Background The efficacy and safety of mesh/graft in surgery for anterior or posterior pelvic organ prolapse is uncertain.Objectives To systematically review the efficacy and safety of mesh/graft for anterior or posterior vaginal wall prolapse surgery.Search strategy Electronic databases and conference proceedings were searched, experts and manufacturers contacted, and reference lists of retrieved papers scanned.Selection criteria Randomised controlled trials (RCTs), nonrandomised comparative studies, registries, case series involving at least 50 women, and RCTs published as conference abstracts from 2005 onwards.Data collection and analysis One reviewer screened titles/ abstracts, undertook data extraction, and assessed study quality. Data analysis was conducted for three subgroups: anterior, posterior, and anterior and/or posterior repair (not reported separately).Results Forty-nine studies involving 4569 women treated with mesh/graft were included. Study quality was generally high. Median follow up was 13 months (range 1-51 months). In anterior repair, there was short-term evidence that mesh/graft (any type) significantly reduced objective prolapse recurrence rates compared with no mesh/graft (relative risk 0.48, 95% CI 0.32-0.72). Nonabsorbable synthetic mesh had a significantly lower objective prolapse recurrence rate (8.8%, 48/548) than absorbable synthetic mesh (23.1%, 63/273) and biological graft (17.9%, 186/1041), but a higher erosion rate (10.2%, 68/666) than absorbable synthetic mesh (0.7%, 1/147) and biological graft (6.0%, 35/581). There was insufficient information to compare any of the other outcomes regardless of prolapse type.Conclusions Evidence for most outcomes was too sparse to provide meaningful conclusions. Rigorous long-term RCTs are required to determine the comparative efficacy of using mesh/graft.