Factors associated with fatigue in children with enthesitis/spondylitis-related juvenile idiopathic arthritis: an exploratory cross-sectional study

Abstract

Background

Enthesitis/spondylitis-related juvenile idiopathic arthritis (ESR-JIA) is a subtype of JIA characterized by peripheral arthritis and enthesitis, with possible axial skeleton involvement. Fatigue is a common and challenging symptom in JIA, often persisting despite pharmacological treatment, and may be particularly pronounced in ESR-JIA. It can impair physical and psychological well-being, daily activities, and academic performance. However, factors associated with fatigue in children with ESR-JIA remain poorly understood.

Methods

This cross-sectional study recruited 77 children diagnosed with ESR-JIA. Fatigue was assessed using the Pediatric Quality of Life Inventory Multidimensional Fatigue Scale (PedsQL-MFS) where higher scores indicate less fatigue. Associations between fatigue and clinical disease characteristics, objective clinical measures, subjective patient-reported factors, and blood biomarkers were examined. Spearman correlation analyses and simple and stepwise multivariable linear regression were performed to identify factors associated with fatigue.

Results

Univariable linear regression showed body composition parameters (free fat mass, skeletal muscle mass, and skeletal muscle index) and muscle strength (ankle dorsiflexion strength and ankle plantarflexion strength) were positively correlated with PedsQL-MFS total scores. In contrast, pain intensity, depression-related symptoms, and sleep efficiency were negatively associated with PedsQL-MFS total scores. Subsequent stepwise multivariable regression analysis showed pain intensity, depression-related cognitive changes, ankle plantarflexion strength, and sleep efficiency were independently associated with PedsQL-MFS total scores, collectively explaining 33.6% of the variance.

Conclusions

Fatigue in children with ESR-JIA was associated with multiple psychological and physical factors. Lower muscle mass, weaker ankle plantarflexor strength, higher pain intensity, depression-related cognitive changes, and poorer sleep efficiency were associated with more severe fatigue. These findings highlight the multidimensional nature of fatigue in ESR-JIA and support the need for comprehensive fatigue assessment and management.

Introduction

Juvenile idiopathic arthritis (JIA) is the most common rheumatic disease in childhood and is characterized by a chronic, relapsing course with a potential risk of long-term disability [1]. JIA is typically defined as arthritis of unknown etiology with onset before the age of 16 years and a disease duration of at least six weeks [2]. According to the Pediatric Rheumatology International Trials Organization (PRINTO) classification of 2018, JIA can be divided into six categories based on clinical features. Enthesitis/spondylitis-related JIA (ESR-JIA) is a form of juvenile spondyloarthritis characterized by peripheral arthritis and enthesitis, and it may involve the axial skeleton [1]. Compared with other JIA subtypes, ESR-JIA is associated with more severe pain and fatigue, poorer quality of life, and a worse overall prognosis [3]. With the widespread use of biologic therapies, clinical outcomes in children with JIA have improved markedly, leading to a substantial reduction in long-term disability [1]. However, several subjective symptoms, most notably fatigue, often show limited responsiveness to pharmacological treatment and have become major unmet concerns for patients and their families in the biologic era [4].

Fatigue is widely acknowledged as an overwhelming, distressing, and persistent sense of exhaustion that results in reduced physical and/or cognitive functioning and is not alleviated by rest or sleep [5]. It adversely affects psychological, physiological, cognitive, emotional, and social functioning. Fatigue is highly prevalent across rheumatic diseases; approximately 80–90% of adults with rheumatoid arthritis (RA) report fatigue [6, 7], with 38–41% experiencing severe fatigue [8], while severe fatigue affects up to 45.6% of patients with psoriatic arthritis [9]. Similarly, fatigue is common among children with JIA and may persist even during disease remission and into adulthood [10, 11]. Previous studies have reported that 60–76% of patients with JIA experience chronic fatigue, with 25.1% suffering from severe fatigue, compared with 9.1% in healthy peers [12]. Notably, moderate to severe fatigue persists in 27% of individuals with a history of JIA after 18–19 years of follow-up [13]. Although fatigue does not directly cause physical disability, it represents a substantial cognitive and emotional burden that adversely affects physical and mental well-being, academic performance, and social participation. Importantly, fatigue often responds poorly to both biologic and conventional antirheumatic therapies, making it one of the most challenging unmet symptoms in the biologic era [14].

Identifying factors associated with fatigue is a prerequisite for developing effective interventions. In 2016, Wineke Armbrust et al. proposed a JIA-specific theoretical model of fatigue based on pediatric chronic disease frameworks, integrating physiological and psychological factors and highlighting their interactions [11]. Furthermore, previous studies have suggested that fatigue in JIA is strongly associated with pain and depressive symptoms, whereas no clear direct association with disease activity [11, 15–18]. Beyond these subjective factors, fatigue in arthritis may also be related to objective clinical characteristics of the disease (e.g., number of active joints and sacroiliac joint involvement) as well as physical structure and function (e.g., muscle mass, strength, and functional capacity) [19–21]. While these associations have been more extensively explored in adult RA [19], evidence in JIA remains limited, and the relationships between fatigue and these parameters are largely theoretical. It has been hypothesized that chronic inflammation and subsequent sedentary behavior may contribute to muscle impairment and subsequently exacerbate fatigue; however, this hypothesis has not yet been systematically tested in the JIA population.

