Author Affiliations: Departments of Ophthalmology (Ms Ku and Drs Ali, Suhler, and Rosenbaum), Medicine (Dr Rosenbaum), and Public Health (Ms Ku and Drs Suhler and Choi), Oregon Health & Science University, and Portland Veterans Affairs Medical Center (Dr Suhler), Portland, Oregon.
Objective To examine the characteristics and visual outcome in 207 patients with retinal vasculitis.
Methods Demographic and visual outcome data were collected retrospectively from the ophthalmologic records of 207 cases (321 affected eyes). Descriptive analysis was performed on all cases and visual outcome analysis was performed for the 114 cases with visual acuity recorded at 2 or more visits. The Kaplan-Meier method and Cox regression were used to examine visual outcome and predictors for prognosis.
Results Patients in our series had a roughly even distribution of sex, were predominantly non-Hispanic white (77.8%), and had bilateral disease (75.7%). The annualized mean visual acuity change for the 203 eyes (114 patients) with some follow-up was 0.01 logMAR unit per year. Although 75 eyes (36.9%) had 20/25 or better visual acuity at baseline, 33.6% of the remaining eyes experienced visual acuity improvement of at least 2 lines on the Snellen chart during follow-up and some continued to improve more than 9 years after the initial evaluation. Cox multivariate analysis demonstrated that patients who were nonwhite, had worse visual acuity at baseline, or who had an ocular infection were more likely to experience improvement by this definition.
Conclusions We believe that this is the first US case series to investigate visual outcome in patients with this diagnosis. Although many patients in our series worsened despite therapy, a subset experienced substantial improvement.
Retinal vasculitis is a poorly understood, potentially sight-threatening inflammatory eye condition characterized by an abnormal appearance of the retinal vasculature.1 The definition of retinal vasculitis varies and remains controversial. The Standardization of Uveitis Nomenclature Working Group reached a “provisional” consensus in November 2004 that retinal vasculitis is a descriptive term for evidence of ocular inflammation and retinal vascular changes and
agreed to consider perivascular sheathing and vascular leakage or occlusion on fluorescein angiogram as evidence of retinal vascular disease for the classification of retinal vasculitis.2(p511)
Retinal vasculitis may occur as an isolated idiopathic condition; as a manifestation of infectious disease such as tuberculosis, Lyme disease, syphilis, toxoplasmosis, or acute retinal necrosis; as part of neurologic disorders such as multiple sclerosis; or in association with a systemic immune-mediated disease.3 Autoimmune processes are thought to be responsible when retinal vasculitis occurs without systemic manifestations, but the exact etiology remains unknown.4 The annual incidence of retinal vasculitis, including the isolated form and with an associated systemic disease, in the United States has been estimated as 1 to 2 per 100 000, but significant regional variation exists.5
The usual clinical manifestations of retinal vasculitis demonstrable by ophthalmologic examinations or fluorescein angiography include perivascular sheathing, vascular leakage, and occlusion in association with intraocular inflammation such as intermediate or posterior uveitis.6,7 The most common symptoms include blurred or decreased vision, floaters, and scotomata, although the disease may be asymptomatic. Causes of poor outcome of retinal vasculitis are multifactorial and prediction of visual acuity is difficult as the course of the disease may vary.8 Although a good prognosis can be expected with adequate treatment with immunosuppressive therapy, poor visual outcome despite therapy often is associated with complications such as macular ischemia, branch retinal vein occlusion, central retinal vein occlusion, branchretinal artery occlusions, persistent neovascularization, vitreous hemorrhage, and tractional retinal detachment.3,7,9
Even though about 1 in every 8 patients with uveitis has an associated retinal vasculitis,10 there are only a limited number of publications related to this disease and well-controlled clinical trials have not been conducted.8 Three studies from London, England, have examined ophthalmological and clinical features of patients with retinal vasculitis.11- 13 One study investigated clinical and immunological findings in patients with retinal vasculitis but excluded cases secondary to an ocular infection, and 2 studies examined ophthalmological/immunological features and visual outcome of idiopathic retinal vasculitis.11- 13 The most recent of these studies was published in 1996. Current treatment guidelines are based on limited studies, and the evaluation and treatment of retinal vasculitis are thus often challenging.8 To our knowledge, no US population-based studies have examined visual outcome of patients with this disease.
To clarify the relationship between retinal vasculitis and systemic vasculitis, we recently reported on 207 patients with retinal vasculitis attending the uveitis clinic at Oregon Health & Science University between 1985 and 2010.10 In the present study, demographic and visual outcome data from the identified 207 cases were analyzed to examine the characteristics of the patient population, visual outcome, and predictors for prognosis.
