Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Brit. Y. Ophthal. (1976) 6o, I92

Episcleritis and A study of their clinical manifestations and association with

D. D. M. McGAVIN, J. WILLIAMSON, J. V. FORRESTER, AND W. S. FOULDS Tennent Institute of and Southern General Hospital, Glasgow W. W. BUCHANAN, W. C. DICK, P. LEE, R. N. M. MACSWEEN, AND K. WHALEY Centre for Rheumatic , Glasgow

Sir William Read, Royal Ophthalmologist during Jayson and Jones, 197I), and with this in mind the the reign of Queen Anne (1702-17I4), who inci- present study differentiates rheumatoid episcieritis dentally was unable to read or write (Guthrie, and scleritis from non-rheumatoid and 1958), gave a choice description of the dangers of scleritis according to the criteria for 'classical' or perforation in deep scleritis: 'For it is most certain 'definite' rheumatoid arthritis (Ropes, Bennett,

that this horny membrane . . . by how much Cobb, Jacox, and Jessar, I959). deeper the blister is hidden in the membrane . . . in danger to make an ulceration by breaking Material and methods through the membrane, whereupon may ensue an utter loss and decay of all the humours'. Further Examination of patients was carried out at two centres: documentation of the sequelae of scleral inflamma- the Centre for Rheumatic Diseases, Glasgow, and the tion awaited a further century for the description of Tennent Institute of Ophthalmology, Western Infir- staphylomata by Scarpa (i8i8), who described two mary, Glasgow. During the years I965 to 1973 any patient seen at cases of posterior staphylomata found at necropsy. the Centre for Rheumatic Diseases with rheumatoid Mackenzie (I830) attributed scleritis to atmospheric arthritis who had any complaint was referred to an conditions and suggested the term 'sclerotitis atmos- ophthalmologist and any case of episcleritis or scleritis pherica' or 'rheumatic ophthalmia'. Later Macken- was examined in detail. Those patients examined before zie was to suggest the name 'sclerotitis idio- to or (I854) I97I who were still alive were recalled hospital http://bjo.bmj.com/ pathica'. Wolfe (I882), lecturing to the students visited in their homes. It was at the beginning of 1971 of Anderson College, Glasgow, used the currently that a scleritis clinic was instituted at the Tennent accepted nomenclature of episcleritis and scleritis. Institute, and a pro forma outlined the investigations to The later terminology has been excellently recorded be documented for patients at both centres. In all Sevel (I967). patients these investigations included examination by by an ophthalmologist and by a rheumatologist; this was Earlier authors often associated of supplemented by laboratory investigations appropriate the with rheumatic (Mackenzie, to ocular inflammation and rheumatoid disease. I830; Littell, I840; Wolfe, I882; von Arlt, i885). The purpose of the study was to compare patients on October 2, 2021 by guest. Protected copyright. Besides 'rheumatism', scleritis was attributed to who presented with rheumatoid disease, who subse- (Cleobury, i826; Walton, I875), mercury quently developed episcleritis and scleritis, with patients poisoning (Morgan, I839), venereal disease, parti- who presented with episcleritis or scleritis in whom cularly (Morgan, I839; von Arlt, I885); rheumatoid disease was incidental to their referral. scrofula (Mackenzie, I854); excessive exposure to varying atmospheric conditions (Mackenzie, I830; CENTRE FOR RHEUMATIC DISEASES de Schweinitz, I899), and menstrual derangement (de Schweinitz, I899). Between I965 and 1973, 42Io patients with rheumatoid The most consistent disorder associated arthritis were examined at the Centre for Rheumatic systemic Diseases. These patients satisfied the diagnostic criteria with episcleritis and scleritis, particularly the latter, of the American Rheumatism Association (Ropes and is rheumatoid arthritis (Van der Hoeve, 1934; others, 1959). Seven of these patients had episcleritis Verhoeff and King, 1938; Smoleroff, 1943; Edstrom and 28 had scleritis. and Osterlind, I948; Mundy, Howard, Stillman, and Bevans, 1951; Hollenhorst and Henderson, I951; Watson and Lobascher, I965; Watson, TENNENT INSTITUTE OF OPHTHALMOLOGY I966; Lyne and Pitkeathly, I968; Fowler, 1970; During 197I and 1972, 62 patients were examined at the scleritis clinic: 35 of these patients had episcleritis Address for reprints: J. Williamson, Tennent Institute of Ophthal- and 27 had scleritis. Two of the patients with episcleritis mology, Glasgow and nine of the patients with scleritis had rheumatoid Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 193

arthritis according to the criteria already cited (Ropes predominantly superficial inflammation affecting pri- and others, 1959). marily the episclera will be contrasted with features of The presentation of results initially gives the incidence deeper involvement affecting primarily the sclera. of episcleritis and scleritis in rheumatoid arthritis, and Ophthalmic examination took note of any corneal also the incidence of rheumatoid arthritis in episcleritis complications, anterior and , particularly and scleritis. Thereafter, unless it is otherwise indicated, posterior subcapsular cataract. The fundi were examined the patients are divided into those with rheumatoid after instillation of cyclopentolate hydrochloride BP disease and those without rheumatoid disease. This I per cent. A Haag-Streit applanation tonometer re- results in four groups of patients with: corded the in patients seen in i. Rheumatoid episcleritis (nine cases) hospital, and a Perkins hand-held applanation tono- 2. Rheumatoid scleritis (37 cases) meter was used for those patients who were visited in 3. Non-rheumatoid episcleritis (33 cases) their homes. 4. Non-rheumatoid scleritis (i8 cases) Any suggestion of limitation of extraocular muscle Comparisons are made between these four groups of movement or history of prompted referral to patients. In the section dealing with the systemic the orthoptist for a more exact record. manifestations of rheumatoid disease and also in some All patients were screened for evidence of kerato- of the laboratory investigations a further comparison is sicca (KCS) by Schirmer's tear test made, that is, between the combined groups of rheu- (Schirmer I) using standardized filter-paper strips*. matoid episcleritis and scleritis and a group of rheuma- The length of wetting of the filter paper at the end of toid controls, patients with rheumatoid arthritis matched 5 minutes was measured. A normal tear secretion was for age and sex with the rheumatoid episcleritis/scleritis accepted as I5 mm wetting or more for patients under groups but without episcleral or scleral inflammation. the age of 40 years, and io mm wetting for patients This gives two further groups: over the age of 40 years, as standardized by Halberg 5. Rheumatoid episcleritis/scleritis (46 cases) and Berens (I967). All patients with reduced tear flow 6. Rheumatoid controls (46 cases) had a drop of rose bengal i per cent instilled into the Statistical analysis is presented throughout the study. eye, and punctate staining indicated the presence of However, the numbers, in what is a relatively uncommon dead or dying cells and excess mucus. Confirmation of type of ocular inflammation, are often very limited. The KCS was given by Schirmer's test (Schirmer II) in biological meaning of statistical evidence therefore which a io per cent ammonia solution is held approxi- requires critical appraisal. mately 15-2 cm below the nose to stimulate tear produc- tion, which is recorded as in Schirmer I. OPHTHALMIC EXAMINATION Complaints of eye symptoms were elicited with as little GENERAL EXAMINATION prompting as possible. The duration of the condition The following investigations were designed to measure http://bjo.bmj.com/ was estimated from the first episode of inflammation. the severity of any disease of the locomotor system, and Accurate documentation of duration is difficult as it is complications of rheumatoid arthritis and other connec- not always certain that a condition which may persist tive-tissue disorders. intermittently during a period of some years has truly For each patient were recorded four grades of func- settled. Visual acuity was assessed by Snellen's test tional incapacity (Steinbrocker, Traeger, and Batterman, chart at the Tennent Institute with or without glasses I949). The articular index, a summation index, esti- as appropriate. Refraction was not done routinely. mated the number of involved joints and the severity of The anterior segment of each eye was examined in their involvement as judged by tenderness to pressure on October 2, 2021 by guest. Protected copyright. diffuse light, often in ordinary daylight, and by detailed over the joint margin. Radiological evidence of the examination with a slit-lamp microscope. Patients seen degree of bone involvement was categorized in four in their homes were examined with a single loupe and stages of increasing severity (Steinbrocker and others, focal light. Slit-lamp examination served to reveal the I949). depth of inflammation, determining which network of A number of extra-articular complications of rheu- vessels was predominantly affected. It is occasionally matoid disease were noted: subcutaneous - difficult to be certain whether an apparently nodular tous nodules, peripheral neuropathy, manifest episcleritis is in fact a scleritis; the diagnosis is often as skin ulceration or bruising, and tendon rupture. more readily apparent in retrospect when, for example, Atrophy of skin was recorded as a recognized side- an area of scleral translucency will indicate a scleritis. effect of treatment with systemic The greyish-blue colour of deep scleritis is due to the (Scarborough and Shuster, 1960; Greenwood, I966). dark becoming evident through translucent sclera. An enlarged spleen suggested the possibility of Felty's In contrast, episcleritis is usually bright red in colour. syndrome (Felty, 1924), and evidence of amyloidosis A careful slit-lamp examination after observation in was noted, as rheumatoid arthritis is now the com- diffuse light, with a conclusive diagnosis made only after monest cause of secondary amyloidosis in western observing the inflammation for varying periods of time, countries (Boyle and Buchanan, 197I). is sufficient in most cases to provide a diagnosis. Never- Xerostomia and salivary gland enlargement were theless, there remain a few cases in which it is impossible recorded as features associated with Sj6gren's syndrome to differentiate with certainty an episcleritis from a (Sj6gren, 1933). superficial scleritis. The figures that are given, therefore, *Developed by Halberg and Berens (Crookes-Barnes Laboratories, are not absolute, but the features associated with Inc., Wayne, NJ, USA) Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 194 British Journal of Ophthalmology

An electrocardiogram was used in most patients sera giving a positive result in the rheumatoid latex test, (unless bedridden and visited at home) to exclude and by the tanned red-cell haemagglutination technique cardiac disease, which has been noted as a complication for thyroglobulin antibody (Fulthorpe, Roitt, Doniach, in patients with scleritis and rheumatoid arthritis and Couchman, I96I). All the remaining autoantibody (Smoleroff, 1943; Jayson and Jones, I97I). Blood investigations applied the indirect immunofluorescent pressure recordings and cardiac auscultation screened technique. These were tests for thyroid microsomal patients for aortic incompetence, which has been asso- antibody, antinuclear factor, salivary duct antibody, ciated with scleritis (Pitkeathly, Howitt, and Lyne, mitochondrial antibody, gastric parietal-cell antibody, I970). and smooth-muscle antibody (Table I). All patients had a chest x-ray examination to exclude The assay of complement C3 (ic/ia globulin) used evidence of active or and to specific immunodiffusion plates (Hyland Laboratories) demonstrate any pulmonary complication of rheumatoid and serum concentrations of IgG, IgA, and IgM arthritis. globulins were similarly measured by radial immuno- diffusion using commercially-prepared plates (Behring- LABORATORY METHODS werke) according to the method of Mancini, Carbonara, and Heremans (I965). In each patient the haemoglobin concentration, erythro- cyte sedimentation rate, and serum proteins, although nonspecific in nature, reflected the severity of disease. EYE SWABS Haemoglobin estimation at the Tennent Institute of Platinum loop scrapings and swabs taken from the Ophthalmology was given as a percentage, but in the of the inflamed eye (the more inflamed in Centre for Rheumatic Diseases it was given in grams bilateral disease) in each case seen at the Tennent per ioo ml. For comparison of data each haemoglobin Institute of Ophthalmology were sent for investigation value in g/Ioo ml was converted to a percentage reading, of bacteria, fungi, and viruses. Platinum-loop scrapings taking 14 gIo00 ml as a value of ioo per cent. Alkaline were plated on chocolate agar for bacteriology. Isolation phosphatase, serum glutamic oxaloacetic transaminase, of fungi was effected by platinum-loop scrapings plated and serum glutamic pyruvic transaminase were recorded on 2 or 3 per cent malt extract agar with chloromycetin to assess any association of liver disease and scleritis. 0o05 mg/ml. Swabs for virus studies were placed Liver disease has been noted in 6 per cent of patients immediately in Eagle's tryptose phosphate medium with with Sjogren's syndrome without rheumatoid arthritis calf serum io per cent. and in I-5 per cent of patients with Sj6gren's syndrome and rheumatoid arthritis (Whaley, Williamson, Dick, TYPE I ALLERGY () Goudie, Nuki, and Buchanan, 1970). Serum uric acid levels were estimated to exclude gout. In view of the Lyne and Pitkeathly (1968) with J. W. Kerr and J. M. relevance of scleritis in venereal disease, usually syphilis MacLeod noted that, in 25 of 33 cases having single http://bjo.bmj.com/ (Cleobury, i826; von Arlt, I885; Holloway and Fry, attacks of episcleritis, the inflammation affected the 1931), the Wassermann reaction, Venereal Disease interpalpebral area. The findings in this present study Reference Laboratory (VDRL) slide test and Reiter endorse their report and it seemed worth considering protein complement-fixation test (RPCFT) were per- external aggravating factors in the category of immediate formed. Further investigations included the antistrepto- hypersensitivity. Altogether 23 patients, all of them with lysin-O titre, and tests for evidence of and non-rheumatoid inflammation, and all except six cases . The antistreptolysin-O titre may be of episcleritis, were subjected to skin testing for house raised in erythema nodosum, a condition associated with dust, grass pollen, altemaria, feathers, house dust mite, on October 2, 2021 by guest. Protected copyright. episcleritis (McCarthy, I96I). Brucellosis, a well recog- Aspergillus, Candida albicans; and cat and dog hair, if nized cause of uveitis (Woods and Guyton, I944; one of these pets was in the home. Hewson, I964), has been incriminated in a case of scleritis (Thomas, Helluy, Cordier, and Algan, I953) and to exclude it the complement-fixation, standard Table I Autoantibody investigations agglutination (direct), and antihuman globulin Autoantibody Tissue substrate Reference (Coombs') tests were performed. Cat, Marinoni, Giraldi, and Costa (I97I) reported that two of their cases of Thyroid Human toxic Holborow, Brown, Roitt, juvenile rheumatoid arthritis had high false positive microsomal thyroid and Doniach (I959) Antinuclear Rat liver Beck (Ig6I) titres for toxoplasmosis, and accordingly the toxoplasma factor dye test (Sabin and Feldman, 1948) was routinely Salivary duct Human salivary MacSween, Goudie, investigated. gland Anderson, Armstrong, A multiplicity of autoantibodies may occur in the Murray, Mason, Jasani, Boyle, Buchanan, and connective-tissue (collagen) disorders (Copeman, I964). Williamson (I967) Scleritis has been called a collagen disorder (Hayreh Mitochondrial Rat kidney Goudie, MacSween, and and Watson, 1970). In view of this, autoantibody screen- Goldberg (1966) tests were Gastric parietal Human gastric Adams, Glen, Kennedy, ing undertaken in each case. Tests were mucosa McKenzie, Morrow, carried out for rheumatoid factor, using the latex Anderson, Gray, and agglutination technique (Hyland Laboratories) (Ziff, Middleton (I964) Brown, Lospalluto, Badin, and McEwan, 1956), the Smooth muscle Rat stomach Johnson, Holborow, and R3 smooth muscle latex agglutination test (Denver Laboratories) using Glynn (I965) Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis I95

MORTALITY All cases with rheumatoid episcleritis were female and this was consistent with the An initial impression suggested that more of the patients higher with rheumatoid episcleritis or scleritis had died since incidence of rheumatoid arthritis in women. examination as compared with the group of rheumatoid Scleritis was found in twice as many women as controls. The control group was carefully matched for men (67'6 to 32-4 per cent). It is interesting to note age, sex, and year of presentation at the Centre for that of the non-rheumatoid patients, episcleritis Rheumatic Diseases. Contact with the general practi- was found equally in women and men (48-5 and tioner of each patient determined which patients had 5I 5 per cent), but scleritis showed a marked died by February 1974. The Registrar of Births, preponderance in women (83-3 to i6'7 per cent). Deaths and information Marriages, Glasgow, provided The 9 patients with rheumatoid episcleritis as to the cause of death in each case.

Results Table IIIa Age (rheumatoid episcleritis and INCIDENCE (Table II) rheumatoid scleritis) 46 cases Between I965 and 1973, 4210 patients with rheu- Age matoid arthritis were examined at the Centre for Disease Cases range Mean SD SEM Rheumatic Diseases. Seven of these patients had (years) (years) Episcleritis 9 54-69 58-7 ±5-9 ±2-0 Scleritis 37 28-77 58 3 ±iI-8 ±t.9 Table II Incidence of episcleritis and scleritis in rheumatoid arthritis Statistical analysis: t=o-o840; P =NS

No. ofpatients Clinic Disease Cases with rheumatoid Per- Table IIIb Age (non-rheumatoid episcleritis and arthritis centage non-rheumatoid scleritis) 5I cases Centre for Episcleritis 7 4210 0-17 Rheumatic Scleritis 28 4210 o-67 Age Diseases 1965-73 Disease Cases range Mean SD SEM Scleritis Clinic, Episcleritis 35 2 5.7 (years) (years) Tennent Scleritis 27 9 33-3 Institute of Episcleritis 33 20-75 45.1 ±i4-6 ±2 5 Ophthalmology Scleritis s8 32-75 53 2 ±13-7 ±3-2 1970-71 Statistical analysis: t = I 8746; P =NS http://bjo.bmj.com/ Statistical analysis (rheumatoid versus non-rheumatoid): Episcleritis: t = 2-7092; P = < o oo5 Scleritis: t=s5044; P==NS episcleritis (o0 I7 per cent) and 28 patients had scleritis (o-67 per cent). Thirty-five cases of episcleritis and 27 cases of Table IIIc Sex ratio (rheumatoid episcleritis and scleritis presented at the Tennent Institute of rheumatoid scleritis) 46 cases Ophthalmology during 197 I and 1972. Two of on October 2, 2021 by guest. Protected copyright. the cases of episcleritis (5 7 per cent) and nine of Females Males the cases with scleritis (33.3 per cent) had clinical Disease Cases Ratio and radiological evidence of rheumatoid arthritis No. Per cent No. Per cent according to the diagnostic criteria of the American Episcleritis 9 9 I00 0 - - Rheumatism Association (Ropes and others, 1959). Scleritis 37 25 67.6 12 32 4 2.I/1 Each of these cases was sero-positive for rheumatoid 34 73.9 12 26-5 2-8/I factor. One further case of episcleritis and six further cases of scleritis were weakly sero-positive for rheumatoid factor, but had no clinical or radio- Table Illd Sex ratio (non-rheumatoid episcleritis logical evidence of rheumatoid arthritis. and non-rheumatoid scleritis) 51 cases (Unless otherwise indicated in the remaining data patients from both clinics with rheumatoid Females Males arthritis are combined and compared with the Disease Cases Ratio findings of patients without rheumatoid arthritis.) No. Per cent No. Per cent Episcleritis 33 I6 48-5 17 515 I/I-I Scleritis I8 I5 83.3 3 I6-7 5/I AGE, SEX, AND BILATERALITY (Tables IIIa-f) 31 6o-8 20 39.2 i-6/i The only group of patients with a mean age outside the sixth decade was the non-rheumatoid episcleritis Statistical analysis (female): X2=59337; P= <0-02 Statistical analysis (rheumatoid versus non-rheumatoid (female): group which had a mean value of 45 1 years. Scleritis: t= 5175; P=NS Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 196 British Yournal of Ophthalmology

