The assessment of functional disability is a fundamental measurement in children with juvenile idiopathic arthritis (JIA). Self-reported or parent proxy-reported physical functioning with questionnaires is one of the core outcome measures in randomised clinical trials1 as well as in long-term outcome studies,2 3 and has been recommended for incorporation in standard patient care.4 5

A number of studies have investigated the functional ability of children with JIA, and most of them have documented decreased levels compared to healthy children. However, all analyses have been based on a cross-sectional assessment, either at a single point in time along the disease course6^–10 or at last follow-up observation in the context of a long-term outcome survey (see Oen2 and Ravelli3 for reviews). However, because physical functional disability is known to fluctuate over time, cross-sectional assessments of physical function at a single point in time may provide a misleading picture of the burden of disability. For example, patient outcomes estimated solely on cross-sectional measurements are unable to distinguish early development of disability from late. Furthermore, it is unknown whether the physical limitations observed at a single point in time reflect short-lived, episodic, or chronic disability. From a clinical perspective, it would be more useful to understand the pattern of physical disability over time. Studies in adult patients with chronic arthritis have shown that some patients experience a steady decline in function over time, and others experience either steady improvement or a highly variable course.11 12 Furthermore, the reversible component of functional limitations has been found to be much larger than expected.13 Because children have a much greater potential for recovery than adults, it is likely that the latter phenomenon is much more pronounced in patients with JIA. It is therefore important to evaluate the course of functional disability over time to document whether and to what extent the degree of functional impairment reflects irreversible (or, at least, persistent) or reversible phenomena.

Several studies have sought for predictive factors of physical disability in children with JIA, but the results are conflicting (reviewed in Ravelli and Martini).14 Factors that have been suggested include female sex, young age at onset, presence of IgM rheumatoid factor, symmetric polyarthritis, greater severity of joint disease, long duration of active disease, hip involvement and early radiographic progression. Male sex, persistent extra-articular symptoms, higher platelet count and the carriage of the −173 single-nucleotide polymorphism of the macrophage migration inhibitory factor (MIF) gene have all been identified as specific determinants of disability in systemic JIA, whereas a high erythrocyte sedimentation rate and the presence of ankle and/or wrist disease have been associated with worse functional outcome in oligoarticular JIA. However, the relationship between the level of initial disability and the degree of long-term functional impairment has never been examined.

The primary objective of the present study was twofold: first, to investigate the time course of functional disability in patients with JIA who received multiple longitudinal assessments; and second, to identify early predictors of long-term functional outcome, including initial functional status among candidate predictors.

PATIENTS AND METHODS

Study design and patient selection

The clinical charts of all consecutive patients seen at the study units between January 1987 and December 2002 who had JIA, as defined by the International League of Associations for Rheumatology (ILAR),15 were followed for at least 1 year and had two or more functional ability questionnaires completed by a parent reviewed. All patients seen before the publication of ILAR criteria for JIA were re-classified using such criteria. Due to their peculiar clinical features, particularly the presence of enthesitis and reduced spinal mobility, patients with enthesitis-related arthritis or juvenile ankylosing spondylitis were believed to deserve a distinctive functional ability measure and were thus excluded from the study.

Assessment of functional ability

Functional ability assessment was carried out by asking a parent of each patient (generally the mother) to complete the Italian version of the Childhood Health Assessment Questionnaire (C-HAQ).16 17 The C-HAQ measures the child’s ability in performing functions in eight areas: dressing and grooming, arising, eating, walking, hygiene, reach, grip and activities. Each question is scored from 0 to 3 (0 = no difficulty, 1 = some difficulty, 2 = much difficulty, 3 = unable to complete task). The question with the highest score determines the score for that functional area. If aids or devices are used or help is needed to complete tasks in a certain area, a minimum score of 2 is recorded for the corresponding functional area. The scores for each of the eight functional areas are averaged to calculate the C-HAQ disability index, which ranges from 0 to 3 (0 = best; 3 = worst).

In the years prior to 1994 (when the C-HAQ was published), functional ability was measured using either the modified Lee Index18 or the Juvenile Arthritis Functional Assessment Report (JAFAR).19 In order to standardise scores from all functional ability tools, scores from the modified Lee Index and the JAFAR were proportionally converted to the 0–3 scale of the C-HAQ. Combination of scores from the different instruments for the purpose of the analysis was felt to be justified as a very high correlation among the three instruments, when administered to the same patient on the same day, has been previously observed.20 For this reason, and for simplicity, functional ability scores obtained from the different instruments will be collectively referred to as C-HAQ scores. All C-HAQ assessments completed from the first to the last study visit were collected. The design of the study was approved by the independent Institutional Review Board of the Gaslini Institute of Genoa, Italy.

Assessment of predictor variables

At the time of first C-HAQ completion, the following predictor variables were recorded for each patient by reviewing clinical charts: sex; age at disease onset; ILAR category; age at visit; disease duration; count of joints with swelling, pain upon movement/tenderness, restricted motion and active disease;20 presence of antinuclear antibodies; erythrocyte sedimentation rate (ESR); and C-reactive protein (CRP).

