Tumour necrosis factor-α levels are elevated in adolescent patients with juvenile idiopathic arthritis on etanercept therapy

Methods Serum TNF-a was measured in samples derived from 200 adolescent and young adult patients with JIA attending the adolescent and young adult rheumatology clinic at University College London Hospital using a commercial enzyme linked immunosorbent assay (ELISA) kit (eBioscience). Samples were tested in duplicate. Median age at sampling and median disease duration were 18 years and 8 years 9 months, respectively. Male:female ratio was 1:1.2. Equal numbers of patients with polyarticular (n=64) and enthesitis related arthritis (ERA, n=64) were tested in addition to 48 with oligoarticular, 16 systemic onset, and 8 psoriatic arthritis. Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) measurements were also collected. Furthermore, an L929 cell viability bioassay was used to determine if the addition of etanercept abrogates the cytotoxic effects of TNF-a in L929 cells. Results Surprisingly, TNF-a serum levels were shown to be markedly elevated in patients on etanercept (median TNF-a on etanercept= 134.2pg/ml, IQR [49.4-207.1], median not on etanercept = 4.2pg/ml, IQR [1.4-11.0], p<0.0001). TNF-a levels were also higher in patients on etanercept compared to those on other biologics (adalimumab, infliximab, abatacept, or tocilizumab, median= 4.4pg/ml, IQR [1.8-9.1]) or disease modifying anti-rheumatic drugs alone (median = 4.2 pg/ml, IQR [1.1-12.9]), p<0.0001. In addition, ESR and CRP levels had a negative correlation with high TNF-a levels in patients on etanercept (p=0.0018 and p=0.0034 respectively). Etanercept was included at its therapeutic serum concentration (2.4ug/ml) to ensure there was no cross reactivity with the assay. Finally, we showed that the addition to TNF-a to human serum leads to cytotoxicity in a TNF-a sensitive cell line, while adding etanercept at its therapeutic concentration along with TNF-a significantly reduces cell death (p = 0.0277).