Annals of the Rheumatic Diseases 1985, 44, 592-598

Arthritis and coeliac diseaseJ T BOURNE,' P KUMAR,2 E C HUSKISSON,' R MAGEED 3D J UNSWORTH,3 AND J A WOJTULEWSKI4

From the Departments of 'Rheumatology and 2Gastroenterology, St Bartholomew's Hospital, WestSmithfield, London ECIA 7BE; the 3Bone and Joint Research Unit, London Hospital Medical College,London El; and 4St Mary's Hospital, Eastbourne

SUMMARY We report six patients with coeliac disease in whom arthritis was prominent atdiagnosis and who improved with dietary therapy. Joint pain preceded diagnosis by up to threeyears in five patients and 15 years in one patient. Joints most commonly involved were lumbarspine, hips, and knees (four cases). In three cases there were no bowel symptoms. All wereseronegative. X-rays were abnormal in two cases. HLA-type Al, B8, DR3 was present in fiveand B27 in two patients. Circulating immune complexes showed no consistent pattern before orafter treatment. Coeliac disease was diagnosed in all patients by jejunal biopsy, and jointsymptoms in all responded to a gluten-free diet. Gluten challenge (for up to three weeks) failedto provoke arthritis in three patients tested. In a separate study of 160 treated coeliac patientsattending regular follow up no arthritis attributable to coeliac disease and no ankylosingspondylitis was identified, though in a control group of 100 patients with Crohn's disease theexpected incidence of seronegative polyarthritis (23%) and ankylosing spondylitis (5%) wasfound (p<0.01). Arthritis appears to be a rare manifestation of coeliac disease. This relationshipmay provide important clues to the role of gastrointestinal antigens in rheumatic diseases.

Key words: gluten, enteropathic synovitis.

The role of gut-derived factors in provokingrheumatic disease was first proposed by Rea Smithin 1922.1 Since then the link between inflammatorybowel disease and arthritis has become firmlyestablished,2 and antigens derived from the gut havebeen implicated in the pathogenesis of severalrheumatic disorders, including reactive arthritis,ankylosing spondylitis, and even rheumatoidarthritis.3

Coeliac disease (gluten-sensitive enteropathy) isknown to be associated with abnormal intestinalpermeability.4 Musculoskeletal manifestationsusually recognised are secondary to metabolic bonedisease or electrolyte depletion.5 Recently Adelizziet al.6 reported a patient with coeliac disease whopresented with seronegative oligoarthritis whichresolved on a gluten-free diet. We report a furthersix patients with coeliac disease in whom arthritiswas prominent at diagnosis and who improved withdietary therapy. In addition we determined theAccepted for publication 20 Feburary 1985.Correspondence to Dr J T Bourne, Dept of Rheumatology, StBartholomew's Hospital, West Smithfield, London EClA 7BE.

incidence of specific arthritis in a defined treatedcoeliac population.

Patients and methods

Six patients with coeliac disease (CD) in whomarthritis was prominent at diagnosis were seen at StBartholomew's Hospital, London, between 1970and 1983. One patient (No 1) was referred from StMary's Hospital, Eastbourne. Coeliac disease wasdiagnosed in all patients by jejunal biopsy. Returnof jejunal mucosa to normal after institution of agluten-free diet was confirmed by further jejunalbiopsy.

