TNF blockade contributes to restore lipid oxidation during exercise in children with juvenile idiopathic arthritis

Thirty children (15 non-TNF blockade and 15 under TNF blockade medication) aged 7 to 18 years old with JIA (according to the International League of Associations for Rheumatology criteria) and 15 healthy controls were enrolled in the study. Participants in the three groups were matched by age, pubertal stage, and gender. All participants were followed between January 2017 and June 2019 at the pediatric unit of the Clermont-Ferrand University Hospital in France. We firstly included patients treated by TNF blockade, then we included matched JIA patients without TNF blockade and healthy controls. Healthy controls were recruited from the entourage (classmates, family, sports club) of the patients.

Participants were excluded if they had a physician-diagnosed infection, had received oral corticosteroids within the previous 3 months, had chronic disease (other than JIA for patients). All treatments had been administered for at least 3 months at the time of evaluation. Disease status (active or inactive) was evaluated according to the American College of Rheumatology criteria [22]. The physical activity level (PAL) was determined according to the International Physical Activity Questionnaire for Adolescents (IPAQ-A), and individuals were classified into one of three levels of physical activity (PA): low, moderate, or high.

The study was approved by the governing ethics committee (Comité de Protection des Personnes [CPP] Sud-Est VI – Clinical trial number NCT 02977416). All participants and their parents gave their informed consent and were free to withdraw from the study at any time.

Participants were asked to refrain from consuming any food or liquid, with the exception of water, 3 h before the visit. Participants were in minimal fasting state of 3 h. They also avoided high fat food and refrained from strenuous physical activity for at least 24 h before the visit. Participants arrived the hospital laboratory for indirect calorimetry test in the morning. After sitting quietly for 20 min, participants performed a graded exercise test to volitional fatigue on an electromagnetically-braked cycle ergometer with continuous gas collection and heart rate monitoring. Following a 2-min warm-up involving unloaded pedaling, Tanner stages 1 and 2 participants started cycling at 10 W and the work rate was increased by 10 W every 3 min. Tanner stages 3 and 4 participants started at 20 W and the work rate was increased by 15 W every 3 min. In cases where the heart rate was unstable, this stage was extended for up to 5 min to obtain a stable heart rate of ±5 beats. When the respiratory exchange ratio (RER) was ≥1.00—indicating the absence of fat oxidation—, the work rate was increased by the same increments at 1-min intervals until volitional fatigue was reached. The VO_2peak was considered to have been reached when the RER was ≥1.05 and the participant achieved his or her age-predicted maximal heart rate (HR_max: 220 − age), according to the methodology validated by Riddell et al. [23].

All the tests were performed on a Cyclus 2 ergometer (RBM Elektronik-Automation GmbH, Leipzig, Germany). The O₂ consumption (VO₂) and CO₂ production (VCO₂) were measured breath-by-breath through a mask connected to an O₂ and CO₂ analyzer (MetaMax 3b, Cortex Biophysik, Leipzig, Germany).

Ventilatory parameters were averaged every 1 min during the submaximal exercise test and during the subsequent 10-min recovery period. The heart rate was continuously monitored over the duration of the tests (Polar RS800cx monitor, Polar, Finland).

Indirect calorimetry is known to be the standard method to quantify substrate oxidation rates at rest and during exercise [24]. The VO₂ and VCO₂ values were averaged over the last minute of each work rate, with the results then being used to calculate fat oxidation over a wide range of exercise intensities for each participant [25] by employing Péronnet and Massicotte’s equation [26]. Lipids (g.min^− 1) = 1.6946 x VO₂–1.7012 x VCO₂ and CHO (g.min^− 1) = 4.585 VCO₂–3.2255 VO₂.

For each individual, a best-fit polynomial curve was constructed for their fat and CHO oxidation rate (expressed as mg.min^− 1) vs. exercise intensity (expressed as % VO_2peak). Each individual curve was then used to determine the peak fat oxidation rate (MFO) and the exercise intensity that was associated with the Fatmax [25].

