OBJECTIVE The authors systematically reviewed the management of treatment-refractory depression in older people (defined as age 55 or older). METHODS The authors conducted an electronic database search and reviewed the 14 articles that fit predetermined criteria. Refractory depression was defined as failure to respond to at least one course of treatment for depression during the current illness episode. The authors rated the validity of studies using a standard checklist and calculated the pooled proportion of response to any treatment reported by at least three studies. RESULTS All the studies that met inclusion criteria investigated pharmacological treatment. Most were open-label studies, and the authors found no double-blind randomized placebo-controlled trials. The overall response rate for all active treatments investigated was 52% (95% CI=42-62; N=381). Only lithium augmentation was assessed in more than two trials, and the response rate was 42% (95% CI=21-65; N=57). Only two studies included comparison groups receiving no additional treatment, and none of the participants in these groups responded. In single randomized studies, extended-release venlafaxine was more efficacious than paroxetine, lithium augmentation more than phenelzine, and selegiline more than placebo. CONCLUSIONS Half of the participants responded to pharmacological treatments, indicating the importance of managing treatment-refractory depression actively in older people. The only treatment for which there was replicated evidence was lithium augmentation. Double-blind randomized controlled trials for management of treatment-refractory depression in older people, encompassing pharmacological and nonpharmacological therapies and populations that reflect the levels of physical and cognitive impairment present in the general older population with depression, are needed.