Juvenile arthritis damage index predicts poor response to biological treatment: A prospective cohort study

Abstract

BACKGROUND

Juvenile arthritis damage index (JADI) is a tool that measures the degree of aggressiveness of the juvenile idiopathic arthritis (JIA) course and assesses articular [JADI - articular damage (JADI-A)] and extraarticular [JADI - extraarticular damage (JADI-E)] damage. While aggressive JIA often requires early biologic disease-modified antirheumatic drugs (bDMARDs), the utility of JADI as a predictor of treatment response remains underexplored.

AIM

To evaluate the potential of JADI as a predictor of bDMARD treatment response in JIA patients.

METHODS

This prospective study included 112 highly active non-systemic JIA biologic-naïve patients with a mean age of 12.2 ± 4.6 years and a median disease duration of 2.5 (interquartile range: 1-5) years. Their clinical and radiological assessment, juvenile arthritis disease activity score 71, JADI-A, and JADI-E, were evaluated twice: Before the biologic initiation (baseline) and 12 months after (end of study). At baseline, 50% had any damage, with 43% with articular damage and 23% with extraarticular damage.

RESULTS

During the study, JADI-A/JADI-E improved (33.9%/9.8%), worsened (8.9%/5.4%), or remained unchanged (57.1%/84.8%). Patients with baseline damage had higher markers of JIA activity: Polyarticular course, earlier onset age, ANA-positivity, and more active joints. Patients without initial structural damage (JADI“-“) were more likely (odds ratio = 3.8, 95% confidence interval: 1.6-9.0, P < 0.004) to achieve a low degree of activity or remission (46.2%), while on biological therapy, their scores were comparable to JADI-positive (18.3%). Pre-biological joint damage according to the JADI-A index (P = 0.003), wrist (P = 0.035), elbow (P = 0.027), cervical spine limitation of motion (P = 0.051), and erosions confirmed by magnetic resonance imaging (P = 0.002), were associated with poor response to biological treatment and follow-up JIA activity.

CONCLUSION

Baseline structural damage in JIA is associated with diminished bDMARDs efficacy, increased disability, and shorter remission duration. JADI enhances conventional clinical risk stratification by facilitating timely initiation of bDMARDs, adherence to treat-to-target strategy and tailored patient care.

Key Words: Juvenile idiopathic arthritis; Juvenile arthritis damage index; Biologic disease-modified antirheumatic drugs; Joint contractures; Treatment response prediction; Pediatric rheumatology

Core Tip: This prospective study of biological-naïve juvenile idiopathic arthritis patients demonstrates that baseline articular/extraarticular damage (assessed via the juvenile arthritis damage index) predicts reduced response to biological disease-modifying antirheumatic drugs. Patients with baseline damage demonstrated higher disease activity (polyarticular course, elevated juvenile arthritis disease activity score 71, magnetic resonance imaging erosions) than those without damage. The observed juvenile arthritis damage index reduction in juvenile idiopathic arthritis patients showed significant associations with baseline joint damage status, treatment stability, and absence of uveitis/osteonecrosis. This index encourages early biologic intervention and tailoring personalized treatment strategies to minimize long-term disability and enhance clinical outcomes.

INTRODUCTION

Juvenile idiopathic arthritis (JIA) is one of the most common chronic rheumatic diseases in children. It plays a significant role in developing functional disorders of the musculoskeletal system and extraarticular damage to structures that can be both temporary and irreversible[1]. Due to progress in JIA treatment, previously achieving remission or minimal disease activity has become a realistic and important goal of therapy, which is accompanied by a reduction in joint damage and a decrease in the degree of functional disorders[2].

According to the literature, factors influencing the effectiveness of biological therapy in patients with JIA include: The female sex[3,4], systemic JIA category[3,5], later age of disease onset[5-7], previous inefficiency, and the number of disease-modifying antirheumatic drug (DMARD)[5,8], high baseline score according to the questionnaire assessment of the functional status of the Childhood Health Assessment Questionnaire (CHAQ)[3], high disease activity according to the physician’s assessment, absence or short periods of remission in the past[9], the presence of uveitis[5], long disease duration[6], the presence of human leukocyte antigen B27 and enthesitis-related JIA category[10], rheumatoid factor (RF) and antibodies to cyclic citrullinated peptide positivity[3], high baseline level of pro-inflammatory cytokines[11], antirheumatic therapy compliance[8], and hip involvement[4].

One poorly studied but critically important factor determining the effectiveness of biological therapy in achieving remission or low disease activity is the aggressive course of arthritis, accompanied by the development of irreversible joint and extraarticular changes. According to research, early detection and prevention of joint damage are key in optimizing treatment outcomes and improving the long-term prognosis for patients with juvenile arthritis[3,8,12].

Disease activity in JIA is routinely monitored using composite measures such as [Juvenile arthritis disease activity score (JADAS)][13], ACRpedi criteria[14], the CHAQ[3], and criteria for Wallace et al’s study[15] inactive disease status. None of these tools provides a universal method for objective assessment of irreversible cumulative joint and organ damage over time, relying instead on subjective physician or patient evaluations. Psychological distress in children with JIA and their parents, along with the increased prevalence of anxiety and depression among parents of JIA patients, may distort the true clinical picture of the disease[16,17].

