic Diseases, 1982, 41, 295-307

Occasional survey

Clinical associations between arthritis and liverdiseasePETER R. MILLS AND ROGER D. STURROCK

From the Gastroenterology Unit, Royal Infirmary and Centre for Rheumatic Diseases, Baird Street Hospital,Glasgow, Scotland

The purpose of this paper is to review clinical associa-tions between diseases of the liver and joints. Sincethe publication of an excellent review of arthropathyand liver disease by Whelton' in 1970 there has beenan increasing awareness of the systemic mani-festations of many primary liver or joint diseases.Recognition of the pattern of systemic involvementof a disease may be of considerable assistance inmaking a firm diagnosis and may help elucidate thecause of these diseases where so little understandingof the basic pathophysiology exists. In addition, areview of the more common examples of liver diseaseinduced by antirheumatic drugs will be included.The first observation of a possible link between

arthritis and liver disease was the recognition of anameliorating effect of jaundice in patients withrheumatoid arthritis reported by Still2 in 1897 andWishart3 in 1903. Philip Hench at the Mayo Clinic inthe 1930s confirmed that 31 out of 45 patients withrheumatic conditions obtained complete (71 %) orincomplete (29%) remission from joint symptomsfor a mean of 19-5 weeks following the spontaneousonset of jaundice.4 The remission was dependent onthe degree of jaundice, requiring serum bilirubin ofconcentration greater than 8 mg/100 ml (137 ,umoIl)rather than the cause, as viral hepatitis or obstructivejaundice seemed equally effective. Attempts toreproduce this phenomenon artificially by intraven-ous infusion of bile salts, human bile, liver extracts, orhighly jaundiced blood were all initially unsuccess-ful.5 However, an infusion of a combination of biliru-bin and dehydrocholic acid was then reported to giveprolonged pain relief and reduction in joint swellingto 10 patients with rheumatoid arthritis.6 Infusion ofchenodeoxycholic acid alone was, much later, alsoshown to produce mild transient relief of pain in 6 outof 10 rheumatoid arthritis patients.7 Interest in theCorrespondence to Dr P. R. Mills, Gastroenterology Unit, RoyalInfirmary, Glasgow G4 OSF.

therapeutic response obtained continued, with delib-erate induction of viral hepatitis resulting in tempor-ary remission of symptoms in most rheumatoidarthritis patients, the improvement in several casespreceding the development of jaundice.8The search for 'Nature's' active agent in this

response which could be induced by jaundice, butprobably also by pregnancy,'0 finally culminated inthe hypothesis that corticosteroid hormone metabol-ism was deranged in these conditions. The firstdramatic responses to corticosteroid therapy inrheumatoid arthritis were then reported." However,while the rate of hepatic cortisol conjugation hassubsequently been demonstrated to be reduced inpatients with acute and chronic liver disease, this inturn results in diminished ACTH release, and asteady state is re-established in which cortisol synth-esis is reduced and plasma levels remain normal.'2Further exploration of the mechanism underlying thebeneficial effect of acute hepatic dysfunction onchronic arthritis therefore continues.

In an ingenious study Pinals'3 demonstrated that inrats the hepatotoxin dimethylnitrosamine mayrapidly produce a severe but transient necrotisinghepatitis without causing jaundice, as the rat liver canhandle large bilirubin loads. If dimethylnitrosaminewas given to the rat on the fourth day after an injec-tion of Freund's adjuvant, which would normallyproduce adjuvant arthritis within a few weeks, thenpartial suppression of the subsequent developmentof arthritis resulted. Injection of necrotic rat liverextracts, but not normal rat liver, also produced thesame effect, suggesting that some anti-inflammatorymaterial was being released from necrotic liver andinhibiting the development of adjuvant arthritis.Acute hepatic injury rather than jaundice itself there-fore appears to be all that is necessary to obtain therequired effect in adjuvant disease. The mechanismthus still remains unsolved and must surely be a

295

on March 4, 2020 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.41.3.295 on 1 June 1982. Dow

nloaded from

http://ard.bmj.com/

296 Mills, Sturrock

rewarding area for further experimentation. Thepossible immunosuppressive effect of acute liverinjury has not yet been invoked as the cause of rapidremission of arthritis symptoms.The review will be divided into the following sec-tions:(A) Primary diseases of the liver in which joints maybecome involved.

Primary biliary cirrhosis.Chronic active hepatitis.Acute virus hepatitis.Haemochromatosis.Wilson's disease.

(B) Primary diseases of joints in which the liver maybecome involved.Rheumatoid arthritis/Sjogren's syndrome.Felty's syndrome.Juvenile chronic arthritis.Adult Still's disease.Systemic lupus erythematosus.Systemic sclerosis.Polymyalgia rheumatica.Polyarteritis nodosa.

(C) Antirheumatic drug-induced liver disease.

(A) Primary diseases of the liver in which joints maybecome involved

PRIMARY BILIARY CIRRHOSISPrimary biliary cirrhosis (PBC) is an uncommon formof chronic liver disease, of unknown aetiology, whichpredominantly affects middle-aged women. The dis-ease usually presents with pruritus followed by slowlyprogressive jaundice but is clearly a multisystem dis-order, the majority of patients having associated siccasyndrome, renal tubular acidosis, and pancreatichyposecretion.'4'16Rheumatoid arthritis has been described in

5-9 7% of PBC patients. 7 9 Rheumatoid factor waspresent in 73 (64%) out of 114 patients, but in lowtitre, being s 1/64 in 72% of the positive cases."8 Aradiological study of hands in 13 PBC patientsshowed that 12 had distal small joint erosions whichwere largely asymptomatic.20 However, in a largerseries with more extensive radiological screening theprevalence of joint erosions was 13 (31 %) out of 42patients, 8 of whom had a positive rheumatoid factorbut only 4 suffered from symptomatic arthritis.21 Cir-culating immune complexes were detected by Clqbinding assay in 31 (62%) of 50 PBC patients, 17 ofwhom had arthritis.22 This took the form of aseropositive inflammatory polyarthritis in 12 of 18patients with Clq binding >20%, whereas a milderseronegative arthritis associated with sclerodermaand Raynaud's phenomenon was found in 5 of 13patients with Clq binding <20%, suggesting that

circulating immune complexes may be involved in thedevelopment of inflammatory polyarthritis in PBCpatients.The sicca complex of dry eyes and mouth was

found in 72 % of patients with PBC, 42 % withchronic active hepatitis, and 38% with cryptogeniccirrhosis, although less than 5 % had accompanyingarthritis completing Sjogren's syndrome.23 If investi-gated with a battery of tests, including Schirmer'stest, rose Bengal ocular staining, parotid gland sec-retory sialography, salivary gland scintigraphy, andlip biopsy, then all of 14 PBC patients had 2 or moreabnormal tests indicating features of the sicca com-plex.14

Scleroderma was first described in association withPBC in 1964,24 since when there have been severalconfirmatory reports.19 25-27 Later the association ofthe whole CRST syndrome of calcinosis cutis,Raynaud's phenomenon, sclerodactyly, and telan-giectasia was noted.26 In a detailed study of 83 PBCpatients 14 (17%) were found to have scleroderma,which was usually confined to mild cutaneouschanges and loss of oesophageal peristalsis.27 How-ever, one patient had extensive progressive systemicsclerosis which resulted in death from cardiopulmo-nary failure. Of the 14 patients with scleroderma 10had Raynaud's phenomena, 7 had telangiectasia ofhands and face, and only 2 had calcinosis. The com-plete CRST syndrome has been noted in 2 4-5 *4 % ofseries of PBC patients.17 19 27

Hypercholesterolaemia is a common biochemicalfinding in PBC patients and may be associated withthe formation of symptomless periarticular bonecysts, especially of the small bones of the hands.28 Inaddition hypercholesterolaemia may cause a trans-ient, flitting polyarthritis such as is found in patientswith familial hyperbetalipoproteinaemia.29 Hyper-cholesterolaemic arthropathy has been described in apatient with PBC as a transient flitting polyarthritis,often initiated by unusual exercise, with joint effu-sion and erythema of the overlying skin.30 Choles-tyramine is recommended in the long-term manage-ment.

Polymyalgia rheumatica has been noted in 3patients thought to have PBC who responded well tocorticosteroids." However, 2 of these patients didnot have a liver biopsy, the diagnosis of PBC restingon elevation of serum alkaline phosphatase and IgMand a positive mitochondrial antibody. With somejustification it has been argued that the diagnosis ofPBC was not adequately confirmed.32The association of cirrhosis with digital clubbing is

well established, especially in PBC and chronic activehepatitis.' Hypertrophic osteoarthropathy was con-sidered to be a very rare accompanying feature ofchronic liver disease until the description of 20 cases



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

Occasional survey: Clinical associations between arthritis and liver disease 297

by Epstein et al.33 in 1979, 10 of which had PBC.Subsequently prospective radiological surveys of theskeleton have revealed that 35% and 38% of PBCpatients have periosteal new bone formation.2 34Thedistribution of periostitis between these 2 series dif-fered, for in the former the lower tibia and fibula andin the latter the first metacarpal were the most fre-quently affected sites. Many of these patients have nosymptoms, though marked tenderness over the distalleg bones, when present, may be a reliable sign ofunderlying periostitis. Periostitis is found less com-monly in other forms of chronic liver disease.2 34Thereason for the development of periosteal new boneformation in chronic liver disease remains unknown.No difference could be established between patientswith liver disease with and without periostitis in rela-tion to cardiac index, degree of right to left pulmo-nary arteriovenous shunting, presence ofoesophageal varices or surgical portacaval shunts,arterial Po, blood levels of growth hormone, para-thyroid hormone, and vitamin A, and urinary excre-tion of oestrogen.."

