OBJECTIVE: To summarize evidence on the use of upper extremity casting designed to achieve reductions in contracture, tone, pain, function, oedema or spasticity in the elbow, wrist or hand of adults and children with neurological conditions. DATA SOURCES: A search was conducted of the Cochrane Database of Systematic Reviews; the electronic databases MEDLINE, EMBASE, CINAHL, PEDro, OT-Seeker; Google Scholar; reference lists of retrieved trial reports and review articles. REVIEW METHODS: Two independent reviewers determined whether retrieved study abstracts met inclusion criteria: human subjects; >50% of participants children or adults described as having brain injury, cerebral palsy or stroke. Methodological quality of randomized controlled trials was rated using the PEDro scale (1-10 highest). Thirty-one papers were retrieved and 23 studies appraised: three were randomized controlled trials and four were systematic reviews. Over three-quarters of the studies, excluding systematic reviews, were lower level evidence (n = 4 level V; n = 4 level IV; n = 1 level III). Methodological quality of randomized controlled trials was high (PEDro 8, 8 and 9) and there were modest positive short-term outcomes for two trials, although they did not include no-stretch comparison conditions. Safety issues typically included pain or skin breakdown; two adverse events were not cast related. CONCLUSION: While theoretical rationales suggest upper limb casting should be effective there is insufficient high-quality evidence regarding impact or long-term effects to either support or abandon this practice. High variability in casting protocols indicates little consistency or consensus in practice. As maximum or low-load stretch are rationales for cast application, the absence of no-stretch conditions in existing trials is a major weakness in current evidence.