We obtained ophthalmological records of 207 patients with retinal vasculitis identified from our previous review of 1390 patients with uveitis.10 Cases of retinal vasculitis were identified by active inflammation in the anterior chamber of the eye and/or the vitreous humor along with perivascular exudates, intraretinal hemorrhage, or cotton-wool spots documented on clinical examination or by vascular occlusion or leakage as identified by fluorescein angiogram at any time while attending the clinic.10
Study data were collected retrospectively by medical record review. No minimum follow-up time was required and if there were multiple visits, all available records were reviewed. Sex, race, laterality of the disease, age at initial consultation, and age at disease onset were recorded. Disease onset was defined as timing of initial clinical symptoms as documented by history at the first consultation. Best-corrected Snellen visual acuity with pinhole improvement for all affected eyes was recorded; visual acuity at the initial visit was recorded for all 207 patients and visual acuity at the final visit was recorded for 114 patients who had 2 or more visits. The duration between the initial evaluation and the final visit was recorded as follow-up time. Visual acuity of counting fingers and hand motion were recorded as 20/2000 and 20/20 000, respectively.14 Because the lines on the Snellen visual acuity chart follow a geometric progression, the logMAR notation was used to compute visual acuity change and average.14 Data on treatment used during follow-up were collected for 99 patients who had at least 1 follow-up visit beyond the initial consultation and did not have an ocular infection; records on treatment were unavailable for 3 patients.
All statistical analysis was performed using Stata version 11.0 (StataCorp). Frequencies of variables collected on the 207 cases were tabulated for descriptive analysis. Analysis of visual outcome was performed on 114 patients (203 eyes) who had visual acuity measured at 2 or more different times. Excluded from the analysis of visual outcome were 22 patients who had unknown laterality (ie, 1 eye or both eyes) or visual acuity of light perception with and without projection because these are simply the detection of a stimulus, not actual visual acuity measurements.14 Average visual acuity change per year for the 203 eyes with varying lengths of follow-up was computed by summing the change in logMAR visual acuity from the initial consultation to the final visit and dividing by the sum of follow-up time across all individual eyes to avoid ceiling and floor effects of visual acuity change. Because of this computation method, conventional parametric tests such as the t test could not be used for estimation of the mean and standard deviation for the annualized average visual acuity change and P value for difference in each of the 2-group comparisons. Instead, we used the bootstrapping method, which is a computer-intensive resampling method that uses the sampling distribution of statistics of interest from multiple resampled data sets by random sampling with replacement from the original samples.15
Because the incidence of visual acuity change varied over time, the proportions of visual acuity improvement were computed using the Kaplan-Meier method to illustrate the cumulative hazard function of visual acuity improvement as a time-to-event outcome. An event was defined as improvement in visual acuity of at least 2 lines on the Snellen chart. In this analysis, each affected eye was treated as 1 independent individual. The Cox proportional hazards model was used to examine the relationship between time to 2-line improvement in visual acuity and other covariates while adjusting for potentially confounding variables such as demographic characteristics. Because each affected eye was treated as an independent individual, the Cox proportional hazards model with shared frailty was performed to account for potential within-individual correlation.
This study was reviewed and approved by the institutional review board at the Oregon Health & Science University.
As recently reported,10 our cohort of patients with retinal vasculitis included 35 with primary retinal vasculitis. The other relatively common diagnoses in this cohort included pars planitis (n = 36), idiopathic uveitis (n = 18), Behçet disease (n = 14), sarcoidosis (n = 13), and birdshot choroidopathy (n = 9). Twenty-nine patients in this series had an ocular infection including 7 with acute retinal necrosis, 5 with AIDS/cytomegalovirus, 1 with endophthalmitis, 4 with herpes simplex or zoster keratitis, 3 with syphilis, 7 with toxoplasmosis, and 2 with tuberculosis. Patients in our series consisted of 91 males (46.2%) and 106 females (53.8%) and were mostly non-Hispanic white (144; 77.8%) (Table 1); the sex of 10 patients was unknown because of missing records and race was unknown for 22 patients. One hundred forty patients (75.7%) had both eyes affected and 45 (24.3%) had only 1 eye affected; laterality of disease was unknown for 22 patients. The mean (SD) age at initial consultation was 39.3 (19.3) years (range, 3.0 to 79.0 years) and the mean (SD) age at onset was 36.6 (19.5) years (range, 3.0 to 78.0 years).
