Annals of the Rheumatic Diseases, 1988; 47, 218-223

The arthropathy of cystic fibrosisJOSH DIXEY,' ANDREW N REDINGTON,2 ROBIN C BUTLER,'MICHAEL J SMITH,2 JOHN R BATCHELOR,3 DAVID F WOODROW,'MARGARET E HODSON,2 JOHN C BATTEN,2AND DERRICK A BREWERTON'

From 'Charing Cross and Westminster Medical School; 2Brompton Hospital; and the 3Royal PostgraduateMedical School, London

SUMMARY Musculoskeletal symptoms are frequent in cystic fibrosis (CF). Here the clinicalfeatures of 29 patients with CF who had significant arthropathy are described. Twelve hadepisodic arthritis (EA) characterised by repeated short attacks of severe, incapacitatingpolyarthritis, which in seven was associated with fever and erythema nodosum. Ten patients hadhypertrophic pulmonary osteoarthropathy (HPOA). The onset of symptoms in the group withHPOA was usually later (mean age 20 years v 16 years for EA) and was associated withsignificantly worse lung function than in patients with CF, either without arthropathy or with EA.Seven patients had arthropathies which could not be classified as EA or HPOA.

Key words: episodic arthritis. pulmonary osteoarthropathy.

In recent years the life expectancy of patients withcystic fibrosis (CF) has improved,' and manypatients survive to adult life. As a consequence,rheumatic disorders are now commonly observed inolder patients. Two distinct arthropathies have beendescribed in CF: episodic arthritis (EA), whichappears to be more common in children3 4 but hasbeen reported in adults5; and hypertrophic pulmon-ary osteoarthropath5y (HPOA), which is normallyconfined to adults but has been described inchildren.6 7 Both EA and HPOA have been thesubjects of reports of individual cases.813 So far,there have been no prospective studies of thesearthropathies in patients with CF. To define theclinical and pathological characteristics in moredetail and to prepare for a continuing researchprogramme we studied all patients attending ourclinic who complained of musculoskeletal symptoms.

Patients and methods

Approximately 250 patients attend the BromptonHospital adult cystic fibrosis clinic. Since 1983 allpatients with rheumatic symptoms have been re-ferred for rheumatological assessment.

Accepted for publication 27 July 1987.Correspondence to Professor D A Brewerton, RheumatologyDepartment, Westminster Hospital, 17 Page Street, London SWI P2AP.

DIAGNOSISEpisodic arthritis was diagnosed when there was ahistory of repeated attacks of severe polyarthropathywith clear resolution of symptoms between attacks.HPOA was diagnosed on the basis of the followingclinical features: (a) finger clubbing, (b) chronicsymmetrical pain or swelling, or both, of the wrist,knee, or ankle, and (c) the presence of periostitis onradiographs or a positive 99mTc hydroxydiphos-phonate (HDP) bone scan, or both. A significantnumber of patients with rheumatic symptoms didnot conform to EA or HPOA, and they areconsidered separately.

INVESTIGATIONSWhenever possible the following investigations wereperformed while the patient was symptomatic: fullblood count, erythrocyte sedimentation rate (ESR),biochemical profile including plasma uric acid con-centration, RAHA (rheumatoid arthritis haemagglu-tination test) rheumatoid factor (RF), antinuclearfactor (ANF), immunoglobulin concentrations, C3and C4 complement levels, immune complex (IC)titre (enzyme linked immunosorbent assay (ELISA),Flexia system; Pharmacia (UK) Ltd), and HLAanalysis (class I antigens). Radiographs were takenof the affected joints, and in six of the patients withHPOA 99mTc HDP bone scans were performed.Knee arthroscopy was performed under local

218

copyright. on O

ctober 16, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.47.3.218 on 1 M

arch 1988. Dow

nloaded from

anaesthesia using a Storz paediatric arthroscope inone patient with EA and in one patient with HPOA.The synovial biopsy specimens obtained were fixedfor routine light and electron microscopy, and tissuewas snap frozen for immune studies. In all patientsthe forced expiratory volume in one second (FEVy)and the forced vital capacity (FVC) were recordedon a spirometer and values expressed as percentagepredicted for age, height, and sex. 14 Lung functiondata were expressed as mean (SD). The EA andHPOA groups were compared with each other andwith a group of 170 patients with CF withoutarthropathy (mean age 19 (SD 1-2) years) using theMann-Whitney U test. Permission for this study wasobtained from the Brompton Hospital ethics com-

mittee.