In 1999, Len et al[18] proposed the pediatric Escola Paulista de Medicina Range of Motion (ROM) Scale to reduce the reliance on subjective assessments and associated bias. This scale measures joint mobility in children with JIA by evaluating ROM. However, it cannot differentiate between ROM limitations caused by active inflammation vs persistent contractures resulting from chronic inflammation and irreversible joint and periarticular soft tissue damage.

The pediatric Escola Paulista de Medicina ROM Scale does not assess ROM in small joints of hands, feet, or temporomandibular joints, which are frequently affected in JIA patients, thus limiting comprehensive functional evaluation[18]. These limitations prompted development of the juvenile arthritis damage index (JADI), a user-friendly instrument specifically designed for evaluating cumulative JIA-related damage[19].

This index comprises both JADI - articular damage (JADI-A) and JADI - extraarticular damage (JADI-E) assessments. The JADI was designed for rapid and convenient calculation based on physical examination and retrospective review of patient medical records. The JADI evaluates damage severity, defined as persistent changes resulting from either previous active disease or treatment-related complications[19].

The articular component (JADI-A) specifically assesses three principal forms of joint damage that persist for ≥ 6 months and are unrelated to current active arthritis: (1) Limited ROM; (2) Ankylosis; and (3) Prior surgical interventions, including joint replacement[19].

The JADI was validated in a cross-sectional study of 158 JIA patients meeting International League of Associations of Rheumatology classification criteria (excluding enthesitis-related arthritis) with ≥ 5 years disease duration. Both JADI-A and CHAQ showed comparable strong correlations with Steinbrocker functional class. However, JADI-A demonstrated: (1) Stronger association with number of joints with limited ROM (ρ = 0.73 vs 0.61) and Poznanski score (ρ = 0.68 vs 0.52); and (2) Weaker correlation with active joint count than CHAQ (ρ = 0.42 vs 0.61).

The extra-articular damage index (JADI-E) significantly correlated with functional class[19]. The JADI has been utilized to identify JIA predictors and establish criteria for biologic therapy initiation. In their landmark study, Russo and Katsicas[20] utilized JADI assessment to demonstrate that cervical spine involvement and corticosteroid use within the first 6 months of disease constitute independent predictors of both global and articular damage in JIA. Key predictors for early biologic therapy identified through JADI evaluation include: (1) Polyarticular JIA subtype; (2) Upper extremity, hip, or ankle joint involvement within 6 months of diagnosis; and (3) > 35% disease activity time during the first year[21,22].

JADI has proven to be a simple, practical tool for comprehensive assessment of: (1) Chronic articular damage; (2) Extra-articular manifestations; and (3) Long-term inflammatory consequences in JIA. Notably, no prior studies have investigated the relationship between joint damage, as quantified by JADI, and treatment response patterns. This study aimed to evaluate the JADI damage index’s prognostic value in predicting the response to biological therapy among patients diagnosed with JIA.

MATERIALS AND METHODS

In the prospective cohort study analysis, 112 biologic-naïve patients with non-systemic JIA categories were included.

Inclusion criteria

(1) Diagnosis of JIA according to the ILAR (International League of Associations for Rheumatology) criteria[23]; (2) Non-systemic JIA categories: Oligoarthritis (persistent and extended), polyarthritis (RF positive and negative), enthesitis-related arthritis, and psoriatic arthritis; (3) Treatment with either non-biologic (nb) or biologic disease-modified antirheumatic drugs (bDMARDs) or their combination; and (4) Age at study inclusion of 2-17 years.

Exclusion criteria

JIA with systemic onset, undifferentiated JIA category.

Study assessments

The following parameters were evaluated in each patient.

Demography: Sex, onset age, disease duration, JIA category.

Clinical assessment: Number of active joints, swollen joints, painful joints, joints with limited ROM, clinical assessment of JIA activity with visual-analogue scale (VAS): Physician’s VAS (MDVAS) and patients/parental VAS, JADAS 71[13]. CHAQ[3,24], assessment of JADI-A and JADI-E[19]. According to the JADI assessment method, the patient was marked as positive if he/she had at least one sign of articular or extraarticular damage[19]. Uveitis was confirmed via split-lamp examination with an ophthalmologist, and the previous records were used if uveitis was detected earlier.

The joint was considered damaged only in the case of persistent limitation ROM unrelated to current active arthritis and present for at least 6 months, despite previous therapy, including exercise and rehabilitation. The damage is often irreversible and cumulative, and therefore, the index value is most often expected to increase or remain stable over time. However, some forms of damage may regress or even disappear over time in children[19].

The articular and extraarticular damage indices JADI-A and JADI-E were measured at the time of each patient examination by the same researcher (Kolkhidova ZA). The methodology for calculating the JADI-A index was based on a routine clinical examination using a goniometer to measure the angles of contractures and calculate the total points according to the table described by the index developers (Supplementary Table 1). This study used the Russian version of the electronic calculator based on the Yandex Forms service (Supplementary material) to simplify the JADI-A calculation.

The JIA activity according to JADAS71 was determined according to previously published cut-offs[13]: (1) Oligoarthritis Polyarthritis; (2) Inactive disease ≤ 1 ≤ 1; (3) Low disease activity 1.1-2.0, 1.1-3.8; (4) Moderate disease activity 2.1-4.2, 3.9-10.5; and (5) High disease activity > 4.2 > 10.5.