Statistical analysis

For purposes of the analysis, C-HAQ scores were divided into three categories: 0–0.49; 0.5–1.5; and 1.51–3, representing physical functional disability states 1, 2 and 3, respectively. State 1 represented the absence of disability, while states 2 and 3 represented “mild to moderate” and “severe” disability, respectively. These cut-offs were consistent with past studies, where C-HAQ scores less than 0.50 represented no disability (ie, few difficulties in performing daily activities) and scores above 1.50 reflected a severe level of disability (ie, considerable difficulties or assistance required in performing daily activities).21 22 At each C-HAQ administration, a patient was assigned to one of these three disability states.

The course of physical functional ability over time was characterised first, and it was anticipated that patients would show steady deterioration or improvement in physical functional disability, or fluctuations between functional disability states, over their follow-up period.

Subsequently, early predictors of long-term functional outcome were searched for. Only patients with a minimum disease duration between disease onset and last C-HAQ administration of 3 years (n = 178) were included in this analysis. The separate (univariate) and joint (multivariate) effects of selected demographic and clinical variables recorded at the time of the first C-HAQ administration on the physical functional disability state at last C-HAQ administration (dichotomised to a binary outcome as state 1 or state 2 or 3) were examined. Predictor variables were sex, onset age, ILAR category, disease duration, joint counts, antinuclear antibody (ANA) status, functional disability state at first C-HAQ administration, ESR and CRP. In univariate analysis, comparison of quantitative variables was made by means of the Mann–Whitney U test, whereas comparison of qualitative variables was performed by means of the χ² test or the Fisher exact test in case of expected frequencies less than 5. For each predictor variable, unadjusted odds ratios (OR) and 95% CIs were calculated. A multiple logistic regression analysis was then carried out entering predictor variables as explanatory variables and the physical functional disability state at last follow-up observation, dichotomised as indicated above, as the outcome variable. Cases with missing data were excluded from multivariate analyses.

Before performing univariate and multivariate analyses, continuous predictor variables were converted to binary variables using the cut points obtained through the receiver operating characteristic (ROC) curve analysis. Variables that were significantly associated with the outcome in bivariate analyses were entered in multivariate procedures. Multivariate analysis was adjusted for disease duration between disease onset and final C-HAQ administration, which was widely variable across patients, by forcing it in the model even if it was not significantly associated with the outcome in bivariate analysis. Using a backward selection procedure, predictor variables that were significantly associated with the outcome were identified. The effect was expressed in terms of adjusted odds ratio (OR_adj) and 95% CI were calculated; statistical significance was tested by means of the likelihood ratio test (LR test). The area under the ROC curve (AUC-ROC) of the best-fitting model was used as an indicator of its predictive ability.

All statistical tests were two sided; a p value of less than 0.05 was considered as statistically significant. The statistical packages used were Statistica (StatSoft Corp., Tulsa, Oklahoma, USA) and Stata, V.7 (Stata Corporation, College Station, Texas, USA).

RESULTS

A total of 227 patients who had 1356 C-HAQ completed and 949.7 patient years of follow-up between first and last C-HAQ administration were identified. Patients in this series represent 50–60% of all patients with JIA attending the authors’ institutions in the study period. Demographic and clinical features of patients at their first C-HAQ administration are shown in table 1. The initial disability state was 1 in 113 patients (49.8%), 2 in 91 patients (40.1%) and 3 in 23 (10.1%) patients. There was a relatively greater proportion of patients with systemic arthritis in state 3 and of patients with persistent oligoarthritis in state 1. As expected, the severity of joint disease increased progressively from state 1 to state 3. Patients in state 3 had higher values of ESR and CRP than those in states 1 and 2.

Table 2 summarises the number and types of transitions in disability states observed over the follow-up period. The following three longitudinal patterns of physical disability were observed: (1) a stable state of disability throughout the study period, with 63 (27.8%) and 8 (3.5%) patients experiencing continued absence of disability or persistently moderate disability, respectively; (2) a steady improvement or deterioration in disability over time, which was observed in 52 (22.9%) and 13 (5.7%) patients, respectively; (3) a fluctuating course of disability, with deterioration and improvement, which occurred in 91 (40.1%) patients. For the vast majority of patients, the observed transitions in disability occurred between states 1 and 2. Steady improvement in disability was observed in 48.3% and 34.8% of patients with initial disability state 2 and 3, respectively. Patients with initial disability state 3 experienced more frequently a fluctuating course of disability than those with initial disability states 2 or 1 (65.2% vs 41.8% and 33.6%, respectively). Very few individuals were observed moving directly from states 1 and 3. At the end of follow-up, the disability state was 1 in 168 patients (74%), 2 in 52 patients (22.9%) and 3 in 7 (3.1%) patients (table 3). Patients who were in state 1 at initial assessment were more likely than those who were in state 2 or 3 to remain in this state at final visit; patients who were in state 2 or 3 initially had a similar likelihood of being in state 1 or 2 at final visit.