In a separate study 160 coeliac patients attendingroutine gastroenterology follow up were questionedand examined for evidence of rheumatic disorder(by JTB and ECH). As a control 100 patients withCrohn's disease were similarly assessed (by JTB). Inboth groups rheumatological investigations wereconfined to symptomatic subjects only. Ankylosingspondylitis was diagnosed by the New York (1966)criteria.7

592

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from

Arthritis and coeliac disease 593

The erythrocyte sedimentation rate (ESR) wasmeasured by the Westergren method. The IgMrheumatoid factor was assayed by latex fixation(Ortho Diagnostics). A modified National Institutesof Health microlymphocytotoxicity test was used forHLA-A,B,C, typing,8 and a two-colour fluor-escence technique was used for DR typing.9 For

immune complex studies patients' sera were rou-tinely stored at -70°C pre- and postinstitution of agluten-free diet. Antigliadin (AGA), antireticulin(ARA), and antiovalbumin (AOA) antibodieswere measured as previously described. 10-13 Im-mune complexes were measured by the Clq bindingassay of Zubler et al. (Clq polyethylene glycol

Fig. 1 Patient No 1 hand x-ray(March 1975). Soft tissue swellingis seen around several proximaland distal interphalangeal joints.

Fig. 2 (a) Patient No 1 handx-ray (March 1976). Soft tissueswelling is again seen aroundseveral proximal and distalinterphalangeal joints.Periarticular osteoporosis at themetacarpophalangeal joints, smallerosions at several distalinterphalangeal joints, andnumerous carpal cysts are evident.(b) Close 'up of Fig. 2a distalinterphalangeal joint left middle

1 finger to show small erosions.

Fig. 2a

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from

594 Bourne, Kumar, Huskisson, Mageed, Unsworth, Wojtulewski

(PEG))14 and the Clq solid phase assay of Hay etal.15 IgG- and IgA-containing complexes were iden-tified by a solid phase (Fab)2 anti-C3 assay afterbinding of peroxidase-conjugated (Fab)2 anti-IgG or

(Fab)2 anti-IgA (unpublished data). The signifi-cance was assessed by Fisher's exact probability test.

Results

CASE STUDY OF SIX PATIENTS WITH

COELIAC-RELATED ARTHRITIS (CRA)

Case historyAn 18-year-old female student (patient No 1)developed pain and swelling of both elbows andankles in July 1974. Symptoms partially subsidedover the following months but were accompanied byan erythematous papular rash over face, forearms,and legs. By March 1975 she was unwell with activesynovitis of several proximal and distal interpha-langeal joints, right wrist, and both knees. The ESRwas 28 mm/h. The following investigations werenormal or negative: haemoglobin, white cell count,rheumatoid arthritis latex test, antinuclear anti-bodies, calcium, phosphate, and alkaline phospha-tase. Hand x-rays showed soft tissue swelling only(Fig. 1).Her joint symptoms continued, accompanied by

worsening morning stiffness, malaise, and weightloss, until March 1976 when further investigationsrevealed an ESR of 48 mm/h, haemoglobin 9-9 g/dl,mean corpuscular volume (MCV) 120 fl, and lowserum iron and folate. Immunoglobulins werenormal. X-rays of the hand showed periarticularosteoporosis and small erosions at the distal inter-phalangeal joints (Figs 2a and b). Barium followthrough suggested malabsorption, and jejunalbiopsy showed subtotal villous atrophy consistentwith coeliac disease.She was placed on a gluten-free diet and within

one week noted improvement in her arthritis. BySeptember 1976 her joint symptoms had resolvedcompletely and she had gained weight. The onlyfinding was residual swelling of the distal interpha-langeal joints. She has remained well subsequently.Hand x-ray in 1983 showed healing of the erosions(Fig. 3).Of the initial six patients (Table 1) four were

female and two male. Age at diagnosis of coeliacdisease ranged from 20 to 60 years (mean 39-3years).

Duration of symptoms (Table 1). Joint painpreceded diagnosis by up to three years in fivepatients and by 15 years in one patient. In threecases there were no bowel symptoms, the diagnosisbeing suspected because of associated anaemia.

Pattern ofjoint involvement (Fig. 4). The distribu-tion of arthritis varied widely. Lumbar spine, hip,knee, and shoulder were affected most commonly.Pain at rest (six patients), morning stiffness of morethan 30 min (four patients), symptomatic flares (fivepatients), and joint swelling (two patients) wereprominent symptoms. Painful limitation of move-

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from

Arthritis and coeliac disease 595

Fig. 3 Patient No 1 hand x-ray(October 1983). Erosions at thedistal interphalangeal joints haveresolved.

ment (five patients) and joint tenderness (fourpatients) were the commonest physical signs.

