LETTERS

Cladribine in the treatmentof systemic lupuserythematosus nephritisSystemic lupus erythematosus (SLE) nephri-tis often requires treatment with cyclophos-phamide, which carries the risk of major sideeVects including infection, ovarian failureand bladder malignancy. Therapeutic strate-gies that would specifically target lym-phocytes are appealing. Following the firstreport of the use of the purine nucleosideanalogue cladribine (2-chloro-2’-deoxyadenosine), a selective lymphocyte de-pleting agent, in the treatment of lupusnephritis,1 we report our experience in twopatients with severe renal involvement.

CASE 1A 32 year old woman was diagnosed withSLE at age 28, with polyarthritis, photosensi-tive rash, subcutaneous nodules, fatigue andlymphopenia. ANA, anti-dsDNA, anti-Smand anti-RNP antibodies were present. Vari-ous immunosuppressants and corticosteroidsfailed to maintain a sustained remission. Twoand a half years after presentation, she devel-oped haematuria and proteinuria and renalbiopsy revealed WHO Class III lupus nephri-tis. Treatment with pulsed intravenous cyclo-phosphamide and methylprednisolone had tobe stopped after four months and a totalcyclophosphamide dose of 9 g because of ananaphylactic reaction during an infusion.Despite azathioprine, 150 mg/day, and pred-nisolone, up to 20 mg/day, she developedsevere hypertension (210/120 mm Hg) andbiopsy confirmed lymphocytic cutaneousvasculitis. Cladribine (0.05 mg/kg/day forseven days as continuous intravenous infu-sion) and prednisolone 60 mg/day werestarted. Cutaneous vasculitis resolved withinfive days and serum creatinine fell from 190to 120 µmol/l in five weeks. Cladribine waswell tolerated apart from a herpes simplexinfection in the natal cleft that responded toacyclovir. She relapsed three months later,with a new rise in creatinine (154 µmol/l) andrecurrence of cutaneous vasculitis.

A further infusion of cladribine was given,keeping prednisolone at 5 mg/day. Althoughthe vasculitic rash again resolved, renal func-tion and proteinuria continued to deteriorate.

She has subsequently been maintainedwith mycophenolate mofetil 1 g twice dailyand oral prednisolone. Serum creatinine has

returned to 98 µmol/l and proteinuria to5.3 g/24 h and remained stable despitegradual reduction of prednisolone dose to 15mg daily.

CASE 2

A 35 year old woman was diagnosed with SLEat age 31, with fever, pancytopenia, and neph-rotic syndrome (proteinuria 6.65 g/24 h).ANA and anti-dsDNA antibodies werepresent. Renal biopsy revealed WHO Class IIIlupus nephritis. In the next four years sherequired three treatment cycles of intravenouscyclophosphamide (total dose per six monthcycle: 9–10 g). Azathioprine, methotrexate,cyclosporin A and prednisolone 5–40 mg/dayin the interim had failed to control herdisease. Cyclophosphamide, additionally, hadresulted in premature ovarian failure. Repeatrenal biopsy showed progression to Class IVnephritis with focal necrosis and crescents.Cladribine (continuous IV infusion of 0.05mg/kg/day for seven days) and prednisolone40 mg/day proved ineVective as creatininerose from 149 to 243 µmol/l in two months.She also developed a perineal herpes simplexinfection but drug was otherwise well toler-ated. Pulse intravenous cyclophosphamideand methylprednisolone were subsequentlyreintroduced and creatinine has again fallen to118 µmol/l.

Table 1 shows the results of investigationsbefore and after cladribine infusions for bothcases.

In the initial study by Davis et al,1 three ofseven patients treated with continuous cladri-bine infusion for a week responded com-pletely and renal function did not deterioratein any of the seven patients. Our limitedexperience suggests that cladribine may beeVective in other manifestations of SLE (thatis, cutaneous vasculitis), but it does not seemto have a consistent eVect in severe nephritis.Good tolerability of the drug was confirmedand although herpes simplex infections oc-curred in both patients the role of corticoster-oids cannot be ignored.

Further studies are required to establishthe position of cladribine in the treatment ofSLE especially in the presence of otherlymphocyte depleting agents such as myco-phenolate mofetil, which is reported to beeVective in lupus nephritis,2–4 even in casesrefractory to cyclophosphamide.5

VLASSIS KONTOGIANNISPETER C LANYON

RICHARD J POWELLClinical Immunology Unit, University Hospital,

Queens Medical Centre,Nottingham NG7 2UH

Correspondence to: Dr V Kontogiannis.

1 Davis CJ, Austin H, Boumpas D, Fleisher TA,Yarboro C, Larson A, et al. A pilot study of2-chloro-2’deoxyadenosine in the treatment ofsystemic lupus erythematosus-associatedglomerulonephritis. Arthritis Rheum1998;41:335–43.

2 Briggs W, Choi M, Scheel PJ. Succesful mycofe-nolate mofetil treatment of glomerular disease.Am J Kidney Dis 1998;31:213–17.

3 Ho A, Magder L, Petri M. The eVect ofazathioprine/mycophenolate on systemic lupuserythematosus activity. [Abstract]. ArthritisRheum 1998;41(suppl.):S281.

4 Pashinian N, Wallace D, Klinenberg J. Myco-phenolate mofetil for systemic lupus erythema-tosus (abstract). Arthritis Rheum 1998;41(suppl):S110.

5 Glicklich D, Acharya A. Mycophenolate mofetiltherapy for lupus nephritis refractory tointravenous cyclophosphamide. Am J KidneyDis 1998;32:318–22.

Leg bone pain syndromein a kidney transplantpatient treated withtacrolimus (FK506)Patients with chronic renal failure oftendevelop musculoskeletal problems such asrenal osteodystrophy and amyloidarthropathy,1 and in successful renal trans-plantation other complications may ensue,particularly avascular necrosis.2 Since theavailability of immunosuppressive agents forrejection, there has been a decrease inmusculoskeletal problems, however, newcomplications have been described such as asymmetrical bone pain syndrome and reflexsympathetic dystrophy syndrome (RSDS),some of them related to cyclosporin.3–6

Tacrolimus is a novel macrolide withpotent immunosuppressive eVects and with avery similar mechanism of action to cy-closporine A—that is, calcineurin phos-phatase inhibition.7 8 We report on a patienttreated with tacrolimus, who developed a legbone pain syndrome, two months after kidneytransplantation.

The patient was a 50 year old woman withsevere hypertension, treated with atenolol(100 mg/day), verapamil (240 mg/day) andclonidine (0.150 mg/day). She developedchronic renal failure and was treated withperitoneal dialysis in 1995. In 1997 sheunderwent a kidney transplant from a cadaverand immunosuppressive treatment with tac-rolimus (4 mg/day) and prednisone (15mg/day) was started. Two months after trans-plantation she reported progressive bilateralsymmetric pain in the knees. Because of painand diYculty in walking she was readmittedto our unit. At this time, the patient wasreceiving tacrolimus (4 mg/day) and pred-nisone (5 mg/day). Clinical examinationrevealed pain on movement and tendernessover the bone and joint line, without swelling

Table 1 Results of investigations before and after cladribine infusions

Patient 1 Patient 2

First infusion Second infusion

Before AfterBefore After Before After

Proteinuria 12.25 g/24 h 4.2 g/24 h 5.2 g/24 h 12.4 g/24 h 4 g/24 h 7.2 g/24 hSerum creatinine 190 µmol/l 120 µmol/l 154 µmol/l 163 µmol/l 149 µmol/l 243 µmol/lAnti-ds DNA 132 IU/ml 58 IU/ml 292 IU/ml >300 IU/ml 171 IU/ml 49 IU/mlC3 0.51 g/l 0.67 g/l 0.72 g/l 0.51 g/l 0.39 g/l 0.60 g/lC4 0.13 g/l 0.14 g/l 0.15 g/l 0.12 g/l 0.12 g/l 0.16 g/lC3d 22 units/ml 10 units/ml 23 units/ml 20 units/ml 13 units/ml 12 units/mlUrine analysis red cells, hyalogranular, cellular casts red cells, a few casts

Reference ranges: serum creatinine 50–100 µmol/l, anti-dsDNA: 50–300 IU/ml positive, >300 IU/ml strongly positive, C3: 0.63–1.19 g/l, C4: 0.11–0.43 g/l, C3d: upto 12 units/ml.

Ann Rheum Dis 1999;58:653–660 653

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

or increased temperature. She had no signs ofautonomic vasomotor disturbances and ar-ticular mobility was normal. Examination ofthe remaining peripheral and axial joints wasnormal.

