Efficacy of DMARDs

Although DMARDs have improved health outcomes for children with JIA, few data are available to evaluate the comparative effectiveness of either specific DMARDs or general classes of DMARDs (e.g., non-biologic vs. biologic, or by mechanism of action). Not surprisingly, methotrexate, the oldest of the DMARDs used for children with JIA, is the most studied DMARD. Because it is frequently used, methotrexate is often considered to be a component of conventional treatment both in clinical care and in research studies. Good-quality studies support the efficacy of methotrexate. The paucity of evidence precludes head-to-head comparisons of the newer DMARDs against each other, with or without methotrexate.

Research on the effectiveness of treatments for JIA is challenging because it is a rare condition that includes multiple categories, which could potentially respond differentially to therapy. Furthermore, the health impact of JIA fluctuates over time. Therefore, trials require large sample sizes with long follow-up periods.

Developing a summary estimate of effectiveness of the DMARDs is challenging because there is:

• Heterogeneity in the study population. Changes in the definition of JIA (e.g., JRA, JCA) may have led to the inclusion in studies of individuals who may respond differently to treatments. Similarly, differences by disease category (e.g., polyarticular, pauciarticular, systemic) might lead to different conclusions about the effectiveness of treatment. This may be particular relevant when examining response rates for systemic JIA given its similarities to auto-inflammatory diseases.
• Variation in comparators. Over time, the standard of care for JIA has changed. For example, relatively recent studies of biologic DMARDs often allow methotrexate, a DMARD, in the comparator group, while older studies do not include methotrexate in the comparator groups. Some older studies included systemic corticosteroids as a comparator.
• Outcome measures vary across the studies and are sometimes incompletely described. For example, some studies report the percentage improvement from baseline without providing baseline data or an estimate of variability. Among the six randomized discontinuation trials, for example, four reported laboratory measures of inflammation,^24,26,29,30 four reported whether a flare occurred,^24-26,30 three reported active joint count,^24,28,29 and four reported quality of life as measured by CHAQ.^24,26,29,30 Of those that reported the CHAQ score, one²⁶ reported only the percentage change from baseline without the absolute value or measure of dispersion (e.g., range, standard deviation), and two^29,30 gave average values without measures of dispersion.

Future trials in this domain should consider:

• The challenge of the appropriate comparator. Trials are needed to evaluate the effectiveness of DMARDs compared to conventional therapy as well as against other DMARDs. Defining conventional therapy is challenging because it evolves with advances in the field. Factorial designs involving multiple treatments are a potential solution. Patient-level meta-analysis, pre-planned across different trials, may also help address this issue.
• The issue of treatment-by-category interaction. To fully explore comparative effectiveness, larger studies will be needed. In addition, patient-level meta-analysis may help address this challenge.
• The need for study populations who are representative of typical patients with JIA. Subjects from the studies included in this review were identified through specialty clinics, which is appropriate for rare conditions. However, baseline characteristics varied. Studies should be designed to reflect the comparative effectiveness for typical subjects at various points along the disease spectrum (e.g., at presentation, after failing conventional treatment). Furthermore, most evidence regarding treatment impact is from patients with poly-articular forms of JIA with fewer data on persistent oligoarthritis, a common type of JIA.
• The variable course of JIA. Trials that evaluate the efficacy of treatment should be sufficiently long with frequent assessment of health status to capture the natural variability of the disease course.
• The need for standardized outcome measures. In addition to providing a better understanding of the impact of the trial, standardized outcome measures would facilitate high-quality meta-analysis.
• The need for standardized definitions for, and systematic ascertainment and reporting of, adverse events possibly associated with therapeutic interventions in the treatment of JIA.
• The impact of DMARDs on the specific health conditions associated with JIA, including uveitis and macrophage activation syndrome.

