Contents lists available at SciVerse ScienceDirect
The Knee
Review
Little clinical advantage of computer-assisted navigation over conventionalinstrumentation in primary total knee arthroplasty at early follow-up
Tao Cheng a, Xiao-Yun Pan a,b, Xin Mao a, Guo-You Zhang c,d, Xian-Long Zhang a,⁎a Department of Orthopaedic Surgery, Shanghai Sixth People's Hospital, Shanghai Jiao Tong University School of Medicine, 600 Yisan Road, Shanghai 200233, People's Republic of Chinab Department of Orthopaedic Surgery, The Second Affiliated Hospital of Wenzhou Medical College, 109 Xueyuan West Road, Wenzhou 325027, Zhejiang Province, People's Republic of Chinac Department of Hand and Plastic Surgery, The Second Affiliated Hospital of Wenzhou Medical College, 109 Xueyuan West Road, Wenzhou 325027, Zhejiang Province, People's Republic of Chinad Department of Dermatology, University Hospital Schleswig-Holstein, University of Lübeck, 160 Ratzeburger Allee, Lübeck 23538, Germany
Received 17 March 2011Received in revised form 25 September 2011Accepted 16 October 2011
Keywords:Joint arthroplastyTotal knee replacementComputer-assisted surgeryNavigationMeta-analysisEvidence-based medicine
Purpose: Even though computer-assisted navigation systems have been shown to improve the accuracy ofimplantation of components into the femur and tibia, long-term results are lacking and there is little evidenceyet that navigation techniques also improve functional outcomes and implant longevity following total kneearthroplasty (TKA). The aim of this study was to summarize and compare the clinical outcomes of total kneearthroplasties (TKAs) performed using navigation-assisted and conventional techniques.Methods: The study was conducted according to the guidelines described in the Cochrane Handbook for System-atic Reviews of Interventions and Preferred Reporting Items for Systematic Reviews and Meta-Analyses State-ments. Methodological features were rated independently by two reviewers. A meta-analysis of randomizedcontrolled trials (RCTs) or quasi- randomized controlled trials (qRCTs) was carried out to evaluate the efficacyof CAS versus conventional TKA. Data were pooled in fixed and random effects models and the weighted meandifference (WMD) and odds ratio (OR) were calculated. Heterogeneity across studies was determined, and sub-
group analyses by the type of navigation system (image-based or image-free navigation system)were conducted.Results: Twenty-one studies that included 2333 knees were collected from different countries. The surgical timewas longer for CN TKA than for the conventional procedure. There was no significant difference in the Knee So-ciety Score between the two groups at the 3-month and 6-month follow-up. The rates of postoperative compli-cations in patients who had CN TKA were similar to those in the patients who had conventional TKA.Conclusion: No significant differences in short-term clinical outcomes were found following TKAs performedwith and without computer navigation system. However, there is clearly a need for additional high-quality clin-ical trials with long-term follow-up to confirm the clinical benefits of computer-assisted surgery.
Total knee arthroplasty (TKA) is a common orthopedic procedure,generally performed in elderly arthritis patients [1–3]. TKAs are con-ventionally performed with the use of intramedullary or extramedul-lary jig-based alignments and achieve a high rate of success [4–6].However, recent studies have shown that achieving optimal implantalignment can be difficult with current techniques, even for experi-enced surgeons [7–10]. Based on the theoretical assumption that theuse of computer-assisted navigation in TKAs may improve implantalignment and increase implant longevity [11], computer-assisted sur-gery is becoming increasingly common in TKAs [12–24].
Significant improvements in component orientation and mechani-cal leg axis have been reported when using image-based and image-free computer navigation systems, particularly in TKAs with knee de-formities [10,25–41]. However, the better alignment achieved incomputer-navigated (CN) TKAs did not necessarily result in betterclinical outcomes [18,32,42–45]. Exact informations on the clinical re-sults are needed to determine potential advantages of CN TKAs. Thisinformation would be even more valuable when obtained by meta-analysis and systematic review, which combines or integrates the re-sults of several independent clinical trials to increase statistical power.
