key: cord-007580-qwh8ei60
authors: Knopf, Harry L.S.; Hierholzer, John C.
title: Clinical and Immunologic Responses in Patients with Viral Keratoconjunctivitis
date: 2014-09-26
journal: Am J Ophthalmol
DOI: 10.1016/0002-9394(75)90398-0
sha: 
doc_id: 7580
cord_uid: qwh8ei60

nan

Ocular viral infections manifest them selves in a variety of clinical pictures. Most commonly, the clinician finds follicular con junctivitis, with or without accompanying keratitis. Epithelial and subepithelial corneal infiltrates are frequently observed with adenovrrus mWnons. hut -also occur with hprppg simpW and nfhpr sriraijnduced kerajilis-ia&ctions. 1 ' 2 In addition to keratoconjunctivitis, viruses can (directly or in directly) cause deep ocular involvement, in cluding uveitis and optic nerve inflamma tion. 1 In an effort to learn more about the etiolo gies of ocular viral infections and the pathogenesis of viral keratitis, we studied two groups of patients with epidemic and nonepidemic viral infections. All of them had follicular conjunctivitis, and many suffered from keratitis during at least one stage of their illness.

We had three major goals: to determine the rate of viral recovery from cases of nonepidemic acute or subacute follicular conjunc tivitis; to compare the accuracy of clinical diagnoses with that achieved by culture or serologic studies, or both; and to explore the possible role of serum antibody in the forma tion and maintenance of corneal subepithelial infiltrates. We examined one group of patients who were involved in an epidemic caused by a single viral agent, and a second group of patients who were infected with many differ ent viruses. Our results indicate that a variety of agents can cause similar illness patterns, and that clinical observation cannot always distinguish one from another.

Furthermore, serologic tests are a more accurate diagnostic method than clinical course in ocular viral illness, as is the case in systemic viral illnesses. Finally, viral anti body titers may correlate directly with spe cific clinical changes in adenovirus-induced keratoconjunctivitis.

Patients-We studied two groups of pa tients by the following methods. The first was a population of 56 industrial employees who had epidemic keratoconjunctivitis. These patients were described in detail elsewhere, 8 but the data concerning their ocular findings were not presented.

The second group of 64 patients comprised a heterogeneous population seen during an 18-month period. The only criterion for se lection was follicular conjunctivitis. All but nine patients were seen in the emergency room, clinic, or staff offices at Barnes Hos pital from November 1971 to December 1972. The other nine were followed at the National Eye Institute from July 1970 to June 1971. One of us (H.L.S.K.) saw every patient during the acute or subacute phase of the illness and followed them until signs and symptoms abated. There were 32 males and 32 females, ranging in age from 3 to 77 years old (mean, 31.8 years). Twenty-nine patients were black, 33 were white, and two were orientals. All were ambulatory and in good health.

Collection of virus culture specimens-Fifty-eight of the 64 nonepidemic patients had viral cultures. A cotton swab was rolled in the inferior cul-de-sac and transferred to a carrier medium (sodium chloride, potas sium chloride, dipotassium phosphate, char coal, and agar, pH 7.0). The contents of this tube were either stored at -70°C for a short period or immediately transferred to tissue culture tubes of human diploid fibroblast, primary human embryo kidney cell, or human epidermoid carcinoma. Stan dard culture methods have been described previously.* All viruses isolated in this initial attempt were passed and compared to ref erence prototype virus strains for indication of atypicality.

The methods for viral culture in the 56 epidemic keratoconjunctivitis patients have been described elsewhere. 3 A single patient was cultured for trachoma inclusion con junctivitis (TRIC) agent by other meth ods. 8 Serologic studies-Of the original 64 study patients, 13 did not return for a convalescent serum. Therefore these serologic results are based on 51 patients. We collected sera dur ing the acute phase and once or twice during the convalescent phase. There were two family groups in which a specimen from one uninf ected member was considered acute and another infected member's specimen was considered convalescent. We took a con valescent specimen from 34 of the 51 pa tients within 30 days of the initial visit; 14 within 60 days; and three patients had speci mens that were separated by more than 60 days.

