Abstract: BACKGROUND: Total knee arthroplasty is one of the most successful methods for treating knee osteoarthritis in the elderly. Compared with fixed-bearing prosthesis, mobile-bearing prosthesis has the theoretical advantages of reducing the wear of the prosthesis and increasing the range of motion. Most of the above studies found no significant difference between the two. 
OBJECTIVE: To evaluate the differences of clinical effects between fixed-bearing and mobile-bearing prosthesis in total knee arthroplasty through meta-analysis.  
METHODS: The databases of CNKI, Wanfang, VIP, CBM, PubMed, EMBase and Cochrane Library were systematically retrieved. The literature was strictly selected according to the inclusion criteria. We collected high-quality randomized controlled trials of total knee arthroplasty patients, who were operated with fixed-bearing or mobile-bearing prosthesis. The Cochrane Collaboration manual was used to evaluate the risk of bias and the modified Jadad score scale was utilized to evaluate the quality of the literature. The relevant data were extracted according to the primary outcome measures: Knee Society score, knee Knee Society score function score; and secondary outcome measures: range of motion, Oxford knee score and revision rate. The relevant data were analyzed with RevMan 5.3 software.  
RESULTS: Fifteen randomized controlled trials were eligible for meta-analysis. According to the modified Jadad rating scale, there were 4 articles with 6 points, 10 articles with 5 points and 1 article with 4 points. There are 1 277 samples in the fixed-bearing group and 1 244 in the mobile-bearing group. Meta-analysis results demonstrated that (1) there was no statistically significant difference in Knee Society score-knee score, range of motion, Oxford knee score, and revision rates between the two groups using fixed-bearing and mobile-bearing prostheses in the short-term follow-up (< 5 years) subgroup (P > 0.05). The Knee Society score-function score in the mobile-bearing group was superior to that in the fixed-bearing group (MD=-2.26, 95%Cl: -4.71 to -0.34, P=0.02). (2) In the subgroup of long-term follow-up (≥ 5 years), there was no significant difference between the two groups in Knee Society score-knee score, Knee Society score-function score, and revision rate (P > 0.05). Range of motion in the mobile-bearing group was better than that in the fixed-bearing group (MD=-3.60, 95%Cl: -6.99 to -0.21, P=0.04).
CONCLUSION: in total knee arthroplasty, whether in short-term follow-up or mid-to-long-term follow-up, mobile-bearing prosthesis is not much different from fixed-bearing prosthesis in most respects. However, there is weak evidence that the improvement of knee joint function after knee arthroplasty with mobile-bearing prosthesis is slightly better than that with fixed-bearing prosthesis. Most of the articles were followed up in short and medium terms. More high-quality, large-sample and multi-center randomized controlled trials are needed to verify the reliability of the conclusions.

Key words: bone, knee, knee joint, joint arthroplasty, mobile-bearing, fixed-bearing, platform, meta-analysis, randomized controlled trial