Acute retinal necrosis caused by reactivation of herpes simplex virus type 2

Acute retinal necrosis is a severe form of necrotizing retinitis. Acute retinal necrosis has been demonstrated to be caused by varicella-zoster virus and herpes simplex virus type 1. We treated three patients with acute retinal necrosis apparently caused by recrudescence of latent herpes simplex virus type 2. Primary viral infection was probably congenital, with documented perinatal herpes simplex virus type 2 infection in two patients. Bilateral chorioretinal scars were present in two patients, neither of whom had a history of ocular herpetic infection, suggesting that earlier subclinical chorioretinitis had occurred. In each case, periocular trauma preceded the development of retinitis by two to three weeks. These cases are evidently caused by trauma-induced reactivation of latent virus rather than the onset of a primary infection.

Immunolocalization of integrins in proliferative retinal membranes

PURPOSE

Integrins are cell surface adhesion molecules that serve as receptors for extracellular matrix components or for other cells. Integrins help regulate processes such as cell proliferation, migration, and differentiation. These processes are thought to have fundamental roles in the pathogenesis of proliferative retinal membranes in diseases such as proliferative vitreoretinopathy (PVR) and proliferative diabetic retinopathy (PDR). Therefore, the authors sought to determine the expression pattern of integrins in human proliferative membranes.

METHODS

Tissue was obtained from two patients with PVR, two with PDR, and one subretinal neovascular membrane from a patient with presumed ocular histoplasmosis. Integrins were detected with an avidin-biotin-complex immunohistochemical technique using nine different monoclonal antibodies specific for alpha subunits 2, 3, 4, 5, 6, and V, and beta subunits 1, 2, and 3.

RESULTS

All integrin subunits studied were detectable to varying degrees in proliferative membranes. beta 1 and alpha 6 were especially prominent at the edges of most PVR and PDR membranes. Pigmented cells expressed up to nine different integrin subunits, in contrast to normal RPE cells, which immunostained for only alpha 4 and beta 2. Proliferative diabetic retinopathy vessels expressed all nine integrin subunits examined, including alpha 4, which was poorly expressed in vessels of nondiabetic retinas.

CONCLUSIONS

Integrin subunits are readily detectable in pathologic membranes. Both PVR and PDR are associated with altered integrin expression in vascular endothelium and pigmented cells. The distribution of integrins at the edge of a membrane suggests a role in the growth or contraction of the membrane, presumably by participating in the interaction between cells and substances such as vitreous collagen. Therefore, integrin antagonists may hold promise for the treatment of proliferative retinopathies.

Immunolocalization of integrins in the human retina

PURPOSE

Integrins are cell surface proteins that participate in interactions between cells and with extracellular matrix. Binding of integrins to their ligands influences cell activities including proliferation, migration, and differentiation. Expression of integrin subunits from three different subfamilies were examined in human retina.

METHODS

Integrins were detected in frozen sections of two human retinas with an avidin-biotin-complex immunohistochemical technique, using nine different monoclonal antibodies specific for alpha 2, alpha 3, alpha 4, alpha 5, alpha 6, alpha v, beta 1, beta 2, and beta 3. One retina was from a patient who had conjunctival squamous cell carcinoma, and the other was from uninvolved regions of an eye with a choroidal melanoma.

RESULTS

All integrins tested were detectable in consistent patterns in two retinas. All except alpha 2 and alpha 4 were stained vibrantly in retinal and choroidal vessels. All alpha subunit staining of vessels showed overlap or close proximity to beta 1 staining. In addition to vessels, beta 1 was also present in the internal limiting membrane; alpha 2, alpha 3, alpha 4, alpha 5, and beta 2 were all found throughout much of the neural retina, albeit with distinctive staining patterns. Other than in association with vessels, alpha 6 and alpha v were not detected in neural retina, and beta 3 was only weakly detected in the nerve fiber layer; alpha 4 and beta 2 were expressed in the retinal pigment epithelium; beta 1 and beta 2 were strongly expressed in drusen present in one of the eyes.

CONCLUSION

Nine integrin subunits have been found to have unique distributions in adult human retina. An understanding of the distribution in normal retina can serve as a useful contrast to patterns of staining associated with retinal diseases.

Cytokine mRNA levels in rat ocular tissues after systemic endotoxin treatment

PURPOSE

An intertwined cascading network of cytokines is believed to direct the development of endotoxin-induced uveitis (EIU). This study investigated mRNA levels of interleukin (IL) 1 alpha, IL-1 beta, IL-6, tumor necrosis factor (TNF)-alpha, interferon (IFN)-gamma, transforming growth factor (TGF)-beta 1, and the helper T lymphocyte marker, CD4, during the course of EIU in rats.

