Background During primary total knee arthroplasty (TKA), synovectomy as a part of the procedure has been recommended to relieve pain and inflammation of the synovium, but there is a controversy about it due to increased bleeding. In this meta-analysis, the aim is to answer whether synovectomy should be performed routinely during TKA for symptomatic knee osteoarthritis (KOA). Methods Relevant randomized controlled trials (RCTs) on synovectomy were retrieved through database searches of PubMed, Embase, Web of Science, and Cochrane Library up to February 2019. Studies that compared postoperative pain, clinical Knee Society Score (KSS), functional KSS, range of motion (ROM), drainage, pre- and postoperative hemoglobin difference, transfusion rate, operative time, and/or complications were included in the meta-analysis. Review Manager 5.3.0 was used for meta-analysis. Results We included 5 RCTs with 542 knees. Pooled results indicated that the synovectomy group was associated with more blood loss via drainage (WMD = - 99.41, 95% CI - 153.75 to - 45.08, P = 0.0003) and pre- and postoperative hemoglobin difference (WMD = - 0.93, 95% CI - 1.33 to - 0.5, P < 0.00001), compared with the non-synovectomy group. No statistically significant differences were demonstrated between both groups in postoperative pain, clinical KSS, functional KSS, ROM, transfusion rate, or complications (P > 0.05). Conclusions The current evidence demonstrates that performing synovectomy in primary TKA for symptomatic KOA does not have any clinical benefit. It increases postsurgical blood loss. Surgeons routinely undertaking synovectomy should deliberate whether this is clinically indicated and consider limiting resection, if possible.