Background
Juvenile Idiopathic Arthritis associated Uveitis (JIA-U), with an estimated incidence at 8.3%, represents its most frequent extra-articular manifestation and is the main cause of uveitis in pediatrics in developed countries [
1]. JIA-U is a serious and disabling sight-threatening disease, accounting for up to 10% of pathologies leading to blindness. Patients with JIA develop uveitis in approximately 50 and 90% within 3 months and 4 years, respectively, after the diagnosis of arthritis. Uveitis precedes the onset of arthritis in 2–7% of patients. Uveitis, usually affecting anterior chamber, is asymptomatic in the majority of JIA children, aside entesitis-related arthritis and late onset psoriatic arthritis JIA type, which may instead experience symptomatic uvea involvement. Any way, periodic ophthalmologic examination is mandatory required for screening [
2]. In JIA-U care, to define related and unified outcome measures are for sure quintessential to assess the effect of the treatment. Observations published worldwide on the outcomes of JIA-U are surprisingly diverse and different. The broad variety of outcome measures utilized makes the comparison of the disease course, risk for structural complications, levels of impairment in visual function, and responses to treatment quite difficult.
Of note, standardizing outcomes allows data comparison, and a core set of outcomes, agreed upon by both different involved researchers and patients, provides a common focus for interventional studies. Disease-specific and universally agreed upon outcomes are likely to reduce selective reporting and reporting bias [
3].
The Standardization of Uveitis Nomenclature (SUN) Working Group has provided a standardized nomenclature of uveitis, inflammation grading, and outcome measures [
4]. This instrument is of importance for the comparison of outcomes and trials, and for the optimal design of future trials.
Since specific outcome measures have not been established for JIA-U till 2010, a multinational interdisciplinary working group of ophthalmologists and paediatric rheumatologists gathered to consider this issue for the first time and published the first JIA-U specific proposal [
5].
They reviewed the recommendations of the SUN Working Group with special consideration given to their applicability for reporting clinical outcomes in studies of JIA-U. The proposed outcome measures were: measurement of uveitis activity (grade of cells in the AC, grade of flare in the AC, number of visits with active uveitis), visual acuity (measurement of visual acuity, amblyopia), development of structural complications (synechiae formation, cataract formation, macular edema, epiretinal membrane formation, band keratopathy, ocular hypertension and glaucoma, ocular hypotony, vitreous haze and cells), quality of life assessments, sparing of corticosteroids and other immunosuppressive drugs, surgery requirement, biomarkers.
To our knowledge no others studies have been conducted on this issue. Aim of our study was to summarize evidence regarding the current availability of outcome measures in JIA-U. We aimed to provide a complete, exhaustive, systematic literature review to describe the cutting edge of knowledge about outcome measures for use in JIA-U. The literature relating to outcome measures used in JIA-U studies in childhood and adolescence was reviewed.
Discussion
The aim of this study was to review the available evidence regarding outcome measures for JIA-U. Assessment of childhood JIA-U needs urgently to develop better outcome measures, as trial endpoints in clinical studies for development and licensing drug, as well as in driving treatment decisions in routine clinical practice. Nonetheless any attempt to judge the quality and the performance of the reported outcome measures has been acted, the present systematic review provides evidence that paediatric outcomes in JIA-U are particularly scarce and heterogeneous [
42]. Twelve outcome measures with distinct dimensions have been identified: disease activity (i.e. grade of flare, grade of cells); visual function performance (i.e. visual acuity); disease-associated tissue damage or complications (i.e. cataract, amblyopia); disease features (i.e. bilateral disease; uveitis subtype); laboratory features (i.e. biomarkers); JIA features (i.e. persistence of disease); quality of life (i.e. EYE-Q); management (i.e. use and sparing of oral corticosteroids and other immunosuppressive drugs; surgery requirement). It may be argued that the different distribution of the different entered outcome measures and dimensions may reflect a different point of view of the clinician, making treatment decisions primarily based on disease activity, visual acuity and disease damage. The impact of the disease on function and quality of life, probably the primary concern of parents and/or patients, seems indeed quite fare and neglected. Surely, all dimensions are inter-related, but the relationship among them is quite complex and mostly it may be hard to weight each one of them independently. For example, worsening visual acuity is associated with worsening patient-reported visual function, but the relationship is not necessarily direct, and the related correlation not really true.
