Annals of the Rheumatic Diseases, 1977, 36, 428432

Ulceration of the cornea in rheumatoid arthritisMALCOLM I. V. JAYSON,' AND DAVID L. EASTY2

From the Department of Medicine, University of Bristol, and the Royal National Hospital for RheumaticDiseases, Bath,' and Bristol Eye Hospital,2 Bristol

SUMMARY Five patients with melting of the cornea in association with rheumatoid arthritis are

described. The arthritis was often inactive and without systemic manifestations, in contrast to thatseen in association with scleritis. In 3 there was evidence of reduced tear formation, but in none

was tear production absent. In 3 patients the lesions healed during treatment with azathioprine or

penicillamine.

Scleritis is a well recognized complication ofrheumatoid arthritis (Jayson and Jones, 1971;McGavin et al., 1976; Watson and Hayreh, 1976).Sometimes inflammation from the sclera can spreadinto and involve adjacent cornea (scierokeratitis);this may be associated with marginal thinning.However, disease of the cornea can occur alonewithout involvement of the sclera with little or noinflammatory change. Occasionally marginal thin-ning may be gross with corneal melting (keratolysis)which may proceed to perforation. The associationof keratolysis with rheumatoid arthritis is wellrecognized by ophthalmologists (Brown and Grayson,1968; Watson, 1975), but not by rheumatologistsand no attention has been paid to the rheumato-logical features of the disease. We present data on 5patients with marginal gutter ulcers of the cornea inassociation with rheumatoid arthritis.

Case reports

CASE 1This woman had a 22-year history of rheumatoidarthritis starting in the fingers and wrists and lateraffecting the feet, ankles, and knees. She had beenmaintained on methylprednisolone 4 mg a day forseveral years and a variety of anti-inflammatorydrugs. By November 1975 she showed gross changesof advanced inactive rheumatoid arthritis affectingthe proximal interphalangeal (PIP), metacarpo-phalangeal (MCP), and wrist joints, both shoulders,the metatarsophalangeal (MTP) joints of both feet,and both knees with fixed flexion deformities. Therewas no synovial thickening nor increased temperature

Accepted for publication March 3, 1977Correspondence to Prof. M. I. V. Jayson, Department ofRheumatology, University of Manchester, Hope Hospital,Eccles Old Road, Salford, Manchester M6 8HD

over any of the joints. Investigations showed anormal plasma viscosity, haemoglobin, and whiteblood count. Rose-Waaler and latex tests werenegative, as were tests for antinuclear factor andother autoantibodies; no LE cells were found.X-rays showed gross changes of rheumatoidarthritis.

In November 1974 she had been seen elsewherewith a complaint of dryness of the eyes, but not ofthe mouth, and was given artificial tears. In January1975 the corrected visual acuity was 6/60 in theright eye and 6/36 in the left; there was some sore-ness and irritation of both eyes, and peripheralcorneal thinning was noted, more marked on theright than the left. There was minimal evidence ofinflammation with only slight hyperaemia of thesclera in the area adjacent to the cornea. In February1975 the cornea perforated (Fig. 1) and was repairedwith a peripheral annular corneoscleral graft. Theinitial result was satisfactory but 2 weeks aftersurgery the donor tissue became infiltrated withinflammatory cells. It was thought that the appear-ance was consistent with an acute homograftreaction. In order to keep the donor tissue inposition, a conjunctival pedicle flap was fashionedand sutured over the grafted area. Further meltingof central corneal tissue occurred at the edge of theconjunctival flap.

Similar corneal thinning developed in the left eye(Fig. 2) which failed to respond to intensive localtherapy including subconjunctival heparin but notsteroids. Eventually the cornea perforated withprolapse of the iris. In April 1975 this was repairedby mobilizing a flap from the adjacent sclera andoverlaying this with a conjunctival bridge flap. InNovember 1975 the scleral flap retracted andperforation seemed imminent. During these surgicalprocedures, peripheral corneal melting progressed

428

copyright. on F

ebruary 25, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.36.5.428 on 1 O

ctober 1977. Dow

nloaded from

Ulceration of the cornea in rheumatoid arthritis 429

remorselessly in both eyes despite 2-hourly treatmentwith 1-cysteine eye drops.

