A potential application of therapeutic drug monitoring is to predict efficacy after switch to another biological in the case of inefficacy of the previous TNF-inhibitor (TNFi) in rheumatoid arthritis (RA) patients. It has been shown that when antidrug antibodies against adalimumab are detected (resulting in lower drug serum concentrations) in patients failing adalimumab, a normal response to a next TNF blocker can be anticipated. However, when clinical response is unsatisfactory and no antidrug antibodies against the first TNFi are detected (generally drug levels are adequate in this case), this predicts a lower response to a next TNFi. This means drug resistant failure in the former, compared to class resistant failure in latter category of patients. The current RA treatment strategy after failure of the first TNF-inhibitor is to start either a second TNFi or a non-TNFi. However, by channelling patients with sufficient adalimumab concentration to a non-TNFi will provide higher chance of disease control. Patients with very low or undetectable drug levels have an equal or potential higher chance of disease control with a drug of the same class (i.e. another TNFi).

Over the last decades biopharmaceuticals such as agents against tumor necrosis factor (TNF), are frequently prescribed to optimize rheumatoid arthritis treatment. Although TNFi such as adalimumab, etanercept and infliximab, have improved the treatment of rheumatoid arthritis, a proportion of patients discontinue the treatment because of inefficacy or intolerance.

Where TNFi have failed, mainly two treatment approaches are available: switch to another TNFi or to a biological with a different mode of action (notably rituximab, abatacept or tocilizumab) or to a target synthetic DMARDs. The EULAR recommendation for the management of rheumatoid arthritis advocate that any biologic agent including a subsequent TNFi can be used with equal chance for effect in case of non-response to a previous TNFi. This recommendation is based on three randomized controlled trails, but it should be noted that systematic reviews also including non-randomized controlled studies sometimes concluded that non-TNFi are more effective after TNFi failure than a second TNFi.

Of note, currently there are no strong predictors available for response to bDMARDS in RA. This study focuses on non-response after adalimumab, as this is the most prescribed TNFi worldwide.

Although it seems that indeed on a group level response to a non-TNFi is comparable to a second TNFi after the first TNFi has failed, using therapeutic drug monitoring could identify subgroups of patient who would benefit more from either a non-TNFi or a TNF-i as next treatment. The underlying pathophysiological mechanisms for this hypothesis are explored in this study.

Nonresponse on adalimumab in RA can have different causes. Firstly, the patient might not be sensitive to TNF blockade at all, or the patient develops this trait later on (primary nonresponse or secondary nonresponse). In these patients, switching to a non-TNFi might conceptually be superior to starting a second TNFi. However, in other patients nonresponse (either primary or secondary) might be caused by inefficient drug concentration because of development of antidrug antibodies against adalimumab. In these patients a TNFi might be just as effective as a non-TNFi, as these patients have drug- but not class failure. Thus, testing of adalimumab levels might be helpful in channelling patients to their most optimal treatment.

Above mentioned hypothesis has been tested in three studies (5) of which one concerned adalimumab in RA. This study showed that response to a second TNFi was indeed higher in patients with antidrug antibodies (ADA) and low adalimumab levels, and lower in patients with adequate adalimumab levels. However, a diagnostic study comparing a serum concentration guided versus usual care switching strategy has not yet been performed.

In this multi-centre, double blinded randomized controlled trial it will be evaluated whether clinical outcome after switching to a subsequent biological treatment is superior with a switching strategy using adalimumab concentration compared to usual care switching strategy.

Patients with rheumatoid arthritis starting another bDMARD (biologic disease modifying antirheumatic drugs) after adalimumab failure (defined as DAS28-CRP >2.9) are randomly assigned to usual care (with randomized switch to etanercept or to a non-TNFi) or 'drug concentration' guided switch.

Data regarding disease status, functioning, adverse events and use of co-medication/rescue medication will be collected during this study.

Inclusion Criteria:

• rheumatoid arthritis patient, according to American College of Rheumatology (ACR) 1987 and/or EULAR/ACR 2010 criteria;
• recently failed treatment with adalimumab (defined as DAS28-CRP >2.9) and not treated with a subsequent bDMARD or target synthetic DMARD (tsDMARD)
• Received adalimumab for at least 10 weeks in standard dosing (40mg subcutaneously every other week, either in monotherapy or combined with methotrexate or leflunomide)
• Stop adalimumab due to inefficacy, either alone or combined with side effects
• who has agreed to participate (written informed consent);
• age 16 years or older.

Exclusion Criteria:

• Treatment with another TNF-inhibitor prior to adalimumab
• Treatment with all non-TNFi options (abatacept, rituximab, sarilumab and tocilizumab) prior to adalimumab
• scheduled surgery during the follow-up of the study or other pre-planned reasons for treatment discontinuation
• life expectancy shorter than follow-up period of the study;
• no possibility to safely receive an TNF-inhibitor or a non TNF-inhibitor

To avoid duplication of research, the gathered data will be shared once all desirable data analysis have been performed and the results are published

Six months after the study is published the data will be shared

Researchers with demonstrable interest in autoimmunity, biologicals, or therapeutic drug monitoring (TDM) can contact the investigators of the trial if they are interested in gaining access to the data. Depending on their research objectives the data will be shared

Rheumatologists and patients remain blinded for allocation.