METHODS

Reverse transcription followed by polymerase chain reaction amplification was used to determine relative mRNA levels in four ocular tissues (iris/ciliary body, cornea, lens, and neuroretina) at 0, 1, 3, 6, 24, and 48 hours after subcutaneous injection of 200 micrograms of Escherichia coli endotoxin.

RESULTS

Four general patterns of mRNA expression were observed: (1) constitutively expressed and unaffected by endotoxin; (2) constitutively expressed but further induced by endotoxin, reaching peak levels at 3 hours postinjection; (3) initially undetectable or marginally detectable and induced by endotoxin, with peak levels occurring 3 hours postinjection; and (4) never present at appreciable levels. The most dramatic responses were seen in the mRNA levels of IL-1 alpha, IL-1 beta, IL-6, TNF-alpha, and IFN-gamma in iris/ciliary body. Lesser mRNA level responses were found for IL-1 beta and IL-6 in cornea and for IL-1 alpha, IL-1 beta, IL-6, TNF-alpha, and IFN-gamma in neuroretina. Little or no changes in mRNA levels were observed for CD4 or TGF-beta 1 in any tissue or for any mRNA examined in lens.

CONCLUSIONS

These data show that subcutaneous endotoxin induces cytokine mRNA expression differentially in ocular tissues. These data support the hypothesis that induction of cytokine expression in iris/ciliary body plays a major role in the development of EIU.

Local proliferation of fibroblast-like synoviocytes contributes to synovial hyperplasia. Results of proliferating cell nuclear antigen/cyclin, c-myc, and nucleolar organizer region staining

OBJECTIVE

To test the hypothesis that local proliferation contributes significantly to the hyperplasia of rheumatoid synovium.

METHODS

Immunohistologic and chemical staining was used to identify 3 markers of cell proliferation: proliferating cell nuclear antigen, c-myc proto-oncogene, and nucleolar organizer regions. Synovium from 21 patients with rheumatoid arthritis, 34 with degenerative joint disease, and 7 with joint trauma was examined.

RESULTS

All 3 markers indicated substantial, active proliferation of synovial lining cells in synovium with hyperplasia. Proliferating cells showed type I procollagen immunoreactivity but were negative for CD68, a monocyte/macrophage marker. Proliferation was greater in rheumatoid arthritis than in the other conditions evaluated.

CONCLUSION

In situ proliferation of fibroblast-like synoviocytes in the synovium lining contributes considerably to the increase in cell numbers in rheumatoid synovium.

Treatment of severe refractory uveitis with intravenous cyclophosphamide

OBJECTIVE

We assessed the efficacy of intravenous (iv) cyclophosphamide therapy for patients with uveitis.

METHODS

Eleven patients with uveitis were treated with iv cyclophosphamide. Patients were from a tertiary care, referral practice, had failed oral corticosteroid therapy, and had disease that interfered with activities of daily living.

RESULTS

Only 5 patients benefitted and the improvement was sustained without additional immunosuppression in only 2 patients, one of whom had only mild eye disease and was receiving cyclophosphamide for a central nervous system vasculitis. Four patients worsened during the therapy.

CONCLUSION

Although iv cyclophosphamide may have fewer risks than daily, oral cyclophosphamide, the majority of patients with uveitis do not experience a sustained benefit from this therapeutic approach.

Efficacy of antibodies to adhesion molecules, CD11a or CD18, in rabbit models of uveitis

Adhesion molecules play a critical role in leukocyte emigration to a site of inflammation. In order to assess the potential therapeutic benefit of blocking adhesion molecule function in anterior uveitis, the efficacy of antibodies to specific adhesion molecules was tested in 3 separate rabbit models of anterior uveitis. Antibodies to two different leukocyte molecules, CD11a and CD18, and antibodies to the endothelial ligand for CD11a/CD18, ICAM-1 (intercellular adhesion molecule-1, CD54), were studied in inflammation after intravitreally injected interleukin-1, intravitreally injected endotoxin, or an ocular reversed passive Arthus reaction. The CD18 antibody (2 mg/kg intravenously) reduced the cellular infiltrate with each of these 3 models. The antibody to CD11a was equally effective but was tested only in the IL-1-induced model. The antibody to ICAM-1 reduced the cellular infiltrate associated with this model, but the results did not reach statistical significance. None of the antibodies was able to reduce the associated increase in vascular permeability as measured by protein in the aqueous humor. The antibody to CD18 failed to reduce the inflammation if it was administered 24 hours after the intravitreally injected endotoxin. These observations demonstrate that leukocyte migration into the anterior segment of the eye is dependent on the CD11a/CD18 complex.