Other prominent clinical features. Five patientscomplained of malaise and weight loss. Proximalmyopathy was noted in two cases.

Laboratory data (Table 2). Useful screening testsincluded evidence of anaemia (five patients), mac-rocytosis (four patients), and low serum folate (allsix patients). ESR was clearly raised in two patientsand marginally raised in a further three patients.Biochemical evidence for coexisting osteomalaciawas found in three patients (Nos 2, 4, and 6).All were latex negative and had normal immuno-globulins.

X-ray changes. X-rays were normal in fourpatients. One patient showed congenital fusion of

Lumbar spineHipKneeShoulderElbowWristAnkleCervical spineSacroiliac jointProximal I P jointDistal I P joint

Table 1 Six patients with coeliac-related arthritis (CRA): presentation

Patient Sex Age at diagnosis Duration of bowel symptoms -Duration of joint symptomsof coeliac disease (years) (years)

1 F 20 - 22 F 49 0-5 33 M 60 - 134 M 47 2-5 35 F 20 - 36 F 40 15+ 15+

4

444

Fig. 4 Pattern of joint involvement in six patients withcoeliac-related arthritis. (IP joint=interphalangeal joint.)

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from

596 Bourne, Kumar, Huskisson, Mageed, Unsworth, WojtulewskiTable 2 Six patients with CRA: laboratory data

Patient Haemoglobin MCV Serum folate ESR Calcium Phosphate Alkaline(mm/h) (mmol/l) (mmol/l) phosphatase* (lUll)

1 4 54 2<30 1-26 1222 4 T 1 26 2-37 0-84 6043 N t 4 17 2-40 1-13 644 4 N 4 22 1-68 0-91 2645 4 N 4 80 2-28 1-16 1056 1 7 1-70 1-23 24

*=Normal range=50-150 IU/I.N=normal; I =low; T =raised.

cervical vertebra C2 and C3 and asymptomaticsacroiliitis. The x-ray findings of patient No 1 havebeen presented above.

Tissue type (Table 3). HLA-type Al, B8, DR3was present in five patients. HLA-B27 was found intwo patients, neither of whom had clinical evidenceof ankylosing spondylitis. X-ray evidence of sac-roiliitis (asymptomatic) was present in one of these.

Outcome. All patients were placed on a gluten-free diet after demonstration of coeliac disease byjejunal biopsy. Three patients experienced completeand lasting remission of their arthritis within threemonths. Two patients noted marked improvementover a similar time period but continue to experi-ence residual joint symptoms. A further patientnoted moderate improvement only.

Gluten challenge. Three patients agreed to under-go gluten challenge for the purposes of: (a) confirm-ing the diagnosis of coeliac disease; and (b)observing the effect of a normal diet on the course oftheir arthritis. Each patient received a normal dietplus 20 g additional gluten per day. Jejunal biopsiesafter three weeks in patients No 1 and 5 and after oneweek in patient No 3 all showed relapse, confirmingthe diagnosis of coeliac disease. No patient experi-enced a return of joint symptoms over this period. Afourth patient (No 4) reported exacerbation of painand swelling in his left knee 24 hours after lapsingfrom his gluten-free diet, though this episode wasnot witnessed by us.

Table 3 Tissue type of six patients with CRA

Patient A B C DR

1 1,2 8,w44 w7,- 2,32 1,- 8,- w7,- 3,-3 1,11 8,27 w2,w7 3,44 1,3 8,w51 w7,- 3,75 2,w32 w44,w50 ND ND6 1,2 8,27 wl,w7 3,7

ND=not done.

