Characteristics of a cohort of children with Juvenile Idiopathic Arthritis and JIA-associated Uveitis

Background Juvenile idiopathic arthritis-associated uveitis (JIA-U) can lead to poor visual outcomes and impact a child’s quality of life (QOL) and function. Our aim is to identify risk markers of JIA-U and examine differences in the QOL of children with JIA and JIA-U. Methods Rheumatology and ophthalmology record reviews and questionnaires were completed every 4–6 months on 287 children with JIA. We collected arthritis, uveitis, and QOL data. We examined data through last study visit. Results There were 52/287 (18 %) children with JIA-U who were younger at arthritis diagnosis, had oligoarticular persistent JIA, and ANA positive. Confirmed uveitis predictors were age at JIA diagnosis (OR = 0.86) and oligoarticular subtype (OR = 5.92). They had worse vision specific QOL and function, but similar general QOL. Blindness occurred in 17.5 % of children but was more common in African American children compared to non-Hispanic Caucasian children ((5/7 (71 %) vs. 2/29 (7 %), p <0.001) despite a similar uveitis prevalence (22 % vs. 16 %). Both races had similar complications, although band keratopathy was more frequent in African Americans (75 % vs. 15.6 %, p = 0.003). Conclusions We confirm young age at JIA diagnosis and the oligoarticular JIA subtype as predictors of uveitis development. Although we were unable to identify predictors of ocular complications or blindness, AA children appeared to have a more severe disease course manifested by increased ocular complications, vision loss and blindness. Potential causes that warrant additional study include underlying disease severity, access to medical care and referral bias. Further investigation of the risk factors for vision-compromising uveitis and its’ long-term effects should be conducted in a large racially diverse population. Electronic supplementary material The online version of this article (doi:10.1186/s12969-015-0018-8) contains supplementary material, which is available to authorized users.

Vision loss and ocular complications have been reported in 3-66 % of children [9, 6, 10-13, 3, 2]. These devastating ocular sequelae may be prevented by identifying children at greatest risk for severe disease.
The American Academy of Pediatrics guidelines recommend ophthalmology screening every 3-4 months in children with oligoarticular or polyarticular rheumatoid factor (RF) negative arthritis who are antinuclear antibody (ANA) positive, <7 years of age, and diagnosed with arthritis for ≤4 years since they are at high risk for developing uveitis [14]. These factors have been confirmed in several studies [14, 5, 15-17, 11, 9]. Risk factors for vision loss and/or ocular complications include male gender, short duration between arthritis and uveitis diagnoses, a uveitis diagnosis prior to arthritis diagnosis, young age at uveitis onset, an anterior chamber cell score ≥1+, initial visual acuity (VA) of 20/200 or worse, and presence of complications at first ophthalmology examination [18][19][20][21]. Hence, early detection of ocular disease and timely aggressive treatment in high risk patients can prevent poor visual outcomes.
Our primary objective is to identify risk markers for uveitis development and a severe ocular course in children with JIA by (1) comparing those with JIA alone and those with JIA-associated-uveitis (JIA-U), and (2) analyzing children with severe uveitis. We defined severe disease as a history of ocular complications, vision loss (VA 20/50 or worse) and legal blindness (VA 20/200 or worse).

Methods
This is a retrospective analysis of prospectively collected data. It was approved by the Institutional Review Board of Emory University and conformed to the requirements of the US Health Insurance Portability and Privacy Act. Informed consent was obtained from the parent/legal guardian, and assent was obtained from the children if applicable. Research adhered to the tenets of the Declaration of Helsinki.

Subjects
Eligible children presenting with a diagnosis of JIA or uveitis were enrolled from the pediatric rheumatology outpatient clinics from November 2011 to January 2015. They were followed prospectively from time of enrollment and returned for their usual follow up appointment every 3-6 months. Inclusion criteria included: 1) a diagnosis of JIA with or without uveitis and 2) English speaking. Exclusion criteria included: 1) a diagnosis of another chronic or autoimmune disease associated with uveitis; 2) major developmental disorders (i.e. cerebral palsy, autism); 3) other chronic disorders that can affect quality of life (QOL) and function; and 4) refusal to participate. Some patients were diagnosed with JIA or uveitis months or years prior to enrollment.

Data collection
We reviewed rheumatology and ophthalmology medical records from time of presentation to every 3-6 months during the usual pediatric rheumatology follow up visit. Baseline data collected included date of birth, gender, parents self-described race and ethnicity, arthritis characteristics (onset, date of diagnosis, JIA subtype, joints with tenderness swelling and/or limitation, radiographs), uveitis characteristics (onset, date of diagnosis, laterality, location, ocular complications, surgeries), ocular exam (best corrected visual acuity (VA), intraocular pressure, anterior chamber cells and flare), labs (ANA, RF, erythrocyte sedimentation rate (ESR), HLA-B27, and anti-cyclic citrullinated peptide (anti-CCP)) and all medications used currently and in the past. Follow up data on arthritis, uveitis, labs and medications were collected at 3-6 month intervals (time of last study visit to current study visit) and during uveitis flares. Data from the ocular exam was recorded from every ophthalmology visit with varied intervals depending on presence of uveitis and disease activity.