Results

CLINICAL FEATURESEpisodic arthritisTwelve patients had a clinical diagnosis of EA. Theage of onset varied considerably (mean 16 years,range 5-29), but attacks usually started in adolesc-ence. In those with an early onset parents reportedthat there had been complaints of severe widespreadjoint pains of sudden onset with flu-like symptoms.Frequently, an acute infection was diagnosed andantibiotics were prescribed. In most episodes wehave observed, the patients were temporarily unableto walk because of severe lower limb pain. Manylooked miserable because of the pain and lay still inbed. Typically, there was erythema over the fingerjoints, with tenderness over most peripheral jointsbut little swelling. Although active joint movementwas limited, the passive range was usually normal.Attacks lasted an average of four days, and thejoints then returned to normal. One of the 12patients had severe symptoms which persisted for upto two weeks. Seven patients developed a high,swinging fever and a rash which was typical oferythema nodosum, being present on the legs andoccasionally also on the forearms. Eight patientscontinued to have attacks one to three years afterdiagnosis, and in two patients the attacks haveoccurred at frequent intervals over a much longerperiod (11 and 17 years). In a further two, attacksstarted in childhood and stopped after puberty.Only one patient was not taking pancreatic supple-ments. This patient stopped his enzyme supplementsbecause he believed that they provoked his attacksof EA. He was unwilling to undergo a provocationtest. In most patients non-steroidal anti-inflammatorydrugs (NSAIDs) led to rapid symptomatic relief. Inone patient, however, attacks of EA were unrespon-sive to NSAIDs but responded to prednisolone.

The arthropathy of cystic fibrosis 219

Hypertrophic pulmonary osteoarthropathyTen patients had HPOA. The age of onset tended tobe later than in the patients with EA (mean20 years, range 15-21 years). The onset was usuallyinsidious, with bone pain in the region of wrists,knees, and ankles. At first, the pain was not severebut progressed steadily to a continuous and trouble-some ache with symmetrical swelling and periarticulartenderness of the large joints. In most patients thejoint disease had a symmetrical pattern affectingwrists (six), knees (nine), and ankles (seven). Kneeeffusions were common; in severe disease thepatients experienced difficulty in walking because ofankle pain. The small joints of hands and feet wereinvolved in one patient. In three patients there wasan exacerbation of joint symptoms during chestinfections, and, in one, the size of the knee effusionregularly reflected the severity of the acute ex-acerbations of his chest disease. All the patientswere taking pancreatic supplements. At the onset ofHPOA the pain usually responded to NSAIDtherapy, but drug therapy became less effective asthe disease progressed.

MiscellaneousSeven patients had rheumatic disorders which couldnot be classified as HPOA or EA. Three had clinicalfeatures similar to HPOA but with no periostealreaction on radiography. One women had intermit-tent myalgia induced by exercise, and one man hadthe clinical features of ankylosing spondylitis. A 16year old woman had suffered from a chronic painlesstenosynovitis over the dorsum of both wrists andankles for 10 years. A man developed classicalseropositive rheumatoid arthritis (RA) at the age of31 years.

In addition to the 29 patients included in thisreport, many patients with CF were examined whocomplained of chronic knee pains. No significantarthropathy could be demonstrated and in mostpatients the symptoms were mild and not incapaci-tating. These patients were not investigated further.

PATHOLOGICAL FINDINGSEpisodic arthritisRheumatoid factor and ANF were not detected andthe plasma urate was normal. Blood was takenduring attacks of EA in five patients with thefollowing abnormal findings: a raised white bloodcell count (WBC) count in two (12 and 19x 109/l); araised ESR in four (range 12-55 mm/h); a raised ICtitre in one (141 IU/l, normal range (NR) <20 IU/l).The immunoglobulin and complement levels werenormal in all five patients.One patient with EA was investigated in more

detail during an acute attack. Plasma C reactive

copyright. on O

ctober 16, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.47.3.218 on 1 M

arch 1988. Dow

nloaded from

220 Dixey, Redington, Butler, Smith, Batchelor, Woodrow, Hodson, Batten, Brewerton

protein was raised to 152 mg/l (NR <10 mg/i). Kneearthroscopy was normal apart from a small quantityof synovial fluid with minimal evidence of inflamma-tion (20 neutrophils/ml, protein 16 g/l). A synovialbiopsy specimen showed no abnormality on lightmicroscopy, and immunofluorescent staining wasnegative apart from a few small deposits of IgM onendothelial cells. On ultrastructural examinationthere was a conspicuous reduplication of the en-dothelial basal lamina.