Table IlIe Bilaterality (rheumatoid episcleritis and inflammation were more often admitted to the rheumatoid scleritis) 46 cases wards, or does it indicate that there was more clinical indication for electrocardiography in the Eye inflammation episcleritis/scleritis groups? Certainly clinical ex- Disease Cases Percentage Unilateral Bilateral amination of the control group had given little evidence of severe cardiovascular disease. The No. Per cent No. Per cent Episcleritis 9 I9-6 5 55-6 4 44-4 Scleritis 37 8o-4 12 32-4 25 67-6 Table IVa Past and associated disease (rheumatoid episcleritis and scleritis) 46 cases Episcleritis Scleritis Table IlIf Bilaterality (non-rheumatoid episcleritis (9 cases) 37 (cases) and non-rheumatoid scleritis) 5 I cases Per Per Eye inflammation No. cent No. cent Disease Cases Percentage Unilateral Bilateral Ischaemic heart Ischaemic heart disease, 7 cases disease, 30 cases No. Per cent No. Per cent (2 unknown) 3 42-9 (7 unknown) I8 6o-o Rheumatic fever 2 22-2 Hypertension, Episcleritis 33 64-7 13 39-4 20 6o-6 Hypertension, 32 cases Scieritis i8 35 3 Is 6I-I 7 38 9 7 cases (5 unknown) 8 25-0 (2 unknown) I 14-3 Angina pectoris Aortic stenosis I (electrocardio- Cerebrovascular gram normal) 3 8-I accident I Cerebrovascular showed a similar number of unilateral and bilateral Raynaud's accident 3 phenomenon I Pericarditis I 2.7 cases (5:4) but rheumatoid scleritis (37 cases) was Pneumonia I Mitral stenosis I more commonly bilateral (67-6 per cent). Non- Pleural effusion I Aortic stenosis was more Pleurisy I and incompe- rheumatoid episcleritis (33) commonly Sinusitis I tence I bilateral (39-4 to 6o06 per cent) and non-rheumatoid Osteoarthritis I1 Rheumatic fever scleritis (I8) was more often unilateral (6i-i to Chronic renal Pulmonary embolus disease I1 Deep venous 38-9 per cent). Oral pemphigoid I thrombosis I Gastric ulcer I Pneumonia 5 13-5 Recurrent colitis I Pleural effusion 5 PAST MEDICAL HISTORY AND ASSOCIATED DISEASE Hemicolectomy I Sinusitis 4 Jo-8 http://bjo.bmj.com/ Pelvic I Rheumatoid (Tables IVa, b) Diabetes I nodules of lung 3I Bilateral otitis Tuberculosis of A breakdown of medical history and associated media I lung 2 disease revealed a high incidence of associated Emphysema 2 Asthma abnormality of the cardiovascular and respiratory Pneumoconiosis systems. Granulomatous Of 30 cases with rheumatoid scleritis (6o0o per (collagen) disease 4 cent) in whom electrocardiograms were taken, i8 Osteoarthritis 4 on October 2, 2021 by guest. Protected copyright. had ischaemic heart disease. Three further patients Dupuytren's contracture I complained of angina pectoris but had normal I tracings. Six patients with non-rheumatoid epis- Duodenal ulcer 4 cleritis (21 9 per cent) had electrocardiographic Cholecystectomy 3 Hysterectomy 3 changes and one further patient had a possible Diverticulosis I pericarditis. Three patients with non-rheumatoid Adenocarcinoma of bowel I scleritis (I6'7 per cent) had ischaemia of the Appendicectomy myocardium. There is a significant difference Prostatectomy Prolapsed bowel between the incidence of ischaemia in rheumatoid Inguinal hernia scleritis and that in non-rheumatoid scleritis Hepatitis The combined episcleritis/scleritis Herpes zoster (P=o00005). (trunk) 3 groups showed a total incidence of myocardial Psoriasis ischaemia in 21 of 37 known cases (56-7 per Rodent ulcer 5 Thyrotoxicosis cent) in comparison with the rheumatoid control Diabetes group, eight of 20 known cases (40o0 per cent). It is Trigeminal neuralgia I interesting to note that only seven in the episcleritis/ Pernicious anaemia scleritis groups did not have an electrocardiogram Puerperal fever I compared with i6 in the matched control group. Chronic otitis Does this reflect the fact that the groups with eye media Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 197

statistically insignificant difference between these of respiratory complications - that is, three of nine two groups might well be misleading. cases in episcleritis (33-3 per cent) and I3 of 37 Excluding respiratory complaints such as bron- cases in scleritis (35.I per cent). The incidence of chitis, emphysema and asthma, pleurisy and respiratory disease in non-rheumatoid episcleritis rheumatoid pneumonia, pleural effusion, pleurisy and scleritis was much lower; two of 33 cases in and rheumatoid nodules of the lung, both rheuma- episcleritis (6'i per cent) and one of i8 cases in toid episcleritis and scleritis had a similar incidence scleritis (5'6 per cent). The difference in each case is statistically significant (P=o'05 and P=o0o05, Table IVb Past medical history and associated respectively). The combined rheumatoid episcleri- disease (non-rheumatoid episcleritis and non- tis/scleritis groups gave a total of i6 with rheuma- rheumatoid scleritis) 5i cases toid respiratory complaints out of 46 cases (34'8 per cent) as compared with five of 45 rheumatoid

Episcleritis Scleritis controls (ii i per cent). The difference is statistic- (33 cases) (i8 cases) ally significant (P=o'oi). Per Per Other associated systemic disorders are listed in No. cent No. cent Tables IVa, b. Hypertension, Hypertension, 30 cases I6 cases EYE SYMPTOMS (Tables Va, b) (3 unknown) 8 26-7 (2 unknown) 4 25-0 Ischaemic heart Ischaemic heart Redness and ocular pain are by far the commonest disease, 32 cases disease 3 Ii67 (I unknown) 6 x8-8 Rheumatic fever 5.5 complaints in both episcleritis and scleritis. A red Valvular heart Aortic incompe- I eye was referred to only slightly less often in rheu- disease 2 6-I tence Rheumatic fever 2 Msgraine Pericarditis I 3-0 Sinusitis 5 27-8 Cerebrovascular Asthma 2 I I* I Table Va Eye symptoms (rheumatoid episcleritis accident I Tuberculosis of Sinusitis 8 chest and rheumatoid scleritis) 44 cases Pneumonia 2 Bronchitis Asthma 2 Pneumonia Episcleritis Scleritis Bronchitis I Osteoarthritis 3 I6.7 (35 cases, Osteoarthritis 5 I5-2 'Fibrositis' 2 Symptoms (9 cases) 2 unknown) 'Rheumatism' 5 'Slipped disc' 'Back trouble' 2 'Rheumatism' I No. Per cent No. Per cent 'Fibrositis' 2 Gout I 'Knee synovitis' I Menopausal Red, inflamed, bloodshot 8 88-9 29 82-9 http://bjo.bmj.com/ Cervical haematuria Ocular pain 3 33 3 I6 45-7 spondylosis I Cholecystectomy Gritty, sandy 2 22-2 II 3I.4 Paget's disease of Appendicectomy 2 Burning I III 5 54.3 spine and Hysterectomy 2 Itching 0 3 8-6 ischium I Gastritis Watering I 3 Coccygodynia I 'Gastric Irritation I 2 5-7 Patellar haemorrhage' 0 2 displacement I Duodenal ulcer Oedema of lids 0 2 Gout I Apical dental Headache 0 2

Pyelitis 2 abscess Discharge, mattering I I 2-9 on October 2, 2021 by guest. Protected copyright. Tuberculosis of Systemic Dryness I I kidneys I erythematosus I Blurring of vision 0 I Hydronephrosis I Erythema nodosa I Renal stone I Multiple rodent Duodenal ulcer 4 12-I ulcers I Hiatus hernia 2 Jaundice I Table Vb Eye symptoms (non-rheumatoid Cholecystitis I Depression I Appendicectomy I episcleritis and non-rheumatoid scleritis) 51 cases Perianal sinus I Apical dental Episcleritis Scleritis abscess I (33 cases) (i 8 cases) Jaundice I Symptoms Thyroidectomy I No. Per cent No. Per cent Erythema nodosa 2 Chronic discoid Red, inflamed, bloodshot 31 93-9 I6 88-9 lupus erythema- Ocular pain i6 48-5 13 72-2 tosus I Gritty, sandy 8 24-2 3 I6'7 Psoriasis I Headache 4 s2s- 5 *6 Pityriasis rosea I Oedema of lids 4 o Ulcer inguinal Irritation 3 9-I 3 region I Discharge, mattering 3 0 Acne I Burning 2 6-I 2 II*I Herpes zoster Watering 2 0 ophthalmicus I Dryness I 3-0 I Diabetes I Blurring of vision I I Glandular fever I Itching o I Puerperal fever I Photophobia o o Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from I98 British Journal of Ophthalmology

matoid episcleritis (88-9 per cent) than in non- VISUAL ACUITY (Table VII) rheumatoid episcleritis (93-9 per cent), and in Visual acuity was recorded in the Tennent Institute rheumatoid scleritis (82z9 per cent) compared with of Ophthalmology but it was not recorded routinely non-rheumatoid scleritis (88&9 per cent). A greater in patients seen at the Centre for Rheumatic disparity of complaint was notable with regard to Diseases. An arbitrary dividing line of visual acuity ocular pain in rheumatoid episcleritis (33-3 per cent) of or a for compared with non-rheumatoid episcleritis (48 5 6/12 (Snellen), worse, provided parameter per cent), and in rheumatoid scleritis (45-7 per cent) compared with non-rheumatoid scleritis (72-2 Table VII Visual acuity (Tennent Institute of per cent). The lower incidence of complaint by Ophthalmology) 62 cases rheumatoid patients probably reflects the measure No. of with of discomfort they experience elsewhere. Disease Eyes visual acuity Percentage (total no.) < 6/9 LENGTH OF HISTORY OF EYE INFLAMMATION Episcleritis 56 8 14-3 (Tables VIa, b) Scleritis 41 I5 36.6 Figures giving the duration of ocular inflammation Statistical analysis (episcleritis versus scleritis): must be interpreted cautiously, as it is impossible x2=6 5070; P= <0 025 to know whether the inflammation is truly quiescent. However, in each of the four groups of patients the level of visual acuity achieved by each patient. the mean duration (with a large standard deviation Of the total of 56 eyes with episcleritis, eight eyes in each case) showed no significant difference. It (14-3 per cent) had a visual acuity of 6/12 or worse, was interesting to note that the mean duration of and of the 4I eyes with scleritis, I5 (36-6 per cent) ocular inflammation in the rheumatoid patients was had a visual acuity of 6/12 or worse. The difference shorter than that in the non-rheumatoid patients; between these two groups is statistically significant 36-2 months (rheumatoid episcleritis) and 35-o (P=0o025). It must be emphasized that refraction months (rheumatoid scleritis) compared with 42-4 was not done as a routine in these cases and that months (non-rheumatoid episcleritis) and 48-9 the difference is based on the patient's best visual months (non-rheumatoid scleritis). There was a acuity either with distance glasses or without. wide range in each group, from a few days or weeks to 23 years. http://bjo.bmj.com/ TYPE OF INFLAMMATION (Tables VIIIa, b) All nine cases of rheumatoid episcleritis were of Table VIa Length of history of eye inflammation the 'simple' variety. Nodular episcleritis was noted (rheumatoid episcleritis and rheumatoid scleritis) in six cases (I8-2 per cent) of the patients with non- 40 cases rheumatoid inflammation (Fig. I) and simple episcleritis was recorded in 27 (818 per cent). Mean

Disease Cases Range (months) SD SEM Diffuse scleritis (Fig. 2) was more common in on October 2, 2021 by guest. Protected copyright. rheumatoid scleritis (83 3 per cent) and non- Episcleritis 8 3 days/ 36-2 ±6o-o ±2I+2 (I unknown) I5 years rheumatoid scleritis (6i1i per cent). A nodular Scleritis 32 I week/ 35s0 +54-0 +9-6 scleritis was slightly less common in rheumatoid (5 unknown) 20 years scleritis (I6-7 per cent) as compared with non- Statistical analysis: t=0o0534; P=NS rheumatoid scleritis (33-3 per cent). One case of non-rheumatoid necrotizing scleritis (Fig. 3) pre- sented in a patient who had developed gout some Table VIb Length of history of eye inflammation months earlier, but the area of necrotic sclera was (non-rheumatoid episcleritis and non-rheumatoid scleritis) 49 cases Table VIlla Type of inflammation (rheumatoid Mean episcleritis and rheumatoid scleritis) 45 cases Disease Cases Range (months) SD SEM Per Per Episcleritis 31 4 weeks/ 42.4 ±74 2 ±I3 3 Disease Cases cent Type cent (2 unknown) 23i years Scleritis I8 7 weeks/ 48-9 ±74-2 ±17-5 Episcleritis 9 Nodular o 20 years Simple 9 Scleritis 36 (I unknown) *Nodular 6 I6-7 Statistical analysis: t=0o2938; P =NS Anterior 36 I00 Diffuse 30 83-3 Statistical analysis (rheumatoid versus non-rheumatoid): Posterior 8 22-2 Necrotizing o Episcleritis: t-=OII39; P=NS Scleritis: t=o05713; P=NS *Scleral translucency in 29 cases (8o-6 per cent) Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis I99

Table VIIIb Type of inflammation (non-rheumatoid episcleritis and non-rheumatoid scleritis) 51 cases

Per Per Disease Cases cent Type cent Episcleritis 33 Nodular 6 I8-2 Simple 27 8I-8 Scleritis i8 Nodular 6 33-3 Anterior I8 IoO* Diffuse i I 6I-I Posterior I 5 6 Necrotizing s 5 6

*Scleral translucency in two cases (i I-I per cent) probably determined by a subconjunctival injection of some days previously. Co-existent posterior scleritis was more common in rheumatoid scleritis (22z2 per cent) than in non- FIG. 3 Necrotizing scleritis rheumatoid scleritis (5-6 per cent). The figure of eight cases in rheumatoid inflammation was a tentative one, based on two cases of choroiditis, or muscle involvement. These figures do not include one case of , and five cases of patients with macular disturbance which might extraocular muscle imbalance, suggesting tendon have been due to other causes. The section on posterior segment abnormality will deal with this in more detail. One of the most interesting findings was the difference between rheumatoid and non- rheumatoid scleritis in the incidence of scleral translucency (Fig. 2). Of 36 cases with rheumatoid scleritis, 29 (8o-6 per cent) showed scleral translu- cency as compared with two of i8 cases with non- rheumatoid scleritis ( II i per cent).

DISTRIBUTION OF INFLAMMATION

(Tables IXa, b) http://bjo.bmj.com/ The distribution of ocular inflammation in episcleri- tis is more commonly in the interpalpebral area, suggesting the possibility of external aggravating factors. Scleritis will also affect the superior epis- quadrants of the eye in particular, although less FIG. I Nodular non-rheumatoid episcleritis markedly so in non-rheumatoid scleritis. Rheuma- on October 2, 2021 by guest. Protected copyright.

FIG. 2 Bilateral diffuse scleritis Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 200 British Journal of Ophthalmology

Tables IXa, b Distribution of inflammation Distribution of inflammation (rheumatoid episcleritis & rheumatoid scleritis) 38 cases Episcieritis Scleri tis 9 cases 29 cases (8 unknown) Right (7eyes) Left (7 eyes) Right (31 eyes) Left (30 eyes) 22 2%/ 00/0 68.90/ 72 80/a A 55 60 44 50/o 5560/a 33.40/a 78.9/ 55 20/a 655 a58.60/ /22.20/ 1/ 2%/ /58.60/a 51.7%\ Distribution of inflammation (non-rheumatoidepisceritis &non-rheumatoid scleritis) 51 coses Episcleritis Scleritis 33 cases 17 cases (one posterior scieritis) Right (23 eyes) Left (29 eyes) Right (14 eyes) Left (12 eyes) \3-0°h 18.20/ \29.30/ 29.40/ B 273/ 48.-50/ 57.6 5.5o5% 47.j/ 6<4.70/a 58. 41.20/ /18.20/ao /27 30/a /5.9o0/a /I1180/a toid scleritis has a fairly even distribution affecting not statistically significant. Corneal complications all quadrants of the anterior sclera. showed a higher incidence in rheumatoid as compared with non-rheumatoid disease, which was CORNEAL COMPLICATIONS (Tables Xa, b) statistically significant for episcleritis (P=0ooo5) but not for scleritis. An unexpected finding among the rheumatoid The predominant corneal complication is a patients was a higher percentage of corneal compli- sclerosing which is usually localized at the cations in episcleritis (69z2per cent) compared with site nearest to the inflamed sclera or episclera. scleritis (43-5 per cent). The small number of Occasionally, in diffuse inflammation, the condition I3 http://bjo.bmj.com/ eyes with episcleritis, however, may have given an is annular (Fig. 4). The scar of sclerosing keratitis inaccurate picture when one considers the anatomy is permanent but in itself has little effect on visual of the structures involved. The difference is not acuity. However, of io patients with annular sclero- statistically significant. Non-rheumatoid inflam- sing keratitis in whom visual acuity was assessed, mation showed a lower percentage of corneal six (6o0o per cent) had markedly reduced visionI involvement in episcleritis (I8-9 per cent) than in due to changes. scleritis (28o per cent) although the difference was Limbal guttering occurred in one eye with rheu- on October 2, 2021 by guest. Protected copyright.

FIG. 4 Annular sclerosing keratitis in a patient wit/ non- rheumatoid scleritis Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 201

Table Xa Corneal complications (rheumatoid matoid episcleritis, three eyes with rheumatoid episcleritis and rheumatoid scleritis) 46 cases scleritis (Fig. 5), and one eye with non-rheumatoid episcleritis. Apart from the last case, each occurred Eye inflammation at the site of an area of sclerosing keratitis. A loca- Disease Eyes Unilateral Cases Bilateral Cases Per- lized stromal keratitis occurred deep in the mid- (no.) (no.) (no.) centage stroma in one patient with rheumatoid episcleritis. with rheumatoid scleritis had a Episcleritis 13 Localized Localized One patient per- stromal I* sclerosing 3 69-z forating keratolysis, which is the comeal counter- Localized sclerosing I part of a perforating scleritis. Annular sclerosing I Scleritis 62 Localized Localized ANTERIOR UVEITIS (Tables XIa, b) sclerosing 2 sclerosing 6 43-5 Annular Annular The percentage of eyes with active uveitis or sclerosing I sclerosing 4 evidence of old uveitis was slightly lower in rheu- Keratolysis I Grossly matoid I with degenerate episcleritis (23 per cent) compared blind I rheumatoid scleritis (28-3 per cent). The difference Localized is not statistically significant. A similar picture sclerosing I * emerged in non-rheumatoid episcleritis (17.0 per Three eyes with scleritis (4-9 per cent) and one with episcleritis cent) and non-rheumatoid scleritis (20o per cent). (7-6 per cent) developed a limbal gutter at the site of sclerosing Posterior synechiae suggesting a more active and keratitis Statistical analysis: X2=-8399; P =NS plastic type of uveitis were found in seven of I7 In all unilateral cases and one bilateral case (marked with asterisk) eyes with rheumatoid scleritis (41I2 per cent) as only one eye was affected compared with one eye with posterior synechiae associated with non-rheumatoid scleritis (20z0 Table Xb Corneal complications (non-rheumatoid per cent). episcleritis and non-rheumatoid scleritis) 5 I cases CATARACT (Table XII) Eye inflammation The findings as regards the development of cataract Disease Eyes Unilateral Cases Bilateral Cases Per- in patients with rheumatoid scleritis who had (no.) (no.) (no.) centage

Episcleritis 53 Localized http://bjo.bmj.com/ sclerosing 2* I8-9 Table XIa Anterior uveitis (rheumatoid episcleritis Localized and rheumatoid scleritis) 45 cases sclerosing 2 Annular sclerosing I* Eye inflammation Limbal guttering I Disease Eyes Unilateral Bilateral Per- Band kera- (no.) - centage topathy I Affected Cases Affected Cases

Scleritis 25 Localized Annular eyes (no.) eyes (no.) on October 2, 2021 by guest. Protected copyright. sclerosing 3 sclerosing 2* 28-o Episcleritis I 3 - 2* I 23-1 5 I Statistical analysis: X2=o-83II; P=NS Scleritis 6o I I 2 6 28-3 Statistical analysis (rheumatoid versus non-rheumatoid): (2 unknown) I 4 Episcleritis: X2 =I2-gi6i; P= <00005 Scieritis: X2 =I-8oI; P=NS *Both eyes were affected Statistical analysis: X'=0'1484; P=NS