Blood tests showed creatinine levels of 1.3mg/dl, calcium of 10.1 mg/dl, phosphate of3.5 mg/dl and urate of 7.2 mg/dl. Other labo-ratory findings were normal. Patchy oste-oporosis in the knees was seen radiographi-cally. Bone scintigraphy showed intenseuptake in both the osseous and vascularphases in the knees (fig 1). Calcitonintreatment was begun (three monthly cycles of100 intramuscular units/day during 20 days)without clinical improvement. Because of thehigh serum concentrations of tacrolimus (15µg/ml) and the ineVective calcitonin treat-ment, tacrolimus was reduced to 2 mg/day.Nine months after transplantation, she wasfree of symptoms and radiographs andtacrolimus concentration (5.1 µg/ml) werenormal. Changes in plasma tacrolimus con-centrations subsequent to the resolution ofsymptoms did not occur and the patient con-tinued asymptomatic.

We describe a complication in a patienttreated with tacrolimus after kidney trans-plantation that is similar to that described byother authors in transplanted patients treatedwith cyclosporin.4–6 Although the radiographicand bone scintigraphy findings suggestedRSDS, the symptoms of this patient were notthe classic features of this entity. The eYcacyof corticosteroids in the treatment of uncom-plicated RSDS has been demonstrated,4 5 so itis possible that corticosteroids might have aprotective role against a full RSDS develop-ment, as she was treated with high doses ofprednisone after the renal transplantation.

The early onset of symptoms after theadministration of the drug and the clinicalimprovement after the reduction of theimmunosuppressant dose, are features thatsupport a possible relation between tac-rolimus and the leg bone pain syndrome. Thepatient had high plasma tacrolimus concen-trations at the onset of the clinical symptomsand the improvement appeared only whenthe drug doses went down. Although recur-rence of knee symptoms with an increase intacrolimus dose would be much strongerproof of this association, it is not ethicallyjustifiable. Furthermore, she was treated withverapamil in addition to other drugs for con-trolling hypertension. Verapamil might haveplayed a part in a possible increased risk forthis clinical complication, because it de-creases tacrolimus clearance.10 However,there are reports that calcium channel block-

ers (albeit of the dihydropyridine type) canimprove the bone pain syndrome.11

Although leg bone pain syndrome inkidney transplant patients who have receivedcyclosporin A is very rare, there are casereports described in the literature.4–6 To ourknowledge, this is the first case of a renaltransplant patient with pain in the lowerlimbs, related to tacrolimus treatment. Addi-tional case reports are needed to support thisassociation.

VIRGINIA VILLAVERDEMIGUEL CANTALEJO

ALEJANDRO BALSAEMILIO MARTIN MOLA

Department of Rheumatology, HospitalUniversitario La Paz, Madrid, Spain

AURELIO SANZDepartment of Nephrology, Hospital

Universitario La Paz, Madrid, Spain

Correspondence to: Dr V Villaverde, RheumatologyUnit, Hospital Universitario La Paz. Paseo de laCastellana 265, 28046 Madrid, Spain.

1 Muñoz-Gomez J, Bergada E, Gomez R, LlopartE, Subias R, Sole M. Amyloid arthropathy inpatients undergoing periodical haemodialysisfor chronic renal failure. Ann Rheum Dis1985;44:729–33.

2 Bonomini V, Bortolotti GC, Feletti C, ScolariMP. Serial histomorphometric and histochemi-cal bone biopsy studies in dialysis and trans-plantation. J Urol Nephrol 1975;12:941–50.

3 Barbosa LM, Gauthier VJ. Bone pain intransplant patients. Arthritis Rheum 1993;36(suppl):91.

4 Muñoz-Gomez J, Collado A, Gratacos J,Campistol JM, Lomeña F, Llena J, et al. Reflexsympathetic dystrophy syndrome of the lowerlimbs in renal transplant patients treated withcyclosporin A. Arthritis Rheum 1991;34:625–30.

5 Naredo E, Balsa A, Sanz A, Pantoja L, MartinMola E, Gijon Baños J. Leg bone painsyndrome due to cyclosporin A in a renaltransplant patient. Clin Exp Rheumatol 1994;12:653–6.

6 Dubost JJ, Fourcade J, Soubrier M, Ristori JM,Sauvezie B, Deteix P. Epiphyseal pain fromcyclosporine in renal transplants. Nephrologie1997;18:17–22.

7 Sugiyama E, Suzuki H, Tunru I, Yamashita N,Hori T, Kobayashi M. FK506, an immunosup-pressant partially inhibits interleukin 6 produc-tion by adherent rheumatoid synovial cells. JRheumatol 1994;21:1597–601.

8 Suzuki N, Kaneko S, Ichino M, Mihara S,Wakisaka S, Sakane T. In vivo mechanisms forthe inhibition of T lymphocyte activation bylong-term therapy with tacrolimus (FK506).Arthritis Rheum 1997;40:1157–67.

9 Kozin F, Ryan L, Carrera G, Soin J, WortmannR. The reflex sympathetic dystrophy syndrome(RSDS): scintigraphic studies, further evidencefor the therapeutic eYcacy of systemic corti-costeroid and proposed diagnostic criteria. AmJ Med 1981;70:23–30.

10 Christians U, Schmidt G, Bader A, Lampen A,Schottmann R, Linck A, et al. Identification ofdrugs inhibiting the in vitro metabolism of tac-rolimus by human liver microsomes. Br J ClinPharmacol 1996;41:187–90.

11 Redlich K, Pietschmann P, Stule T, Peterlik M.Comparative study on the eVect of calciumchannel blockers on basal and parathyroidhormone-induced bone resorption in vitro.Pharmacol Toxicol 1997;80:262–5.

Spleen haemorrhagicinfarction and hazards ofanticoagulation inWegener’s granulomatosisIn the largest cohort published to date, nosplenic involvement is described in Wegener’sgranulomatosis (WG).1 We report on twopatients who required splenectomy for symp-tomatic spleen infarction in the course ofWG.

CASE 1A 42 year old man was admitted with an eightmonth history of arthritis and lower limbdysesthesia. Examination showed an acutelyill patient with a 39°C fever, oral ulcers,haemorrhagic gingival hyperplasia, bilateralhaemorrhagic nasal discharge with crusts,diVuse necrotic purpura, neuritis, and blackdiscoloration of some fingers and toes. Thespleen was not palpable. Silent anterior myo-cardial infarction was diagnosed because ofraised MB-CK levels and ST-segment in-crease with loss of R waves in leads V1,V2,V3on electrocardiogram.2 Antineutrophil cyto-plasmic antibodies (c-ANCA) were disclosedin serum and necrotising vasculitis was shownon skin biopsy specimen.3 No antiphospholi-pid antibody or coagulation protein abnor-mality could be disclosed. Treatment con-sisted of intravenous administration ofprednisolone, cyclophosphamide, sodiumheparinate, diltiazem, dinitrosorbide andenalapril. His short-term course was un-eventful. At day 14, the patient suddenlydeveloped a severe haemorrhagic shock.Echotomography of the abdomen showed asplenic mass. At laparotomy, the spleen wasalmost disrupted by voluminous haematoma.Histological analysis of the spleen showedwidespread necrotising vasculitis with haem-orrhagic infarction. After five years of followup, the patient is in complete remission withoral cotrimoxazole treatment.

CASE 2A 23 year old young man was admitted inAugust 1996 because of repeated otitismedia, sinusitis, epistaxis, headache, arthral-gia with fever and weight loss. Despite a shortcourse of oral corticosteroids and antibiotics,his general condition worsened. Antiprotein-ase 3 c-ANCA were disclosed in serum.Chest computed tomography showed pulmo-nary nodules. Intranasal endoscopic biopsiesdemonstrated necrotising vasculitis with epi-thelioid and giant cells. Treatment includedoral prednisone and intravenous cyclophos-phamide pulses. After a few days, serum cre-atinine concentrations abruptly increased to198 µmol/l and urine analysis showed micro-scopic haematuria and proteinuria. Highdose methylprednisolone pulses were thengiven, intravenous cyclophosphamide waschanged to a 100 mg oral daily regimen andthe patient eventually achieved remission. InOctober 1996, abdomen computed tomogra-phy showed an intrasplenic lesion that wasconsistent either with a splenic infarct or hae-matoma (fig 1). The later course was markedby a WG flare in January 1997, which wascomplicated with massive thrombosis of theleft iliofemoral vein and the inferior venacava. No thrombophilic disorder could befound. Intravenous heparin then oral antico-agulation with acenocoumarol were given.Because of persistent left hypocondrium ten-derness, splenectomy was performed in Sep-tember 1997. Histological examinationshowed spleen infarction with organised hae-matoma and sequelae of vasculitis (fig 2).

COMMENT

Because they are vessels without collateralflow, occlusion of distal parenchymal splenicarteries leads invariably to splenic infarction.Of note, two of the three patients describedby Wegener in 1936 had spleen involvement.4

The frequency of spleen involvement rangesfrom 50% to 100% of WG cases atnecropsy.5–7 Histological data frequently

Figure 1 Bone scintigraphy, showing increaseduptake in both knees.