Hypertrophic pulmonary osteoarthropathyThe WBC count was raised in five patients (range14-21 x 109/l) and the ESR was raised in eight (range22-95 mm/h). Immune complexes were present inthree (63, 92, and 884 IU/l). Rheumatoid factor andANF were not detected. Serum IgM was normal.Seven patients had raised IgG (range 16-8-23.1 g/l,NR 6-3-16-0 g/l) and IgA (range 3*2-6-9 g/l, NR0*7-3.2 g/l). Complement C3 levels were all normal,but complement C4 levels were depleted in fivepatients (range 0*12-0-23 g/l, NR 0-25-0-75 g/l).Knee arthroscopy was performed in one patient.

The synovial membrane was normal in appearance,apart from the presence of a fibrinous exudate. Thesynovial fluid contained 560 WBCs/ml (95% lym-phocytes), protein 32 g/l. On light microscopy therewas engorgement of blood vessels with mononuclearinfiltrate. Specific immunofluorescent stainingshowed a prominent coating of IgM on endothelialcells, occasional granules of C3 on vessel walls, andpositive staining for fibrinogen on the lining layer.

Ultrastructural examination showed a conspicuousreduplication of capillary basal laminae and focalloss of endothelial cells with partial occlusion ofsome capillaries by fibrin and platelets (Fig. 1).

HLA TYPINGIn both groups of patients there was a normaldistribution of class I HLA antigens when comparedwith normal controls in a Caucasian population andwith patients with CF without arthropathy.

DIAGNOSTIC IMAGINGIn EA the radiographs taken of the most symptoma-tic joints (normally hands, wrists, and knees) wereinvariably normal. A 99mTc HDP bone scan, per-formed in one patient during an acute attack of EA,was normal.

Periostitis at the distal ends of long bones waspresent in nine patients with HPOA (Fig. 2a). Ofthe six 99'Tc HDP bone scans performed, four had'hot' areas corresponding to the periostitis seen onradiography (Fig. 2b). One patient had a positivebone scan accompanied by normal radiographs andone had a normal scan but a marked periostealreaction on radiography.

LUNG INVESTIGATIONSIn EA the percentages of predicted normal valuesfor FEV1 and FVC (mean (SD) 58 (14)% and 73(25)% respectively) were not significantly differentfrom the values for control patients with CF withoutarthropathy (Fig. 3). Pseudomonas aeruginosa was

....~ ~ ~~ aro.cpz~ walshwn

Fig. 1 Electron micrograph of- t̂ w < 0 Part ofa capillary wall showing

it # ~~~~discontinuity in the endothelial cell- ffi~~4 lining (long arrow) and multiple

) t ~~~basal laminae (short arrows) in apatient with CFand HPOA.

i L O ~~~~~(Uranyl acetate and lead citrate.)

copyright. on O

ctober 16, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.47.3.218 on 1 M

arch 1988. Dow

nloaded from

The arthropathy of cystic fibrosis 221

0

0

S

00

0@

I000

S.000

000

p<v.001

p<O. 05NS

CF CF+HPOA CF+EA(n=170) (n=10) (n=12)

Fig. 3 FEV, values (percentage predictedfor weight,height, and sex) in patients with CF without arthropathy(n=1 70), CF+HPOA (n=10), and CF+EA (n=12);p values refer to Mann-Whitney U tests. (FEV,=forcedexpiratory volume in one second; CF=cystic fibrosis;HPOA=hypertrophic pulmonary osteoarthropathy;EA=episodic arthritis.)

Fig. 2 (a) Wrist radiograph and(b) bone scan ofa patient with CFandHPOA showing periosteal newboneformation at the distal ends ofradius and ulna bones (a) andincreased uptake ofradioisotopeindicative ofperiostitis (b).

