Viera Kalinina Ayuso
Thursday 7 February 2013
Uveitis in childhood. Clinical and fundamental developments.
Promotor: Prof. dr. A. Rothova
Defence: 7 February 2013
Chapter 1 of this thesis provides a general overview of uveitis in childhood. Although uveitis in childhood is a relatively rare condition, it has a significant impact in ocular morbidity and visual disability in children suffering from this disorder. Uveitis in childhood is a potentially blinding disease due to frequently occurring vision-threatening complications in these patients. Its onset is often asymptomatic, but even in advanced cases young children are unaware of their visual disability and usually do not complain. In consequence, the diagnosis of uveitis in a child is regularly delayed and severe complications can be already present at the first ophthalmologic presentation. However, detection of uveitis signs in a child is only a first step in the diagnosis of the cause of uveitis. The subsequent work up should focus on the detection of underlying systemic or infectious causes of uveitis. The main systemic association of uveitis in children is juvenile idiopathic arthritis (JIA), which accounts for at least 30% of uveitis cases in childhood. In the other cases, even after an extensive diagnostic work-up, an idiopathic nature of uveitis is being frequently concluded. Second most common uveitis entity in children is intermediate uveitis, which is almost always idiopathic in pediatric population. The exact pathogenesis of intraocular inflammation in children is not yet clarified, however it is generally accepted that it is an autoimmune disorder in which T cells play the main role. Management of uveitis in children is usually prolonged, complex and challenging. Although locally administered corticosteroids represent the first line treatment, systemic treatment with immunosuppressive drugs is warranted in many cases. Methotrexate is frequently the systemic agent of the first choice, especially in patients with JIA. Complications of childhood uveitis frequently require surgical treatment.
This thesis aimed to gain new insights regarding the course and prognosis of uveitis in childhood (chapter 3-5;9), the pathogenesis of JIA-associated uveitis (chapter 7-8) and the treatment of uveitis in children (chapter 6;11).
Finally we wanted to study prospectively ocular complications in immunocompromised children after hematopoietic stem cell transplantation (HSCT) and specifically to evaluate their risk of uveitis development during systemic viral reactivations. Further, we investigate whether a routine ophthalmologic screening is warranted for pediatric HSCT population (chapter 12).
Chapter 2 is devoted to the complex pathogenesis of JIA-associated uveitis. We review the literature and summarize the current knowledge regarding the pathogenesis of JIA-associated uveitis. We discuss the possible role of immune responses and cytokine involvement, genetic associations and the influence of external triggers in this disease, supported by data obtained from experimental uveitis models.
All of the combined data point to the development of aberrant immune responses involving both the innate and adaptive immune responses. CD4+ T cells are the primary cell type involved in the pathogenesis of JIA-associated uveitis. Specifically, CD4+ T cells serve as both effectors (T helper (Th) 1 and Th17) and regulators of the inflammatory process (T regulatory cells (Tregs)). An imbalance between pro-inflammatory Th1/Th17 cells and anti-inflammatory Tregs is believed to play the crucial role in the inflammation in JIA. In contrast, the role of B cells has been studied less extensively and it is yet unresolved, but these cells may also play a role in the pathogenesis of JIA. B cells and plasma cells (terminally differentiated B cells) are being described as the most abundant cell type infiltrating the eye in JIA-uveitis. However, these findings are primarily based on enucleated eyes, thus representing the end-stage of the disease process, which might be distinct from the earlier active stages of intraocular inflammation in JIA. JIA-associated uveitis is commonly marked by antinuclear antibody positivity, which reflects the production of abnormal self-targeting antibodies in this disorder; however, the nature of the self-antigen remains unknown, and the role of these antibodies in the pathogenesis of the disease is still unclear.