The results of bivariate analysis of baseline predictors of disability state al final visit in the 178 patients who had a disease duration ⩾3 years between disease onset and last C-HAQ administration are shown in table 4. A disability state 2 or 3 at last follow-up observation was significantly associated with an age at disease onset younger than 7.3 years, a count of tender and restricted joints at first C-HAQ administration >8 and >10, respectively, and an initial disability state 2 or 3. Table 5 presents the best-fitting model obtained through logistic procedures, which was controlled for the disease duration at last C-HAQ administration. According to this model, patients who had a younger age at disease onset and a greater number of restricted joints at first C-HAQ administration carried the highest risk of developing long-term physical disability. The AUC-ROC of the model was 0.69, meaning moderate predictive power.

DISCUSSION

This study is the first to examine the longitudinal course of physical functional disability in JIA. A wide within-patient and between-patient variability in the time course of disability was found. Nonetheless, three longitudinal patterns could be characterised. The first reflected a stable state of disability throughout the entire study period, with 27.8% and 3.5% of the patients experiencing no disability or moderate disability, respectively. The second pattern was one of either steady improvement or deterioration in disability over time and was observed in 22.9% and 5.7% of the patients, respectively. The third pattern was characterised by multiple transitions and fluctuating states of disability over the study period and was seen in 40.1% of the patients. In spite of the wide variability in disability course, at last follow-up observation three quarter of the patients had no disability, 23% had moderate disability and only 3% had severe disability. That only a small proportion of patients had serious functional impairment at final visit confirms the tendency toward a marked improvement in functional outcome seen in recent studies.2 3 23 Furthermore, this finding, together with the high frequency of the fluctuating pattern of disability, suggests that C-HAQ scores in children with JIA is largely affected by acute reversible factors (ie, functional limitations due to current disease activity) throughout the whole disease course. Studies of the longitudinal correlation between functional disability measures and clinical indicators of disease damage, such as scores of radiographic progression24^–27 or the Juvenile Arthritis Damage Index,28 are needed to dissect the acute reversible components from the chronic irreversible components of physical disability. This information could have important implications for standard patient care and clinical trials because it would help identify the portion of the disability process that is likely or not likely to respond to aggressive treatment with antirheumatic drugs.

In bivariate analysis of risk factors for disability, an age younger than 7.3 years at disease onset, a count of tender and restricted joints greater than 8 and 10, respectively, at baseline, and an initial disability state of 2 or 3 were found to predict a disability state of 2 or 3 at last follow-up observation. Multivariate analysis, which was controlled for duration of follow-up between first and last C-HAQ administration, showed that a younger age at disease onset and a greater count of restricted joints were the strongest independent predictors of worse physical disability at last follow-up. Younger onset age and severity of polyarthritis have been identified as predictors of worse functional outcome in previous analyses.21 29 This means that children who have an early disease onset and experience polyarthritis with prominent restriction in joint motion carry the greatest risk of developing long-term functional impairment. These clinical features may, thus, identify the subgroup of patients with JIA that deserves the most precocious and aggressive therapeutic interventions aimed to prevent progression to irreversible physical disability.

This study must be viewed in light of several potential limitations. We should recognise that re-classification of patients seen before the publication of ILAR criteria might have lead to an inaccuracy in the definition of some of the categories of JIA. Furthermore, the use of different functional ability tools over the years might have affected the validity of the results. During the 16-year time frame of the study (1987–2002) there have been a number of major shifts in the treatment of JIA. Until the end of the 1980s there were virtually no drugs of proven benefit for this disease, whereas in the last decade there have been major therapeutic advances including the widespread use of methotrexate and intra-articular corticosteroids, the earlier introduction of these drugs and, in recent years, the availability of new biological agents.30^–32 These progresses have been accompanied by less use of long-term steroid treatment. Although an influence of these changes on the study findings cannot be excluded, it is unlikely that differences in the therapeutic choices across the attending doctors have biased the results because in the authors’ group these choices have been quite consistent during the study period. The effect of exercise or physical therapy, which could influence the level of disability, or of additional factors that may predict changes in the level of disability, such as comorbid conditions, radiographic joint damage, coping strategies, psychological distress, or parental depression, was not investigated. Most children in this study had minor rather than major disabilities, which may have limited the generalisability of findings. However, the study patients are representative of the authors’ clinic population and are likely to be representative of the patients seen in most tertiary paediatric rheumatology centres. The study patients represent only a fraction of those followed at the authors’ institutions in the study period, possibly resulting in sampling bias (ie, potentially skewing results toward patients having prolonged contact with the authors’ institutions). However, it is likely that most of the patients discarded had a benign, short-lived disease, whereas the majority of the sickest children with arthritis were included. Therefore, the clinical meaning of the study findings is unlikely to have been affected by the sampling method.

In summary, the time course of physical functional disability in children with JIA was found to be widely variable within and between patients. After a median follow-up of 3.4 years, 75% of the patients had no disability and only 3% had severe disability, which confirms that currently most children with JIA have a fairly good prognosis for functional outcome. Patients with a younger age at disease onset and a greater number of joints with restricted motion at first C-HAQ administration were found to have the highest risk of developing long-term physical impairment. These clinical characteristics may identify the patient subgroup that deserves the most aggressive and precocious therapeutic interventions aimed to prevent progression to non-reversible functional disability.