Laboratory assessment of JIA activity: RF, anti-cyclic citrullinated peptide (antibodies to cyclic citrullinated peptide), and antinuclear antibody (ANA). ANA was considered positive if its titer was at least 1/320 in the human epithelial type 2 immune-fluorescent assay.

Imaging assessment: Joint X-ray and magnetic resonance imaging (MRI) were performed according to the attending physician’s opinion. An independent radiologist performed interpretation, and the following patterns were analyzed: Early signs of arthritis (periarticular osteoporosis), presence of bone erosions and ankyloses on X-ray, and bone marrow edema, suggesting osteitis or the presence of erosions on MRI. In cases of multiple patterns, the worst indicator was considered.

Treatment: Previous treatment with nbDMARDs (number of drugs and name of the drugs) and corticosteroid use were evaluated before the baseline and during the study.

Methodology

Individual case report forms and an electronic database were developed and completed for all patients. The patients were examined at two control points: (1) At the time of the biological treatment initiation (baseline); and (2) 6-12 months after its start (end of study) in real clinical practice. The baseline examination, which included a detailed assessment of clinical and anamnestic parameters, joint status, and determination of the JADI damage index, was directly performed by the same contractor (Kolkhidova ZA).

The attending physician determined the scope and choice of clinical laboratory and instrumental assessments because this study was based on data from real clinical practice and implied the use of a set of routine laboratory and instrumental tests, as well as additional assessments by the decision of the attending physician, which were determined by a specific clinical situation.

Statistical analysis

The sample size for a studied population with a margin of error of 5% and 95% significance is 87 patients. All statistical calculations were performed with the Statistica 10 software package (Tulsa, OK, United States). For all numerical data, preliminary testing for the normality of the distribution was performed using the Shapiro-Wilk test, and the quantitative data were presented as means with standard deviations or medians with interquartile ranges (IQR) if the distribution was normal or not, respectively. The absolute number (n) and the percentage (%) are reported for categorical variables. Missing data were not included in the analysis. The sets of independent quantitative variables were compared using the Mann-Whitney and χ² tests used for categorical variables. In cases where the expected frequency was lower than 5, Fisher’s exact test was applied. Wilcoxon matched pairs test was performed for two dependent numerical variables. Sensitivity and specificity were assessed for each of the studied parameters to assess the ability of each feature to distinguish patients with dichotomous characteristics. For quantitative variables, critical values were calculated using the area under the receiver-operating curve with a 95% confidence interval (CI), and the odds ratio was calculated using a 2 × 2 table without considering the time of the event of interest. Independent predictors of damage were identified as the event of interest using binary logistic regression by including quantitative and qualitative indicators related to the dependent variable in the analysis. A prognostic model for the risk of a certain outcome (damage) was constructed using multivariate logistic regression. Independent variables were selected using stepwise direct selection with the Wald statistic as an exclusion criterion. All significant predictors were included in the following multiple regression analysis, excluding the duplicated factors to avoid multicollinearity. Multicollinearity was assessed with correlation analysis and clinically meaningful overlap. Differences were considered statistically significant at P < 0.05.

Ethics

Written informed consent was obtained according to the Declaration of Helsinki Study, and participation was voluntary. The local Ethics Committee of the V.A. Nasonova Institute of Rheumatology approved the trial protocol, No. 22, dated 2 December 2021.

RESULTS

Study population characteristics

The cohort (n = 112) had a mean age of 12.6 ± 4.5 years with female predominance (58% vs 42%). Median disease duration was 1.9 years (IQR: 1.0-3.9). All patients exhibited high disease activity [JADAS71: 22 (16.1-27.9)] and received either biologic or synthetic DMARDs.

Biologic/synthetic DMARDs

Biologic/synthetic DMARDs including: Etanercept: 50 (45%); adalimumab: 36 (32%); golimumab: 14 (12%); secukinumab: 3 (3%); and tofacitinib: 9 (8%).

Prior nbDMARD exposure (n = 104, 93%)

Prior nbDMARD exposure (n = 104, 93%) included: (1) Single nbDMARD (n = 84, 75%): Methotrexate: 81 (72%); and sulfasalazine: 3 (2.7%); (2) Sequential nbDMARD combinations (n = 19, 17%): Methotrexate + sulfasalazine: 14 (12.5%); methotrexate + leflunomide: 2 (2%); methotrexate + cyclosporine: 3 (2.7%); and methotrexate + hydroxychloroquine: 1 (0.9%); and (3) Triple therapy (n = 1, 0.9%): Methotrexate + sulfasalazine + leflunomide.

Corticosteroid use

Corticosteroid use: (1) Historical: 17 (15%); and (2) Baseline: 4 patients - 1.25 mg/day (n = 3) and 2.5 mg/day (n = 1).

JADI-positive cohort analysis

The JADI-positive group comprised 60 patients (53.5%), with: (1) 52 (46%) showing articular damage (JADI-A+); (2) 25 (22%) having extra-articular damage (JADI-E+); and (3) 17 (15%) demonstrating both damage types. Damage severity (median scores) were JADI-A: 2 (IQR: 1-4) and JADI-E: 0 (IQR: 0-1). Distribution of joint damage in JADI-positive patients is presented at Table 1.