Therefore, this exploratory cross-sectional study aimed to systematically investigate fatigue-related risk factors in children with ESR-JIA by integrating clinical, physiological, and psychological domains, in order to improve understanding of fatigue in this population and to provide scientific insight for the future.

Methods

Study design

This study was an exploratory cross-sectional investigation conducted between August 2023 and December 2024, focusing on children with ESR-JIA to characterize fatigue-related factors. Patients aged 6–18 years meeting the 2018 PRINTO classification criteria for ESR-JIA were eligible for inclusion (Fig. 1). All participants underwent comprehensive assessments of clinical disease characteristics, objective physiological measures, subjective patient-reported outcomes, quality of life, and inflammatory and immunological blood biomarkers. This exploratory design was intended to systematically examine physiological and psychological factors associated with fatigue in the ESR-JIA population.

Fig. 1.

Study design flow diagram

Participant recruitment

Participants were recruited from both inpatient wards and outpatient clinics at Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, China. After routine clinical evaluation and confirmation of an ESR-JIA diagnosis by an experienced rheumatologist, eligible patients and their guardians were approached by the research team and invited to participate. All patients who met the inclusion criteria and did not meet any exclusion criteria underwent the full study assessments. Written informed consent was obtained from both the participants and their legal guardians prior to enrollment. The study protocol was approved by the Ethics Committee of Xinhua Hospital, Shanghai Jiao Tong University School of Medicine (Approval No. XHEC-C-2023-059-2).

Inclusion criteria

1. Age between 6 and 18 years;

2. Diagnosis of ESR-JIA according to the 2018 PRINTO classification criteria [22];

3. Ability to complete all required study assessments (e.g., questionnaires, body composition measurements, and muscle strength testing);

4. Provision of written informed consent by the patient and their legal guardian.

Exclusion criteria

1. Presence of other severe chronic diseases (e.g., congenital heart disease, diabetes mellitus), particularly other rheumatic diseases;

2. History of major surgery or hospitalization within the previous 6 months;

3. Current use of glucocorticoids;

4. Cognitive or language impairment preventing completion of questionnaires;

5. Presence of conditions associated with fatigue severity, such as psychiatric disorders.

Data collection

Data were collected from participants who met the inclusion criteria, did not meet any exclusion criteria, and provided written informed consent. Data collection covered sociodemographic characteristics, clinical disease features, objective physiological measures, subjective psychological outcomes, and inflammatory and immunological blood biomarkers, which were assessed at the same baseline visit before any subsequent intervention or treatment. All data were collected by trained researchers to ensure consistency and reliability. Participants’ personal information was kept strictly confidential in accordance with ethical approval requirements.

Sociodemographic information

Sociodemographic data collected included age, sex, and height.

Clinical disease characteristics

Clinical disease characteristics included Human leukocyte antigen B27 (HLA-B27) status, the number of affected joints, the presence of sacroiliac joint involvement, disease duration, and medication use, including the use of disease-modifying antirheumatic drugs and biologic agents. Data were obtained from medical records and verified through reports from parents and treating physicians.

1. Affected joints were defined as swollen joints or joints with a limited range of motion accompanied by pain on motion and/or tenderness;

2. The number of affected joints was defined as the number of joints with peripheral arthritis;

3. Sacroiliac joint involvement: MRI-defined sacroiliitis was defined by the presence of inflammation (bone marrow edema, joint space enhancement, or erosions/sclerosis) in the sacroiliac joint visualized by MRI, with or without the manifestation of clinical sacroiliitis [23];

4. Enthesitis was defined as inflammation localized to a tendon, ligament, or joint capsule insertion site to bone, with the clinical symptoms of tenderness and swelling at specific sites, which could be confirmed by MRI [24].

Objective clinical measures

Body composition

Muscle mass was assessed using a bioelectrical impedance analysis device (InBody 770, JM1143-741435, Korea). Participants stood barefoot on the platform with feet shoulder-width apart and shoulders abducted to 30 degrees, remaining still throughout the assessment. The measured parameters included body mass (BM), body mass index (BMI), percentage body fat (PBF), fat-free mass (FFM), skeletal muscle mass (SMM), and skeletal muscle index (SMI).

Muscle strength

Lower-limb muscle strength:

Lower-limb muscle strength was assessed using a handheld dynamometer (K-Force, Kinvent, France) to measure muscle strength around the hip, knee, and ankle joints. Testing procedures followed the manufacturer’s guidelines, with different body positions adopted for specific muscle groups. Prior to testing, participants performed two familiarization contractions at approximately 50% of maximal effort. During formal testing, participants were instructed to perform a maximal isometric contraction of the target muscle group for 3 s. Each test was repeated three times with 30-second rest intervals, and the mean value was used for analysis [25, 26]. Muscle groups assessed included hip flexors (HFS) and extensors strength (HES), knee flexors (KFS) and extensors strength (KES), and ankle dorsiflexors (ADS) and plantarflexors strength (APS).

Handgrip strength:

Handgrip strength was measured using a handheld dynamometer (K-Force, Kinvent, France). Participants were seated comfortably with the back supported, the elbow flexed at approximately 90 degrees, the forearm parallel to the floor, and the wrist in a neutral position. Participants were instructed to perform a brief maximal grip contraction for 3 s. Three trials were conducted with 30-second rest intervals, and the mean value was recorded as the handgrip strength [26].