Measures of quality of life and function
We administered parent-and patient-based questionnaires at each study visit -the Childhood Health Assessment Questionnaire (CHAQ) to measure physical function, the Pediatric Quality of Life Inventory Version 4.0 (PedsQL) to measure overall QOL, and the Effects of Youngsters Eyesight on Quality of Life (EYE-Q) to measure vision related QOL and function [22][23][24][25]. Childreport forms were administered as appropriate per age based on the specific measure. Questionnaires were read by the study team to children who had significant visual impairment.
The CHAQ evaluates functional disability and has well-documented reliability and validity [22]. There are 20 questions encompassing 8 functional components: 1) dressing and grooming, 2) arising, 3) eating, 4) walking, 5) hygiene, 6) reach, 7) grip, and 8) activities. There are three parameters within each area: 1) difficulty in performing daily functions, 2) use of special aids or devices, and 3) activities that require assistance from another person. Scores range from 0 to 3 and lower scores indicate better physical function.
The PedsQL is a validated measure of health-related QOL in children and adolescents from 2 to 18 years of age [23]. Four core scales are measured on a 5-pointscale: 1) physical functioning (8 items), 2) emotional functioning (5), 3) social functioning (5), and 4) school functioning (5). Scores range from 0-100 with higher scores indicating better QOL. For this study, we included measures of overall QOL (PedsQL Total) and psychosocial related QOL (PedsQL Psychosocial).
The EYE-Q is a measure of vision related QOL and function in children [25,24,26]. It consists of both parent and patient self-reports for children 5-18 years old. There are 19 items that query on tasks that rely on distance, near, color, and night vision; functionality; and photosensitivity. QOL items inquire about feelings regarding medication use, missing school, and lab draws using a 5-point Likert scale response format. There is a question about ocular symptoms, the use of visual aids, and a subjective assessment of vision severity.

Statistical analysis
Statistical analyses were conducted using SAS v9.3 for Windows (Cary, NC). Statistical significance was assessed at the 0.05 level unless otherwise noted. Data are summarized using means and standard deviations, medians and interquartile ranges (25 th -75 th percentile) or counts and percentages, when appropriate. Duration of arthritis and/or uveitis were calculated as time between date of diagnosis to last study visit. Duration between uveitis and arthritis diagnosis was calculated by subtracting date of arthritis diagnosis from date of uveitis diagnosis. We compared children with JIA alone and JIA-associated uveitis using Chi-square tests and two sample t-tests. In instances of small expected cell counts (<5), exact Chi-square tests were used. When continuous data were skewed or nonnormal, a nonparametric test (Mann Whitney-U or Kolmogorov-Smirnov test) was used in place of the two sample t-test.
We conducted further analysis in children with uveitis and compared those with and without ocular complications, those with a VA of 20/200 and worse, and African American (AA) and Caucasian (W) children with uveitis. Subgroups (i.e., complicated vs. uncomplicated, vision loss vs. no vision loss, AA vs. W) were compared using similar methodology described above. Multivariable logistic regression was used to simultaneously model multiple risk factors for development of uveitis. Similar models were constructed for the outcome of complicated disease. Candidate predictors identified in the univariate analyses were entered into the model. Stepwise backward elimination was used to obtain the final model. Non-significant variables were systematically removed until all variables in the model were significant at the 0.05 level or a significant reduction in model fit was observed.

Demographics and clinical characteristics of cohort
There were 287 children with JIA of whom 52 (18 %) had uveitis at the last study visit (Table 1). Of the 52 children with uveitis, 5 with JIA alone developed uveitis after enrollment, and 1 child with uveitis alone developed JIA. Our cohort consisted of 71 % females, 77 % W, and 12 % AA children. The median (IQR) age of arthritis diagnosis was 6.4 years (2.9 -11.7), and 41 % were of the oligoarticular persistent JIA subtype.
On sub-analysis, we then considered 91 children who had JIA for ≥4 years since approximately 80 % of the cases of uveitis in children with JIA are diagnosed within 4 years of JIA diagnosis. Even after restricting the analyses to this subgroup, those with uveitis were still likely to be younger at arthritis diagnosis (p < 0.001), have oligoarticular persistent JIA (p <0.001), be RF negative (p = 0.026) and be anti-CCP negative (p = 0.026). The associations between ANA positivity, and the other JIA categories were no longer significant.
In comparing males and females with uveitis, females were younger at arthritis and uveitis diagnosis (2.3 years vs. 6.6 years, p = 0.002, and 4.5 years vs. 7 years, p = 0.017) and ANA positive (66.7 % vs. 16.7 %, p = 0.003). There were no differences in ocular complications or medication use.

