Article Text

Download PDFPDF

The course of retinal vasculitis
  1. Amro Ali1,
  2. Jennifer H Ku1,2,
  3. Eric B Suhler1,2,3,
  4. Dongseok Choi2,
  5. James T Rosenbaum1,4,5
  1. 1Department of Ophthalmology, Oregon Health & Science University, Portland, Oregon, USA
  2. 2Department of Public Health and Preventive Medicine, Oregon Health & Science University, Portland, Oregon, USA
  3. 3Portland Veterans Affairs Medical Center, Portland, Oregon, USA
  4. 4Department of Medicine, Oregon Health & Science University, Portland, Oregon, USA
  5. 5Devers Eye Institute, Legacy Health Systems, Portland, Oregon, USA
  1. Correspondence to Jennifer Ku, Department of Ophthalmology, Oregon Health & Science University, Mailcode: CEI, 3375 Terwilliger blvd, Portland, OR 97239, USA; kuj{at}


Aims To determine if characteristics of retinal vasculitis correlate with ocular complications, or the response to different lines of treatment.

Materials and methods We performed a computerised database analysis of 56 patients evaluated for uveitis at the Casey Eye Institute from September 1985 until May 2010. All patients had non-infectious retinal vasculitis and at least 1 year of follow-up.

Results Although occlusive vasculitis was rare, retinal neovascularisation occurred much more commonly in the occlusive vasculitis subgroup than among the non-occlusive vasculitis subgroup (p<0.01). Epiretinal membrane (ERM) was found more commonly in the retinal vasculitis patients who presented with cotton wool spots and intraretinal haemorrhage compared to retinal vasculitis patients who presented with sheathing noted on clinical examination (p<0.01). Smoking was significantly related to vision loss. Age at presentation below 40 years correlated with therapy beyond oral corticosteroids

Conclusions The heterogeneity of retinal vasculitis should be considered in providing prognostic information. Neovascularisation occurs more commonly in occlusive retinal vasculitis, and ERM is diagnosed more frequently in conjunction with cotton wool spots and intraretinal haemorrhage rather than just vascular sheathing. Cigarette use predicts visual loss and patients who are relatively young often receive treatment beyond oral corticosteroids.

  • Epidemiology
  • Retina
  • Treatment Medical
  • Vision

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The concept of retinal blood vessel inflammation was first introduced in 1784, when John Hunter published a paper on ‘Observations on the Inflammation of the Internal Coats of the Veins’.1 Retinal vasculitis represents a group of diseases characterised by inflammation affecting the retinal vasculature usually in the context of evidence for inflammation elsewhere within the eye.2 Retinal vasculitis may also occur in the presence of a variety of entities, such as immune-mediated conditions, infection, or malignancy. It can occur in the context of a systemic inflammatory disease or as an isolated idiopathic condition.3–5 Clearly, retinal vasculitis is not a single entity. Understanding the prognosis of retinal vasculitis could help in recommending appropriate therapy, but information on the outcome of retinal vasculitis is limited and the heterogeneity within this diagnosis further confounds our understanding.

Ironically, despite the uncertainties in knowledge about retinal vasculitis, nowhere else in the body can microcirculation be visualised with such precision as in retinal circulation. The vessels are distributed on the flat surface of the retina, allowing for accurate photographic documentation and clear differentiation to be made between arteries and veins. Ophthalmologists routinely diagnose retinal vasculitis based on the presence of pathognomonic clinical findings, such as vascular sheathing, cotton wools spots and intraretinal haemorrhage.

Some patients with retinal vasculitis may have vision-threatening conditions that might respond to immunosuppressants, but many patients with retinal vasculitis have a more benign process that can be treated with local corticosteroids or even by observation. Retinal vasculitis results in vascular leakage on fluorescein angiogram, and it may lead to retinal oedema, exudation and cystoid macular oedema, which is a significant contributing factor for vision loss in retinal vasculitis. When adequately treated with immunosuppressive therapy, cystoid macular oedema is associated with good prognosis.6 Late changes secondary to vascular occlusion and remodelling include telangiectasias, microaneurysms, and ischemia-induced neovascularisation, with sequelae, such as recurrent vitreous haemorrhage, traction retinal detachment, rubeosis iridis, and neovascular glaucoma, all of which can lead to loss of vision.7–14

We recently analysed a large cohort of patients with retinal vasculitis who had been seen at the Casey Eye Institute, Oregon Health & Science University.15 ,16 We found that many patients maintained excellent visual acuity, while others did less well. In this report, we analysed a subset of these patients for whom we had at least 1 year of follow-up information. We sought to determine if specific clinical signs, such as vascular sheathing, could aid in the prediction of prognosis.