JIA-associated uveitis appears to have a genetic predisposition, particularly with respect to human leukocyte antigen (HLA) class II genes; to date, however, only a protective role of HLA-DR1 has been independently confirmed. JIA-associated uveitis has a relatively low concordance rate among twins, which suggests the involvement of external (non-genetic) factors. Although many external factors—including a variety of infections—have been proposed to increase the risk of developing JIA, none of these factors has been adequately confirmed. The development of chronic and recurrent inflammation in the eyes and joints suggests a similarity between these two organs, although this mechanism and eventual similarities remain unclear.
In chapter 3-4 we analyze the role of baseline factors in long-term development of ocular complications (chapter 3) and in visual outcomes (chapter 4) in JIA-associated uveitis. We performed retrospective analysis of data from 117 affected eyes of 65 patients with JIA-associated uveitis with median follow-up of 7.6 years (minimally 1 year). The prognostic roles of gender and age of onset and atypical manifestation of JIA with initial uveitis before arthritis have been specifically analyzed. In both chapters we document the association of male gender with poor prognosis in JIA-associated uveitis. We found significantly more complication in boys, compared to girls: at 5 years of follow-up boys suffered more frequently from cystoid macular edema (50% vs 4%) and papillitis (31% vs 2%), and needed cataract surgery more frequently than girls (59% vs 32%).
Uveitis was diagnosed before arthritis in 23% of the patients of this study. So, in these patients uveitis represented the first sign of JIA. Children with initial uveitis had at 5 years of follow-up significantly more frequently posterior synechiae, band keratopathy, and cystoid macular edema, but less secondary glaucoma than children with classic presentation of JIA with initial arthritis and later uveitis. Interestingly, boys presented with uveitis as initial manifestation of JIA significantly more frequently than girls (44% of boys vs 15% of girls).
In multivariate analysis male gender appeared to be independently associated with cataract surgery, cystoid macular edema and papillitis, while initial uveitis was an independent factor associated with development of posterior synechiae, which is also known as a risk factor for a complicated course of JIA-associated uveitis from the literature.
Visual outcome of boys compared to girls was also significantly worse at 1 and 3 years of follow-up. But this was also the case for children with initial uveitis before arthritis, who had significantly worse visual acuity compared to children with initial arthritis, up to 3 years of follow-up.
In the whole series, blindness, defined as best corrected visual acuity of 20/200 or worse, was identified in 10% of affected eyes in our series. This means that 15% of patients in our study were affected by uni- or bilateral blindness. Secondary glaucoma was the most common cause of visual loss in these eyes (50%). Boys’ eyes were significantly more frequently affected by blindness than girls’ eyes (21% vs 6%). The majority of patients with blind eyes had initial uveitis instead of arthritis (58% vs 42%) and this difference was statistically significant. However, in a multivariate analysis blindness was independently associated with male gender and not with initial uveitis.
No significant difference in development of complications and vision between younger-onset (<7 years) and older-onset (>7 years) groups was noted in this study.
We conclude that male gender and uveitis as initial manifestation of JIA are associated with a complicated course in JIA-associated uveitis. Male gender also appears to be a risk factor for poor visual prognosis in JIA-associated uveitis. Age of onset does not seem to have independent prognostic value for the course of JIA-associated uveitis. We recommend to ophthalmologists to be aware of the role of these factors and to take them in their prognostic consideration in an individual patient with JIA-associated uveitis.
In chapter 5 we investigate the course and activity of JIA-associated uveitis in childhood and puberty in a retrospective analysis of clinical data of 62 patients with JIA and uveitis. It is remarkable that although girls with JIA are known to have a higher risk of developing uveitis, boys with JIA-associated uveitis, appear to have a more serious progression of disease, with a poorer prognosis, as described in chapter 3-4. The cause of these differences between genders is unknown, but it might suggest an involvement of hormonal influence.
In general, high uveitis activity was found during early years of age (4-6 years of age), followed by a quiet stage around the age of 9 and 10, after which activity rose again. This association was significantly compatible with a quadratic model. We also saw an increase in systemic medication with age in these children.