Table 1 The predictors of damage in patients with juvenile idiopathic arthritis.

Parameter	Se	Sp	OR (95%CI)	RR (95%CI)	P value
Polyarthricular JIA category	35.0	86.5	3.5 (1.3-9.0)	1.6 (1.2-2.2)	0.017
ANA positivity	45.0	75.0	2.5 (1.1-5.5)	1.5 (1.1-2.1)	0.027
Arthritis of ≥ 5 joints at onset	25.0	90.4	3.1 (1.1-9.3)	1.5 (1.1-2.1)	0.034
Active joints > 8 at the study inclusion	55.0	76.9	4.1 (1.8-9.3)	1.8 (1.3-2.6)	< 0.001
Parental/patient VAS > 6 cm	53.3	71.2	2.8 (1.3-6.2)	3.2 (1.6-6.2)	0.009
MDVAS > 6 cm	88.3	50.0	7.6 (2.9-19.7)	3.2 (1.6-6.2)	< 0.001
JADAS71 > 17.7 points	81.7	48.1	4.1 (1.8-9.7)	2.1 (1.3-3.6)	0.001
CHAQ > 0 points	68.3	59.6	3.2 (1.5-6.9)	1.7 (1.2-2.6)	0.003
Ankylosis, confirmed by X-ray	29.8	97.9	19.5 (2.4-155.8)	2.2 (1.7-3.0)	0.008
MRI-confirmed erosions	49.0	90.9	12.5 (3.8-40.9)	2.2 (1.6-3.1)	< 0.001

Se: Sensitivity; Sp: Specificity; CI: Confidence interval; OR: Odds ratio; RR: Relative risk; JIA: Juvenile idiopathic arthritis; ANA: Antinuclear antibody; VAS: Visual-analogue scale; MDVAS: Physician’s visual-analogue scale; JADAS: Juvenile arthritis disease activity score; CHAQ: Childhood Health Assessment Questionnaire; MRI: Magnetic resonance imaging.

Positive extra-articular manifestations were observed in 25 patients (22%): (1) Leg length discrepancy (> 1 cm): 12 (11%); and (2) Imaging-confirmed osteonecrosis: 9 (8%). No cases of growth retardation, pubertal delay, osteoporosis, fractures, diabetes mellitus, or morphologically confirmed amyloidosis were documented. The pattern of JADI-assessed damage across JIA categories is presented in Figure 1.

Figure 1 Baseline juvenile arthritis damage index item distribution in non-systemic juvenile idiopathic arthritis patients. JADI-A: Juvenile arthritis damage index - articular damage; JADI-E: Juvenile arthritis damage index - extraarticular damage; JIA: Juvenile idiopathic arthritis; oJIA: Oligoarthritis (persistent and extended); pJIA: Polyarthritis (rheumatoid factor positive and negative); PsJIA: Psoriatic arthritis; ERA: Enthesitis-related arthritis; GC: Glucocorticosteroids.

Predictors of damage in JIA

Univariate analysis demonstrated the following significant predictors of any structural damage included: (1) Disease characteristics: Polyarticular JIA category; ANA positivity; and baseline arthritis affecting ≥ 5 joints; (2) Disease activity markers: > 8 active joints at enrollment; Parent/patient VAS > 6 cm; MDVAS > 6 cm; JADAS71 score > 17.7; and CHAQ score > 0; and (3) Structural findings: Radiographically confirmed ankylosis; MRI-documented erosions.

Sensitivity, specificity, odds ratios, and relative risk calculations are presented in Table 2. Multivariate regression analysis (r² = 0.33; P = 4 × 10^-8) identified three independent predictors (Table 3): (1) > 8 active joints at enrollment; (2) MDVAS > 6 cm; and (3) MRI-confirmed erosions.

Table 2 The final predictors of damage development (multiple regression analysis data).

Predictor	Β	SE	P value
Intercept	0.108	0.084	0.200
Active joints > 8 at the study inclusion	0.191	0.093	0.044
MDVAS > 6 cm	0.318	0.102	0.003
MRI-confirmed erosions	0.355	0.097	< 0.001

MDVAS: Physician’s visual-analog scale; MRI: Magnetic resonance imaging.

Table 3 Baseline comparison of the studied population, with and without damage, n (%).