Although the rheumatoid arthritis seemed com-pletely inactive, she was started on azathioprine50 mg bd (2- 5 mg/kg) and her progress was closelyfollowed by slit lamp examinations. Over the suc-ceeding weeks the melting ceased, and by June 1976re-epithelialization and stromal regeneration hadoccurred over the previously repaired iris prolapsein the left cornea (Fig. 3). The final visual acuity inthe right eye was no better than 'counting fingers',

I,

3

5

a

but with the left eye an acuity of 6/12 and N5 wasachieved.

CASE 2A less severe example of marginal corneal ulcerationoccurred in this 75-year-old man with a 2-yearhistory of rheumatoid arthritis affecting principallythe PIP and MCP joints of both hands and bothwrists. Plasma viscosity was raised at 2 42 cP,haemoglobin 12-5 g/dl, and white cell count8- 3 x 109/l. Rhematoid factor was present in a titre

2

4

6

IFig. 1 Case 1. Right cornea, showing prolapsed iris through a perforated marginal ulcer. An annular corneo-scleral graft was performed.Fig. 2 Case 1. Left eye, showing gross marginal corneal thinning. Perforation occurred and the prolapsed iris wascovered by operation with a conjunctival bridge flap.Fig. 3 Case 1. Left cornea. The conjunctival flap has retracted leaving an area of dense scarring surrounding acentral area of stromal regeneration. The corneal melting process is no longer active at this stage.Fig. 4 Case 3. Right cornea at presentation, showing active scleritis, peripheral corneal thinning, and inflam-matory cell infiltration.Fig. 5 Case 3. Right cornea showing further progress in the degree of marginal ulceration.Fig. 6 Case 4. Diffuse corneal thinning which first occurred 15 months after a cataract extraction. A centralisland of normal cornea is shown.

copyright. on F

ebruary 25, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.36.5.428 on 1 O

ctober 1977. Dow

nloaded from

430 Jayson, Easty

of 1: 512 but tests for antinuclear factor and otherautoantibodies were negative. X-ray of the handsshowed typical rheumatoid changes.For one year before presentation he had been seen

in the ophthalmic clinic for progressive marginalthinning affecting the periphery of the lower half ofeach cornea. There was minimal associated inflam-mation in the sciera, and Schirmer's test showedsome reduction of tear secretion, being 5 mm oneach side.He was treated with D-penicillamine 250 mg a

day for the first month, 500 mg for the secondmonth, and 750 mg for the third month. By the endof 3 months the eye changes had improved dramatic-ally. The left cornea showed little evidence ofmarginal thinning, and in the right eye re-epithe-lialization of the thinned areas was developing. Atthe same time the rheumatoid arthritis improvedconsiderably.

CASE 3In March 1976 this 73-year-old man developed atransient inflammatory polyarthritis affecting thePIP and MCP joints of both hands, and theMTP joints of both feet and both ankles. Plasmaviscosity was raised at 1-90 cP with a normalhaemoglobin and white cell count. Rose-Waaler testwas positive at a titre of 1: 128; antinuclear factorand other autoantibody tests were negative. Hisarthritis persisted for several months. But by June1976 there was little clinical evidence of jointdisease.

In May 1976 he presented to the ophthalmicdepartment with marginal thinning of the superiorpart of the right cornea. There was some associatedvascularization and patchy inflammatory cell in-filtration. In contrast to the 2 patients alreadydescribed, there was an active scleritis adjacent tothe corneal disease (Fig. 4). Tear secretion wasreduced, Schirmer's test for the right eye being4 mm and for the left 2 mm. Rose bengal showedconsiderable punctate staining more marked on theright than on the left.The corneal and scleral disease showed no response

to intensive topical therapy with 1-cysteine andsteroid drops in either eye, and the marginal thinningprogressed with increasing local inflammation(Fig. 5). At the same time the patient becamegenerally unwell, with a leucocytosis and a pro-ductive cough. Examination showed the changes ofchronic bronchitis and emphysema. Despite in-tensive investigations, no other cause for his illnesswas discovered, and his condition improved onantibiotic therapy.

Because of the progressive corneal change, he was4_+--...-A -_+4;: N C __ A -.fl. C __ /1-1

but after several weeks there was no evidence ofarrest of the corneal thinning and scleritis. Sub-sequently he was again treated with topical corti-costeroid drops (Guttae Betnesol) and this timeshowed a rapid improvement.

CASE 4

This 78-year-old man had a 12-year history ofrheumatoid arthritis affecting the PIP and MCPjoints in both hands, both wrists, shoulders, andfeet with gross hallux valgus and MTP subluxation.He was not receiving any medication. Examinationshowed advanced but inactive disease. Plasmaviscosity was raised at 2 12 cP with a haemoglobinof 1 lg/dl and white cell count of 5. 5 x 109/l. Rheum-atoid factor, antinuclear factor, and other auto-antibody tests were negative. X-ray showed advancedrheumatoid changes.

