Introduction
Juvenile idiopathic arthritis (JIA) is a chronic rheumatic disease of childhood, characterized by persistent joint inflammation and functional impairments, including reduced mobility, muscle strength, and cardiorespiratory fitness [1]. Children with JIA often experience prolonged physical inactivity due to pain and fatigue, leading to exacerbated health issues such as obesity and cardiovascular disease, ultimately affecting their functional capacity and health-related quality of life (HRQoL) [2, 3].
Despite significant advancements in pharmacologic therapies, particularly biologic disease-modifying antirheumatic drugs (DMARDs), many patients continue to experience functional impairments, such as muscle weakness, reduced aerobic capacity, and compromised physical mobility [4, 5]. Additionally, chronic inflammation and prolonged corticosteroid use further elevate the risk of musculoskeletal deterioration and cardiovascular morbidity, thereby exacerbating physical inactivity and predisposing affected children to long-term disability [6, 7].
Exercise interventions have emerged as a promising adjunctive therapy capable of mitigating these systemic and functional impairments [8]. Growing evidence indicates that appropriately structured exercise regimens, including resistance training, aerobic conditioning, aquatic exercises, and balance-focused interventions, may enhance muscle strength, cardiorespiratory endurance, and HRQoL in children with JIA [9–11]. However, existing randomized controlled trials (RCTs) have yielded heterogeneous and sometimes conflicting results, possibly due to methodological variability in the type of exercise, intervention duration, delivery mode, and outcome measures employed.
While several trials have reported improvements in physical function and selected cardiorespiratory outcomes following exercise interventions, others have found no significant effects, particularly with respect to quality of life or aerobic capacity [12, 13]. These mixed findings highlight the need for a comprehensive synthesis of the available evidence.
Given the growing body of randomized trials and the persistent clinical need for robust, evidence-based rehabilitation strategies, a systematic synthesis of the literature examining the effects of exercise interventions on physical function, cardiorespiratory fitness, and HRQoL is both timely and critical. Accordingly, we perform this meta-analysis to comprehensively evaluate the efficacy of exercise-based interventions targeting these critical health domains in pediatric patients with JIA. The findings are expected to inform clinical practice, guide rehabilitation protocols, and optimize non-pharmacologic management strategies for children with JIA.
Methods
Protocol and registration
This meta-analysis was registered in PROSPERO (CRD420251048687) and conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [14]. The original study protocol was provided as Supplementary file S1.
Search strategy
Two authors independently performed a comprehensive search in MEDLINE, Cochrane Library, Embase, Web of Science, and PEDro from database inception up to 1, October 2025 to identify relevant randomized controlled trials evaluating exercise interventions in pediatric patients with juvenile idiopathic arthritis. A structured search strategy combining medical subject headings (MeSH) and free-text keywords related to “physical function,” “juvenile idiopathic arthritis,” “exercise,” “cardiorespiratory fitness,” and “randomized controlled trial” was used. No restrictions were applied regarding language or publication date.
Additionally, reference lists of included studies and relevant systematic reviews were manually screened to identify further eligible RCTs. Detailed search strategies were summarized in Supplementary file S2.”
Eligibility criteria
Studies were included based on the following criteria: (1) Population: Studies enrolling children or adolescents under 18 years of age with a clinical diagnosis of JIA were eligible. The age criterion referred to age at intervention, while all participants met the standard definition of JIA with disease onset before age 16. (2) Intervention: Structured exercise interventions, including but not limited to aerobic, resistance, balance-proprioceptive, core stability, and aquatic training programs.
(3) Comparison: Control groups receiving usual care or non-exercise interventions. (4) Outcomes: Reporting at least one primary outcome related to physical function (e.g., child health assessment questionnaire [CHAQ], 6-minute walk test [6MWT]), cardiorespiratory fitness (e.g., peak oxygen consumption [VO₂peak], maximum heart rate [HRmax]), or HRQoL. (5) Study Design: randomized controlled trials.
Exclusion criteria included: (1) Studies not reporting outcomes relevant to physical function, cardiorespiratory fitness, or HRQoL. (2) Non-randomized trials, observational studies, case series, or case reports.
Study selection and data extraction
Two authors independently screened titles and abstracts retrieved from the electronic search. Full-text articles of potentially eligible trials were retrieved and independently assessed according to predefined eligibility criteria. Relevant data were extracted independently by two authors using a standardized extraction form, including authors, year of publication, participant characteristics (age, male/female ratio, JIA subtype), exercise intervention details (exercise modality, frequency, duration), control intervention and outcome measures (physical function, cardiorespiratory fitness, HRQoL).