654 Letters, Matters arising

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

showed massive or multiple areas of splenicnecrosis, associated to a variable extent withcentral arteritis, splenic trabeculitis, folliculararteriolitis, disseminated parenchymatousgranulomata and capsulitis (fig 2). Patientswith splenic infarction in WG usually remainasymptomatic. Prominent splenomegaly israre.6 With computed tomography, focalsplenic infarction appears as well defined,peripheral wedge shaped areas of low attenu-ation. In WG, the diVuse vasculitis processoften results in massive hypodense lesionsinvolving the spleen parenchyma (fig 1).8–10 Aperipheral rim of enhancement may be seen,as for spleen abscess, haematoma andlymphoma.11 Splenectomy has been per-formed successfully in some patients.10 12 13

Spleen lesions may also appear to heal onrepeated computed tomography under medi-cal treatment only, consisting of prednisoneand cyclophosphamide.14

Few recent reports of splenic involvementin WG provide histological analysis from livepatients.10 12 13 Inaugural spontaneous ruptureof a normal sized spleen with only sub-capsular neutrophil infiltrate has been de-scribed in a patient who subsequently devel-oped full blown WG.15 In another case, aspontaneous splenic haemorrhage was as-cribed to vasculitis in a patient who hadsevere WG that required haemodialysis.13 Inour two patients, microscopical study of thespleen also showed haemorrhagic infarctioncaused by specific WG related vasculitis

process. A severe splenic haemorrhage oc-curred in patient 1, which was clearly relatedto both necrotising vasculitis and hypocoagu-lable state. Anticoagulation was indicated forinaugural myocardial infarction in case 1 anddeep venous thrombosis in case 2, in bothcases during active WG flare. Splenectomywas required in both our cases.

Our data suggest that antithrombotictreatment entails a specific risk of bleedingcomplications in patients with WG vasculitis.When anticoagulation is necessary in WGpatients, computed tomography of the abdo-men should be systematically performed and,if splenic infarction is disclosed, splenectomyshould be considered.

THOMAS PAPODU LE THI HUONG

JEAN-CHARLES PIETTEInternal Medicine, Hôpital Pitié-Salpêtrière,

Paris, France

MARC ANDREOLIVIER AUMAITRE

Internal Medicine, Hôpital Cébazat,Clermont-Ferrand, France

FRÉDÉRIC CHARLOTTEHistopathology, Hôpital Pitié-Salpêtrière, Paris,

France

JEAN-LOUIS KEMENYHistopathology, Hôpital Cébazat,

Clermont-Ferrand, France

Correspondence to: Dr T Papo, Internal MedicineUnit, Hôpital Pitié-Salpêtrière, 83 Boulevard del’Hôpital 75651 Paris cedex 13, France.

1 HoVman GS, Kerr GS, Leavitt RY, HallahanCW, Lebovics RS, Travis WD, et al. Wegenergranulomatosis: an analysis of 158 patients.Ann Intern Med 1992;116:488–98.

2 Papo T, Piette JC, Laraki R, Bletry O, Le ThiHuong Du, Godeau P. Silent myocardialinfarction in Wegener’s granulomatosis. AnnRheum Dis 1995;54:233–4

3 Leavitt RY, Fauci AS, Bloch DA, Michel BA,Hunder GG, Arend WP, et al. The AmericanCollege of Rheumatology 1990 criteria for theclassification of Wegener’s granulomatosis.Arthritis Rheum 1990;33:1101–7.

4 Wegener F. Uber eine eigenartige rhinogenegranulomatose mit besonderer beteiligung desarteriensystems und der nieren. Beitr PatholAnat 1939;102:36–68.

5 Fahey JL, Leonard E, Churg J, Godman G.Wegener’s granulomatosis. Am J Med 1954;17:168–78.

6 Pinching AJ, Lockwood CM, Pussel BA, ReesAJ, Sweny P, Evans DJ, et al. Wegener’sgranulomatosis: observations on 18 patientswith severe renal disease. Q J Med 1983;208:435–60.

7 Walton EW. Giant-cell granuloma of the respi-ratory tract (Wegener’s granulomatosis). BMJ1958;2:265–70.

8 McHugh K, Manson D, Eberhard BA, LaxerRM, Shore A. Splenic necrosis in Wegener’sgranulomatosis. Pediatr Radiol 1991;21:588–9.

9 Fonner BT, Nemcek AA, Boschman C. CTappearance of splenic infarction in Wegener’sgranulomatosis. AJR 1995;164:353–4.

10 Ruel M, Bobrie G, Jarrousse B, Henry-BiabaudE. Iconographie splénique au cours de lagranulomatose de Wegener. Presse Med 1989;18:725.

11 Kalaitzoglou I, Drevelengas A, Paladas P,Asimaki A. MRI appearance of pulmonaryWegener’s granulomatosis with concomitantsplenic infarction. Eur J Radiol 1998;8:367–70.

12 Morayati SJ, Fink-Bennet D. Indium-111 leu-kocyte scintigraphy in Wegener’s granulomato-sis involving the spleen. J Nucl Med 1986;27:1864–6.

13 Kettritz R, Anders S, Kettritz U, Schneider W,Göbel U, Luft FC. Spontaneous splenichemorrhage in a patient with Wegener’s granu-lomatosis. Am J Kidney Dis 1998;5:860–2.

14 Gregorini G, Campanini M, Tira P, Lancini L,Tincani A, Majorca R. Spleen involvement inWegener’s granulomatosis: two case reports.APMIS 1990;S19:23.

15 Franssen CFM, Ter Maaten JC, Hoorntje SJ.Spontaneous splenic rupture in Wegener’s vas-culitis. Ann Rheum Dis 1993;52:314.

Amiodarone inducedlupusLupus related to amiodarone has not previ-ously been described. We report on a patientwho developed drug induced lupus (DIL) inassociation with amiodarone treatment. Toour knowledge, this is the first report ofamiodarone induced lupus (CD ROM:Medline, USA National Library: 1966–98).

A 71 year old white woman was admittedbecause of two weeks of pleuritic chest pain,dyspnea on exertion, and non-productivecough. She had malaise, intermittent fever,arthralgia, and weight loss for more than sixmonths. There was no history of Raynaud’sphenomenon, oral ulcers or photosensitivity.She had a six year history of arterialhypertension and atrial fibrillation treatedwith amiloride, digital and amiodarone (200mg two times daily) for the past two years.

Physical examination disclosed malar rash,an aortic systolic murmur (grade II/VI), andhypoventilation in both pulmonary bases.Laboratory studies showed an erythrocytesedimentation rate of 90 mm 1st h. Peripheralblood examination revealed a mild normo-chromic and normocytic anaemia (10 g/dl),normal white blood cells count (4000/µl),with lymphopenia (20 per cent), and normalplatelets count (180 000/µl). Coagulationtests were normal. All serum chemistries,including thyroid function tests, creatininephosphokinase, immunoglobulins, comple-ment levels, and urine analysis were withinnormal limits. Coombs’s tests were negative.Rheumatoid factor was 1:320. Circulatingimmune complexes (IgG-C1q) were positive.Antinuclear antibodies (ANAs) were positiveat 1:640; anti-Ro, anti-La, anti-dsDNA, anti-Sm, anti-histone antibodies, antiphospho-lipid antibodies, cryoglobulins, C reactiveprotein, VDRL and Mantoux test were nega-tive. Blood and urine cultures were negative.Electrocardiogram was within normal limits,and the two dimensional echocardiogramshowed mild aortic stenosis. Chest radiogra-phy revealed bilateral pleural eVusions, with-out fibrosis or cardiomegaly. Pleural fluid wasexudative, with lymphocytic predominance,without cytological features for malignancy.Cultures of pleural fluid for bacteria, includ-ing for Mycobacterium tuberculosis, were nega-tive. Bone scan with technetium-99m showedincrease uptake in hands, elbows, and knees.The histopathological examination of biopsyspecimens of the skin, including indirectimmunofluorescence stain, muscle and tem-poral artery did not show abnormal features.

The amiodarone was stopped and thepatient progressively improved. No cortico-steroids were given. On the third week shedeveloped a transient relapse, with fever,malaise and with evidence of unilateral pleu-ral eVusion. One year after no clinical,analytical or radiological findings werepresent, and three years later she stillremained free of symptoms, and the ESR,complete blood count, and radiological datawere normal. The titre of ANA decreased butremained weakly positive at 1:40.

It is estimated that 3–7% of all patientswith systemic lupus erythematous (SLE)might have DIL.1 Clinical features of SLEand DIL are similar but there are certain dis-tinguishing characteristics between them: the

Figure 1 Contrast enhanced upper abdominalcomputed tomography: large hypodense areawith a peripheral rim of normally enhancingtissue.

Figure 2 Cut gross pathological section ofspleen shows changes corresponding to those seenby computed tomography. Large firm yellow(white) area consistent with infarction wassurrounded by dark peripheral zone of splenicparenchyma.