The median duration between disease onset and initial consultation was 32 months (range, 0 to 324 months) and median follow-up time was 4 months (range, 0 to 207 months). The mean logMAR visual acuity at initial consultation for 321 affected eyes (visual acuity unavailable for 4 eyes) was 0.17 (range, −0.12 to 3.00; 20/30 Snellen equivalent). The mean logMAR visual acuity at initial consultation of 118 eyes evaluated only once was also 0.17 (range, −0.12 to 3.00; 20/30 Snellen equivalent) and that of the remaining 203 eyes with 2 or more visits was 0.18 (range, −0.12 to 3.00; 20/30 Snellen equivalent). The mean initial logMAR visual acuity was 0.34 (range, −0.12 to 3.00; 20/44 Snellen equivalent) in non-Hispanic white patients, 0.61 (range, 0.00 to 3.00; 20/81 Snellen equivalent) in other races, 0.38 (range, 0.00 to 1.30; 20/48 Snellen equivalent) in those with an ocular infection, and 0.41 (range, −0.12 to 3.0; 20/51 Snellen equivalent) in those without an ocular infection. Because all patients had varying lengths of follow-up, visual acuity change from the initial to the final visit as a function of follow-up time was examined to look for any trend in visual acuity change dependent on the length of follow-up time; patients with longer follow-up tended to have greater change in visual acuity.
Of the 207 patients, 99 had some follow-up and did not have an ocular infection. Of these 99 patients, 89 (89.9%) received 1 or more treatments, which included topical/oral corticosteroids, disease-modifying antirheumatic drugs, biologic disease-modifying antirheumatic drugs, alkylators, nonsteroidal anti-inflammatory drugs, periocular/intravitreal triamcinolone injections, and surgeries (Figure 1). The remaining 10 patients (10.1%) did not require treatment. Of the 99 patients, 26 had visual acuity improvement of at least 2 lines on the Snellen chart. Eighteen (69.2%) of these patients received 1 or more treatments and the remaining 8 (30.8%) did not require treatment.
Figure 1. Percentage treated and treatment used in 99 patients who had 1 or more visits and did not have ocular infection. *Includes vitrectomy, phacoemulsification, lens implantation, lensectomy, fluocinolone implant, membranectomy, air-fluid exchange, and trabeculectomy. DMARDs indicates disease-modifying antirheumatic drugs; NSAIDs, nonsteroidal anti-inflammatory drugs; and POKI/IVKI, periocular/intravitreal triamcinolone injections.
The annualized mean (SD) change in visual acuity for 203 eyes with 2 or more visits was 0.01 (0.01) logMAR unit per year, corresponding to a decrease in visual acuity of 0.1 line on the Snellen chart (Table 2). Males experienced more worsening than females; the mean change in logMAR visual acuity was 0.02 in males while females experienced no change (P value by the bootstrap for difference = .37). Visual acuity in the non-Hispanic white group decreased by 0.02 logMAR unit per year while visual acuity in patients of other races improved by 0.04 unit on average (P value by the bootstrap for difference = .14). Patients with primary retinal vasculitis experienced a bigger decrease in visual acuity (0.06 log MAR unit) compared with those with retinal vasculitis with an associated disease (no change in visual acuity) (P value by the bootstrap for difference = .04). Those with retinal vasculitis secondary to an ocular infection experienced improvement in visual acuity of 0.05 logMAR unit per year, while visual acuity in those without an infection decreased by 0.01 unit (P value by the bootstrap for difference = .35). The annualized mean change in visual acuity was 0.05 logMAR unit for those with unilateral disease while those with bilateral disease experienced no change (P value by the bootstrap for difference = .20). Forty-seven eyes (23.2%) did not experience worsening or improvement.
Figures 2, 3, and 4 depict the cumulative proportion of eyes achieving visual acuity improvement of 2 or more lines on the Snellen chart as a function of follow-up time. Tick marks represent cases lost to follow-up and dotted lines represent 95% confidence intervals for survival proportions. A Kaplan-Meier event was defined as improvement in visual acuity of at least 2 lines on the Snellen chart. Seventy-five eyes (36.9%) had visual acuity of 20/25 or better at baseline and were unlikely to improve by 2 lines. A total of 43 among the 203 affected eyes experienced improvement overall by this definition. Among those with initial visual acuity worse than 20/25, 35.6% of eyes experienced a 2-line improvement. Improvement was achieved for 15 eyes in the first year of follow-up (8.3%; 54 lost to follow-up), 33 eyes by 5 years (26.1%; 115 lost to follow-up), and 76 eyes by 10 years (53.8%; 151 lost to follow-up). In the non-Hispanic white group, 10 eyes in the first year of follow-up (7.4%; 42 lost to follow-up), 17 eyes by 5 years (16.7%; 88 lost to follow-up), and 22 eyes by 5 years (34.2%; 118 lost to follow-up) experienced improvement. However, in individuals of races other than non-Hispanic white, visual acuity improved in 5 eyes in the first year (10.4%; 12 lost to follow-up) and 16 eyes by 5 years (50.5%; 27 lost to follow-up). In the eyes with retinal vasculitis secondary to an ocular infection, visual acuity improved in 4 eyes in the first year (23.4%; 6 lost to follow-up) and 8 eyes by the fifth year (74.5%; 10 lost to follow-up). Among those without an ocular infection, visual acuity improved in 11 eyes in the first year (6.6%; 33 lost to follow-up), 25 eyes by 5 years (21.6%; 105 lost to follow-up), and 36 eyes by 10 years (51.5%; 139 lost to follow-up).