The sera were tested for antibodies to the prototype strains of many respiratory and other viruses in order to survey all of the likely causes of follicular conjunctivitis (Table 1) . Complement-fixation, hemagglutination-inhibition, indirect hemagglutination, or serum neutralization tests were employed as appropriate for each particular virus. The sera were heat-inactivated (56°C, 30 minutes) for all serologic tests. Antibodies to influenza A, B, and C, parainfluenza 1, 2, and 3, mumps virus "viral antigen," rubeola, respiratory syncytial virus (RS v), coronavirus OC-43 and 229E, herpes simplex 1, varicella, Mycoplasma pneumoniae, Bedsonia (TRIC agent), cytomegalovirus, and adeno virus group-specific hexon-antigen were as sayed by the standardized complement-fixa tion test with overnight fixation of 5 units of complement. 4 ' 8 Antibodies to influenza A and B, parainfluenza 1, 2, 3, 4A, and 4B, Newcastle disease virus, mumps virus, rubeola, coronavirus OC-43, vaccinia, and adenovirus types 1 to 33 were assayed by the standardized hemagglutination-inhibition test with 0.01 M phosphate-buffered saline diluent and spectrophotometrically standard ized 0.4% mammalian or 0.5% avian red blood cells. 7 Antibodies to coronavirus 229E were additionally measured by an indirect hemagglutination test. 8 Neutralizing anti bodies to rubeola, RS v, selected adenoviruses, the six coxsackie B viruses, echovirus types 4 and 9, and coxsackievirus A-10 were assayed by serum neutralization tests in human epidermoid carcinoma or monkey kidney tissue culture.' Neutralizing VOL. 80, NO. 4 VIRAL KERATOCONJUNCTIVITIS 663 antibody titers to vaccinia, herpes, coxsackievirus A-9, A-16, and A-21, and four strains of acute hemorrhagic conjunctivitis virus were similarly assayed by serum neutraliza tion tests in human diploid fibroblast tissue culture. Serum neutralization antibodies in coxsackievirus A-4 and A-7 were assayed in suckling mouse brain. 10 For diagnostic purposes a fourfold titer change (either rise or fall) was consid ered significant for any of the serologic tests. In addition a stable titer of > 1:32 in any serologic test (hemagglutination-inhibition, serum neutralization, or complement-fixa tion) suggested infection with that agent if the clinical course and other serologic tests were confirmatory.

When serologic comparisons were carried out, as with the adenovirus groups described below, all of the testing was done in a stan dardized manner. We used only the hemag glutination-inhibition test, with 0.01 M phos phate-buffered saline and standardized mam malian or avian red blood cells, to compare adenovirus types. Antigens for types 1 to 33 were maintained at the Center for Disease Control and were used for similar serologic studies in the past. 7 Other serologic test re sults did not enter into the calculations for significance of titer levels. We tested all sera simultaneously and in duplicate with the same cell lots whenever possible so that test ing conditions and results would be equiv alent.

Clinical examinations-Those patients who had epidemic keratoconjunctivitis 8 were ex amined during the initial phase of their ill ness and approximately one month after res olution of all symptoms. Visual acuity was re corded and, by slit-lamp microscopy, we made a determination of presence or absence of subepithelial infiltrates. There was no at tempt to quantitate these or to judge sever ity.

The patients in the nonepidemic group were treated in a different manner. All pa tients were seen at least twice for evaluation of subjective and objective evidence of ocu lar illness. During each examination, we at tempted to document severity of symptoms, visual acuity, the presence of pre-auricular nodes (or other adenopathy), and the appear ance of the conjunctiva, cornea, and anterior chamber by slit-lamp microscopy. Detailed drawings of the cornea were made in keratitis cases, but the exact number of lesions was not recorded. Instead, we attempted to depict accurately the areas of cornea in volved for future evaluation. Each patient was followed until keratitis showed signs of stability or actual regression. When possible, those patients with persistent lesions were followed for a longer time period.