Letters, Matters arising 655

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

patients with DIL are usually older; theprevalence of men and women is similar; andthe presenting symptoms are usually mild,with the patient usually complaining ofmalaise, fever and arthralgia, with or withoutarthritis, while skin, central nervous system orrenal involvement is rare. Pleuropericardialdisease is frequent and, as in classic SLE,anaemia and leucopenia may be present.Serum complement components are usuallynormal, ANAs are positive but anti-dsDNAand anti-Sm are negative, while anti-histonesantibodies can be detected in most ofpatients.2

The pathogenic mechanisms proposed forDIL include: cross reactivity between drugand the nucleic acid; hapten complex forma-tion between drug and nucleic acid, or struc-tural damage to the chromosomal DNA;action of drug as an adjuvant or immuno-stimulant, which, in concert with appropriateimmune response genes, triggers polyclonalB/T cell activation; and interference with thecomplement pathway.1

The incidence of side eVects associatedwith amiodarone ranges from 40% to 93%and, in most of cases, these side eVects areconsequence of its potential to be directlytoxic to several organ systems.3 4 However,there is also some evidence of immunologi-cally mediated phenomena related to amio-darone. A positive skin and basophil degranu-lation tests with amiodarone, secretion ofleucocyte inhibitory factor, positive lympho-blastic transformation and circulation of aspecific antibody of the IgG class have beendescribed.5 6 Moreover, several studies sug-gest that various biological and immunologi-cal markers of “systemic” disease activity arepresent in patients taking this drug. Circulat-ing immune complexes, ANAs, and non-specific increase in ESR and white blood cellcount, sometimes with eosinophilia, are com-mon findings.7 8

Low ANA titre is not uncommon in anelderly patient. However, spontaneous SLEin elderly people is not usual and DIL mustalways be considered in the diVerential diag-nosis. This case, presenting with malaise,fever, arthralgia, circulating immune com-plexes, and autoantibodies strongly suggestsan immunological underlying condition.Moreover, this patient meets four SLE cri-teria: malar rash, serositis, haematologicaldisorder (lymphopenia), and positive ANAstest. Imputability criteria of amiodaroneinduced lupus are present on a semiologicalbasis with classic features of DIL and on achronological basis with disappearance ofmost of the symptoms after amiodaronewithdrawal. The relapse could be explainedbecause of the long elimination half time ofthe drug9 and, in consequence, the immuneresponse might progress despite discontinua-tion of the treatment.

RUI SUSANOLUIS CAMINALDAVID RAMOS

BERNARDINO DÍAZDepartment of Medicine, Hospital Central de

Asturias (Universitary Centre), Oviedo, Spain

Correspondence to:Dr R Susano, Internal Medicine, Hospital Centralde Asturias, C/ Julian Claveria s/n, 33006, Oviedo,

Spain.

1 Faubert PF, Porush JG, Venkataseshan VS,Grishman E. Lupus-like syndromes. In: Grish-man E, Churg J, Needle MA, VenkataseshanVS, eds. The kidney in collagen-vascular diseases.New York: Raven Press, 1993:87–119.

2 Dubois EL, Wallace DJ. Drugs that exacerbateand induce systemic lupus erythematosus. In:Wallace DJ, Dubois EL, eds. Dubois’s lupus ery-thematosus. Philadelphia: Lea and Febiger,1987:450–60.

3 Raeder EA, Podrid PJ, Lown B. Side eVects andcomplications of amiodarone therapy. AmHeart J 1985;109:975–83.

4 Martin WJ, Rosenow EC. Amiodarone pulmo-nary toxicity. Recognition and pathogenesis(Part I). Chest 1988;93:1067–74.

5 Akoun GM, Gauthier-Rahman S, Milleron BJ,Perrot JY, Mayaud CM. Amiodarone-inducedhypersensivity pneumonitis. Evidence of animmunological cell-mediated mechanism.Chest 1984;84:133–5.

6 Fan K, Bell R, Eudy S, Fullenwider J.Amiodarone-associated pulmonary fibrosis.Evidence of an immunologically mediatedmechanism. Chest 1987;94:625–30.

7 Manzano L, Yebra M, Merino F, Lladós G,Diego F, Durántez A. Afección pulmonar porla amiodarona y anticuerpos antinucleares. RevClin Esp 1986;179:80–1.

8 Harris L, McKenna WJ, Rowland E, Holt DW,Storey GC, Krikler DM. Side eVects of longterm amiodarone therapy. Circulation 1983;67:45–51.

9 Andreasen P, Agerback H, Bjerregard P,Gotzsche H. Pharmacokinetics of amiodaroneafter intravenous and oral administration. Eur JClin Pharmacol 1981;19:293–9.

Antinuclear antibodies inrelapsing polychondritisThe prevalence of antinuclear antibodies(ANA) in relapsing polychondritis (RP) hasbeen recently reported by Zeuner et al to beas high as 66%,1 usually in a low titre with aspeckled pattern. We report here on ourexperience of ANA testing in patients withRP.

The charts of 180 patients followed up inour institution fulfilling the criteria for RPproposed by Michet et al2 have been recentlyretrospectively reviewed, with special focuson dermatological manifestations and theirrelation with myelodysplasia.3 4 This aim ledus to exclude 36 patients because the associa-tion of RP with potentially confoundingdiseases,3 such as systemic lupus erythemato-sus (SLE) present in nine, mixed connectivetissue disease (MCTD) in five, rheumatoidarthritis in three, Takayasu arteritis in three,mesenteric panniculitis in three, spondyloar-thropathy in two, Crohn’s disease in two,psoriasis in two, or Lichen planus in two.Among the 144 patients remaining, 111 havebeen tested for ANA by using either Hep-2cells or liver sections as substrate, or both.Most patients had repeated ANA determina-tions, including initial testing before onset ofcorticosteroid treatment, and the higher titreswere considered for analysis. Figure 1 showsthe histogram of ANA positivity according toANA titre.

ANA were either absent or present in lowtitres in a majority of patients (73% and 18%,respectively). “Significant” titres—that is,titre > 1/100 ANA—were demonstrated inonly 10 of 111 patients (9%). The pattern offluorescence was as follows: homogeneous infive, homogeneous and speckled in two,speckled and nucleolar in one, perinuclearand nucleolar in one, and not provided inone. Among those 10 patients, five had clini-cal or ophthalmological features, or both,suggestive of an associated Sjögren’s syn-drome including two who had antibodies toboth SS-A and SS-B, and two others had amyelodysplastic syndrome.4 None of these 10patients had antibodies to ds-DNA. Usingthe same methods, ANA > 1/100 were foundin 15 of the 36 patients initially excluded(42%). Among the nine with SLE, ANA >1/100 and antibodies to ds-DNA (by Farrassay or Crithidia fluorescence) were foundin eight patients. All five patients withMCTD had ANA > 1/1000 (in a speckledpattern in four), with positive antibodies toRNP and negative tests for ds-DNA. The tworemainders had either rheumatoid arthritis,or Lichen planus associated with multinodu-lar goitre; both of them also had features sug-gestive of sicca syndrome. Therefore, withinthese 36 patients, strong ANA positivity wasmainly restricted to SLE, MCTD or Sjö-gren’s syndrome, or all three.

Beside the recent article by Zeuner et al,1

the prevalence of ANA has rarely beenreported in RP. McAdam et al found positiveANA in four of their 18 patients tested(22%), and noted 3 of 23 (13%) ANApositivity in a literature review.5 Data regard-ing ANA were not provided in the large seriesof patients with RP followed up at the MayoClinic.2 6 The low prevalence of ANA ob-served in our cases with “pure” RP—that is,RP not associated with another connectivetissue disorder except for a possible Sjögren’ssyndrome—agrees with the negative results oftests for IgG antinucleosome antibodiesrecently reported by our group in thiscondition.7

We conclude that: (a) the prevalence ofANA observed in RP is low and, (b) assuggested by other authors6 8 the finding of asignificant titre of ANA in a patient with RPstrongly suggests the presence of an associ-ated disorder, such as SLE, MCTD, Sjö-gren’s syndrome or acquired myelodysplasia.

JEAN-CHARLES PIETTEROULA EL-RASSIZAHIR AMOURA

Internal Medicine Unit, HôpitalPitié-Salpêtrière, 83 Boulevard de l’Hôpital,

75651 Paris Cedex 13 France

Figure 1 Prevalence of ANA in “pure” RP according to ANA titre.

90

70

80

60

40

50

30

20

0

10

ANA titre

Pre

vale

nce

(%

)

Negative ≤ 1/100 ≤ 1/1000 > 1/1000

656 Letters, Matters arising

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

Correspondence to: Dr J-C Piette.

1 Zeuner M, Straub RH, Rauh G, Albert ED,Schölmerich J, Lang B. Relapsing polychondri-tis: clinical and immunogenetic analysis of 62patients. J Rheumatol 1997;24:96–101.