Parameter	JADI (+ ve), n = 60	JADI (- ve), n = 52	P value
Age at study inclusion, years, IQR	11.6 (7.5-15.4)	13.1 (9-16.1)	0.244
Sex, females	36 (60.0)	29 (55.7)	0.651
JIA category, according to the ILAR[3]
Oligoarthritis	23 (38.0)	19 (36.5)	0.845
Polyarthritis	21 (35.0)	7 (13.5)	0.017
Psoriatic arthritis	4 (7.0)	4 (8.0)	0.834
Enthesitis-related arthritis	12 (20.0)	22 (42.0)	0.019
JIA onset age, years, IQR	8.7 (3-12)	10.1 (6.9-13.1)	0.079
Diagnosis delay, moths, IQR	3.9 (1-12)	4 (1.5-10)	0.560
JIA duration, years, IQR	2.3 (0.8-4.5)	1.6 (0.9-3.2)	0.360
Time before non-biologic DMARDs, months, IQR	5 (2-14)	6 (2.2-23)	0.491
ANA positivity (> 1:320)	27 (45.0)	13 (25.0)	0.027
RF-positivity	6 (10.0)	2 (3.8)	0.207
ACCP-positivity	6 (10.0)	1 (1.9)	0.078
Enthesitis	6 (10.0)	14 (26.9)	0.019
Uveitis	6 (10.0)	4 (7.7)	0.669
Arthritis of ≥ 5 joints at onset	15 (25.0)	5 (9.6)	0.034
Number of past nbDMARDs
No	3 (5.0)	5 (9.6)	0.345
One	47 (78.3)	37 (71.2)	0.512
Two	10 (16.7)	9 (17.3)	0.872
Three	0 (0)	1 (1.9)	0.943
No episodes of nbDMARDs therapy interruption	45 (75.0)	41 (78.8)	0.631
Past corticosteroid treatment	11 (18.3)	6 (11.5)	0.314
Past corticosteroid treatment duration	-	-	0.622
< 3 months	8 (13.3)	5 (9.6)	-
≥ 3 months	3 (5.0)	1 (1.9)	-
Corticosteroid treatment duration, IQR	1 (0.3-6)	1.8 (0.2-2.6)	0.802
Remission duration ≥ 6 months in the past	5 (8.3)	5 (9.6)	0.812
Active joints, IQR	9 (6-17.5)	6 (4-8)	< 0.001
Swollen joints, IQR	8 (6-15)	5 (3-7.5)	< 0.001
Painful joints, IQR	8 (5.5-15)	6 (3.5-8)	0.001
Joints with limited ROM, IQR	3 (2-4.5)	2 (1.5-2)	< 0.001
Parental/patient VAS (cm), IQR	7 (5-7)	2 (1.5-2)	0.045
MDVAS (cm), IQR	7 (7-8)	6 (4-8)	< 0.001
ESR (mm/hour), IQR	10 (5-24)	11 (5.5-19)	0.896
С-reactive protein (mg/L), IQR	1.8 (0.3-10.5)	1.5 (0.2-5.6)	0.512
JADAS71, IQR	23.5 (20-34.2)	19.1 (15-25)	< 0.001
CHAQ, IQR	0.4 (0-0.8)	0 (0-0.3)	0.003
X-ray changes	47/60 (78.3)	32/52 (61.5)	-
Early signs	25/47 (53.2)	29/47 (90.6)	0.002
Erosions	8/47 (17.0)	2/47 (6.3)	0.286
Ankylosis	14/47 (29.8)	1/47 (3.1)	0.008
MRI-changes	51/60 (85.0)	44/52 (84,6)	< 0.001
Bone marrow edema	26/51 (51.0)	40/44 (90.9)	-
Erosions	25/51 (49.0)	4/44 (9.1)	-

+ ve: Positive; - ve: Negative; JADI: Juvenile arthritis damage index; JIA: Juvenile idiopathic arthritis; DMARDs - Disease-modified antirheumatic drugs; ILAR: International League of Associations of Rheumatology; ANA: Antinuclear antibody; RF: Rheumatoid factor; ACCP: Anti-cyclic citrullinated peptide; nbDMARDs: Non-biologic disease-modified antirheumatic drugs; ROM: Range of the motion; VAS: Visual-analogue scale; MDVAS: Physician’s visual-analogue scale; ESR: Erythrocyte sedimentation rate; JADAS: Juvenile arthritis disease activity score; CHAQ: Childhood Health Assessment Questionnaire; MRI: Magnetic resonance imaging; IQR: Interquartile range.

Baseline comparison of JADI-positive vs JADI-negative cohorts

The JADI-positive group exhibited significantly different baseline characteristics compared to patients without damage. JIA polyarticular category, the initial involvement of five or more joints, irrespective of JIA category, ANA positivity, more active, painful, swollen joints and joints with limited ROM, higher parental VAS and MDVAS, higher CHAQ and JADAS71 scores, radiological findings were typical for the JADI positive group. The detailed characteristics are in Table 4.

Table 4 Comparative analysis of clinical parameters of patients whose juvenile arthritis damage index improved vs remained unchanged/deteriorated on biological disease-modified antirheumatic drugs therapy (end of study), n (%).

Parameter	Improved JADI, n = 41	Non-improved JADI, n = 19	P value
Presence of the JADI–A at the baseline	38 (92.7)	14 (73.7)	0.044
Baseline JADI-A, IQR	2 (1-5)	1 (0-2)	0.014
Patients with uveitis	2 (4.9)	4 (21.1)	0.05
Treatment with nbDMARDs at the EOS	40 (97.6)	14 (73.7)	0.004
Stable nbDMARD treatment during the study	39 (95.1)	12 (63.2)	0.001
Stable bDMARD treatment during the study	36 (87.8)	12 (63.2)	0.03
Bone osteonecrosis (JADI-E)	3 (7.3)	6 (31.6)	0.014
Eye damage (JADI-E)	0 (0)	2 (10.5)	0.034
MRI-confirmed bone erosions, n/n performed (%)	14/37 (37.8)	11/15 (73.3)	0.021
bDMARD treatment:			0.297
Etanercept	19 (46.3)	8 (42.1)
Golimumab	2 (4.9)	3 (15.8)
Adalimumab	14 (34.1)	8 (42.1)
Tofacitinib	4 (9.8)	0 (0)
Secukinumab	2 (4.9)	0 (0)

JADI-A: Juvenile arthritis damage index - articular damage; nbDMARDs: Non-biologic disease-modified antirheumatic drugs; EOS: End of study; bDMARD: Biologic disease-modified antirheumatic drugs; JADI-E: Juvenile arthritis damage index - extraarticular damage; MRI: Magnetic resonance imaging; IQR: Interquartile range.