In December 1974 he had bilateral extractions ofsenile cataracts without operative or postoperativecomplications. In March 1976 he presented withbilateral corneal melting with marked marginalguttering but no evidence of inflammatory changes(Fig. 6). On repeated observations the cornealmelting was not progressive, and, in view of his ageand general infirmity, positive therapy was notthought indicated. He is being followed up atfrequent intervals.

CASE 5

This 82-year-old man was referred to the ophthalmicclinic with bilateral corneal thinning. Although hehad no rheumatic complaints examination did showchanges of trivial inactive polyarthritis affecting theMCP joints of both hands and both knees withrheumatoid nodules over both elbows. Plasmaviscosity, haemoglobin, and white cell count werenormal; rheumatoid factor, antinuclear factor, andother autoantibodies were absent.The limbal guttering involved the whole of the

periphery of each cornea, and in the right eye therewas a marked degree of melting on the nasal side.There weie many punctate inflammatory cellinfiltrates in the stroma on each side and someepiscleral hyperaemia on the nasal Fide of the rightcernea. He had previously shown an adverseresponse to topical steroid drops, in that they hadapparently increased the melting process. Other localtherapy had not helped. In view of this, treatmentwith azathioprine 175 mg/day (2- 5 mg/kg) wasintroduced for a period of 2 months. Though theimmediate clinical response was not dramatic,there was eventual improvement, with disappear-ance of the conjuctival hyperaemia and corneal

copyright. on F

ebruary 25, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.36.5.428 on 1 O

ctober 1977. Dow

nloaded from

Discussion

Damage to the cornea in rheumatoid arthritis mayresult from an adjacent scleritis with sclerosingkeratitis-peripheral inflammatory cell infiltrationand vascularization followed by permanent scar-

ring-localized to the site nearest the inflamedsclera (McGavin et al., 1976). Limbal gutteringoccurs in association with rheumatoid scleritis, butmay also develop in the apparent absence of scleralinflammation. It is possible that in some of thesepatients the scleritis was suppressed by local or

systemic steroids. Patients may present with marginalthinning with no inflammatory changes except fora few dilated adjacent blood vessels (Watson, 1975).Occasionally the entire superficial part of theremaining central island of corneal stroma mayslough away. Progressive melting away of the corneamay lead to perforation. If the anterior segment isreplaced by grafting with homologous material, thegraft may melt away in the same way (Brown andGrayson, 1968).

In patients with scleritis associated with rheumatoidarthritis the arthritis is usually severe with nodulesand erosions with positive tests for rheumatoidfactor and a low haemoglobin. Extra-articularmanifestations of rheumatoid disease, particularlyvasculitis, are common (Jayson and Jones, 1971). Incontrast, the patients in this series showed a widerange of expression of rheumatoid disease, but insome, although there were advanced deformities, thearthritis was inactive without systemic manifestationsand with negative tests for rheumatoid factor.The relationship of keratolysis with Sjogren's

syndrome is confused. Bloch et al. (1965) found a

history of corneal ulceration in 3 of 63 patients withSjogren's syndrome. Gudas et al. (1973) found onecase, and Krachmer and Laibson (1974) recorded 6patients with this combination. Some of the patientsin the present series showed evidence of reducedlacrimal secretion but in none was tear productiontotally absent, whereas many patients with Sjogren'ssyndrome have much more severe limitation of tearformation without corneal ulceration. There arecommonly other clinical manifestations of Sjogren'ssyndrome, such as xerostomia, but none was foundin the present series. Moreover, a variety of auto-antibodies are found in Sj6gren's syndrome patients,but these tests were usually negative in the patientswith keratolysis. Circulating antibodies to cornealtissue have been found in patients with idiopathiccomeal ulceration-Mooren's ulcer (Schaap et al.,1969). Whether this is relevant to the corneallesions in association with rheumatoid arthritis isuncertain.Two types of pathological change have been