(b)

FEV1 100

(x)

80 F

60 -

40 F

20 F

OL

(a)

.^I

.

copyright. on O

ctober 16, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.47.3.218 on 1 M

arch 1988. Dow

nloaded from

222 Dixey, Redington, Butler, Smith, Batchelor, Woodrow, Hodson, Batten, Brewerton

present in the sputum of nine patients, with addi-tional Staphylococcus aureus in two. In one patientno pathogens were isolated on repeated sputumculture.

In patients with HPOA the percentage predictedFEV, was 26 (5)%, which was significantly worsethan for those with EA (p<0-05) (Fig. 3). The FVCpercentage predicted was also significantly worsewhen compared with the group with EA (46 (9)% v73 (25)%; p<0-01) and with the control group (70(24)%; p<0.001). All patients with HPOA hadpositive sputum cultures for P aeruginosa.

Discussion

Although arthritis has been little recognised as acomplication of cystic fibrosis (CF), it is nowestablished as a major cause of pain and disability inapproximately 10% of young adults with this disease,and it is evident that arthritis will become morecommon as patients live longer. Most patients haveone of two types of arthritis: episodic arthritis orhypertrophic pulmonary osteoarthropathy, whichcan usually be readily distinguished on clinicalgrounds.

Clinically, EA associated with CF is unique.Repeated attacks of severe, generalised joint painrequiring bed rest, often accompanied by fever anderythema nodosum, and with complete resolution ofjoint symptoms within four days, cannot be satisfac-torily explained by other forms of recurrent arthritis,such as reactive arthritis following genitourinary orgut infections,15 erythema nodosum,16 palindromicrheumatoid arthritis, 17 or intestinal bypass arthritis,18none of which has the same clinical features. Onlyfamilial Mediterranean fever (FMF)'9 is similar inthe severity of the arthritis, the pattern of theattacks, and the fever, but our patients with EAlacked the other clinical features, racial characteris-tics, and family history of FMF, and there is noknown association between CF and FMF, either inindividuals or in families. A metaraminol provocativetest was not indicated.2"

In HPOA the clinical onset is insidious and thecourse is related to the severity of the lung disease.On occasions, exacerbations of the arthritis parallelthe worsening of the chest infection.4 9 P aeruginosais the dominant pathogen in the sputum, but it is nomore common in patients with HPOA than in EA orin CF without arthritis.The pathogenesis of both EA and HPOA is

obscure. From the limited histological material andthe immunological investigations in our study, theevidence of inflammatory and immunological re-sponses is sparse and probably no more than may beexpected in the presence of persistent chest infec-

tion. The absence of an association between EA andHLA-B8 is also important because in sarcoidosisboth erythema nodosum and acute arthritis areassociated with B8.2' 22 The further possibility of agenetic component linked to the CF related gene onchromosome 723 has been considered and will beinvestigated when appropriate techniques are avail-able.

Despite the lack of similarity between EA orHPOA and reactive arthritis following genitourinaryor gut infection (including the absence of anassociation with HLA-B27), further investigation ofthe role of chest or gut infection is indicated. It isalso necessary to exclude the possibility that EAmay be an adverse reaction to pancreatic supple-ments. Finally, because of the evidence that neuro-genic mechanisms may be involved in clubbing ofthe fingers, in periosteal reactions, in HPOA, and inCF,24-6 there is now a new indication and a newopportunity to investigate neurogenic mechanismsin both EA and HPOA.

Dr B Strickland kindly supervised the diagnostic imaging and Dr JMoss the ultrastructural analysis. ANR was supported by theFrances and Augustus Newman Foundation. We are grateful toMiss Amanda Brown and Miss Karen Bailey for their secretarialassistance.

References

1 di Sant'Agnese P A, Davis P B. Cystic fibrosis in adults. 75cases and a review of 232 cases in the literature. AmJ Med 1979;66: 121-32.

2 Shwachman H, Kowalski M, Khaw K-T. Cystic fibrosis: a newoutlook. 70 patients above 25 years of age. Medicine (Balti-more) 1977; 56: 129-49.