Biological therapy outcomes

At the time of the follow-up examination, the clinical activity indicators were significantly lower than before the initiation of biological drugs: The median of the JADAS71 index decreased from 22 (16.1-27.9) to 5 (3-9.5), P < 0.001. Low JIA activity, according to the JADAS71 index, was achieved in 35 (31%) patients, moderate activity in 46 (41%) patients, and high activity in 31 (28%) children the degree of JIA activity.

According to Wallace’s criteria, 14 (12.5%) patients reached the inactive stage of the disease. A positive JADI damage index was found on the follow-up examination in 51 (45%) patients. Joint damage JADI-A“+” was recorded in 34 (30%), extraarticular JADI-E“+” in 16 (14%) children, and the presence of both joint and extraarticular damage was noted in 10 (9%) of patients (Figure 2).

Figure 2 Dynamics of changes in juvenile arthritis damage index - articular damage and juvenile arthritis damage index - extraarticular damage indices compared to baseline after initiation of biological therapy. JADI-A: Juvenile arthritis damage index - articular damage; JADI-E: Juvenile arthritis damage index - extraarticular damage.

The JADI-A index decreased from 2 (1-4.5) in 52 children (46.4%) to 2 (1-3) in 34 children (30.4%), with a statistically significant difference (P < 0.001). The median JADI-E score changed from 1 (1-1) in 25 patients (22.3%) to 1 (1-2) in 16 patients (14.3%), although this difference was not statistically significant (P = 0.421). The JADI-E index increased in four patients was observed following cataract surgery initiated after biologic therapy due to a reduction in JIA activity.

JADI progression analysis

We analyzed factors that potentially influenced the improvement in the JADI damage index (Table 5). Joint contractures at treatment initiation are more amenable to reversal than extraarticular lesions. Specifically, the presence of uveitis with irreversible complications, as well as erosions or osteonecrosis at the initiation of biological treatment, adversely affects the reversibility of damage. Additionally, the stability of both conventional and biological therapies following initiation significantly positively impacts the dynamics of the JADI damage index.

Table 5 Comparative analysis of clinical parameters of patients whose JADI-A improved vs remained unchanged/deteriorated on biologic disease-modified antirheumatic drugs therapy (end of study), n (%).

Parameter	Improved JADI-A, n = 38	Non-improved JADI-A, n = 14	P value
Baseline JADI-A, IQR	2 (1-6)	1.5 (1-2)	0.096
Patients with uveitis	2 (5.3)	2 (14.3)	0.278
Treatment with nbDMARDs at the EOS	38 (100)	9 (64.3)	0.001
Stable nbDMARD treatment during the study	37 (97.4)	8 (57.1)	0.001
Stable bDMARD treatment during the study	33 (86.8)	9 (64.3)	0.067
Bone osteonecrosis (JADI-E)	2 (5.3)	3 (21.4)	0.079
MRI-confirmed bone erosions	13/35 (37.1)	8/12 (66.7)	0.075
bDMARD treatment			0.249
Etanercept	17 (44.7)	5 (35.7)
Golimumab	2 (5.3)	3 (21.4)
Adalimumab	13 (34.2)	6 (42.8)
Tofacitinib	4 (10.5)	0 (0)
Secukinumab	2 (5.3)	0 (0)

JADI-A: Juvenile arthritis damage index - articular damage; nbDMARDs: Non-biologic disease-modified antirheumatic drugs; EOS: End of study; bDMARD: Biologic disease-modified antirheumatic drugs; JADI-E: Juvenile arthritis damage index - extraarticular damage; MRI: Magnetic resonance imaging; IQR: Interquartile range.

The largest decrease in JADI (Figure 3) was observed in patients treated with tofacitinib (-4.0 points). The number of patients with improved JADI ranged from 100% (tofacitinib and secukinumab) to 70.4% (etanercept), 63.6% (adalimumab), and 40% (golimumab). Post-hoc evaluation revealed significant findings that stable therapy bDMARD (P = 0.042) and treatment with nbDMARDs (P = 0.006) at the end of study make the greatest contribution to the regression of articular damage (Table 6).

Figure 3 Dynamics of juvenile arthritis damage index during the study in juvenile idiopathic arthritis patients treated with different biological disease-modified antirheumatic drugs. JADI-A: Juvenile arthritis damage index-articular; BL: Baseline; EOS: End of study.

Table 6 Predictors of articular damage improvement (multiple regression analysis data), n (%).

Predictor	Β	SE	P value
Intercept	-0.026	0.198	0.896
Treatment with nbDMARDs at the EOS	0.343	0.119	0.006
Stable bDMARD treatment during the study	0.249	0.119	0.042

EOS: End of study; bDMARD: Biologic disease-modified antirheumatic drugs; nbDMARDs: Non-biologic disease-modified antirheumatic drugs.