Table XIb Anterior uveitis (non-rheumatoid episcleritis and non-rheumatoid scleritis) 51 cases Eye inflammation Disease Eyes Unilateral Bilateral Per- (no.) - centage Affected Cases Affected Cases eyes (no.) eyes (no.) Episcleritis 53 I 3 2 2 17.0 I 2 Scleritis 25 I 4 I I 20-0 Statistical analysis: x'=o Iosx; P=NS Statistical analysis (rheumatoid versus non-rheumatoid): FIG. 5 Limbal guttering in a patient with Episcleritis: X2=o26o8; P =NS rheumatoid episcleritis Scleritis: X'=o-6388; P=NS Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from

202 British Journal of Ophthalmology

Table XII Posterior subcapsular cataract (PSC) results need to be viewed with caution. Many of in patients receiving systemic corticosteroids the rheumatoid patients received a number of different antirheumatic drugs which might have Rheumatoid Cases Mean age Sex ratio Per- scieritis (no.) ± SD FIM PSC centage disturbed the retinal picture. However, macular disturbance is a recognized complication of scleritis Patients 25 55-89±12-99 3-5/1 9 36-o (Watson and Lobascher, I965). Two eyes showed 'Control' group 148 52-5±11-5 9-6/i I7 II-5 old foci of choroiditis and one eye, in a patient now Statistical analysis: X2 = Io-0639; P = < 0-005 deceased, optic atrophy of obscure origin, although the eye had been grossly inflamed and suffered a subsequent perforating keratolysis. It was interest- received treatment with systemic corticosteroids are ing to find oedema of the head in one of particular interest. The incidence of posterior case, suggesting a possible ischaemic optic neuro- subcapsular cataract in patients with rheumatoid pathy. One patient had an old detached . scleritis who had been treated with systemic cortico- There was an area of old choroiditis in one case steroids was high (36-o per cent). Nine of 25 diagnosed as non-rheumatoid episcleritis (i-9 per patients who had received treatment with systemic cent of 52 eyes). Two eyes with non-rheumatoid corticosteroids had developed posterior subcapsular scleritis had a possibly significant abnormality of cataract. There were three further patients who had the posterior pole (8-o per cent of 25 eyes); in both advanced cataract, two of whom had bilateral cases these were areas of macular disturbance. cataract (one had one aphakic eye), but these are not There was a marked difference in the incidence included for the purposes of statistical analysis. of abnormality between rheumatoid episcleritis A previous study involving two of us (JW and and scleritis. The significant difference between DDMM), of the incidence of steroid-induced posterior segment complications in rheumatoid cataract in patients with rheumatoid arthritis, scleritis and non-rheumatoid scleritis (P=o0o5) found that i7 of 148 patients with rheumatoid may reflect more severe disease in the rheumatoid arthritis (ii.5 per cent) who had received systemic patient. corticosteroids had posterior subcapsular . These patients also attended the Centre for Rheu- matic Diseases, Glasgow. Comparison of the INTRAOCULAR PRESSURE (Tables XIVa, b) incidence of posterior subcapsular cataract in rheu- One patient with rheumatoid episcleritis had matoid scleritis (36-o per cent) with posterior bilateral raised intraocular pressure, but this was http://bjo.bmj.com/ subcapsular cataract in patients with rheumatoid arthritis without scleritis (I I5 per cent), both groups receiving systemic treatment, Table XIIIa Posterior segment abnormality is statistically significant (P=o0oo5). (rheumatoid episcleritis and rheumatoid scieritis) Only three other patients in the study had the 42 cases characteristic form of cataract. One patient with Episcleritis Scieritis Per- rheumatoid episcleritis had bilateral posterior on October 2, 2021 by guest. Protected copyright. subcapsular cataract and had been treated with 9 eyes (4 unknown) 55 eyes (7 unknown)* centage systemic corticosteroids for an uncertain length of - Macular disturbance 1I 200ot time. It is interesting that of with non- Old choroiditis 2 4-2 5I patients Optic atrophy I 2-I rheumatoid episcleritis or scleritis, two patients Oedema of optic nerve head I developed posterior subcapsular cataract, and these Old detached retina I were the only two patients of the non-rheumatoid 29-1 group who were receiving systemic corticosteroid *Two cases were obscured by cataract treatment. fThree eyes were myopic

POSTERIOR SEGMENT ABNORMALITY (Tables XIIIa, b) Table XIIIb Posterior segment abnormality No patient with rheumatoid episcleritis had abnor- (non-rheumatoid episcleritis and non-rheumatoid 5I cases mality of the posterior segment of the eye. However, scleritis) the percentage of abnormal features in association Episcieritis Per- Scieritis Per- with rheumatoid scleritis was 29-I per cent. The 52 eyes (I unknown) centage* 25 eyes centage commonest abnormality was a macular disturbance Old choroiditis I i-9 Macular disturbance 2 8-o in i i eyes (20oo per cent), although three of these eyes were myopic, and this might have affected their *One case obscured by macular state. The origin of macular disturbance Statistical analysis: X'=I-665x; P=NS Statistical analysis (rheumatoid versus non-rheumatoid): is difficult to determine in these cases, and the Scleritis: X2 =4 3844; P = < 0°05 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 203

Table XIVa Raised intraocular pressure* (rheumatoid episcleritis and rheumatoid scleritis) 37 cases Eye inflammation Disease Eyes (no.) Unilateral Bilateral Percentage Cases Eyes affected IOP mmHg Cases Eyes affected lOP mmHg

Episcleritis 9 (4 unknown) s 2 26 - +25 Scleritis 48 (I4 unknown) I I 22 2 2 29 I8-7 +32 26 +23 4 I 22 Absolute 22 32 *The highest reading recorded is given in each case tA c13s of closei-aagla glaucz-n nzt coa;isirsd a3 p3itive as the raised pressure was not due to inflammation

Table XIVb Raised intraocular pressure (non-rheumatoid episcleritis and non-rheumatoid scleritis) 5 I cases Eye inflammation Disease Eyes (no.) Unilateral Bilateral Percentage Cases Eyes affected lOP mmHg Cases Eyes affected IOP mmHg

Episcleritis 53 - I I 22 I-9 Scleritis 25 3 I 26 - 2-0 30 38 2 2* 30 +28 30 +34 *Both tViese cases, although unilateral cases of scleritis, received topical steroids to both eyes. On withdrawing topical steroids all four eyes reverted to normal intraocular pressures. They are con3idered cases of steroid-induced glaucoma and not primarily due to inflammation. Stati3tical analy3is (rheumatoid versus non-rheumatoid): Scleritis, X'=05453; P =NS http://bjo.bmj.com/ due to closed-angle glaucoma and the condition Only one patient, a woman with unilateral rheuma- settled with bilateral peripheral iridectomies. No toid scleritis, required a drainage procedure to other patient with rheumatoid episcleritis had control her intraocular pressure. There was no raised intraocular pressure. Nine eyes with rheuma- flare-up of her scleritis. One patient with a pressure toid scleritis (I8-7 per cent) had raised intraocular at 22 mmHg in one eye is still receiving miotic pressure; in two cases there was absolute glaucoma. treatment. The intraocular pressure of the remain- on October 2, 2021 by guest. Protected copyright. Four eyes had levels recorded by applanation ing patients all settled coincident with the quieten- tonometry at or between 22 and 26 mmHg. One ing of their inflammation, including those patients eye, with a reading at 22 mmHg, was a case of old in whom pressures were recorded in the 30-38 'burnt out' absolute glaucoma. One eye had a mmHg range. reading at 29 mmHg, one at 30 mmHg, and two at 32 mmHg. Only one patient with non-rheumatoid episcleritis OCULAR MOVEMENTS (Tables XVa, b) had one eye with a tension fractionally above normal No patient with rheumatoid episcleritis had at 22 mmHg, a percentage of i 9 per cent. Seven evidence of muscle imbalance. Eight eyes with eyes with non-rheumatoid scleritis had raised rheumatoid scleritis (12-9 per cent) showed limi- levels but in two cases (four eyes) the level of tation of movement of one extraocular muscle or pressure fell from 30 mmHg (right) + 28 mmHg more. Two cases, however, were based on a history (left) and 30 mmHg (right) + 34 nmmHg (left) to of transient diplopia, suggesting tendon or muscle within normal limits on withdrawing topical involvement consistent with the site of scleral steroid treatment. It is considered that these were inflammation. The limitation of movement in the cases of steroid-induced glaucoma. The remaining remaining six cases was substantiated by orthoptic three cases (I2o per cent) include one patient in examination. whom the intraocular pressure was 38 mmHg. In Three cases of non-rheumatoid episcleritis (5 7 each case the highest reading recorded is given. per cent) suggested possible tendon and/or muscle Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from

204 British Journal of Ophthalmology

Table XVa Ocular movements (rheumatoid the Schirmer I tear test was markedly reduced. episcleritis and rheumatoid scleritis) 46 cases All three patients had died before complete exami- nation could be carried out and no Schirmer II test Disease Eyes (no.) Muscle imbalance Percentage or rose bengal staining had been done. Episcleritis I3 - - Scleritis 62 8* 12-9 SJOGREN'S SYNDROME *Two cases based on history of transient diplopia Sjogren's syndrome is defined as a triad of features; KCS, xerostomia (with or without salivary gland Table XVb Ocular movement (non-rheumatoid swelling), and a connective-tissue disease. The episcleritis and non-rheumatoid scleritis) 51 cases presence ofeither KCS or xerostomia with a systemic disorder is accepted as sufficient (Bloch, Buchanan, Disease Eyes (no.) Muscle imbalance Percentage Wohl, and Bunim, I965). Episcleritis 53 3* 5*7 Only two of the 37 cases of rheumatoid scleritis Scleritis 25 6t 24-0 exhibited the triad of dry eyes, dry mouth, and a connective-tissue disorder (in this case rheumatoid *AI1 three cases tentatively based on history of transient diplopia tOne case based on a history of intermittent diplopia arthritis). No case of rheumatoid episcleritis had Statistical analysis (non-rheumatoid episcleritis versus non-rheu- the triad of features of Sj6gren's syndrome. How- matoid scleritis): X2=5-5976: P= <0'025 Statistical analysis (rheumatoid versus non-rheumatoid): ever, two further cases of rheumatoid scleritis and Scleritis: x2 = i-6248; P =NS one case of rheumatoid episcleritis had KCS associated with the systemic disease. Eight further cases of rheumatoid scleritis and two cases of inflammation, but the conclusion is tentative. All rheumatoid episcleritis also had xerostomia with three cases gave a brief history of diplopia which rheumatoid arthritis. This means that the total they associated with an inflamed eye. Six cases with incidence of Sj6gren's syndrome, according to the non-rheumatoid scleritis (24o0 per cent) gave criteria given, is I 2 of 37 cases of rheumatoid evidence of tendon and/or muscle imbalance, one scleritis (32-4 per cent) and three of nine cases of case based on an account of intermittent diplopia rheumatoid episcleritis (33-3 per cent). and the remaining five cases confirmed by an orthoptic assessment. The difference in incidence SICCA SYNDROME between non-rheumatoid episcleritis and non- http://bjo.bmj.com/ rheumatoid scleritis is statistically significant One patient with non-rheumatoid episcleritis (P =oo25). (3 o per cent) had dry eyes combined with a dry One patient, a 34-year-old woman who had non- rheumatoid scleritis presented with a dramatic . Examination revealed a paresis of the Table XVIa sicca (rheumatoid left lateral rectus muscle. This was the third time in episcleritis and rheumatoid scleritis) 42 cases three years that she had had an inflamed eye. On on October 2, 2021 by guest. Protected copyright. the first occasion there had been a paresis of the Disease Cases (no.) No. of KCS cases Percentage right medial rectus muscle. A biopsy (IX3 mm) of tendon and muscle of the Episcleritis 8 (I unknown)* I 14-2 of the middle section Scleritis 34 (3 unknown)t 4 12'9 left lateral rectus muscle confirmed the presence of a tenomyositis. Chemosis was characteristic of each *The 'unknown' case had a markedly reduced Schirmer I test tTwo 'unknown' cases had markedly reduced Schirmer I tests episode after a few days of ocular pain, and on each (one case is included as a positive KCS because of a reduced occasion was thought to herald the onset of an Schirmer I test with widespread rose bengal staining) . Statistical analysis: x2 =ooog6; P =NS

KERATOCONJUNCTIVITIS SICCA (Tables XVIa, b) Table XVIb Keratoconjunctivitis sicca (non- One patient with rheumatoid episcleritis (14-2 rheumatoid episcleritis and non-rheumatoid per cent) had definite KCS, and four patients with scleritis) 51 cases rheumatoid scleritis (I2-9 per cent). Of the patients Disease Cases (no.) No. of KCS cases Percentage with non-rheumatoid inflammation, four with and one with Episcleritis 33 4 12- episcleritis (I2zI per cent), scleritis Scleritis s8 5s6I (5-6 per cent) had KCS. There were three further cases of possible KCS. Statistical analysis: x =o5678; P =NS with rheumatoid Statistical analysis (rheumatoid versus non-rheumatoid): One case was associated episcleritis Episcleritis: X2=00247 P =NS and two cases with rheumatoid scleritis. In each Scleritis: x'=o-67b0; P=NS Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 205

Table XVII Articular features ofpatients with rheumatoid episcleritis and rheumatoid scleritis

Rheumatoid episcleritis/scleritis Rheumatoid controls Features No. of Range Mean No. of Range Mean P value cases (years) (years) SD SEM cases (years) (years) SD SEM Duration of rheumatoid arthritis 46 0°5-50 14.5 ±12-9 ±1-9 46 O-I7-48 9-2 ±10-2 ±I55 <0-05 Functional class 45 I-4 2-7 ±°-9 ±0 I4 46 1-4 2-28 ±o-78 ±O-oI NS Articular index 43 0-78 23I ±20-0 ±31 4I 0-53 I9-0 ±12-9 ±2-0 NS X-ray stage 36 I-4 2.97 ±o-6i ±05- 38 I-4 2.42 ±1503 ±01-7 <00- NS =Not significant

mouth, features of the sicca syndrome, in which the incapacity was the only investigation in which the dryness of mucous membranes is not associated rheumatoid control group showed a more severe with a connective-tissue disorder. degree of abnormality compared with the rheuma- toid episcleritis/scleritis group. The mean value for 45 known cases of rheumatoid episcleritis/ RHEUMATOLOGY scleritis was 2 I 7 and for 46 cases in the rheumatoid The following data refer to those patients with control group was 2z28. rheumatoid episcleritis or rheumatoid scleritis and The articular index records the number of compare them with a group of rheumatoid controls, involved joints and the severity of their involvement matched for age and sex - that is, patients with as judged by tenderness to pressure over the joint rheumatoid arthritis but with no evidence of margin. The matched group of rheumatoid controls episcleral or scleral inflammation. For the purposes had a mean value of i9. The difference between the of this paper, as regards the articular and extra- control group and the combined rheumatoid epi- articular features of rheumatoid disease, the patients scleritis/scleritis group (mean value: 231) was not with rheumatoid episcleritis and scleritis are com- statistically significant. bined as one group and compared with the rheu- The x-ray stage indicates radiologically the matoid control group. degree of bone and joint involvement categorized in four stages of increasing severity. Comparison

of the rheumatoid episcleritis/scleritis group (mean http://bjo.bmj.com/ DURATION OF RHEUMATOID ARTHRITIS; FUNCTIONAL value: 2-97) and the group of rheumatoid controls CLASS; ARTICULAR INDEX, AND X-RAY STAGE (Table XVII) (mean value: 242) indicated that episcleritis and scleritis occurred in a group with more severe The range of duration was huge in both rheumatoid erosive arthritis. The difference is statistically episcleritis/scleritis and in rheumatoid controls, significant (P = OOI). from a few months to 50 years. However, episcleritis and scleritis occurred in patients with longer- SUBCUTANEOUS NODULES; PERIPHERAL NEUROPATHY; on October 2, 2021 by guest. Protected copyright. standing rheumatoid arthritis (mean values 14-5 SKIN ULCERS; SKIN BRUISING; SKIN ATROPHY; years compared with 9-2 years). The difference is statistically significant (P=o0os). AMYLOID; TENDON RUPTURE AND SPLENOMEGALY (Table XVIII) The functional class measures in four grades of increasing severity the degree of functional inca- Of 46 cases with rheumatoid episcleritis or scleritis pacity due to arthritis. The study of functional 23 had subcutaneous granulomatous nodules

Table XVIII Extra-articular features in patients with rheumatoid episcleritis and rheumatoid scleritis

Rheumatoid episcleritislscleritis Rheumatoid controls Extra-articular features No. of cases No. of cases P value No. of cases positive Per cent No. of cases positive Per cent Subcutaneous nodules 46 23 50-0 46 II 23.9 < 005 Peripheral neuropathy 45 3 8-9 46 I 2-2 NS Skin ulcers 45 6 13-3 46 2 4 4 NS bruising 46 Is 23 9 46 5 IO-9 NS atrophy 46 i6 34-8 46 5 IO-9

(S° ° per cent). Comparison with the incidence in rheumatoid episcleritis/scleritis group and five in the rheumatoid control group (23-9 per cent) the rheumatoid control group had been given showed a statistically significant difference systemic corticosteroid treatment. (P=O OI). There were two men considered as having Three of 45 known cases of rheumatoid episcleri- secondary amyloidosis; both had bilateral rheu- tis or scleritis (8-9 per cent) revealed a high inci- matoid scleritis. One case was confirmed by liver dence of peripheral neuropathy as compared with biopsy. No case of amyloidosis was confirmed in only one case in the rheumatoid control group the group of rheumatoid controls. (2-2 per cent). However, the difference is not No patient with rheumatoid episcleritis/scleritis statistically significant. Two of the patients with and none of the rheumatoid control group suffered rheumatoid episcleritis/scleritis had received treat- tendon rupture. ment with systemic corticosteroids and two had The incidence of splenomegaly in the rheumatoid been given gold therapy (one patient had received episcleritis/scleritis group was five of 46 cases both). Either of these treatments could have pro- (IO9 per cent). Only one patient of 46 rheumatoid duced neuropathy. The patient in the rheumatoid controls (2z2 per cent) had a palpable spleen. The control group had not received systemic treatment difference for these numbers is not statistically with corticosteroids or gold. significant. Six of 45 known cases with rheumatoid episcleritis or scleritis (I3 3 per cent) had skin ulceration as Laboratory investigations compared with two of 46 cases in the rheumatoid control group (4.4 per cent). The difference is not The following data record information found in statistically significant for these numbers. Three of six groups of patients; patients with rheumatoid the patients with rheumatoid episcleritis or scleritis episcleritis and rheumatoid scleritis, non-rheuma- had received treatment with systemic cortico- toid episcleritis and non-rheumatoid scleritis, and steroids and both patients in the rheumatoid patients with rheumatoid episcleritis and scleritis control group had received similar . combined as one group for comparison with the Skin bruising occurred more often in patients rheumatoid control group. with rheumatoid episcleritis or scleritis (i i cases; 23-9 per cent) as compared with the rheumatoid HAEMOGLOBIN; ERYTHROCYTE SEDIMENTATION RATE controls (5 cases; IO 9 per cent). The difference (ESR); ALBUMIN/GLOBULIN RATIO; ALKALINE for these numbers is not statistically significant. http://bjo.bmj.com/ Ten of the patients in the rheumatoid episcleritis/ PHOSPHATASE; URIC ACID; SERUM GLUTAMIC OXALOACETIC TRANSAMINASE (SGOT), and SERUM scleritis group had received treatment with systemic GLUTAMIC PYRUVIC TRANSAMINASE (SGPT) corticosteroids and four of cases the five positive (Tables XIXa, b, c) in the rheumatoid control group had been similarly treated. For the purposes of comparison haemoglobin Skin atrophy was found in i 6 patients in the results for both men and women are grouped rheumatoid episcleritis/scleritis group (34-8 per

together. In patients with rheumatoid episcleritis on October 2, 2021 by guest. Protected copyright. cent) as compared with five patients in the rheuma- the mean value was 87-6 per cent but in those with toid control group (Io 9 per cent); this is statistic- rheumatoid scleritis it was 83 4 per cent. The ally significant (P=o'oi). Ten of the patients in the mean haemoglobin value for patients with non-