The sample size estimation was determined according to previous works reported in the literature (Nguyen et al. with 11 patients by group [21]) and for an expected effect size of around 1 (as the standard deviation of the percentage change in the maximum rate of lipid oxidation ranged between 1 and 2.2). For a two-sided type-I error of 1.8% (the correction required due to multiple comparisons) and for a statistical power greater than 80%, it was proposed that 22 participants should be included in each group. Furthermore, due to the lack of data in the literature concerning (1) between- and within- participant variability and (2) the expected differences due to TNF blockade use, an exploratory interim statistical analysis was planned 15 participants by group, which allow to highlight an effect-size greater than 1.2 with aforementioned statistical assumptions.

Statistical analyses were performed using Stata software version 13 (StataCorp, College Station, TX). Tests were two-sided with the type-I error set at α = 0.05. Continuous data were expressed as the mean ± standard deviation (SD) or as the median (interquartile range) according to the statistical distribution (the assumption of normality was assessed with the Shapiro–Wilk test). To account for the between- and within-patient variability due to several measures being taken for the same participant, random-effects models for the correlated data were undertaken rather than the usual statistical tests, as they would have been inappropriate due to an unverified assumption of independence. Time-point evaluations, groups (with/without a TNF blockade and controls) and their interactions were considered as fixed effects, whereas the participant (patient) was considered as a random effect (slope and intercept). A Sidak’s post hoc test for multiple comparisons was applied. The normality of the residuals from these models was studied as described above using the Shapiro–Wilk test. When appropriate, the data were log-transformed to achieve normality of the dependent endpoint. Concerning non-repeated measures, quantitative variables (notably age, body mass index [BMI], disease duration, VO_2peak, VO_2peak.kg^− 1 of body weight, MFO, physical activity level) were compared between groups by ANOVA or by using the Kruskal–Wallis test when assumptions required for the ANOVA were not met (normality and homoscedasticity as analyzed via the Bartlett test). When appropriate (omnibus p-value < 0.05), a post hoc test to account for multiple comparisons was performed: A Tukey–Kramer post-ANOVA and a Dunn test after the Kruskal–Wallis test. As less than 5% of the data were missing, the missing data were not dealt with.

The characteristics of the participants are given in Table 1. No significant difference were observed for age, BMI, VO_2peak, VO_2peak (/kg of body mass), disease activity, physical activity levels, and resting metabolism between the three groups.

The fat and CHO oxidation rates relative to exercise intensities are displayed in Fig. 1a and b. The CHO oxidation rates were the same for each group at the same relative exercise intensities (Fig. 1a). However, lipid oxidation rates were statistically lower in JIA children without TNF blockade than controls from the exercise intensities corresponding to 30% of VO₂ peak up to those corresponding to 70% of VO₂ peak (p = 0.02).

At the relative exercise intensity of 30 to 70% of the VO_2peak, fat oxidation rate was higher in the healthy controls than JIA children treated with TNF blockade (p = 0.14) (Fig. 1b).

The heart rate, power, and %VO_2peak at which the MFO for each group was reached are shown in Fig. 2. MFO was reached at a power of 42.8 ± 17.6, 35.9 ± 20.3, and 29.0 ± 11.2 W (p = 0.19) in the healthy controls, JIA children with TNF blockade and children JIA without TNF blockade, respectively. The heart rate reached at MFO was 135 ± 18, 127 ± 17, and 127 ± 13 beat.min^− 1 in the healthy controls, JIA with TNF blockade and JIA without TNF blockade, respectively (p = 0.40). Furthermore, MFO was reached at 53.6 ± 12.3, 49.9 ± 9.9, and 50.6 ± 11.5 of the %VO_2peak in healthy controls, JIA with TNF blockade and JIA without TNF blockade respectively (p = 0.67).

Maximal fat oxidation rates expressed in mg.min^− 1 and relative to total body mass (Fig. 3) were higher in healthy controls than JIA with TNF blockade and JIA without TNF blockade (225.3 ± 92.9, 163.2 ± 59.0, 134.3 ± 45.2 mg.min^− 1, p = 0.008 and 4.7 ± 1.8, 3.4 ± 0.8, 3.1 ± 1.2 mg.min^− 1.kg^− 1, p = 0.003, respectively).