Disease-associated tissue damage with and/or without complications is the most common used measure of disease activity, and as primary outcome measure is statistically associated with VA to the < 20/50 and 20/200 thresholds. Additionally, since structural damage may reflect previous and current disease activity, it can be considered a suitable candidate as complete outcome measure.
Current measures of disease activity are largely based on subjective clinical estimation. Of note, there are profound prognostic differences between the absence of cells and the presence of even a small number. Additionaly, since most clinical activity in JIA-U occurs within 3 grades of activity (0.5–2), aggregated counts may also lose sensitivity [
5]. The development of quantitative imaging in uveitis in adults is most established using OCT measurement of CMT to measure severity of ME [
22]. Uveitic ME is undoubtedly an outcome measure to be considered: it is the single most important reversible cause of sight-loss, and, since responsive to treatment, it can be the objective focus of most clinical trials. Additionally, ME measurement by OCT is the prototypic demonstration of quantitative imaging in uveitis. However the use of uveitis quantitative imaging, and their development as endpoints, is still in a relatively early phase. Indeed, in the description of a patient with ME, central macular thickness is only one feature; it does not aim to be a comprehensive assessment of the functional, psychological, or societal consequences of uveitis. Even so, it is an extremely useful parameter, which let us assess severity, progression and response to treatment at a tissue level in an objective way, which was never possible before the OCT use. In order to get the most objective and comparable measurement as possible, when clinical clues of ME are present, we suggest obtaining OCT imaging even in children [
11,
12]. Despite the need of further investigations, ME might be used, by itself, as uveitis outcome index apart from the other complications. Quantitative imaging approaches are now being developed and validated for other key inflammatory parameters such as anterior chamber cells, vitreous haze, retino-vascular leakage, and chorioretinal infiltrates [
36‐
39]. The development, validation, and adoption of sensitive and discriminatory measures of disease activity are an unmet need since its tremendous implications in drug development and routine clinical care. Interestingly, in adults, OCT-derived parameters have become standard outcome measures in clinic faster than they have become acceptable as trial endpoints by the regulatory authorities. Therefore, FDA have pointed out that while OCT-derived endpoints are ‘well-defined and reliable’ they do not directly measure how a patient ‘survives, feels or functions’.
Regarding VA, the majority of the studies [
7‐
13,
15‐
20] have identified VA as the main outcome measure for use in JIA-U. However, due its strong relationship with present and past ocular damage [
5], it remains a significant indicator of related disease eye damage, whilst its reliability as measure of activity remains debated. The assessment of uveitis activity by visual acuity its self, without data from slit lamp evaluation, seems inappropriate. Change over time in visual acuity, measured in LogMAR, mostly before and after treatment, seems indeed a more reliable method to assess uveitis activity regarding to visual acuity.
Visual function assessment, appropriately adjusted for age, relevant activities and reporting skills [
22,
27], can expand the utility of quality of life measures. The impact of altered visual function on quality of life may be objectified through the EYE-Q, an uveitis-related quality of life assessment questionnaires specific for children. Before Angeles-Han study [
27], such a paediatric-customized assessment was not available. This and a subsequent related study [
28] make this tool a disease-specific and universally agreed upon outcomes, that leads to reduce selective reporting and reporting bias.
There is no consensus whether the reduction of topical and systemic corticosteroids should be defined as desirable outcome measure. Despite this, many studies [
5,
7,
8,
10,
11,
13,
16,
17,
20] considered it as one of them. Of note, both the available RCTs entered this information, even though used as secondary outcome [
40,
41]. As matter of fact, tapering/stopping steroids, particularly systemic, represents an indirect measure of activity and good control of disease. However, we suggest that it is important information to be collected in trials, even if furthermore data analysis would be necessary.