Table XIXa Laboratory investigations in patients with rheumatoid episcleritis and rheumatoid scleritis Rheumatoid episleritis Rheumatoid scleritis Laboratory -- investigation No. of No. of P value cases Mean SD SEM cases Mean SD SEM Haemoglobin ESR 9 87-6 per cent +15-4 ±5-I 37 83-4 per cent ±i8-2 ±3-0 NS 7 74-4 mm/Ist h ±37-9 ±14-3 30 65-3 mm/ist h ±36-i ±6-6 NS (Westergren) (Westergren) Albumin/globulinAlumnglbi ratio 9 o-88 ±0 37 ±01-2 37 I.03 ±0-4I ±007 NS Alkaline phosphatase 9 12-07 KA units ±4-73 ±i-58 37 14-68 KA units ±12-73 ±2-09 NS Uric acid 8 4-6 mg/Ioo ml ±x-66 +0-59 35 4-88 mg/xoo ml ±1 47 ±0-25 NS SGOT 8 17-88 lU/i ±8-17 +2-89 30 29-33 IU/i ±43-23 ±7-89 NS SGPT 8 15-88 IU/I ±12-86 ±4-55 30 23-87 IU/I ±36-3 ±6-63 NS Normal values: haemoglobin 90-125 per cent male, 80oI0o per cent female; ESR 5-12 mm/sat h male, 0-7 mm female; A/G ratio 1-03- 2-36 g/Ioo ml; alkaline phosphatase 5-15 KA units; uric acid 22-7-2 mg/soo ml male, 0-5-5-7 mg/ioo ml female; SGOT IU/I; SGPT 4-28 IU/I 11-35 Episcleritis and scleritis 207 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from

Table XIXb Laboratory investigations in patients with non-rheumatoid episcleritis and non-rheumatoid scleritis Non-rheumatoid episcleritis Non-rheumatoid scleritis Laboratory No. of No. of P value investigations cases Mean SD SEM cases Mean SD SEM Haemoglobin 31 99-9 per cent ±9-2 ±i-6 8 98-4 per cent 8-o ±1-9 NS ESR 27 13-3 mm/ist h ±12-8 ±25 I5 I9-9 mm/Ist h ±13-0 ±3 4 NS (Wintrobe) (Wintrobe) Albumin/globulin ratio 33 x-75 ±0-38 ±0 07 x8 I553 ±0°33 o-o8 < 0o05 Alkaline phosphatase 33 7-36 KA units ±2-6i ±0'45 x8 8-o6 KA units ±3-33 ±0o79 NS Uric acid 33 5-39 mg/0oo ml ±1-90 ±0 33 x8 4 41 mg/0oo ml ±1i-25 ±0 29 NS SGOT 32 20z56 IU/I ±7-9 ±'t-4 x8 22-67 IU/i ±i6-04 +3-78 NS SGPT 32 I9-88 IU/I ±9 13 ±i-6I i8 19-39 IU/i ±12-7 ±2 99 NS Normal values are given with Table XIXa

Table XIXc Laboratory investigations in patients with rheumatoid episcleritis and rheumatoid scleritis: comparison with rheumatoid controls

Rheumatoid episcleritis/scleritis Rheumatoid controls Laboratory No. of No. of P value investigations cases Mean SD SEM cases Mean SD SEM Haemoglobin 46 84-3 per cent ±17-6 ±2-6 45 90-I per cent ±12-3 ±i-83 NS ESR 37 67-1 mm/Ist h ±36-I ±5-9 45 48-4 mm/ist h ±34' ±5-I <0-025 (Westergren) (Westergren) Albumin/globulin ratio 46 1I0 ±0°4 ±o-6 40 PI-2 ±0°39 ±o-o6 NS Alkaline phosphatase 46 14-2 KA units ±ii-6 ±17 4I ii-5 KA units ±3-7 ±o-58 NS Uric acid 43 4-8 mg/ioo ml ±I-5 ±0-23 40 4-36 mg/ioo ml ±I.14 ±o-i8 NS SGOT 38 26-9 IU/I ±38-7 ±6-3 27 14-6 IU/1 ±5-I ±I-0 NS SGPT 38 22-2 IU/I ±32-8 ±5 3 23 12-0 IU/l ±5 5 ±I-I NS Normal values are given with Table XIXa http://bjo.bmj.com/ rheumatoid episcleritis was 99 9 per cent, signifi- had a mean value of I3-3 mm/Ist hour and those cantly higher than in those with rheumatoid epi- with non-rheumatoid scleritis gave a mean value of scleritis (P=o0oos). A similar significant difference I9-9 mm/Ist hour (Wintrobe). The ESR in rheu- was found forpatients with non-rheumatoid scleritis, matoid episcleritis/scleritis patients was higher than in whom the mean value for haemoglobin was 98-4 in the rheumatoid control group; mean values were per cent, compared with patients with rheumatoid 67 i mm compared with 48-4 mm. The difference

scleritis (P=o0oo5). Comparison of the rheumatoid in these values is statistically significant (P=0 025). on October 2, 2021 by guest. Protected copyright. episcleritis/scleritis group with rheumatoid controls The mean value of the albumin/globulin ratio in suggested that the anaemia might be more severe rheumatoid episcleritis was o88 while that for in patients with eye inflammation (84-3 to 90 I per rheumatoid scleritis was I103. There was a high cent) but the difference is not statistically significant percentage of reversed ratios in both rheumatoid for these numbers. episcleritis (66-7 per cent) and scleritis (62-2 per For the purposes of comparison ESR readings cent). Non-rheumatoid patients with episcleritis for both men and women are grouped together. showed a mean value of I-75 and those with The ESR in rheumatoid inflammation was very scleritis a mean value of I-53. The difference much higher than that in patients with non- between these two figures is significant (P=o0o5). rheumatoid episcleritis or scleritis. Unfortunately Neither of the groups with non-rheumatoid inflam- the methods differed in the two groups. The mation had any patient with a reversed ratio. There Westergren method was used at the Centre for was a highly significant difference between rheuma- Rheumatic Diseases and the Wintrobe method at toid and non-rheumatoid episcleritis (P=o-ooi) and the Tennent Institute. There is, however, an rheumatoid and non-rheumatoid scleritis (P = o-oo i). obvious difference in severity which does not The mean value of the albumin/globulin ratio in require exact comparison. The mean value in the rheumatoid episcleritis/scleritis group (i -o) rheumatoid episcleritis was 74-4 mm/ist hour and was more abnormal than the group of rheumatoid in rheumatoid scleritis 65-3 mm/ist hour (Wester- controls (I i2), but the difference was not statisti- gren). Patients with non-rheumatoid episcleritis cally significant. Consistent with these findings, the Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 208 British Journal of Ophthalmology

number of patients with a reversed albumin/globulin One patient with rheumatoid episcleritis (I4-3 ratio was greater in the combined episcleritis/ per cent) had a raised ASO titre at 724 units and scleritis group (63-o per cent) than in the rheuma- six with rheumatoid scleritis (20 7 per cent) had toid control group (47.5 per cent). raised titres at 320 units (2), 400 units (2), 500 None of the mean values in all six groups units, and 640 units. Three patients with non- studied is outside the normal range for alkaline rheumatoid episcleritis (9 i per cent) showed raised phosphatase of 5-I5 King Armstrong units, ASO titres, two readings at 320 units and one at although it is interesting to note that the mean I280 units. One case of non-rheumatoid scleritis value for rheumatoid scleritis is at the upper limit (5-6 per cent) had a raised titre at 320 units. None of normal, and also that IO of 37 cases with rheuma- of the patients with a raised ASO titre had erythema toid scleritis (27-o per cent) had raised readings. nodosum. Further, it is noteworthy that there is a statistically Only one of 28 patients with rheumatoid scleritis significant difference between the levels for rheu- (3y5 per cent) had a tentative positive result for matoid episcleritis and non-rheumatoid episcleritis brucellosis. The standard agglutination test gave a (P=O ooi) and between rheumatoid scleritis and titre of i /8o for Brucella abortus, but the Coombs' non-rheumatoid scleritis (P = o0o5). test was negative. Serum was anti-complementary. As with alkaline phosphatase, none of the mean One patient with non-rheumatoid episcleritis (3-1 values for serum uric acid levels in the six groups of per cent) had a positive titre for Brucella abortus patients investigated was outside the normal range. at i/i6o in both the standard agglutination and A similar pattern emerged for SGOT and Coombs' test. However, the complement-fixation SGPT, where all mean values recorded were within test was negative. normal limits. The only group of patients with a percentage incidence of positive toxoplasmosis dye test titres the normal adult range, approximately WASSERMANN REACTION; VDRL SLIDE TEST, AND outside 30-40 per cent (Ludlam, 1972), was the very small REITER PROTEIN COMPLEMENT-FIXATION TEST group of seven cases with rheumatoid episcleritis The following data relate only to those patients in which the percentage was 71I4 per cent (Tables with eye inflammation, both rheumatoid and non- XXa, b). It is interesting to note that 8-6 per cent rheumatoid. Seven of the patients with rheumatoid of the patients with rheumatoid episcleritis or episcleritis and 28 of the patients with rheumatoid rheumatoid scleritis had a titre of I/256 or more. scleritis had these investigations and no patient The group with non-rheumatoid episcleritis and http://bjo.bmj.com/ gave a positive result. All the patients with non- non-rheumatoid scleritis showed only one in the rheumatoid episcleritis and with non-rheumatoid same range (I/1024), 2-0 per cent. It is probable scleritis gave similarly negative results. that in the normal adult population the incidence of antibodies at a titre of 1/256 is not more than I1O ANTISTREPTOLYSIN-O TITRE (ASO); BRUCELLOSIS AND per cent (Beattie, 1958; Ludlam and Beattie, I963). TOXOPLASMOSIS AUTOANTIBODIES (Tables XXIa, b) The following data relate only to those patients on October 2, 2021 by guest. Protected copyright. with rheumatoid episcleritis and scleritis, and non- The following results refer only to those patients rheumatoid episcleritis and scleritis. with rheumatoid and non-rheumatoid eye inflam-

Table XXa Toxoplasma dye test titres in 35 cases (rheumatoid episcleritis and rheumatoid scleritis)

Disease Cases < i/8 +I/8 +i/i6 +I/32 +±/64 +I/I28 +1/256 +1/512 +I/I024

Episcleritis 7 (2 unknown) 2 - I 2 - I - - I Total+ 5 (7I-4 per cent) Scleritis 28 (9 unknown) I5 I 6 Total+ 13 (46-4 per cent)

Table XXb Toxoplasmosis (non-rheumatoid episcleritis and non-rheumatoid scleritis) 5 I cases

Disease Cases < I/8 + I/8 +I/i6 +I/32 + I/64 + I/I28 + I/256 +1/512 + I/I024

Episcleritis 33 17 5 - 8 2 I - - - Total + i6 (48-5 per cent) Scleritis I8 I2 - I I 3 - - - I Total+ 5 (27-8 per cent) Episcieritis and scleritis zog Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from

Table XXIa Autoantibodies (rheumatoid rheumatoid factor in patients with non-rheumatoid episcleritis and rheumatoid scleritis) 46 cases scleritis, six of i8 cases (33-3 per cent). Also, six patients had a positive antinuclear factor (33-3 Episcleritis (9 cases) per cent); in two of them this was weakly positive. Autoantibody Cases (no.) Per- Three patients had a positive rheumatoid factor + Weakly + centage together with a positive antinuclear factor. The

Rheumatoid latex 9 9 0 I00 remaining autoantibody investigations in the non- Antinuclear factor 9 2 0 22'2 rheumatoid patients were in the range accepted for Thyroglobulin TRC 9 0 0 - a Thyroid microsomal 9 I 0II-I general population (see Discussion). Smooth muscle 8 (I unknown) I 12-5 Salivary duct 9 4 0 44-4 Gastric parietal cell 9 3 33-3 R3 TITRE (Table XXII) Mitochondrial 8 (I o unknown) Sixteen of the rheumatoid episcleritis/scleritis Scleritis (37 cases) group (42-I per cent) had titres of 1/5I2 or higher Rheumatoid latex 37 35 I 97.3* compared with eight of the rheumatoid controls Antinuclear factor 37 8 I 24 3 (2I*6 per cent). The difference was just outside Thyroglobulin TRC 37 3 0 8-i for these numbers Thyroid microsomal 37 4 2 56-2 significance (X2=3-6148). Smooth muscle 34(3unknown) 2 6 23 5 Salivary duct 35(2unknown) 5 5 28-6 Gastric parietal cell 37 5 I 56-2 IC/ IA GLOBULIN (HUMAN COMPLEMENT C3); Mitochondrial 34 (3 unknown) o IMMUNOGLOBULIN IgG, IMMUNOGLOBULIN IgA, AND IMMUNOGLOBULIN IgM (Tables XXIIIa, b) *One ca3e is included in which the Rose-Waaler test was used These data also relate only to those patients with rheumatoid and non-rheumatoid eye inflammation. Table XXIb Autoantibodies (non-rheumatoid The mean values for rheumatoid episcleritis and episcleritis and non-rheumatoid scleritis) 5I cases rheumatoid scleritis, and non-rheumatoid episcleri- Episcleritis (33 cases) tis and non-rheumatoid scleritis, all fall within the Autoantibody normal range for each of these investigations. + Weakly + Percentage Tables XXIIIa, b record the occasional readings

Rheumatoid latex 0 I 3-0 which do not fall within the normal range. These Antinuclear factor I 2 9-I variants predominated in the investigation of ic/ia Thyroglobulin TRC 0 l 3-0 Thyroid microsomal 3 I 12-1 globulin, where six patients with rheumatoid http://bjo.bmj.com/ Smooth muscle 0 0 episcleritis or rheumatoid scleritis and six patients Salivary duct 0 0 Gastric parietal cell I 4 1552 with non-rheumatoid episcleritis and non-rheuma-

Mitochondrial 0 toid scleritis had depressed levels. Scleritis (i8 cases) Rheumatoid latex 0 6 33-3 BACTERIOLOGY; VIROLOGY, AND MYCOLOGY Antinuclear factor 4 2 33.3 Thyroglobulin TRC 0 0 Examination of the conjunctiva for bacteria, viruses,

Thyroid microsomal I 0 5*6 on October 2, 2021 by guest. Protected copyright. Smooth muscle I I and fungi was carried out only in those patients Salivary duct o o I I*I seen at the Tennent Institute of Ophthalmology. Gastric parietal cell 2 0 One patient with non-rheumatoid episcleritis Mitochondrial o 0 had a coagulase-negative staphylococcus. One patient with rheumatoid scleritis and one patient mation. Apart from the relatively high incidence of salivary duct antibody in the small group with Table XXII R3 titres rheumatoid episcleritis, where a positive result was given in four of nine cases (44-4 per cent), and an Episcleri tis/scleritis controls 20- Rheumatoid apparently high incidence of smooth-muscle anti- 38 coses (8 unknown) 37 cases (6 unknown)* body in rheumatoid scleritis, where eight of 34 5 cases were positive (23-5 per cent) (although six of -these were only weakly positive) the positive autoantibodies listed in Table XXIa in patients v0-9 with rheumatoid episcleritis and rheumatoid scleri- 6 61 z r, In-1 5 5 tis is generally in the range expected for patients with rheumatoid arthritis (see Discussion). The possibly significant finding in the non-rheumatoid patients is the high incidence of a weakly positive Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 2IO British Journal of Ophthalmology

Table XXIIIa 3lc/[la globulin and immunoglobulins IgG, igA, and IgM in rheumatoid episcleritis and rheumatoid scleritis Rheumatoid episcleritis Rheumatoid scleritis Outside normal Outside normal Globulin or Cases Mean SD SEM Below Above Cases Mean SD SEM Below Above immunoglobulin (no.) (mg/ioo ml) (no.) (mglioo ml) - - Per Per Per No. cent No. cent No. cent 3lc/13la globulin 8 I26-5 ±44.I ±i56 2 250- 33 142-3 ±36-7 ±6-4 4 12-1 2 6-i Immunoglobulin igG 7 1365-7 ±546-4 ±2o6-5 - - 33 I717-3 ±790-2 ±137-6 I 3-0 2 6-i igA 7 3914 ±i52-8 ±57-8 - - 33 488-5 ±259'3 ±452 - 3 9-1 IgM 7 141-7 ±57-8 ±21-9 - - 33 193-3 ±112-0 ± 9-5 - - Normal values: i3lc/,3la globulin Ioo-i9o mg/ioo ml; immunoglobulin igG 560-2800 mg/Ioo ml; immunoglobulin igA I04-845 mg/OO ml; immunoglobulin IgM 54-443 mg/1oo ml

Table XXIIIb Ilc/lla globulin and immunoglobulins igG, igA and IgM in non-rheumatoid episcleritis and non-rheumatoid scieritis Non-rheumatoid episcleritis Non-rheumatoid scleritis Outside normal Outside normal Globulin or Cases Mean SD SEM Below Above Cases Mean SD SEM Below Above immunoglobulin (no.) (mg/loo ml) (no.) (mg/Ioo ml) Per ~~~~~~Per Per No. cent No. cent No. cent

,3lc/3la globulin 29 1345 ±319 ±5-9 3 10-4- I6 148-8 ±44-2 ±11-0 3 i8-8 i 6.3 Immunoglobulin igG 29 1225-2 ±491-7 ±91-3 - - 17 13341 ±339-4 ±82 3 - - igA 29 341-5 ±179-7 ±33-4 I 3 4 17 403-2 ±253-6 ±6i-5 I 5-9 I 5.9 xgM 29 I69-0 ±86-o i±6-o I 3-4- I7 202-3 i09-8 ±267 - I 59 Normal values are given with Table XXIIIa http://bjo.bmj.com/ with non-rheumatoid scleritis was a carrier of Table XXV Cause of death and associated disease Stayhylococcus albus. recorded on death certificates (17 cases) No case in the series had any virus cultured from the conjunctiva. Rheumatoid episcleritis/ Rheumatoid scleritis controls Two patients with non-rheumatoid episcleritis (iI cases; 4 unknown) (6 cases) had fungi isolated from the conjunctiva, in both cases the Penicillium species. Two patients with No. Percentage No. Percentage rheumatoid scleritis had fungi present, in one case Myocardial infarction and on October 2, 2021 by guest. Protected copyright. Penicillium and in one Aspergillus. One further case congestive cardiac failure 5 45.5 4 66-7 Peptic ulceration 3 27-3 0 with non-rheurnatoid scleritis produced a fungus Cerebrovascular accident 9-' 2 33-3 of the Paecilomyces species. Bronchopneumonia I6.7 Pulmonary embolus 00 Pulmonary oedema 0 Hypertension 0 MORTALITY (Tables XXIV and XXV) Aortic stenosis and incompetence 0 The following data relate to those patients with Hepatitis 0 0 rheumatoid episcleritis and rheumatoid scleritis, 'Natural causes' I 0 Myeloproliferative disease I1 0 Renal failure 0 I1 Perforated diverticulum of colon 0 I Table XXIV Mortality (rheumatoid episcleritis, Diabetes (maturity onset) 0 rheumatoid scleritis and rheumatoid controls) 66 cases Disease Cases Deceased Percentage Rheumatoid episcleritis/ comparing them with the rheumatoid control scleritis 33 (2 unknown)* I5 45-5 group, matched for age, sex, and year of presenta- Rheumatoid controls 33 (2 unknown)* 6 18-2 tion. Statistical analysis: X2=5-6571; P= <0-025 By February 1974, 15 of 33 known cases of *These four patients could not be traced rheumatoid episcleritis or rheumatoid scleritis Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 2Ir

Table XXVI Type I allergy (atopy) 23 patients House dust Age Sex Diagnosis Control House mite Grass Alternaria Feathers Aspergillus Candida Cat Dog (years) dust (scratch test) pollen albicans hair hair