At the end of the study period, all of the drawings were re-examined. For each pa tient we estimated the percentage of cornea involved with infiltrates during the peak phase of the illness. We assigned values be fore the serologic data were available. When all of the serologies were complete, the adeno virus groups were investigated separately. We calculated the severity of keratitis vs. sero logic titer by hemagglutination-inhibition test values and the percentage of corneal involve ment arrived at previously.

Therapeutic regimen-Almost all of the patients were treated with topical antibiotics, compresses, or astringents. Two of the pa tients in the adenovirus group received « corticosteroids after cessation of the acute phase and when visual acuity decreased. Pa tients in the herpes group were given various antiviral preparations. Although these were generally effective in controlling the acute disease, we did not attempt statistical corre lations to determine efficacy.

Ancillary studies-We took swabs for rou tine bacterial cultures in 26 nonepidemic pa tients simultaneously with the viral cultures. Although most of these were taken at ran dom, occasionally patients presented with signs and symptoms suggestive of bacterial conjunctivitis complicating the viral illness; these patients were cultured specifically to identify possible infectious agents. In addi tion, conjunctival scrapings were examined 664 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER, 197S in 27 patients. These were processed with Giemsa or Gram stains, or both.

Tear studies-Tears were collected from ten patients with acute and chronic, nonepidemic viral infections or chronic noninfectious conjunctivitis. The details for collect ing, processing, and quantitating immunoglobulins in tears have been described pre viously. 11 ' 12 The viral serologic studies on the tear specimens were similar to those done on the sera from these individuals, but the tear specimens were not heat-inactivated and were diluted starting at 1:2 instead of 1:8. (Other studies involving the epidemic keratoconjunctivitis group have been de scribed previously. 8 )

Although more than 90% of the infected eyes were cultured, the yield of positive viral cul tures was low. Table 2* shows the difference between the number of positive isolates ob tained by culturing before or after the eighth * All of the data for the epidemic keratoconjunc tivitis group were published elsewhere, except for the corneal quantitative material in clinical studies.' day of illness. Culturable virus tended to dis appear after the first week of illness in most of the virus groups examined.

Adenovirus isolates (types 3, 7a, 8, and 16) were typical prototype-like strains but three of the adenovirus type 7a viruses were not; these were neutralized or inhibited less well with prototype reference antisera than with the patient's own convalescent serum. 18 This was not clinically significant.

Serologic studies (nonepidemic patients)-Since positive viral cultures were sparse, the restrospective study of acute and convales cent sera constituted the major means of diag nosis. If cultures were positive or if clinical evidence pointed to a specific diagnosis-the presence of herpetic dendrites-the total bat tery of serologic tests (Table 1) was not al ways carried out.

We compared the clinical and serologic diagnoses in the 51 patients whose sera we examined (Table 3 ). All patients with posi tive cultures showed the expected serologic changes. Two patients appeared to have ex perienced infections with two agents, either simultaneously or in close temporal proxim ity. In one patient, the hemagglutination-inhibition titer to adenovirus type 7 was stable at 1:128, indicating a recent infection; at the same time, the culture for TRIC agent was positive. 8 The second patient had an ill ness characterized by initial upper respiratory infection and secondary chronic uniocular keratitis, with comeal ulceration and recur rences. Over the three months of observation during the chronic phase, he developed a gradual fourfold rise in adenovirus type 7 hemagglutination-inhibition antibody (1:32 to 1:128). Almost eight months after his initial illness he demonstrated signs of an terior uveitis after an upper respiratory ill ness (nonspecific) and increased his herpes simplex antibody titer fourfold (1:32 to 1:128) in both complement-fixation and serum neutralization tests. These data sup port a premise of double infections in both patients.

Five additional patients demonstrated sero-VOL. 80, NO. 4 VIRAL KERATOCONJUNCTIVITIS 665 logic changes suggestive of either sequential or simultaneous infection with two or three agents. However, in these patients we could not differentiate clinically among the various agents. Therefore, the serology that best fitted the clinical picture and time sequence of infection was chosen as the final diagnosis.

