|Home | About | Journals | Submit | Contact Us | Français|
To describe the disease characteristics and visual outcome of pediatric uveitis.
Retrospective longitudinal observation.
527 pediatric uveitis patients from the National Eye Institute, University of Illinois, Chicago and Oregon Health Sciences University.
Retrospective chart review.
Demographics, uveitis disease characteristics, complications, treatments, and visual outcomes and were determined at baseline and at 1, 3, 5 and 10 year time points.
The patient population was 54 % female; 62.4% White, 12.5% Black, 2.7% Asian, 2.1 % multiracial and 14.61% Hispanic. Median age at diagnosis was 9.4 years. The leading diagnoses were idiopathic uveitis (28.8%), juvenile idiopathic arthritis-associated uveitis (20.9%), and pars planitis (17.1%). Insidious onset (58%) and persistent duration (75.3%) were most common. Anterior uveitis was predominant (44.6%). Complications were frequent, and cystoid macular edema (Odds Ratio: OR 2.94, p=0.006) and hypotony (OR 4.54, p=0.026) had the most significant visual impact. Ocular surgery was performed in 18.9% of patients. The prevalence of legal blindness was 9.23% at baseline, 6.52% at 1 year, 3.17% at 3 years, 15.15% at 5 years and 7.69% at 10 years. Posterior uveitis and panuveitis had more severe vision loss. Hispanic ethnicity was associated with a higher prevalence of infectious uveitis and vision loss at baseline.
The rate and spectrum of vision threatening complications of pediatric uveitis are significant. Prospective studies using standard outcome measures and including diverse populations are needed to identify children most at risk.
Children account for only 2.2 to 13.8 % of patients in many uveitis clinics, and most published series of pediatric uveitis are limited to a small number of patients.1, 2, 3 Some authors have proposed that uveitis in children has a relatively severe course and is more likely to lead to vision loss.4 Delayed diagnosis, extended burden of disease over a lifetime, limited treatment options in children, difficult examinations, and the risk of amblyopia are all challenges specific to childhood uveitis.5, 6 The purpose of this study is to describe the etiologies, disease activity, complications and visual outcome of pediatric uveitis in a large, diverse cohort of patients. A previous lack of standardization of outcome measures and definitions in uveitis prohibits large scale comparison and meta-analysis of smaller studies. The development of standardized uveitis nomenclature (SUN) in 2005 provided definitions of outcome measures related to disease activity, onset and course of uveitis.7
Databases from the National Eye Institute (NEI), University of Illinois at Chicago (UIC) and Oregon Health Sciences University (OHSU) were reviewed to identify uveitis patients diagnosed from 0 to 18 years of age, with at least one visit prior to August 2005. Data were retrieved retrospectively from visits at standard time intervals, i.e., baseline and 1, 3 and 5 years (+/-3 months) and 10 years (+/-2 years); review periods were therefore non-contiguous. No minimum follow-up period was required and if there were multiple visits within the interval; the visit with the most complete data on six specific variables (visual acuity, manifest refraction, intraocular pressure (IOP), slit lamp examination, dilated fundus examination and vitreous haze) was selected. Approval from the relevant Institutional Review Boards was obtained for this study; the work was Health Insurance Portability and Accountability Act (HIPAA)-compliant. A standardized data sheet of outcome variables was developed, and trained reviewers abstracted the study data from all available charts at each site. Data retrieved included date of birth, age at presentation, date of diagnosis, sex, self-reported race and ethnicity, uveitis diagnosis, anatomic location, and laterality. Detailed clinical information regarding the nature of uveitis, including presenting symptoms, intraocular inflammation, anterior segment complications, posterior segment complications, optic nerve related complications, disease course and duration, visual acuity, treatments, complications, treatment-related side effects, and reviewer-assigned activity status based on overall assessment (active or inactive) were recorded in a standardized manner. Best available visual acuity was recorded for the better eye for bilateral cases and for the affected eye for unilateral cases. Free text and “not specified” options were available for all time points and variables such that missing data did not result in data being carried forward.