3 Newman A J, Ansell B M, Episodic arthritis in children withcystic fibrosis. J Pediatr 1979; 94: 594-6.

4 Schidlow D V, Goldsmith D P, Palmer J, Huang N N. Arthritisin cystic fibrosis. Arch Dis Child 1984; 59: 377-9.

5 Rush P J, Shore A, Coblentz C, Wilmot D, Corey D, LevisonH. The musculoskeletal manifestations of cystic fibrosis. SeminArthritis Rheum 1986; 15: 213-25.

6 Grossman H, Denning C R, Baker D H. Hypertrophicosteoarthropathy in cystic fibrosis. Am J Dis Child 1964; 107:39-44.

7 Nathansan I, Riddlesberger M M. Pulmonary hypertrophicosteoarthropathy in cystic fibrosis. Pediatr Radiol 1980; 135:649-51.

8 Athreya B H, Borns P, Rosenlund M L. Cystic fibrosis andhypertrophic osteoarthropathy in children. Am J Dis Child1975; 129: 634-7.

9 Matthay M A, Matthay R A, Mills D M, Lakshminarayan S,Cotton E. Hypertrophic osteoarthropathy in adults with cysticfibrosis. Thorax 1976; 31: 572-5.

10 Vaze D. Episodic arthritis in cystic fibrosis. J Pediatr 1980; 96:346.

11 McGuire S, Monaghan H, Tempany E. Arthritis in childhoodcystic fibrosis. Ir J Med Sci 1982; 151: 253-4.

12 Blau H, Yahav J, Katznelson D. Episodic arthritis in cysticfibrosis. Progress in Rheumatology 1984; 2: 357-60.

13 Phillips B M, David T J. Pathogenesis and management ofarthropathy in cystic fibrosis. J R Soc Med 1986; 79 (suppl 12):44-50.

copyright. on O

ctober 16, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.47.3.218 on 1 M

arch 1988. Dow

nloaded from

The arthropathy of cystic fibrosis 223

14 Cotes J E. Lung function: principles and application inmedicine. Oxford: Blackwell Scientific, 1971.

15 Keat A. Reiter's syndrome and reactive arthritis in perspective.N Engl J Med 1983; 309: 1606-15.

16 Truelove L H. Articular manifestations of erythema nodosum.Ann Rheum Dis 1960; 19: 174-80.

17 Hench P S, Rosenberg E F. Palindromic rheumatism: newoft-reccurring disease of joints (arthritis, peri-arthritis, para-arthritis) apparently producing no articular residues; report of34 cases. Proceedings of the Staff Meeting of the Mayo Clinic1942; 16: 808-15.

18 Fagan E A, Elkon K B, Griffin G F, et al. Systemicinflammatory complications following jejuno-ileal bypass.Q J Med 1982; 51: 445-60.

19 Heller H, Gafni J, Michaeli D, et al. The arthritis of familialMediterranean fever. Arthritis Rheum 1966; 9: 1-17.

20 Barakat M H, El-Khawad A 0, Gumaa K A, El-Sobki N I,Fenech F F. Metaraminol provocative test: a specific diagnostic

test for familial Mediterranean fever. Lancet 1984; i: 656-7.21 Brewerton D A, Cockburn C, James D G, Neville E. HLA

antigens in sarcoidosis. Clin Exp Immunol 1977; 27: 227-9.22 Hedfors E, Lindstrom F. HLA-B8/DR3 in sarcoidosis. Tissue

Antigens 1983; 22: 200-3.23 Scambler P J, Wainwright B J, Farrall M, et al. Linkage of

COL1A2 collagen gene to cystic fibrosis, and its clinicalimplications. Lancet 1985; ii: 1241-2.

24 Schneerson J M. Digital clubbing and hypertrophic osteoarthro-pathy: the underlying mechanisms. Br J Dis Chest 1981; 75:113-31.

25 Davis P B, Shelhamer J R, Kaliner M. Abnormal adrenergicand cholinergic sensitivity in cystic fibrosis. N Engl J Med 1980;302: 1453-6.

26 McPherson M, Dormer R L, Bradbury N A, Dodge J A,Goodchild M. Defective I-adrenergic secretory responses insubmandibular acinar cells from cystic fibrosis patients. Lancet1986; ii: 1007-11.

copyright. on O

ctober 16, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.47.3.218 on 1 M

arch 1988. Dow

nloaded from