Additionally, surgery may not be a primary outcome measure since indications for surgery and surgical methods used are definitively biased by the practice patterns of the treating ophthalmologist.
At the present, there is no consensus either about biomarker role, as outcome measure. Aside from ANA, the other biomarkers resulted from this systematic review have not been duplicated in larger and multicenter cohorts, thus limiting overall their potential role. Of course, their use in clinical practice as outcome measure seems currently not applicable. ANA may reflect a risk for a more severe course, but it does not seem a sensible outcome measure, related to a change over treatment. Further investigation are needed to proof the other mentioned biomarkers as part in the disease activity.
The development of complications and a blind visual outcome seems to be associated with the type of uveitis, especially panuveitis [
20] and bilateral disease [
7,
8,
17‐
19]. Therefore it can be suggested that type and disease features may be part of a composite outcome measures, but again the type of uveitis reflects the severity and not the activity of the disease. Persistent JIA oligo-arthritis results a statistically significant risk factor of having ocular complications [
8], but further studies would be necessary to assess if it could be a useful measurement of uveitis activity or severity.
While our systematic review does not attempt to provide the solution to overcome the heterogeneity in uveitis studies, it does provide an estimate of the scale of the problems and provides data to inform this important debate. The broad variation in outcome measures chosen by the investigators of eligible analysed articles is, in itself, an indicator that there is likely to be no easy answer to this issue. At the present, the strongest evidence available in treating JIA-U comes from SYCAMORE and ADJUVITE trials. However, as the broad heterogeneity reported in the rest of eligible studies, they failed to use the same measure for the primary end-point. Nonetheless achieving the same conclusions, thus making tremendous step-forwards in JIA-U clinical setting, they used different instruments –slit lamp versus laser flare photometry- to assess and measure, apparently, the same outcome variable of inflammation. In our opinion, this fact represents the paradigm of the need to harmonize the choice of the outcome measures.
Considering a composite score, we would however, underscore the potential risk that one may fail to detect therapeutic benefit due to the high level of “noise” introduced by joining in one unique score a too-wide range of clinical entities. Considering the “composite” feature of an outcome measure, while it may be argued that such outcome measures provide a comprehensive “holistic” assessment of the patient state, frequently including visual acuity, it is likely that the frequent use of composite measures in these series of reviewed is just simply driven by the lack of a single outcome measure suitable for all patients. In other words, lacking something better, of necessity, virtue.
In adults, it has been tackled the issue of heterogeneous outcome measures in clinical trials by establishing “core outcome sets” (COS). This approach provides a standardized set of outcome measures that are reported in all clinical trials of a condition under consideration, while still allowing the investigator discretion to choose his or her own primary or secondary outcome measures [
43]. The use of COS may enhance evidence synthesis by reducing heterogeneity, outcome-reporting bias and improving the statistical power of any meta-analysis. COS development is supported by a number of initiatives, such as COMET (Core Outcome Measures in Effectiveness Trials) and has been endorsed by the Cochrane Library, the GRADE (Grading of Recommendations Assessment, Development and Evaluation) working group, and the WHO [
43,
44]. In summary our systematic review formally surveys the heterogeneity present around outcome measures in recent and current articles related to JIA-U in children.
According the OMERACT filter criteria [
45], probably, none of the 12 selected outcomes, alone, can fulfil all the required variables of validity. Surgery, and structural complications represent more risk factors of worst outcome rather than a reliable and feasible measure. Oligo-arthiritis persistent JIA is of course a risk factors but not an outcome measure. Grade of cells or grade of flare, probably the most “true” outcome measure for active disease, are, however, lacking of discrimination and reliability, since they can differ in repeated measures, and may change if performed by different ophthalmologists. Visual acuity might be convergent with the other variables, but it cannot be really sensitive to change, since it cannot be responsive to small significant clinical changes. In addition, its reliability, as stability when measured repeatedly, can be affected in children. QOL, as any childhood measures of this type, require data acquisition on disease impact from parents and from caregivers and patients, both. Therefore it appears that multiple and composite domains seem necessary to capture all aspects of such complicating, sight and visual-threatening a disease.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.