The results of this study suggest that JIA-associated uveitis encompasses a biphasic course; a high initial disease activity, followed by a quiet stage and a new wave of activity during early teenage years. This increase in uveitis activity during early teenage years was also indirectly reflected by the increased use of immunosuppressive medication in estimated puberty years in patients with JIA-associated uveitis.
Chapter 6 is a study that attempts to answer the question about the effect, optimal duration of methotrexate therapy and about the risk of relapse of uveitis after its discontinuation. In this retrospective study we assessed clinical data of 22 pediatric JIA patients, treated with methotrexate for active uveitis. Eighty-two percent of patients showed improvement of their uveitis activity with at least a 2-step decrease of the anterior chamber cells within the first year of methotrexate therapy. A topical steroid–sparing effect was observed when methotrexate was administered for a period from 3 to 9 months. Methotrexate was discontinued because of inactive uveitis in 13 patients. In sixty-nine percent of these patients a relapse of uveitis was observed after a mean time of 7.5 months after the withdrawal of methotrexate. In 46% of patients a relapse occurred within the first year after the withdrawal of methotrexate. Longer inactivity under methotrexate therapy was independently protective for relapses after the withdrawal; to be specific, 1-year increase of duration of inactive uveitis before the withdrawal of methotrexate resulted in a decrease of hazard for a new relapse of 93%.
In conclusion, our study confirms the efficacy of methotrexate in the management of JIA-associated uveitis. After the withdrawal of methotrexate, the relapse of uveitis occurs in the majority of JIA patients within the first year. Our results indicate that prolonged treatment with methotrexate and especially a longer period of inactivity before the withdrawal may be desirable to minimize new relapses after the withdrawal of methotrexate. Our results indicate that the period of inactivity before withdrawal should be preferably 2 years or longer.
In Chapter 7 we search for intraocular biomarkers in JIA-associated uveitis, using Surface Enhanced Laser Desorption/Ionization Time of Flight Mass Spectrometry (SELDI-ToF MS). Keeping in mind that uveitis can be first sign of JIA, as described in chapter 3-4, and the absence of a definitive diagnostic test, there is a strong rationale for efforts to identify specific biomarkers in the ocular fluid of children with JIA-associated uveitis. We investigate protein profiles in aqueous humor (n = 73) and serum (n = 105) samples from a total of 116 children with and without uveitis. The study population includes patients with JIA-associated uveitis (definitive JIA), suspected JIA (uveitis typical for JIA, but not meeting criteria of International League against Rheumatism for diagnosis of JIA), other pediatric uveitis entities and non-inflammatory pediatric controls.
We found that in the definitive JIA group 3 protein peaks in aqueous humor at mass/charge (m/z) 6,672; 8,725 and 13,762 had qualitative and quantitative differences in expression compared to the other uveitis entities and the controls. Interestingly, definitive JIA and suspected JIA samples did not differ in their protein profiles in aqueous humor. The quantitative expression of m/z 13,762 peak in definitive and suspected JIA samples was associated with activity of uveitis. By the intraocular presence of the protein peak at m/z 13,762 definitive JIA samples could be distinguished from other uveitis entities and controls with a sensitivity of 64% and a specificity of 95%. Using this algorithm, 63% patients with suspected JIA could be clustered together with definitive JIA patients. Mass spectrometric analysis revealed the peak at m/z 13,762 as transthyretin. Increasing evidence suggests involvement of transthyretin in various biologic processes. A plentitude of isoforms have been reported for transthyretin, many of which are causally associated with formation of amyloid fibrils. Intraocular production of transthyretin by ciliary pigment epithelium has been described in a rabbit model in the literature. Considering its suggested production by ciliary pigment epithelium, local production of transthyretin in aqueous humor in our study seems plausible. In our study the ion at m/z 13,762 was not detected in serum, which might suggest its intraocular origin.
We conclude that aqueous humor of patients with JIA-associated uveitis has a distinct expression profile of proteins compared with controls and other uveitis entities. These profiles are comparable with profiles of uveitis patients suspected for JIA-association on clinical grounds, but who do not meet criteria of International League against Rheumatism completely. These findings suggest common intraocular molecular processes in patients with JIA-associated uveitis and chronic anterior uveitis without arthritis. Transthyretin, the protein identified, could be involved in the pathogenesis of JIA-associated uveitis; however its role and the associated mechanism require further investigation.