34 Female Episcleritis - 40 Female Scleritis 4- 34 Female EpiScleritis 37 Female Episcleritis - 2-4-- so Female Episcleritis - 22 Male Episcleritis 76 Female Scleritis - 61 Female EpiScleritis - 35 Female Episcleritis - 36 Female Episcleritis -

5' Female Scleritis +- t- - 46 Male Episcleritis 47 Female Scleritis - 67 Female Episcleritis - 40 Male Episcleritis 40 Male Episcleritis - 21 Male Episcleritis + 34 Female Scleritis - 49 Female Episcleritis - 5o Male Episcleritis - 70 Female Episcleritis 43 Male Scleritis - 50 Male Episcleritis - _-

(45-5 per cent) had died since examination. The Lyne and Pitkeathly (I968), Fowler (1970), Jayson control group of patients by February 1974 gave and Jones (I97I), with follow-up studies by Jones only six of 33 known cases (i8'2 per cent) who and Jayson (I973), Watson (1974), and Lyne were deceased. The difference is statistically (1974), are the sum total of reviews. The incidence significant (P=0'025). of scleritis in rheumatoid arthritis in the study The details of cause of death, as recorded on the presented has some similarity with the majority of death certificates of these patients, give the expected reports previously documented (Table XXVII). high incidence of cardiovascular and respiratory Jayson and Jones (I971) in a group of I42 patients http://bjo.bmj.com/ disease. One unexpected feature was the finding had a notably dissimilar figure and Sevel's estimate that three patients of the rheumatoid episcleritis/ of one case of scleritis in 3000 patients with rheuma- scleritis groups (27-3 per cent) died after gastro- toid arthritis is much lower than our own conclu- intestinal haemorrhage. sions. As with the review of Jayson and Jones (I971) this present series was a prospective study for incidence. The incidence of episcleritis in TYPE I ALLERGY (ATOPY) (Table XXVI) with rheumatoid arthritis has patients presenting on October 2, 2021 by guest. Protected copyright. Skin tests in 23 patients, all with non-rheumatoid not been documented elsewhere, which is under- eye inflammation, and all but six of them cases of standable as the condition is relatively uncommon, episcleritis, did not suggest that there is an in- at least in so far as presentation is concerned. Only creased incidence of atopic disease in these patients. seven patients, an average of approximately one for Of the 23 patients, four had a mild positive reaction each year of the study, are included in this group. to house dust or house dust mite, and five had positive skin tests to more than one antigen. Three of the patients with positive skin tests gave a clear Table XXVII Scleritis in rheumatoid arthritis history of either hay fever or asthma; the remaining patients with positive skin tests had no clinical Patients with Authors Date Percentage history of type I allergy. Rheumatoid arthritis Sdleritis Discussion Smith 1957 465 2 0-4 Bywaters I 964 1000 1-2 015 OPHTHALMOLOGY Hart I964 900 ±7 o-8 Richardson I964 100 ±3 0-3 Sevel I965 3000 00-03 There have been few reports reviewing episcleritis (Estimate) and scleritis; most papers have been on individual Jayson and cases of interest. Studies by Watson and Lobascher Jones 1971 142 9 6.3 (I965), Sevel (I965), Watson (I966), Sevel (I967), Present study 1976 4210 28 o-67 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 212 British 7ournal of Ophthalmology

Reports giving the incidence of rheumatoid but these were eyes which required enucleation, arthritis in episcleritis and scleritis (Table XXVIII) and so were particularly severe examples. Nine are consistent with the findings of this study. It patients in our study with rheumatoid episcleritis will be interesting to follow up the patients without gave an almost exactly similar mean age as the apparent rheumatoid disease some years hence, patients with rheumatoid scleritis (Table IIIa). particularly in view of the finding of one patient Non-rheumatoid inflammation presented a lower with episcleritis and six with scleritis, each of mean age, particularly for episcleritis (Table IIIb). whom had a weakly positive rheumatoid factor. Lyne and Pitkeathly (I968) reported a similar A very interesting question posed by Sevel picture in 55 cases of episcleritis with men having (I966) is whether scleritis has been seen more often a mean age of 45 years and women 48 years, but in recent years. A histogram of his 43 cases which their figures included three cases of rheumatoid had presented at the Pathology Department, Insti- arthritis. Their 3I cases of scleritis, including nine tute of Ophthalmology, London, between 1950 cases of rheumatoid arthritis, had a mean age of and I964 suggests an increased incidence of scleritis 49 years for men and 50 years for women. It may since I954. Studies in rheumatoid arthritis suggest be that rheumatoid episcleritis and scleritis tend to that the incidence of rheumatoid vasculitis has affect a slightly older age-group of patients with increased since the advent of treatment with sys- rheumatoid disease. Two of us (JW and DDMM) temic corticosteroids (Ferguson and Slocumb, examined a group of 148 patients with rheumatoid 196I; Smyth and Gum, 196I), although this view arthritis for evidence of steroid-induced cataract, does not have unquestioned acceptance (Schmid, and their mean age was 52-5 years. The control Cooper, Ziff, and McEwen, I96I). It may be that group of 159 patients with rheumatoid arthritis, -the same vasculitis that is possibly related to who had never been treated with systemic cortico- systemic corticosteroid treatment, in particular to steroids, had a mean age of 49-3 years (William- variations in dosage (Slocumb, 1953), has been son, Paterson, McGavin, Jasani, Boyle, and Doig, -instrumental in initiating a greater frequency of I969). In the progression of the disease the most scleritis. Systemic corticosteroids were first used in severe cases affect an even older age-group; this the treatment of rheumatoid arthritis in I949 was evident in the pathological series of Sevel (Hench, Kendall, Slocumb, and Polly, 1949) and (I966). The predominance of women with rheu- in the treatment of scleritis in 195I (Talkov, matoid ocular inflammation (Table IMIc) reflects Colpoys, Davis, Papper, and Fienberg, 1951). the higher incidence of women with rheumatoid The only reports giving the mean age of a of or and patients arthritis, ratio 2 3:1 (Boyle Buchanan, http://bjo.bmj.com/ with scleritis in rheumatoid arthritis are those of I97I), and this agrees with previous studies (Sevel, Sevel (I966), and Jayson and Jones (I97i). The I965; Jayson and Jones, 197I). Non-rheumatoid nine patients of Jayson and Jones had a mean age inflammation presents us with an intriguing prob- of 56-5 years, which is similar to the figure given lem. The report of Lyne and Pitkeathly (I968), here (Table IIIa). Sevel's series of II cases with with only three cases of rheumatoid arthritis in rheumatoid scleritis had a mean age of 64-I years, their 55 cases of episcleritis, supports the evidence of this study that non-rheumatoid episcleritis occurs equally in the sexes (Table IIId). If episcleri- on October 2, 2021 by guest. Protected copyright. Table XXVIIIa Rheumatoid arthritis in scleritis tis is due to immune-complex disease and has a Patients with similar basis in this respect to scleritis, as suggested Authors Date Percentage by Jones (197I), why is there such a notable differ- Rheumatoid ence in the respective sex ratios? The sex ratio in Scieritis arthritis non-rheumatoid scleritis (Table HId), together Watson I966 Io8 35 32-4 with the finding that one-third of these patients Lyne and Pitkeathly I 968 31 9 29-0 had a weakly positive rheumatoid factor (one was a Present study 1976 27 9 33-3 man), gives some concern for the future of these patients. An unusual connective-tissue disorder seen in Table XXVIIIb Rheumatoid arthritis in episcleritis the present study, in a 45-year-old man with rheumatoid scleritis, presented with widespread Patients with Authors Date Percentage of orbits, nasal septum, sinuses, and Rheumatoid lungs, which had occurred intermittently for more Episcleritis arthritis than Io years. The clinical picture was less severe Watson i966 I17 4 3 4 than that of a true Wegener's disease, and lacked Lyne and the glomerulonephritis typically present in this Pitkeathly I968 55 3 5-5 Present study 1976 35 2 5-7 disease (Godman and Churg, 1954). A chronic eventually required dacryocystec- Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 213

tomy; histological examination showed no necro- streptococcal infection, sarcoidosis, tuberculosis, tizing vasculitis but did show that finger-tip drug reactions (for example, to sulphonamides), vasculitis, subcutaneous nodules, and skin ulcera- rheumatic fever, gonorrhoea, lymphogranuloma tion had occurred. This case may fall into the inguinale, and coccidioidomycosis (Fry, I952). category of a limited form of Wegener's disease, as James (I96I) could not give any aetiological factor described by Carrington and Liebow (I966) in i6 in only 23 of I70 cases of erythema nodosum. He patients, nine of whom had pulmonary lesions only. found evidence of sarcoidosis in I26 and infection, Four similar cases have been reported more recently usually streptococcal, in 2I. None of the three (Cassan, Coles, and Harrison, 1970). Other connec- patients in this study had a raised ASO titre or any tive-tissue disorders found in the non-rheumatoid definite history of the conditions listed. The epi- groups of patients were chronic discoid lupus scleritis in both cases was 'simple', but the patient erythematosus associated with episcleritis, and with scleritis had a nodular lesion. An anterior systemic lupus erythematosus associated with uveitis was present in the case of nodular scleritis, scleritis (one case of each). There is some doubt and this has been documented previously (Allan, whether these two conditions are separate entities, I90I). Duke-Elder and Leigh stated that scleral although systemic manifestations are rare in the keratitis was a rarity, citing Palich-Szanto (I959). chronic discoid form of lupus erythematosus (Boyle One of the cases of episcleritis, which was bilateral, and Buchanan, I971). Systemic lupus erythemato- had associated bilateral sclerosing keratitis. sus has been described previously in two cases of It is uncertain why many patients with rheuma- episcleritis by Harvey, Shulman, Tumulty, Conley, toid episcleritis and rheumatoid scleritis had a and Schoenrich (I954) and in isolated reports by marked increase in electrocardiographic abnor- Larson (I96I) and Spaeth (I967). Associated malities (Table IVa). It is possible that the changes, scleritis has been reported in two cases (Sevel, I965; which include bundle-branch block myocardial Watson, 1966). No previous paper gives episcleritis infarction, and non-specific S-T wave changes, as a complication of chronic discoid lupus erythe- might have been in part, at least, due to rheumatoid matosus. involvement of the heart, but without necropsy In the late nineteenth and early twentieth it is impossible to differentiate between rheumatoid century, gout was considered to be an important disease, induced ischaemia, and other causes of cause of ocular inflammation, but the association ischaemic heart disease (Hart, I969). Pitkeathly, in most cases of episcleritis or scleritis is vague and Howitt, and Lyne (1970) documented three cases indefinite (Duke-Elder and Leigh, I965). During of aortic incompetence in patients with scleritis, http://bjo.bmj.com/ this century reports of episcleritis have been re- one of whom had a history of polyarthritis of corded by Gilbert (1914), Savin (1938), and Mc- uncertain origin. They presented the scleritis and Williams (I952). In recent years only Watson and aortic incompetence as two manifestations of Lobascher (1965) reported one case with an epi- connective-tissue disease. Both patients had right scleritis. Two male patients with gout were seen in bundle-branch block; this conduction defect was the present study; one with non-rheumatoid epi- found in three patients with rheumatoid scleritis. scleritis and one with a A further case of aortic non-rheumatoid necrotizing incompetence associated on October 2, 2021 by guest. Protected copyright. scleritis. The necrotizing element developed some with rheumatoid scleritis was described by Jones weeks after the onset of scleritis, and might have and Jayson (I973) in their follow-up study (Jayson been iatrogenic in origin. The patient had a diffuse and Jones, 1971). In one male patient in this study scleritis which progressed to develop widespread with diffuse rheumatoid scleritis aortic stenosis and scleral translucency in the upper half of the anterior incompetence were discovered post mortem. One sclera. A subconjunctival injection in the upper and with rheumatoid episcleritis had aortic stenosis and lower fornices was followed by an area of necrosis one with non-rheumatoid scleritis had aortic above (Fig. 3). The eye subsequently settled satis- incompetence. None of these patients had a history factorily. Subconjunctival injections of of rheumatic fever. According to Sokoloff (I972) have been incriminated in the aetiology of scleritis rheumatoid disease may occasionally cause a (Tooker, 1931; Swan and Butler, 195I). Three of valvulitis with incompetence (Van Valkenburgh, the five cases of Swan and Butler had skin sensi- Georges, and Irby, I972). Pericarditis has been tivity reactions to atropine. described in association with rheumatoid arthritis Erythema nodosum, a condition previously and scleritis (Smoleroff, I943; Edstrom and associated with episcleritis (McCarthy, I96I) and Osterlind, 1948; Jayson and Jones, 197I). Two reported with scleritis (Watson and Lobascher, patients in the present study had clinical evidence 1965) was evident in two women with non-rheuma- of pericarditis; one had rheumatoid scleritis and toid episcleritis, and one woman with non-rheuma- the other non-rheumatoid episcleritis. Pericarditis toid scleritis. There are many common disorders is a recognized complication of rheumatoid disease associated with erythema nodosum, including (Harrold, I968; Bacon and Gibson, 1974) and it Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 214 British Journal of Ophthalmology

was noted by Bonfiglio and Atwater (I969) that I965) and colitic arthritis or acute toxic arthritis patients with pericarditis more often had severe (Fernandez-Herlihy, 1959). Regional ileitis (Crohn's joint disease. Kirk and Cosh (I969) found that io disease) exhibits the same type of peripheral of I00 patients with rheumatoid arthritis had arthritis (Ferguson, 1972). This form of arthritis clinical features of pericarditis. was first described by Hench (I935). It begins A striking increase in pulmonary complications - abruptly and reaches its height of intensity within such as pleurisy, pneumonitis, pleural effusion, the first day or so, involving the small joints of and rheumatoid nodules - was apparent in patients hands or feet less often than does rheumatoid with rheumatoid episcleritis and rheumatoid scleritis arthritis (Wright and Watkinson, I965). (Table IVa). Pleural effusion is a recognized entity It has been suggested that peptic ulceration in rheumatoid disease (Carr and Mayne, I962; occurs more commonly in rheumatoid arthritis Walker and Wright, I 967). This complication occurs (Atwater, Mongan, Wieche, and Jacox, I965; in severe rheumatoid disease as manifested by the Bowen, Mayne, Caine, and Bartholomew, I960; presence of subcutaneous nodules (Boyle and Meltzer, Bockman, Kanenson, and Cohen, 1958), Buchanan, 1971). Pulmonary nodules are themselves but both are such common diseases that it is unwise not uncommon (Noonan, Taylor, and Engleman, to be dogmatic. Four of our patients with rheuma- I963). This significant finding of pulmonary disease toid scleritis had peptic ulceration and three of the in patients with rheumatoid eye inflammation as II known to have died since examination had compared with rheumatoid controls, together with gastrointestinal haemorrhage recorded on their the possibility of greater cardiac involvement, death certificates. If the association is valid, is it emphasizes the serious nature of extra-articular caused by a vasculitis or is it a therapeutic conse- rheumatoid disease affecting vital structures with quence of, for example, salicylates? Certainly all the probable final seal of significance in the higher three patients who died of gastrointestinal haemorr- incidence of mortality in patients so affected. hage had been receiving salicylates. Against this it There is an increasing appreciation of the asso- is noteworthy that four patients with non-rheuma- ciation of ocular inflammation and joint disorders toid episcleritis, that is without severe disease, in patients with inflammatory bowel disease. also had a history of peptic ulceration. Wright, Lumsden, Luntz, Sevel, and Truelove Skin diseases reported with episcleritis are (I965) found that II8 per cent of I44 patients with eczema (Watson and Lobascher, I965), contact had iritis, and 17A4 per cent dermatitis and acne rosacea (Watson, I966),

had sacro-iliac joint abnormalities. Billson, de psoriasis, lichen planus, and erythema elevatum http://bjo.bmj.com/ Dombal, Watkinson, and Goligher (1967) substan- diutinum (Lyne and Pitkeathly, I968); and with tiated the association of eye inflammation and scleritis, epidermolysis bullosa dystrophica (Mazza ulcerative colitis, and Jayson, Salmon, and Harrison and Panagis, I967). Erythema nodosum, which (1970) found that 50 per cent of patients with has already been discussed, has been associated and colitis had an anterior with both episcleritis and scleritis. Psoriasis was uveitis. Lyne and Pitkeathly (I968) reported two evident in two women in the study, one with patients with ulcerative colitis associated with rheumatoid scleritis and one with non-rheumatoid scleritis, but without any connective-tissue disorder. episcleritis. There were no secondary eye complica- on October 2, 2021 by guest. Protected copyright. 'Watson (I966) described one case of regional tions in either of the patients with psoriasis. A ileitis associated with episcleritis, and Hopkins, further patient with non-rheumatoid episcleritis Horan, Burton, Clamp, de Dombal, and Goligher had a history of pityriasis rosea. (1974) recorded two patients with episcleritis Reviewing the literature of scleritis in rheuma- associated with regional ileitis. One patient in the toid arthritis, Sevel (I966) found that the condition present study with rheumatoid episcleritis had was more often bilateral. The ratio of 2:I agrees recurrent colitis, and one with rheumatoid scleritis with the findings given here (Table IIIe). In his had diverticulosis; both were women. The patient own series of II enucleated eyes, the disease was with recurrent colitis had an associated anterior unilateral in eight, an inconsistency which Sevel uveitis; the patient with diverticulosis had a was unable to explain. Extra-articular manifesta- particularly severe bilateral scleritis with sclerosing tions of rheumatoid disease are significantly keratitis, anterior uveitis, and secondary glaucoma apparent in this present study and the fact that in in one eye. However, while it is reported that most cases rheumatoid scleritis is bilateral reflects patients with ulcerative colitis and uveitis have a the widespread nature of extra-articular involve- threefold increase in peripheral joint disease ment. Non-rheumatoid scleritis was more often (Wright and Watkinson, I965; McEwen, I968) unilateral and whether or not these patients will the arthritis is regarded in most instances as distinct later develop autoimmune disease, it seems that from rheumatoid arthritis. It has been called they have a more localized disorder. Non-rheuma- arthritic ulcerative colitis (Ruhl and Sokoloff, toid episcleritis in contrast was more often bilateral Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 215

(Table IIIf) and characteristically recurrent, with affected mainly by anterior segment disease, parti- inflammation alternating from one eye to the other. cularly corneal inflammation and cataract. The It is generally recognized that the symptoms of review of posterior subcapsular cataract/advanced episcleritis are less distressing than those of scleritis cataract in rheumatoid patients with scleritis who (Watson and Lobascher, I965; Jayson and Jones, were receiving systemic corticosteroids also sug- 1971). One of the objects of this presentation was gests, although without supporting figures, a severe to provide evidence that these accepted differences effect on visual acuity (Table XII). Furthermore, do exist. Ocular pain was more apparent in scleritis. the group of patients with rheumatoid scleritis who The patients complained of a more often had a posterior segment abnormality would have in both rheumatoid and non-rheumatoid episcleritis some disturbance of vision (Table XIIIa). Fortu- than did those with scleritis (Tables Va, b). A nately, a cataract can readily be extracted and this distinctive finding was the higher incidence of should provide some reassurance to the clinician complaints of redness and pain in non-rheumatoid who considers that treatment with systemic cortico- inflammation. Perhaps the rheumatoid patient steroids is imperative. The male patient who suffers to such a degree elsewhere that the spon- developed gout, and subsequently scleritis, re- taneous complaint of a red and painful eye is not corded a visual acuity which dropped from 6/6-2 always given. However, Jones and Jayson (1973) to 6/i8-i in four weeks, due to the development of a pointed out that the fewer complaints in patients variable myopic . This eventually with rheumatoid eye inflammation might be due to settled coincident with quiescence of his scleritis the many anti-inflammatory drugs which these as a mixed astigmatism which, on refraction, gave patients were receiving, or possibly the higher a visual acuity of 6/5, a reminder that reduction in pain threshold which has been described in rheu- vision might have had a very simple solution. matoid arthritis (Huskisson and Hart, 1972). In his The abundance of terms used to describe epi- early description of 'sclerotitis', Mackenzie (I830) scleritis and scleritis requires a new simplicity of reported that a number of patients described their approach, and this was provided by the classification eyes as hot and dry at the onset of an episode of of Watson, Hayreh, and Awdry (I968a, b), which inflammation. It was interesting to discover a few was used in this study (Tables VIIIa, b). Clearly patients in this study who said that a dry, gritty, the pathological activity in scleritis and nodular or burning eye would often precede the flare-up of episcleritis has the same basis (Manschot, I96I; inflammation. Sevel, I967), and it only remains to give a descrip-

