Background:Several recent trials indicate low-dose ketamine produces rapid antidepressant effects. However, uncertainty remains in several areas: dose response, consistency across patient groups, effects on suicidality, and possible biases arising from crossover trials.Methods:A systematic search was conducted for relevant randomized trials in Medline, Embase, and PsycINFO databases up to August 2014. The primary endpoints were change in depression scale scores at days 1, 3 and 7, remission, response, suicidality, safety, and tolerability. Data were independently abstracted by 2 reviewers. Where possible, unpublished data were obtained on treatment effects in the first period of crossover trials.Results:Nine trials were identified, including 201 patients (52% female, mean age 46 years). Six trials assessed low-dose ketamine (0.5mg/kg i.v.) and 3 tested very low-dose ketamine (one trial assessed 50mg intra-nasal spray, another assessed 0.1–0.4mg/kg i.v., and another assessed 0.1–0.5mg/kg i.v., intramuscular, or s.c.). At day 3, the reduction in depression severity score was less marked in the very low-dose trials (P homogeneity <.05) and among bipolar patients. In analyses excluding the second period of crossover trials, response rates at day 7 were increased with ketamine (relative risk 3.4, 95% CI 1.6–7.1, P=.001), as were remission rates (relative risk 2.6, CI 1.2–5.7, P=.02). The absolute benefits were large, with day 7 remission rates of 24% vs 6% (P=.02). Seven trials provided unpublished data on suicidality item scores, which were reduced on days 1 and 3 (both P<.01) but not day 7.Conclusion:Low-dose ketamine appears more effective than very low dose. There is substantial heterogeneity in clinical response, with remission among one-fifth of patients at 1 week but most others having benefits that are less durable. Larger, longer term parallel group trials are needed to determine if efficacy can be extended and to further assess safety.