Polyarticular JIA is defined as the JIA type where five or more joints are involved with either a rheumatoid factor-positive (on two consecutive tests) or a rheumatoid factor-negative further subdivision [1,2]. JIA is said to be polyarticular if at any time in the first 6 months from onset the cumulative number of involved joints reaches five. If this additive total effect only occurs after 6 months from onset then the type is changed from oligoarticular (up to four joints involved) to extended oligoarticular (five or more joints after 6 months). In addition, the JIA subtypes of systemic onset JIA and enthesitis related JIA, as well as psoriatic JIA, can also involve five or more joints. If the other associated defining features of any these particular groups are missing or overlooked, it is possible for a clinical case to be assigned to the incorrect group. Indeed, there may be some false divisions between the various subtypes if polyarticular arthritis is the main feature. However, there are still too few studies with large enough subgroups of patients with all the subtypes to know if the ILAR classification criteria for subtypes is truly clinically relevant from a therapeutic point of view in disease with polyarticular involvement. There is no single diagnostic test for any of the subtypes of JIA apart from a persistently raised rheumatoid factor level in the absence of connective tissues disorders.

The incidence of JIA is approximately 1:10,000, with a prevalence in the population of around 1:1,000 and about one-third of cases being polyarticular in nature [3,4]. This subtype may affect from five to most joints, can cause erosive disease, joint deformities with damage to growth plates, reduced mobility and general debility from chronic inflammation. It is associated with long-term disease activity into adult life in over 60% of cases. Treatment to achieve clinical remission is therefore vital, but the exact algorithm of care to achieve full remission has still not been established and there is considerable variation between centers based on clinical experience.

The ambitious primary outcome was set as clinical inactive disease [18] at 6 months with clinical remission on medication within 12 months of treatment also assessed (defined as a continuous 6 months of clinically inactive disease while receiving medication).

The strength of this study was that it attempted to use the 'window of opportunity' approach for long-lasting disease control from aggressive initial treatment [13]. Rather than simply looking for improvement using the usual American College of Rheumatology (ACR) Paediatric 70 response rates (70% improvement) the much more stringent primary outcome of clinically inactive disease was used. Assessment of two active treatments versus two concurrent placebos was also a strength, as it involved analyzing a treatment algorithm rather than looking at a single drug effect.

The main finding of the study was that after 12 months clinical remission on treatment was achieved in significantly more patients in the intensive treatment group than in the placebo and methotrexate group (P = 0.0534).

The randomization process did allocate more severe patients to the placebo arm in terms of more active joints and a significantly higher erythrocyte sedimentation rate (ESR) in the placebo arm, but this should really have served to accentuate the effect of the full treatment regime. This study demonstrated how effective injected weekly methotrexate can be given early in the disease course.

The relatively large numbers of patients discontinuing complex studies such as this provides a warning to investigators to include larger numbers of patients than the strict power calculations allow in order that the final numbers in each group remain meaningful. Larger numbers also allow for more subanalyses to be carried out in the explanation of results.