Submaximal exercise capacity

Submaximal exercise capacity was assessed using the Six-Minute Walk Test (6MWT), which reflects endurance and functional exercise capacity [27]. The 6MWT has been demonstrated to be valid and reliable in patients with juvenile idiopathic arthritis. The test was conducted in a straight 20-m corridor, during which participants were instructed to walk as fast as possible for 6 min. Participants were allowed to slow down, stop, or rest if necessary, but the timer continued to run. The total distance walked over 6 min was recorded and used as the outcome measure of submaximal exercise capacity [27, 28].

Functional ability

Functional ability was evaluated using the Childhood Health Assessment Questionnaire (CHAQ), a widely used instrument with established reliability and validity that assesses functional status over the preceding week [29]. The CHAQ evaluates joint-related functional ability across eight domains of daily living (dressing and grooming, arising, eating, walking, hygiene, reach, grip, and activities) through 30 items. Patients rated the degree of difficulty experienced in completing each task during the past week due to their disease. Each item is scored on a 4-point scale (0 = no difficulty, 1 = some difficulty, 2 = much difficulty, 3 = unable to do). The highest item score within each domain was used as the domain score, and the mean of the eight domain scores was calculated to yield a total score ranging from 0 (no disability) to 3 (severe disability) [29].

Subjective clinical measures

Fatigue level

Fatigue was assessed using the Pediatric Quality of Life Inventory Multidimensional Fatigue Scale (PedsQL-MFS), which has demonstrated good reliability and validity in children with chronic conditions and includes both child self-report and parent-proxy report versions [30]. Items are rated on a 5-point Likert scale ranging from 0 (“never”) to 4 (“almost always”). Raw scores were reverse-scored and linearly transformed to a 0–100 scale (0 = 100, 1 = 75, 2 = 50, 3 = 25, 4 = 0). Mean scores were calculated for the total scale and each subscale, with lower scores indicating greater fatigue. The scale comprises three domains—general fatigue, cognitive fatigue, and sleep-related fatigue—the mean of which yields the total fatigue scores [31].

Pain level

Pain intensity was assessed using the Numerical Rating Pain Scale (NRPS), with scores ranging from 0 (no pain) to 10 (worst imaginable pain) [32]. Participants reported the maximum pain intensity experienced during the preceding week.

Emotional status (depression and anxiety level)

Depressive symptoms were evaluated using the Children’s Depression Inventory (CDI), a validated self-report measure for children aged 7–17 years [33]. Participants rated the frequency of each item over the past two weeks on a 3-point scale (0–2) [33]. Anxiety symptoms were assessed using the Screen for Child Anxiety Related Emotional Disorders (SCARED), a self-report instrument for children aged 8–18 years that evaluates anxiety over the past week across five domains (panic/somatic symptoms, generalized anxiety, separation anxiety, social phobia, and school phobia) [34].

Sleep quality

Sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI), a widely used self-report questionnaire developed by Buysse et al. that has demonstrated good reliability and validity across populations [35]. The PSQI consists of 19 items covering seven components: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. The global score ranges from 0 to 21, with higher scores indicating poorer sleep quality, and reflects sleep patterns over the previous month.

Blood biomarkers

Blood biomarkers assessed in this study included inflammatory cytokines—interleukin-8 (IL-8), tumor necrosis factor-α (TNF-α), and soluble interleukin-2 receptor (IL-2R)—as well as a nonspecific inflammatory marker, erythrocyte sedimentation rate (ESR), and immunological indicators (CD4%, CD8%, and CD4/CD8 ratio). All laboratory analyses were performed by the Department of Clinical Laboratory at Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, China.

Health-Related Quality of Life (HRQoL)

Health-related quality of life was assessed using the Pediatric Quality of Life Inventory™ 3.0 Rheumatology Module (PedsQL™ 3.0-RM), a widely used instrument for evaluating HRQoL in patients with JIA [36]. The Chinese version of the PedsQL™ 3.0-RM has been validated and shown to be reliable and valid for assessing HRQoL in children with JIA [37]. The scale consists of 22 items across five domains: pain and hurt, daily activities, treatment, worry, and communication, and includes both child self-report and parent-proxy report versions. Items are rated on a 5-point Likert scale ranging from 0 (“never”) to 4 (“almost always”) and are linearly transformed to a 0–100 scale (0 = 0, 1 = 25, 2 = 50, 3 = 75, 4 = 100), with higher scores indicating better HRQoL [37].

Sample size

As this study focused on a specific JIA subtype, ESR-JIA, and no previous cross-sectional studies have specifically examined fatigue and its associated factors in this population, a formal a priori sample size calculation based on effect size estimation was not feasible. Given the exploratory nature of the study, the target sample size was determined with reference to prior cross-sectional studies investigating fatigue in broader JIA populations. In particular, a cross-sectional study by Armbrust et al. examining fatigue and its associated factors in children with JIA included approximately 80 participants [15]. Based on this reference and considering the relatively low prevalence of ESR-JIA, a target sample size of around 80 patients was deemed appropriate to allow exploratory analyses of fatigue-related factors. During the recruitment period, 100 patients with ESR-JIA were enrolled. After data cleaning and exclusion of cases with missing key variables, 77 participants were included in the final analysis.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics version 27.0 (IBM Corp., Armonk, NY, USA) and R version 4.0.0 (R Foundation for Statistical Computing, Vienna, Austria). Data are presented as mean ± standard deviation for continuous variables and n (%) for categorical variables. The PedsQL-MFS total score was analysed as a continuous outcome according to the standard scoring method, with higher scores indicating lower fatigue. Normality of continuous variables was assessed using the Shapiro–Wilk test. Correlations between PedsQL-MFS scores and clinical, physiological, psychological, and biomarker variables were examined using Pearson correlation for approximately normally distributed continuous variables and Spearman correlation for non-normally distributed, skewed, or ordinal variables. Correlation results were visualized using heatmaps. Candidate variables were selected based on previous literature, clinical relevance, and data availability. Variables associated with the PedsQL-MFS total score at p < 0.10 in correlation or univariable linear regression analyses were considered for multivariable analysis. Multivariable linear regression was used to examine factors associated with the PedsQL-MFS total score. Regression coefficients (β), 95% confidence intervals (CIs), and p values were reported. Multicollinearity among variables included in the final multivariable linear regression model was assessed using tolerance and variance inflation factor (VIF) values. To assess the robustness of the final model, a sensitivity analysis was performed using a forced-entry linear regression model including the variables retained in the final multivariable model. Statistical significance was set at p < 0.05.

