Vision Loss in Uveitis

Uveitis accounts for 10% to 20% of blindness in the United States and Europe, and perhaps as much as 25% in the developing world.^1–4 Vision Loss in patients with uveitis is common and appears to be related most directly to the severity, location and duration of the inflammation, and to the occurrence of secondary, vision-threatening complications such as cataract, macular edema, and glaucoma.^5–14 Prompt and complete control of inflammation is important, therefore, to limit both vision loss and its psychosocial impact. This issue of Ocular Immunology & Inflammation (OII) contains four articles^15–18 related to the occurrence of, risk factors for, and functional outcomes related to vision loss in patients with uveitis.

Oh et al.¹⁵ retrospectively reviewed a systematically stratified, randomly selected sample of 9,036 new-onset uveitis patients collected from approximately one million medical care claims submitted to the Korean National Health Insurance Service (KNHIS) between 2003 and 2013. The authors investigated the incidence and clinical and socioeconomic risk factors for blindness in uveitis. All cases had at least two visits to an ophthalmology clinic and were broadly classified as anterior or non-anterior uveitis. Subjects who developed uveitis within three months of ophthalmic surgery were excluded. For each subject, five control subjects without uveitis, but with baseline macular degeneration, diabetic retinopathy or glaucoma were randomly selected from the same nationwide cohort database over the same study period and were stratified for sex, age, income, region, and other ocular co-morbidities. The authors identified 93 subjects with anterior uveitis (1.0%) and 48 with non-anterior uveitis (0.53%) who met their criteria of blindness in at least one eye, and 13 with anterior uveitis and 7 with non-anterior uveitis who met their blindness criteria bilaterally. The incidence of blindness in those with anterior vs non-anterior uveitis were 2.41 (95% CI 1.94–2.95) and 4.58 (95% CI 3.45–5.92) per 1000 person-years, respectively, whereas the incidence of bilateral blindness in the two subgroups was 0.36 (95% CI 0.20–0.59) and 0.61 (95% CI 0.26–1.17)) per 1000 person-years, respectively. The hazard ratios (HR) for unilateral and bilateral blindness were 7.15 (95% CI 5.57–9.19) and 6.39 (95% CI 3.34–12.20), respectively. Of note, blindness was defined as a best-corrected vision of less than 20/400 as used in South Korea, which was worse than the standard recommended by many countries and the World Health Organization, which use less than or equal to 20/200. The specific causes of blindness in eyes with uveitis as well as whether or not their vision loss was reversible were not specified in the study. Household income at or below the 30^th percentile was a risk factor for unilateral blindness for the entire uveitis cohort (nominal p-value < 0.001; HR 1.50, 95% CI 1.02–1.98)), whereas age over 60 years (nominal p-value < 0.001) and a diagnosis of Behçet’s disease (nominal p-value = 0.005; HR 4.49, 95% CI 1.59–12.71) were risk factors for unilateral blindness in eyes with anterior and non-anterior uveitis, respectively. The incidence of unilateral blindness in subjects with Behçet’s disease was 15.46 (95% CI 4.8–35.91) per 1000 person-years. The authors emphasized that the incidence of uveitic blindness in their population-based cohort was six to seven times higher than blindness observed in control subjects with macular degeneration, diabetic retinopathy or glaucoma, but was materially less than rates reported from uveitis referral clinics.

Pistilli et al.¹⁶ retrospectively evaluated visual acuity outcomes over time in 8,799 study eligible eyes from 5,530 subjects with noninfectious uveitis without HIV infection seen at four participating tertiary referral centers in the United States for up to five years after first presentation between 1978 and 2007 as part of the Systemic Immunosuppressive Therapy for Eye Disease (SITE) study. The median follow up was 183 days (interquartile range 49 to 613 days). The median age of the eligible cohort was 41 years, 65% were female, and 73% were Caucasian. Mean Snellen vision at presentation differed statistically (p < .001) by primary location of the inflammation and was lowest for panuveitis (20/84, 95% CI 20/78-20/90), followed in turn by posterior (20/64, 95% CI 20/60-20/68), intermediate (20/47, 95% CI 20/47-20/49), and anterior uveitis (20/37, 95% CI 20/36-20/38). These anatomical subgroups all showed modest improvement at month 1 to 20/74, 20/56, 20/43, and 20/35, respectively, followed by further modest mean improvement at month 24 to 20/66, 20/50, 20/37, and 20/32, respectively. Between 24 months and five years vision remained stable for eyes with panuveitis (20/59, p < .001), intermediate uveitis (20/38, p = .016), and anterior uveitis (20/31, p = .014), but tended to decline back to baseline for posterior uveitis (20/64, p = .94). In addition to the primary SITE study analysis summarized above, the authors compared the occurrence of 10 letter gains or losses among their retrospectively analyzed eyes to similar vision changes observed in the Multicenter Uveitis Steroid Treatment (MUST) trial, a randomized, prospective study of treatment with corticosteroids plus immunosuppression as indicated vs fluocinolone acetonide implant in 479 eyes with noninfectious uveitis. The Kaplan-Meier curves of 10 letter gain and loss were similar in the two studies, supporting the validity of the retrospective, somewhat less standardized testing situations utilized in the SITE study. Specifically, analysis of both the SITE and MUST studies found that about 50% of eyes either gained, lost, or both gained and lost 10 letters of vision of some point over the four years following presentation. Moreover, in both studies the proportion of eyes that either lost or gained 10 letters at four years approached 20% and 30%, respectively. The authors suggested that the early and sustained vision gains in the SITE study reflected generally improved care provided by tertiary referral centers and suggested that patients with active uveitis might benefit from early referral whenever comparable care is not otherwise available.