Ulceration of the cornea in rheumatoid arthritis 431

described in patients with marginal guttering(Iwamoto et al., 1972). The first is inflammatorywith marked vascularization in and around thecorneal lesion and dilatation of the adjacent con-junctival blood vessels. Often these patients aretreated with local steroid eye drops which reducethe inflammation, but the cornea continues to meltaway. Electron microscopy of the cornea showschanges resembling hypersensitivity reactions withaccumulation of lymphocytes, neutrophils, andoedema, and fibrinoid necrosis ofsmall blood vessels.The second type is the so-called quiescent group,with much less vascularization and no evidence ofinflammation. The electron microscopic appearancesshow fatty degenerative changes in the stroma. Thisdescription of the pathological changes must becompared with the classification of Aronson et al.(1970) who suggested that inflammatory diseases ofthe cornea could be separated into an infiltrativegroup secondary to some inflammatory process, andan ischaemic group which they divided into a primarysubgroup which followed obstruction of bloodvessels, possibly by deposition of immune complexesand intravascular clotting, and a secondary groupfollowing infection, trauma, or chemical damage.There is now good evidence that whatever the

cause, the melting of the cornea is mediated bycollagenases. Berman et al. (1971) showed theproduction of collagenases in rabbit eyes, andsubsequently (Berman et al., 1973a) in humanmaterial. These enzymes cleave the tropocollagenmolecule into two components containing three-quarters and one-quarter of the original moleculerespectively. It seems that corneal fibroblasts canproduce a proenzyme which is activated by lysosomalenzymes or trypsin to develop collagenolyticactivity (Hook et al., 1973). This collagenaseactivity is inhibited by ac-2 macroglobulins withwhich they form complexes (Berman et al., 1975).It is of interest that some ophthalmolgists thinkthat serum will inhibit corneal ulceration as ac-2macroglobulins are present in serum (Berman et al.,1973b). oc-2 macroglobulins are normally present intears, perhaps explaining the association of marginaldegeneration with reduced tear production and whythe corneal ulceration was not relieved by the use ofartificial tears. Berman et al. (1973b) experimentallyfound a relationship between a-2 macroglobulinactivity in tears and the extent of corneal ulcers.With regard to treatment of corneal ulceration,

the evidence is confused by the failure to examineseparately corneal ulceration of different causes.Hook et al. (1973) thought that local steroidswould potentiate collagenolytic activity, so pro-moting keratolysis. Several of the patients in thepresent series had received steroids systemically and

copyright. on F

ebruary 25, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.36.5.428 on 1 O

ctober 1977. Dow

nloaded from

432 Jayson, Easty

locally, and it is not possible to know whether thisproduced or exacerbated the corneal disease orwhether it was simply the patients with most severedisease who were receiving more aggressive formsof therapy. It is our impression that topical steroidshad a deleterious effect in one patient and arecontraindicated over a long period. However, in onepatient improvement only occurred with localsteroid therapy. Understanding of the part playedby collagenases in the pathogenesis of cornealmelting has led to trials of cysteine, acetylcysteine,and penicillamine drops locally (Slansky andDohlman, 1970; Francois et al., 1973; Mehra andSingh, 1975) as they are thought to inhibit collagen-ase activity. Although these agents seem effective inexperimental trials, they proved disappointing inthe present study. Elliott et al. (1972) proposed theuse of local heparin for peripheral corneal ischaemiclesions. In the one patient who received this treat-ment, we found it of no value. In any event, ontheoretical grounds it seems unlikely to help. Afurther form of therapy is to apply contact lenseswhich protect the ulcerated area and act as anartificial epithelium. They may lead to healing of thedefect beneath.

Operative measures may not be successful. It iswell known that peripheral ulceration can recur inpatients who have cornea scleral replacements(Watson, 1975), and in addition an enhanced re-jection ofdonor tissue has been noted. Attempts havealso been made to cover corneal perforations withconjunctival flaps or scleral autografts, but theresults have not been encouraging. Many suchprocedures result in visual loss. However, a report ofgood healing in several types of progressive marginalulceration after excision and recession of adjacentlimbal conjunctiva has recently appeared (Wilsonet al., 1976). We feel that iris prolapse is not neces-sarily an indication for surgery, as a thin layer ofcorneal tissue can reform over the knuckle of irisand afford protection against pathogenic micro-organisms.We found 3 patients with progressive disease

which came under control only when systemictherapy was started with azathioprine or penicil-lamine. In one patient with scleritis and cornealulceration there was no response to azathioprine.In any uncontrolled small series of this type it isimpossible to draw valid conclusions, but we suggestthat with incipient or repeated perforation of thecornea it is worth trying these agents.

ReferencesAronson, S. B., Elliott, J. H., Moore, T. E., Jr., and O'Day,D. M. (1970). Pathogenetic approach to therapy of peri-pheral corneal inflammatory disease. American Journal ofOphthalmology, 70, 65-90.