Chapter 8 describes the immunohistochemical investigation of cellular infiltration in uveitis in JIA and other pediatric uveitis entities. Iris specimens were obtained during trabeculectomy, from 24 eyes with uveitis diagnosed before the age of 16 years and 6 eyes of children with open angle non-uveitic glaucoma. Histologically, mild to moderate signs of inflammation were found in 64% of specimens with JIA-associated uveitis while clinically only 9 % of these patients had mild uveitis activity shortly before surgery. All but 1 iris specimens from patients with JIA-associated showed a non-granulomatous type of inflammatory reaction. One of the JIA specimens was characterized by the abundance of giant cells typical for a granulomatous inflammatory process. This is a remarkable finding since uveitis in JIA is generally considered as a non-granulomatous entity. Two non-granulomatous JIA specimens with moderate inflammatory reaction were studied immunohistochemically. Both specimens showed presence of CD4+, CD68+ and CD138+ cells, while CD8+ and CD20+ cells could be detected only in one of them. Five different uveitis specimens including JIA, idiopathic anterior uveitis and idiopathic intermediate uveitis, were consistent in the presence of CD4+ T cells in the inflammatory infiltrates, while presence of other cell types in these specimens was discrepant.
We conclude that CD4+ T cells are present in inflammatory infiltrates in different childhood uveitis entities, including JIA-uveitis. Uveitis in JIA is characterized by a mixed inflammatory infiltrate in the iris with involvement of CD4+ T cells, plasma cells and histiocytes. These findings could form be the basis for further research for the specification of the subsets and the roles of these cells in the pathogenesis of JIA-associated uveitis.
In Chapter 9 we investigate prognostic factors in the second most common uveitis entity in children, intermediate uveitis by retrospective analysis of clinical data of 35 children with intermediate uveitis. Visual outcomes and development of complications were analysed in relation to age of onset (< 7 and > 7 years) and ocular signs at presentation. We found that younger onset of intermediate uveitis was associated with higher risk of development of secondary glaucoma, vitreous haemorrhage and cataract. The younger-onset group also had worse visual outcomes up to 3 years of follow-up compared to the older onset group. Unilateral legal blindness occurred in 20% of the patients, no difference in frequency of blindness was found between younger and older onset groups. Cystoid macular edema and/or maculopathy represented the main cause of blindness in this population. Development of cystoid macular edema was independently associated with papillitis and snowbanking at initial presentation. Other complications at onset were not predictive for future complications. Over time, anti-inflammatory (and immunopuppressive) therapy (topical and systemic) could be stopped completely in 69% of patients. Continuation of any kind of anti-inflammatory therapy was significantly longer and more frequently needed in children with a younger age of onset compared with the older onset group. In the group with onset of intermediate uveitis < 7 years, remission was reached 25% versus 58% in patients with older onset of intermediate uveitis.
We conclude that children with young onset of intermediate uveitis carry a higher risk of complications and worse visual outcome. Papillitis and snowbanking at initial presentation are unfavourable factors (unrelated to age of onset), associated with development of cystoid macular edema, the main cause of visual loss in pediatric intermediate uveitis.
In chapter 10 we report on 3 pediatric cases with coexistence of idiopathic intermediate uveitis and alopecia areata. These are the first 3 cases with this coexistence of which we are aware in the literature. Alopecia areata preceded the diagnosis of bilateral intermediate uveitis in 1 child and followed within several months after intermediate uveitis diagnosis in 2 children. All 3 children were otherwise healthy. The results of extensive diagnostic evaluations were negative in all 3 cases. However, family history of all 3 children was positive for other autoimmune conditions. The severity of uveitis ranged from mild to sight-threatening, and hair loss ranged from local lesions in 2 cases to total alopecia in 1 case. The exact pathogenesis of intermediate uveitis and alopecia areata has not yet been revealed. Both conditions are considered to be autoimmune and T-cell-mediated with CD4+ cells playing a major role in their pathogenesis. Both the eye and the hair follicle are considered to be immune-privileged sites with similar mechanisms of immune privilege. So, an immune-privilege collapse or an escape from peripheral tolerance might be key factors associated with the induction of both conditions.