Patients and methods


We have recently reported on 207 patients with retinal vasculitis who were evaluated at the Oregon Health & Science University, Casey Eye Institute, Portland, between the years 1985 and 2010.15 ,16 This report is based on a subset of that group, 56 patients with 101 affected eyes (45 patients were affected bilaterally and 11 patients were affected unilaterally). We excluded 151 patients for various reasons, such as ocular infection, ocular malignancy, associated ocular conditions that may affect visual outcome, such as diabetic retinopathy and macular degeneration, and short duration of follow-up (less than 1 year). Retinal vasculitis was recognised by the presence of different clinical findings of vasculitis, such as perivascular sheathing, intraretinal haemorrhage, or cotton wool spots documented on clinical ocular examination. In 39 eyes, a fluorescein angiogram was also used to document the presence of retinal vasculitis, such as vascular occlusion or leakage. All patients with retinal vasculitis had accompanying evidence for intraocular inflammation, such as leukocytes in the anterior chamber or vitreous humour.

Age, gender, race, laterality, clinical findings, ocular complications, and follow-up time were documented. Age at initial consultation and age at onset of disease were recorded. One-year minimum of follow-up in our clinic was required for inclusion in this study. Best-corrected Snellen's visual acuity of all affected eyes was recorded, as was vision at initial visit, one-year visit follow-up, and at the final visit. Vision of counting fingers and hand motion were recorded as 20/2000 and 20/20 000, respectively. As the lines on the Snellen visual acuity chart follow a geometric progression, Log Mar (log of the minimum angle of resolution) notation was used to compute the change in visual acuity. Complete data about different lines of treatment during follow-up were collected for the 56 patients.

Statistical analysis

Statistical analyses were performed using Stata V.11.0 (Stata, College Station, Texas, USA). Frequencies of variables collected on the 56 cases were tabulated for descriptive analysis. Linear regression model was employed to examine the association between change in Log Mar vision and demographic/clinical variables, such as gender, race and angiographic findings in the first year of follow-up, and from baseline evaluation to final visit. Likewise, logistic regression model was used to illustrate the demographic and clinical predictors of ocular complications and the use of steroid-sparring therapy. Complications examined in the modelling included cataract, glaucoma, cystoid macular oedema, epiretinal membrane and neovascularisation. Oral corticosteroid-sparring therapy included biologic/non-biologic disease modifying antirheumatic drugs and periocular/intravitreal triamcinolone injections.

Using Fisher's exact test, we compared different subsets of retinal vasculitis: A) occlusive versus non-occlusive retinal vasculitis and B) retinal vasculitis presenting with vascular sheathing, versus retinal vasculitis presenting with sheathing, cotton wool spot and intraretinal haemorrhage based on correlation with ocular complications.

This study was reviewed and approved by the Institutional Review Board at the Oregon Health & Science University.



We analysed 56 patients with 101 affected eyes. Of these 56 patients, 25 (44.6%) were male and 31 (55.4%) were female. Mean age at onset of presentation was 31.9 years (range 3–70). The mean age at first visit was 32.2 years and the mean duration of follow-up was just over 60 months. Specific diagnoses included pars planitis 17 (30.4% of entire group), Behcet's disease 9 (16.1%), birdshot choroidopathy 8 (14.3%), primary retinal vasculitis: 7 (12.5%), sarcoidosis 4 (7.1%), Vogt–Koyanagi–Harada syndrome 4 (7.1%), juvenile idiopathic arthritis 1 (1.8%), idiopathic uveitis 3 (5.4%), multiple sclerosis 1 (1.8%), psoriatic arthritis 1 (1.8%), and ulcerative colitis 1 (1.8%). Nine patients (16.1%) were smokers, 36 (64.3%) were non-smokers, and the remaining 11 (19.6%) patients had no recorded information about their smoking history. History of alcohol use was positive in 8 patients (14.3%), negative in 40 (71.4%), and the remaining 8 (14.3%) had no recorded information about their history of alcohol use. Race distribution in our study was mostly non-Hispanic Caucasian (40, 71.4%) followed by Hispanic (6, 10.7%), Asian (4, 7.1%), African–American (2, 3.6%), multicultural (2, 3.6%) and unknown (2, 3.6%). Overall, there were no statistically significant differences found with respect to age, sex, race, or alcohol use with regard to an impact on the course of retinal vasculitis or the response to different lines of treatment.