Immune complex and antibody studies. Immunecomplexes were detected in four of the six patients,but no overall pattern emerged. Representative datafrom patient No 5 are illustrated in Fig. 5. IgA-containing complexes were not found in any patient.Immune complexes detected by the remainingmethods showed great variability on repeated sam-pling both before and after treatment. No significantrise in levels was seen during the period of glutenchallenge. Antigliadin and antireticulin antibodiestended to be high in untreated patients falling tonormal with institution of a gluten-free diet andrising again during gluten challenge. Antiovalbuminantibodies (reflecting immune response to non-specific dietary proteins) were detected in threepatients both before and repeatedly after institutionof a gluten-free diet.

Patient No 5Antibodies *-* Anti reticulin (ARA)

t o----o Anti gliadin(AGA)1/5120 o S --_ Ari ovabmin (AOA)

1/1280 _M _

164

o----o Clq Solid phase 6; r ne complxes *-* IgG i-c

A lg1A I-C4

z ol _ t.' Untreated GFdiet GFdiet Normal diet

4/12 19/12 3/52(Pre challenge) (Post challenge)

Fig. 5 Antigliadin (AGA), antireticulin (ARA),antiovalbumin (AOA) antibodies and immune complexesdetected byfour methods (see text) in patientNo 5 beforetreatrnent, after 4 months and 19 months on agluten-free (GF) diet, and after a three-week glutenchallenge. The hatched area indicates the normal ranges.

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from

Arthritis and coeliac disease 597

SEARCH FOR ARTHRITIS IN A TREATEDCOELIAC POPULATIONThe age range of the 160 treated coeliac patientsexamined was 15 to 81 years (mean 46.5 years),female to male ratio 3-3:1, and time from diagnosisone month to 67 years (mean 11*4 years). Thedistribution of locomotor disease is shown in Table4. Evidence of synovitis for which no clear diagnosiscould be made ('undifferentiated arthritis', Table 4)was found in six patients. The small joints of thehands and wrists were affected most commonly andsymptoms in all subjects settled within three years ofonset. This arthritis was not thought to be attribu-table to coeliac disease. No specific arthritis wasidentified and there was no evidence of ankylosingspondylitis.The age range of the 100 patients with Crohn's

disease examined was similar: 16 to 82 years (mean38-3 years), female to male ratio 2:1, and time fromdiagnosis five months to 35 years (mean 9-1 years).The distribution of locomotor symptoms is shown inTable 5. Seronegative polyarthritis was found in23% and ankylosing spondylitis in 5% (p<0-01).

Discussion

Well recognised musculoskeletal manifestations ofcoeliac disease (CD) include metabolic bone dis-

Table 4 Pattern of locomotor disease in 160 patients withtreated coeliac disease

Osteoarthritis, cervical/lumbar spondylosis 20Musculoskeletal pain without signs 14Undifferentiated arthritis 6

Right MCPJs, neck, shouldersRight+left wristsRight+left PIPJs and MCPJsIPJ right thumbRight+left PIPJsRight wrist and knee

Osteomalacia 3Hypermobility 1Gout 1

PIPJ=proximal interphalangeal joint; MCPJ=metacarpophalan-geal joint.

Table 5 Pattern of locomotor disease in 100 patientswith Crohn's disease

Seronegative polyarthritis 23Osteoarthritis. cervical/lumbar spondylosis 15Musculoskeletal pain without signs 2Ankylosing spondylitis 5Symptomatic sacroiliitis without x-ray signs 3Polymyalgia rheumatica 2Gout 1Palindromic rheumatismHypermobility 1

ease, muscle weakness (e.g., due to hypokalaemia)and paraesthesiae (due to hypocalcaemia).5 Inaddition there have been isolated reports of