2 Michet CJ Jr, McKenna CH, Luthra HS,O’Fallon WM. Relapsing polychondritis. Sur-vival and predictive role of early disease mani-festations. Ann Intern Med 1986;104:74–8.

3 Piette JC, El-Rassi R, Wechsler B, Laporte JL,Rybojad M, Francès C. Dermatological mani-festations in relapsing polychondritis. Associa-tion with myelodysplasia. [Abstract]. ArthritisRheum 1996;39 (suppl 9):S230.

4 Piette JC, Papo T, Chavanon P, Du LTH,Francès C, Godeau P. Myelodysplasia andrelapsing polychondritis. J Rheumatol 1995;22:1208–9.

5 McAdam LP, O’Hanlan MA, Bluestone R,Pearson CM. Relapsing polychondritis. Pro-spective study of 23 patients and a review of theliterature. Medicine (Baltimore) 1976;55:193–215.

6 Isaak BL, Liesegang TJ, Michet CJ Jr. Ocularand systemic findings in relapsing polychondri-tis. Ophthalmology 1986;93:681–9.

7 Amoura Z, Cacoub P, Chabre H, Amoura I,Wechsler B, Musset L, et al. Prevalence of anti-nucleosome IgG in connective tissue diseases.[Abstract]. Lupus 1998;7 (suppl 1):82.

8 Trentham DE, Le CH. Relapsing polychondri-tis. Ann Intern Med 1998;129:114–22.

There is no associationbetween polymyalgiarheumatica and acuteparvovirus B19 infectionParvovirus B19 has been associated with agrowing number of diseases. Besides thefrequent manifestations such as erythemainfectiosum, aplastic crisis in persons withunderlying haemolytic anaemia, hydrops feta-lis in pregnant women and acute or chronicarthritis a range of rather rare diseases havebeen described in recent reports.1 Amongthem are case reports on persistent parvovirusB19 infection in immune incompetent people,encephalitis, myocarditis, systemic lupus ery-thematosus (reviewed by Anderson andYoung1) and rheumatoid arthritis.2 Further-more, parvovirus B19 has been suspected toplay a part in the aetiology of polymyalgiarheumatica (PMR).3 4 Because of the acuteonset of PMR and its systemic symptoms aninfectious cause may be a relevant factor.Additionally, autoimmune processes havebeen demonstrated in both, PMR and parvo-virus B19 infection.5 6 As the receptor for par-vovirus B19, the P-blood group antigen(globoside), is also present on endothelial cellsan interrelation between parvovirus B19 andgiant cell arteritis or PMR may be possible.

Parvovirus B19 can only replicate in eryth-roid precursor cells in human bone marrow,but it is known that infection of cellsnon-permissive for viral replication leads toan excess production of the viral non-structural protein (NS1) without productionof capsid proteins.7 As the NS1 protein iscytotoxic and able to induce apoptosis, itprobably plays a part in the pathogeneticprocess of the parvovirus B19 induced tissuedamage. This is confirmed by the fact thatantibodies against NS1 of parvovirus B19 arepreferentially produced during chronic orpersistent parvovirus B19 infections, forexample in parvovirus B19 associatedchronic arthritis.8 9

To test the hypothesis whether PMR isassociated with acute parvovirus B19 infec-tion, we tested the seroprevalence of IgGantibodies against the two structural proteinsVP1 and VP2 and against the non-structuralprotein NS1 in 110 PMR patients (patients

with giant cell arteritis excluded; mean(SEM) age 67.0 (0.8) years, range: 48—77)and, for comparison, in 135 healthy controlsof diVerent ages. At the time point of bloodsampling (median disease duration at thetime point of blood sampling: 0.6 years,range: 0—7.3, mean (SEM): 1.4 (0.2) years),35 patients had no corticosteroids and 75patients received on an average 15.2 (1.8) mgprednisolone/day. Furthermore, we investi-gated the interrelation between age,symptoms10 or laboratory parameters and thepresence of NS1 specific antibodies inhealthy controls and patients with PMR.Non-parametric Kruskal-Wallis one wayanalysis was used to compare means ofdiVerent subgroups. The significance levelwas p<0.05.

Subjects in the control group had variousages between 18 to 75 years. Overall sero-prevalence of IgG against the capsid proteinsVP1 and VP2 was 78% (fig 1). Overall IgGseroprevalence against VP1 and VP2 was88% in patients with PMR (not significantlydiVerent versus the age matched controlgroup). With respect to the NS1 IgGantibody, overall seroprevalence in the con-trol group was 22% (fig 1) and in patientswith PMR 20% (p=0.057 versus the agematched control group: 10%; fig 1).

Furthermore, we investigated the associa-tion between the presence of NS1 IgGantibodies and PMR related symptoms10 orlaboratory parameters (patients with NS1 ascompared with patients without NS1 IgGantibodies were not diVerent in age, sex, andmedication). The symptoms were assessedusing standard record forms from the medi-cal histories (at the time serum wascollected).10 We asked the patients for muscu-lar pain in the left/right shoulder, left/rightupper arms, left/right neck, left/right glutealmuscle, and left/right thigh. If one musclegroup was painful, the corresponding itemwas scored with one point (the sum of theitem points was the overall muscle score).10 InPMR patients with NS1 IgG as comparedwith patients without NS1 IgG, arthralgiawas more frequent (with versus without: 73%v 40%, p=0.024). However, the overall mus-cle score was lower in NS1 positive than inNS1 negative patients (0.5 (0.2) SEM v 1.6(0.3) SEM score points; p=0.021). Withrespect to other PMR related symptoms, no

significant diVerences were found. In patientswith a positive NS1 IgG antibody, interleukin6 (4.6 (0.9) SEM v 11.3 (2.2) SEM;p=0.037) and soluble ELAM (48.2 (4.8)SEM v 71.4 (5.2) SEM; p=0.024) weresignificantly lower as compared with patientswithout NS1 IgG. No significant diVerenceswere found with respect to erythrocytesedimentation rate, C reactive protein, tu-mour necrosis factor, interleukin 2, andinterleukin 1â.

In view of these data, there was a positiveassociation between NS1 and arthralgia.However, other symptoms and parameters ofinflammation such as erythrocyte sedimenta-tion rate were not associated with thepresence of NS1 IgG. Furthermore, diseaserelated immune mediators such as interleukin6 or soluble ELAM were lower in patientswith as compared with patients without NS1IgG. As a positive NS1 IgG titre indicates anactive infection, an acute parvovirus B19infection does not seem to be a pathogeneticfactor in our patients with PMR.

ANDREA HEMAUERSUSANNE MODROW

Institute for Medical Microbiology and Hygiene,University Medical Centre Regensburg,

Germany

JOACHIM GEORGIDepartment of Internal Medicine, Ostseeklinik,

Damp, Germany

KLAUS HELMKEDepartment of Internal Medicine IV, Hospital

Munich-Bogenhausen, Germany

PETER VAITHDepartment of Internal Medicine, University

Medical Centre Freiburg, Germany

BERNHARD LANGRheumazentrum Baden-Baden, Germany

JÜRGEN SCHÖLMERICHRAINER H STRAUB

Department of Internal Medicine I, UniversityMedical Centre Regensburg, Germany

Correspondence to: Dr R H Straub, Department ofInternal Medicine I, University Medical Centre,D-93042 Regensburg, Germany.

1 Anderson LJ, Young NS Human parvovirusB19. In: Parks WP, ed. Monographs in virology.Vol 20. New York: Karger, 1997.

2 Murai C, Munakata Y, Takahashi Y, Ishii T,Shibata S, Muryoi T, et al. Rheumatoid arthri-tis after human parvovirus B19 infection. AnnRheum Dis 1999;58:130–2.

3 Cimmino MA. Genetic and environmental fac-tors in polymyalgia rheumatica. Ann RheumDis 1997;56:576–7.

4 Cimmino MA, Negrari A, Grazi G, Franconi G,Accardo S. Parvovirus B19 infection in poly-myalgia rheumatica (PMR). [Abstract]. Arthri-tis Rheum 1997;40:238.

5 Kerr JR, Boyd N. Autoantibodies following par-vovirus B19 infection. J Infect 1996;32:41–7.

6 Soloninka CA, Anderson MJ, Laskin CA. Anti-DNA and antilymphocyte antibodies duringacute infection with human parvovirus B19. JRheumatol 1989;16:777–81.

7 Liu JM, Green SW, Shimada T, Young NS. Ablock in full length transcript maturation incells nonpermissive for B19 parvovirus. J Virol1992;66:4686–92.

8 von Poblotzki A, Gigler A, Lang B, Wolf H,Modrow S. Antibodies to parvovirus B19 NS-1protein in infected individuals. J Gen Virol1995;76:519–27.

9 von Poblotzki A, Hemauer A, Gigler A,Puchhammer-Stöckl E, Heinz FX, Pont J, et al.Antibodies to the nonstructural protein of par-vovirus B19 in persistently infected patients:Implications for pathogenesis. J Infect Dis1995;172:1356–9.

