• uveitis
• juvenile idiopathic arthritis
• children

The inflammatory joint disease currently referred to as juvenile idiopathic arthritis (JIA) has endured many labels including Still's disease,¹ juvenile chronic arthropathy,² and juvenile rheumatoid arthritis.³ From the work of Edelsten and colleagues published in this issue of the BJO (p 51), we learn that the ocular prognosis for this disease is as diverse as the names for the entity itself. Importantly, we also discover that certain clinical factors, most significantly, the severity of uveitis at onset, can help us predict prognosis.

Establishing a diagnosis is a means to an end that will guide both the physician and the patient in gauging prognosis and choosing among treatment options. But how is one to interpret a diagnosis that has a broad spectrum of outcomes? How does a clinician offer advice for a disease such as uveitis if one patient might suffer 48 hours of mild inflammation, followed by complete resolution, whereas another patient might endure 48 years of chronic inflammation and consequent blindness? Can research help recognise clues that will guide physician and patient in devising a treatment protocol such that “the punishment,” otherwise known as the treatment, truly fits “the crime,” also known as the disease.

The outcome of a disease is determined by many factors besides what is implied by the diagnosis itself. Both patients and physicians believe that treatment is a major determinant of outcome. The gold standard in assessing the effect of therapy is the randomised controlled clinical trial. In the case of uveitis, the clinician can rarely rely on this gold standard because it has seldom been undertaken. Since we believe that treatment is important, we naturally believe that compliance with treatment is important. Sex and other genetically determined factors, socioeconomic status, education, and the results of laboratory tests are additional factors that may influence or predict prognosis. The microarray gene chip technology that allows an investigator to determine expression of thousands of genes simultaneously promises to revolutionise the ability to determine prognosis. Gene array can be used to show that “one disease” based on clinical features and histopathology is more than one disease based on gene expression, as exemplified by diffuse large B cell lymphoma.⁴ Furthermore, gene array promises to identify host genetic factors that may influence clinical outcome, just as the presence of sickle cell trait influences the course of malaria.

The report by Edelsten and colleagues is especially laudable for its approach. To derive its valuable conclusions, the authors needed a large database, recording careful observations over nearly two decades. A provocative conclusion is that the prognosis for the eye disease associated with JIA has improved over the past 17 years. Why should this be? It is tempting to congratulate ourselves that more aggressive treatment, such as with methotrexate, has been responsible. But one must remember that we do not have rigorous proof to demonstrate the efficacy of methotrexate for this or any other form of uveitis. It is tempting to hope that the new systemically administered inhibitors of tumour necrosis factor will improve prognosis even further, but early reports have been somewhat contradictory about the role for these drugs in the treatment of inflammatory eye disease.⁵

“The severity of uveitis . . . can help us predict prognosis of juvenile idiopathic arthritis”

Sometimes improved prognosis is a mirage induced by early detection. The prognosis for uveitis associated with JIA would mistakenly appear improved if more vigilant screening in 1998 allowed the ophthalmologist to diagnose the child with a mild cellular response in the anterior chamber that was destined to disappear regardless of treatment, while this same scenario in 1982 did not come to the attention of an ophthalmologist. Shifting diagnostic criteria may also explain an apparent change in disease prognosis over time. The term “juvenile idiopathic arthritis,” was coined by the International League Against Rheumatism Task Force for Classification Criteria in 1995, when, recognising the inadequacy of previous terminology, it developed a system for classification of childhood arthritis.⁶ This umbrella term acknowledges that different forms of arthritis are subsumed within it. Despite revision in 1997,⁷ the system continues to receive criticism and further revision is suggested.⁸ Further, it is clear that clinicians and scientists remain confused in the classification. For example, one might argue that in the study by Edelsten and co-workers, patients classified as “non-standard” because they had uveitis associated with psoriatic arthritis did, in fact, belong to the “standard” group, if psoriatic arthritis is considered a form of JIA as recommended by current classification criteria.⁷

Edelsten and colleagues appropriately recognise the limitations of their study. Extrapolations to an individual practice must be made cautiously if the physician has a different threshold for initiating methotrexate, if follow up visits are scheduled less frequently, and if the definition of “severe disease” varies from that used in this report. Prognostic information is always based on probability. Patients are rarely interested to learn that the probability of blindness is 6%, if the unfortunate individual in the examination chair today is that unlucky one person out of 17.

With the publication of this report by Edelsten and co-workers, we know more about prognosis of JIA associated uveitis than we ever have before. We also know that we have much more to learn about this eye disease.