poly , ,16 17an cunespolymyositis, pericarditis, and cutaneousvasculitis, 18 which resolved with institution of agluten-free diet. Adelizzi et al. first reported arthri-tis as a presenting symptom of CD.6 Their patient, a28-year-old white male, developed an acute asym-metrical arthritis involving knee, ankle, and subtalarjoints. CD was confirmed by jejunal biopsy. Noother cause of arthritis was found, and permanentresolution occurred with institution of a gluten-freediet. Subsequently a group of French workers19 havereported two patients with seronegative polyarthritisin whom CD was diagnosed and whose arthritisresolved completely on a gluten-free diet.We have identified a further six patients with CD

in whom arthritis was prominent at diagnosis. Thecausal relationship of CD and arthritis is supportedby the absence of other disease and by the strikingresponse to a gluten-free diet. Only two gastrointes-tinal conditions associated with a flat jejunal biopsymay cause arthritis - hypogammaglobulinaemia andlymphoma. Immunoglobulin levels and follow upover a minimum of four years have excluded these.Similarly no evidence of Crohn's disease, Whipple'sdisease, or causes of reactive arthritis were found.Osteomalacia was thought to contribute to (but notadequately to account for) symptoms in threepatients.The very varied pattern of arthritis in these

patients makes identification of CD difficult, par-ticularly since 50% of them had no bowel symptoms.Malaise and weight loss were useful clinical pointersand low serum folate was found in all six patients.The value of this test in tracing occult adult CD hasrecently been emphasised.2tThe pathogenesis of the arthritis is unclear, as

indeed is the pathogenesis of CD itself. Currentevidence favours immunologically mediated muco-sal injury in the context of marked genetic suscepti-bility. The high frequency of HLA-B8,-DR3 hasbeen repeatedly reported.2 It is possible that thearthropathy may be due to absorption through dam-aged jejunal mucosal of either immune complexesdirectly or gut-derived antigens which provoke ahumoral or cell-mediated response resulting in jointinjury. Intestinal permeability is known to beabnormal in untreated coeliac disease,4 returning tonormal within six months of the institution of agluten-free diet. In keeping with the findings ofother workers,22 23 we have shown the presence ofcirculating immune complexes in the majority of ourpatients, but the levels do not correlate well with theonset of arthritis. The failure to provoke arthritisafter gluten challenge despite relapse of jejunal

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from

598 Bourne, Kumar, Huskisson, Mageed, Unsworth, Wojtulewski

mucosa suggests that gluten or its constituents arenot directly responsible for promoting arthritis. Wehave not excluded this possibility, however, since thebowel may relapse more quickly than the joints. Itwas considered unethical to extend the period ofgluten challenge beyond three weeks.We postulate that abnormal mucosal permeability

in untreated coeliac disease allows access to as yetunidentified provoking antigens which may varywidely between individuals. This would account forthe varied clinical picture and the absence ofcoeliac-related arthritis in the treated population.We have not investigated the incidence of arthritisamongst untreated coeliacs. The lengthy time periodover which our six patients were identified suggeststhat the incidence amongst this population is verylow. This implies that altered intestinal permeabilityis only one of several factors leading to arthritis. Abetter understanding of the mechanism of this rareform of arthritis may provide important clues to thewider role of gastrointestinal antigens and geneticsusceptibility in rheumatic disease.

References

1 Smith R. The surgical relief of intestinal foci of infection inarthritis deformans. Ann Surg 1922; 76: 515-8.

2 Wright V, Watkinson G. The arthritis of ulcerative colitis.Medicine (Baltimore) 1959; 38: 243-59.

3 Neumann V, Wright V. Arthritis associated with bowel disease.Clin Gastroenterol 1983; 12: 767-95.

4 Hamilton I, Cobden I, Rothwell J, Axon A T R. Intestinalpermeability in coeliac disease: the response to gluten with-drawal and single-dose gluten challenge. Gut 1982; 23: 202-10.