Failure to inhibit endotoxin-induced uveitis with antibodies that neutralize tumor necrosis factor

Many of the effects of bacterial endotoxin are attributed to the cytokines that it induces. Tumor necrosis factor alpha (TNF alpha) may be responsible for many of the pathophysiologic effects of endotoxin because antibodies that neutralize TNF alpha reduce endotoxin-induced mortality in experimental animals and many of the activities of endotoxin can be mimicked by the infusion of TNF alpha. Accordingly, we tested the ability of an antibody that neutralizes TNF alpha in a well characterized rat model, endotoxin-induced uveitis. Rats receiving subcutaneous endotoxin had a prompt rise in serum TNF alpha that was eliminated by the monoclonal antibody. Small quantities of TNF alpha were detectable in aqueous humor after either subcutaneous or intravitreal injection of endotoxin. Endotoxin-induced uveitis as quantitated by cells and protein in aqueous humor was, however, not affected by the antibody. Similarly, intraocular inflammation was not reduced by the intravitreal injection of the neutralizing antibody. Although TNF alpha may mediate many of the systemic effects of endotoxin, its role in the intraocular inflammatory effects of endotoxin should be questioned.

Ocular inflammatory effects of intravitreally injected interleukin-2

The primary, known physiologic effect of interleukin-2 (IL-2) is to act as a T lymphocyte growth factor. We investigated the potential contribution of IL-2 to intraocular inflammation by studying the inflammation resulting from the intravitreal injection of recombinant, human IL-2 in New Zealand white rabbits. Serial slit lamp observations indicated that 40 microgram of intravitreally injected IL-2 induced an anterior uveitis which was maximal 5 days after the injection. Inflammation was less marked but still significant with amounts of IL-2 as low as 400 ng. Direct examination of aqueous humor confirmed elevations of protein, prostaglandin E2, and mononuclear cells which correlated with the clinical observations. The kinetics of the response to intravitreal IL-2 distinguished it from the responses to other intravitreally injected cytokines such as interleukins 1, 6, or 8 as well as tumor necrosis factor. Intramuscular injection of cyclosporine A significantly reduced the protein extravasation associated with IL-2 injection, but cyclosporine had no effect on inflammation secondary to an intravitreal injection of interleukin-1. These observations implicate IL-2 as a potential contributor to uveitis. In addition, the studies with cyclosporine indicate the heterogeneity of inflammation such that pharmacologic agents which affect one cause of uveitis are not necessarily efficacious in another model.

Retinal pigment epithelial cells produce interleukin-1 beta and granulocyte-macrophage colony-stimulating factor in response to interleukin-1 alpha

The retinal pigment epithelium (RPE) is clinically involved in diverse ocular inflammatory diseases. Because perturbed RPE cells produce a variety of inflammatory substances, RPE cells may play an integral part in these diseases. Interleukin-1 (IL-1) and granulocyte-macrophage colony-stimulating factor (GM-CSF) are pleiotropic cytokines with the ability to trigger numerous inflammatory responses. This report shows that cultured human RPE cells synthesize interleukin-1 beta (IL-1 beta) and GM-CSF in response to the potentially inflammatory cytokine, IL-1 alpha, but not to E. coli endotoxin. Control RPE cells made little or no mRNA or protein for either IL-1 beta or GM-CSF. Upon stimulation of the cells by IL-1 alpha, both IL-1 beta and GM-CSF mRNAs were readily apparent by 3 hours, persisted for over 24 hours, and were translated into immunologically detectable proteins. GM-CSF protein was secreted into the culture medium, whereas IL-1 beta protein remained cell associated. The IL-1 alpha-induced mRNA and protein production were inhibited by dexamethasone. These observations provide additional evidence that RPE cells are capable of playing a pivotal role during ocular inflammation.