Most of the clinical misdiagnoses were made in the cases of presumed epidemic keratoconjunctivitis. Adenovirus type 7 was the most prominent masquerader, and herpes simplex passed for epidemic keratoconjunctivitis in one case.

We listed the geometric mean titers for the major adenovirus groups encountered in this study in Table 4 . Although the average fold-rise in geometric mean titers was simi lar, the final absolute titers were different by more than twofold. This difference was sta tistically significant.

In an attempt to explain the higher final titer of the Group 1 adenovirus-infected pa tients, we looked at two different factors. The first was a consideration of a possible differ ence in the acute and convalescent phase se rum samples ( Table 5 ). The two groups were similar in makeup in that nine of 16 Group 1 patients and ten of 13 Group 2 were examined during the acute phase of their illness: they showed a fourfold rise in their titer to the infecting adenovirus.

There was a moderate disparity between patient history and the serologic data. Three Group 1 patients (Cases 8,9, and 10) claimed they were ill for a few days, but their titer reflected a long-term illness (> 32 initial or no change with time, or both). We ob served one patient (Case 8) with an un changing 1:16 titer throughout his illness, and we could not explain the lack of change in titer. This disparity was less evident in Group 2 (Case 26).

The difference in acute phase geometric titer for each group was nearly significant .05 < P < .06). However, when we con sidered only patients proven to be in the acute phase, the difference was not signifi cant (5.6 vs. 9.3; P > .1). There was a difference in the number of days between acute and convalescent samples for the two groups. Overall, the mean num ber of days was 21 for Group 1 and 29 for Group 2 (not significant). However, if only the 19 patients in the acute phase were again considered, the averages were 19 days and 32 days, respectively (P <.05; t = 2.34; df, 17). (From these data, a higher titer should have been expected for Group 2 be cause of the longer interval between acute phase and serum peak.) Thus, neither the titer of acute phase sera nor the sampling time of convalescent phase sera seemed to ac count for the difference in final antibody levels.

We also considered the degree of upper respiratory or generalized illness among pa tients in both groups. Of the 21 patients with serologic evidence of Group 1 infection, only 16 gave accurate histories; nine of these reported an associated upper respiratory in fection with conjunctivitis. In contrast, one (with adenovirus type 9) of the 13 patients in Group 2 exhibited signs or symptoms of upper respiratory infection. This differ ence was again significant (P < .01, \ 2 test).

Thus, those with Group 1 nonepidemic adenovirus infection were afflicted with more upper respiratory infections, but these in- fections were not necessarily responsible for the significantly higher serologic titers. In fact, there was no significant relationship be tween the presence of an upper respiratory infection and a serologic response that equaled or exceeded 1:128 (Table 6 ). This negative correlation held true for changes in both homologous and heterologous virus titers. Therefore, the explanation for the higher serum titers in patients with Group 1 infections may not be related directly to the concomitant upper respiratory infections.

We compared the sensitivity of the two most commonly employed serologic tests (Table 7) . Hemagglutination-inhibition tests demonstrated fourfold rises in many cases where complement-fixation tests failed to show significant changes. Hence, primary de pendence on the usual complement-fixation hexon-antigen test would have resulted in fewer positive diagnoses, even in acute infec tions.

Ancillary laboratory studies (nonepidemic patients)-We microscopically examined conjunctival smears from 21 patients, and routine bacterial cultures were taken from 11 of these patients plus ten other individuals. Cultures from five patients showed coagulasepositive staphylococci; these patients had various diagnoses: adenovirus type 8 epi demic keratoconjunctivitis, adenovirus type 7 pharyngoconjunctival fever, herpes sim plex keratitis, and a case of chronic "fac titious" conjunctivitis.