Predictors of uveitis development
On univariate analysis, oligoarticular persistent JIA subtype, ANA positivity, age at arthritis diagnosis, and anti- .75]) and were the best combination of predictors of uveitis among the models we considered (AUC = 0.77).

Quality of life and function
Children with uveitis had worse vision related QOL and function compared to children with JIA alone as measured by the EYE-Q child and parent reports (p < 0.001) ( Table 1). However, all children had similar overall QOL, psychosocial QOL and physical function as measured by the PedsQL and CHAQ.

Medication use
Most children with uveitis were treated with subcutaneous and/or oral methotrexate (85 %

Epidemiology of uveitis
In our 52 children with uveitis, we had data on time to uveitis diagnosis in 50 patients in relation to arthritis diagnosis (Table 3). Approximately 24 % of children were diagnosed with uveitis prior to JIA, and 22 % within the 1 st year of their JIA diagnosis. Cumulatively, 86 % of children were diagnosed with uveitis within 4 years of their JIA diagnosis, and 4 % by 5 years.

Children with uveitis compared by race
In examining our data on the 8 children with legal blindness (VA of 20/200 or worse), we noted an increased frequency of AA children (5/7 (71 %) with a blind outcome compared to W children (2/29 (7 %)) despite a similar uveitis prevalence (22 % vs. 16 %) when restricting to non-Hispanic Caucasians and excluding children with missing data (Table 4). Although our results on race are underpowered due to a small number of AA children with uveitis, we conducted sub-analysis to compare AA . They appeared to have more ocular complications, and upon further examination, AA had on average 3 times as many complications as NHW (p < 0.001). There were no differences in insurance status, age at arthritis diagnosis, JIA subtypes, duration between uveitis and arthritis diagnoses, treatment, or QOL. Table 5 examines the characteristics of the 8 AA children with uveitis based on age of JIA diagnosis. Four children were diagnosed with JIA at <7 years of age (AAP cut off age for frequent ocular examinations) of whom three were diagnosed with uveitis within 4 years of arthritis. They had a history of a VA of worse than 20/ 200 and ocular complications. The four that were diagnosed with JIA ≥7 years of age had uveitis either at presentation or prior to a JIA diagnosis and three had better VA and less complications. The one child diagnosed at 7 years of age had disease that was similar to the younger group, and it is possible she had an earlier onset that was unrecognized.