Results

A total of 100 children with ESR-JIA who met the diagnostic criteria were recruited for this study. All participants provided written informed consent and completed the study assessments at Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, China. Complete datasets were obtained from 77 participants and included in the final analysis (Fig. 1), while 23 participants were excluded for the following reasons: (1) missing body composition data (n = 13), as the bioelectrical impedance analysis device (InBody) was formally implemented in the study only after December 12, 2023, and data collected prior to this date were excluded; and (3) incomplete blood biomarker data (n = 10), primarily due to missing laboratory results among outpatient participants.

Baseline characteristics of the study population

As shown in Table 1, a total of 77 participants were included in the final analysis, comprising 38 males and 39 females. The mean age of the cohort was 12.05 ± 2.68 years, with a mean height of 151.33 ± 14.88 cm and a mean body weight of 46.52 ± 15.74 kg. The prevalence of HLA-B27 positivity was 32.5%, and sacroiliac joint involvement was present in 88.7% of participants. Overall, 59.7% of patients received combination therapy with conventional disease-modifying antirheumatic drugs and biologic agents. The mean disease duration was 3.23 ± 1.92 years. The mean total fatigue score of the study population was 68.33 ± 17.79 (Table 2).

Table 1.

Baseline demographic and clinical characteristics of the study participants

Variables	Overall N = 77	Male N = 38	Female N = 39	P
Age, years	12.05 ± 2.68	11.79 ± 2.80	12.31 ± 2.57	0.400
Height, cm	151.33 ± 14.88	152.85 ± 16.94	149.84 ± 12.62	0.379
Weight, kg	46.52 ± 15.74	46.24 ± 16.72	46.79 ± 14.94	0.879
BMI, kg/m2	19.78 ± 4.18	19.21 ± 3.97	20.34 ± 4.35	0.555
HLA-B27				< 0.001
Negative	52 (67.5%)	17 (44.7%)	35 (89.7%)
Positive	25 (32.5%)	21 (55.3%)	4 (10.3%)
Sacroiliac joint involvement				0.168
Positive	68 (88.3%)	36 (94.7%)	32 (82.1%)
Negative	9 (11.7%)	2 (5.3%)	7 (17.9%)
Number of affected joints, n	3.73 ± 1.71	3.08 ± 1.48	4.36 ± 1.69	< 0.001
Medication type				0.211
DMARDs	22 (28.6%)	11 (26%)	11 (32%)
Biologic agents	9 (11.7%)	2 (15%)	7 (7.9%)
DMARDs+biologic agents	46 (59.7%)	25 (59%)	21 (61%)
Disease duration, years	3.23 ± 1.92	1.58 ± 0.92	2.08 ± 1.11	0.919

Data are presented as mean ± standard deviation or n (%)

BMI, body mass index; HLA-B27, human leukocyte antigen B27; DMARDs, disease-modifying antirheumatic drugs

Table 2.