CHRONIC ACTIVE HEPATITISChronic hepatitis is defined as chronic liver diseaseexisting for longer than 6 months and is dividedpathologically into 2 categories, chronic persistentand chronic active hepatitis. Chronic active hepatitismay have several causes, including hepatitis B virusinfection, drugs, Wilson's disease and alpha-one anti-trypsin deficiency, and even excess alcohol may causesimilar changes. However, this discussion will belimited to chronic active hepatitis of unknown cause,predominantly affecting young women and constitut-ing the largest subgroup found in Britain. This dis-ease is postulated to have an immunological basis,and multisystem organ involvement is common.3'

Joint symptoms are common in chronic activehepatitis but take the form of arthralgia and stiffnessmore often than true inflammatory arthritis.' Thearthralgia may be of acute onset, transient or recur-ring, migrating from joint to joint, particularly involv-ing large joints such as knees, ankles, wrists, andelbows and tending to reflect disease activity of theliver. Occasionally joint symptoms may precede theonset of liver disease. The arthralgia is often accom-panied by swelling and erythema of the periarticularstructures, but joint effusion is uncommon. Thesechanges resolve without causing joint deformity.Occasional patients with all the features ofrheumatoid arthritis are seen.

Joint symptoms were described in 11-28% ofpatients with chronic active hepatitis in the 4 largestseries.' 1 3 3 Rheumatoid factor and antinuclear fac-tor were present in 95% and 84% of patients witharthralgia and were slightly less commonly found at

73% and 64% in patients without arthralgia.' In 13patients with arthralgia who had radiological studiesone had changes of rheumatoid arthritis, one showedearly osteoarthrosis of both elbows, and 2 showederosion of the ulnar styloid together with erosion of aphalangeal head in one patient.' No radiologicalabnormality was seen in the remaining 9 patients.A striking case of painless seronegative destructive

arthritis of small joints of the hands and feet has beendescribed in association with chronic activehepatitis.38 The patient developed marked swellingand warmth of the involved joints, and radiologyshowed erosions and periosteal new bone formation.The arthritis resolved spontaneously after 10months, leaving the joints clinically appearing nor-mal. Synovial biopsy showed moderate villous prolif-eration and a striking abundance of surface fibrinoidmaterial resembling a rheumatoid nodule. Immuno-fluorescent studies revealed IgG and C3 complementin the surface fibrinoid, suggesting the possibility ofimmune complex deposition.

VIRAL HEPATITISArthritis has been associated with the prodromalstages of several viral illnesses, including rubella,mumps, and arboviruses.39 The earliest description ofarthritis occurring in viral hepatitis was given byRobert Graves, an Irish physician, in 1843, and areview of the substantial literature summarising ourclinical knowledge was published by Fernandez andMcCarty in 1971.40Acute viral hepatitis is often associated with a

serum-sickness-like illness characterised by arthral-gia, frank arthritis, skin rash, angioneurotic oedema,and occasionally haematuria and proteinuria.4'These clinical features usually precede the onset ofjaundice but, as 80% of patients with viral hepatitisare anicteric,42 may be the sole indicators of the ill-ness, thus causing difficulty in establishing the correctdiagnosis unless serum transaminase values arechecked. The arthritis is usually mild, transient, andleaves no residual joint deformity. In occasionalpatients, however, the presentation may be moresevere and prolonged, imitating the clinical picture ofrheumatoid arthritis. The pattern of joint involve-ment is variable but usually symmetrical, affectingperipheral joints, especially the proximal inter-phalangeal joints, but large joints and the spine mayalso be involved. Joint swelling and morning stiffnessare both recorded.

Joint symptoms are reported to have occurred in118 (36%) of 324 patients at some stage during thecourse of viral hepatitis.43 In contrast to an earlierstudy,44 in which all 18 patients had virus B hepatitis,there was no difference in the incidence of jointsymptoms between HBsAg positive and negative



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

298 Mills, Sturrock

patients. However, 38% of the HBsAg positivepatients suffered joint symptoms for more than 2weeks compared with 15% of those in whom theantigen was not detected. The frequency of jointsymptoms in the HBsAg negative group increasedwith age from an incidence of only 18% in children to45 % in adults over 30 years.

The remarkable similarity of this syndrome toserum sickness suggests that immune complexesmight play a role in its pathogenesis. There is strongevidence that the hepatitis B virus may produce com-plement-binding immune complexes, which are

detected only during the period of activity of the jointdisease, whether in association with acute viralhepatitis or chronic active hepatitis.4446 During theacute phase of joint disease both total serum

haemolytic complement (CH 50) and C4 were

severely depressed." Cryoprecipitates extractedfrom serum contained IgG, IgM, and IgA, comple-ment components C3, C4, and CS, HBsAg andhepatitis B surface antibody (anti-HBs) in high con-

centration in comparison with serum alone and couldactivate both the classical and alternate complementpathways in vitro.45 These immune complexes disap-pear with the resolution of joint symptoms. However,direct virus infection of synovial cells has also beensuggested because of the finding of virus-like parti-cles deep in synovial cells in cases of HBsAg positiveassociated arthritis.47

HAEMOCHROMATOSIS AND WILSON'SDISEASEThe joint symptoms associated with these 2 rare

inherited disorders of iron and copper metabolismrespectively have been the subject of a recent com-prehensive review by Dr E. B. D. Hamilton48 andhave also been well covered in a standard textbook.49This short account is therefore limited to recent addi-tions to the literature.

In haemochromatosis a further warning is giventhat chronic arthropathy may be the only initialsymptom of the disease.50 In 3 patients, 2 of whompresented with arthritis in the hands and one withpain in both shoulders, there were no other abnormalphysical signs and liver function tests were normal.The finding of high serum iron levels suggested thecorrect diagnosis, which was confirmed by observingexcess quantities of iron in liver biopsy specimens.Phlebotomy had no effect on the joint symptoms, butthe importance of recognising and treating iron over-

load before the development of overt haemo-chromatosis is stressed.

In a review of arthropathy in Wilson's diseaseGolding and Walshe found symptoms or signs relat-ing to the locomotor system in 24 (75 %) of 32 con-secutive hospital admissions of patients with Wilson's

disease.5" The most common symptoms were painand stiffness in the knees (9 patients) and spine (16patients) associated with premature osteoarthrosis.Joint hypermobility was found in 9 patients in theabsence of any features of Marfan's syndrome orhomocystinuria. Five patients had suffered attacks ofacute polyarthritis resembling acute rheumatoidarthritis which were thought to be related to penicil-lamine therapy. No clinical, biochemical, or radiolog-ical evidence of osteomalacia or rickets wasobserved. A radiological survey revealed generalisedloss of bone density (21 cases), premature osteoarth-rosis (8 cases-mean age 28 years), spinalosteochonditis (5 cases) and a tendency to narrowingof intervertebral joint spaces and squaring of ver-tebral bodies.Recurrent polyarthritis has been reported in a

child with Wilson's disease before starting penicil-lamine therapy.52 Technetium scintigraphy of jointswas described in 25 patients with Wilson's disease, 11of whom had joint pains.53 Evidence of synovitis wasreported to have been found in 22 joints of 10patients and postinflammatory articular degenera-tion in 15 joints of 8 patients. Unfortunately thesefindings showed little correlation with the presenceor absence of joint symptoms.

(B) Primary diseases of joints in which the liver maybecome involved

RHEUMATOID ARTHRITIS/SJOGREN'SSYNDROMERheumatoid arthritis and Sjogren's syndrome will beconsidered together. Rheumatoid arthritis is a com-mon chronic, predominantly articular, conditionwhich has many systemic manifestations. The ques-tion of liver involvement has been examined by manyclinicians, and the evidence for and against will bepresented.The clinical evidence for liver disease in

rheumatoid arthritis is limited. Hepatomegaly wasreported in 23 (10-6%) of 216 patients withrheumatoid arthritis, 5 (15 6%) of 32 patients withSjogren's syndrome, and only one (0 6%) of 170patients with osteoarthrosis.54 Liver scintigraphy hasbeen used to assess liver size, when 7 (22 %) out of 32rheumatoid arthritis patients were shown to havehepatomegaly.55 Other clinical signs of chronic liverdisease are usually lacking.Abnormal biochemical tests of liver function have

been reported in a large proportion of patients. Theconcept of a 'rheumatoid liver' was based on thefinding that 26% of rheumatoid arthritis patients hadraised serum alkaline phosphatase levels, which wereconfirmed to be of hepatic origin by correlation withelevated serum 5-nucleotidase concentrations.56-58