It was surprising to find that the history of eye tion of the appearance of any inflammation. There http://bjo.bmj.com/ inflammation was slightly longer in the non- has been much confusion regarding the terms rheumatoid group of patients, both for episcleritis 'scleromalacia' and 'scleromalacia perforans' (Van and scleritis (Table VIb). A simple explanation der Hoeve, 1934). 'Scleromalacia' is often used for might be the problem of recollection. Many patients a diffuse scleritis with scleral translucency, but this were examined early in the disease, but not all. is not the eye which is the most likely to perforate, With so many physical disabilities in what is gener- and so confusion results. According to the literature ally severe rheumatoid disease, the rheumatoid the eye which commonly perforates has a typically patient might simply have misjudged how long the necrotizing sclera, and the inflammation is necrotic on October 2, 2021 by guest. Protected copyright. eye trouble had lasted. Lyne and Pitkeathly (I968) and the sclera opaque but not translucent (Fig. 3). gave a mean duration for all forms of episcleritis of Most reports of scleromalacia perforans give a 20 months and a mean duration for all forms of distinct destructive picture to the scleral lesion, scleritis of 6o months. The difficulty is knowing describing such features as necrosis, sloughing, or when an eye is truly quiet and will remain so. ulceration (Van der Hoeve, 1934; Oast, I937; Many patients in this study had quiescent periods Verhoeff and King, 1938; Harbater, I949; Shaned- between inflammation lasting many months, and ling, 1950; Francois, I951; Talkov and others, in one case seven years. Nevertheless, a rather I951; Ashton and Hobbs, I952; Renard, Lelievre, uneasy impression given by the rheumatoid patients and Mazel, 1953; Ellis and Holtz, 1953; Armstrong is of the disease 'burning out' after two or three and McGovern, 1955; Marquard, I956; Boberg- years. Ans, 1958; Williams and Rosenthal, 1959; Taffet As expected, the deeper inflammation of scleritis and Carter, I96I; Francois, Victoria-Troncoso, has a more significant effect on visual acuity (Table Hanssens, and Bakskulin, I969; Gombos, I967). VII). The study of visual acuity was confined to Blum and Salmoun (I963) and Merz (I964) patients presenting at the Tennent Institute of reported cases where scleral translucency with Ophthalmology as cases of episcleritis or scleritis, alarming progressive ectasia necessitated surgical and did not include patients presenting with intervention. In these cases there was no necrosis rheumatoid arthritis at the Centre for Rheumatic or ulceration. Diseases. Visual acuity in these patients was The importance of these observations is the Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 216 British yournal of Ophthalmology

appreciation that areas of scleral translucency only ment in one, and extraocular muscle imbalance in rarely give cause for concern. Rheumatoid nodule five. In a paper on the clinical manifestations of of the sclera was the term used by Ashton and scleritis, Watson (I974) found that of 217 eyes Hobbs (I952) to describe the scleritis found in with episcleritis, 170 had the simple variety and 47 rheumatoid arthritis. It gives an accurate descrip- had nodular episcleritis. Of 30I eyes with scleritis, tion of the granulomatous reaction, similar to that II9 had diffuse anterior scleritis, 134 had nodular found in the subcutaneous nodule of rheumatoid anterior scleritis, 42 showed necrotizing scleritis, arthritis. However, a sub-division of the lesions into and six revealed features of a posterior scleritis. nodular, diffuse, or necrotizing scleritis provides a This series, easily the largest general study of clear and simple picture of the scleral reaction with episcleritis and scleritis, is particularly significant the premise that the inflammation is granulomatous. for its record of the clinical features of these The overall descriptive term for scleritis which conditions. satisfies the pathological uniformity of the condition Lyne and Pitkeathly (I968) noted that single is necrogranulomatous scleritis (Sevel, i967). It is attacks of episcleritis affected the interpalpebral interesting that none of our cases of rheumatoid area in 25 of 33 cases. This distribution of inflam- episcleritis was nodular in character, whereas in mation is substantiated by the present study non-rheumatoid episcleritis nearly one-fifth of (Tables IXa, b). Lyne and Pitkeathly suggested cases were nodular (Tables VIIIa, b). Two cases of that minor trauma might account for this distribu- rheumatoid inflammation initially diagnosed as tion and found a foreign body in one case. Certainly nodular episcleritis both progressed to a scleritis. external factors would seem to play a part, whether Only one patient in the study presented with a trauma or some other factor or factors. Patients necrotizing scleritis. A 63-year-old man developed sometimes complain of a dry, gritty, and burning a non-rheumatoid gouty scleritis, but the necro- eye and was seldom a problem. An eye tizing lesion discussed earlier was probably induced which is inflamed primarily deep to the conjunctiva by a subconjunctival steroid injection. Duke- may slightly increase in temperature. Evaporation Elder and Leigh (i965) described the possibility of of may ensue more rapidly in consequence and recurrent nodular episcleritis causing the circum- so irritation due to exposure will affect the inter- comeal region of the sclera eventually to become palpebral area preferentially. However, this could slate-coloured. We submit that any appearance of account only for aggravation of an already inflamed this nature is due to involvement of the sclera and area. It was thought that type I allergy (atopy) is no longer an episcleritis but a scleritis. The two might play a part in the external aggravation of an http://bjo.bmj.com/ patients diagnosed initially as having nodular episcleritis. However, in the present study no rheumatoid episcleritis subsequently had a distinct increase in atopic disease could be found in 23 slate-coloured zone at the site of inflammation, patients with non-rheumatoid eye inflammation, I7 and were therefore cases of scleritis. of whom had episcleritis. Inflammation affecting Transparency of the sclera which may accompany the anterior sclera in rheumatoid arthritis affects scleritis is often described as scleral thinning. This the upper half of the sclera more often than the

description may create apprehension in the mind lower half (Jayson and Jones, 197I), and Sevel's on October 2, 2021 by guest. Protected copyright. of the clinician, and a term such as scleral trans- review of the literature confirms this view. Sevel lucency is preferable. In a recent presentation (1965) in his iI cases examined histologically, involving one of us (DDMM), ultrasonographic found that i8-o per cent had disease of the posterior assessment of the thickness of areas of post-scleritic segment, and his investigation of the literature with scleral translucency showed that these areas were regard to eyes enucleated because of scleritis gave not invariably associated with thinning. Only three a figure of izvo per cent. These figures are not of eight patients had distinct evidence of thinning of dissimilar to our own for rheumatoid scleritis sclera (Rooney, McGavin, Sutherland, and Railton, (Table VIIIa). However, the distribution of 1974). The presence of translucency of the sclera anterior scleritis in this report, although more often reflects the depth and extent of the scleral inflam- affecting the superior aspect of the sclera, does not mation, which in this study was very much more have the considerable emphasis of Sevel's cases evident in rheumatoid scleritis, indicating that (Table IXa). Jayson and Jones (I 970I) in their scleritis in rheumatoid disease is generally more group of 14 patients found a distribution similar severe (Tables VIlla, b). Posterior scleritis is more to that seen in our study. common in patients with rheumatoid arthritis Corneal complications in scleral inflammation which also suggests a more extensive inflammation have been described for over a century (Mackenzie, in rheumatoid disease. Inflammation of the posterior I830). The predominant disturbance is a sclerosing sclera is deduced from its secondary effects on keratitis which, according to Watson, Hayreh, and neighbouring tissues. Posterior inflammation re- Awdry (I968a), occurs in nearly half of those sulted in choroiditis in two patients, retinal detach- patients with scleritis but seldom in those with Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 217

episcieritis. In i i cases which were enucleated and ported in Io05 to 6o per cent of patients with comprised the series of rheumatoid scleritis (Sevel, rheumatoid arthritis receiving treatment with long- I965), 36-o per cent had some form of keratitis. term oral corticosteroid therapy (Black, Oglesby, However, rheumatoid episcleritis and rheumatoid von Sallmann, and Bunim, I960; Giles, Mason, scleritis in this investigation showed a high incidence Duff, and McLean, I962; Crews, I963; Irby, of corneal complications (Table Xa), in particular Toone, Wittkamp, and Wiesinger, I964; Spencer sclerosing keratitis (Fig. 4), and it was interesting and Andelman, I965; Furst, Smiley, and Ansell, to find that anterior uveitis although less common I966). A similar study at the Centre for Rheumatic had a similar incidence in rheumatoid episcleritis Diseases, Glasgow, provided an incidence of and rheumatoid scleritis, unexpected findings as steroid-induced cataract of I V5 per cent (William- regards episcleritis. The anticipated difference son and others, I969). The combination of treat- between episcleritis and scleritis was confirmed in ment with systemic corticosteroids and the presence regard to cataract, glaucoma, and posterior segment of anterior segment inflammation in the form of disease, but it will be recalled that only 13 eyes scleritis carries a high risk of the development of had rheumatoid episcleritis. Corneal ulceration cataract; in this study, an incidence of 36-o per occurred in two forms; limbal guttering (Fig. 5), cent (Table XII). While it is helpful to be aware of which has been called marginal furrows (Brown this possible complication, it is reassuring that and Grayson, I968), and keratolysis. While limbal where systemic corticosteroid treatment is clearly guttering may persist for a considerable time, and indicated, any resulting cataract can be removed during the active phase of scleral inflammation by surgery. appears to 'creep' around the limbus, keratolysis Posterior scleritis can provoke various intraocular is a more dramatic 'melting' of the which side-effects including choroiditis with exudative may result in loss of the anterior chamber with retinal detachment (Radn6t, 1948; Harper, I960; prolapse. Watson and Holt-Wilson (I974) described Hurd, Snyder, and Ziff, 1970), macular oedema a further corneal complication of scleritis which (Watson and Lobascher, I965), retinal striations they called deep keratitis, in which a deep sheet (Sears, I964) and oedema (Bertelsen, opacity with deep vascularization after long-stand- I960). Orbital myositis is an extraocular manifes- ing recurrent scleritis was evident. We have recently tation (Gorduren, I962; Maria, I956). In his seen a young woman with non-rheumatoid scleritis, pathological series, Sevel (I965) found choroiditis who is not included in this study, where this (64-0 per cent), choroidoretinitis (9-0 per cent) and striking corneal corrplication was present. retinal detachment (64-0 per cent). Rheumatoid The main feature of interest in the patients with scleritis provides disturbing indications of deep- http://bjo.bmj.com/ anterior uveitis was the higher incidence of posterior seated inflammation, posterior segment abnor- synechiae in patients with rheumatoid scleritis as mality affecting nearly one-third of the patients compared with non-rheumatoid scleritis. This (Table XIIIa). Although macular disturbance would suggest a more active and tenacious form of could have other causes, such as drug-induced uveal inflammation, which is consistent with the , senile degeneration (Bedell, 1950; impression of severe ocular disease in rheuma- Favre, I960) and vascular disease (Norn, I96I) it toid arthritis. However, the presence of anterior is recognized that it may occur with on October 2, 2021 by guest. Protected copyright. uveitis in both rheumatoid and non-rheumatoid (Manschot and Von Winning, I953) or iridocyclitis inflammation was very nearly as common in epi- (Samuels, 1930; Wolff, I93I). It is possible, there- scleritis as in scleritis. Anterior uveitis occurred in fore, that the large number of rheumatoid patients approximately one-fifth of patients with episcleritis with a macular disturbance may be attributed, and it occurred in about one-quarter of those with in part at least, to the consequence of severe scleritis (Tables XIa, b). Iridocyclitis has often scleritis, the incidence being higher than in the been described with scleritis (Wolfe, i882; Kiehle, non-rheumatoid patients with scleritis, both groups 1946; Duke-Elder and Leigh, I965). Sevel (I965) having a mean age in the sixth decade. The uveitis in his series of i i rheumatoid cases found histologi- accompanying a scleritis is more often an anterior cal evidence of cyclitis in 46-o per cent. The figures uveitis, but whether anterior or posterior it is given in this study are objective in that they indicate from this serious complication that most of the evidence of active or old uveitis (for example, consequent ocular problems arise, problems which posterior synechiae), and so the true figure will be affect vision - such as cataract, glaucoma, exudative higher. A deep scleritis will invariably be accom- retinal detachment, and macular disturbance. Optic panied by uveitis-whether iritis, cyclitis, or nerve disease after scleritis has had little documen- choroiditis-but a number of cases, particularly tation. Bertelsen (I960) described I2 patients with in the anterior segment, may settle without any a posterior scleritis and two of these had oedema residual evidence of inflammation. of the optic nerve head. The cases of Gorduren Posterior subcapsular cataracts have been re- (I962) and Maria (1956) of sclerotenonitis and Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 2I18 British Journal of Ophthalmology

orbital myositis both had optic nerve oedema with inflammation as a result of surgical intervention. associated retinal oedema. One patient in the The incidence of Sj6gren's syndrome in rheu- present series with a rheumatoid scleritis had matoid arthritis has been estimated at between optic disc oedema. 9 and 34 per cent, and in the sicca syndrome there Watson (I966) found that seven of io8 cases with is a 6o to 75 per cent incidence of connective- scleritis (6-5 per cent) had raised intraocular tissue disease (Henkind and Gold, 1973). Talal pressure, and Sevel (I965) in his study of enucleated (1972) considers that keratoconjunctivitis sicca eyes reported that 45-0 per cent of 43 cases had a develops in I o to I 5 per cent of patients with history of secondary glaucoma. Watson's cases rheumatoid arthritis and cites the report of Sten- settled with the good response of the scleritis to stam (I947). While the incidence of keratocon- systemic treatment. He stated that surgery was junctivitis sicca in the present study in patients with avoided because of the dangers of aggravating the rheumatoid episcleritis and rheumatoid scleritis inflammation. Apart from one case of non-rheu- falls within the accepted range for rheumatoid matoid episcleritis with an intraocular pressure arthritis, the incidence of Sjogren's syndrome is fractionally above the upper limit of the normal considerably higher than that found by Williamson range, all cases of secondary glaucoma in this study (1970) in patients with rheumatoid arthritis, also were in patients with scleritis (Tables XIVa, b). attending the Centre for Rheumatic Diseases. He As in Watson's series (I966), the intraocular found that between 6 and 7 per cent of patients pressure settled coincident with quietening of the with rheumatoid arthritis had Sj6gren's syndrome. scleritis in most cases. One patient required surgery The incidence given here in rheumatoid episcleritis and one patient, with a minimally raised pressure (33.3 per cent) and rheumatoid scleritis (32z4 (22 mmHg), is still receiving medical treatment. If per cent) reflects the extent of extra-articular mani- the indications for operative intervention become festations in patients with rheumatoid episcleral compelling, surgery should be carried out despite and scleral inflammation. the presence of inflammation. The patient with rheumatoid arthritis and severe scleritis who RHEUMATOLOGY required surgery underwent a drainage procedure without any exacerbation of inflammation. It is not the purpose of this paper to discuss the The deeper and more extensive inflammation of rheumatological findings in any detail. However, scleritis affects ocular movement considerably some of the results of this aspect of our study lead

more often than episcleritis, and one suspects that us to suspect that the underlying disease of scleritis http://bjo.bmj.com/ the latter may affect movement only because of the in these patients may be a vasculitis. discomfort sustained in moving the eye to the The arthritis in patients with rheumatoid epi- affected side (Tables XVa, b). There is a close scleritis and rheumatoid scleritis was of longer association of scleritis, in particular posterior duration and revealed more severe erosive changes scleritis, with tenonitis, myositis, and pseudo- on x-ray examination as compared with the rheu- tumour formation. (Although orbital myositis has matoid controls (Table XVII). On the other hand, been called a form of pseudotumour the differen- the functional capacity and the degree of joint on October 2, 2021 by guest. Protected copyright. tiation in terms seems more appropriate.) It is tenderness were similar in the rheumatoid patients sometimes difficult to reach the correct diagnosis with and without inflammatory . Sub- in these conditions. The patient presents with cutaneous nodules in rheumatoid arthritis are found pain, limitation of ocular movement, chemosis, in approximately one-third of patients (Boyle and and sometimes lid oedema. The indirect conse- Buchanan, 1971). Of the patients with rheumatoid quence of inflammation provides the clue to episcleritis and rheumatoid scleritis 50 per cent diagnosis, and this appearance will often simulate had subcutaneous nodules, twice the percentage orbital cellulitis. Bertelsen (I960) confesses to this found in the control group of patients (Table mistake in six of his 12 patients with sclerotenonitis XVIII). Histological studies of early nodule and orbital myositis. Both Gorduren (I962) and formation by Sikoloff, McCluskey, and Bunim Maria (1956) began treatment with systemic (1953) suggest that the nodule develops around an antibiotics in their cases before reaching the true inflamed blood vessel. It is significant that biopsy of diagnosis. Coop (I96I) in a series of 47 cases of episcleral nodules in patients with rheumatoid orbital pseudotumour found nine which presented arthritis has given a pathological picture similar to with chemosis as part of the clinical picture. In the that of the subcutaneous nodule of rheumatoid woman with non-rheumatoid scleritis who pre- arthritis (Edstrom and Osterlind, 1948; Mundy, sented with strabismus there was the same confusion Howard, Stillnan, and Bevans, I951; Ferry, I969), in diagnosis. A biopsy of the central strip of tendon/ and examination of enucleated eyes with scleritis muscle of the lateral rectus muscle confirmed the has confirmed a uniform basic pathology (Manschot, diagnosis. There was no untoward effect on the I96I; Sevel, I965). Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 219