The Standardized Uveitis Nomenclature (SUN) criteria were used to report the clinical data.7 Treatments specific for uveitis were recorded; the options were: corticosteroids (topical, local injection, oral, intravenous or combinations thereof), nonsteroidal anti-inflammatory drugs or sulfasalazine, methotrexate, azathioprine, mycophenolate mofetil, leflunomide, cyclosporine, tacrolimus/ FK506, cyclophosphamide, chlorambucil, etanercept, infliximab, intravenous immunoglobulin, alpha interferon, adalimumab, and daclizumab. The presence of treatment limiting side effects for medications other than corticosteroids were noted (though not specified). Complications were recorded if they occurred at any time point and categories included: band keratopathy, posterior synechiae, cataract, vitreous hemorrhage, cystoid macular edema (CME), epiretinal membrane, ocular hypertension (defined as IOP > 21 mm Hg requiring intervention), glaucoma (defined as ocular hypertension and optic nerve damage or visual field deficit), hypotony (defined as IOP ≤ 5 mm Hg), and disc edema. Ocular surgery was divided into cataract, glaucoma, cornea, posterior segment (including laser and cryotherapy) and other.
The proportions of patients who met the disease outcome variables were calculated at baseline, 1, 3, 5 and 10 year time points, and examined for association with visual outcome and clinical variables. Statistical analysis methods included descriptive statistics, chi square, and multiple logistic regression analyses, based on the type of variable and data distributions.
Patient characteristics are summarized in Table 1. 527 pediatric patients with uveitis were identified, of whom 285 (54 %) were female, 329 (62.4%) White, 66 (12.5%) Black, 14 (2.7%) Asian, 11 (2.1 %) multi-racial, and 105 (19.9%) were of unknown race. Across all races, 77 (14.61%) were Hispanic, 398 (75.52%) were non-Hispanic and 52 (9.9%) were of unknown ethnicity. There were no significant female/male differences among sites; however, the proportion of White patients differed significantly by site (p< 0.001): 78% at NEI, 64.2% at OHSU and 51% at UIC. Further, the prevalence of Hispanic ethnicity varied between sites: 7.9% at NEI, 10% at OHSU and 26.1% at UIC. Within the patients of Hispanic ethnicity, the racial distribution was 19 (24.7%) White, 3 (3.9%) multiracial, 2 (2.6%) Black, and 53 (68.8%) of unknown race. Other baseline characteristics are summarized in Table 1.
The most common primary reasons for referral to eye care provider were red eye (25.0%), reduced vision (23.0%), routine examination (16.5%), pain (6.8%), not specified (8.9%), floaters (6.0%), school eye exam (6.0%), other (5.0%) and photophobia (3.4%). The three most common diagnoses were idiopathic uveitis (28.8%), juvenile idiopathic arthritis (JIA)-associated uveitis (20.9%), and pars planitis (17.1%). Anterior uveitis was most common (235/527, 44.6%), followed by intermediate uveitis (148/527, 28.0%), posterior uveitis (76/527, 14.4%) and panuveitis (68/527, 13.0%). The most common etiology was idiopathic for anterior and panuveitis, pars planitis for intermediate, and infection for posterior uveitis (toxoplasmosis: 5.6% and other infections: 4.2%), see Table 2 (available at http://aaojournal.org). The majority of cases, 399 (75.7%), were bilateral and this predominance over unilateral presentation was demonstrated in anterior, intermediate and panuveitis; however, the frequency of unilateral presentation significantly exceeded that of bilateral presentation in posterior uveitis (62% unilateral versus 38% bilateral), (p< 0.001). Further, bilateral presentation was significantly more common in the most prevalent uveitis diagnoses, i.e., idiopathic (79.6%), JIA-associated (89.1%) and pars planitis (85.5%), as compared to other etiologies combined (overall 58.8%) (p< 0.001).
Of patients in whom disease activity could be classified, uveitis was most commonly insidious in onset (295, 55.9%) as compared to sudden (192, 36.4%). Duration was most commonly persistent (397, 75.3%) as compared to limited (68, 12.9%). A chronic disease course (139, 26.4%) was more common than acute (52, 9.9%) or recurrent (51, 9.7%). Applying the SUN definition of chronic disease7, 144, (27.3%) cases relapsed after discontinuation of therapy and 39 (26.4%) were unable to discontinue therapy due to persistent inflammation, see Table 3. In a logistic regression model including duration (persistent, limited) and course (acute, recurrent, chronic), with limited disease used as reference category, acute disease was significantly associated with baseline vision worse than 20/200 as compared with persistent, recurrent or chronic presentations (Odds Ratio: OR 5.89, p=0.021). Posterior and panuveitis were associated with greater risk for vision loss at all levels at baseline. Odds ratios ranged from 2.69-9.40, (p< 0.0001-0.012), (Table 4).