Recognition of posterior scleritis and its treatment with indomethacin

Posterior scleritis is a rare disease, the clinical signs of which may vary. There is no consensus on the appropriate method of treatment for this disease. Some have advocated treatment with nonsteroidal antiinflammatory drugs, whereas others have reported variable success with potentially more toxic therapy. Patients with posterior scleritis constitute approximately 1% of the uveitis clinic population at the Casey Eye Institute in Portland, Oregon. The cases of 6 patients with posterior scleritis, the diagnosis of which was elusive, are reported. Posterior scleritis was generally confirmed by ultrasound examination. Each of the patients responded to treatment with indomethacin, which was usually the sole method of nontopical therapy. Thus, recognition of this relatively rare disease had marked implications for treatment.

Intraocular lymphoma. Immunopathologic analysis of vitreous biopsy specimens

Immunologic analysis of cell surface markers (immunophenotyping) has become a standard procedure in the evaluation of systemic lymphomas. However, attempts to apply these techniques to intraocular lymphoma have not been uniformly successful. We successfully immunophenotyped five consecutive cases of intraocular lymphoma using immunoperoxidase surface marker analysis of cytocentrifuged specimens in two cases and flow cytometry in three. In all five cases, a monoclonal B-cell population was unequivocally present. Contrary to previous reports, we found surface marker analysis of vitreous biopsy specimens to be helpful in the diagnosis and treatment of intraocular lymphoma. Not only did it support the cytologic diagnosis but it allowed comparison of the immunophenotype of vitreous infiltrates with that of previous or subsequent lymphomatous lesions from nonocular sites.

Activity of an interleukin 1 receptor antagonist in rabbit models of uveitis

Interleukin 1 has been implicated in intraocular inflammation. The availability of a cloned, recombinant interleukin 1 receptor antagonist has enabled us to test the role of interleukin 1 in specific models of uveitis in New Zealand white rabbits. Seventy-five micrograms of interleukin 1 receptor antagonist injected intravitreally resulted in a 97% reduction in aqueous humor cells present 6 hours after intravitreal injection of 10 ng of human interleukin 1 alpha. Disruption of the blood aqueous barrier was prevented by the receptor antagonist (mean +/- SD aqueous humor protein of 0.6 +/- 0.1 g/L in rabbits treated with interleukin 1 receptor antagonist vs 32.2 +/- 9.9 g/L in controls). Lower doses of interleukin 1 produced more modest but significant inhibition. Despite the activity of interleukin 1 receptor antagonist in inhibiting interleukin 1-induced inflammation, interleukin 1 receptor antagonist did not produce significant reduction in inflammation subsequent to an active Arthus reaction or subsequent to the intravitreal injection of 125 ng of endotoxin. A potential explanation of these observations is that cytokines in addition to interleukin 1 may be present in sufficient quantities to produce intraocular inflammation or that the effects of interleukin 1 may be primarily intracellular (intracrine) and therefore resistant to the activity of exogenously administered receptor antagonist.

Reduction of endotoxin-induced vascular permeability by monoclonal antibodies against lipopolysaccharide determinants

Endotoxin, a bacterial lipopolysaccharide implicated in the pathogenesis of septic shock, markedly alters vascular permeability following intravenous injection in rabbits. We investigated the ability of murine monoclonal antibodies to confer protection against endotoxin-induced increases in a rabbit model of ocular vascular permeability. Four monoclonal antibodies of differing specificities as well as polymyxin B were compared for their effects on endotoxin from either Escherichia coli or Pseudomonas aeruginosa. Preincubation of endotoxin with antibodies directed against Pseudomonas O side chain or core glycolipid resulted in marked attenuation of vascular permeability due to Pseudomonas endotoxin, but not E. coli endotoxin. Antilipid A antibodies were not significantly effective in neutralizing either endotoxin with in vitro preincubation. Low avidity of the antilipid A antibody, low density of lipid A binding sites, or inaccessibility of the lipid A may have prevented more marked interactions. When administered intravenously prior to endotoxin challenge, none of the antibodies demonstrated the ability to provide specific protection to subsequent endotoxin in this model. They did provide partial nonspecific protection against endotoxins regardless of epitope specificity. When administered prophylactically, polymyxin B, an antibiotic that binds to lipid A, was highly effective in neutralizing the toxic effects of endotoxin. Since antibodies to lipid A reduce mortality in septic shock, the failure to demonstrate efficacy in this study may be due to the marked sensitivity of the rabbit eye to endotoxin. Alternatively, beneficial effects from antiendotoxin antibodies in septic shock may be unrelated to the inhibition of vascular permeability. Some protection from antiendotoxin antibodies may be due to enhancement of nonspecific mechanisms.