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


These results have been confirmed with serumalkaline phosphatase levels being elevated in 18%,35 %, and 51 % of 3 large series of rheumatoid arth-ritis patients,54 59 60 the isoenzyme being of liverorigin in the majority of cases. In 100 rheumatoidarthritis patients serum alkaline phosphatase waselevated in 47% of patients with features of the siccacomplex but in only 19% of those with normal saliv-ary and lacrimal gland function.59 Dynamic tests ofhepatic function, in particular the bromsulphthalein(BSP) excretion test were abnormal in 5 6% ofrheumatoid arthritis patients and 18 8% of Sj6gren'ssyndrome patients.54 Serum bilirubin and the trans-aminases are, in contrast, almost invariably nor-mal.' 59

Serum autoantibody tests may be useful markersfor autoimmune liver disease. Mitochondrial anti-body was found in 6 (0 9 %) of 997 rheumatoid arth-ritis patients, 1 (1-5%) of 71 Sjogren's syndromepatients, and 3 (6%) of 50 patients with the siccasyndrome.6" The prevalence of mitochondrial anti-body in 0-5 % of a healthy population62 was probablyonly significantly exceeded by the sicca syndromepatients, a condition which is known often to accom-pany chronic liver disease.'A review of liver biopsy findings in rheumatoid

arthritis does not reveal any consistent structuralabnormality, the majority of reports demonstratingonly minor nonspecific changes. 63-67 For example,in an unselected series of 117 patients withrheumatoid arthritis liver histology was normal in35 %, showed nonspecific reactive hepatitis in 43 %,and fatty change in 22%.'4 However, when liverbiopsy was carried out in 31 rheumatoid arthritispatients who had clinical and/or biochemical evi-dence of hepatic dysfunction, 4 (13 %) had definablechronic liver disease, 23 (74%) nonspecific reactivechanges, and only 4 were normal.67 The liver diseasediscovered included one example each of primarybiliary cirrhosis, chronic active hepatitis, alcoholiccirrhosis, and amyloidosis. The only useful markerfor the presence of liver disease in this study was thefinding of a positive mitochondrial antibody.Amyloid deposition is not often specifically soughtbut probably occurs in 15-21 % of rheumatoid arth-ritis cases,6869 though the frequency and distributionmay be identical with an age- and sex-matched con-trol population.70 The only 'specific' histologicallesions seen in the liver in rheumatoid arthritis arenecrotising arteritis, which may accompany extensivevisceral disease7" and nodular regenerative hyper-plasia, which has only once been described in un-complicated rheumatoid arthritis,72 but appears inassociation with Felty's syndrome."Therefore occasional hepatic enlargement, eleva-

tion of serum alkaline phosphatase, and impaired

BSP excretion and nonspecific reactive histologicalchanges in the liver all appear to be merely a featureof the systemic disease process in rheumatoid arth-ritis. The biochemical abnormalities have beenshown to correlate with rheumatoid disease activity,high ESR, low serum albumin, and high serum globu-lin54 5 and seem to be improved by successful anti-inflammatory drug therapy.' Nonspecific reactivehepatitis may represent either the residuum of pre-vious inflammatory intrahepatic disease or a responseto a variety of extrahepatic disease processes includ-ing infections, gastrointestinal tract disorders, col-lagen diseases, or drug reactions.74 In effect thesenonspecific changes represent 'wear and tear' from avery large number of possible causes, and it is notsurprising that it should be a common feature in adisease such as rheumatoid arthritis, which is associ-ated with generalised systemic manifestations.Patients with rheumatoid arthritis are treated withmany potentially hepatotoxic therapeutic agents, butno relationship between alteration in hepatic func-tion and any single drug or combination of drugs hasbeen identified.59 67 Arguments against a drug effectare the lower prevalence of hepatic abnormalitiesseen in patients with seronegative arthritis who takesimilar quantities of the same anti-inflammatorydrugs as the rheumatoid arthritis patients,54 and theapparent improvement in biochemical abnormalities,and occasionally in liver histological changes, inpatients whose disease activity was reduced by drugtherapy, irrespective of the drug used.60

Recent developments of interest in theimmunological field are the demonstration of IgGantibodies in the sera of 31 % of rheumatoid arthritispatients which, when preincubated with normalhuman lymphocytes, caused antibody-dependentlymphocyte cytotoxicity of human Chang liver cells.75Also patients with rheumatoid arthritis and the siccacomplex have been demonstrated to have lympho-cyte sensitisation to salivary antigens which maycross-react with antigenic determinants present in thenormal human biliary tract.7678 Further studies arerequired to determine whether cellular immuneresponses to biliary antigens are to be found inpatients with rheumatoid arthritis.

FELTY SSYNDROMEFelty's syndrome is considered separately because ofits association with nodular regenerative hyperplasiaof the liver, portal hypertension, and haemorrhagefrom oesophageal varices.

Felty's syndrome is defined as rheumatoid arthritisin association with splenomegaly and neutropenia.Abnormal liver function tests were found in 8 out of12 consecutive patients with this syndrome, 5 ofwhom had histological changes in the liver consisting



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

300 Mills, Sturrock

of pronounced sinusoidal lymphocytic infiltration,Kupffer cell prominence, and periportal fibrosis andone of whom had macronodular cirrhosis in addi-tion.79 Nodular regenerative hyperplasia was thendescribed in 5 patients with Felty's syndrome, one ofwhom died from haemorrhage from oesophageal var-ices.73 This is a rare lesion of the liver, as examinationof 51 necropsies on patients with rheumatoid arthritisrevealed no other cases.73 Nodular regenerativehyperplasia of the liver, as originally described bySteiner80 in 1959, is a difficult histological diagnosisto make and may be missed in a small needle biopsy.The surface of the liver may appear normal, finelygranular, or nodular. Histologically there is diffusenodulation of the liver, the size of the nodules usuallyranging from 0 1 to 0-2 cm but occasionally reachingas large as 2-0 cm. The nodules are not outlined bytrue fibrosis but by condensed reticulin fibres giving anegative stain for collagen with Mallory's trichrome,and there is reversed lobulation, so that the hepaticvein radicle is found in the periphery of the nodulewithin compressed columns of hepatocytes.8"The interest in nodular regenerative hyperplasia of

the liver is the association with the development ofportal hypertension. Of 26 cases of nodular regenera-tive hyperplasia so far reported 12 had Felty'ssyndrome and one uncomplicated rheumatoid arth-ritis.8' Five of these rheumatoid patients developedmassive haemorrhage from oesophageal varices,which was fatal in 3 cases. Portal hypertension in thiscondition probably has 2 causes: first, an increase insplenic and hence portal blood flow, and secondly, anincrease in postsinusoidal venous resistance prob-ably caused by compression and distortion ofintrahepatic venous radicles.8" The importance ofbeing aware of nodular regenerative hyperplasia inFelty's syndrome is that after variceal haemorrhagethe optimal surgical treatment should be splenec-tomy, with consideration given to a stunt procedurein addition only if the patient is fit.

JUVENILE CHRONIC ARTHRITIS

Since the original description of chronic joint diseasein children by Still2 in 1897 there has been confusionover terminology and definitions, which has beenpartly resolved recently.82 At a conference in 1977the term chosen was juvenile chronic arthritis, whichincluded subgroups consisting of juvenile ankylosingspondylitis, seropositive juvenile rheumatoid arth-ritis, seronegative chronic arthritis (the main sub-group, still known as juvenile rheumatoid arthritis inthe USA), psoriatic arthritis, and arthropathiesassociated with inflammatory bowel disease.82 Thisdiscussion will be limited to that group of patientswith seronegative chronic arthritis whose clinicalpresentation at onset may be classified into 3 groups:

systemic illness, polyarthritis, and pauciarticulardisease (-4 joints involved).83The systemic disease presentation of seronegative

juvenile chronic arthritis was found in 32 (26%) of124 patients.83 These patients all had high intermit-tent fever of greater than 2 weeks' duration; 30 hadthe typical evanescent pale red macular skin rash, 29(91 %) had hepatosplenomegaly and/or generalisedlymphadenopathy, 12 had pleuritis, 10 pericarditis,29 showed a peripheral blood leucocytosis, and allhad seronegative polyarthritis. Six of these 32patients had evidence of mild hepatic dysfunctiontogether with massive hepatomegaly, 5 ofwhom havebeen reported in detail.84 Liver function tests showedincreased serum SGOT (mean 115 IU/l) in 4,increased serum bilirubin (mean 1 6 mg/100 ml=27,umolll) in 3, and impaired BSP excretion (mean9 5 % retention at 45 minutes) in 3. Liver biopsy in 4patients showed similar mild changes, includingperiportal lymphocyte infiltrates and Kupffer cellhyperplasia. The massive hepatomegaly was there-fore not associated with any great distortion of hepa-tic architecture or cell necrosis and normally regres-sed with remission of other systemic disease mani-festations.Acute hepatic dysfunction, probably due to inter-

current viral hepatitis, has been described in 7 chil-dren with juvenile chronic arthritis.85 86 All the chil-dren had a striking clinical remission from jointsymptoms, fever, and skin rash during the period ofhepatic dysfunction. The remissions were transient infour (< 1 month) but lasted 2 months, 7 months, and5 years in the remaining 3 children. These observa-tions were similar to those of Hench in adultrheumatoid arthritis.5

ADULT STILL'S DISEASEIn 1971 it was first recognised that adults may presentwith an illness similar to seronegative juvenilechronic arthritis with systemic manifestations.87 Thissymptom complex is currently known as adult Still'sdisease, and 58 cases have been reviewed.88Adult Still's disease has the same clinical features

as juvenile chronic arthritis, including intermittentfever, skin rash, seronegative polyarthritis, lym-phadenopathy, splenomegaly, pleuritis, and pericar-ditis.88 Hepatic abnormalities were reported in 3 outof 10 cases, consisting of mild hepatomegaly or mod-est abnormalities of liver function tests.89 Thesehepatic abnormalities could not be attributed to drugtherapy, as they returned to normal shortly after thepatients started anti-inflammatory therapy and reap-peared with relapse of the disease. All 3 patientsshowed periportal infiltration with inflammatorycells on liver biopsy. Overall, 19 (33%) of the 58 casesof adult Still's disease had hepatic abnormalities.88



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


More disturbing preliminary reports have beenpublished concerning acute hepatic failure in adultStill's disease, which was life-threatening in 2 cases9'and fatal in one other.88 However, both of the 2surviving patients were taking high doses of aspirin,and one indomethacin in addition, so drug-inducedhepatotoxicity was also a possible explanation.