Figure 2. Kaplan-Meier curve demonstrating the cumulative proportion of eyes that experienced visual acuity improvement of 2 or more lines on the Snellen chart as a function of follow-up time in 203 eyes in 114 patients with retinal vasculitis who had visual acuity recorded at 2 or more points. Tick marks represent cases lost to follow-up and dotted lines represent 95% confidence intervals for survival proportions.
Figure 3. Kaplan-Meier curve demonstrating the cumulative proportion of eyes that experienced visual acuity improvement of 2 or more lines on the Snellen chart as a function of follow-up time by race groups. “White” denotes 150 eyes with retinal vasculitis in non-Hispanic white patients and “other races” denotes 43 eyes in patients with retinal vasculitis who are either black, Asian, Hispanic, or multiracial. Ten eyes from patients of unknown race were excluded. Tick marks represent cases lost to follow-up and dotted lines represent 95% confidence intervals for survival proportions.
Figure 4. Kaplan-Meier curve demonstrating the cumulative proportion of eyes that experienced visual acuity improvement of 2 or more lines on the Snellen chart as a function of follow-up time by infectious vs noninfectious cause. “Infectious” denotes 20 eyes in patients with retinal vasculitis secondary to an ocular infection and “noninfectious” denotes 183 eyes with retinal vasculitis without an infectious cause. Tick marks represent cases lost to follow-up and dotted lines represent 95% confidence intervals for survival proportions.
Multivariate Cox regression analysis was performed, evaluating whether demographic or clinical characteristics could be identified as factors predictive of visual outcome (Table 3). A multivariate Cox proportional hazard ratio model with shared frailty was used to account for the within-individual correlation (ie, correlation between the eyes within the same individual). Compared with non-Hispanic white patients, patients of other races were more likely (P = .04) to experience improvement in visual acuity (hazard ratio [HR] for visual acuity improvement adjusted for within-individual correlation, 3.54; 95% CI, 1.08-11.61). Those with worse visual acuity (greater logMAR value) at the initial consultation were more likely (P < .01) to achieve visual acuity improvement than those with better visual acuity (HR, 6.68 per 1 logMAR unit; 95% CI, 3.16-14.14) (eg, an eye with 20/200 visual acuity at baseline was 6.68 times more likely to have improved visual acuity of 2 lines compared with an eye with 20/20 visual acuity). A ceiling effect likely exists because 75 eyes (36.9%) had visual acuity of 20/25 or better at baseline and thus were unlikely to improve by 2 lines. Those with retinal vasculitis secondary to ocular infection were more likely (P < .01) to have improved visual acuity (HR, 11.72; 95% CI, 2.45-56.10) than those without an infection. However, sex, duration between disease onset and baseline examination, age at disease onset, or unilateral as opposed to bilateral disease were not significantly associated with visual outcome. Univariate analysis was suggestive of visual acuity improvement occurring less (P = .04) in cases with primary retinal vasculitis than those with an associated disease (HR, 7.78; 95% CI, 1.07-56.63), but this was no longer significant in the model adjusted for within-individual correlation.
This study of 207 patients with retinal vasculitis was undertaken to examine characteristics of the patient population, visual outcome, and factors predictive of outcome. A limited number of previously published studies have examined characteristics and ophthalmological features of patients with retinal vasculitis. A point-prevalence study from London described ophthalmological features of 150 patients with this diagnosis.11 The study demonstrated different patterns of retinal vasculitis occurring in different systemic inflammatory diseases and in isolated retinal vasculitis and poor visual outcome associated with macular edema and branch vein occlusion.11 In a longitudinal study of 52 patients with relapsing retinal vasculitis, these investigators observed visual improvement in 12 of 26 patients with retinal vasculitis alone and 10 of 26 patients with retinal vasculitis associated with a systemic inflammatory disease.12 The same group examined visual outcome of 53 patients with idiopathic retinal vasculitis and observed significantly worse visual outcome in ischemic retinal vasculitis compared with nonischemic retinal vasculitis.13 Sex distribution, time between disease onset and referral, and age at consultation of the patients in our series were similar to what was described in the literature and to the demographics previously reported from 236 patients with uveitis seen in our clinic.16 To our knowledge, no studies have examined the visual outcome of retinal vasculitis with a known cause (ie, systemic inflammatory disease and ocular infection) as well as in the idiopathic and isolated form. Also, no studies have attempted to examine potential clinical and demographic predictors of visual outcome.
