The introduction of medications that target specific proinflammatory cytokines has revolutionized the management of patients with rheumatoid arthritis. The agents that antagonize the effects of tumour necrosis factor (TNF)-alpha -- infliximab, etanercept and adalimumab -- have consistently shown very good efficacy for controlling the clinical and radiographic manifestations of the disease. However, it has become apparent that some patients will receive no clinical benefit, gradually lose the effect over time or experience adverse effects with the TNFalpha antagonists. The management of these patients is challenging and there are no clear guidelines. The concomitant administration of a disease-modifying antirheumatic drug, such as methotrexate, has been shown to improve outcomes. Optimization of the methotrexate or TNFalpha antagonist dose may lead to improved responses, as demonstrated in some dose escalation studies. Switching to another TNFalpha antagonist is a step that is supported by small, mostly uncontrolled studies. Finally, the T-cell co-stimulation antagonist abatacept, as well as the B-cell depleting agent rituximab, are also available for use in patients who have had an inadequate response or intolerance to the TNFalpha antagonists.Genotypic studies have identified TNF and TNF receptor polymorphisms that appear to predict independently whether a patient will respond to a TNFalpha antagonist, but genotyping is not available for routine use in clinical practice. Until such tools for predicting response are widely available, the management of patients with poor responses to TNFalpha antagonists will have to depend upon the wishes of the patient regarding medication dosage schedules and adverse effect profiles, as well as how comfortable the treating physician is with the available biological medications. In this article, we review the current data and construct an algorithm to help guide clinicians in the management of patients with inadequate responses to the TNFalpha antagonists.