Discrepancies were resolved through discussion and consensus, with involvement of a third author if needed. The final dataset was reviewed for consistency and completeness prior to analysis.
Quality assessment
The methodological quality of included studies was independently assessed by two authors using the Cochrane tool to assess the risk of bias in randomized trials (RoB 2) [15]. The following domains were evaluated: 1) randomization process; 2) deviations from intended interventions; 3) missing outcome data; 4) measurement of the outcome; 5) selection of the reported results. Disagreements were resolved through consensus, and unresolved differences were adjudicated by a third author.
Statistical analysis
Meta-analysis was conducted using Stata version 16.0 (College Station, Texas, USA). For continuous outcomes measured on the same scale, the weighted mean difference (WMD) with 95% confidence intervals (CIs) was calculated. Standardized mean differences (SMDs) with 95% CIs were computed for outcomes measured on different scales.
The pooled effect estimates were derived from the differences between baseline and post-intervention values, along with their standard deviations (SDs) and the number of participants. Additionally, a correlation coefficient of 0.5 was applied to adjust for the variability in standard deviations before and after the intervention [16]. Statistical heterogeneity among studies was assessed using the I² statistic [17].
A random-effects model was applied if significant heterogeneity was present (I² ≥50%), otherwise, a fixed-effect model was utilized (I² <50%) [18]. To complement statistical significance with estimates of clinical relevance, we calculated effect sizes (Cohen’s d) for each continuous outcome. For outcomes reported as SMD, the SMD was considered equivalent to Cohen’s d.
For outcomes reported as WMD, we estimated effect sizes using pooled standard deviations derived from the included studies where available. Thresholds for interpretation followed Cohen’s conventional criteria: small (d = 0.20), medium (d = 0.50), and large (d = 0.80) [19]. Statistical significance was defined as p < 0.05.
Results
Study selection and characteristics
The initial search yielded 727 records. After removing 216 duplicates, 511 titles and abstracts were screened, of which 471 were excluded. Forty full-text articles were retrieved and assessed for eligibility, and 32 studies were excluded for the following reasons: non-RCT design (n = 16), lack of outcomes of interest (n = 7), unavailable data despite efforts to contact authors (n = 4), and comparisons involving different exercise interventions (n = 3).
Reasons for exclusion of potentially relevant studies were summarized in Supplementary file S3. Finally, eight randomized controlled trials were included in the meta-analysis [9, 20–26]. The study selection process was summarized in Fig.
1.
Fig. 1Flow diagram of the study selection process based on PRISMA guideline
The included trials published between 2003 and 2024, comprising a total of 405 children and adolescents with juvenile idiopathic arthritis. Participant ages ranged from 8.66 to 15.70 years, with a total of 132 males and 273 females included across all studies, reflecting the known sex distribution in JIA populations. Regarding the JIA subtypes, polyarticular JIA was included in all studies, while other subtypes were represented to varying extents.
The studies investigated diverse exercise modalities, including aerobic training (n = 3) [9, 22, 23], Pilates or core stability exercises (n = 2) [20, 21], aquatic training (n = 1) [25], home-based land exercise (n = 1) [26], and inspiratory muscle training (n = 1) [24]. Most interventions were performed 3 times per week, with session durations ranging from 20 to 60 min, over 8 to 20 weeks. Control conditions included conventional physical therapy, non-exercise comparators, wait-list, or assessment-only.
Outcomes commonly assessed were physical function (e.g., CHAQ, 6MWT), cardiorespiratory fitness (e.g., VO₂peak, HRmax), and health-related quality of life. All included trials reported at least one primary outcome relevant to the meta-analysis. Detailed study characteristics were presented in Table 1.