Berman, M., Dohiman, C. H., Gnadinger, M., and Davison,P. (1971). Characterisation of collagenolytic activity in theulcerating cornea. Experimental Eye Research, 11, 255-257.

Berman, M. B., Kerza-Kwiatecki, A. P., and Davison,P. F. (1973a). Characterisation of human corneal col-lagenase. Experimental Eye Research, 15, 367-373.

Berman, M. B., Barber, J. C., Talamo, R. C., and Langley,C. E. (1973b). Corneal ulceration and the serum anti-proteases. I. cxl-antitrypsin. Investigative Ophthalmology,12, 759-770.

Berman, M., Gordon, J., Garcia, L. A., and Gage, J. (1975).Corneal ulceration and the serum antiproteases. II. Com-plexes of human corneal collagenases and o-macro-globulins. Experimental Eye Research, 20, 231-244.

Bloch, K. J., Buchanan, W. W., Wohl, M. J., and Bunim,J. J. (1965). Sjogren's syndrome. A clinical, pathologicaland serological study of sixty-two cases. Medicine, 44,187-231.

Brown, S. I., and Grayson, M. (1968). Marginal furrows.A characteristic corneal lesion of rheumatoid arthritis.Archives of Ophthalmology, 79, 563-567.

Elliott, J. H., Aronson, S. B., Moore, T. E., and Williams,F. C. (1972). Heparin therapy of peripheral cornealischemic syndromes. Symposium on the Cornea (Trans-actions of the New Orleans Academy of Ophthalmology),pp. 78-104. Ed. by R. Castroviejo. Mosby, St. Louis.

Francois, J., Cambie, E., Feher, J., and Van den Eeeckhout,E. (1973). Collagenase inhibitors (penicillamine). Annalsof Ophthalmology (Chicago), 5, 391-408.

Gudas, P. P., Jr., Altman, P., Nicholson, D. H., and Green,W. R. (1973). Corneal perforations in Sjogren syndrome.Archives of Ophthalmology, 90, 470-472.

Hook, R. M., Hook, C. W., and Brown, S. I. (1973). Fibro-blast collagenase; partial purification and characterisatlon.Investigative Ophthalmology, 12, 771-776.

Iwamoto, T., De Voe, A. G., and Farris, R. L. (1972).Electron microscopy in cases of marginal degeneration ofthe cornea. Investigative Ophthalmology, 11, 241-257.

Jayson, M. I. V., and Jones, D. E. P. (1971). Scleritis andrheumatoid arthritis. Annals of the Rheumatic Diseases,30, 343-347.

Krachmer, J. H., and Laibson, P. R. (1974). Corneal thinningand perforation in Sjogren's syndrome. American Journalof Ophthalmology, 78, 917-920.

McGavin, D. D. M., Williamson, J., Forrester, J. B., Foulds,W. S., Buchanan, W. W., Dick, W. C., Lee, P., MacSween,R. N. M., and Whaley, K. (1976). Episcleritis and scleritis.A study of their clinical manifestations in association withrheumatoid arthritis. British Journal of Ophthalmology,60, 192-226.

Mehra, K. S., and Singh, R. (1975). Cysteine in cornealulcer. I. An experimental study. Annals of Ophthalmology,7, 1329-1331.

Schaap, 0. L, Feltkamp, T. E. W., and Breebaart, A. C.(1969). Circulating antibodies to corneal tissue in a p3tientsuffering from Mooren's ulcer (ulcus rodens corneae).Clinical and Experimental Immunology, 5, 365-370.

Slansky, H. H., and Dolhman, C. H. (1970). Collagenase andthe cornea. Survey of Ophthalmology, 14, 402-416.

Watson, P. G. (1975). Connective tissue disorders and theeye. Recent Advances in Ophthalmology, pp. 214-277.Ed. by P. D. Trevor-Roper. Churchill Livingstone,Edinburgh and London.

Watson, P. G. and Hayreh, S. S. (1976). Scleritis andepiscleritis. British Journal of Ophthalmology, 60, 163-191.

Wilson, F. M., Grayson, M., and Ellis, F. D. (1976). Treat-ment of peripheral corneal ulcers by limbal conjuncti-vectomy. British Journal of Ophthalmology, 60, 713-719.

copyright. on F

ebruary 25, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.36.5.428 on 1 O

ctober 1977. Dow

nloaded from