Production and modulation of interleukin 6 synthesis by synoviocytes derived from patients with arthritic disease

Interleukin 6 (IL-6) is a potent cytokine, the biological activities of which include the stimulation of immunoglobulin secretion, T cell activation, induction of the acute phase response, activation of megakaryocytes, and pyrogenicity. These biological activities make it a plausible contributor to rheumatoid arthritis. The ability of synoviocytes to synthesise this potential mediator of inflammation was tested. Cultures of fibroblast-like cells were established from joint tissue from patients with rheumatoid arthritis, degenerative joint disease, or trauma. Supernatants from synoviocytes from each diagnostic category contained IL-6-like activity as detected in a B9 plasmacytoma cell proliferation assay. Supernatants from IL-1 stimulated synoviocytes from patients with rheumatoid arthritis (n = 5) contained an average of 70,000 U/ml IL-6. Western blot analysis confirmed that these supernatants contained peptides that reacted with a highly specific antibody to IL-6. A cDNA probe specific for IL-6 hybridised with mRNA derived from synoviocytes representative of each disease state. Interleukin 6 mRNA expression increased by culturing synoviocytes in the presence of 10% calf serum, IL-1 (30 U/ml), insulin (166 ng/ml), or basic fibroblast growth factor (16 ng/ml). In contrast, dexamethasone (10(-6) mol/l) suppressed the ability of IL-1 to increase the expression of IL-6 mRNA. Recombinant IL-6 itself did not detectably upregulate its own message. The regulation of production of IL-6 by synoviocytes may be important in the pathogenesis of joint inflammation.

Acute anterior uveitis and spondyloarthropathies

An acute onset, unilateral anterior uveitis occurs during the course of either Reiter's syndrome or ankylosing spondylitis. Conversely, many patients who suffer from an acute anterior uveitis are HLA-B27-positive and have associated joint disease. The consistent presentation of the uveitis can aid in the process of differential diagnosis. This article includes a discussion of the recognition of the characteristic presentation, the complications, the role of B27 testing, the relevance of animal models, the pathogenesis, and treatment.

Retinal pigment epithelial cells secrete interleukin-6 in response to interleukin-1

Interleukin-6 (IL-6) is a peptide whose properties include the ability to activate T-lymphocytes, stimulate the secretion of immunoglobulin, induce neuronal differentiation, and trigger the release of acute phase proteins. We have detected IL-6-like activity in conditioned medium from cultured human retinal pigment epithelial (RPE) cells with a bioassay based on the ability of IL-6 to induce the proliferation of murine B-9 plasmacytoma cells. Biologic activity increased approximately 90-fold when the cells were cultured in the presence of IL-1 alpha (30 units/ml). Western blot analysis confirmed that conditioned medium from IL-1 alpha-stimulated RPE cells contained peptides with molecular weights ranging between 19,000 and 30,000 and reactive with antibody to IL-6. Finally, Northern blot analysis indicated that cells cultured in the presence of interleukin-1 contained a 1.2 kilobase transcript that hybridized to a cDNA probe specific for IL-6 messenger RNA. IL-6 peptide on Western blots and mRNA on Northern blots were undetectable unless cells were cultured in the presence of IL-1 alpha. Although IL-6 is synthesized by a variety of cell types, this report is the first to detect its synthesis by an eye-specific cell type. Furthermore, these observations indicate that retinal pigment epithelial cells respond to IL-1, a cytokine that previously has been implicated in ocular inflammation.

Use of a soluble interleukin-1 receptor to inhibit ocular inflammation

Interleukin-1 (IL-1) has been strongly implicated as an inflammatory mediator in anterior uveitis. Recently, solubilized receptors have been utilized to block the binding of viruses to cell membranes or to inhibit cytokine activity. We have tested the activity of an intravitreally injected soluble, human interleukin-1 receptor in a rabbit model of IL-1-induced inflammation. 3 ug of the soluble receptor markedly inhibited both the cellular infiltration and the protein extravasation that followed 6 hours after an intravitreal injection of 10.5 ng of recombinant human interleukin-1 alpha. The efficacy of the soluble receptor was less marked 24 hours after the IL-1 injection. The cellular infiltrate was not reduced at all if the IL-1 receptor was injected 2 hours after the IL-1. The activity of the soluble receptor deserves further study as a therapeutic modality for inflammatory eye disease.