Baseline outcome measures of the study participants

Variables	Overall N = 77	Male N = 38	Female N = 39	P
Body composition
PBF, %	25.76 ± 9.69	22.38 ± 9.52	29.05 ± 8.79	0.002
FFM, kg	33.81 ± 10.26	35.49 ± 12.36	32.17 ± 7.50	0.157
SMM, kg	17.98 ± 6.15	19.03 ± 7.43	16.97 ± 4.45	0.143
SMI, kg/m²	5.23 ± 1.34	5.45 ± 1.55	5.01 ± 1.07	0.150
Muscle strength
HFS, kg	13.87 ± 4.11	13.99 ± 3.84	13.75 ± 4.40	0.801
HES, kg	26.85 ± 4.67	28.61 ± 3.83	25.13 ± 4.81	< 0.001
KFS, kg	13.83 ± 4.02	14.83 ± 3.90	12.86 ± 3.94	0.031
KES, kg	19.88 ± 6.43	21.05 ± 7.39	18.73 ± 5.18	0.113
ADS, kg	26.15 ± 3.82	27.25 ± 3.48	25.07 ± 3.87	0.011
APS, kg	15.45 ± 3.86	16.24 ± 4.14	14.68 ± 3.45	0.077
HGS, kg	10.19 ± 6.43	11.26 ± 7.20	9.15 ± 5.47	0.152
NRPS	3.26 ± 1.92	3.11 ± 2.09	3.41 ± 1.76	0.490
6-WMT, m	559.83 ± 86.34	556.99 ± 91.01	562.60 ± 82.61	0.777
Fatigue
General fatigue	70.77 ± 21.88	72.49 ± 23.55	69.10 ± 20.30	0.501
Cognitive fatigue	70.84 ± 21.68	72.11 ± 22.19	69.59 ± 21.38	0.613
Sleep fatigue	63.43 ± 19.57	65.24 ± 20.553	61.67 ± 18.69	0.427
Total fatigue score	68.33 ± 17.79	69.91 ± 19.09	66.79 ± 16.52	0.445
Depression
Emotional condition	3.09 ± 1.40	3.05 ± 1.59	3.13 ± 1.20	0.814
Cognitive changes	1.17 ± 1.90	0.71 ± 1.33	1.62 ± 2.24	0.035
Behavioral changes	6.70 ± 4.80	7.32 ± 5.07	6.10 ± 4.50	0.270
Physical condition	1.75 ± 1.84	1.61 ± 1.87	1.90 ± 1.83	0.491
Interest loss	2.64 ± 1.81	2.97 ± 1.76	2.31 ± 1.81	0.106
Total depression score	13.97 ± 8.04	14.61 ± 7.78	13.36 ± 8.34	0.500
Total anxiety score	59.65 ± 15.09	56.89 ± 15.27	62.33 ± 14.62	0.115
Sleep quality
Sleep quality	0.81 ± 0.81	0.66 ± 0.85	0.95 ± 0.76	0.117
Sleep duration	0.08 ± 0.48	0.16 ± 0.68	0.00 ± 0.00	0.150
Sleep efficiency	0.19 ± 0.40	0.18 ± 0.39	0.21 ± 0.41	0.820
Sleep disturbances	0.91 ± 0.75	0.79 ± 0.74	1.03 ± 0.74	0.167
Daytime dysfunction	1.22 ± 1.13	1.00 ± 0.00	1.44 ± 1.14	0.091
PSQI total score	4.34 ± 3.10	3.68 ± 2.99	4.97 ± 3.11	0.067
Health-related quality of life
Pain	74.30 ± 21.84	78.62 ± 19.62	70.09 ± 23.29	0.087
Daily activities	95.13 ± 10.54	95.92 ± 9.51	94.36 ± 11.54	0.519
Treatment	90.26 ± 15.02	89.61 ± 16.25	90.90 ± 13.90	0.709
Communication	80.63 ± 25.09	84.87 ± 26.06	76.50 ± 23.71	0.144
Worry	81.60 ± 23.43	80.26 ± 23.36	82.91 ± 23.72	0.624
Total score	84.72 ± 12.45	85.84 ± 14.07	83.62 ± 10.71	0.437
Blood biomarkers
ESR	11.13 ± 12.57	10.16 ± 12.00	12.08 ± 13.19	0.507
IL-8	33.78 ± 61.45	19.23 ± 36.47	47.95 ± 76.42	0.039
TNF-α	183.91 ± 225.04	186.91 ± 237.82	180.98 ± 214.95	0.909
CD4%	34.82 ± 7.42	34.12 ± 7.21	35.51 ± 7.65	0.414
CD8%	27.60 ± 6.77	27.31 ± 6.46	27.88 ± 7.13	0.714
CD4/CD8	1.30 ± 0.45	1.29 ± 0.43	1.31 ± 0.47	0.847

Data are presented as mean±standard deviation or number (percentage)

BM, body mass; BMI, body mass index; PBF, percentage body fat; FFM, fat-free mass; SMM, skeletal muscle mass; SMI, skeletal muscle index; HFS, hip flexion strength; HES, hip extension strength; KFS, knee flexion strength; KES, knee extension strength; ADS, ankle dorsiflexion strength; APS, ankle plantarflexion strength; HGS, handgrip strength; NRPS, Numerical Rating Pain Scale; 6-MWT, six-minute walk test; ESR, erythrocyte sedimentation rate; IL-8, interleukin-8; TNF-α, tumor necrosis factor-alpha

Correlation between fatigue and outcome variables

As shown in Fig. 2, none of the clinical disease characteristics included in this study were significantly correlated with PedsQL-MFS total scores (p > 0.05). NRPS was negatively correlated with general fatigue (r = − 0.28, p < 0.05), sleep-related fatigue (r = − 0.36, p < 0.01), and PedsQL-MFS total scores (r = − 0.31, p < 0.01) (Fig. 3).

Fig. 2.

Correlation heatmap between disease-related characteristics and fatigue dimensions

Fig. 3.

Correlation heatmap between objective clinical factors and fatigue dimensions

As shown in Fig. 3, body composition measures, including FFM, SMM, and SMI, were positively correlated with PedsQL-MFS total scores (FFM: r = 0.40, p < 0.01; SMM: r = 0.39, p < 0.01; SMI: r = 0.38, p < 0.01).

In terms of muscle strength, ADS and APS were positively associated with PedsQL-MFS total scores (ADS: r = 0.31, p < 0.01; APS: r = 0.31, p < 0.05).

As shown in Fig. 4, depression scores (cognitive changes) were negatively associated with PedsQL-MFS total scores (cognitive changes r = − 0.33, p < 0.01). Total anxiety scores were also negatively associated with PedsQL-MFS total scores (r = − 0.24, p < 0.05).

Fig. 4.

Correlation heatmap between subjective clinical factors and fatigue dimensions

Simple linear regression analysis

As shown in Table 3, simple linear regression analyses were conducted to examine the associations between PedsQL-MFS total scores and disease-related outcomes, objective and subjective factors, blood biomarkers, adjusting for sex, age, disease duration, and the number of affected joints.