JADI-treatment response relationship

The relationship between the severity of JIA based on the JADAS71 score and the frequency of JADI damage has been identified, allowing the JADI score to predict the future effectiveness of biological treatment (Figure 4). We conducted a comparative analysis of the activity parameters at the follow-up assessment, considering initial damage (JADI-positive) at the time of biological treatment initiation. One year after therapy initiation, patients with initial damage exhibited higher values for the JADAS71 and the CHAQ [median 0 (0-0.4) vs 0 (0-0), P = 0.005]. The VAS scores for patients and their parents/guardians [2.5 (2-4) vs 2 (1-3), P = 0.026] and MDVAS also showed increased values [3 (2-4) vs 1.5 (1-3), P = 0.004]. Additionally, patients with initial damage were less likely to achieve disease remission according to the Wallace criteria [n = 4; 6.7% vs n = 10; 19.2%, odds ratio (OR) = 3.3, 95%CI: 0.98-11.4, P = 0.045]. Furthermore, they required intra-articular corticosteroid injections more frequently during the follow-up period (n = 14; 23.3% vs n = 2; 3.8%, OR = 7.6, 95%CI: 1.6-35.3, P = 0.004).

Figure 4 Distribution of damage in patients with different disease activities, according to the juvenile arthritis disease activity score 71 grade at the end of the study, P < 0.001. JADI: Juvenile arthritis damage index; JADAS: Juvenile arthritis disease activity score.

Patients without initial structural damage (JADI “-”) were more likely (OR = 3.8, 95%CI: 1.6-9.0, P < 0.004) to achieve a low degree of activity (n = 24/52; 46.2%) after starting biological therapy, compared with JADI-positive patients (n = 11/60; 18.3%). When treatment was initiated, children with JADI“+” showed a moderate level of disease activity during the follow-up assessment, which was significantly higher compared to JADI“-” (n = 32, 29% vs n = 14, 12.5%, P < 0.05). The high activity degree in patients being on biological therapy did not depend on the initial structural damage (JADI“+” n = 17; 15% vs JADI“-” n = 14; 12.5%) before the start of biological treatment (Figure 5).

Figure 5 The juvenile arthritis disease activity score 71 activity index dynamics during the study in Juvenile idiopathic arthritis patients with and without baseline damage. ^aP < 0.05. JADI: Juvenile arthritis damage index.

A comparative analysis of patients with a low degree of activity and patients with a moderate/high degree of activity according to the JADAS71 index was conducted to find predictors of refractory to targeted therapy. A moderate-high degree of activity (JADAS71) at the follow-up assessment was significantly associated with the following indicators at the time of biological treatment initiation: Joint damage according to the JADI-A index (n = 9; 26% vs n = 43; 56%, P = 0.003), wrist contracture (n = 0; 0% vs 9; 12%, P = 0.035), elbow joints (n = 1; 3% vs n = 14; 18%, P = 0.027), cervical spine limitation of motion (n = 1; 3% vs n = 12; 16%, P = 0.051), and erosions confirmed by MRI (n = 2; 7% vs n = 27; 39%, P = 0.002).

DISCUSSION

Many studies have been conducted globally to predict the response to biological therapy in JIA[25-34]. However, the role of damage indices in this context has been inadequately explored despite compelling evidence suggesting their pivotal significance. These indices are crucial, as they reflect the long-term outcomes of therapeutic interventions[6,13,21,22,35,36]. We assumed that the damage index is the most significant indicator for assessing long-term effectiveness of therapy, integrating all aspects of its effects. At the same time, the number of papers devoted to this problem is extremely limited[6,29,37,38], and most of them are of a one-step, cross-sectional design, rarely dividing patients into those who had no experience with biological therapy and those who had already undergone such treatment. The distinctive feature of our research is the establishment of a prospective follow-up cohort, wherein the alterations in the JADI index were systematically observed during the administration of biological therapy. Additionally, we assessed JADI’s contribution to predicting this therapeutic intervention’s efficacy.

Among 112 patients who did not receive biological therapy, 53.5% had long-term changes manifested by a positive JADI index. JADI-A“+” was diagnosed in 46% of the participants, JADI-E“+” in 22%, and a combined (JADI-A and JADI-E) was detected in 15% of the patients. Similar rates of irreversible changes among children before the biological therapy have been described in studies by Italian and Indian colleagues: JADI-A“+” was registered in 32.4% and 60.7%, and JADI-E“+” - in 26.1% and 39.3%, respectively[39-41]. However, these studies were conducted 16 years earlier, in 2008, when the use of biological drugs had just appeared and was not generally available for children with JIA. The median duration of arthritis in these studies was 6.7 (5.4-9.3) and 5 (1-20) years, while in our cohort, this indicator was significantly lower - 1.9 (1.0-3.9) years. In more modern studies performed in Thailand in 2023[42] and India in 2018[43], the prevalence of JADI-A“+” among patients was 36.6% and 30.6%, and JADI-E“+” was 22.2% and 24%, respectively. At the same time, the median duration of the disease in these studies was comparable to that in our study: 2.7 (0.9-4.9) and 2 (0.6-5.0) years.