Approximately 25 per cent of patients with LABORATORY INVESTIGATIONS rheumatoid arthritis have vasculitis on post-mortem examination (Cruickshank, I954). Rheumatoid Laboratory data confirm the expected evidence of vasculitis, unlike classical , widespread disorder in patients with rheumatoid spares the kidneys, and hypertension and asthma disease as compared with patients with non- are rare (Ferguson and Slocumb, I96I). It is inter- rheumatoid eye inflammation. A significant differ- esting that while respiratory disease and possibly ence was noted in the findings comparing rheuma- cardiac involvement were features notable in the toid and non-rheumatoid patients as regards patients with rheumatoid eye inflammation, renal haemoglobin and albumin/globulin ratio. The ery- disease was not a problem. Rheumatoid vasculitis throcyte sedimentation rate was measured by two affects the peripheral nerves (Hart, Golding, and different methods but the readings in patients with McKenzie, 1957; Ferguson and Slocumb, I96I; rheumatoid disease were clearly far higher (Tables Wilkinson and Torrance, I967), giving rise to a XIXa, b). It is interesting to note that the mean sensory and/or motor neuropathy (Hart and values for haemoglobin, erythrocyte sedimentation Golding, I960; Schmid and others, I96I) and to rate, and albumin/globulin ratio were all more dermal infarction (Bywaters, 1957; Golding, Hamil- abnormal in the patients with rheumatoid epi- ton, and Gill, I965). In the group of patients with scleritis and rheumatoid scleritis than in the rheumatoid episcleritis and rheumatoid scleritis, rheumatoid control group, although only the 8-9 per cent had a peripheral neuropathy, presum- findings for the erythrocyte sedimentation rate ably due to vasculitis affecting the vasa nervorum, were statistically significant (Table XIXc). and i3'3 per cent had skin ulcers, both incidences The normal adult range for positive titres to very much higher than expected, although for Toxoplasma gondii is in the region of 30 to 40 per these numbers there was no statistically significant cent. Only the small group of patients with rheuma- increase in the incidence of peripheral neuropathy toid episcleritis had a figure that was very much or skin ulcers as compared with that in rheumatoid higher (Table XXa). Beattie (1958) found 0-2 per controls (Table XVIII). cent of normal adults with a titre of 1/256 or more Many patients with rheumatoid scleritis also and Ludlam and Beattie (I963) reported only one manifested skin bruising and skin atrophy (Table titre at I/512 in 478 women (0o2 per cent). In the XVIII), features almost certainly attributable to patients with rheumatoid eye inflammation three of systemic corticosteroid treatment, which most of 35 patients (8-6 per cent) had titres of 1/256 or the patients were more. A similar increased incidence of high titres, receiving (Boyle and Buchanan, http://bjo.bmj.com/ 1971). It has been suggested that systemic cortico- noted in this paper in patients with rheumatoid eye steroids produce alteration in the dermal connective inflammation, has been recorded in primary tissue, particularly in collagen (Scarborough and biliary cirrhosis (MacSween, 1973), reticulosis or Schuster, I960; Greenwood, I966). after immunosuppressive or antimetabolite treat- Patients with rheumatoid ocular inflammation ment (Broughton, 1970), pregnancy (Purtilo, also had a high incidence of splenomegaly (i0o9 Hallgren, and Yunis, I972) and where a choroido- per cent) (Table XVIII). All five patients with retinitis due to toxoplasmosis may relapse and toxo- splenomegaly had anaemia and two of the patients plasma antibody levels rise (Beattie, 1958). These on October 2, 2021 by guest. Protected copyright. had a neutropenia. The syndrome of chronic are conditions associated with impaired cell- rheumatoid arthritis, splenomegaly, and neutro- mediated immunity, and there is recent evidence of penia, first described by Felty (1924), is said to depression of cell-mediated immunity in some occur in fewer than 5 per cent of patients with patients with rheumatoid arthritis (Waxman, rheumatoid arthritis (Walker and Wright, I968). Lockshin, Schnapp, and Doneson, 1973). The picture that emerges in the patient with The titre of rheumatoid factor reflects the severity rheumatoid episcleritis and rheumatoid scleritis is of the disease (Ziff, 1957; Kellgren and Ball, 1959) that of long-standing rheumatoid disease and and the consequent prognosis (Duthie, Brown, severe erosive joint changes on x-ray examination, Truelove, Barager, and Lawrie, I964). It is not with many extra-articular manifestations of rheu- surprising, therefore, that some patients with matoid disease, many of which reflect an underlying rheumatoid episcleral or scleral inflammation had rheumatoid vasculitis. Episcleritis and scleritis in relatively high titres (Table XXII). Higher titres these patients is further evidence of widespread of rheumatoid factor are also associated with many and severe disease affecting eyes, peripheral nerves, extra-articular manifestations of rheumatoid disease skin, lungs, and possibly the heart. These findings such as subcutaneous nodules (Kellgren and Ball, substantiate the reports of Jones and Jayson (I973), I959), vasculitis, peripheral neuropathy, and intrac- who found a significant incidence of rheumatoid table skin ulcers (Epstein and Engleman, 1959). subcutaneous nodules and other features of micro- Autoantibody investigations in rheumatoid epi- vasculitis in patients with rheumatoid scleritis. scleritis and rheumatoid scleritis generally show Fl Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 220 British Journal of Ophthalmology

results consistent with the expected incidence in a The results of the autoantibody tests in patients general rheumatoid population (Table XXIa). In with episcleritis and scleritis, but with no clinical tests of patients with rheumatoid arthritis, Pollak or radiological evidence of rheumatoid arthritis, (I964) found that 24 per cent had positive antinuclear are of particular interest (Table XXIb). One-third results, and Weir, Holborow, and Johnson (I96I) of the patients with scleritis gave positive results to found that 14 per cent were similarly positive. tests for rheumatoid factor, and this might be an Thyroid antibodies in patients with rheumatoid early manifestation of rheumatoid disease. Alter- eye inflammation showed no striking increase and natively, it is conceivable that plasma cells in the indeed did not match the incidence of I5 of 62 scleral infiltrate could produce rheumatoid factor. cases (24-2 per cent) of rheumatoid arthritis found It is known that plasma cells in the lymphoid by Anderson, Goudie, Gray, and Buchanan (I96I). tissues and in lymphoid infiltrates of synovium The incidence of thyroid microsomal antibody in induce rheumatoid factor which then circulates in the present study of rheumatoid patients with the blood stream (Mellors, Nowoslawski, and eye inflammation is only slightly lower than that in Korngold, I96I). It is possible, therefore, that the the female group reported by Buchanan, Crooks, low titre of rheumatoid factor in the blood of those Alexander, Koutras, Wayne, and Gray (I96I), patients with scleritis, but no rheumatoid arthritis, where 26- i per cent of 46 women had thyroid merely reflects local scleral inflammation. However, microsomal antibody. Johnson, Holborow, and only careful long-term follow-up studies will Glynn (I965) found an incidence of smooth-muscle answer this question. The incidence of antinuclear antibody in 6-4 per cent of 47 patients with rheuma- factor in the serum of patients with non-rheumatoid toid arthritis and Doniach, Roitt, Walker, and scleritis could also indicate potential systemic Sherlock (I966) an incidence of i6 per cent in disorder. Previous studies give a lower figure for 32 patients with rheumatoid arthritis. The figure the presence of antinuclear factor in the serum of given here in rheumatoid scleritis (23-5 per cent) is patients with rheumatoid arthritis (I4 per cent: slightly higher than in these two series, but it is Weir and others, I96I; 24 per cent: Pollak, I964) difficult to reach any conclusion at this stage and it is known that the prevalence and titre of anti- regarding this finding. High incidences of smooth- nuclear factor is highest in patients with severe muscle antibody have been found in liver disease, arthritis (Hall, Bardawil, Bayles, Mednis, and particularly active chronic hepatitis (67 per cent), Galins, I960), vascular and granulomatous compli- primary biliary cirrhosis (50 per cent), and crypto- cations and Felty's syndrome (Alexander, Bremner, genic cirrhosis (28 per cent) (Doniach and others, and Duthie, I960; Ward, Johnson, and Holborow, http://bjo.bmj.com/ I966). Certainly liver disease has been noted in I964). Thyroglobulin antibody in patients with association with rheumatoid arthritis, but the inci- non-rheumatoid episcleritis and non-rheumatoid dence was small; o-66 per cent of patients with scleritis shows no increase beyond figures pre- rheumatoid arthritis alone and I-5 per cent of viously recorded for normal populations (5 per patients with Sjogren's syndrome, in whom the cent: Roitt and Doniach, 1958; 6 per cent: Owen connective-tissue disorder was rheumatoid arthritis and Smart, 1958; 4 per cent: Blizzard, Hamwi,

(Whaley and others, 1970). Salivary duct antibody Skillman, and Wheeler, 1959; io per cent: Beutner on October 2, 2021 by guest. Protected copyright. was found in one-quarter of patients with rheuma- and Witebsky, I962). Similarly, the incidence of toid arthritis, apparently uncomplicated by Sj6- thyroid microsomal antibodies is similar to the gren's syndrome (MacSween, Goudie, Anderson, figure of 8 per cent given by Beutner and Witebsky Armstrong, Murray, Mason, Jasani, Boyle, Bucha- (I962) for a normal population. Other autoantibody nan, and Williamson, I967), and it would appear tests in patients with non-rheumatoid eye inflam- that the present findings are consistent with their mation show little variation from general population observations without drawing any conclusion from studies for smooth-muscle antibody (I3-5 per cent: the higher percentage of antibodies in the very MacSween, 1973) and gastric parietal-cell antibody, small group with rheumatoid episcleritis. Arm- in which a general population study of hospital strong (I967) considered that there was no evidence inpatients older than 50 years gave a figure of io that gastric parietal-cell antibodies occurred more per cent with the higher figure of I3 6 per cent in commonly in rheumatoid arthritis. In his investi- women (Anderson, Buchanan, and Goudie, I967). gation of patients with rheumatoid arthritis, i6'2 MacSween (I973) in his group of II4 control per cent of 246 females and 6-3 per cent of 79 males patients in a study of primary biliary cirrhosis had positive results for gastric parietal-cell antibody. found an incidence of I0 per cent in women and The small group of patients with rheumatoid 5 per cent in men. episcleritis precludes any accurate conclusion and The complement system consists of a group of the group of patients with rheumatoid scleritis serum proteins which amplify the effects of some had a percentage incidence of gastric parietal-cell antigen-antibody interactions (Austen, 1972). In antibody of i6'2 per cent. rheumatoid arthritis, serum complement levels Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 221

have been found to be normal or slightly raised higher incidence of atopy than would have been (Ellis and Felix-Davies, I959) and also reduced expected in the general population. The incidence (Mongan, Cass, Jacox, and Vaughan, I969; is within the range found in controls during studies Weinstein, Peters, Brown, and Bluestone, I972). of type I allergy (Johansson, I967). This finding in a Six rheumatoid patients in this study show reduced small series suggests that any external factor which values (Table XXIIIa) and it may be that hypo- might predispose to the interpalpebral distribution complementaemia reflects active vasculitis, as has of inflammation found in many patients with been shown previously in rheumatoid arthritis episcleritis (Lyne and Pitkeathly, I968) (Table (Mongan, Cass, Jacox, and Vaughan, I969; IXb) does not have its origin in a susceptibility to Weinstein and others, 1972). Six of the patients atopic disease. with non-rheumatoid eye inflammation also had In a recent publication reviewing ocular mani- serum complement levels below ioo mg/IOO ml festations of rheumatic disorders, Henkind and (Table XXIIIb) and it is difficult to suggest causes Gold (1973) plead for the inclusion of examination of this finding. One of these patients gave a history by an ophthalmologist in the protocol for any of gout and essential hypertension, and another had prospective study of rheumatic disease. This study, erythema nodosa. However, it should be noted which has anticipated their request by some years, that there is some dubiety as to what is the range attempts to collate the ocular manifestations of of normal for serum complement readings. The episcleritis and scleritis with the rheumatic dis- normal mean value in this study was taken as orders, in particular rheumatoid arthritis. 145 mg/IOO ml (+22 mg/10o ml) (Shanbrom, Khoo, and Lou, 1967). Thompson found a (1972) Conclusions mean value of ii8 mg/ioo ml (+26 mg/ioo ml) in 66 healthy individuals. i. Episcleritis occurred in approximately two Immunoglobulins IgG, IgA, and IgM show patients per iooo with rheumatoid arthritis little variation from normal figures in either rheu- (O0I7 per cent). Scleritis occurred in approxi- matoid or non-rheumatoid eye inflammation. mately seven patients per IOOO with rheuma- toid arthritis (o-67 per cent). Rheumatoid arthritis is common in patients presenting MORTALITY with scleritis, in this series occurring in one- The mortality in rheumatoid patients with epi- third of patients (33.3 per cent), but is less scleral and scleral inflammation was soberingly

common in patients presenting with episcleri- http://bjo.bmj.com/ high, 45.5 per cent (Table XXIV). Fifteen of 33 tis, occurring in approximately one patient known cases of rheumatoid episcleritis or rheu- in 20 (5 7 per cent). matoid scleritis had died since examination, com- 2. All patients with rheumatoid episcleritis, pared with six of 33 known rheumatoid control rheumatoid scleritis, and non-rheumatoid patients (i8-2 per cent). A decreased life expec- scleritis had a mean age in the sixth decade, tancy has been observed in rheumatoid arthritis and the disease was more prevalent in women; (Reah, I963) and the outlook and severity of the whereas patients with non-rheumatoid epi- condition is always more grave in sero-positive scleritis had a mean age in the fifth decade on October 2, 2021 by guest. Protected copyright. disease (Duthie and others, I964). Jones and Jayson and the disease occurred equally between the (I973) found that five of their original 14 patients sexes. with rheumatoid scleritis had died within three 3. Rheumatoid episcleritis and rheumatoid years, a mortality of 36 per cent. Only i8-5 per cent scleritis were associated with widespread of their 133 control patients had died during the systemic disease, particularly of the cardio- same period. The figures for mortality suggest that vascular and respiratory systems. the conclusion of Decker (1972) was accurate: 4. The most common complaints in episcleritis 'death due solely to rheumatoid arthritis is rare and scleritis were of redness and pain. Ocular indeed. When it occurs it is due to the extra- pain was more evident in scleritis. Eye symp- articular manifestations'. toms were more readily noted in non-rheu- matoid inflammation. TYPE I ALLERGY (ATOPY) 5. All groups of patients had a very wide range in the duration of intermittent ocular inflam- The simplest methods of screening patients for type mation, from a few days or weeks to I5 years I allergy are either to carry out a series of skin or more. tests, as has been done here, or to measure the 6. Visual acuity is more severely depressed in blood levels of IgE. A group of 23 patients, each scleritis than in episcleritis as a result of the patient with non-rheumatoid eye inflammation, associated ocular complications. failed to give any indication on skin testing of a 7. In a classification of scleritis as nodular, Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 222 British 7ournal of Ophthalmology

diffuse, and necrotizing, diffuse scleritis was raised titres compared with studies in normal the most common form in both rheumatoid patients. and non-rheumatoid scleritic inflammation. i6. Autoantibody studies in patients with rheu- Posterior scleritis was more common in matoid episcleritis and with rheumatoid rheumatoid scleritis than in non-rheumatoid scleritis showed little variation from results inflammation. Scleral translucency was found expected in rheumatoid arthritis. Autoanti- considerably more often in rheumatoid than body studies in patients with non-rheumatoid in non-rheumatoid scleritis. scleritis showed a high incidence of rheuma- 8. Ocular inflammation in episcleritis usually toid factor (33-3 per cent) and antinuclear affected the interpalpebral area. Rheumatoid factor (33-3 per cent). scleritis had a fairly even distribution in all 17. The prognosis for life is poorer in patients four quadrants of the anterior sclera. with rheumatoid disease complicated by 9. Episcleritis and scleritis associated with episcleritis or scleritis when compared with rheumatoid disease had more associated rheumatoid controls matched for age, sex, ocular complications than episcleritis and and year of presentation; 45.5 per cent of scleritis unassociated with rheumatoid disease. those with eye inflammation had died since I0. Posterior subcapsular cataract was common the examination, compared with i8-z per cent in rheumatoid scleritis in cases in which the of the control group. patient had been given systemic cortico- I 8. There is no clear evidence of any higher steroids, occurring in 36 per cent of these incidence of atopic disease in patients with patients. non-rheumatoid episcleritis and non-rheuma- I I. Rheumatoid episcleritis and rheumatoid toid scleritis as compared with the expected scleritis generally presented in patients with incidence in a general population. rheumatoid disease of long duration and in patients with radiological evidence of more Many colleagues supported our study, and for this we advanced joint disease when compared with express our warmest gratitude. A number of ophthal- mologists in the west of Scotland took interest in the rheumatoid control patients. study and referred patients to the scleritis clinic. With- 12. Subcutaneous granulomatous nodules and out their help the investigation would have been in- skin atrophy were significantly more common complete. Dr G. B. Ludlam kindly provided data for in patients with rheumatoid episcleritis and our study of positive titres for Toxoplasma gondii. Mrs

rheumatoid scleritis when compared with Hazel Hughes, Mrs Catriona Breen, and Mrs Mary http://bjo.bmj.com/ rheumatoid control patients. Although not Mason of the Immunopathology Laboratory, Glasgow, statistically significant, peripheral neuropathy, examined sera for autoantibody studies and Miss skin ulceration, skin bruising, and spleno- Tineke Simmes gave us information regarding immuno- megaly were all more apparent in patients globulin estimations. Miss Nancy Wylie of the library, with episcleritis and rheumatoid Royal College of Physicians and Surgeons, Glasgow, rheumatoid has been a constant help in obtaining papers and check- scleritis than in the rheumatoid control group. ing references. The illustrations for the paper were 13. Sj6gren's syndrome was more evident in from the Department of Medical Illustration, Southern on October 2, 2021 by guest. Protected copyright. patients with rheumatoid episcleritis and General Hospital (Mr R. Loudon Brown, Mr Robin rheumatoid scleritis when compared with Ewart, and Mr Alastair Irvine), and from the Tennent general rheumatoid patients also referred to Institute of Ophthalmology (Mr John Watt and Mrs the Centre for Rheumatic Diseases. Anne Currie). Mr R. Clelland, Registrar of Births, 14. The erythrocyte sedimentation rate was Deaths and Marriages, Glasgow, obtained details of significantly higher in patients with rheuma- death certificates for patients who had died since toid episcleritis and rheumatoid scleritis when examination. The expenses of the study were supported by two compared with the rheumatoid control grants, from B.D.H. Pharmaceuticals (/Jioo) and patients. Geigy Pharmaceuticals (Cio20). I 5. Antibodies to Toxoplasma gondii in patients Tables VIII, XII, XVII, XVIII, XXIV, XXVII, with rheumatoid episcleritis and rheumatoid XXVIII, are reproduced in whole or in part by kind scleritis showed a higher percentage of permission of Churchill Livingstone, Edinburgh. References ADAMS, J. F., GLEN, A. I. M., KENNEDY, E. H., MCKENZIE, I. L., MORROW, J. M., ANDERSON, J. R., GRAY, K. G., and MIDDLETON, D. G. (I964) Lancet, I, 401 ALEXANDER, W. R. M., BREMNER, J. M., and DUTHIE, J. J. R. (I960) Ann. rheum. Dis., 19, 338 ALLAN, A. W. (I9OI) Brit. med. JI., 2, I267 ANDERSON, J. R., BUCHANAN, W. w., and GOUDIE, R. B. (I967) In 'Autoimmunity'. Thomas, Springfield, Ill. , GOUDIE, R. B., GRAY, K. G., and BUCHANAN, w. w. (I96I) Scot. med. J7., 6, 449 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 223