Discussion
Our study confirms known risk factors associated with uveitis development in children with JIA-young age at JIA diagnosis, oligoarticular JIA subtype, ANA positivity, RF negativity and anti-CCP negativity; only the oligoarticular JIA subtype and young age at arthritis diagnosis remained significant predictors in our model [6,27,12]. When analysis was restricted to children at lower risk for uveitis (a diagnosis of JIA >4 years), these two predictors remained significant. We also confirm reports that most children are diagnosed within the first 4 years of arthritis (86 %), hence regular screening within the first 4 years of a JIA diagnosis is crucial. We were unable to identify predictors of severe uveitis (increased ocular complications, blindness) and found no significant differences in gender, duration between arthritis and uveitis diagnoses, age at uveitis onset, and a uveitis diagnosis prior to arthritis diagnosis in children with severe disease.
Uveitis can have a negative impact on a child's daily function and ability to perform visual tasks in the home and school [28]. Additionally, children need regular ophthalmology and rheumatology physician visits, phlebotomy, frequent ophthalmic drops, and systemic immunosuppressive therapy consisting of oral medications, injections and infusions. Investigation into the effects of uveitis on a child's QOL and function is important. Most studies use general QOL measures, arthritis specific measures, and the ocular exam as a measure of outcome. The EYE-Q is the only patient reported outcome measure specific to uveitis which we are currently validating in a pediatric uveitis population [24,25]. In our cohort, children with uveitis appeared to suffer from worse vision specific function and QOL but no differences were apparent in general QOL or physical function. We expected those with both JIA and uveitis to experience worse general QOL due to involvement of two systems. It is possible that any diagnosis of chronic autoimmune disease affects QOL equally since care and treatment are similar, irrespective of number of diagnoses and disease activity. This emphasizes the importance of a global and comprehensive assessment of outcomes that takes into consideration vision specific measures, and not limited to general measures.
Race can be associated with various autoimmune conditions, including JIA and systemic lupus erythematosus. In JIA, more children of European descent develop oligoarticular disease, whereas non-European children develop polyarticular rheumatoid factor (RF) positive JIA with worse arthritis outcomes [29][30][31]. Since JIA subtypes differ in their association with uveitis, and race predisposes to different JIA categories, race may influence uveitis risk.
Few studies describe the association of race with pediatric uveitis. In adults, the Pacific Ocular Inflammation Study noted that of 224 uveitis cases, 38 % were Asian, 37 % were white, and only 2 % were Black [32]. However, Blacks had the highest incidence of uveitis at 39.6 cases per 100,000 patient years. In a JIA cohort in Toronto, the incidence of uveitis was similar in children of European and non-European ancestry (14.4 % vs. 12.8 %) although when compared to the general Toronto population, risk was increased in European children and decreased in non-European ancestries [29].
Uveitis has been reported in approximately 6-8 % of AA children with JIA and ANA positivity varies although other reports no associations with race (Table 6) [33,34,29,16,35]. In our cohort, uveitis frequency was similar in AA and W children (22 % vs. 17 %), with only   [16]. However, our population may differ due to our location in the Southeast leading to ascertainment bias. A 1984 study demonstrated that 8.3 % of 42 South African children with JIA developed uveitis, but none were ANA positive [36]. In 1997, of 172 children with juvenile rheumatoid arthritis (JRA), 20 % were of African descent, 8.7 % of these had uveitis, and none were ANA positive [34]. In 2007, 4 % of 758 children with JIA were of African descent and had a lower risk of developing JIA-U [29]. In 2010, Adelowo examined 23 children with JIA and noted that only two (8.6 %) were diagnosed with uveitis [37]. In 2013, we reported that only 13 of the 220 (5.6 %) AA children in the Childhood Arthritis and Rheumatology Research Alliance (CARRA) registry developed uveitis and 4 of them were ANA positive [16]. Hence, only small cohorts of AA children with JIA and uveitis have been examined.
There is less information on the uveitis course of AA children with JIA, and one study describes severe outcomes in African children with uveitis secondary to various etiologies (idiopathic, Still's disease, and sympathetic ophthalmia) [38]. Woreta et. al. report that in their JIA-U cohort, there were no associations between race and ocular complications, a VA of 20/50 or worse, or a VA of 20/200 or worse, although they only had 5 blacks and 64 white patients [39]. In our cohort, AA and W children had similar arthritis characteristics since all AA children were primarily of the oligoarticular or polyarticular RF negative JIA subtype, and uveitis was diagnosed within the first 4 years of JIA. However, AA children were  diagnosed at an older age and suffer from more ocular complications, vision loss, and blindness. Examining the characteristics of each AA child with JIA-U, we note that those diagnosed with JIA at 7 years and younger had a more severe course, hence, the AAP recommended screening guidelines are crucial and more frequent monitoring may be important (Table 5). We are unable to ascertain whether this is secondary to biologic differences or variances in access to health care. This needs further study in a large racially diverse multi-center population.
There are otherwise no other reports on the course of disease in children of African descent.

Strengths and limitations
Our current cohort may not accurately reflect uveitis prevalence since not all children have had JIA for 4 years and may still develop uveitis. Thus, longer follow up is needed, especially in young children with early disease since they are at highest risk for uveitis. We continue to follow our cohort prospectively and recruit patients with newly diagnosed JIA. Exploration of the association of race with uveitis should be performed in a larger and more diverse cohort, likely in a multi-center study. There appeared to be differences in uveitis course based on race, but our analysis was underpowered. However, although our cohort of AA children is small, the number of those with JIA-U is similar to previously published reports, including our analysis of the large CARRA registry which had 13 (6 %) AA children with uveitis (Table 5). We had an increased frequency of uveitis compared to other studies, and this may be due to our institution being a tertiary center in the Southeast where there is a larger AA population. A report of uveitis prevalence in adults in the Southeastern US in 1997 demonstrated that of 385 patients, 31 % were African American which is also of increased frequency [40]. Focusing on regions with diverse populations could help us better appreciate the differences in outcomes associated with race, and our results suggests the need for further evaluation.
We were unable to make conclusions on the role of Hispanic ethnicity since non-English speaking patients were excluded. However, we have translated all our instruments and are actively recruiting children from this population.

Conclusions
In conclusion, our results confirm age at JIA diagnosis and the oligoarticular JIA subtype as predictors of uveitis development, but we were unable to confirm predictors of severe disease. We demonstrate that visual disability may be underestimated in JIA outcome studies since few studies look at vision specific QOL and function in children with uveitis. Vision specific measures may reveal areas of vision that are not quantified by general measures or the ocular exam. A comprehensive assessment of disability that incorporates uveitis specific measures may be a more clinically sensitive assessment of visual outcome.
Our study also highlights the potential role of race in uveitis and the need for further exploration in a large racially diverse cohort, likely in a multi-center study. Identifying high risk children will enable early, aggressive treatment and potentially prevent visual complications associated with severe disease.