Uveitis precipitated by nonpenetrating ocular trauma

Although penetrating trauma is a well-recognized cause of uveitis, the role of nonpenetrating trauma in initiating uveitis is not defined. We analyzed the records of 496 patients seen at the uveitis clinic at our institution. Twenty-four of these 496 patients (4.8%) suspected that the cause of their intraocular inflammation was related to previous nonpenetrating trauma. In contrast, only one of 251 patients (0.4%) attending the general ophthalmology clinic for routine care provided a history of recent trauma or attributed the present ocular complaint to trauma (P less than .02). Patients with posttraumatic uveitis were usually male (19 of 24, 79%), younger (31 +/- 16 years) than the average patient examined in the uveitis clinic, and more likely to have unilateral disease. In ten (42%) of the patients the trauma was work-related. Bilateral inflammation was seen in eight (one third) of the patients and 17 of 28 patients (71%) had a considerable degree of inflammation posterior to the lens. Many patients had an identifiable cause of uveitis such as ankylosing spondylitis, Reiter's syndrome, sarcoidosis, or acute retinal necrosis; but most patients had no known predisposition. The role of nonpenetrating trauma in initiating uveitis has implications for diagnosis and treatment.

Synthesis of platelet activating factor by ocular tissue from inflamed eyes

Platelet activating factors (PAFs) are a family of ether lipids with properties that suggest a major role in inflammation. We have previously implicated PAFs in ocular inflammation based on the inhibition of several rabbit models of iritis with a specific PAF receptor antagonist. We have tested ocular tissues for the ability to synthesize PAF. Iris, ciliary body, cornea, and/or retina were carefully dissected from New Zealand white rabbits, and tissue from four eyes was pooled. Tissues were stimulated with calcium ionophore (10 mumol/L), and supernatants were extracted with chloroform-methanol. Platelet-aggregating activity was found in the chloroform phase in 2 of 9, 1 of 8, 0 of 9, and 3 of 9 studies involving iris, retina, ciliary body, or cornea, respectively. Twenty-four hours after the intravitreal injection of 125 ng of endotoxin, aggregating activity was consistently detectable from supernatants of stimulated iris and ciliary body, occasionally present from stimulated retina but not detectable from cornea. The shape of the aggregation curve resembled that produced by 0.5 to 2.0 ng of authentic PAF. Moreover, the aggregation could be completely inhibited by a PAF receptor antagonist and the aggregating activity chromatographed identically on high-performance liquid chromatography to a PAF standard. These studies indicate that PAF-like activity could be detected from several ocular tissues subsequent to inflammation. Iris, ciliary body, retina, vascular endothelium, and/or leukocytes could each contribute to the presence of this inflammatory mediator.

Retinal vasculitis--a primer

Retinal vasculitis is a diagnosis that is generally suggested by an ophthalmologist. Frequently patients with the disorder are referred to nonophthalmologists for further diagnostic evaluation or treatment. The criteria for defining vasculitis differ greatly between ophthalmologists and other physicians. To facilitate collaboration between ophthalmologists and their colleagues, we have sought to clarify the term "retinal vasculitis" by discussing its subcategories, the potential role of antiphospholipid antibodies, and the etiology of retinal vasculitis. We offer guidelines for evaluating the disorder and treating patients.

Systemic sulfonamides as a cause of bilateral, anterior uveitis

Between September 1976 and May 1989, 12 cases of uveitis attributed to the systemic use of sulfonamide derivatives were reported to the National Registry of Drug-Induced Ocular Side Effects and the US Food and Drug Administration. We evaluated these reports in addition to one case previously reported in the literature and one patient seen at the Uveitis Clinic, Oregon Health Sciences University, Portland. The patients' median age was 34 years. Twelve of 14 patients were treated with trimethoprim-sulfamethoxazole. All patients for whom the location of the eye disease was specified presented with an iritis. Six reports included a description of ocular symmetry, with all patients having bilateral inflammation. Of the nine patients for whom data on the duration of drug use was available, seven experienced adverse effects within 8 days of beginning trimethoprim-sulfamethoxazole therapy and four showed effects within 24 hours. Three patients had histories of rechallenge with trimethoprim-sulfamethoxazole, and in each case acute iritis recurred within 24 hours of reinstitution of therapy. Five patients had additional evidence of an adverse reaction manifested as Stevens-Johnson syndrome, erythema multiforme, diffuse macular or vesicular rashes, stomatitis, glossitis, conjunctival and scleral injection, and granulomatous hepatitis. The consistent presentation including bilateral, anterior inflammation and the recurrence with rechallenge strongly indicate a cause-effect relationship. Although uveitis secondary to sulfonamides is a rarely diagnosed clinical event, recognition of the distinct presentation of this entity is important in the differential diagnosis of uveitis.