The smears were helpful only in cases of herpetic infections. In three patients with herpes simplex and one with herpes zoster, 3 Twenty-eight patients showed signs of keratitis during their illness and had chronic subepithelial infiltrates when examined one month after resolution of their illness (16 of these experienced some degree of decreased visual acuity as well). The 28 remaining patients, who were also infected during the first wave of the epidemic were seen at the same time and had no subepithelial infiltrates. The peak geometric mean hemagglutination-inhibition titer of the for mer group with residual infiltrates was 1:34.5 (range 1:10 to 1:160). The data for the un affected group were 1:102.5 geometric mean titer (range 1:20 to 1:320). This difference was significant (P < .001). (The collection of sera and intervals between examinations were similar for patients because they were part of an epidemic group. 3 The serologic tests were exactly the same as those carried out in the nonepidemic group, and often were done simultaneously t Estimated maximum percent of cornea with epithelial or subepithelial infiltrates. * Titers did not change significantly (> fourfold) during the observation period (compare with Table 5) analysis of corneal disease. Group 1 was composed of five patients with adenovirus type 3 and 11 patients with adenovirus type 7 infections. Group 2 contained 11 patients with adenovirus type 8 and one each with adenovirus type 9 and adenovirus 29. The data concerning the corneal involve ment and antibody levels of both groups were compiled (Table 8, compare with Table 5 ). Analysis by x 2 testing showed several sig nificant correlations. First, the number of eyes involved with corneal changes differed significantly in the two groups. In Group 1, there were five patients with no involve ment, seven with one eye, and four with both eyes affected. Group 2 had only one patient without corneal changes, three pa tients with a single eye involved, and nine patients with both eyes involved (x 2 = 5.94; df, 2; PS2.05). When this same factor VOL. 80, NO. 4

(number of eyes) was used, and the antibody levels were compared for all adenovirus infections (Groups 1 and 2) , a more signifi cant difference was noted. At a level of 1:64, there were 12 doubly involved patients, four with one eye, and three with no corneal in volvement. At a level of 1:128, only one pa tient showed infiltrates in both eyes, six had single eye involvement, and three had no ocu lar involvement. The difference between the number of patients in these three categories proved to be statistically significant (x 2 = 7.65;df,2;P<.025).

Correlations of the degree of corneal in volvement with antibody level were also car ried out. There was no difference between the two groups in the number of patients with central lesions vs. peripheral lesions of the cornea. This was reflected in the patients with decreased visual acuity (Table 8) .

We devised a comparison of the two groups in terms of fold-rise in antibody titer (Fig  ure) . (The fold-rise was a reflection of the rate of antibody production as well as the amount of antibody.) As evidenced by the graph, the responses of the two groups were almost diametrically opposed. For each foldrise in titer, the Group 1 patients had pro gressively less corneal involvement. For the same change in titer, the Group 2 patients showed progressively more corneal infiltra tion. However, the final level of antibody achieved was lower in Group 2 (compare with Table 4 ).

A similar difference in the two groups was noted at the 40% level of corneal infiltration. Only two patients (of 16) in Group 1 had infiltrates that involved > 40% of the corneal surface area. In Group 2, six of 13 patients were affected in the same way ( x 2 = 5.73; df, 1; P<.025).

Finally, age was not a factor differentiat ing the two groups. The average age of Group 1 patients was 29.9 years; Group 2 patients averaged 30.5 years. The difference was not statistically significant (t = .128; df, 25).