449 patients (85.2%) were treated with corticosteroids (topical, periocular, intraocular, oral or intravenous), 78 patients (14.8%) with non-steroidal anti-inflammatory medications or sulfasalazine, 31 patients (5.9%) with biologic agents, and 141 patients (26.8%) with systemic immunosuppressive agents other than steroids or biologics. Systemic immunosuppression with immunomodulatory or biologic agents or combinations thereof was employed in 172 patients (32.6%). Treatment limiting side effects (excluding those corticosteroid-related) were reported in 27 (5.2%). Ocular surgery was performed in 18.9% of total 527 patients, with cataract extraction being the most commonly performed (13.4%); other surgeries included posterior segment surgeries (9.7%) glaucoma surgeries (6.5%), corneal procedures (2.5%), and other operations (1.9%).
Complications were highly prevalent at baseline and throughout the follow-up period: at the first visit, the most common were posterior synechiae (177/527, 33.8%), cataract (104/527, 19.9%), and band keratopathy (109/527, 20.8%), (Table 5). The prevalence of ocular complications, particularly cataract, followed by ocular hypertension, glaucoma and CME, increased at subsequent time points. For example, one third of subjects had CME (15/63, 23.8%) and over half of subjects developed cataract (31/63, 49.2%) at 3 years. At baseline, the mean number of complications per person, regardless of visual acuity, was significantly greater for panuveitis and anterior uveitis compared to posterior uveitis; however, when adjusted for duration of disease, this difference became insignificant. There was no significant difference in rate of complications with regard to sex, race or ethnicity with the exception that White patients were more likely to develop glaucoma at 5 years and CME at 1 year compared to non-White patients (p=0.02 and p=0.01, respectively). Small numbers in these groups and lack of correction for multiple comparisons limit the certainty of these findings. The mean number of complications in patients with vision of 20/100 or worse was 2.29, Standard Deviation: SD 1.49 (n=69); the mean number of complications for those with vision better than 20/100 was significantly lower at 1.29, SD 1.36 (n=450), p= 0.0005. Further, the mean difference in number of complications for patients with vision of 20/50 or worse versus better than 20/50 was 0.57, 95% 95% Confidence Interval: CI 0.30-0.84, p< 0.0001, (Table 5). Multiple logistic regression of odds of vision worse than 20/50 in relation to complications revealed that CME (OR 2.94, p=0.006) and hypotony (OR 4.54, p=0.026) had the most significant impact on vision.
The proportion of patients with vision 20/200 or worse was 9.2% (48/ 527) at baseline, 6.5% (9/139) at 1 year, 3.2% (2/63) at 3 years, 15.1% (5/22) at 5 years and 7.7 % (1/13) at 10 years, (Table 5). Posterior uveitis and panuveitis were both associated with statistically significantly greater risk of vision loss at all levels: 20/200 or worse, 20/100 or worse and 20/50 or worse at baseline. The odds ratios for posterior uveitis ranged from 5.62 to 9.40, and those for panuveitis ranged from 2.69 to 4.02, see Table 4. Subjects whose uveitis met the SUN definition of “limited” disease course were significantly more likely to have presented with severe vision loss (20/200 or worse) at baseline than those with a “persistent” disease course (17.6% (12/68) vs 7.9 % (31/391), p = 0.007). Similarly, while 17.6% of patients with a limited course had vision of 20/200 or worse at baseline, no patients with limited course had this degree of vision loss at subsequent follow-up.