Table 3.

Results of simple linear regression analyses for total fatigue score

Variables	β	SE	t	p
Body composition
FFM, kg	0.79	0.29	2.76	0.01*
SMM, kg	1.32	0.48	2.73	0.01*
SMI, kg/m2	4.84	2.15	2.25	0.03*
Muscle strength
KFS, kg	0.38	0.67	0.57	0.57
ADS, kg	1.20	0.59	2.02	0.05*
APS, kg	1.22	0.62	1.95	0.06*
HGS, kg	0.39	0.42	0.93	0.35
NRPS	-2.78	1.18	-2.35	0.02*
6-WMT	-0.01	0.03	-0.39	0.70
Blood biomarkers
ESR	0.11	0.17	0.67	0.51
IL-8	0.01	0.04	0.33	0.74
TNF-α	-0.01	0.01	-1.05	0.30
IL-2R	-0.003	0.01	-0.59	0.56
CD3%	-0.12	0.33	-0.36	0.72
CD4%	0.13	0.29	0.45	0.66
CD8%	-0.05	0.33	-0.17	0.87
CD4/CD8	6.17	4.78	1.29	0.20
Depression
Emotional condition	-0.95	1.54	-0.62	0.54
Cognitive changes	-2.83	1.18	-2.41	0.02*
Behavioral	-0.54	0.45	-1.19	0.24
Physical condition	-2.63	1.16	-2.27	0.03*
Interest Loss	-0.22	1.21	-0.18	0.86
Total depression score	-0.50	0.27	-1.87	0.07*
Anxiety
School anxiety	-1.98	1.15	-1.72	0.09
Total anxiety score	-0.19	0.16	-1.23	0.22
Sleep quality
Sleep quality	0.10	2.92	0.03	0.97
Sleep efficiency	-4.60	2.00	-2.30	0.03*
Sleep disturbances	-2.29	3.27	-0.70	0.49
Daytime dysfunction	-7.44	5.79	-1.29	0.20
PSQI total score	-1.01	0.84	-1.21	0.23
Health-related quality of life
Pain	0.27	0.10	2.72	0.01*
Daily activities	0.11	0.20	0.58	0.57
Treatment	0.18	0.14	1.29	0.20
Communication	0.15	0.09	1.60	0.12
Worry	0.24	0.10	2.42	0.02*
Total score	0.47	0.17	2.74	0.01*

FFM, fat-free mass; SMM, skeletal muscle mass; SMI, skeletal muscle index; KFS, knee flexion strength; ADS, ankle dorsiflexion strength; APS, ankle plantarflexion strength; HGS, handgrip strength; 6-WMT, six-minute walk test; ESR, erythrocyte sedimentation rate; IL-8, interleukin-8; TNF-α, tumor necrosis factor–α; IL-2R, interleukin-2 receptor; CD3%, percentage of CD3 + T cells; CD4%, percentage of CD4 + T cells; CD8%, percentage of CD8 + T cells; CD4/CD8, CD4/CD8 ratio; PSQI, Pittsburgh Sleep Quality Index

Simple linear regression analyses were adjusted for sex, age, disease duration, and number of affected joints

β represents the unstandardized regression coefficient

Variables with p < 0.10 in univariable linear regression were considered for inclusion in the multivariable linear regression model. Statistical significance was defined as p < 0.05

Disease-related outcome

NRPS was negatively associated with PedsQL-MFS total scores (β = −2.78, p < 0.05).

Objective clinical factors

In terms of muscle mass, including FFM (β = 0.79, p < 0.05), SMM (β = 1.32, p < 0.05), and SMI (β = 4.84, p < 0.05), were positively associated with PedsQL-MFS total scores. In terms of muscle strength, ADS (β = 1.20, p = 0.05) and APS (β = 1.22, p = 0.06) showed associations with PedsQL-MFS total scores approaching statistical significance.

Subjective clinical factors

Within the depression domains, cognitive changes (β = −2.83, p < 0.05) and physical condition (β = −2.63, p < 0.05) were both significantly negatively associated with PedsQL-MFS total scores. Regarding sleep-related variables, only sleep efficiency was significantly associated with PedsQL-MFS total scores (β = −4.60, p < 0.05). In contrast, total HRQoL scores (β = 0.47, p < 0.05) were positively associated with PedsQL-MFS total scores.

Blood biomarkers

No significant associations were observed between blood biomarkers (including ESR, IL-8, and TNF-α) and PedsQL-MFS total scores (p > 0.05).

Multivariable linear regression analysis

As shown in Table 4, in the multivariable linear regression analysis, the final model was statistically significant, F = 9.143, p < 0.001, explaining 33.6% of the variance in PedsQL-MFS total scores, with an adjusted R² of 0.300. After adjustment for the variables included in the model, depression-related cognitive changes (β = −0.27, p < 0.05), NRPS (β = −0.30, p < 0.01), APS (β = 0.29, p < 0.01), and sleep efficiency (β = −0.21, p < 0.05) were significantly associated with PedsQL-MFS total scores. No substantial multicollinearity was detected among the variables included in the final model, with all VIF values below 5. A forced-entry sensitivity analysis including the variables retained in the final model produced similar results. The model explained 33.7% of the variance in PedsQL-MFS total scores, with an adjusted R² of 0.300, which was comparable to the final stepwise model.

