ObjectiveTo assess the effect of surgical treatment or conservative treatment on the clinical outcomes for displaced 3-part or 4-part proximal humeral fractures in elderly patients. MethodsWe searched CNKI (between January 1979 and April 2014), Wanfang Data (between January 1982 and April 2014), Medline (between January 1946 and April 2014), PubMed (between January 1966 and April 2014), Embase (between January 1974 and April 2014) and Cochrane Database of Systematic Reviews (between January 2005 and April 2014) for randomized controlled trials comparing surgical treatment and conservative treatment for displaced 3-part or 4-part proximal humeral fractures in elderly patients. The data were extracted and a Meta-analysis was made using RevMan 5.2. ResultsFive randomized controlled trials with a total of 226 patients were accepted in this Meta-analysis. The differences of Constant scores[WMD=0.48, 95%CI (-4.39, 5.35), P=0.85], incidences of avascular necrosisr of the humeral head[RR=0.66, 95%CI (0.37, 1.16), P=0.15], incidences of nonunion[RR=0.56, 95%CI (0.20, 1.58), P=0.27], incidences of osteoarthritis[RR=0.35, 95%CI (0.10, 1.22), P=0.10] between the two groups were not statistically different. The difference of incidences of additional surgery after primary treatment was statistically significant[RR=3.52, 95%CI (1.18, 10.45), P=0.02]. ConclusionThe results does not support the surgical treatment to improve the shoulder function when compared with conservative treatment for treating elderly patients with displaced 3-part or 4-part proximal humeral fractures. But surgical treatment increases the incidence of additional surgery after primary treatment. Considering surgery can increase the trauma and economic burden of patients, so the conservative treatment is suggested.