The effect of Hispanic ethnicity on clinical outcome measures was examined (Table 6). There was no difference in the mean number of complications, median age at diagnosis or presentation or female/male ratio between Hispanic and non-Hispanic children. There were important differences, however, in characteristics of uveitis in the Hispanic group when compared to non-Hispanics. For example, the distribution of anatomic location differed significantly with a lower prevalence of anterior (26/77, 33.8%, Hispanic, 185/398, 46.5%, non-Hispanic), intermediate (16.9% vs. 29.6%) and panuveitis (10.4% vs 13.8%), but a higher prevalence of posterior uveitis at 39.0% in the Hispanic group as compared to 10.0% in non-Hispanics (p< 0.0001). Further, while the prevalence of idiopathic uveitis was comparable at 27.4% in Hispanic and 29.8% in non-Hispanics, the prevalence of JIA-related uveitis and pars planitis were both significantly lower in Hispanics at 12.3% vs. 24.1% and 9.6% vs 19.2%, respectively (p=0.0004). In contrast, the ratio of infectious to non-infectious etiologies for uveitis was significantly greater in Hispanics than in non-Hispanics; the prevalence of toxoplasmic uveitis was 14/72, 19.2% vs. 11/398, 2.8% and that of uveitis due to other infectious causes was 9/72, 12.3% in Hispanics as compared to 23/ 398, 6.0% in non-Hispanics (p<0.0001). There were no cases of ankylosing spondylitis, Crohn's disease, sympathetic ophthalmia or Blau syndrome in the Hispanic group, while these diagnoses were seen in the non-Hispanic group.
In the Hispanic group non-infectious uveitis accounted for 21.7% of all unilateral cases, while toxoplasmosis and other infectious accounted for 43.5% and 34.8% of unilateral cases, respectively. This differed significantly from the non-Hispanic group in which non-infectious uveitis accounted for a larger proportion of unilateral disease at 69.2% and toxoplasmosis and other infectious accounted for a smaller fraction at 11.0% and 19.8% respectively (p<0.001). However, there were no differences in causes of bilateral uveitis presentation, the vast majority of which was due to non-infectious causes in both Hispanic (90.0%) and non-Hispanic children (98.0%). Hispanic ethnicity had no impact on the tendency of non-infectious uveitis to present as bilateral disease. Logistic regression analysis revealed that Hispanic ethnicity was associated with significantly greater risk of vision loss at baseline: 20/200 or worse (OR 2.83, 95% CI 1.45-5.55, p=0.002), 20/100 or worse (OR 2.59, 95% CI 1.43-4.69, p=0.002) and 20/50 or worse (OR 2.76, 95% CI 1.67-4.57, p < 0.0001, (Table 7). When uveitis diagnoses of idiopathic, JIA-related, pars planitis, toxoplasmosis, and other infections were added to the model, Hispanic ethnicity no longer predicted poor visual outcome, and toxoplasmosis (OR 8.86, 95% CI 2.35-33.44, p=0.0013) and infections (OR 9.81, 95% CI 2.82-34.15, p=0.0003) emerged as significant risk factors. However, when examining the effect of ethnicity and non-infectious etiologies only, Hispanics were still more likely to have visual acuity of 20/50 or worse at baseline (p=0.03) and test for linear trend indicated that at baseline Hispanics were more likely to have worse vision for all cutoffs (p=0.026).
The results from this large cohort show the spectrum of disease as well as the breadth of complications and visual outcomes in a study population that is geographically, racially and ethnically diverse, with a sizeable fraction of non-White patients. Table 8 shows the results of a comparison of our data to that of other studies that included 100 or more children published after 2000.
