16 randomized controlled trials (RCTs) with a total of 1073 patients (1097 knees) were included in this meta-analysis to compare the effects of early versus late tourniquet release in total knee arthroplasty. Outcomes assessed included calculated blood loss (n=4 studies), drops in haemoglobin (n=6 studies) and haematocrit (n=4 studies), measured postoperative blood loss (n=7 studies), total measured blood loss (n=11 studies), operative time (n=12 studies), transfusion rate (n=8 studies), incidence of postoperative deep vein thrombosis (n=6 studies), and incidence of wound complications (n=8 studies). Findings indicated significantly greater total measured blood loss and transfusion rate, but significantly lower incidence of wound complications in the early release group compared to the late release group. All other outcomes were similar between groups.