A comparison of the percent corneal involvement with fold-rise in antibody titer among patients infected with Group 1 (AV 1) and 2 (AV 2) adenoviruses. r represents the correlation coefficient of the least-squares line calculated to fit the data. Only patients with at least a fourfold rise are included. GROUP)-Rises in titers to heterologous vi ruses were observed in Groups 1 and 2 infec tions. Thus, in patients infected with adeno virus type 7 (Group 1), a significant change in titer to this agent occurred, as well as smaller changes to such related types as 3, 11, and 21 (Table 8 ). However, rises in heterologous titers were more consistently seen in patients with Group 2 infections. For example, adenovirus type 8 infections produced rises to 12 related viral types in the Group 2 series. This was reflected in the higher heterologous average titers (Table  8 ). In particular, we noted a tendency for heterologous responders within family groups. In one of these families, all of the blood relatives showed changes in titer to several other Group 2 subtypes after adeno virus type 8 infection. However, the spouse of one member, exposed to the same virus, had a serologic change to adenovirus type 8 alone (homologous responder). 670 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER, 1975 We observed patients with keratoconjunctivitis from other causes during this study (Table 3) , but only adenovirus patients formed homogenous groups that lent them selves to statistical analysis. The herpes group was large (eight herpes simplex, two herpes zoster, two possible mixed herpes simplex-zoster, and three presumed cytomegalovirus), but heterogenous in terms of acute and chronic illness, treatment with var ious antiviral drugs or corticosteroids, and poor serologic responses. Variability in his tory, follow-up, and other factors occurred among the patients with adenovirus so the number of patients who met the criteria for various analyses often differed from one test to another. In terms of treatment, however, almost all of the patients with adenovirus required only therapy with topical vasocon strictors or antibiotics, or both. Only two patients (one each from Groups 1 and 2) required corticosteroids for symptomatic re lief, and this was instituted after achieving maximum corneal scores.

Analysis of tears-Ocular external secre tions (tears) were analyzed in ten patients. In general, there were low antibody titers. Four patients with serologic diagnoses of herpetic infection represented chronic, re current keratoconjunctivitis or kerato-uveitis. Only one patient with disci form keratitis and uveitis had detectable tear-borne neutralizing antibody to herpes simplex. His serum titer, however, was not greater than 1:8 in a simul taneous neutralization test. One other patient showed tear antibody to herpes zoster and se rologic changes consistent with simplex. Whether this represented true simultaneous infection or the presence of cross reaction was not determined.

Two patients had high titers (1:8, 1:16) to coronavirus OC-43 in their external secretions. Upper respiratory infection has produced antibody in tears 11 and this may represent that kind of relationship.

The predominant immunoglobulin species in the tears was IgA, in agreement with pre vious work. 11 ' 12 However, we made no at tempt to prove that the measurable hemagglutination-inhibition or serum neutralization antibodies were unique to IgA. DISCUSSION Because all patients in the second part of this study were seen during a nonepidemic time period, the cases of keratitis were varied in their etiologies (Table 3 ). This observa tion agrees with previous studies that de scribed a similar variety of viral agents re sponsible for epidemic keratoconjunctivitis keratitis. 16 " 18 The behavior of the individual viral infec tions also conformed to expected patterns. Thus, Group 1 adenoviruses caused more upper respiratory and generalized illness, but less keratitis, than Group 2 infections. 1 ' 2 - [19] [20] [21] Herpes viruses caused more severe keratitis and kerato-uveitis, but clinical illnesses cor related poorly with serologic changes in anti body titer. 28 ' 23 We also encountered difficulty in culturing any virus after seven to ten days of symptoms. This has been reported before for epidemic keratoconjunctivitis, but it is probably true for many other viral syndromes. 2,24- 28 We, like others, also found the hemagglutination-inhibition test to be supe rior to complement-fixation as a serum diag nostic procedure in adenovirus infection. 16 No procedure gave consistent results with tears.

There were differences in the extent of corneal involvement observed in Group 1 and 2 infections. The correlations indicate that Group 2 viruses produced a lower level of specific antibody titer (despite a longer time interval between the collection of acute and convalescent sera), and induced more corneal disease in the affected indi viduals.

These data suggest that the higher anti body titer in Group 1 protected the cornea, supported by examination of the keratocon junctivitis epidemic of adenovirus type 8, a Group 2 virus. In the latter patients, a range of antibody was observed that was somewhat higher than that of the nonepidemic Group 1. (The reason for this difference is not readily apparent, but does not lie in testing differences, since both sets of sera were ex amined under identical conditions and often simultaneously.) What is more important, however, is the range of corneal involve ment with this single viral agent. Once again, the patients with higher antibody titer ex perienced less corneal involvement with this usually virulent agent. 17 ' 19 The corneal infiltrates may only represent a pattern of infection and the antibody re sponse to it. Certainly more than one viral type can cause this pattern, as shown in the nonepidemic group. Also, the same virus can cause different degrees of involvement (epi demic study), perhaps depending on the level of antibody illicited by the infection.

