Surgical Interventions For Unicompartmental Knee Osteoarthritis:A Network Meta-analysis

ObjectiveNetwork meta-analysis(NMA)is a relatively new evidence-based medical method in clinical work.It simultaneously compares multiple interventions and ranks sequence of outcomes to provide all-round evidence supporting for the selection of clinical treatment.In this study,we systematically evaluated the efficacy of total knee arthroplasty(TKA),unicompartmental knee arthroplasty(UKA)and high tibial osteotomy(HTO)in the treatment of unicompartmental knee osteoarthritis using the Bayesian network meta-analysis method.MethodsWe electronically searched PubMed,Embase,Cochrane Library from inception to August 31,2018 to collect randomized controlled trials(RCTs)and cohort studies between TKA,UKA and HTO,choosing patients with regard to unicompartmental knee osteoarthritis.Studies comparing two or three indicators were included.Two reviewers independently screened literatures according to the inclusion and exclusion criteria,extracting data and assessing the risk of bias of included studies.Using the Cochrane Collaboration Risk Assessment Tool and the Newcastle-Ottawa Scale to assess methodological quality and risk of bias.Finally,network meta-analysis was performed using R software.ResultsA total of 5 RCTs and 16 cohort studies involving 966 relevant literatures(2104 knees)were included.5 trials were included in the analysis of operation time.Network meta-analyses indicated that there was no difference between TKA,UKA and HTO.Pairwise meta-analyses showed that compared with UKA,HTO(MD=-26.00,95%CI=-31.97~-20.03,P<0.00001)significantly decreased the operation time,TKA(MD=22.11,95%CI=19.03~25.20,P<0.00001)significantly increased the operation time.Gait analysis was included in three trials.Network meta-analyses indicated that there was no difference between TKA,UKA and HTO.Pairwise meta-analysesshowed that compared with UKA,TKA(MD=-0.18,95%CI=-0.33~-0.03,P=0.02)significantly decreased the natural walking velocity.Eight trials included ROM(Range of motion).Network meta-analyses indicated that compared with knee arthroplasty(UKA and TKA),HTO significantly increased the ROM.Pairwise meta-analyses showed that compared with UKA,HTO(MD=10.4,95%CI=7.97~12.82,P<0.00001)significantly increased ROM.Comparing with UKA,the postoperative ROM of TKA(MD=-3.89,95%CI=-5.60~-2.17,P < 0.00001)was significantly lower.4 trials included KSS(Knee society score).Network meta-analyses indicated that there was no difference between TKA,UKA and HTO.Pairwise meta-analyses showed that compared with UKA,TKA(MD=-2.96,95%CI=-4.57~-1.35,P=0.0003)significantly decreased KSS.In terms of postoperative revision rates and complications,3 interventions above showed no significant difference.Pairwise meta-analyses showed that compared with UKA,HTO(MD=3.24,95%CI=1.90~5.54,P<0.001)significantly increased revision rates.Compared with UKA,HTO(MD=2.84,95%CI=1.72~4.68,P<0.0001)and TKA(MD=2.70,95%CI=1.72~4.25,P<0.0001)significantly increased postoperative complications.Conclusions1.Concerning gait,HTO was not significantly different from UKA,TKA was not significantly different from HTO.2.As for improving ROM,HTO was superior to UKA.3.In terms of KSS,HTO was not significantly different from UKA.4.In terms of revision rates,TKA was not significantly different from UKA and HTO.5.In terms of complications,TKA was not significantly different from HTO.