NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Kemper AR, Coeytaux R, Sanders GD, et al. Disease-Modifying Antirheumatic Drugs (DMARDs) in Children With Juvenile Idiopathic Arthritis (JIA) [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2011 Sep. (Comparative Effectiveness Reviews, No. 28.)
This publication is provided for historical reference only and the information may be out of date.
A succinct summary of the results of this review of the comparative benefits and harms of DMARDs for children with JIA is presented in the tables that follow. First, we provide an aggregated view of the strength of evidence and brief conclusions (Table 12). Next, we describe the nature and quality of the evidence for Key Questions 1, 2, and 4 in a format recommended by the GRADE committee (Tables 13-16).7 We then provide a tabular summary of the evidence for Key Question 5 (Table 17). Finally, we comment on the applicability of our findings.
GRADE summary tables were developed to describe the strength of evidence. For Key Question 1, separate GRADE summary tables are presented for the biologic and non-biologic DMARDs. We identified six randomized discontinuation trials that were conducted for the biologic DMARDs. Unlike RCTs or prospective cohort trials, randomized discontinuation trials evaluate the risk of worsening disease among those who initially responded to therapy. Because of this fundamental difference, we present a separate GRADE strength of evidence rating for the randomized discontinuation studies for each outcome. GRADE summary tables do not apply to Key Question 3 or Key Question 5. Findings from Key Question 3 are summarized in Tables 7 (Parts 1-3), under Results, and in Appendix E. Findings from Key Question 5 are summarized by outcome measure in Table 16.
Applicability was assessed for Key Question 1 only. Insufficient evidence was available to rate applicability for Key Questions 2 and 3, and Key Question 4 and 5 were not amenable to assessment of applicability. For Key Question 1, we assessed applicability as follows:
- Population: There was variation across studies in the definition of JIA and both duration and severity of illness, likely reflecting the range of patients seen in usual practice. However, six of the studies of the biologic DMARDs were randomized discontinuation studies, which include patients who have responded to the intervention.
- Intensity or quality of treatment: With the exception of methotrexate, the non-biologic DMARDs are less often used than the newer biologic DMARDs. The intensity of treatment in the studies of the biologic DMARDs is consistent with current recommendations.
- Choice of, and dosing of, the comparator: Methotrexate, a non-biologic DMARD, is a standard treatment for JIA. Six of the studies of the biologic DMARDs included methotrexate as a comparator; none of the studies of the non-biologic DMARDs included methotrexate as a comparator. The reasons for use or dose escalation of the comparator drugs were usually not described.
- Outcomes: The most commonly reported outcome measures were laboratory indicators of inflammation (e.g., ESR) or the ACR Pediatric 30. The ACR Pediatric 30 blends several relevant outcomes (e.g., active joint count, functional status, pain), but is not normally used in daily clinical practice. As described for Key Question 5, new instruments to better assess response to therapy and changes in health-related quality of life are in development.
- Timing of followup: Five of the studies of biologic DMARDs and five of the studies of non-biologic DMARDs actively followed subjects for more than 6 months. This would allow sufficient time to detect clinically important benefits and may be long enough to identify important harms.