Table 1The characteristics of eligible studiesAuthor/yearParticipants (M/F)Age, mean ± SDyearJIA typeMedication useExercise intervention detailsControl conditionOutcomesModalityTime/sessionFrequencyDuration1Azab et al., 2022(20)IG: 7/12CG: 4/14IG: 12.32 ± 1.67CG: 11.56 ± 1.46PolyarticularStable dosage of medicationPilates + stretching, ROM and weight bearing activities25 min3 times weekly12 weeksStretching, ROM and weight bearing activitiesCHAQ, VAS, VO2peak, HRmax, HRQoL2Elnaggar et al., 2020(21)IG: 5/12CG: 3/13IG: 12.11 ± 1.65CG: 11.31 ± 1.35PolyarticularStable doses of medicationsCore stability exercises + static stretching and contraction45 min3 times weekly12 weeksStatic stretching and contraction6MWT3Grewal et al., 2007(9)IG: 6/35CG: 10/29IG: 11.7 ± 2.5CG: 11.5 ± 2.4Polyarticular; Oligoarticular; Systemic; Enthesitis-related; Psoriatic; OtherStable doses of medicationsAerobic exercise50 min3 times weekly12 weeksRelaxation (qigong)CHAQ, VO2peak, HRQoL4Lelieveld et al., 2010(22)IG: 2/15CG: 2/14IG: 10.6 ± 1.5CG: 10.8 ± 1.4Polyarticular; Oligoarticular; Systemic; Enthesitis-relatedNAInternet-based aerobic exerciseNANA17 weeksWaiting-listHRmax5Sandstedt et al., 2013(23)IG: 8/25CG: 4/17IG: 13.3 ± 2.8CG: 14.9 ± 3.0Polyarticular; Oligoarticular; Entesites related; PsoriaticA local and/or systemic arthritis-related therapyAerobic exercise20 min3 times weekly12 weeksRoutine medical treatmentsCHAQ, HRmax6Sarac et al., 2024(24)IG: 12/5CG: 11/5IG: 15.12 ± 2.23CG:15.70 ± 1.59Polyarticular; Oligoarticular; Entesites relatedStable dosage of medicationInspiratory muscle trainingNA2 times daily8 weeksRoutine medical treatmentsVO2peak, HRmax, 6MWT, HRQoL7Takken et al., 2003(27)IG: 11/16CG: 3/24IG: 8.66 ± 2.29CG: 8.88 ± 1.86Polyarticular; Oligoarticular; SystemicMethotrexate, TNF blockers, and/or prednisoneAquatic exercise60 min1 time weekly20 weeksAssessment-onlyCHAQ, VO2peak, 6MWT8Tarakci et al., 2013(26)IG: 25/18CG: 19/19IG: 10.0 ± 3.4CG: 10.8 ± 4.0Polyarticular; Oligoarticular; Systemic; PsoriaticBiologic drug, DMARDsLand-based home exercise20–45 min4 times weekly12 weeksWait-listCHAQ, 6MWT, VAS, HRQoL6MWT 6 min Walk Test, CG Control Group, CHAQ Childhood Health Assessment Questionnaire, DMARDs Disease- Modifying Antirheumatic Drugs, F Female, HRmax Maximum Heart Rate, HRQoL Health-related Quality of Life, IG Intervention Group, JIA Juvenile idiopathic arthritis, M Male, ROM Range of motion, TNF Tumor Necrosis Factor, NA Not Available, VAS Visual Analogue Scale, VO2peak Peak Oxygen Consumption
Risk of bias
Three studies were judged as overall high risk of bias relating to missing outcome data and selection of the reported results [9, 23, 25]. Another five studies were judged to have ‘some concerns’ relating to bias mainly arising from deviations from intended interventions and selection of the reported results [20–22, 24, 26]. The detailed risk of bias assessment for the included studies was presented in Fig.
2.
Fig. 2Risk of bias summary for 8 RCTs included in this systematic review
Meta-analysis of outcome measures
Physical function
Four of the eight included studies reported data on CHAQ scores (Fig. 3) [9, 20, 25, 26]. As moderate heterogeneity was observed (I² = 56.4%, p = 0.08), a random-effects model was applied.
The pooled analysis showed a significant improvement in physical function favoring exercise interventions (WMD = -0.41; 95% CI: -0.80 to -0.03; p = 0.03), corresponding to a medium effect size (Cohen’s d ≈ 0.50).
Fig. 3Forest plot of the effect of exercise interventions on CHAQ scores
Four studies also assessed physical performance using the 6-minute walk test (Fig. 4) [21, 24–26]. Given the absence of heterogeneity (I² = 0.0%, p = 0.59), a fixed-effect model was used.
The combined result indicated a significant increase in walking distance among participants in the exercise group (WMD = 26.70 m; 95% CI: 1.70 to 51.71; p = 0.04), corresponding to a small-to-medium effect size (Cohen’s d ≈ 0.35).