Visual acuity has been characterized in various ways in previous publications of pediatric uveitis cohorts. Edelsten et al defined vision loss as visual acuity of 6/12 or worse in at least one eye and found a prevalence of vision loss of 17% (43/249 patients)3. In that study, performed in the United Kingdom, a high rate of toxoplasmosis and idiopathic posterior uveitis resulted in a higher prevalence of posterior uveitis than reported in other studies. We used similar vision cutoffs and time points to a study performed in Miami 8 but found lower prevalence of vision loss with a baseline prevalence of legal blindness (vision of 20/200 or worse in the better eye) of 9.23% (48/ 527) as compared to 18% (11/61 bilateral cases) and 53% (17/32 unilateral cases) reported by Rosenberg and co-authors. 8 Edelsten, Rosenberg and our study all found that posterior uveitis is a risk factor for poor visual outcome. 3, 8
In addition to looking at the prevalence of legal blindness, we examined the prevalence of vision of 20/100 or worse and 20/50 or worse, since this defines the visual requirement for an unrestricted driver's licenses in many states and may be likely to affect vision-specific pediatric activities of daily living. We found the proportion of children with vision worse than 20/200 remained about 10% (mean 8.4%, range 3.4-13.3%) at each of the follow-up time points; however, on average, the proportion of children with moderate visual compromise with visual acuity of 20/50 or worse was 26.4% (18.1% - 36.4%). Children with vision of 20/100 or worse may be more likely to have difficulty with school work and the prevalence of this visually impaired group was 69/527 (13.3%) at baseline, 9/139 (9.4%) at 1 year, 5/63 (7.9%) at 3 years and 5/22 (15.1%) at 5 years. These data highlight the need for close monitoring and aggressive treatment of children with uveitis, as well as the need to ascertain the impact of uveitis on age appropriate activities of daily living relevant to children. The more common diseases, i.e., idiopathic, JIA-associated and pars planitis, were more often bilateral, and therefore lack of disease control may be more likely to result in bilateral loss of vision, with major functional impact. Infectious uveitis was more likely to present as posterior or panuveitis, and was associated with higher risk for poor visual outcome. In fact, when we looked at infectious and toxoplasmic uveitis in a logistic regression model including the leading 3 uveitis diagnoses and ethnicity, both toxoplasmosis and other infections combined were significantly associated with visual outcome of visual acuity of 20/200 or worse at baseline. Posterior and panuveitis in children may therefore be viewed as a poor prognostic marker and warrant further specific study and close intervention.
We did find some differences to suggest that race and ethnicity were associated with different patterns of etiology, disease presentation and progression, and this observation warrants further study. Our findings suggest that Hispanic patients were more likely to have infectious uveitis and to present with vision loss even with non-infectious uveitis. Because toxoplasmosis represented a sizeable proportion of infectious uveitis and due to its propensity to affect the macula, it was particularly associated with loss of visual acuity, a public health issue of import. With the changing demographics of the United States and globalization, patterns of uveitis in the population may be affected. There are global variations in the epidemiology of uveitis. A meta-analysis of worldwide studies of uveitis that included 7% pediatric patients reported parasitic anterior uveitis (49.3%) as the most common etiology for disease in children but idiopathic was the second most common at 25.5%. 9
The weaknesses of our study include its retrospective nature and examination by different ophthalmologists. Despite the large number of patients at baseline, data on fewer patients were available at each subsequent time point. We recognize that patients with longer follow-up especially at tertiary centers tend to be those with more severe disease, which could increase the prevalence of complications and poor outcomes. Estimates of severity and complications from academic referral centers are likely to exceed community-based prevalence. Strengths of our study include the geographic distribution of study sites, ethnic and racial diversity, large patient population, the use of standardized outcome variables at specific time points and the availability of data for up to 10 years of follow-up. The effectiveness of the SUN definitions of disease activity to separate patients into groups that behave differently supports the application of standardized outcome variables and highlights the need for additional codification of outcome measures in uveitis.7
While several studies have sought to identify prognostic factors for pediatric uveitis, relatively small numbers, lack of standardized definitions of outcome variables, and non-uniform, retrospective data collection have made this challenging. By combining 3 uveitis centers with different referral patterns and geographic locations, we attempted to maximize the number and diversity of subjects, and this is the largest study of pediatric uveitis to date. Standardized composite outcome measures such as those developed by the American College of Rheumatology, and the core outcome measures developed by the Pediatric Rheumatology International Trials Organization have enhanced the ability to critically evaluate new treatments.10 There is a need for standardized outcome measures and evaluation tools to measure the functional impact of pediatric uveitis and age-appropriate quality of life assessment tools to assess activities of daily living relevant to children.
Financial Support: Supported in part by grants from Research to Prevent Blindness (Career Development Award to JRS, unrestricted grant to Casey Eye Institute and unrestricted grant to University of Illinois, DAG). Supported in part by the intramural research program of the National Eye Institute / National Institutes of Health/ Department of Health and Human Services (JAS). The funding organizations participated in conducting the study, data collection, data management, and data analysis for the manuscript.
Meeting presentation: Presented as a poster at the AAO Annual Meeting, November 2007, New Orleans, Louisiana.
This article contains online-only material. The following should appear online-only. Table 2: Prevalence of uveitis diagnoses by anatomic location.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.