We conclude that theoretically, the coexistence of intermediate uveitis and alopecia areata might be based on the similarities in their complex pathogenesis. However, more research is needed to evaluate if the coexistence is based on an association between 2 autoimmune disorders or if it is a coincidence.
In Chapter 11 we assess the effect of an Ahmed glaucoma valve implant on corneal endothelial cell density in children with uveitic glaucoma. The results of this cross-sectional study (80 eyes of 42 patients included) show that previous Ahmed glaucoma valve implantation is associated with a lower corneal endothelial cell count in eyes with secondary glaucoma due to uveitis. Endothelial cell density was significantly lower in eyes with an Ahmed glaucoma valve implant compared to eyes with uveitis without the implant. The mean time following an Ahmed glaucoma valve implantation contained 3.5 years in our study. Factors significantly associated with decreased endothelial cell density in our study included presence of an Ahmed glaucoma valve implant, previous intraocular surgery, age, duration of uveitis and history of corneal touch by the implant tube. Following a multivariate analysis, presence of an Ahmed glaucoma valve implant remained associated with decreased endothelial cell density independently from the effect of age. Importantly, within the implant group, the time interval following Ahmed glaucoma valve implantation was highly correlated with decreased corneal endothelial cell density (adjusted for age).
We conclude that Ahmed glaucoma valve implants in children with uveitic glaucoma have negative effect on endothelial cell density, and this effect is associated with the time interval following Ahmed glaucoma valve implantation. This is the first study investigating the effect of Ahmed glaucoma valve implants on endothelial cell density in children with uveitis which can therefore serve as a starting point for further prospective research in this direction.
Chapter 12 is a prospective study, investigating development of ocular complications in a cohort of 49 children within 1 year after hematopoietic stem cell transplantation (HSCT) by performing systemic ophthalmologic examinations of these children before HSCT, before leaving the HSCT unit and after 3, 6 and 12 months following HSCT. Specifically, we attempted to investigate the risk of development of viral uveitis in these immunocompromised children by performing additional examinations during systemic viral reactivations within the first year after HSCT. In our study 27% of children developed an ocular complication following HSCT. Systemic viral reactivations occurred in 41% of patients, however no new ocular abnormalities were observed during viral reactivation. New onset chorioretinal scars were observed after systemic viral reactivations in 2 patients (4%), however the etiologic relationship between the viruses and the scars could not be confirmed in this study due to the absence of active intraocular inflammation.
Ocular complications following HSCT included: dry eye syndrome (14%), (sub)retinal hemorrhage (12%), optic disc edema (6%), chorioretinal scars (4%), vitritis (2%) and increased intraocular pressure (2%). All ocular complications, except dry eye syndrome, were detected within the first 3 months following HSCT. Time point of development of dry eye syndrome in these children varied from 3 weeks to 12 months with the median of 5 months following HSCT. In most cases the symptoms of all complications were mild and self limiting. Children with a malignant indication for HSCT had higher risk for development of ocular complications compared to children with non-malignant diseases.
We conclude that ocular complications in pediatric HSCT-patients are common, although mostly mild. The risk of viral uveitis development during systemic viral reactivations is very low, however, potential risk of vision-threatening complications in this population can not be ruled out. Based on our findings, there is no strong evidence for routine ophthalmologic screening of these patients. However, we recommend performing ophthalmologic examination pre-HSCT, in case of systemic infectious reactivations and apparently when visual symptoms are recorded. Groups of patients at high risk of ocular complications, namely patients with malignancies and children of pre-amblyopic age, deserve higher awareness.