Objective：To compare the clinical outcome between closed suction drainage and nondrainage following primary total knee arthroplasty（TKA） on the basis of the best available evidence. Methods：Randomized controlled trils（RCTs） between 1950 and De－cember 2013 comparing primary TKA patients managed with closed-suction drainage and those without drainage in the databases of Medline，EMBASE，Cochrane Library and CNKI were enrolled. The Review Mananger was used to analyze selected studies. Twenty-two CONSORT criteria were used to evaluate the the quality of the literature. Subgroup analysis was carried out when significant het－erogeneity occured. Results：Sixteen RCTs involving 1 417 knees（713 knees in closed-suction drainage group and 704 in nondrainage group） satisfied the inclusion criteria. In closed-suction drainage group，the incidence of soft tissue ecchymosis was significantly lower compared with that in nondrainage group. Subgroup analysis showed greater differences between the two groups in the setting of simul－taneous bilateral TKA. However，results favored nondrainage group with respect to the total blood loss volume and blood transfusion rates. No significant difference was demonstrated in the prevalence of infection，incidence of deep venous thrombosis（DVT），or postoperative range of motion（ROM） between two groups. Conclusion：Current RCTs demonstrate that the use of closed suction drainage after primary TKA provides no benefit in terms of infection，DVT，and postoperative ROM when compared with nondrainage. Besides，it increases perioperative total blood loss volume and elevates the risk of homologous blood transfusion. The only proven advantage of drainage is a decreased incidence of blood leakage from the wound，especially when simultaneous bilateral TKA is performed.