Clinical presentation

Retinal vasculitis presented with vascular sheathing, cotton wool spots, and intraretinal haemorrhage. Vascular sheathing occurred without cotton wool spots or intraretinal haemorrhage in 86 eyes (85.2%); vascular sheathing and intraretinal haemorrhage were found in 11 eyes (10.9%); vascular sheathing and cotton wool spots were found in 4 eyes (3.9%); (table 1). Vasculitis occurred only with arterial involvement in 2 eyes (2.0%), venous involvement occurred in 44 (43.6%), mixed involvement occurred in 13 eyes (12.9%), and the remaining 42 eyes (41.6%) were not specified according to the chart review. Vascular occlusion, defined as complete or near-complete blockage of a vessel as was documented with fluorescein angiography, occurred in 6 eyes (5.9%) in comparison with 95 (94.1%) eyes without occlusion. Retinal vasculitis occurred bilaterally in 45 patients (80.4%) and unilaterally in 11 (19.6%) patients (a total of 101 eyes).

Table 1

Using the Fisher's exact test to compare ocular complications between vasculitis presenting with intraretinal haemorrhage and cotton wool spots versus vasculitis presenting with only sheathing

Using the simple linear regression model, Log Mar changes from the baseline visit to the final visit, and from the baseline visit to one-year follow-up were evaluated. Log Mar changes were not related to the following factors: age, sex, race, age of onset, laterality, clinical presentation, obtaining of fluorescein angiography, occlusive vasculitis, and the possible use of an immunomodulator, except for smoking, which was significantly associated with worsening vision, based on Log Mar changes from baseline to final visit (p<0.01) (table 2). By contrast, we could not show a statistically significant effect for any of these variables, including smoking, for an evaluation based on acuity at 1 year.

Table 2

Simple Linear regression model for Log-Mar change for assessment of visual acuity (VA) changes from baseline to final visit and its correlation to many other variables

We studied the correlation between the decision to use a steroid sparing agent (SSA) and different factors, such as age, sex, race, age of vasculitis onset, time between onset of disease and the time of the baseline evaluation, laterality, clinical presentation, alcohol use and smoking. All these had an insignificant correlation except for age. Age at onset of presentation below 40 years is associated with use of therapy beyond steroids. (OR 3.74, 95% CI 1.29 to 12.52, p=0.01) (table 3). Also, we assessed the correlation between the presence of ocular complications and other variables such as smoking, alcohol, race, sex, use of fluorescein angiography, occlusive vasculitis and we found that the presence of ocular complications was significantly associated with the use of SSA (OR 4.59, 95% CI 1.78 to 12.34, p<0.01) (table 4).

Table 3

Univariate exact logistic regression model for therapy used beyond steroids, that is, steroid sparing agents (SSAs)

Table 4

Univariate exact logistic regression model for assessment of ocular complications and its correlation to many other variables

Neovascularisation occurred in 7 eyes (7.4%) in the non-occlusive group compared to 6 eyes (100%) in the occlusive group (p<0.01 for group difference, the Fisher's exact test). ERM, noted in 14 eyes (table 5), or 16.3% of the group which presented with vascular sheathing compared to 9 eyes (60.0%) of the group that presented with sheathing, cotton wool spots and the intraretinal haemorrhage (p<0.01 for group difference, the Fisher's exact test) (table 1).

Table 5

Using the Fisher's exact test to compare ocular complications between occlusive vasculitis versus non-occlusive vasculitis

Thirty-eight of 56 patients (67.9%) were treated with systemic corticosteroid (oral prednisone), with an average dose over time of 27 mg (range 100–20) and average duration of 13.9 months (range 1–78), 28 (27.7%) eyes were treated with periocular triamcinolone injections, and 4 (4.0%) eyes were treated with intravitreal triamcinolone injections.