SYSTEMIC LUPUS ERYTHEMATOSUSHepatomegaly has been found in 23-39 % of patientswith systemic lupus erythematosus (SLE),9'-9'though clinically significant liver disease had notbeen thought to be common.' Most hepatic interestin SLE had centred around the increased sensitivityto aspirin hepatotoxicity,95 the unconfirmed highchronic carrier rate for HBsAg as a possible reflec-tion of impaired immunological competence,96 97 andthe rare complication of arteritis causing hepaticinfarction and spontaneous rupture of the liver.98 99

Cirrhosis had been reported in 3 (2%) of 138patients9' and hepatic insufficiency was the cause ofdeath in 4 (3 %) out of 135 SLE patients.93 Systematicstudies of hepatic histology in SLE are found in 2papers.92 '0 Of 19 patients who had hepatomegaly orabnormal liver function tests 6 showed normal orinsignificant changes in liver histology; 11 showedminor alterations including fatty change, portal tractfibrosis, and mild to moderate cellular infiltration;and 2 had chronic active hepatitis which had pro-gressed to cirrhosis.92 Of 18 unselected SLE patients5 had normal liver histology and 13 showed the sameminor changes described above."

In 1980 Runyon et al. reported on a retrospectivereview of the spectrum of liver disease in 206 SLEpatients."'0 Clinical changes were as follows:hepatomegaly (39%), splenomegaly (6%), andjaundice (24%), and 43 (21 %) of the patients weredefined as having liver disease on the basis of abnor-mal liver histology or the repeated 2-fold or greaterincrease in 2 or more liver function tests. All patientswere HBsAg negative, and aspirin was being takenmore frequently by patients without liver disease.Liver histology was available in 33 patients andincluded the following findings: steatosis (12 cases),cirrhosis (4), chronic active hepatitis (4), primarybiliary cirrhosis (1), hepatic granulomas (3), chronicpersistent hepatitis (2), haemochromatosis (1), cen-trilobular necrosis (3), microabscesses (2), choles-tasis (1), and nonspecific reactive changes (1). There-fore 9 (4 *4 %) of the SLE patients had serious chronicliver disease and 3 had died from hepatic failure atthe time of review. Three possible conclusions can bedrawn: that SLE and the chronic liver disease werecoincidental, that chronic liver disease is a manifesta-tion of SLE, or that chronic active hepatitis may beassociated with systemic manifestations suggesting a

diagnosis of SLE. The latter conclusion seems mostsatisfactory, as 20% of chronic active hepatitispatients may have features which would fulfil theAmerican Rheumatism Association's criteria for thediagnosis of SLE.'02 Patients with chronic activehepatitis and features of SLE should probably beconsidered as having both diseases, so that the liverlesion will be treated appropriately while careful vig-ilance is maintained for other life-threatening mani-festations of SLE.

SYSTEMIC SCLEROSISLiver disease appears to be uncommon in systemicsclerosis, with only 8 (1.1 %) instances of hepaticinvolvement reported in a review of 727 patients.24Moreover, in a retrospective review of necropsy find-ings in systemic sclerosis hepatomegaly and cirrhosiswere both more common in a matched controlgroup.103

However, in a prospective assessment of the extentof visceral involvement in systemic sclerosis 16(52%) of 31 patients were shown to have abnormalliver function tests or lengthened prothrombintimes.'04 There is an association between systemicsclerosis and primary biliary cirrhosis, in particularwhen all the features of the CRST syndrome arepresent.2' 26 Portal hypertension, leading to haemor-rhage from oesophageal varices, has been describedin 4 patients, 3 of whom had cirrhosis.'0' Extrahepa-tic biliary disease has been recorded in systemicsclerosis consisting of fibrosis of the gall-bladder'06and hepatic duct obstruction caused by ulcerationassociated with vasculitis.'07

POLYMYALGIA RHEUMATICASerum alkaline phosphatase of hepatic origin is ele-vated in 20-62 % of patients with polymyalgiarheumatica,'08110 when it parallels disease activitytogether with the erythrocyte sedimentation rate andreturns to normal with steroid treatment.'08 Serumtransaminases and BSP excretion may also be mildlyabnormal, but jaundice is unusual."' A limitednumber of reports of liver histology have appeared,but the features described include: normal liver (2cases), mild fatty change (2),"' hepatic necrosis(1),"' chronic persistent hepatitis (1),"' and hepaticgranulomas (2).".. ".. An association with primarybiliary cirrhosis has also been recorded.3'An initial report of a high prevalence of serum

anti-HBs in polymyalgia rheumatica"6 has not beensubsequently confirmed...7 ... The prevalence ofserum HBsAg and anti-HBs did not differ frommatched control populations.

POLYARTERITIS NODOSAPolyarteritis nodosa is a multisystem disease thataffects almost all organs in the body. It is included in



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

302 Mills, Sturrock

this review because arthralgia or arthritis are domin-ant features of the disease and because of the associa-tion with the hepatitis B virus and accompanying liverdisease.

Abnormalities in liver function tests are commonin polyarteritis and the presentation may often mis-leadingly suggest the diagnosis of primary liver dis-ease."19 Since the initial description in 1970 by Gockeet al. 120 the association between the hepatitis B virusand polyarteritis has become firmly established.121Serum HBsAg has been found in 45-71 % of cases ofpolyarteritis recently described,121-122 but a slightlylower frequency is the general UK experience. It isclear that exposure to the hepatitis B virus may pro-duce 2 distinct connective tissue syndromes. Trans-ient polyarthritis may precede the onset of acute virusB hepatitis as has already been described. However, asubacute or chronic illness featuring arthralgia/arth-ritis, mononeuritis multiplex, fever, renal disease,hypertension, and central nervous system diseasemay also occur, and this illness does not differ inclinical presentation or target organ involvementfrom polyarteritis not associated with the hepatitis Bvirus...2 124 The only differences noted betweenHBsAg positive polyarteritis and those without evi-dence of hepatitis B virus infection are that theformer are more frequently associated with liverinvolvement, usually have a good response to cor-ticosteroid therapy, and may have a better prog-nosis."2 However, the relationship between HBsAgpositive polyarteritis and liver disease is very vari-able. In a series of 27 cases of HBsAg positive polyar-teritis 8 had acute hepatitis, 10 chronic hepatitis, 3cirrhosis, and in 6 there was no hepatic abnormal-ity. 125

Other hepatic abnormalities which may occur inpolyarteritis are the formation of multiple hepaticartery aneurysms and hepatic infarction due toarteritis. Polyarteritis nodosa is by definition associ-ated with the formation of nodular aneurysms inmedium-sized arteries, so that renal or hepaticarteriography may be the simplest way to achieve aclear diagnosis of polyarteritis.123 126 Liver biopsy car-ries a risk of aneurysm puncture and haemorrhageand is therefore best avoided. Polyarteritis is thecommonest single cause of hepatic infarction whichmay lead to jaundice or intraperitoneal haemor-rhage. 127

(C) Antirheumatic drug-induced liver diseaseThe range of antirheumatic drugs currently in use isextensive, but fortunately hepatotoxicity is a rareoccurrence. Despite a succession of eleganthypotheses the mode of action of these antirheumaticdrugs remains unknown,128 and any hepatotoxicity isusually not predictable or dose-related. However, in

view of the frequency of hepatic involvement in therheumatic diseases and the danger of missing a drugtoxicity reaction it was considered important brieflyto review accumulated experience of those anti-rheumatic drugs in which hepatotoxicity has beenrecorded.The following major antirheumatic drugs will be

reviewed:First-line drugs. (i) Aspirin and related drugs: aspi-

rin, diflunisal. (ii) Indomethacin and related drugs:indomethacin, sulindac. (iii) Propionic acid deriva-tives: ibuprofen, naproxen,. fenoprofen, ketoprofen.(iv) Pyrazolones: phenylbutazone.

