The aim of the present study was to critically appraise current conceptual approaches; demographic, neurobiological and clinical correlates; and management strategies of treatment-resistant depression (TRD), especially in light of recent research findings. To this end, a review of the relevant English-language literature was undertaken using Medline, Embase and Psychinfo. TRD has been defined in conceptually restrictive terms as symptomatic non-response to physical therapies alone, with little systematic study of aetiology made. It is likely that a range of sociodemographic (such as higher socioeconomic status), genetic (such as variation in functional monoamine polymorphisms) and clinical variables (such as signal hyperintensities seen on structural neuroimaging scans) are responsible for non-response in individuals. There is insufficient evidence to suggest that TRD is associated with specific subtypes of depression, physical comorbidity, personality or chronicity. The large-scale Sequenced Treatment Alternatives to Relieve Depression (STAR*D) and other studies have suggested that a structured psychotherapy such as cognitive behaviour therapy may be as effective as medication in initial drug non-responders. Also conventional alternatives such as the use of older antidepressant classes, pharmacological augmentation or electroconvulsive therapy in established cases of TRD are not as effective as traditionally thought. There is insufficient preliminary evidence to make formal recommendations about the use of novel brain stimulation techniques in TRD. TRD should be re-defined as the failure to reach symptomatic and functional remission after adequate treatment with physical and psychological therapies. Treatment resistance may be more usefully conceived within the context of well-defined cohorts such as patients with specific subtypes of depression. Although neurobiological markers such as gene polymorphisms, which are potentially predictive of medication tolerance and efficacy, may be used in the future, it is likely that sociocultural variables such as beliefs about depression, and evidence-based treatments for it, will also determine treatment resistance.