Greater DMARD use, improved clinical but not patient-reported outcomes in juvenile idiopathic arthritis: what are we missing?

This editorial refers to ‘A decade of progress in juvenile idiopathic athritis treatments and outcomes in Canada: results from ReACCh-Out and the CAPRI registry', Nguyen et al. 2024;63:SI173–9.

The last two decades have seen wider access and greater choice in biological DMARDs for children and young people with JIA [1]. As a result, better outcomes and experiences for these young people are expected.

In this issue of Rheumatology, Nguyen et al. aimed to quantify new DMARD starts and outcome attainment within 70 weeks of diagnosis in two Canadian JIA inception cohorts a decade apart: 2005–10 (Research in Arthritis in Canadian Children emphasizing Outcomes (ReACCh-Out) study, n = 1128) and 2017–21 (Canadian Alliance of Pediatric Rheumatology Investigators (CAPRI) registry, n = 721). The authors reported increases in both conventional DMARDs (cDMARDs) and biological DMARDs (bDMARDs) at 70 weeks (43–60% and 6–26%, respectively). While large increases in minimal (69–84%) and inactive disease (64–83%) attainment was reported over this timeframe, little improvement in patient-reported outcome followed [pain <1/10: 57–63%; good health-related quality of life (HRQoL): 52–54%].

The recent increase in bDMARD use is expected given the rapid increase in the number of bDMARDs approved by the Food and Drug Administration and European Medicines Agency for JIA. However, overall use of cDMARDs also increased, particularly in the oligoarticular and psoriatic subtypes [and with the exception of systemic JIA (sJIA)], perhaps reflecting an increasing interest in the window of opportunity and increasing recognition of the benefits of achieving remission within a shorter time period.

It is entirely expected that the increase in use of DMARDs leads to improvement in clinical outcomes. However, Nguyen et al. demonstrated little or no improvement in patient-reported outcome measures assessing quality of life, pain and parent global assessment over the last decade in their Canadian population, despite the improvements in clinical outcome. Rates of inactive disease using the Clinical ten joint count Juvenile Arthritis Disease Activity Score (cJADAS10; 50% for 2005–10; 69% for 2017–21) were lower than using the Wallace criteria (64% for 2005–10; 83% for 2017–21), likely reflecting the inclusion of parent global assessment within cJADAS10.

There is a well-recognized proportion of children and young people whose symptoms (pain, fatigue, functional ability) do not improve, or only improve marginally, when active inflammation appears resolved. Indeed, recent work has demonstrated that in the 3 years following diagnosis, one in five children have improvements in joint counts and physician assessment of disease, but maintain consistently poor patient-reported scores, often reflecting elevated pain and functional disability [2]. Nevertheless, one would still hope that the improved clinical outcomes led to at least some improvement in pain and quality of life for children and young people with JIA. Although sensitivity analyses lowering the threshold for ‘good’ HRQoL did not alter results, it would have been informative to understand trends in overall HRQoL and pain, including average values and spread. This may have uncovered clinically relevant changes in trends for poor outcome, rather than good outcome. However, the authors suggest several reasons for the disconnect between ‘good’ clinical and patient-reported outcomes, including factors resulting directly from increased bDMARD use (increased hospital visits; greater number of side effects), and it is likely multifactorial. It would be hoped that longer term follow-up studies may demonstrate improved quality of life (relating to less long-term damage) but this requires further evaluation. The availability of newer oral treatment with Janus kinase inhibitors, not included in this study, provides further hope that quality of life outcomes may be improved through avoiding need for injections. In addition, evolving input from across the multi-disciplinary team must also be interrogated to understand the changing landscape and impacts of treatments for JIA beyond DMARD therapies.

Nguyen et al. showed the increase in cDMARD and bDMARD use in this study did not correspond to a reduction in prednisolone use (16% vs 15% in overall cohort). Whilst the authors comment on overall relatively low use of CS in Canada compared with other countries, cumulative use was assessed in the study. Therefore, the use of lower doses and/or shorter treatment regimes would not be reflected in these figures. Morbidity relating to CS use remains of concern to clinicians and families and consideration of judicious use of CS within newer treatment approaches is essential.

The huge increase in cDMARD and bDMARD use along with the rapidly increasing spectrum of bDMARDs available leads to challenges in defining the best treatment approach. Current treatment approaches outside sJIA and ERA generally involve escalating through steroids, cDMARDs and a first (and often second) TNF inhibitor before other biological classes are considered [3–5]. This approach mirrors ‘treating to target’ in adult rheumatology [6], with a rapid step-up approach if a ‘target’ has not been met. There is also increasing interest in starting early combination therapy with cDMARD and bDMARD and a ‘step-down’ approach [7]. Whilst Nguyen et al. do not study treatment approaches their work clearly demonstrates that strategies of earlier cDMARD/bDMARDs use has transferred into clinical practice.

It is unlikely that a ‘one-size-fits-all’ treatment approach will be applicable in such a heterogeneous disease as JIA. Stratifying treatment depending on currently defined JIA classification is already recommended for sJIA with early IL-1/IL-6 blockade and ERA with early anti-TNF blockade. This change in practice is reflected in Nguyen et al.’s work, showing largest increase in use of bDMARDs in sJIA and ERA subtypes. However, more precise subgrouping and early targeted bDMARD therapy is likely to lead to ongoing improvement in clinical outcomes, and may improve patient outcomes through avoiding cycling through multiple treatment choices. More precise subgrouping of disease is currently underway with novel machine learning methods unlocking subgroups of disease with similar underlying features and outcome [2, 8]. While the current study by Nguyen et al. does not explore outcome from specific therapies, further work will need to explore which subgroups of JIA should be targeted with bDMARDs with specific modes of action. Such subgroups may be based on a combination of clinical criteria, biomarkers and patient-reported symptoms, with head-to-head comparisons of potential treatments needed across these groups for adaptation into treatment guidelines.

The increase in available bDMARDs is likely to continue, with adult-onset inflammatory rheumatic diseases already having far greater variety of biologic therapies to choose from [9]. Demonstrating similarities between child and adult-onset disease, such as through new Paediatric Rheumatology International Trials Organisation (PRINTO) classification criteria for JIA [10], may expedite drugs for children whose disease cannot be readily distinguished from their adult-onset counterparts. However, even in these groups, novel therapies are often tested in placebo-controlled trials. These trials can demonstrate an overall benefit of such therapies, but they do not show us benefit above the myriad of other biological therapies becoming available, nor in which subgroup of children (often subgroups within ILAR/PRINTO categories) the drugs would be most beneficial in. In the coming years, choice of biologic therapy will likely be a key challenge that must be overcome for optimal treatment success in JIA.

Data availability

There are no original data in this article.

Funding

S.J.W.S.-W. is supported by the Medical Research Council (MR/W027151/1).

Disclosure statement: The authors have declared no conflicts of interest.