Second-line drugs. (i) Gold. (ii) Penicillamine.Third-line drugs. (i) Azathioprine.Aspirin. This was first recognised to cause abnor-

malities of liver function in patients with acuterheumatic fever in 1956.129 Since then aspirin hasbeen frequently reported to produce elevation ofserum transaminases in patients with rheumatoidarthritis,95 systemic lupus erythematosus,95 juvenilechronic arthritis,130 131 and Reiter's syndrome132 butnot apparently in normal individuals.95 133

In a prospective study aspirin given in sufficientdose to produce a serum salicylate level of 25-30mg/100 ml (1 -81-2-17 mmol/l) caused an increase inserum transaminases in 4 of 20 rheumatoid arthritispatients and 7 of 16 systemic lupus erythematosuspatients after 2 weeks of therapy.95 Serum trans-aminases were elevated considerably higher in thepatients with systemic lupus erythematosus butreturned to normal in all patients when therapy wasstopped. These enzyme abnormalities are rarelyassociated with symptoms of liver disease,hepatomegaly, alteration in serum bilirubin orprothrombin time, or evidence of severe hepaticdysfunction.131 While enzyme abnormalities aremore common on high-dose aspirin, when serumsalicylate levels exceed 25 mg/100 ml (1 -81 mmolI),they also occur with lower-dose therapy.130 Liverbiopsy reveals a nonspecific hepatitis with mild focalhepatocyte changes including necrosis, ballooning ofcells, and lobular infiltrates with mononuclearcells.134 These abnormalities are all rapidly reversibleafter stopping therapy, and there are no firmlydocumented examples of fatality or progression tochronic liver disease.More recently there have been 3 reports of aspirin

hepatotoxicity associated with short-lastingencephalopathy in 5 children.135-137 These childrenwere given high-dose aspirin and developed highserum transaminases, occasional slight elevation ofserum bilirubin and prolongation of prothrombintime, altered level of consciousness, and electro-encephalographic (EEG) changes of a metabolicencephalopathy, all of which rapidly improved when



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


aspirin was stopped. However, the evidence for acutehepatic failure is limited as the EEG changes arenonspecific, and hypoprothrombinaemia may beinduced by salicyclic acid alone."38 Acute salicylatepoisoning with associated mild hepatitis is an alterna-tive explanation for these clinical changes.

Diflunisal. A single example of reversible mildcholestatic jaundice has been reported with thisdrug.139

Indomethacin. There is little evidence of hepato-toxicity of indomethacin in the literature despite itswide usage.'40 A 12-year-old boy developed fulmin-ant and fatal hepatitis while on indomethacin 100 mgdaily, which he had taken for 6 months,.4' Biliver-dinaemia, biliverdinuria, and a toxic hepatitisappeared after taking 75 mg indomethacin daily for 3weeks.'42 The patient presented with a greenish hueto his body, and his urine was green. Recoveryensued, but it remained possible that he had a partialdefect in biliverdin reductase which had been madesymptomatic by indomethacin hepatitis.

Sulindac. Three reports of a toxic hepatitis due tosulindac therapy exist.'43- 4' The 3 patients survived,but 2 had a severe hepatitis which came on soon afterstarting sulindac. In the third case mild hepatitis wasconfirmed on rechallenge.Ibuprofen. A closely allied compound, ibufenac,

which is a precursor of ibuprofen, was associated withsevere hepatotoxicity and has been withdrawn."'Ibuprofen appears to be a markedly safer drug withno hepatotoxicity recorded in one series.'47 However,there have been 3 isolated examples of hepatotoxic-ity, in 2 cases reversible hepatitis'48 149 and in the thirdmassive and fatal fatty change associated with bilat-eral pleural effusions."'Naproxen. Three examples of transient toxic

hepatitis causing jaundice are reported with nap-roxen."'"'l53 All patients developed jaundice withinone to 6 weeks of starting therapy. None were rechal-lenged, and only one had a liver biopsy.Fenoprofen. A single case of a jaundiced illness

with high serum transaminases followed the use offenoprofen for 7 weeks.'54 The toxic hepatitis wasmild and settled quickly. No liver biopsy wasobtained and no challenge carried out.

Ketoprofen. No record of clinical hepatotoxicityhas been found.Phenylbutazone. This is a toxic agent and has been

recognised to cause toxic hepatitis, with severalfatalities, and the formation of hepatic granulomas.In a series of 800 patients receiving phenylbutazone 2developed a toxic hepatitis, given an incidence of0-25 %."' The first detailed report on phenyl-butazone hepatotoxicity described 5 patients inwhom jaundice appeared 6-40 days after the start oftherapy."6 There have subsequently been many

reports of toxic hepatitis, including 8 fatalities."7Hepatic granulomas have been recorded in 11patients in association with phenylbutazone therapy,the illness being accompanied by fever, pruritus,jaundice, and occasionally eosinophilia."7

Gold. So-called 'gold hepatitis' has disappeared asa clinical problem since the introduction of dispos-able syringes. However, gold therapy has now beenreported to cause intrahepatic cholestasis in 5patients.".8 ".. The jaundice was marked and accom-panied by eosinophilia in 2 cases. Cholestasis is anearly feature of gold toxicity, appearing after a meanof only 110 mg of gold had been given. All patientsrecovered, although jaundice lasted 2 months in onecase.

Penicillamine. In a series of 99 rheumatoid arthritispatients receiving penicillamine therapy 6 developedelevation of serum transaminases and 2 had evidenceof toxic hepatitis on liver biopsy.'" These abnor-malities reverted to normal in 5 patients whenpenicillamine was stopped. Penicillamine has beenreported to cause fatal cholestatic jaundice in associ-ation with renal failure'6' and 2 examples of revers-ible toxic hepatitis.'62

Azathioprine. Azathioprine was found to causemarked jaundice in 2 patients with no change inserum alkaline phosphatase and either normal orslightly elevated serum transaminases.'63 " Liverbiopsy in one case showed evidence of cholestasis butno necrosis. The other patient was challenged on 2occasions with azathioprine and repeatedly showedhypersensitivity reactions, even to a very low dose of5 mg. It was postulated that azathioprine mightinterfere with bilirubin excretion rather than causehepatic damage.'"

ConclusionIt is apparent that joint symptoms are frequent andimportant markers of the systemic nature of manyliver diseases. In particular, the occurrence of jointsymptoms in patients with haemochromatosis,chronic active hepatitis, or virus hepatitis may be ofconsiderable diagnostic value. Conversely, there islittle evidence of serious hepatic disease in patientswith rheumatic conditions where minor changes pre-dominate and tend to reflect systemic disease activ-ity. Awareness of the significance of these minorhepatic abnormalities will often prevent unnecessaryinvestigation and the occasional unwarranted opera-tion. However, if patients with rheumatic conditionshave persistent signs of hepatic dysfunction or a posi-tive mitochondrial antibody, liver biopsy will reveala variety of subclinical liver diseases.67 " As yetthere is little evidence that treatment of underlyingliver disease prevents the progression of the associ-ated joint changes.



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

304 Mills, Sturrock.

Considerable overlap between hepatic andrheumatic diseases may occur, as exemplified by theoccasional difficulty in distinguishing betweensystemic lupus erythematosus and chronic activehepatitis. In such circumstances it would seem advis-able to manage the patients as though they had bothconditions. The association between the hepatitis Bvirus and polyarteritis nodosa is so far the only estab-lished example of a chronic rheumatic disorder whichmay be caused by a viral infection but offers somehope that in the future a viral aetiology may be foundto underlie more common disorders such asrheumatoid arthritis.l"s

Antirheumatic drugs, with the exception ofphenylbutazone, have a low level of reported hepato-toxicity and probably account for few of the minorhepatic abnormalities seen in the rheumatic diseases.However, with an increasing number of new agentsbecoming available, vigilance in this field must bemaintained.

References

1 Whelton M J. Arthropathy and liver disease. Br J Hosp Med1970; 3: 243-7.

2 Still G F. On a form of chronic joint disease in children. MedChir Trans 1897; 80: 47-60.Wishart J. A link between rheumatoid arthritis and jaundice.Br Med J 1903; i: 252.

4 Hench P S. Effect of jaundice on chronic infectious (atrophic)arthritis and on primary fibrositis. Further observations:attempts to reproduce the phenomenon. Arch Intern Med1938; 61: 451-80.Hench P S. Effect of spontaneous jaundice on rheumatoid(atrophic) arthritis. Attempts to reproduce the phenomenon.Br MedJ 1938; ii: 394-8.

6 Thompson H E, Wyatt B L. Experimentally induced jaundice(hyperbilirubinaemia). Report of animal experimentation andof the physiologic effect of jaundice in patients with atrophicarthritis. Arch Intern Med 1938; 61: 481-500.

7Bruusgaard A, Andersen R B. Effect of an intravenouslyadministered bile acid (chenodeoxycholic acid) on rheumatoidarthritis. Scand J Rheumatol 1975; 4: 169-73.Gardner F, Stewart A, MacCallum F 0. Therapeutic effect ofinduced jaundice in rheumatoid arthritis. Br Med J 1945; u:677-80.Rennie J B, Fraser T N. Transmission of acute infectivehepatitis. Therapeutic effect on rheumatoid arthritis. GlasgowMed J 1946; 27: 157-66.Hench P S. Ameliorating effect of pregnancy on chronicatrophic (infectious) rheumatoid arthritis, fibrositis and inter-mittent hydrarthrosis. Mayo Clin Proc 1938; 13: 161-7.Hench P S, Kendall E C, Slocumb C H, Polley H F. The effect ofa hormone of the adrenal cortex (17-hydroxy-11-dehydro-corticosterone: compound E) and of pituitary adrenocortico-trophic hormone on rheumatoid arthritis: preliminary reportAnn Rheum Dis 1949; 8: 97-104.