Table 4.

Multivariable linear regression analysis of factors associated with total fatigue score

Variables	Standardized β	t	p	VIF
Cognitive changes (depression)	-0.27	-2.77	0.01*	1.14
NRPS	-0.30	-3.11	0.003**	1.11
APS	0.29	3.01	0.004**	1.16
Sleep efficiency	-0.21	-2.15	0.035*	1.03

NRPS, Numerical Rating Pain Scale (maximum pain intensity)

Standardized β coefficients are presented

The multivariable linear regression model was constructed using a stepwise approach and adjusted for sex, age, disease duration, and number of affected joints

The final model explained 33.6% of the variance in total fatigue score

p < 0.05 was considered statistically significant

* p < 0.05; ** p < 0.01

Discussion

This exploratory cross-sectional study systematically investigated disease-related, psychological, and physiological factors associated with fatigue in children with ESR-JIA. To our knowledge, this is one of the first studies to comprehensively investigate fatigue-related factors in this specific JIA subtype. The main findings were as follows: (1) Higher pain intensity was associated with more severe fatigue (lower PedsQL-MFS total scores), other clinical disease characteristics did not seem to be associated with fatigue intensity; (2) Lower muscle mass and weaker lower-limb muscle strength were associated with more severe fatigue; (3) more severe depressive symptoms and poorer sleep efficiency were associated with more severe fatigue; and (4) multivariable linear regression identified depression-related cognitive changes, pain intensity, ankle plantarflexor strength, and sleep efficiency as factors independently associated with fatigue, collectively explaining 33.6% of the variance in fatigue scores.

The mean PedsQL-MFS total scores in this study were 68.33 ± 17.79, which is lower (indicating more severe fatigue) than values reported in previous studies of mixed JIA populations (all JIA subtypes: 77.8; polyarticular JIA: 76.38) [15, 38]. This discrepancy may be attributable to differences in JIA subtype composition. The study by Armbrust et al. included all JIA subtypes, the present study focused exclusively on ESR-JIA. Consistent with prior evidence indicating that ESR-JIA is associated with greater fatigue than other JIA subtypes [39], our findings further indicate that the fatigue burden in children with ESR-JIA may be particularly pronounced.

In this study, fatigue was not significantly associated with conventional clinical disease characteristics, which is consistent with previous reports in JIA populations and adult RA [12, 38–40]. Beyond this, pain showed a significant association with fatigue severity, with higher pain levels associated with more severe fatigue. This association remained significant in the multivariable model. This finding is broadly consistent with previous evidence showing a robust association between pain and fatigue in JIA across both cross-sectional and longitudinal studies [15, 41, 42]. These findings suggest that fatigue may be more closely related to subjective pain burden than to traditional disease characteristics, such as the number or pattern of affected joints. This may partly explain why fatigue often responds poorly to pharmacological treatment and why some patients continue to experience substantial fatigue even when joint inflammation is clinically controlled. Moreover, pain may also mediate the relationship between disease activity and fatigue level. Ringold et al. demonstrated that after adjusting for pain, disease activity was no longer associated with fatigue, suggesting that pain may mediate the relationship between disease activity and fatigue [16]. Similar conclusions have been reported in other studies [12]. Persistent pain may contribute to fatigue through several pathways, including disrupted sleep, reduced physical activity, emotional distress, and increased perceived effort during daily activities [43]. From a neurophysiological perspective, chronic pain and fatigue may also share mechanisms involving increased central nervous system sensitivity to pain-related signals [44]. Clinically, these findings suggest that fatigue assessment in ESR-JIA should not rely only on inflammatory or structural disease indicators. Pain should be routinely evaluated when children report fatigue, and future studies should examine whether comprehensive pain management, together with sleep, psychological, and physical function interventions, may help reduce fatigue burden in this population. Nevertheless, the lack of standardized disease activity measures, such as the JADAS, BASDAI, or ADAS, means that our assessment captured only limited aspects of disease activity. Consequently, our findings should be interpreted with caution, and a potential relationship between disease activity and fatigue cannot be excluded. Future prospective studies incorporating validated disease activity measures are needed to further clarify this association.

Furthermore, we observed a significant positive association between PedsQL-MFS total scores and muscle condition, indicating that greater muscle mass and strength (especially ankle) were associated with less severe fatigue in children with ESR-JIA. This finding adds to the emerging view that fatigue in ESR-JIA may be related not only to disease-related and psychological factors, but also to objective physical status. Greater muscle mass and muscle strength may reflect better physiological reserve, physical capacity, and tolerance to daily activities, which may be associated with a lower fatigue burden. Notably, ankle plantarflexor strength was significantly associated with fatigue in the present study. This finding may be particularly relevant in ESR-JIA, as ankle involvement is common in this subtype and has been linked to poorer outcomes in JIA [45]. The ankle plantarflexors may also be susceptible to weakness in children with ESR-JIA because of pain, altered loading, reduced physical activity, or avoidance of weight-bearing tasks [45, 46]. To our knowledge, no previous JIA studies have directly examined the relationship between muscle mass or strength and subjective fatigue; therefore, comparisons were extended to the RA literature. In a study by Vlietstra et al., appendicular skeletal muscle mass showed a weak association with self-reported fatigue in RA [19]. Similarly, studies in systemic lupus erythematosus have reported associations between reduced muscle strength and fatigue [47]. This association may be partly explained by the fact that inflammation in arthritis has been linked to reduced exercise tolerance, which may lead to earlier onset of fatigue even during low-intensity daily activities [48]. This finding is cross-sectional design and exploratory study, but it is clinically relevant because muscle condition is a potentially modifiable physical factor. Future longitudinal and interventional studies are needed to determine whether improving muscle mass and physical function can reduce fatigue in children with ESR-JIA.