Patients with retinal vasculitis were treated with different immunomodulator agents or SSAs, such as methotrexate (MTX) (n=22), mycophenolate (n=12), infliximab (n=8), cyclosporine (n=7), azathioprine (n=7), tacrolimus (n=2), adalimumab (n=2), cyclophosphamide (n=1), sulfasalazine (n=1). mycophenolate+adalimumab (n=1), mycophenolate+cyclosporine (n=2), cyclosporine+MTX (n=2), infliximab+MTX (n=2).


Retinal vasculitis typically occurs in association with intraocular inflammation as in sarcoidosis or as an isolated, idiopathic entity, but surprisingly, it rarely occurs in association with systemic forms of vasculitis.15 Patients with idiopathic retinal vasculitis are typically young adults with no signs or symptoms suggestive of an underlying systemic or ocular disease.17 The most common symptoms reported by patients with retinal vasculitis are blurred vision, flashes, floaters and scotoma.3 ,18 Less common symptoms may include alteration of colour vision, metamorphopsia, and, rarely, pain.3 ,19 ,20 On fundus examination, retinal vasculitis usually reveals the presence of perivascular sheathing, intraretinal haemorrhage, cotton wool spots, or vascular occlusion.

In this study, we investigated 56 patients with an established diagnosis of retinal vasculitis. In our study, we found no gender predilection, although female sex steroid hormones are significantly related to certain immune-mediated diseases, such as rheumatoid arthritis and systemic lupus erythematosus. The mean age at onset of presentation was 31.9 (range 3–70 years).

Studying the effect of age, sex, race, smoking, or alcohol use on log Mar changes revealed no statistically significant differences from baseline to the visit 1 year later. On the other hand, smoking was significantly related to Log Mar changes from baseline visit to final visit. This is an interesting phenomenon, as smoking is well known to be associated with different forms of vasculopathy. Still, this information is limited due to lack of any quantitative details with regard to use of cigarettes, and the number of smokers in our study is small.

Occlusive vasculitis is more likely to be associated with ocular complications such as epiretinal membrane, cystoid macular oedema and neovascularisation (p<0.01 for group difference, the Fisher's exact test). Neovascularisation was more commonly found with occlusive vasculitis, which would be attributed to more ischaemic changes and increased levels of HIF (hypoxia-inducible factor-1 α) and VEGF (Vascular Endothelial Growth Factor). Steroid-sparing agents (SSA) are more commonly used in retinal vasculitis with ocular complications (OR 4.59, 95% CI 1.78 to 12.34, p<0.01). This presumably relates to the use of SSA for more refractory disease rather than SSAs causing complications.

Appropriate therapy for retinal vasculitis depends, in part, on prognosis. Patients who are likely to have a poor outcome may be candidates for more aggressive therapy or they might be candidates for novel approaches to inflammation or immunosuppression. In the absence of infection, corticosteroids remain the mainstay of treatment for retinal vasculitis.21–23 Oral prednisone was found to be the most common systemic treatment as it was used in 38 of the 56 patients (67.8%).

In this study, we were able to identify several factors that correlate with a worse prognosis: A) smoking was significantly related to worsening of Log Mar changes from baseline visit to final visit; B) age at onset of presentation below 40 years was associated with greater use of SSAs; C) occlusive vasculitis was associated with retinal neovascularisation; D) vasculitis with cotton wool spots and intraretinal haemorrhage is associated with epiretinal membrane formation.

While we do not know that stopping smoking will change the outcome, our observations encourage us to counsel patients with retinal vasculitis on smoking cessation. Smoking has also been associated with a worse outcome in Graves’ ophthalmopathy.24 It exacerbates cystoid macular oedema25 ,26 which indicates that it promotes vascular leakage.

This study has a few limitations, among which are: it is retrospective; a history of smoking or alcohol use was not recorded for all subjects; the number of patients in the sample is small; not all patients had FAs. The more frequent use of fluorescein angiography might have resulted in more frequent diagnosis of retinal vasculitis, especially if wide-field angiography had been employed. Despite these limitations, our observations are novel and indicate that clinically significant subsets of retinal vasculitis can be recognised. This recognition has the potential to lead to more tailored and effective therapy.



  • Contributors All coauthors shared in the responsibility of study creation, data analysis, and manuscript preparation. All coauthors approve of this manuscript and its submission.

  • Funding This work was supported by Research to Prevent Blindness, through a core grant from the National Eye Institute, and by funds from the Stan and Madelle Rosenfeld Family Trust the William and Mary Bauman Foundation.

  • Competing interests None.

  • Ethics approval OHSU Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.