12 Peterson R S. Adrenocortical steroid metabolism and adrenal-cortical function in liver disease.J Clin Invest 1960; 39: 320-31.

13 Pinals R S. Effect of hepatic injury on adjuvant arthritis. AnnRheum Dis 1973; 32: 471-4.

4 Alarcon-Segovia D, Diaz-Jouanen E, Fishbein E. Features ofSjogren's syndrome in primary biliary cirrhosis. Ann IntemMed 1973; 79: 31-6.

15 Golding P L, Smith M, Williams R. Multisystem involvement inchronic liver disease. Studies on the incidence andpathogenesis. Am J Med 1973; 55: 772-82.Epstein 0, Thomas H C, Sherlock S. Primary biliary cirrhosis isa dry gland syndrome with features of chronic graft-ver-sus-host disease. Liocet 1980; i: 1166-8.

17 Sherlock S, Scheuer P J. The presentation and diagnosis of 100patients with primary biliary cirrhosis.N EnglJ Med 1973; 289:674-8.

16 Mills P R, MacSween R N M, Watkinson G. Arthritis andprimary biliary cirrhosis. Br Med J 1977; ii: 1224.

19 James 0, Macklon A F, Watson A J. Primary biliary cir-rhosis-a revised clinical spectrum. Lancet 1981; i: 1278-81.

2 Marx W J, O'Connell D J. Arthritis of primary biliary cirrhosis.Arch Intern Med 1979; 139: 213-6.

21 Mills P R, Vallance R, Birnie G, et al. A prospective survey ofradiological bone and joint changes in primary biliary cirrhosis.Clin Radiol 1981; 32: 297-302.

22 Crowe J P, Molloy M G, Welis I, et al. Increased Clq bindingand arthritis in primary biliary cirrhosis. Gut 1980; 21: 418-22.

2 Golding P L, Bown R. Mason AM S, Taylor E. 'Sicca complex'in liver disease. Br Med J 1970; ii: 340-2.

"Bartholomew L G, Cain J C, Winkelmann R K, Baggenstoss AH. Chronic disease of the liver with systemic scleroderma. Am JDig Dis 1964; 9: 43-55.Murray-Lyon I M, Thompson R P H, Ansell I D, Williams R.Scleroderma and primary biliary cirrhosis. Br Med J 1970; iii:258-9.

26 Reynolds T B, Denison E K, Frankl H D, Lieberman F L,Peters R L. Primary biliary cirrhosis with scleroderma,Raynaud's phenomenon and telangiectasia. New syndrome.Am J Med 1971; 50: 302-12.

27 Clarke A K, Galbraith R M, Hamilton E B, Williams R.Rheumatic disorders in primary biliary cirrhosis. Ann RheumDis 1978; 37: 42-7.

"Ansell B M, Bywaters E G L. Histiocytic bone and joint dis-ease. Ann Rheum Dis 1957; 16: 503-10.

29 Rooney P J, Third J, Madkour M M, Spencer D, Dick W C.Transient polyarthritis associated with familial hyperbeta-lipoproteinaemia. Q J Med 1978; 47: 249-59.NMills P R, Rooney P J, Watkinson G, MacSween R N M.Hypercholesterolaemic arthropathy in primary biliary cir-rhosis. Ann Rheum Dis 1979; 38: 179-80.

31 Robertson J C, Batstone G F, Loebl W Y. Polymyalgiarheumatica and primary biliary cirrhosis. Br Med J 1978; ii:1128.

32 Ogilvie A L, Toghill A J. Polymyalgia rheumatica and primarybiliary cirrhosis. Br Med J 1978; ii: 1501.

33EpsteinO, AjdukiewiczA B, Dick R, Sherlock S. Hypertrophichepatic osteoarthropathy. Clinical, roentgenologic, bio-chemical, hormonal and cardiorespiratory studies, and a reviewof the literature. Am J Med 1979; 67: 88-97.Epstein 0, Dick R, Sherlock S. Prospective study of periostitisand finger clubbing in primary biliary cirrhosis and other formsof chronic liver disease. Gut 1981; 22: 203-6.

3S Sherlock S. Chronic hepatitis. Gut 1974; 15: 581-97.36 Read A E, Sherlock S, Harrison C V. Acute juvenile cirrhosis

considered as part of a systemic disease and the effect ofcorticosteroid therapy. Gut 1963; 4: 378-93.

37 Mistilis S P, Blackburn C R B. Active chronic hepatitis. Am JMed 1970; 48: 484-95.

38 Barnardo D E, Vernon-Roberts B, Currey H L. A case of activechronic hepatitis with painless erosive arthritis. Gut 1973; 14:800-4.

39 Smith J W, Sanford J P. Viral arthritis. Ann Intern Med 1967;67: 651-9.

40 Fernandez R, McCarty D J. The arthritis of viral hepatitis. AnnIntern Med 1971; 74: 207-11.

41 Levo Y. Extrahepatic manifestations of hepatitis B virus infec-tion. Isr J Med Sci 1979; 15: 276-82.



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


42 Shumaker J B, Goldfinger S E, Alpert E, Isselbacher K J.Arthritis and rash. Clues to anicteric viral hepatitis. Arch InternMed 1974; 133: 483-5.

4 Stewart J S, Farrow L J, Clifford R E, et al. A three-year surveyof viral hepatitis in West London. Q J Med 1978; 47: 365-84.

4Alpert E, Isselbacher K J, Schur P H. The pathogenesis ofarthritis associated with viral hepatitis. Complement-compo-nent studies. N Engl J Med 1971; 285: 185-9.

4S Wands J R, Mann E, Alpert E, Isselbacher K J. Thepathogenesis of arthritis associated with acute hepatitis-B sur-face antigen-positive hepatitis. Complement activation andcharacterisation of circulating immune complexes.J Clin Invest1975; 55: 930-6.

46 Wands J R, Alpert E, Isselbacher K J. Arthritis associated withchronic active hepatitis: complement activation and character-isation of circulating immune complexes. Gastroenterology1975; 69: 1286-91.

47 Schumacher H R, Gail E P. Arthritis in acute hepatitis andchronic active hepatitis. Pathology of the synovial membranewith evidence for the presence of Australia antigen in synovialmembranes. Am J Med 1974; 57: 655-64.

4Hamilton E B D. The arthritis of hyperparathyroidism, haemo-chromatosis and Wilson's disease. In: Bywaters E G L, ed.Clinics in Rheumatology. London: Saunders, 1975: 1: 109-23.

49 Hamilton E B D. Other metabolic arthropathies. In: Scott J T,ed. Copeman's Textbook of the Rheumatic Diseases. 5th ed.London: Churchill Livingstone, 1978: 692-706.

5 M'Seffar A, Fornasier V L, Fox I H. Arthropathy as the majorclinical indicator of occult iron storage disease. JAMA 1977;238: 1825-8.

51 Golding D N, Walshe J M. Arthropathy of Wilson's disease.Study of clinical and radiological features in 32 patients. AnnRheum Dis 1977; 36: 99-111.

5 Bakos S N. Polyarthritis associated with Wilson's disease. ArchDis Child 1979; 54: 985.Canelas H M, Carvalho N, Scaff M, Vitule A, Barbosa E R,Azevedo E M. Osteoarthropathy ofhepatolenticulardegenera-tion. Acta Neurol Scand 1978; 57: 481-7.

u Webb J, Whaley K, MacSween R N M, Nuki G, Dick W C,Buchanan W W. Liver disease in rheumatoid arthritis andSjogren's syndrome. Prospective study using biochemical andserological markers of hepatic dysfunction. Ann Rheum Dis1975; 34: 70-81.Tiger L H, Gordon M H, Ehrlich G E, Shapiro B. Liverenlargement demonstrated by scintigraphy in rheumatoid arth-ritis. J Rheumatol 1976; 3: 15-20.

56 Kendall M J, Cockel R, Becker J, Hawkins C F. Rheumatoidliver? Br Med J 1970; i: 221.

57 Kendall M J, Cockel R, Becker J, Hawkins C F. Raised serumalkaline phosphatase in rheumatoid disease. An index of liverdysfunction? Ann Rheum Dis 1970; 29: 537-40.

5 Cockel R, Kendall M J, Becker J, Hawkins C F. Serum bio-chemical values in rheumatoid disease. Ann Rheum Dis 1971;30: 166-70.

S Fernandes L, Sullivan S, McFarlane I G, et al. Studies on thefrequency and pathogenesis of liver involvement in rheumatoidarthritis. Ann Rheum Dis 1979; 38: 501-6.

"Rau R. The liver in rheumatoid arthritis. In: Eberl R,Rosenthal M, eds. Organic Manifestations and Complications inRheumatoid Arthritis. Stuttgart: Schatteuer, 1976; 155-69.

6 Whaley K, Goudie R B, Williamson J, Nuki G, Dick W C,Buchanan W W. Liver disease in Sjogren's syndrome andrheumatoid arthritis. Lancet 1970; i: 861-3.