10 Straub RH, Herfarth HH, Rinkes B, KonecnaL, Glück T, von Landenberg P, et al. Favorablerole of interleukin 10 in patients with polymy-algia rheumatica. J Rheumatol 1999;26:1318−25.

Figure 1 Seroprevalence of IgG antibodiesagainst VP1/VP2 (black bars) and NS1 (openbars) in control subjects and patients withpolymyalgia rheumatica (PMR). The numberof healthy controls in each subgroup of normalsubjects was at least 20. *p=0.057 vseroprevalence of IgG NS1 in age matchedhealthy controls (groups 56–65 and 66–75 with42 controls v 110 patients with PMR).

100

80

60

40

20

0–25

26–3

5

36–4

5

46–5

5

Age (y)

*

56–6

5

66–7

5PM

R0

Ser

op

reva

len

ce (

% o

f al

l hea

lth

yco

ntr

ols

in t

he

sam

e ag

e g

rou

p)

Letters, Matters arising 657

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

Neutrophil chemotaxis inBehçet’s syndromeIt has been suggested that the marked cellularinflammatory response, which characterisesBehçet’s syndrome (BS), may be attributableto increased neutrophil locomotion.1–4 How-ever, others disagree.5 6 We have re-evaluatedchemotaxis of polymorphonuclear leucocytes(PMNs) in BS among a greater number ofpatients in a controlled setting.

Fifty four male BS patients, nine malepatients with ankylosing spondylitis, eightwith psoriasis and 37 male healthy controlswere studied with 28 female patients with BSand 16 healthy female controls. Behçetpatients with severe disease were those withactive major vessel and/or eye involvement.

We measured chemotaxis with the “underthe agarose method”.7 8 The measurementswere masked with the assessors not knowingthe diagnoses. An inverted microscope fittedwith an ocular micrometre disc to measurethe migration of neutrophils from middlewells to outer (chemotaxis) and inner wells(chemokinesis) was used. Zymosan activatedsera (patients or controls) were used as asource of C5a. Results were expressed asmicrometre square (1 mm=8 squares). Addi-tionally the plates were evaluated macro-scopically for observation of the migrationbetween neutrophil wells.

Tables 1 and 2 show the results. There wereno significant diVerences between the chemo-tactic indices of the various groups of patientsand controls studied of either sex. Maximalchemotaxis rates in the groups varied from67% to 100%.

The Boyden millipore filter system hasextensively been used for chemotaxisexperiments.9 10 The agarose method is sim-ple and cheap. This method can preferentiallybe used to diVerentiate chemokinesis fromchemotactic migration.7

There is marked heterogeneity in diseaseexpression in men and women in BS11 and wereasoned that some of the confusion in theliterature about neutrophil activity might berelated to this. Thus we analysed our dataseparately for either sex. Although there was atendency for male patients with severe diseaseto have higher chemotactic indices this wasnot statistically significant (p=0.62). We didnot study any diseased controls for femalepatients with BS.

Abdulla and Lehner6 observed decreasedchemotaxis in BS. Fordham et al,2 on the otherhand reported increased chemotaxis, but nor-mal random migration. While Wilkinson,5

similar to our experience, observed normalchemotaxis in BS, more recently Carletto et al3

reported augmented chemotaxis especially inthe active phases of the disease. Finally, BenEzra et al, among a group of Behçet patientswith uveitis could demonstrate increasedchemotactic activity only among a few of thesepatients, compared with that observed amongpatients with other forms of uveitis. They con-cluded that increased chemotactic activity was

not a regular feature of ocular BS (personalcommunication).

In vivo assays do not diVerentiate chemo-taxis from chemokinesis. In the Carletto studyclinically active Behçet patients demonstratedincreased chemotaxis to sera by Senn’s modi-fied in vivo assay.3 Others had found hyperche-motaxis to neutrophil cytoplasmic fractionsagain by using an in vivo assay.4 Although it isdiYcult to compare the results of in vitro andin vivo assays, we thought these reportedincreases might have resulted from increasedchemokinesis. In our experiments we observedmaximal chemotaxis (3 mm) frequently, how-ever we did not find any significant diVerencesin chemotactic indices between diseased andhealthy subjects.

An interesting aspect of our study was themigration between neutrophil wells that wasobserved in many of the Petri dishes. This wasobserved even though we had not used cellularmaterials as chemotactic agents. Presumablythe gravity of the cellular materials overcamethe chemical gradient of zymosan activatedsera in some Petri dishes. Because of theobserved migration between neutrophil wells,we suggest that there should be only one “tri-ple well rank” in a Petri dish. On the otherhand our method of preincubation of thewhole blood for 45 minutes at 37°C beforeharvesting the PMNs (intended for betterviability) might have been responsible for thisphenomenon by increasing the chemotacticactivity in all groups studied. Further studiesare needed to clarify these issues.

BINNUR TÜZÜNDepartment of Dermatology, Trakya University

Medical Faculty, Edirne, Turkey

YALÇIN TÜZÜNCEM MAT

Department of Dermatology, Istanbul UniversityCerrahpasa Medical Faculty, Istanbul, Turkey

SEBAHATTIN YURDAKULVEDAT HAMURYUDAN

HASAN YAZICIDepartment of Rheumatology, Istanbul

University Cerrahpasa Medical Faculty,Istanbul, Turkey

YILMAZ ÖZYAZGANDepartment of Ophthalmology, IstanbulUniversity Cerrahpasa Medical Faculty,

Istanbul, Turkey

Correspondence to: Dr B Tüzün, BüyükciftlikSokak Belde Apt 24/3, Nisantasi, Istanbul 80200,Turkey.

Supported by TÜBITAK (Turkish Scientific andTechnical Research Council) (TAG 754).

1 Efthimiou J, Addison IE, Johnson BV. In vivoleukocyte migration in Behçet’s syndrome. AnnRheum Dis 1989;48:206–10.

2 Fordham JN, Davies PG, Kirk A, Currey HL.Polymorphonuclear function in Behçet’s syn-drome. Ann Rheum Dis 1982;41:421–5.

3 Carletto A, Pacor ML, Biasi D, Caramaschi P,Zeminians S, Bellavite P, et al. Changes of neu-trophil migration without modification of invitro metabolism and adhesion in Behçet’s dis-ease. J Rheumatol 1997;24:1332–6.

4 Takeuchi A, Kobayashi K, Mori M, MizushimaY. The mechanism of hyperchemotaxis inBehçet’s disease. J Rheumatol 1981;8:40–4.

5 Wilkinson PC. Comments on enhanced leuko-cyte locomotion in Behçet’s disease. In: Lehner

T, Barnes CG, eds. Recent advances in Behçet’sdisease. London: Royal Society of MedicineServices, International Congress and Sympo-sium Series, no 103, 1986:97–8.

6 Abdulla YH, Lehner T. The eVect of immunecomplexes on chemotaxis in Behçet’s syn-drome and recurrent oral ulcers. In: Lehner T,Barnes CG, eds. Behçet’s syndrome. London:Academic Press, 1979:55–75.

7 Nelson RD, Bauman MP, Gracyk JL, FiegelVD, Herron MJ. Leukocyte chemotaxis: migra-tion under agarose method. In: Douglas SD,Quie PG, eds. Investigation of phagocytes indisease. Edinburgh: Churchill-Livingstone,1981:20–31.

8 Nelson RD, Quie PG, Simmons RL. Chemo-taxis under agarose: A new and simple methodfor measuring chemotaxis and spontaneousmigration of human polymorphonuclear leuco-cytes and monocytes. J Immunol 1975;115:1650–6.

9 Fernandes AC, Anderson R, Ras GJ. An objec-tive filter-based, enzymatic method for the invivo measurement of the migration of humanpolymorphonuclear leukocytes. J ImmunolMethods 1985;83:259–71.

10 Fernandez HN, Henson PM, Otani A, HugliTE. Chemotactic response to human C3a andC5a anaphylatoxins. J Immunol 1978;120:109–15.

11 Yazici H, Tüzün Y, Pazarli H, Yurdakul S,Özyazgan Y, Özdogan H, et al. Influence of ageof onset and patient’s sex on the prevalence andseverity of manifestations of Behçet’s syn-drome. Ann Rheum Dis 1984;43:783-7.

MATTERSARISING

Neuropsychiatric systemiclupus erythematosusThe considerable diYculties in making senseof the literature on patients with lupusinvolving the central nervous system arere-emphasised in the paper by Rood et al.1

The authors, who to be fair take a sensiblycautious approach to their results, neverthe-less seek to persuade us that the IL10 locus isassociated with neuropsychiatric lupus on thebasis of a historical case notes review of 42lupus patients with neuropsychiatric disease,compared with 50 who lack such involve-ment.