Tables
Table 12Summary of the evidence on comparative effectiveness and harms of DMARDs for childhood JIA
| Key question | Strength of evidence | Conclusions |
|---|---|---|
| 1. In children with JIA, does treatment with DMARDs, compared to conventional treatment: | ||
| a. Improve laboratory measures of inflammation? | Low | Trials of DMARDs usually report changes in laboratory measures of inflammation (e.g., ESR). However, ESR is inconsistently associated with treatment. This conclusion is based on 14 studies of 1060 subjects. |
| b. Improve radiological progression? | Insufficient | Insufficient data are available to evaluate the impact of DMARDs on radiological progression. Only one cohort study of 63 subjects reported data on radiological progression. |
| c. Improve symptoms? | Moderate | Among children who have responded to a biologic DMARD, randomized discontinuation trials show that continued treatment for from 4 months to 2 years decrease the risk of having a flare (RR 0.46, 95% CI 0.36 to 0.60). This conclusion is based on four studies of 322 subjects. Among the non-biologic DMARDs, there is some evidence that methotrexate is superior to conventional therapy and oral corticosteroids, based on two randomized trials of 215 subjects. |
| d. Improve health status? | Low | Changes in health status were reported in 12 studies involving 927 subjects. Health status improved inconsistently with treatment with DMARDs. |
| 2. In children with JIA, what are the comparative effects of DMARDs on: | ||
| a. Laboratory measures of inflammation? | Low | Trials of DMARDs usually report changes in laboratory measures of inflammation (e.g., ESR). However, ESR is inconsistently associated with treatment. This is based on 4 RCTs of 448 subjects and 1 cohort study of 72 subjects. |
| b. Radiological progression? | Insufficient | No study addressed radiologic progression. |
| c. Symptoms? | Low | The non-biologic DMARDs that were compared directly (penicillamine vs. hydroxychloroquine, sulfasalazine vs. hydroxychloroquine, and leflunomide vs. methotrexate) had similar efficacy. Changes in symptoms between the treatment arms were not measured with significant precision to detect a difference. This is based on 4 RCTs of 448 subjects and 1 cohort study of 72 subjects. One poor-quality RCT of 94 subjects found that etanercept was similar to infliximab. |
| d. Health status? | Low | The non-biologic DMARDs that were compared directly (penicillamine vs. hydroxychloroquine, sulfasalazine vs. hydroxychloroquine, and leflunomide vs. methotrexate) had similar efficacy. Changes in health status between the treatment arms were not measured with significant precision to detect a difference. This is based on 4 RCTs of 448 subjects and 1 cohort study of 72 subjects. One poor quality RCT of 94 subjects found that etanercept was similar to infliximab. |
| 3. In children with JIA, does the rate and type of adverse events differ between: | ||
| a. The various DMARDs? | Insufficient | Three RCTs directly compared two DMARDs; two compared penicillamine to hydroxychloroquine, and one compared leflunomide to methotrexate. The rate and type of adverse events did not differ between treatment groups in these studies. High variability across studies in the ascertainment and reporting of adverse events preclude valid comparisons of the rate and type of adverse events among the various DMARDs. Recently published studies of adverse event reporting databases provide indirect evidence that suggests a possible relationship between cancer and exposure to tumor necrosis factor α blockers. |
| b. DMARDs and conventional treatment with or without methotrexate? | Insufficient | No RCT directly compared a DMARD to conventional treatment. Thirteen trials directly compared a DMARD to placebo. The rate and type of adverse events were generally similar between intervention and placebo groups, with the notable exceptions of infliximab plus methotrexate being associated with more serious adverse events (32% vs. 5% over differing lengths of followup), and methotrexate being associated with higher rates of laboratory abnormalities (35% vs. 13%). |