6 Bullock J Y, Booth R J, Wilson J D. Tissue antibodies in ahealthy New Zealand population. N Z Med J 1979; 89: 11-3.

"Movitt E R, Davis A E. Liver biopsy in rheumatoid arthritis.Am J Med Sci 1953; 226: 516-20.Rau R, Pfenninger K, Boni A. Proceedings: Liver function testsand liver biopsies in patients with rheumatoid arthritis. AnnRheum Dis 1975; 34: 198-9.

"Otto W, Klugmann H J, Geiler G. Supplementary comment:liver affection in rheumatoid arthritis. In: Eberl R, RosenthalM, eds. Organic Manifestations and Complications inRheumatoid Arthritis. Stuttgart: Schatteuer, 1976: 171-3.

SDietrichson 0, From A, Christofferson P, Juhl E. Morpho-logical changes in liver biopsies from patients with rheumatoidarthritis. Scand J Rheumatol 1976; 5: 65-9.

7Mills P R, MacSween R N M, Dick W C, More I A, Watkin-son G. Liver disease in rheumatoid arthritis. Scott Med J 1980;25: 18-22.

"Fingerman D L, Andrus F C. Visceral lesions associated withrheumatoid arthritis. Ann Rheum Dis 1943; 3: 168-81.

9 Lender M, Wolf E. Incidence of amyloidosis in rheumatoidarthritis. Some clinical observations. Scand J Rheumatol 1972;1: 109-12.

70 Ozdemir A J, Wright J R, Calkins E. Influence of rheumatoidarthritis on amyloidosis of aging. Comparison of 47 rheumatoidpatients with 47 controls matched for age and sex.N EnglJ Med1971; 285: 534-8.Sinclair R J G, Cruikshank B. A clinical and pathological studyof sixteen cases of rheumatoid arthritis with extensive visceralinvolvement (rheumatoid disease) Q J Med 1956; 25: 313-32.

7 Harris M, Rash R M, Dymock I W. Nodular, non-cirrhotic liverassociated with portal hypertension in a patient withrheumatoid arthritis. J Clin Pathol 1974; 27: 963-6.

73 Blendis L M, Parkinson M C, Shilkin K B, Williams R. Nodularregenerative hyperplasia of the liver in Felty's syndrome. Q JMed 1974; 43: 25-32.

7 Popper H, Schaffner F. Liver: Structure and Function. NewYork: McGraw-Hill, 1957.

7 Panayi G S. Antibody-mediated leucocyte cytotoxicity toChang human liver cells in rheumatoid arthritis and other dis-eases. Ann Rheum Dis 1976; 35: 27-31.

76 McFarlane I G, Wojcicke B M, Tsantoulas D C, et al. Cellularimmune responses to salivary glands in autoimmune liver dis-ease with sicca syndrome. Clin Exp Immunol 1976; 25: 389-95.

7 Sullivan S, Fernandes L, McFarlane I G, et al. Impairment oflachrymaland salivary secretion and cellular immune responsesto salivary antigens in rheumatoid arthritis.,Ann Rheum Dis1978; 37: 164-7.Wojcicka B M. Separation and characterisation of bile anti-gens. In: Eddleston A L W F, Williams R, eds. Immune Reac-tions in Liver Disease. London: Pitman, 1979: 154-6.

7 Blendis L M, Ansell I D, Jones K L, Hamilton E, Williams R.Liver in Felty's syndrome. Br Med J 1970; i: 131-5.

8Steiner P E. Nodular regenerative hyperplasia. Am J Pathol1959; 35: 943-51.

8 Blendis L M, Lovell D, Barnes C G, Ritland S, Cattan D, VesinP. Oesophageal variceal bleeding in Felty's syndrome associ-ated with nodular regenerative hyperplasia. Ann Rheum Dis1978; 37: 183-6.

82 Ansell B M. Chronic arthritis in childhood. Ann Rheum Dis1978; 37: 107-20.

83 Schaller J, Wedgwood R J. Juvenile rheumatoid arthritis: areview. Pediatrics 1972; 50: 940-53.

8Schaller J, Beckwith B, Wedgwood R J. Hepatic involvement injuvenile rheumatoid arthritis. J Pediatr 1970; 77: 203-10.

8 Kornreich H, Malouf N N, Hanson V. Acute hepatic dys-function in juvenile rheumatoid arthritis. J Pediatr 1971; 79:27-35.

8 Schaller J. The liver and arthritis.J Pediatr 1971; 79: 139-41.8 Bywaters E G L. Still's disease in the adult. Ann Rheum Dis

1971; 30: 121-33.8Esdaile J M, Tannenbaum H, Hawkins D. Adult's Still's

disease. Am J Med 1980; 68: 825-30.69 Bujak J S, Aptekar R G, Decker J L, Wolff S M. Juvenile

rheumatoid arthritis presenting in the adult as fever ofunknown origin. Medicine 1973; 52: 431-44.

sBaker D G, Schumacher H R, Reginato A J. Fifteen patients



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

306 Mills, Sturrock

with Adult Still's disease: life threatening liver failure in two.Arthritis Rheum 1979; 22: 590.

91 Harvey A McG, Shulman L E, Tumulty P A, Conley C L,Schoenrich E H. Systemic lupus erythematosus: review of theliterature and clinical analysis of 138 cases. Medicine 1954; 33:291-437.

92 MacKay I R, Taft L I, Cowling D C. Lupoid hepatitis and thehepatic lesions of systemic lupus erythematosus. Lancet 1959;i: 65-9.

9 Dubois E L, Tuffanelli D L. Clinical manifestations of systemiclupus erythematosus. Computor analysis of 520 cases. JAMA1964; 190: 104-11.

9 Decker J L, Steinberg A D, GershwinM E, et al. Systemic lupuserythematosus-contrasts and comparisons. Ann Intern Med1975; 82: 391-404.

9 Seaman W E, Plotz P H. Effect of aspirin on liver tests inpatients with RA or SLE and in normal volunteers. ArthritisRheum 1976; 19: 155-60.

" Alarcon-Segovia D, Fishbein E, Diaz-Jouanen E. Presence ofhepatitis-associated antigen in systemic lupus erythematosus.Clin Exp Immunol 1972; 12: 9-19.Shorey J, Cheatum D E, Combes B, Ziff M. Hepatitis B antigenand antibody in patients with systemic lupus erythematosus.Arthritis Rheum 1974; 17: 583-9.

9 Haslock I. Spontaneous rupture of the liver in systemic lupuserythematosus. Ann Rheum Dis 1974; 33: 482-4.

9 Levitin P M, Sweet D, Brunner C M, Katholi R E, Botton W K.Spontaneous rupture of the liver. An unusual complication ofSLE: Arthritis Rheum 1977; 20: 748-50.

'°° Chwalinska-Sadowska H, Milewski B, Nazarewicz T. Clinicaland immunomorphological evaluation of pathological changesin the liver in collagenoses. Mater Med Pol 1976; 8: 421-8.Runyon B A, LaBreque D R, Anuras S. The spectrum of liverdisease in systemic lupus erythematosus. Report of 33histologically-proved cases and review of the literature. Am JMed 1980; 69: 187-94.

02 Chwalinska-Sadowska H, Milewski B, Maldyk H. Diagnostictroubles connected with differentiation of systemic lupuserythematosus against chronic hepatitis. Mater Med Pol 1977;9: 60-4.

03 D'Angelo W A, Fries J F, Masi A T, Shulman L E. Pathologicobservations in systemic sclerosis (scleroderma). A study offifty-eight autopsy cases and fifty-eight matched controls. Am JMed 1969; 46: 428-40.

104 Barnett A J, Coventry D A. Scleroderma. 2. Incidence ofsystemic disturbance and assessment of possible aetiologicalfactors. Med J Aust 1969; i: 1040-7.Morris J S, Htut T, Read A E. Scleroderma and portal hyper-tension. Ann Rheum Dis 1972; 31: 316-8.

106 Copeman P W M, Medd W E. Diffuse systemic sclerosis withabnormal liver and gall-bladder. Br Med J 1967; Hi: 353-4.

107 Wildenthal K, Schenker S, Smiley J D, Ford K L. Obstructivejaundice and gastrointestinal haemorrhage in progressivesystemic sclerosis. Arch Intern Med 1968; 121: 365-8.

100 Glick E N. Raised serum-alkaline phosphatase levels inpolymyalgia rheumatica. Lancet 1972; ii: 328.

109 Gibbs P. Polymyalgia rheumatica and liver disease. Lancet1974; i: 351-2.

110 von Knorring J, Wasastjerna C. Liver involvement inpolymyalgia rheumatica. ScandJ Rheumatol 1976; 5: 179-204.Dickson E R, Maldonado J E, Sheps S G, Cain J A. Systemicgiant-cell arteritis with polymyalgia rheumatica. Reversibleabnormalities of liver function. JAMA 1973; 224: 1496-8.

112 LeongA S-Y, Alp M H. Hepatocellular disease in the giant cellarteritis/polymyalgia rheumatica syndrome. Ann Rheum Dis1981; 40: 92-5.