5 Falchuk Z M. Gluten-sensitive enteropathy. Clin Gastroenterol1983: 12: 475-93.

6 Adelizzi R A, Pecora A A, Chiesa J C. Celiac disease: casereport with an associated arthropathy. Am J Gastroenterol1982; 77: 481-5.

7 Bennet P H, Burch T A. In: Bennet P H, Wood P H N, eds.Population studies of the rheumatic diseases. Amsterdam:Exerpta Medica, 1968: 305.

8 Festenstein H, Adams E, Burke J, Oliver R T D, Sachs J A,Wolf E. The distribution of HL-A antigens in expatriates fromEast Bengal living in London. 5th international symposium of

the histocompatibility workshop. In: Histocompatibility testing,Copenhagen: Munksgaard, 1972: 175-8.

9 Van Rood J J, van Leeuwen A, Ploem J S. Simultaneousdetection of two cell population by two-colour fluorescence andapplication to the recognition of B cell determinants. Nature1976; 262: 795-7.

10 Unsworth D J, Leonard J N, McMinn R M H, Swain A F,Holborow E J, Fry L. Antigliadin antibodies and smallintestinal mucosal damage in dermatitis herpetiformis. Br JDermatol 1981; 105: 653-8.

11 Unsworth D J, Manuel P D, Walker-Smith J A, Campbell C A,Johnson G D, Holborow E J. A new immunofluorescent bloodtest for gluten sensitivity. Arch Dis Child 1981; 56: 864-8.

12 Seah P P, Fry L, Hoffbrand A V, Holborow E J. Tissueautoantibodies in dermatitis herpetiformis and adult coeliacdisease. Lancet 1971; i: 834-6.

13 Kendrick R G, Walker-Smith J A. Immunoglobulins anddietary protein antibodies ip childhood coeliac disease. Gut1970; 11: 635-40.

14 Zubler R H, Lange G, Lambert P H, Miescher P A. Detectionof immune complexes in unheated sera by a modified 25f-Clq binding test. Effect of heating on the binding of Clq byimmune complexes and application of the test to systemic lupuserythematosus. J Immunol 1976; 116: 232-5.

15 Hay F C, Nineham L J, Roitt I M. Routine assay for thedetection of immune complexes of known immunoglobulinclass using solid phase Clq. Clin-Exp Immunol 1976; 24:396400.

16 Henriksson K G, Hallert C, Norrby K, Walan A. Polymyositisand adult coeliac disease. Acta Neurol Scand 1982; 65: 301-19.

17 Dawes P T, Atherton S T. Coeliac disease presenting asrecurrent pericarditis. Lancet 1981; i: 1021-2.

18 Meyers S, Dikman S, Spiera H, Schultz N, Janowitz H D.Cutaneous vasculitis complicating coeliac disease. Gut 1981; 22:61-4.

19 Prier A, Le Quintrec Y, Camus J P, Gendre J P, Pradel B.Polyarthrites inflammatoires associees a une atrophie villosi-taire du grele. Rev Rhum Mal Osteoartic 1980; 47: 661-5.

20 Haliert C, Tobiasson P, Walan A. Serum folate determinationsin tracing adult coeliacs. Scand J Gastroenterol 1981; 16: 263-7.

21 Scott H, Brandtzaeg P, Thorsby E, Baklein K, Fausa 0, Ek J.Mucosal and systemic immune response patterns in coeliacdisease. Ann Allergy 1983; 51: 233-9.

22 Doe W F, Booth C C, Brown D L. Evidence for complement-binding immune complexes in adult coeliac disease, Crohn'sdisease and ulcerative colitis. Lancet 1973; i: 402-3.

23 Mohammed I, Holborow E J, Fry L, Thompson B R,Hoffbrand A V, Steward J S. Multiple immune complexes andhypocomplementaemia in dermatitis herpetiformis and coeliacdisease. Lancet 1976; ii: 487-90.

copyright. on M

ay 28, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.44.9.592 on 1 S

eptember 1985. D

ownloaded from