Pistilli et al.¹⁷ also retrospectively evaluated risk factors for reduced vision in the same SITE study cohort referred to five tertiary referral centers over the same three-decade study period. As noted above, two-thirds of the subjects were female, nearly three-quarters (73.0%) were white and just over one in six (15.0%) were black. Approximately one in four (25.8%) had an associated systemic disease, most commonly sarcoidosis (6.8%), followed by spondyloarthropathy (4.8%), juvenile idiopathic arthritis (4.7%), inflammatory bowel disease (2.6%), Behçet’s disease (2.5%), or one or more less common autoimmune conditions (4.2%). Factors associated with worse vision over time included age over 65 years (1.9 lines worse as compared to 36–45 years; p < .001), Hispanic race (1.1 line worse than non-Hispanics; p < .001); current smoking (0.99 and 0.76 lines worse than never and past smokers, respectively; p < .001); a diagnosis of Behcet’s disease (1.0 line worse; p < .001), increasing time from onset of uveitis to referral to a one of the participating tertiary referral centers (up to 1.2 lines worse for > 5 years, as compared to ≤ 6 months, to referral; p < .001), and ocular surgery prior to referral, including retinal detachment surgery (3.6 lines worse; p < .001), cataract surgery (3.0 lines worse than phakic eyes; p < .001), pars plana vitrectomy (PPV) excluding for retinal detachment repair (1.3 lines worse; p < .001), and glaucoma surgery (0.68 lines worse; p ¸0.001). Eyes that underwent retinal detachment repair, PPV, or glaucoma surgery after referral had similar visual outcomes, whereas eyes that underwent cataract surgery after surgery did better than phakic eyes (2.9 letters; p < .001). Clinical findings associated with worse vision outcomes included worse anterior inflammation (up to 1.0 line; p < .001), vitreous inflammation (up to 0.59 line; p < .001), and vitreous haze (up to 4.2 lines; p < .001), the presence of hypotony (IOP ≤ 5 mmHg; p < .001), ocular hypertension (up to 0.34 lines for IOP ≥ 30 mm Hg, both as compared to IOP > 5 and < 24 mm Hg; p < .001), the presence of macular edema (0.90 lines; p < .001), and the presence or history of choroidal neovascularization (1.75 lines worse; p < .001), posterior synechiae formation (0.46 lines; p < .001), or peripheral anterior synechiae formation (0.42 lines; p < .001). The authors concluded that their results supported the generally held notion that enhanced control of active inflammation is associated improved outcomes, and that such outcomes may be more likely when patients are managed by uveitis subspecialists – particularly when the inflammation is moderate to severe and associated with cataract requiring surgery.

Jin et al.¹⁸ performed a cross-sectional, retrospective study of 245 uveitis patients and 105 controls with similar age, sex, income and education distributions to assess the quality of life and mental health status of subjects with uveitis vs controls seen in a tertiary referral center in Wenzhou, China between July 2018 and February 2019. Among subjects with uveitis, 16.7% had evidence of anxiety, whereas 26.5% had evidence of depression. Noteworthy predictors of anxiety identified after controlling for confounding using a multivariate logistic regression model included lower annual household income, poor self-reported visual function (VFQ-25), poor best-corrected visual acuity (BCVA) in the better seeing eye, and uveitic glaucoma, whereas predictors of depression included poor self-reported visual function, lower educational level, lower household income, current treatment with systemic biologic agents, and the occurrence of ocular complications of uveitis, such as complicated cataract, secondary glaucoma and macular edema. The authors concluded that both anxiety and depression were common in patients with uveitis and that important risk factors in their cohort included lower socioeconomic status and the occurrence of ocular complications known to lessen or limit visual function, with self-perceived visual function having a greater impact than BCVA.

Together, these studies highlight both the prevalence and impact of vision loss in patients with uveitis. They also support the generally held view that the severity, location and duration of inflammation, as well as the occurrence of common complications - such as macular edema, cataract, and secondary glaucoma, are important risk factors for the eventual development of limited or lessened vision. The risk of such vision loss seems to be less when patients are treated by uveitis subspecialists, leading some to suggest that early subspeciality referral might bode best for all but the most mild or routine cases, particularly since both perceived and measured vision loss appear to materially diminish the quality of life of patients with uveitis.

Disclosure Statement

No potential conflict of interest was reported by the author.