Study designs other than RCTs will be important in understanding the role of DMARDs in JIA. Randomized discontinuation trials have helped to define the risk of flare in patients who respond to a particular DMARD. Large cohort studies will be important for evaluating the risk of adverse events associated with DMARDs. Such studies could also be important for better characterizing long-term outcomes in JIA. Disease registries could be an important strategy for developing such cohort studies. In addition, such registries could provide indirect evidence about the benefits of treatment.

Safety of DMARDs

Few high-quality data are available regarding the adverse events associated with DMARDs. Because JIA is a chronic illness, understanding the long-term effects of these drugs is critical. One solution to evaluating risk would be to develop registries for DMARDs when used for childhood JIA. Understanding such risk will also provide information about the sequence in which these drugs should be used for difficult-to-treat JIA, or the impact of using multiple drugs.

Our findings suggest that short-term mortality rates associated with DMARDs are very low (we identified only a single patient among several thousand treated who died shortly after receiving a DMARD). The incidence of malignancies during a short course of DMARD treatment also appears to be very low. Simard et al.⁸⁷ have demonstrated that the incidence of malignancies among children with JIA appears to be higher, in general, over the past 20 years, than in the two decades prior to the advent and utilization of biologic DMARDs. These data are, however, confounded by numerous factors, most notably the frequent concurrent use of immunosuppressants in children undergoing treatment for JIA.

The available evidence is inadequate to determine whether the rates and types of adverse events differ between the various DMARDs or between DMARDs and conventional treatment. The findings from RCTs do not reveal a clear pattern pertaining to adverse events associated with the treatment of JIA with DMARDs compared to placebo. Our wider review of the adverse events literature revealed marked differences in the rate and type of adverse event by DMARD, but these findings should be interpreted with caution for several reasons, including: variable definitions of adverse events across studies; non-systematic methods of ascertaining adverse events; nearly universal lack of standard reporting of serious adverse events; a predominance of case reports and uncontrolled series; small sample sizes in most series and RCTs; a limited number of studies for many individual DMARDs; and frequent use of multiple medications and other co-interventions.

Impact of DMARDs on Health Outcomes

Our findings suggest the need for better clinical outcomes measures that are responsive to change across the full spectrum of disease severity. Consistent use of such outcomes measures would facilitate comparative effectiveness research.

The heterogeneity in disease severity and the broad impact of the disease on both physical and psychosocial aspects of children's lives make it difficult to accurately assess children using one instrument or measure. Given the complex nature of JIA, with the potential for both chronic and acute functional limitations and pain, it is difficult to find one tool or instrument that can be responsive to all the facets of disease. Efforts to develop a more standardized composite measure which could incorporate articular indices, severity, and a broader assessment of functional limitations and psychosocial impact would be useful to better differentiate levels of disease activity and overall impact of disease. The current response criteria of the ACR Pediatric 30 definition of improvement, a composite measure which includes articular indices, functional status, laboratory measure, and global assessments, takes into account the various measures most commonly used. However, the responsiveness of several of these measures, including functional status and parent/patient global assessment, are poor to moderate, and they may not adequately reflect changes in disease state. Furthermore, given that the ACR Pediatric 30 is a relative measure of disease activity, the impact of JIA category on percent improvement is unclear, as certain instruments, such as the CHAQ, appear to have differential responsiveness by extent of disease at baseline. The ACR Pediatric 30 is also a relative measure and not a measure of current disease state.

The outcomes measured and reported should be tailored to the questions a study is investigating. If the question is whether a new therapy reduces active arthritis, utilizing outcome measures that focus on factors felt to reflect active disease, such as the JADAS, rather than overall disease status (active disease, disease damage, functional status, and quality of life) may prove particularly useful in more accurately addressing articular response to treatment. In addition, focusing on the most responsive outcome measures to assess treatment effects would enhance our ability to detect promising new treatments. Reporting functional status and quality of life are also important, especially given that many of our current treatments are delivered by infusion or injection and have varying side effects that can negatively impact one's quality of life. However, by reporting articular measures separately from functional status and quality of life measures, one may actually improve our understanding of the overall impact of JIA, including the influence of active arthritis, articular damage, and various treatment regimens.