Fig. 4Forest plot of the effect of exercise interventions on 6MWT
Cardiorespiratory fitness
VO₂peak was evaluated in four of the included trials (Fig. 5) [9, 20, 24, 25]. No heterogeneity was detected (I² = 0.0%, p = 0.61), and a fixed-effect model was applied.
The pooled analysis did not reach statistical significance (WMD = 0.20; 95% CI: -0.07 to 0.47; p = 0.16), and the estimated effect size was small (Cohen’s d ≈ 0.20).
Fig. 5Forest plot of the effect of exercise interventions on VO₂peak
Another four studies reported data on HRmax (Fig. 6) [20, 22–24]. With no detected heterogeneity (I² = 0.0%, p = 0.63), a fixed-effect model was used.
The meta-analysis revealed a significant increase in HRmax in favor of the exercise group (WMD = 7.27; 95% CI: 3.69 to 10.84; p < 0.01), corresponding to a medium effect size (Cohen’s d ≈ 0.55).
Fig. 6Forest plot of the effect of exercise interventions on HRmax
Health-related quality of life
HRQoL was assessed in five of the eight studies (Fig. 7) [9, 20, 24–26]. Due to moderate heterogeneity (I² = 66.7%, p = 0.02), a random-effects model was applied.
The pooled analysis showed a non-significant trend toward improved HRQoL with exercise interventions (SMD = 0.35; 95% CI: − 0.07 to 0.76; p = 0.10), corresponding to a small-to-medium effect size (Cohen’s d ≈ 0.35).
Fig. 7Forest plot of the effect of exercise interventions on HRQoL
Discussion
This meta-analysis synthesized evidence from eight randomized controlled trials to evaluate the effects of exercise interventions on physical function, cardiorespiratory fitness, and HRQoL in children and adolescents with juvenile idiopathic arthritis. Overall, the findings indicate that exercise interventions can improve physical function, although their effects on cardiorespiratory outcomes and HRQoL appear more limited or inconsistent. Our findings are broadly consistent with those of a recent meta-analysis by Liu et al., which included five RCTs and reported no significant improvements in CHAQ, VO₂peak, or HRQoL, although a significant reduction in VAS pain was observed.
While our analysis included more studies (eight RCTs) and identified significant benefits in CHAQ and 6MWT, the overall pattern of heterogeneity and modest effect sizes persisted. These similarities suggest that the current evidence base, despite differences in study selection and sample size, consistently points to modest and heterogeneous benefits of exercise in JIA. Such variability is likely driven by differences in intervention design, patient characteristics, and outcome definitions.
Among all outcome domains, the most robust and clinically relevant improvements were observed in physical function, as evidenced by significant reductions in CHAQ scores, reflecting better self-perceived functional ability, and increased 6MWT distances, reflecting enhanced physical performance. These results are consistent with previous meta-analysis, which have reported the positive impact of exercise on functional outcomes in chronic musculoskeletal pain, particularly in JIA populations [28]. A recent systematic review by Iversen et al. also emphasized the role of targeted physical activity programs in preserving mobility and joint function in children with rheumatic diseases [29].
Improvements in 6MWT suggest potential gains not only in endurance but also in confidence, proprioception, and dynamic joint control, especially when the interventions incorporate neuromuscular elements such as balance or core training [13, 21]. Notably, the pooled effect for this outcome appears to be primarily driven by the study of Elnaggar et al., [21] which implemented a structured core stability program focusing on deep abdominal and trunk muscle activation to enhance postural control and dynamic alignment. However, it is important to note that most studies used CHAQ to assess physical function, and a notable ceiling effect in the CHAQ measure was evident, especially in children with JIA treated with biologics, who often have CHAQ scores of 0 or very low.
This limitation makes it difficult to detect mild functional impairments. Therefore, future research should consider incorporating additional or complementary measures to capture a broader range of functional impairments and participation in physical activity.
In terms of cardiorespiratory fitness, the evidence was mixed. While exercise interventions were associated with significant improvements in HRmax [20, 22–24], the pooled effect on VO₂peak was non-significant [9, 20, 24, 27]. This divergence may reflect physiological distinctions between the two indicators: HRmax can be influenced by moderate-intensity training, whereas VO₂peak represents maximal aerobic capacity and typically requires higher-intensity, longer-duration interventions to achieve measurable change [30].
Moreover, methodological variability in VO₂peak testing protocols (e.g., treadmill vs. cycle ergometer) and disease-specific limitations (e.g., joint pain or fatigue during maximal testing) may have affected both performance and measurement consistency. A prior study also reported similar challenges in eliciting meaningful improvements in VO₂peak in children with JIA, even when using structured aerobic training programs were implemented [31].
