Due to the advancement of antidepressant medication, more patients receive treatment in everyday practice, and response/remission is not achieved in the majority of the patients. Several attempts to validate the definition of treatment resistance have been undertaken and the criteria usually encompased to treatment sequences are that two different classes of antidepressants given in a sufficient dosage for a sufficient time have to be administered. Clinically, it is evident that this defintion does not include the treatment refractory patients which could be defined as not responding to numerus trials including electroconvulsive therapy. The goal to improve treatment through the use of predicitve biomarkers has not as yet been obtained and will be the challenge for the future. The clinical variables predicting treatment resistance include comorbid anxiety disorders as well as melancholic features. Although there is a plethora of hints in textbooks that switching the mechanism of action should be obtained when a patient does not respond to one medication, the few controlled trials which have been undertaken to challenge this notion revealed that staying on the same antidepressant mechanism of action for a longer time is more beneficial. In previous times, the conjunctive use of agents such as lithium or thyroid augmentation have been proposed, however the new data achieved with atypical antipsychotics demonstrate a more rapid improvement. Techniques such as brain stimulation or vagus nerve stimulation have also shown promising early results. Treatment resistance continues to be a major barrier achieving remission in patients with major depressive disorder and a thorough re-evaluation of the condition and its treatment is needed.