Their conclusion needs to be treated withcaution. Does it make sense to lump together42 highly diverse patients and make the kindof claim they have made? The authors suggestthat CNS lupus is attributable to eitherantiphospholipid antibody related throm-botic events, or “immune mediated” disease.This division is artificial. There is a consider-able literature on CNS lupus that proposesthat a wide variety of immunopathogenicmechanisms may be responsible in individualcases. These mechanisms include thromboticeVects, which may be linked to antiphos-pholipid antibodies, a true vasculitis, a crossreaction between antibodies that recognisethe lymphocyte surface targets and neuro-logical antigens, and antibodies to a widevariety of neurological targets. A considerablylarger number of patients will have to bestudied before any claims of links to an IL10promoter haplotype can be truly convincing.

We agree with the authors that patientswith SLE have a higher innate production ofIL10 than controls. However, as there is nosignificant diVerence in the frequency of theIL10 promoter single nucleotide polymor-phisms (SNP)s in SLE patients when

Table 1 Chemotactic indices in men*

Groups Number Mean (SD)

Behçet’s 54 6.4 (8.0)Severe Behçet’s 9 11.3 (12.1)Ankylosing spondylitis 9 5.9 (7.1)Psoriasis 8 10.8 (11.0)Healthy controls 37 10.0 (10.3)

*Kruskal-Wallis one way analysis of variance,corrected for ties: ÷2=2.1381; DF=4; p>0.05.

Table 2 Chemotactic indices in women*

Groups Number Mean (SD)

Behçet’s 28 7.2 (8.2)Healthy controls 16 10.1 (10.2)

*t=1.032, p>0.05.

658 Letters, Matters arising

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

compared with controls in their study, wesuspect that the diVerence in IL10 produc-tion is not attributable to functional diVer-ence between patients with SLE and controlsin terms of the IL10 SNP allelesfrequencies.2–4 DiVerences have been de-scribed with respect to microsatellites5 andone awaits confirmation from other popula-tions or family studies. To our knowledge, adiVerence in IL10 production between pa-tients with neuropsychiatric disease SLE andnon-neuropsychiatric disease SLE has notbeen described. The described associationswould be biologically meaningless if IL10production is similar between these twogroups.

The authors suggest that the -1082A alleleis associated with a higher innate IL10production, however, they appear to ignorethe only published study to date that showedthat the A allele was associated with lowerIL10 production.6 In addition we haveconfirmed that the A allele is associated withlower IL10 production in transient transfec-tion studies and the ATA/ATA genotype isassociated with lower IL10 production inwhole blood culture.4 The increase in the Aallele is mainly accounted for by an increasein the ATA haplotype in their neuropsychiat-ric disease patients and therefore they aredescribing an association with a low IL10producing haplotype, not a high IL10 pro-ducing haplotype. One interpretation of thiswould be that patients with neuropsychiatricdisease symptoms are unable to adequatelycontrol inflammation from a variety of diVer-ent pathological mechanisms because of lowIL10 production.

DAVID ISENBERGCentre for Rheumatology, UCLH, London

ESTHER CRAWLEYDepartment of Molecular Pathology, UCLMS

PAT WOODepartment of Molecular Pathology, UCLMS

1 Rood MJ, Keijsers V, van der Linden MW, TongTQT, Borggreve SE, Verweij CL, et al.Neuropsychiatric systemic lupus erthematosusis associated with imbalance in interleukin 10promoter haplotypes. Ann Rheum Dis1999;58:85–9.

2 Mok CC, Lanchbury JS, Wai Chan D, Sing LaoC. Interleukin 10 promoter polymorphisms insouthern chinese patients with systemic lupuserythematosus. Arthritis Rheum 1998;41:1090–5.

3 Lazarus M, Hajeer AH, Turner D, Sinnott P,Worthington J, Ollier WER, et al. Genetic vari-ation in the interleukin 10 gene promoter andsystemic lupus erythematosus. J Rheumatol1997;24:2314–17.

4 Crawley E, Kay R, Sillibourne J, Patel P, Hutch-inson I, Woo P. Polymorphic haplotypes of theinterleukin-10 5 ’flanking region determinevariable interleukin-10 transcription and areassociated with particular phenotypes of juve-nile rheumatoid arthritis. Arthritis Rheum (inpress).

5 Mehrian R, Quismorio FP, Strassmann G,Stimmler MM, Horwitz DA, Kitridou RC, etal. Synergistic eVect between IL-10 and bcl-2genotypes in determinig susceptibility to sys-temic lupus erythematosus. Arthritis Rheum1998;41:596–602.

6 Turner DM, Williams DM, Sankaran D,Lazarus M, Sinnott PJ, Hutchinson IV. Aninvestigation of polymorphism in theinterleukin-10 gene promoter. Eur J Immuno-genet 1997;24:1–8.

Authors’ replyWe thank Drs Isenberg, Crawley and Woo fortheir interest in our paper.1

They argued that the dichotomy of thepathogenesis of CNS lupus in “immune

mediated” and tromboembolic disease is toorigid, because a wide variety of immuno-pathogenetic mechanisms can be deemedresponsible for CNS lupus. As the hallmarkof SLE is the production of autoantibodies, itseems to be justified to assume that thepathogenesis underlying CNS lupus is B cellmediated. Based upon this assumption weclustered the individual neuropsychiatric dis-ease SLE patients and tested the hypothesisthat a genetic marker in the promoter of theIL10 gene is associated with the phenotype ofCNS-SLE.

In general, a positive result in a geneticassociation study is only possible after a cor-rect definition of the phenotype. After all, ifthe phenotype is inadequately defined, themagnitude and statistical significance of theassociation will be less or lost because of therandom distribution of the genetic marker inthe misclassified patients. If misclassificationoccurred in the sense that CNS lupuspatients were attributed to the non-neuropsychiatric disease SLE population, thefact that we still found a positive resultstrengthens our conclusions instead of weak-ens it.

It might be argued that thromboembolicevents do not fit in the pathogenetic model ofB cell mediated CNS lupus. But, as statedclearly in the article, even after exclusion ofthese ambiguous patients, the distribution ofthe frequencies in the neuropsychiatric dis-ease SLE and non-neuropsychiatric diseaseSLE patients remains the same.

Of course we agree with the notion that ourfindings must be repeated in another group ofpatients. Interestingly, the increased preva-lence of ATA in neuropsychiatric disease SLEpatients has already been reported by Mok ina group of Chinese SLE patients.2 Currentlywe are investigating the distribution of theIL10 promoter haplotypes of neuropsychiat-ric disease SLE patients in an ethnicallydiVerent population.

In our article we have elaborated on twopossible explanations of our findings. Firstly,the increased frequency of the ATA haplotypemight be associated with an increased pro-duction of IL10. We made this assumption inthe light of previous studies stating that SLEas a whole is characterised by an increasedinnate IL10 production.3 4 It is wrong toextrapolate these conclusions to our popula-tion. Because of the retrospective character ofour study, we were not able to measure IL10production in our populations and thereforecannot say whether or not IL10 production inour SLE patients as a whole was similar to ordiVerent from the control population. Itmight well be that diVerences in IL10production would only emerge after stratify-ing into neuropsychiatric disease SLE andnon-neuropsychiatric disease SLE patients.Furthermore, it might be that in the popula-tions mentioned before, there was an excessof patients with neuropsychiatric diseaseSLE.

The second explanation for the skewingfound in IL10 promoter polymorphismsmight be that the increased susceptibility toneuropsychiatric disease SLE in the ATApatients is not conferred via an increasedIL10 production at all, but that it is merely amarker for the real neuropsychiatric diseaseSLE susceptibility allele.

It is not clear whether or not IL10promoter SNPs are associated with low orhigh IL10 production, because of the am-biguous reports in the literature. In our labo-ratory, the -1082 A allele has been found to

be associated with high IL10 production.5 Inthis light we have speculated about the possi-ble link of high IL10 production and thepathogenesis of neuropsychiatric diseaseSLE. Isenberg et al have referred to anothergroup stating that -1082 A is associated witha low ex vivo IL10 production and they inter-pret our results with this finding in mind.6 Inconclusion, we do not know the relevance ofthe IL10 promoter in the in vivo regulation ofIL10 production and therefore both explana-tions are equally speculative.

M J ROODT W J HUIZINGA

Leiden University Medical Centre, Departmentof Rheumatology, Building 1, C4-R, PO Box

9600, 2300 RC Leiden, the Netherlands

1 Rood MJ, Keijsers V, Van der Linden MW, TongTQT, Borggreve S, Verweij CL, et al. Neu-ropsychiatric systemic lupus erythematosus isassociated with imbalance in interleukin10promoter haplotypes. Ann Rheum Dis1999;58:85–9.

2 Mok CC, Lanchbury JS, Wai Chan D, Sing LauC. Interleukin-10 promoter polymorphisms insouthern Chinese patients with systemic lupuserythematosus. Arthritis Rheum 1998;41:1090–5.