ARMSTRONG, E. M. (I967) Ann. rheum. Dis., 26, 499 ARMSTRONG, K., and MCGOVERN, V. J. (1955) Trans. ophthal. Soc. Aust., I5, IIO ASHTON, N., and HOBBS, H. E. (1952) Brit. J. Ophthal., 36, 373 ATWATER, E. C., MONGAN, E. S., WIECHE, D. R., and JACOX, R. F. (I965) Arch. int. Med., II5, I84 AUSTEN, K. F. (I972) In 'Arthritis and Allied Conditions' by J. L. Hollander and D. J. McCarty. Lea & Febiger, Philadelphia BACON, P. A., and GIBSON, D. G. (1974) Ann. rheum. Dis., 33, 20 BEATTIE, C. P. (1958) Trans. ophthal. Soc. U.K., 78, 99 BECK, J. S. (I96I) Lancet, I, 1203 BEDELL, A. J. (I950) Amer. J. Ophthal., 33, I68I BERTELSEN, T. I. (I960) Acta ophthal. (Kbh.), 38, 136 BEUTNER, E. H., and WITEBSKY, E. (I962) J. Immunol., 88, 462 BILLSON, F. A., DE DOMBAL, F. T., WATKINSON, G., and GOLIGHER, J. C. (1967) Gut, 8, 102 BLACK, R. L., OGLESBY, R. B., VON SALLMANN, L., and BUNIM, J. J. (I960) 7. Amer. med. Ass., 174, i66 BLIZZARD, R. M., HAMWI, G. J., SKILLMAN, T. G., and WHEELER, W. E. (1959) New Engl. J. Med., 260, 112 BLOCH, K. J., BUCHANAN, W. W., WOHL, M. j., and BUNIM, J. J. (I965) Medicine (Baltimore), 44, I87 BLUM, F. G., and SALMOUN, S. G. (I963) Arch. Ophthal., 69, 287 BOBERG-ANS, J. (1958) Acta ophthal. (Kbh.), 36, 33 BONFIGLIO, T., and ATWATER, E. C. (1969) Arch. intern. Med., 124, 714 BOWEN, R. JR., MAYNE, J. G., CAIN, J. C., and BARTHOLOMEW, L. G. (I960) Proc. Mayo Clin., 35, 537 BOYLE, J. A., and BUCHANAN, W. W. (1971) 'Clinical Rheumatology'. Blackwell, Oxford and Edinburgh BROUGHTON, C. R. (1970) Med. J. Aust., 2, 418 BROWN, S. I., and GRAYSON, M. (I968) Arch. Ophthal., 79, 563 BUCHANAN, W. W., CROOKS, J., ALEXANDER, W. D., KOUTRAS, D. A., WAYNE, E. j., and GRAY, K. G. (ig6I) Lancet, I, 245 BYWATERS, E. G. L. (1957) Ann. rheum. Dis., I6, 84 (I964) Personal conununication to D. Sevel. Trans. ophthal. Soc. U.K. (I965), 85, 357 CARR, D. T., and MAYNE, J. G. (I962) Amer. Rev. resp. Dis., 85, 345 CARRINGTON, C. B., and LIEBOW, A. A. (I966) Amer. J. Med., 41, 497 CASSAN, S. M., COLES, D. T., and HARRISON, E. G. JR. (1970) Ibid., 49, 366 CAT, I., MARINONI, L. P., GIRALDI, D. j., and COSTA, 0. (1971) Lancet, 2, II56 CLEOBURY, W. (I826) 'A review of the different operations performed on the eyes for the restoration of lost and the improvement of imperfect vision. Also a full account of the various diseases of the eyes and their appendages'. Underwood, London Ophthal., 45, 513 COOP, M. E. (I96I) Brit. Y. http://bjo.bmj.com/ COPEMAN, W. S. C. (I964) 'Textbook of the Rheumatic Diseases'. Livingstone, Edinburgh and London CREWS, S. j. (I963) Brit. med. J., I, i644 CRUICKSHANK, B. (1954) Ann. rheum. Dis., 13, 136 DECKER, J. L. (1972) In 'Arthritis and Allied Conditions' by J. L. Hollander and D. J. McCarty. Lea & Febiger, Philadelphia DE SCHWEINITZ, G. E. (I899) 'Diseases of the Eye'. Saunders, Philadelphia DONIACH, D., ROITT, I. M., WALKER, J. G., and SHERLOCK, S. (I966) Clin. exp. Immunol., I, 237 DUKE-ELDER, SIR s., and LEIGH, A. G. (I965) 'System of Ophthalmology', vol. VIII, part 2. Kimpton, London DUTHIE, J. J. R., BROWN, P. E., TRUELOVE, L. H., BARAGER, F. D., and LAWRIE, A. J. (I964) Ann. rheum. Dis., 23, 193 on October 2, 2021 by guest. Protected copyright. EDSTROM, G., and OSTERLIND, G. (1948) Acta ophthal. (Kbh.), 26, I ELLIS, H. A., and FELIX-DAVIES, D. (1959) Ann. rheum. Dis., I8, 215 ELLIS, 0. H., and HOLTZ, M. J. (1953) Calif. Med., 78, 6o EPSTEIN, w. v., and ENGLEMAN, E. P. (1959) Arthr. and Rheum., 2, 250 FAVRE, M. (I960) Ophthalmologica (Basel), 140, 94 FELTY, A. R. (I924) Bull. Johns Hopk. Hosp., 35, i6 FERGUSON, R. H. (1972) In 'Arthritis and Allied Conditions'. J. L. Hollander and D. J. McCarty. Lea & Febiger, Philadelphia , and SLOCUMB, C. H. (I96I) Bull. rheum. Dis., II, 251 FERNANDEZ-HERLIHY, L. (1959) New Engl. 3. Med., 26I, 259 FERRY, A. P. (I969) Arch. Ophthal., 82, 77 FOWLER, P. (1970) Ann. rheum. Dis., 29, 195 FRAN9OIS, J. (I95I) Trans. ophthal. Soc. U.K., 71, 6i , VICTORIA-TRONCOSO, V., HANSSENS, M., and BAKSKULIN, j. (I969) Ophthalmologica (Basel), 159, 71 FRY, J. (1952) Brit. med. _J., I, 529 FULTHORPE, A. J., ROITT, I. M., DONIACH, D., and COUCHMAN, K. (I96I) Y. clin. Path., 14, 654 FURST, C., SMILEY, W. K., and ANSELL, B. M. (I966) Ann. rheum. Dis., 25, 364 GILBERT, W. (I914) Arch. Augenheilk., 76, iiI GILES, C. L., MASON, G. L., DUFF, I. F., and MCLEAN, J. A. (1962) J. Amer. med. Ass., 182, 719 GODMAN, G. C., and CHURG, J. (I954) Arch. Path., 58, 533 GOLDING, J. R., HAMILTON, M. G., and GILL, R. S. (I965) Brit. _J. Derm., 77, 207 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 224 British Journal of Ophthalmology

GOMBOS, G. M. (I967) Acta ophthal. (Kbh.), 45, 582 GORDUREN, s. (I962) Brit. J. Ophthal., 46, 568 GOUDIE, R. B., MACSWEEN, R. N. M., and GOLDBERG, D. M. (I966) Y. clin. Path., 19, 527 GREENWOOD, B. M. (I966) Ann. rheum. Dis., 25, 272 GUTHRIE, D. (1958) 'A History of Medicine', p. 264. Nelson, London HALBERG, G. P., and BERENS, C. (i967) Schirmer Tear Test. Cooper Laboratories, Wayne, N.J. HALL, A. P., BARDAWIL, W. A., BAYLES, T. B., MEDNIS, A. D., and GALINS, N. (I960) New Engl.3. Med., 263, 769 HARBATER, M. (I949) Arch. Ophthal., 41, I83 HARPER, J. Y. (1960) Ibid., 63, 663 HARROLD, B. P. (1968) Brit. med. 7., I, 290 HART, F. D. (1964) Personal communication to D. Sevel. Trans. ophthal. Soc. U.K. (I965), 85, 357 (I969) Brit. med. Y.,3, 13I and GOLDING, J. R. (I960) Ibid., I, 1594 , and MCKENZIE, D. H. (I957) Ann. rheum. Dis.,I6, 47I HARVEY, A. MCG., SHULMAN, L. E., TUMULTY, P. A., CONLEY, C. L., and SCHOENRICH, E. H. (1954) Medicine (Baltimore), 33, 291 HAYREH, S. s., and WATSON, P. G. (I970) Brit..7. Ophthal., 54, 394 HENCH, P. S. (I935) Nelson's Loose-leaf Surgery. Nelson, New York. Cited by R. H. Ferguson in 'Arthritis and Allied Conditions' by J. L. Hollander and D. J. McCarty. Lea & Febiger, Philadelphia HENCH, P. S., KENDALL, C. E., SLOCUMB, C. H., and POLLY, H. F. (1949) Ann. rheum. Dis., 8, 97 HENKIND, P., and GOLD, D. H. (1973) Ocular Manifestations of Rheumatic Disorders in 'Oculocutaneous Manifestations of Rheumatic Diseases'. Karger; Basel, Miunchen, Paris, London, New York, Sydney HEWSON, G. E. (I964) Trans. ophthal. Soc. U.K., 84, 297 HOLBOROW, E. J., BROWN, P. C., ROITT, I. M., and DONIACH, D. (1959) Brit. 7. exper. Path., 40, 583 HOLLENHORST, R. W., and HENDERSON, J. W. (195I) Amer..7. med. Sd., 221,211 HOLLOWAY, T. B., and FRY, W. E. (I93I) Arch. Ophthal., 6, 136 HOPKINS, D. J., HORAN, E., BURTON, I. L., CLAMP, S. E., DE DOMBAL, F. T., and GOLIGHER, J. C. (1974) Brit.3. Ophthal., 58, 732 HURD, E. R., SNYDER, W. B., and ZIFF, M. (1970) Amer.3. Med., 48, 273 HUSKISSON, E. c., and HART, F. D. (I972) Brit. med..7., 4, 193 IRBY, R., TOONE, E. C., WITTKAMP, B., and WIESINGER, H. (I964) G.P., 29, No. I, 92 JAMES, D. G. (I96I) Brit. med. Y., I,853. JAYSON, M. I. v., and JONES, D. E. P. (1971) Ann. rheum. Dis., 30, 343

,SALMON, P. R., and HARRISON, W. J. (1970) Gut, II,5o6 http://bjo.bmj.com/ JOHANSSON, S. G. 0. (I967) Lancet, 2, 951 JOHNSON, G. D., HOLBOROW, E. j., and GLYNN, L. E. (I965) Ibid., 2, 878 JONES, B. R. (1971) Trans. ophthal. Soc. U.K., 9I, 44I JONES, D. E. P., and JAYSON, M. I. V. (1973) Proc. roy. Soc. Med., 66, i6i KELLGREN, J. H., and BALL, J. (I959) Brit. med.3., I, 523 KIEHLE, F. A. (1946) Amer. 7. Ophthal., 29, 862 KIRK, j., and COSH, J. (I969) Quart. 7. Med., 38, 397

LARSON, D. L. (I96I) 'Systemic Lupus Erythematosus'. Little, Brown, Boston on October 2, 2021 by guest. Protected copyright. LITTELL, S. (i840) 'A Manual of the Diseases of the Eye' (rev. Hugh Houston). Churchill, London LUDLAM, G. B. (1972) Personal communication -, and BEATTIE, C. P. (I963) Lancet, 2, 1136 LYNE, A. J. (I974) Trans. ophthal. Soc. U.K., 94, 58 , and PITKEATHLY, D. A. (I968) Arch. Ophthal., 8o, 171 MCCARTHY, J. L. (I96I) Amer. 7. Ophthal., 5I, 6o MCEWEN, C. (I968) Clin. Orthop., 57, 9 MACKENZIE, W. (I830) 'A Practical Treatise of Diseases of the Eye'. Longman, Rees, Orme, Brown & Green, London (1854) 'A Practical Treatise on the Diseases of the Eye'. 4th ed. Longman, Brown, Green & Longmans, London MACSWEIEN, R. N. M. (I973) Clinico-pathological and Immunological Studies in Liver Disease. MD thesis. University of Glasgow , GOUDIE, R. B., ANDERSON, J. R., ARMSTRONG, E., MURRAY, M. A., MASON, D. K., JASANI, M. K., BOYLE, J. A., BUCHANAN, W. W., and WILLIAMSON, J. (I967) Ann. rheum. Dis., 26, 402 MCWILLIAMS, J. R. (1952) Amer. Y. Ophthal., 35, I778 MANCINI, G., CARBONARA, A. O., and HEREMANS, J. F. (I965) Int. Y7. Immunochem., 2, 235 MANSCHOT, W. A. (I96I) Advances Ophthal., iI, i - and VON WINNING, C. H. 0. M. (1953) Amer. Y. Ophthal., 36, 26I MARIA, D. L. (1956) 7. All-India ophthal. SOC., 14, 94 MARQUARD, H. A. (1956) Acta ophthal. (Kb/h.), 34, 245 MAZZA, C., and PANAGIS, P. (I967) Ann. Ottal., 93, 373 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from Episcleritis and scleritis 225

MELLORS, R. C., NOWOSLAWSKI, A., and KORNGOLD, L. (I96I) Amer. J. Path., 39, 533 MELTZER, L. E., BOCKMAN, A. A., KANENSON, w., and COHEN, A. (1958) Gastroenterology, 35, 351 MERZ, E. H. (I964) Amer. 7. Ophthal., 57, 766 MONGAN, E. S., CASS, R. M., JACOX, R. F., and VAUGHAN, J. H. (I969) Amer. J7. Med., 47, 23 MORGAN, J. (I839) 'Lectures on Diseases of the Eye'. Highley, London MUNDY, W. L., HOWARD, R. M., STILLMAN, P. H., and BEVANS, M. (195I) Arch. Ophthal., 45, 531 NOONAN, C. D., TAYLOR, F. B., and ENGLEMAN, E. P. (I963) Arthr. and Rheum., 6, 232 NORN, M. S. (196I) Acta ophthal. (Kbh.), 39, 817 OAST, S. P. (I937) Arch. Ophthal., 17, 697 OWEN, S. G., and SMART, G. A. (I958) Lancet, 2, I034 PALICH-SZANTO (I959) Ber. dtsch. ophthal. Ges., 62, 315. Cited by S. Duke-Elder and A. G. Leigh. Kimpton, London (I965) PITKEATHLY, D. A., HOWITT, G., and LYNE, A. J. (1970) Ann. rheum. Dis., 29, 477 POLLAK, V. E. (I964) Nezw Engl. J. Med., 271, i65 PURTILO, D. T., HALLGREN, H. M., and YUNIS, E. J. (1972) Lancet, I, 769 RADNOT, M. (1948) Ophthalmologica (Basel), II6, I67 READ, SIR W. (Circa 1702-I714) 'A Short but Exact Account of all Diseases Incident to the Eyes', 2nd ed. Baker, London REAH, T. G. (I963) Proc. roy. Soc. Med., 56, 813 RENARD, G., LELIEVRE, and MAZEL (1953) Bull. Soc. franc. Ophtal., 66, 243 RICHARDSON, A. T. (I964) Personal Communication to D. Sevel. Trans. ophthal. Soc. U.K. (I965), 85, 357 ROITT, I. M., and DONIACH, D. (1958) Lancet, 2, 1027 ROONEY, P., MCGAVIN, D. D. M., SUTHERLAND, G. R., and RAILTON, R. (1974) Thin Sclerae; a Clinical Illusion. HIeberden Society, London ROPES, M. W., BENNETT, G. A., COBB, S., JACOX, R., and JESSAR, R. A. (1959) Bull. rheum. Dis., 9, 175 RUHL, M. j., and SOKOLOFF, L. (I965) Arthr. and Rheum., 9, 95 SABIN, A. B., and FELDMAN, H. A. (1948) Science, Io8, 66o SAMUELS, B. (1930) Arch. Ophthal., 4, 476 SAVIN, L. H. (1938) Trans. ophthal. Soc. U.K., 58, 149 SCARBOROUGH, H., and SHUSTER, S. (I960) Lancet, I, 93 SCARPA, A. (i8i8) 'A Treatise on the Principal Diseases of the Eyes', p. 399 (trans. from the Italian by James Briggs) 2nd ed. Cadell & Davis, London SCHMID, F. R., COOPER, N. S., ZIFF, M., and MCEWEN, C. (I96I) Amer. 7. Med., 30, 56 SEARS, M. L. (I964) Amer. J. Ophthal., 58, 764 SEVEL, D. (I965) Trans. ophthal. Soc. U.K., 85, 357 http://bjo.bmj.com/ (I966) Necrogranulomatous scleritis; a clinical, pathological and experimental study. PhD thesis. University of London (I967) Amer. j. Ophthal., 64, 1125 SHANBROM, E., KHOO, M., and LOU, K. (I967) Clin. Res., 5S, 114 (abstract) SHANEDLING, P. D. (1950) Amer. J. Ophthal., 33, 804 SJOGREN, H. (I933) 'A New Conception of Keratoconjunctivitis Sicca'. English trans. in I943 by J. Bruce Hamilton, Australasian Medical Publication Co., Sydney. Cited by K. J. Bloch, W. W. Buchanan, M. J. Wohl, and J. J. Bunim (I965) Medicine (Baltimore), 44, I87 on October 2, 2021 by guest. Protected copyright. SLOCUMB, C. H. (I953) Proc. Mlayo Clin., 28, 655 SMITH, J. L. (I957) Amer. J. Ophthal., 43, 575 SMOLEROFF, J. W. (I943) Arch. Ophthal., 29, 98 SMYTH, C. J., and GUM, 0. B. (I96I) Arthr. and Rheum., 4, I SOKOLOFF, L. (I972) In 'Arthritis and Allied Conditions' by J. L. Hollander and D. J. McCarty. Lea & Febiger, Philadelphia MCCLUSKEY, R. T., and BUNIM, J. J. (1953) Arch. Path., 55, 475 SPAETH, G. L. (I967) New Engl. J7. Med., 276, i i68 SPENCER, R. W., and ANDELMAN, Y. S. (I965) Arch. Ophthal., 74, 38 STEINBROCKER, O., TRAEGER, C. H., and BATTERMAN, R. C. (I949) 7. Amer. med. Ass., 140, 659 STENSTAM, T. (1947) Acta med. scand., 127, 130 SWAN, K. C., and BUTLER, J. B. V. (I95i) Arch. Ophthal., 45, 630. Cited by K. C. Swan TAFFET, S., and CARTER, G. Z. (I96I) Amer. J. Ophthal., 52, 693 TALAL, N. (1972) In 'Arthritis and Allied Conditions' by J. L. Hollander and D. J. McCarty. Lea & Febiger, Philadelphia TALKOV, R. H., COLPOYS, F. L., DAVIS, R. K., PAPPER, S., and FIENBERG, R. (195i) Arch. intern. Med., 87, 879 THOMAS, C., HELLUY, CORDIER, and ALGAN, B. (1953) Bull. Soc. Ophtal. Paris, 130 THOMPSON, R. A. (I972) Immunology, 22, 147 TOOKER, C. W. (I93i) Amer. J. Ophthal., 14, 911 VAN DER HOEVE, J. (1934) Arch. Ophthal., IIx iii VAN VALKENBURGH, W. G., GEORGES, L. P., and IRBY, R. (1972) Arthr. and Rheum., I 544 Br J Ophthalmol: first published as 10.1136/bjo.60.3.192 on 1 March 1976. Downloaded from 226 British 7ournal of Ophthalmology

VERHOEFF, F. H., and KING, M. J. (1938) Arch. Ophthal., 20, I015 VON ARLT, F. R. (I885) 'Clinical Studies on the Diseases of the Eye' (trans. by Lyman Ware). Pentland, Edinburgh WALKER, W. c., and WRIGHT, V. (I967) Ann. rheum. Dis., 26, 467 , and WRIGHT, V. (I968) Medicine (Baltimore), 47, 501 WALTON, H. (I875) 'A Practical Treatise on the Diseases of the Eye'. Churchill, London WARD, D. J., JOHNSON, G. D., and HOLBOROW, E. J. (I964) Ann. rheum. Dis., 23, 306 WATSON, P. G. (I966) Trans. ophthal. Soc. U.K., 86, I5' -(1974) Ibid., 94, 65 HAYREH, S. S., and AWDRY, P. N. (I968a) Brit. J7. Ophthal., 52, 278 , and (I968b) Ibid., 52, 348 and HOLT-WILSON, A. D. (1974) Trans. ophthal. Soc. U.K., 94, 46 and LOBASCHER, D. (I965) Ibid., 85, 369 WAXMAN, J., LOCKSHIN, M. D., SCHNAPP, j. j., and DONESON, I. N. (1973) Arthr. and Rheum., I6, 499 WEINSTEIN, A., PETERS, K., BROWN, D., and BLUESTONE, R. (1972) Ibid., I5, 49 WEIR, D. M., HOLBOROW, E. j., and JOHNSON, G. D. (I96I) Brit. med. 7., I, 933 WHALEY, K., WILLIAMSON, J., DICK, W. C., GOUDIE, R. B., NUKI, G., and BUCHANAN, W. W. (1970) Lancet, x, 86I WILKINSON, M., and TORRANCE, W. N. (I967) Ann. rheum. Dis., 26, 475 WILLIAMS, G. T., and ROSENTHAL, J. W. (1959) Ann. intern. Med., 5I, 8oi WILLIAMSON, J. (1970) Keratonconjunctivitis sicca in adult rheumatoid arthritis. MD thesis, University of Edinburgh , PATERSON, R. W. W., MCGAVIN, D. D. M., JASANI, M. K., BOYLE, J. A., and DOIG, W. M. (I969) Brit. J7. Ophthal., 53, 36I WOLFE, J. R. (I882) 'On Diseases and Injuries of the Eye'. Churchill, London WOLFF, E. (1931) Proc. roy. Soc. Med., 24, 1617 WOODS, A. c., and GUYTON, J. s. (I944) Arch. Ophthal., 31, 469 WRIGHT, R., LUMSDEN, K., LUNTZ, M. H., SEVEL, D., and TRUELOVE, S. C. (I965) Quart. J7. Med., 34, 229 WRIGHT, V., and WATKINSON, G. (I965) Brit. med. J., 2, 670 ZIFF, M. (B957) 7. chron. Dis.,5, 6B9 , BROWN, P., LOSPALLUTO, J., BADIN, J., and MCEWAN, C. (1956) Amer. _7. Med., 20, 500 http://bjo.bmj.com/ on October 2, 2021 by guest. Protected copyright.