In the present study, depression-related cognitive changes and sleep efficiency remained significantly associated with fatigue in the multivariable model. These findings are consistent with previous evidence suggesting that fatigue in JIA is closely related to subjective symptoms, including depressive symptoms, sleep disturbance, pain, and health-related quality of life. Compared with healthy peers, children with JIA frequently experience depression and sleep disturbance [49], both of which are closely related to functional limitations and persistent pain [48]. A prospective study in children with JIA showed that day-to-day fluctuations in self-reported mood and stress were associated with changes in fatigue and pain [17]. Similarly, a longitudinal study demonstrated that fatigue levels in JIA were associated with quality of life up to 30 years later [50]. The relationship among depressive symptoms, sleep disturbance, pain, and fatigue is likely complex and bidirectional. Pain may contribute to poor sleep quality, while sleep disturbance may increase pain sensitivity and reduce physical and psychological resilience, thereby worsening fatigue. Depressive symptoms may further amplify fatigue through changes in mood, behavior, activity engagement, and central nervous system mechanisms [51] A possible biological explanation is the close bidirectional communication between the immune system and the brain [53, 54]. Pro-inflammatory activity, including increased circulating cytokines such as TNF-α, IL-1, and IL-6, may influence monoaminergic neurotransmitter systems and contribute to sickness behavior, including fatigue. Although the present study did not find a direct association between fatigue and inflammatory biomarkers, these pathways may still help explain the close relationship between psychological symptoms and fatigue. Given the predominantly cross-sectional nature of existing studies, including the present study, the causal relationships among sleep disturbance, pain, depressive symptoms, and fatigue remain unclear [52, 53]. These factors may interact in a vicious cycle that substantially impairs quality of life in children with JIA [54]. Identifying the key drivers within this cycle is an important direction for future research. Notably, emerging evidence suggests that appropriately prescribed physical activity may improve fatigue, pain, mood, sleep quality, and muscle function, highlighting its potential to disrupt the pain-sleep-fatigue cycle.

Several limitations of this study should be acknowledged. First, this study focused exclusively on a specific JIA subtype, ESR-JIA. Given the heterogeneity in pathophysiology and clinical presentation across JIA subtypes, the fatigue-related factors identified in this study may not be generalizable to other JIA subtypes and warrant further investigation. Second, standardized disease activity instruments (e.g., JADAS, BASDAI, or ASDAS) were unavailable for our study, and therefore could not be included in the primary analysis. Although we adjusted for available clinical indicators, including pain, inflammatory markers, and clinical joint involvement, these variables are not adequate substitutes for validated composite disease activity scores. Thus, our findings should be interpreted cautiously and should not be taken as evidence that fatigue is unrelated to disease activity in ESR-JIA. Third, the cross-sectional design limits the interpretation of findings. The observed relationships should be interpreted as associations rather than evidence of causal relationships between fatigue and the identified factors. Fourth, the relatively small sample size limited the complexity of the statistical analyses and precluded the development or validation of a formal prediction model. Therefore, the regression findings should be interpreted as exploratory and hypothesis-generating. Finally, several potentially relevant factors, including socioeconomic status, comorbidities, detailed physical activity levels, and lifestyle-related factors, were not available. The absence of these variables may have limited the interpretation of fatigue-related factors in children with ESR-JIA. Future studies should include more comprehensive assessments to better clarify fatigue-related mechanisms.

Conclusion

This study demonstrated that lower muscle mass, weaker muscle strength, higher pain intensity, depressive symptoms, poorer sleep efficiency, and lower health-related quality of life were associated with more severe fatigue. In the multivariable model, maximal pain intensity, depression-related cognitive changes, ankle plantarflexor strength, and sleep efficiency were independently associated with PedsQL-MFS total scores, collectively explaining 33.6% of the variance. These findings highlight the multidimensional nature of fatigue in ESR-JIA and suggest that psychological, sleep-related, and physical factors should be considered when assessing fatigue in this population. Further longitudinal studies are needed to confirm these findings and explore potential causal pathways.

Author contributions

Mei Zhang and Yue Xing contributed equally as the first authors. Xiangfeng He assisted with manuscript revision. Guimei Guo and Nan Chen contributed equally as corresponding authors.

Funding

This study was supported by the Science and Technology Commission of Chongming District, Shanghai, China (No. CKY2024-60), the “Clinical innovation project” of Xinhua Hospital, “AI Prediction Model for Sarcopenia-Cognitive Impairment Comorbidity in the Elderly Based on Multimodal Medical Big Data” (No.25XHCR06A), the “Research physician” program of Shanghai Jiao Tong University School of Medicine, “Early screening, risk prediction and rehabilitation intervention of sarcopenia and cognitive impairment in the elderly in the community” (No.20250515), and the Shanghai Oriental Talent Program (JS2025098).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of Xinhua Hospital, Shanghai Jiao Tong University School of Medicine (Approval No. XHEC-C-2023-059-2).

Consent for publication

Written informed consent was obtained from the legal guardians of all participants, and assent was obtained from the children themselves prior to enrollment.

Competing interests

The authors declare no competing interests.