3 Llorente L, Richaud-Patin Y, Fior R, Alcocer-Varela J, Wijdenes J, Morel-Fourrier B, et al. Invivo production of interleukin-10 by non-Tcells in rheumatoid arthritis, Sjogren’s syn-drome, and systemic lupus erythematosus.Arthritis Rheum 1994;37:1647–55.

4 Hagiwara E, Gourley MF, Lee S, Klinman DM.Disease severity in patients with systemic lupuserythematosus correlates with an increasedratio of interleukin-10: interferon gammasecreting cells in the peripheral blood. ArthritisRheum 1996;39:379–85.

5 Keijsers V, Verweij CL, Hazes JMW, Westen-dorp RGJ, Breedveld FC, Huizinga TWJ.Innate diVerences in IL-10 production arepresent at the level of transcription and associ-ated with haplotypes: association of IL-10 hap-lotypes and rheumatoid arthritis. [Abstract].Clin Exp Rheumatol 1998;16:200.

6 Turner DM, Williams DM, Sankaran D,Lazarus M, Sinnott PJ, Hutchinson IV. Aninvestigation of polymorphisms in theinterleukin-10 gene promoter. Eur J Immuno-genet 1997;24:1–8.

A man with intermittentfever and arthralgiaKnight and Symmons report a very interest-ing case of a man with Whipple’s disease andprovide us with a timely update on this rarecondition.1 They describe how six monthsafter initial presentation the diagnosis of adultonset Still’s disease (AOSD) was made anddespite regular review at several centres, thisdiagnosis was upheld for a further five years.Multiple investigations were performed add-ing little to the original diagnosis. It is notmentioned whether a serum ferritin wastaken. This may have been useful given theinitial diagnosis of AOSD as it might haveresulted in questioning this diagnosis, permit-ting an earlier diagnosis of Whipple’s disease.

It is widely reported in the literature, as farback as 1975,2 that increased serum ferritinmay be of use in both diagnosis3–5 and moni-toring of AOSD.6 7 Hyperferritinaemia is nothowever exclusive to AOSD, as variousmalignancies, hepatic necrosis andhaemachromatosis can all cause it. However,values up to 25 000 µg/l have been recordedin AOSD and reportedly, values rarely exceed3–5000 µg/l in the above conditions. Thepathogenesis is not clearly understood, but ithas been hypothesised that in AOSD,cytokine upregulation of ferritin mRNA

Letters, Matters arising 659

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from

translation may occur.8 This compares withthe iron regulated pathway of ferritin synthe-sis in haemachromatosis and iron overloadsyndromes.

A comparative study of diagnostic criteriain AOSD by Mason et al9 suggest theYamaguchi criteria are superior to the otherstested, including Cush et al quoted by Knightand Symmons. However, none of the criteriato aid diagnosis make use of serum ferritinmeasurement despite the claims for its use inthe literature and acceptance in clinical prac-tice. Although undoubtedly useful if veryhigh, it is not clear what the relevance of anormal value in AOSD is, in a case satisfyingclinical diagnostic criteria (although we havenever seen such a case). In rare diseases suchas AOSD, it is diYcult to assess and evaluatediagnostic criteria and calculate sensitivityand specificity of possible disease markers. Ifserum were stored on this patient it would beinteresting to know the serum ferritin meas-urement and how, if at all, it would haveaVected this patient’s management.

MARK QUINNANDREW GOUGH

Department of Rheumatology, Old Home, LeedsGeneral Infirmary, Great George Street, Leeds,

West Yorkshire, LS1 3EX.

Correspondence to: Dr M A Quinn.

1 Knight SM, Symmons PM. A man withintermittent fever and arthralgia. Ann RheumDis 1998;57:711–14.

2 Jacobs A, Worwood M. Ferritin in serum; clini-cal and biochemical implications. N Engl JMed 1975;292:951–6.

3 Gonzalez-Hernandez T, Martin-Mola E,Fernandez-Zamirano A, Balso-Criado A,Miguel-Mendieta E. Serum ferritin can be use-ful for diagnosis in adult onset Still’s disease. JRheumatol 1989;16:412–13.

4 Ota T, Higashi S, Suzuki H, Eto S. Increasedserum ferritin in adult Still’s disease. Lancet1987;i:562–3.

5 Schiller D, Mittermayer H. Hyperferritinaemiaas a marker of Still’s disease. Clin Infect Dis1998;26:534–5.

6 Akritidis N, Giannakakis I, Giouglis T. Ferritinlevels and response to treatment in patientswith adult onset Still’s disease. J Rheumatol1996;23:201–2.

7 Schwarz-Eywill M, Heilig B, Bauer H, BreitbartA, Pezzutto A. Evaluation of serum ferritin as amarker for adult onset Still’s disease activity.Ann Rheum Dis 1992;51:683–5.

8 Rogers JT, Bridges KR, Durmowicz GP, Glass J,Auron PE, Munro HN. Translational Controlduring the acute phase response. Ferritinsynthesis in response to interleukin-1. J BiolChem 1990;265:14572–8.

9 Masson C, Loet X, Liote F, Dubast JJ, BoissierM-C, Perroux-Goumy L. Comparative studyof 6 types of criteria in adult Still’s disease. JRheumatol 1996;23:495–7.

Authors’ replyWe thank Drs Quinn and Gough for theirinterest in our paper. Our patient did have hisserum ferritin measured in 1992. It was 197µg/l (normal range 15–200). This was there-fore a situation in which the patient satisfiedclinical diagnostic criteria for adult onsetStill’s disease (AOSD) but had a normal fer-ritin concentration. As the authors point out,had the ferritin concentration been high, thiswould have helped to confirm the diagnosisbut given that it was in the normal range, itcould not actually be used to refute the diag-nosis. It was always felt that this patient’s dis-ease was not typical of AOSD and the variousphysicians who looked after the patient werealways willing to consider alternatives. How-ever, it is diYcult, even with the benefit ofhindsight, to conclude that Whipple’s diseasecould have been diagnosed earlier. Althoughthe normal serum ferritin was not in keepingwith the diagnosis of AOSD it did not pointtowards any other diagnosis in particular.

DEBORAH SYMMONSS M KNIGHT

Rheumatology, Macclesfield District GeneralHospital, Victoria Road, Macclesfield,

Cheshire SK10 3BL

Ear, ear, what’s going onin Norfolk?Having recently started work in the rheuma-tology department of the Norfolk and Nor-wich Hospital I read with great interest thearticle on Hug(h)e(s’) ears: an unusualpresentation.1 Amazingly we have recentlyseen an almost identical, but less catastrophiccase.

A 27 year old white man presented with a24 hour history of ears so swollen andpainful, that he could not lie in bed with themtouching the pillow. One year previously hewas diagnosed with primary antiphospholipidsyndrome (APLS) after recurrent deep veinthromboses and a raised IgG anticardiolipin

antibody at 92 iu/ml. He was subsequentlygiven warfarin.

On admission this time, his INR, whiletaking warfarin 7 mg per day, was subthera-peutic at 1.6. Biopsy from his left ear lobeshowed numerous fibrin thrombi with noassociated vasculitis, consistent with throm-bosis secondary to APLS.

His warfarin dose was increased to obtainan INR between 3 and 4. Within a few dayshe had recovered and was discharged homewell.

Perhaps Hughes’ ears should be renamedNorfolk ears?

EMMA CLARKKARL GAFFNEY

PETER MERRYDepartment of Rheumatology, Norfolk and

Norwich Hospital, Brunswick Road, NorwichNR1 3SR

1 O’Gradaigh D, Scott DG, Levell N. Hug(h)e(s’)ears: an unusual presentation of antiphospholi-pid syndrome. Ann Rheum Dis 1999;58:65–6.

Authors’ replyWe note with interest the report from DrGaVney. While the coincidence is indeedcurious, these cases do suggest possiblemechanisms for activation of thrombosis.The external ear is characterised particularlyby a lower average temperature than corebody temperature, and by its susceptibility totrauma and pressure eVects. In our case, cry-oglobulins were not identified, and nocomment in this regard is made by theauthors. There is no specific reference to anyaural trauma, though presumably, as in ourcase, it is diYcult to assess what pressure wasexerted on the external ear during sleep. It isplausible that such pressure causes a degreeof blood stasis, which together with inad-equate anticoagulation, resulted in thrombo-sis. Such speculation may be interesting, butit is this latter point that deserves emphasis—patients with antiphospholipid syndromewho have had thrombi will do so again,potentially with serious consequences, if theINR is not scrupulously maintained above3.0, a message that must be spread widely:“Friends, Norfolk countrymen, lend me yourears!”

DONNCHA O’GRADAIGHDAVID SCOTT

Department of Rheumatology, Norfolk andNorwich Hospital, Brunswick Road, Norwich

NR1 3SR

660 Letters, Matters arising

on Decem

ber 29, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.58.10.654 on 1 O

ctober 1999. Dow

nloaded from