Background: To evaluate if TNF inhibitor serum drug levels (DL) or anti-drug antibodies (ADAb) can predict successful dose reduction (in patients with high DL) or discontinuation (in patients with no/low DL or ADAb) in rheumatoid arthritis (RA) patients. Research design and methods: RA patients that were using adalimumab (n = 42), etanercept (n = 76) or infliximab (n = 51) and were doing well, were tapered until discontinuation or flare (1–1.5 year follow up). Random timed DL for adalimumab and etanercept and trough DL for infliximab were measured before dose reduction: Receiver-Operator-Curves (ROC) analyses with optimal cut-off DL were determined. Results: No predictive value of adalimumab and infliximab DL for all outcomes were found, except for an inverse association of lower etanercept DL and higher chance for successful dose reduction (Area Under the Curve (AUC) 0.36, 95% CI 0.23–0.49; cut-off <2.6 mg/l). In sub analyses, higher adalimumab trough DL predicted successful dose reduction (AUC 0.86, 0.58–1.00; cut-off >7.8). ADAb were infrequent and not predictive of successful discontinuation. Conclusions: No predictive value of baseline adalimumab, etanercept and infliximab DL or ADAb for successful dose reduction or discontinuation in RA was found in this context, with the possible exception of high adalimumab trough levels for successful dose reduction. |