Since delayed hypersensitivity was not spe cifically assayed, the elevated humoral anti body in epidemic Group 2 or nonepidemic Group 1 patients may reflect a higher degree oi overall immunity, and cellular immunity may be the chief deterrent of corneal dis ease. However, systemic antibody alone may protect the cornea (and even the fellow eye) from severe involvement. Whether it was humoral or cellular, or both, the nonepidemic and epidemic groups of patients displayed a reasonably similar pattern of protection that correlated with high antibody titer.

We only speculate about the pathogenesis of the subepithelial infiltrates. In studies of immune-complex disease, for instance, rheu matoid arthritis and Aleutian mink disease, antigen-antibody complexes in blood vessel walls can attract complement and inflamma tory cells that cause the damage. 87 ' 88 These complexes are made to persist by maintain ing a low titer antibody, and the disease may be completely eliminated by raising the titer. 88 The cornea is similar to the collagenous tis sue in the walls of blood vessels, 29 and these tissues can absorb certain antigens. 80 ' 81 Therefore, it seems feasible that viral par ticles might be bound to the stromal collagen, form complexes with low levels of antibody diffusing within the cornea, and thereby act as a nidus for inflammation. The subsequent fixation of complement and attraction of in flammatory cells completes the subepithelial infiltrate seen by slit-lamp microscopy. Corticosteroids would disrupt the inflammatory response, but it would not remove the com plexes. Hence, discontinuation of the drug would allow inflammatory cells to reappear. 88 A similar set of circumstances was proposed in the pathogenesis of stromal disease in herpes simplex keratitis. 88 Based on these suppositions, one may con clude that the sometimes permanent visual damage observed in epidemic keratoconjunctivitis might be decreased or eliminated by enhancing the immune response of the in fected patient, perhaps during the acute in fection.

Of 120 individuals suffering from follicular conjunctivitis, with or without keratitis, 56 had epidemic keratoconiunctivitis caused by adenovirus type 8. The remaining 64 pa tients had keratoconjunctivitis produced by several different viruses, including herpes simplex, adenoviruses types 3, 7, 8, 16, 21, and 29, and others. Positive viral cultures were difficult to obtain after the first week of illness in most viral infections. Serologic tests were far more successful in identifying causative agents. Hemagglutination-inhibition tests in adenovirus infections proved to be fast, accurate, and more sensitive than other serologic tests. Among the epidemic and nonepidemic adenovirus groups, the de gree of corneal involvement in the infections appeared to be correlated with the level anti body against the infectious agent. Patients with low antibody titers had more severe keratitis than those with high titers. This correlation was true for both the epidemic and nonepidemic patients.

Microbiology of the Eye

Adenovirus type 8 keratoconjunctivitis in the United States. 3. Epidemiologic, clini cal and microbiological features

Epidemic keratoconjunctivitis. A severe industrial outbreak of adenovirus type 8

Adenoviruses

Simplified irradiated McCoy cell culture for iso lation of Chlamydia

Standardized diagnostic com plement fixation method and adaptation to micro test, No. 1228

Standardized viral hemagglutination and hemagglutination-inhibition tests. 2. Description and statistical evaluation

Detection of coronavirus 229E antibody by indirect hemagglutination

Antibody in tears following intranasal vaccination with in activated virus. 2. Enhancement of tear antibody production by the use of polyinosinic: polycytidilic acid (poly I:C)

Epidemic keratoconjunctivitis on Taiwan. Etiological and clinical studies

Adenovirus type 8 infections in the United States

Adenovirus infection of the conjunctiva and cornea

Association of adenovirus type 8 with epidemic keratoconjunctivitis

Diagnosis of adenovirus infections by use of fluorescent antibody technique

Corneal im mune response to herpes simplex virus antigens

The pathogenesis of chronic inflam mation in experimental antigen induced arthritis. 2. Preferential binding of antigen antibody complexes to collagenous tissues

The patho genesis of corneal inflammation due to herpes sim plex virus