In our study, we focused on a group of patients with polyarticular JIA, among whom 75% had a relationship between joint damage and the maximum JADI score of 5 points (2-14). In most cases, the damage index for polyarticular JIA included contractures of the joints of the upper extremities: Elbow (28.5%), proximal interphalangeal (28.5%), and wrist (25%) joints. Such localization of lesions is consistent with the research data of Solari and colleagues, as well as Sarma and coauthors: Involvement of the wrist joints ranged from 23% to 60%, proximal interphalangeal joints - from 24% to 40%, elbow joints - from 22% to 38% in RF-positive and RF-negative poliarticular JIA categories[40,41]. Routine markers of current disease activity, encompassing the number of active joints, the number of joints with limited function, the painful and swollen joints count, the visual analog scale assessments by both the physician and the patient/parent, the JADAS index, and the CHAQ questionnaire, demonstrated a statistically significant association with the presence of joint damage in patients, consistent with the findings of previous studies[27,38,40].

To confirm the validity of the JADI index in evaluating therapy effectiveness, we conducted a comparative analysis between the JADAS71 activity index and the JADI damage index. At the follow-up observation point, a significant correlation was found between these indicators, and a statistically significant change in the damage index during therapy was confirmed.

The prospective study by Susic et al[44] (n = 87, 2-year follow-up) remains the only published work examining JADI progression under biologic therapy (etanercept initiation in 50.6% cohort). The findings revealed paradoxical trends: The proportion of patients with articular damage increased from 36.8% to 41.4%, extraarticular - from 26.4% to 39.1% (P < 0.01). The mean values of JADI-A increased from 3.00 ± 6.51 to 3.62 ± 7.33 (P < 0.001), JADI-E - from 0.59 ± 1.29 to 0.84 ± 1.44 (P < 0.001). The JADI-A raise correlated with radiological progression, disease duration, and glucocorticoid therapy duration. However, the study’s limitations (small sample, lack of control) did not allow us to identify factors influencing the progression of JIA[40,41].

Our study has established the positive dynamics of JADI-A during biological therapy. The prevalence of joint damage decreased from 46.4% to 30.4%, and the median JADI-A decreased from 2 (1-4.5) to 2 (1-3) (P < 0.001). Extraarticular damage at the follow-up assessment was noted in a smaller number of children: 22.3% and 14.3%, respectively. However, the median JADI-E increased from 1 (1-1) to 1 (1-2) (P = 0.421). Considering the potential for reduction of the JADI Index due to the reversibility of contractures, it is advisable to utilize the joint component of the index (JADI-A) to evaluate the dynamic progression of JIA, as opposed to the extraarticular component (JADI-E), which primarily encompasses irreversible changes[19]. The findings of our study with the literature data support the rationale for employing the JADI-A index to assess disease progression and the efficacy of therapeutic interventions.

Study limitations

Despite the significant results obtained, the present study has several limitations to consider when interpreting the data and planning further work. It is an uncontrolled single-center study with real-world data.

Limited sample size and its structure: One of the main limitations of our study is the sample size and the small number of clinical variants of JIA. Although enough patients were included in the analysis, a larger sample could have improved statistical power and provided more reliable results when analyzing subgroups.

Single-center study: Patients undergoing inpatient treatment at one federal center (V.A. Nasonova Institute of Rheumatology) were included; outpatient patients and children observed at their residences by regional rheumatologists were not included. Differences in treatment methods, available resources, and patient populations may affect the generalization of our findings.

Unbalanced sample: The predominance of patients with high disease activity at the start of biological therapy may create a bias in assessing response predictors because initially, more severe cases have less reserve to achieve remission. The article’s authors could not influence the timing of the diagnosis, the appointment of therapy with non-biological DMARDs, or the timing of treatment at the research organization. These factors could have influenced the damage index’s presence and appearance.

Short-term follow-up: The effectiveness of biologic therapy was evaluated after 6-12 months, which is insufficient to analyze long-term outcomes such as damage progression or sustained response to therapy and limits understanding of predicting the course of JIA.

The lack of randomization and a control group of patients on biological therapy or those who do not need it limits the possibility of establishing causal relationships between the JADI index and the response to biological therapy.

Methodological limitations of assessment tools: The JADI index evaluates clinically apparent lesions but does not consider early predictors of progression (for example, subclinical inflammation according to MRI data), leading to an underestimation of the role of preclinical changes.

Statistical limitations: Analyzing multiple predictors (for example, JADI-A“+”, JIA duration > 3 years) without correction for multiple comparisons increases the risk of false positive associations. Insufficient detail of the interaction between risk factors (for example, the collinearity of JADI-A and erosions on MRI) can distort the contribution of individual variables to the model.

The abovementioned factors make it difficult to extrapolate the results to a broader population of patients with various forms and degrees of disease severity.

CONCLUSION

Our findings demonstrate that baseline structural damage assessed by JADI serves as an independent predictor of diminished therapeutic response to bDMARDs, greater functional disability progression and reduced remission rates in JIA. JADI as a risk stratification tool enables timely treatment intensification, including earlier bDMARD initiation in high-risk patients, personalized monitoring protocols when combined with ILAR category, MRI erosion status, and JADAS71 trajectories. Incorporating JADI into standard monitoring improves predictive accuracy beyond conventional tools. The JADI index enhances clinical decision-making through earlier intervention before irreversible damage occurs, objective tracking of structural outcomes and family-centered counseling about long-term risks. This evidence supports JADI’s role as a complementary tool in a modern JIA management paradigm.