| 4. How do the efficacy, effectiveness, safety, and adverse effects of treatment with DMARDs differ among the various categories of JIA? | Insufficient | Only one study – an RCT of methotrexate versus placebo in which each group could also receive oral corticosteroids, intra-articular corticosteroids, and NSAIDs – evaluated efficacy by JIA category. No difference was found among those with extended oligoarticular JIA (n = 43) and systemic JIA (n = 45). We did not identify any studies that provide reliable information on the comparative safety or rates or types of adverse events among the various categories of JIA. |
| 5. What are the validity, reliability, responsiveness, and feasibility of the clinical outcome measures for childhood JIA that are commonly used in clinical trials or within the clinical practice setting? | Insufficient | Most of the studies examining the psychometric properties of the instruments used in JIA were fair-quality cross-sectional or longitudinal non-randomized controlled trials. No one instrument or outcomes measure appeared superior in measuring disease activity or functional status. The current response criteria of the ACR Pediatric 30, a composite measure that 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 depending on extent of disease at baseline. The ACR Pediatric 30 is also a relative measure of disease activity and not a measure of current disease state. |
Abbreviations: ACR = American College of Rheumatology; CHAQ = Childhood Health Assessment Questionnaire; CI = confidence interval; DMARD(s) = disease-modifying antirheumatic drug(s); ESR = erythrocyte sedimentation rate; JIA = juvenile idiopathic arthritis; NSAIDs = non-steroidal anti-inflammatory drugs; RCT = randomized controlled trial; RR = risk ratio
Table 13GRADE summary table for key question 1—biologic DMARDs
| Domains pertaining to strength of evidence | ||||||
|---|---|---|---|---|---|---|
| Number of studies; subjects | Design | Risk of bias/study quality | Consistency | Directness | Precision | Strength of evidence |
| Laboratory measures of inflammation | Low | |||||
| 1; 31 | RCT | Poor (high dropout rate) | NA | Direct | Imprecise | - |
| 4; 322 | Randomized discontinuation trials | Poor to Good | Inconsistent | Direct | Precise | |
| 1; 20 | Cohort | Poor (open-label) | NA | Direct | Imprecise | - |
| Radiologic progression | Insufficient | |||||
| 0; 0 | RCT | - | - | - | - | - |
| 0; 0 | Randomized discontinuation trials | - | - | - | - | - |
| 0;0 | Cohort | -- | -- | -- | -- | - |
| Symptoms | Moderate | |||||
| 3; 165 | RCT | Fair (one study had high dropout rate) | Consistent | Direct | Imprecise | - |
| 6; 341 | Randomized discontinuation trials | Good | Consistent | Direct | Precise | |
| 0; 0 | Cohort | - | - | - | - | - |
| Health status | Low | |||||
| 1; 31 | RCT | Poor (high dropout rate) | NA | Direct | Imprecise | - |
| 4; 272 | Randomized discontinuation trials | Good | Inconsistent | Direct | Imprecise | |
| 0; 0 | Cohort | - | - | - | - | - |
Abbreviations: DMARDs = disease-modifying antirheumatic drugs; NA = not applicable; RCT = randomized controlled trial
Table 14GRADE summary table for key question 1—non-biologic DMARDs
| Domains pertaining to strength of evidence | ||||||
|---|---|---|---|---|---|---|
| Number of studies; subjects | Design | Risk of bias/study quality | Consistency | Directness | Precision | Strength of evidence |
| Laboratory measures of inflammation | Low | |||||
| 7; 624 | RCT | Fair (open-label or unblinded) | Inconsistent | Direct | Imprecise | - |
| 1; 63 | Cohort | Poor (open-label) | NA | Direct | Imprecise | - |
| Radiologic progression | Low | |||||
| 1; 69 | RCT | Good | NA | Direct | Imprecise | - |
| 0; 0 | Cohort | - | - | - | - | - |
| Symptoms | Moderate (MTX) Low (other non-biologic) | |||||
| 7; 624 | RCT | Fair (open-label or unblinded) | Consistent | Direct | Imprecise | - |
| 1; 63 | Cohort | Poor (open-label) | NA | Indirect | Imprecise | - |
| Health status | Moderate (MTX) Low (other non-biologic) | |||||
| 7; 624 | RCT | Fair (open-label or unblinded) | Consistent | Direct | Imprecise | - |
| 1; 63 | Cohort | Poor (open-label) | NA | Indirect | Imprecise | - |
