Aerobic capacity and disease activity in children, adolescents and young adults with juvenile idiopathic arthritis (JIA)

Aerobic capacity (VO_2peak) was assessed using a graded cardiopulmonary exercise test (CPET) to volitional exhaustion performed on an electronically braked cycle ergometer (Lode examiner, Lode BV, Groningen, the Netherlands). The seat height was adjusted to the patient’s comfort. Patients began cycling at a workload of 0 Watts and this increased by 20 Watts every minute until they were no longer able to continue cycling due to volitional exhaustion, despite strong verbal encouragement. Patients breathed through a mouthpiece, which was connected to a calibrated metabolic cart (Oxycon Pro, Care Fusion, Houten, the Netherlands). Expired gases were collected and analysed for breath-by-breath minute ventilation (Ve), oxygen uptake (VO₂), carbon dioxide exhalation (VCO₂) and respiratory exchange ratio (RER; =VCO₂/VO₂) using conventional equations. The highest achieved oxygen uptake averaged over a 30-second period was taken as VO_2peak. Heart rate (HR) was measured continuously via 3-lead ECG. To allow comparison of aerobic capacity with healthy controls, standardized z-scores were determined using: Z-scores = [(observed VO _2peak - predicted VO _2peak ) / SD] where predicted VO_2peak values were obtained from a database of established values for age- and gender- related Dutch controls [37]. A z-score of zero represents a VO₂ similar to that of healthy controls, while a score of ±1.5 is a clinical important difference from healthy controls.

A number of indicators were used to assess disease activity including the total number of swollen joints, number of painful joints, the number of joints with limited range of motion, number of active joints (defined as joints with swelling and/ or pain during motion), the Juvenile Arthritis Disease Activity Score (JADAS 27) [38] and the Disease Activity Index 28 (DAS28)[39]. All of the examinations were performed by an experienced (paediatric) rheumatologist. Furthermore, a global assessment was performed by the patient (well being; VAS patient 0-100 mm) as well by a paediatric rheumatologist (disease activity; PGA 0-100 mm). Given that patients were recruited from both the paediatric and adult clinics, disease activity scores commonly used in both the paediatric and adult departments were computed. In adult Rheumatoid Arthritis (RA) patients, the DAS28 is a commonly used and validated tool to measure disease activity [39]. This is a compound score of the number of a total of swollen joints (out of 28), the number of a total of painful joints (out of 28), assessment of patient’s general health and the erythrocyte sedimentation rate (ESR). Three compound disease activity scores were recently introduced in paediatric rheumatology: Juvenile Arthritis Disease Activity Score (JADAS) 10, JADAS 27 and JADAS 71[38]. We utilised the JADAS 27 compound score because of its feasibility. Moreover, statistical performance of JADAS 27 is comparable to JADAS 71, while the construct validity of JADAS 10 is poor compared to JADAS 71 and JADAS 27 [38]. The JADAS 27 may present an advantage over the DAS 28 because it includes measures for the cervical spine, hips and ankles, joints which are often affected in (adult) patients with JIA.

Past and current medication utilization was obtained from a review of the patient’s medical chart.

All statistical analyses were performed using Statistical Package for Social Science (SPSS) version 15.0 for windows (SPSS Inc., Chicago, IL) version 15.0 for Windows. Because patient data were non normally distributed non-parametric tests were used to analyse the demographic variables and compare the results of the cardiopulmonary exercise tests (Mann–Whitney and Kruskal-Wallis where appropriate). Spearman’s Correlation test was used for determining the correlations between VO_2peak and disease activity markers and standard laboratory values. The level of statistical significance was set at P < 0.05.

Demographic and disease related characteristics are presented in Table 1. In total, 22 boys and 41 girls participated in the study. The longest disease durations were seen in the oldest age group. No significant differences were noted between age groups with respect to disease duration, body mass index (BMI), disease activity markers as measured by number of swollen and tender joints, DAS28, JADAS 27 and laboratory values.

The median peak heart rate was 187 beats per minute (Interquartile range [IQR] 17), indicating an adequate level of maximum achieved exercise. The median aerobic capacity in all patients with JIA was considerably lower compared to healthy peers, with a median z-score of −0.8 (p < 0.01) for absolute VO_2peak (L/min) and −1.5 (p < 0.01) for relative VO_2peak/kg (ml/kg/min) in all patients. VO_2peak did not differ significantly by gender, disease duration or JIA sub-type; therefore all patients were merged for further analysis. There were no statistical significant differences in the z-scores for absolute and relative VO_2peak between the different age groups (Figure 1).

Spearman’s correlation coefficient revealed a significant negative correlation between the number of swollen joints and absolute VO_2peak z-score, but not with relative VO_2peak z-score (Table 2). No significant correlations were seen between VO_2peak values and the number of painful joints. Both disease severity compound scores (DAS28 and JADAS 27) showed significant moderately strong negative associations with all VO_2peak values (Table 2).

A substantial proportion of the patients presented with low disease activity as reflected by low median numbers of swollen and painful joints (Table 2). The patients were divided into 2 groups based on disease activity status, these included patients in remission or inactive disease (defined by Wallace et al. [40], active joint count zero; DAS28 ≤2.6) and a patient group with disease activity (active joint count one and higher; DAS28 > 2.6) [40, 41]. Between these patient groups, significant differences as measured by active joint count and DAS28 score were seen (Figure 2).

Significant negative correlations were observed between laboratory parameters reflecting disease activity (ESR, thrombocytes) and the aerobic capacity (Table 3). Conversely, no relationship was seen between C-reactive protein, an acute phase protein and VO_2peak (Table 3). We found a positive significant correlation between serum haemoglobin and the relative VO_2peak/kg in boys as well in the total patient group. In girls, a positive significant correlation between serum haemoglobin and the absolute VO_2peak (p < 0.05, table 3) was observed. Lower VO_2peak and VO_2peak/kg values were noted in patients with anaemia compared to those who were not anaemic (p < 0.05).

Thirteen of our patients were not using any DMARD at the time of the testing, 38 patients were using one DMARD, 11 patients were using a combination of two DMARD’s and only 1 patient was on a combination therapy of three DMARD’s. Non-parametric tests revealed no significant differences in aerobic capacity between these patients. No significant correlations were found between current use of DMARD and absolute and relative VO_2peak values. When looking at historical DMARD use, only use of anti-Tumor Necrosing Factor biologicals in 10 patients was significantly associated with lower absolute and relative VO_2peak values (p < 0.05). Duration of current and formerly MTX use did not affect aerobic capacity.

A total of 6 patients were currently using corticosteroids at the time of testing, while 10 patients had used them the past and 47 patients never used corticosteroids. Neither current nor former corticosteroid use was significantly related to aerobic capacity.