Treatment-resistant depression (TRD) presents major challenges for both patients and clinicians. There is no universally accepted definition of TRD, but results from the US National Institute of Mental Health's (NIMH) STAR*D (Sequenced Treatment Alternatives to Relieve Depression) programme indicate that after the failure of two treatment trials, the chances of remission decrease significantly. Several pharmacological and nonpharmacological treatments for TRD may be considered when optimized (adequate dose and duration) therapy has not produced a successful outcome and a patient is classified as resistant to treatment. Nonpharmacological strategies include psychotherapy (often in conjunction with pharmacotherapy), electroconvulsive therapy and vagus nerve stimulation. The US FDA recently approved vagus nerve stimulation as adjunctive therapy (after four prior treatment failures); however, its benefits are seen only after prolonged (up to 1 year) use. Other nonpharmacological options, such as repetitive transcranial stimulation, deep brain stimulation or psychosurgery, remain experimental and are not widely available. Pharmacological treatments of TRD can be grouped in two main categories: 'switching' or 'combining'. In the first, treatment is switched within and between classes of compounds. The benefits of switching include avoidance of polypharmacy, a narrower range of treatment-emergent adverse events and lower costs. An inherent disadvantage of any switching strategy is that partial treatment responses resulting from the initial treatment might be lost by its discontinuation in favour of another medication trial. Monotherapy switches have also been shown to have limited effectiveness in achieving remission. The advantage of combination strategies is the potential to build upon achieved improvements; they are generally recommended if partial response was achieved with the current treatment trial. Various non-antidepressant augmenting agents, such as lithium and thyroid hormones, are well studied, although not commonly used. There is also evidence of efficacy and increasing use of atypical antipsychotics in combination with antidepressants, for example, olanzapine in combination with fluoxetine (OFC) or augmentation with aripiprazole. The disadvantages of a combination strategy include multiple medications, a broader range of treatment-emergent adverse events and higher costs. Several experimental pharmaceutical treatment alternatives for TRD are also being explored in combination with antidepressants or as monotherapy. These less studied alternative compounds include pindolol, inositol, CNS stimulants, hormones, herbal supplements, omega-3 fatty acids, S-adenosyl-L-methionine, folic acid, lamotrigine, modafinil, riluzole and topiramate. In summary, despite an increasing variety of choices for the treatment of TRD, this condition remains universally undefined and represents an area of unmet medical need. There are few known approved pharmacological agents for TRD (aripiprazole and OFC) and overall outcomes remain poor. This might be an indication that depression itself is a heterogeneous condition with a great diversity of pathologies, highlighting the need for careful evaluation of individuals with depressive symptoms who are unresponsive to treatment. Clearly, more research is needed to provide clinicians with better guidance in making those treatment decisions--especially in light of accumulating evidence that the longer patients are unsuccessfully treated, the worse their long-term prognosis tends to be.