113 Thompson K, Roberts P F. Chronic hepatitis in polymyalgiarheumatica. Postgrad Med J 1976; 52: 236-8.

114 Long R, James 0. Polymyalgia rheumatica and liver disease.Lancet 1974; i: 77-9

11 Litwack K D, Bohan A, Silverman L. Granulomatous liverdisease and giant cell arteritis. Case report and literaturereview. J Rheumatol 1977; 4: 307-12.

116 Bacon P A, Docherty S M, Zuckerman A J. Hepatitis B anti-body inpolymyalgia rheumatica. Lancet 1975; ii: 476-8.Bridgeford P H, Lowenstein M, Bocanegra T S, Vasey F B,Germain B F, Espinoza L R. Polymyalgia rheumatica and giantcell arteritis: histocompatibility typing and hepatitis-B infectionstudies. Arthritis Rheum 1980; 23: 516-8.Elling H, Skinhoj P, Elling P. Hepatitis B virus and polymyalgiarheumatica: a search forHBsAg,HBsAb, HBcAb, HBeAg, andHBeAb.Ann Rheum Dis 1980; 39: 511-3.

119 Mowrey F H, Lundberg E A. The clinical manifestations ofessential polyangitis (polyarteritis nodosa) with emphasis onthe hepatic manifestations.Ann Intern Med 1954; 40: 1145-64.

139 Gocke D J, Hsu K, Morgan C, Bombardieri S, Lockshin M,Christian C L. Association between polyarteritis and Australiaantigen. Lancet 1970; ii: 1149-53.

121 Trepo C G, Zuckerman A J, Bird R C, Prince A M. The role ofcirculating hepatitis B antigen/antibody immune complexes inthe pathogenesis of vascular and hepatic manifestations inpolyarteritis nodosa. J Clin Pathol 1974; 27: 863-8.

122 Sergent J S, LockshinM D, Christian C L, Gocke D J. Vasculitiswith hepatitis B antigenaemia. Medicine 1976; 55: 1-18.

123 Vertzman L. Polyarteritis nodosa. In: Alarcon-Segovia D, ed.Clinics in Rheumatic Diseases. London: Saunders, 1980: 6:297-317.

124 Duffy J, Lidsky M D, Sharp J T, et al. Polyarthritis, polyarteritisand hepatitis B. Medicine 1976; 55: 19-37.

125 Godeau P, Trepo G, Herreman G, et al. Nriarterite noueuse etantigene Australie. Etude comparative a propos de 54 cas.Sem Hop Paris 1977; 53: 613-9.

126 Bron K M, Gajaraj A. Demonstration of hepatic aneurysms inpolyarteritis nodosa by arteriography.N EnglJ Med 1970; 282:1024-5.

127 Kanter D M. Hepatic infarction. Arch Intern Med 1965; 115:479-81.

126 Rooney P J, Lee P, Brooks P, Buchanan W W. Reflectionsupon possible mechanisms of action of anti-inflammatorydrugs. Curr Med Res Opin 1973; 1: 501-16.

129 Manso C, Taranta A, Nydick I. Effect of aspirin administrationon serum glutamic oxaloacetic and glutamic pyruvic transamin-ases in children. Proc Soc Exp Biol Med 1956; 93: 84-8.

120 Bernstein B N, Singsen B H, King K K, Hanson V. Aspirin-induced hepatotoxicity and its effect on juvenile rheumatoidarthritis. Am J Dis Child 1977; 131: 659-63.

131 Schalier J G. Chronic salicylate administration in juvenilerheumatoid arthritis: aspirin 'hepatitis' and its clinical signifi-cance. Pediatrics 1978; 62 (5 pt 2 suppl): 916-25.

132 RicksW B. Salicylate hepatotoxicity in Reiter's syndrome.AnnIntern Med 1976; 84: 52.

132 Furst D E, Kar N C, Sarkission E S, Gupta N, Paulus H E.Effects of salicylate on liver enzymes in normal young adults.Arthritis Rheum 1976; 19: 267-8.

134 O'Gorman T, Koff R S. Salicylate hepatitis. Gastroenterology1977; 72: 726-8.

13 Sillanpaa M, Makela A-L, Koivikko A. Acute liver failure andencephalopathy (Reye's syndrome?) during salicylate therapy.Acta Paediatr Scand 1975; 64: 877-80.

136 Petty B G, Zahka K G, BernsteinMT. Aspirin hepatitis associ-ated with encephalopathy. J Pediatr 1978; 93: 881-2.

137 Ulshen M H, Grand R J, Crain J D, Gelfand E W. Hepatotoxic-ity with encephalopathy associated with aspirin therapy inrheumatoid arthritis. J Pediatr 1978; 93: 1034-7.

126 Link K P, Overman R S, Sullivan W R, Huebner C F, Scheel LD. Studies on the haemorrhagic sweet clover disease. IXHypoprothrombinaemia in the rat induced by salicylic acid. JBiol Chem 1943; 147: 463-74.

139 Warren J S. Diflunisal-induced cholestatic jaundice. Br Med J1978; ii: 736.



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


140 Boardman P L, Hart F D. Side effects of indomethacin. AnnRheum Dis 1967; 26: 127-32.

141 Kelsey W M, Scharyi M. Fatal hepatitis probably due toindomethacin. JAMA 1967; 199: 586-7.

142 Fenech F F, Bannister W H, Grech J L. Hepatitis with biliver-dinaemia in association with indomethacin therapy. Br Med J1967; iii: 155-6.

143 Anderson R J. Severe reaction associated with sulindac admini-stration. N Engl J Med 1979; 300: 735-6.

144 Wolfe P B. Sulindac and jaundice. Ann Intern Med 1979; 91:656.

145 Dhand A K, LaBrecque D R, Metiger J. Sulindac (Clinoril)hepatitis. Gastroenterology 1981; 80: 585-6.

148 Cuthbert M F. Adverse reactions to non-steroidal antirheuma-tic drugs. Curr Med Res Opin 1974; 2: 600-10.

147 Chalmers T M. Clinical experience with ibuprofen in the treat-ment ofrheumatoid arthritis.Ann Rheum Dis 1969; 28:513-7.

148 Steupel D A, Miller J J. Lymphopenia and hepatic toxicity withibuprofen. J Pediatr 1977; 90: 657-8.

149 Sternlieb P, Robinson R M. Stevens-Johnson syndrome plustoxic hepatitis due to ibuprofen. N Y State J Med 1978; 78:1239-43.Bravo J F, Jacobson M P, Mertens B F. Fatty liver and pleuraleffusion with ibuprofen therapy. Ann Intern Med 1977; 87:200-1.Bass B H. Jaundice associated with naproxen. Lancet 1974; i:998.

112 Law I P, Knight H. Jaundice associated with naproxen.N EnglJMed 1976; 295: 1201.

153 Victorino R M M, Silveira J C B, Baptista A, DeMoura M C.Jaundice associated with naproxen. Postgrad Med J 1980; 56:368-70.

'5Stennett D J, Simonson W, Hall C A. Fenoprofen-inducedhepatotoxicity. Am J Hosp Pharm 1978; 35; 901.

155 Kuzell W C, Schaffarzick R W, Naugler W E, Gaudin G,Mankle E A. Phenylbutazone. Furtherclinical evaluation.ArchIntern Med 1953; 92: 646-61.

158 Englemann E P, Krupp M A, Rinehart J F, Jones R C,Gibson J R. Hepatitis following the ingestion of phenyl-butazone.JAMA 1954; 156: 98-101.

157 Ishak K G, Kirchner J P, Dhar J K. Granulomas andcholestatic-hepatoceliular injury associated with phenyl-butazone. Report of two cases. Am J Dig Dis 1977; 22: 611-7.

l5 Schenker S, Olson K N, Dunn D, Breen K J, Combes B.Intrahepatic cholestasis due to therapy of rheumatoid arthritis.Gastroenterology 1973; 64: 622-9.

159 Favreau M, Tannenbaum H, Lough J. Hepatic toxicity associ-ated with gold therapy. Ann Intern Med 1977; 87: 717-9.Wollheim F A, Lindstr6m C G. Liver abnormalities in penicil-lamine treated rheumatoid arthritis. Scand J Rheumatol 1979;suppl 28: 100-7.Barzilai D, Dickstein G, Enat R, Bassan H, Lichtig C, Geliei B.Cholestatic jaundice caused by D-penicillamine. Ann RheumDis 1978; 37: 98-100.

162 Rosenbaum J, Katz W A, Schumacher H R. Hepatotoxocityassociated with use of D-penicillamine in rheumatoid arthritis.Ann Rheum Dis 1980; 39: 152-4.Sparberg M, Simon N, Del Greco F. Intrahepatic cholestasisdue to azathioprine. Gastroenterology 1969; 57: 439-41.Davis M, Eddleston A LW F, Williams R. Hypersensitivity andjaundice due toazathioprine.Postgrad MedJ 1980; 56: 274-5.Hilton A M, Boyes B E, Smith P J, Sharp J, Dymock I W. Liverdisease in patients with joint symptoms. Ann Rheum Dis 1974;33: 540-7.Denman A M. Rheumatoid arthritis-a virus disease? J ClinPathol 1978; 31 (suppl 12): 132-43.



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

clinical associations betweenarthritis and liver diseasejaundice seemed equally effective. attempts...

Documents