Abbreviations: DMARDs = disease-modifying antirheumatic drugs; MTX = methotrexate; NA = not applicable; RCT = randomized controlled trial
Table 15GRADE summary table for key question 2
| Domains pertaining to strength of evidence | ||||||
|---|---|---|---|---|---|---|
| Number of studies; subjects | Design | Risk of bias/study quality | Consistency | Directness | Precision | Strength of evidence |
| Laboratory measures of inflammation | Low | |||||
| 4; 448 | RCT | Fair (some studies with incomplete blinding) | Inconsistent | Direct | Imprecise | - |
| 1; 72 | Cohort | Poor (insufficient data) | NA | Direct | Imprecise | - |
| Radiologic progression | Insufficient | |||||
| 0; 0 | RCT | - | - | - | - | - |
| 0; 0 | Cohort | - | - | - | - | - |
| Symptoms | Low | |||||
| 4; 448 | RCT | Fair (some studies with incomplete blinding) | Inconsistent | Direct | Imprecise | - |
| 1; 72 | Cohort | Poor (insufficient data) | NA | Direct | Imprecise | - |
| Health status | Low | |||||
| 4; 448 | RCT | Fair (some studies with incomplete blinding) | Inconsistent | Direct | Imprecise | - |
| 1; 72 | Cohort | Poor (insufficient data) | NA | Direct | Imprecise | - |
Abbreviations: NA = not applicable; RCT = randomized controlled trial
Table 16GRADE summary table for key question 4
| Domains pertaining to strength of evidence | ||||||
|---|---|---|---|---|---|---|
| Number of studies; subjects | Design | Risk of bias/study quality | Consistency | Directness | Precision | Strength of evidence |
| Laboratory measures of inflammation | Insufficient | |||||
| 1; 88 | RCT | Good | NA | Direct | Imprecise | - |
| 0; 0 | Cohort | - | - | - | - | - |
| Radiologic progression | Insufficient | |||||
| 0; 0 | RCT | - | - | - | - | - |
| 0; 0 | Cohort | - | - | - | - | - |
| Symptoms | Insufficient | |||||
| 1; 88 | RCT | Good | NA | Direct | Imprecise | - |
| 0; 0 | Cohort | - | - | - | - | - |
| Health status | Insufficient | |||||
| 1; 88 | RCT | Good | NA | Direct | Imprecise | - |
| 0; 0 | Cohort | - | - | - | - | - |
Abbreviations: NA = not applicable; RCT = randomized controlled trial
Table 17Evidence summary table for key question 5
| Number of studies; subjects | Evidence summary |
|---|---|
| Active joint count | |
| 12; 8064 | Shows high responsiveness and moderate correlation with other measures of disease activity and functional status, but poor correlation with psychosocial aspects of quality of life. Lack of inter-rater reliability data. |
| Physician global assessment of disease activity | |
| 12; 8668 | Moderate correlations with measures of disease activity, the CHAQ, and quality-of-life measures. Responsiveness difficult to measure, as often compared to other physician measures of disease activity. No data on inter-rater reliability between providers. |
| Parent/patient global assessment of well-being | |
| 8; 8182 | Moderate correlations with other measures of disease activity, the CHAQ, and physical aspects of the quality of life measures, but poor correlation with psychosocial aspects of the CHQ. Moderate responsiveness and discriminate abilities. |
| CHAQ | |
| 23; 13,374 | Most commonly reported outcome measure with strong reliability, including moderate to strong inter-rater reliability between parent and child. Moderate correlations with other measures of disease activity, but poor responsiveness, which varies depending on how extensive the arthritis is at baseline (ceiling effect). |
| CHQ | |
| 5; 4687 | Limited data for JIA population. Moderate to strong parent to child inter-rater reliability for physical components, but lower for psychosocial aspects. Similarly, moderate correlations with measures disease activity, and the CHAQ for the physical component of the CHQ, but poor for the psychosocial domains. Poor responsiveness. |
| PedsQL/PedsQL-RM | |
| 2; 173 | Insufficient data in JIA populations to evaluate fully. Moderate to strong parent to child inter-rater reliability for physical components, but lower for psychosocial aspects. |
Abbreviations: CHAQ = Childhood Health Assessment Questionnaire; CHQ = Child Health Questionnaire; JIA = juvenile idiopathic arthritis; PedsQL = Pediatric Quality of Life Inventory; PedsQL-RM = Pediatric Quality of Life Inventory-Rheumatology Module