Several novel findings are apparent from our study. While 39.4% of the eyes experienced loss of vision despite therapy, 37.4% experienced some improvement. More than 1 of 3 eyes in our series had visual acuity of 20/25 or better at baseline and were thus unlikely to experience substantial improvement, but visual acuity in 35.6% of the remaining eyes improved by at least 2 lines on the Snellen chart. In no case was improvement due to surgery (ie, cataract surgery). Cox multivariate analysis showed more likelihood to improve for patients who were nonwhite, had an ocular infection, or had worse vision at baseline. However, a ceiling effect may exist because patients with an ocular infection and patients of races other than non-Hispanic white had worse vision at baseline, potentially allowing their vision to improve to a greater degree. Direct comparisons of findings on visual outcome with previously published series are difficult because the prior studies often excluded cases with a known cause such as systemic immune-mediated disease and ocular infection.
A problem inherent in studying retinal vasculitis is the heterogeneity of the disease. The prognosis, for example, of retinal vasculitis associated with sarcoidosis would presumably differ from a retinal vasculitis associated with birdshot choroidopathy. This heterogeneity is especially relevant to the infections associated with retinal vasculitis. Although tuberculosis, syphilis, toxoplasmosis, and herpes family viruses are each commonly associated with retinal vasculitis, the course for each is distinct. Accordingly, one should be extremely cautious in drawing any conclusions about vasculitis secondary to infection in this series because no single infection was associated with an adequate number of subjects to allow definitive conclusions. The heterogeneity of retinal vasculitis also applies to primary retinal vasculitis. As diagnostic sophistication improves, we will undoubtedly be able to define subsets of primary retinal vasculitis and the prognosis for each subset will be more meaningful than trying to determine prognosis for the aggregate group.
There are several limitations to the retrospective approach in this study; accuracy and consistency of collected data were less than could have been achieved in a prospective study. Our findings may also be limited in terms of generalizability because geography affects disease prevalence and subtype, and our series largely represents a patient population in the northwestern United States. For example, Behçet disease, a prominent cause of retinal vasculitis in some locations, was relatively rare in our series. Because the study population was selected from a tertiary referral center, the cases may have been more severe than would have been observed in primary clinics.
We believe that this is the first case series to examine visual outcome of retinal vasculitis in a population from North America. In our earlier published work, we analyzed a variety of subsets of uveitis for the relationship of each to retinal vasculitis. About 1 of 7 patients in our series had an identifiable infectious cause and systemic vasculitides counted for only 1.4%.10 Retinal vasculitis can be divided into subsets based on the involvement of arteries or veins, and in some cases, a specific diagnosis is suggested when the vascular involvement is substantially characteristic.10 Identifying subsets of retinal vasculitis and the prognosis of each subset will add a wealth of information about this disease and will help find optimal treatment options. Further work on the cases in our series to investigate this question is forthcoming.
Correspondence: Jennifer H. Ku, MPH, Casey Eye Institute, Oregon Health & Science University, 3181 Sam Jackson Park Rd SW, Portland, OR 97239 (email@example.com).
Submitted for Publication: November 20, 2011; final revision received April 9, 2012; accepted April 23, 2012.
Published Online: June 11, 2012. doi:10.1001 /archophthalmol.2012.1596
Author Contributions: Drs Choi and Rosenbaum share senior authorship.
Financial Disclosure: None reported.
Funding/Support: This work was supported by Research to Prevent Blindness, a core grant from the National Eye Institute, and funds from the Stan and Madelle Rosenfeld Family Trust, the William and Mary Bauman Foundation, and the William C. Kuzell Foundation.
Online-Only Material: This article is featured in the Archives Journal Club. Go to here to download teaching PowerPoint slides
Thank you for submitting a comment on this article. It will be reviewed by JAMA Ophthalmology editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Sohn EH, He S, Kim LA, et al. Angiofibrotic response to vascular endothelial growth factor inhibition in diabetic retinal detachment. Arch Ophthalmol. 2012;130(9):1127-1134.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 1
Customize your page view by dragging & repositioning the boxes below.
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.