Contrary to expectations, exercise interventions did not yield significant improvements in HRQoL in the pooled analysis [9, 20, 24, 26, 27]. This finding contrasts with results from external trials, such as that by Mendonça et al., in which a Pilates-based regimen produced significant enhancements in both psychosocial and physical dimensions of HRQoL [10]. Several factors may account for this discrepancy, including the relatively short duration of interventions in our included studies (ranging from 8 to 20 weeks), the absence of structured behavioral or motivational components in most protocols, and substantial variability in the instruments used to assess HRQoL.
Moreover, baseline HRQoL levels may have been relatively high in some trials, introducing potential ceiling effects that limit the measurable benefits of intervention. These limitations should be considered when interpreting our findings.
It is also important to consider the inherent heterogeneity among participants with JIA, including differences in disease subtype, treatment exposure, and psychosocial context, all of which may influence both responsiveness to exercise and the interpretation of outcomes. Moreover, variation in adherence, delivery format (home-based versus supervised), and motivational engagement may further contribute to the observed variability in results. Collectively, these multidimensional factors underscore the need for more tailored and stratified approaches to exercise interventions in pediatric rheumatology.
This meta-analysis possesses several strengths, including its exclusive focus on randomized controlled trials, comprehensive evaluation across multiple clinically relevant outcomes, and the use of standardized assessment tools. The inclusion of diverse exercise modalities further enhances the applicability and relevance of the findings to a variety of clinical and rehabilitation settings. Nonetheless, several limitations related to the risk of bias should be acknowledged.
First, substantial heterogeneity was present across the included studies, both in the clinical characteristics of participants and in the design of the exercise interventions. Although we extracted information on JIA subtypes, their uneven distribution and inconsistent reporting precluded subtype-specific analyses and limited our ability to evaluate differential intervention effects across JIA phenotypes. Exercise intensity, a critical determinant of training response, was frequently underreported or described using non-standardized metrics, hindering formal analysis of its impact.
This overall heterogeneity likely contributed to variability in effect sizes and reduced comparability across studies. Second, many studies lacked blinding of participants and/or outcome assessors, which may have introduced bias in outcome measurement and reporting, possibly inflating observed effects. Third, although reduced muscle strength is a hallmark of JIA, few eligible RCTs assessed this outcome using standardized and reliable methods.
Consequently, it was not included in our meta-analysis, limiting the comprehensiveness of our synthesis. This gap likely reflects both inconsistencies in assessment tools and a lack of strength-focused interventions in the current literature. Fourth, although statistical significance was achieved for several outcomes, we were unable to identify established minimal clinically important difference (MCID) values for most of the assessed outcomes in juvenile idiopathic arthritis.
This absence of MCID thresholds limits our ability to interpret whether the reported improvements are clinically meaningful. To address these limitations, future RCTs should prioritize methodological rigor by adopting standardized intervention protocols, implementing blinding procedures, and incorporating validated outcome measures. Moreover, future studies should aim to establish or report MCID values for key outcomes in JIA to facilitate a clearer interpretation of clinical relevance.
Such improvements will help minimize bias, enhance comparability across studies, and strengthen the clinical relevance of future findings.
Although our findings suggest that exercise interventions can improve certain functional outcomes in children with JIA, the observed variability across studies and the lack of consistent benefits in cardiorespiratory fitness and quality of life underscore the need for cautious application in clinical settings. At present, there is insufficient evidence to recommend specific exercise types or intensities tailored to JIA subtypes or disease activity levels. Nonetheless, clinicians may consider incorporating structured and supervised exercise programs as part of individualized management plans, particularly when functional impairments are present and inflammation is well-controlled.
Close monitoring using both subjective measures and objective performance tests may help track progress and guide adjustments to rehabilitation strategies. Future trials with stratified designs and standardized protocols are essential to refine these recommendations and support broader clinical implementation.
In conclusion, this meta-analysis demonstrated that exercise interventions significantly improve physical function in patients with JIA, whereas effects on cardiorespiratory fitness and HRQoL were less consistent. Exercise should be regarded as a core component of multidisciplinary care for JIA, and further high-quality trials are warranted to clarify its long-term benefits and optimal implementation strategies.
Supplementary Information
Supplementary Material 1.
Supplementary Material 2.
Supplementary Material 3.