The primary purpose of this study was to determine if CN TKAmakes possible better clinical outcomes for operative patients as eval-uated Knee Society Score (KSS) and complication rates than the con-ventional technique. The secondary purpose was to ascertain if a CNTKA results in increased operative time and reduced blood loss com-pared with a conventional TKA.
2. Materials and methods
A systematic review was conducted according to the guidelinesdescribed in the Cochrane Handbook for Systematic Reviews of Inter-ventions and PRISMA (Preferred Reporting Items for Systematic Re-views and Meta-Analyses) Statements [46,47]
2.1. Eligibility criteria
Trials included in this paper were those relevant to: (1) those pa-tients undergoing primary TKA with a conventional technique versusa passive navigation technique, (2) Knee scores and post-operativecomplications as its outcome measures, and (3) the study was a pub-lished randomized or quasi-randomized controlled trial (RCT orqRCT). For trials with multiple publications we included only themost complete report for each outcome.
2.2. Literature search
To identify published reports of relevant RCTs or qRCTs we carriedout highly sensitive electronic searches of relevant databases, includingMedline (1995–November 2009), EMBASE (1995–November 2009),Science Citation Index (1995–November 2009), Chinese Biomedical Lit-erature Database (2000–November 2009), Wanfang database (2000–November 2009), and Cochrane Database of Systematic Reviews andthe Central Register of Controlled Trials (2000–November 2009). Thefollowing key words were considered: “computer-assisted”, “naviga-tion”, “navigated”, “total knee arthroplasty”, “total knee replacement”.In addition, we also handsearched five major orthopedic journals, in-cluding The Journal of Bone and Joint Surgery: American and British
Volumes, Clinical Orthopaedics and Related Research, The Journal ofArthroplasty, and TheKnee. Searcheswere not restricted by year of pub-lication or language. The last search was carried out on January 1, 2010.
2.3. Data extraction
Two reviewers independently recorded participants' characteristics,details of the surgical technique and implant used, and clinical out-comes on a data extraction form. The primary outcomes of interest in-cluded postoperative functional outcomes and complications.Secondary outcomes were intraoperative parameters (blood loss, oper-ative time). If necessary, the primary authorswere contacted to retrievefurther information.
2.4. Assessment of methodological quality
Two of us independently assessed the methodological quality ofeach included study with respect to rating of the randomization pro-cedure; allocation concealment; blinding of patients, clinicians, out-come assessors; statistical analysis of individual trials, and numbersof patients lost during follow-up. Any differences that could not be re-solved through discussion were decided by an arbiter.
3. Statistical analysis
For meta-analysis we combined dichotomous outcome data usingthe Mantel–Haenszel odds ratio (OR) method and 95% confidence in-terval (CI). For continuous outcomes we used inverse varianceweighted mean difference (WMD) and 95% CI. Publication bias wasassessed using a funnel plot of the outcome measure recorded inthe largest number of clinical trials. Before analyzing the data, wehypothesized that possible clinical heterogeneity may be due todifferences in interventions (type of navigation system used). Weused a random effects model if heterogeneity existed (Pb0.10) anda fixed effect model otherwise. We used SPSS version 13.0 for Win-dows (SPSS Inc., Chicago, Illinois, USA) and RevMan 5.0 softwarepackage (Cochrane Collaboration, Oxford, United Kingdom) for finalanalyses.
4. Results
4.1. Study identification and study characteristics
Of the 935 articles identified in the literature search, fifty seven were considered po-tentially eligible on the basis of the abstract. Of these, only twenty one were RCTs or qRCTand eligible for inclusion [26,27,29–31,48–63] (Fig. 1). Eleven studies were from Europe[26,27,48,49,51,52,56–59], six were fromAsia [30,50,53,54,60,61], three were fromWest-ern Australia [29,31,63] and one was from North America [55]. In most of these includedstudies, the demographic features of both groups were well balanced at baseline. The CNgroup consisted of 1225 knees (52.5%), whereas the conventional group comprised 1108knees (47.5%). The sample sizes of the trials ranged from 26 to 467 knees. Furthermore,most studies had clear included or excluded criteria. Seventeen studies used image-less navigation system [26,27,29–31,48–52,54,55,57,60–63], whereas the other fivestudies used the image-based navigation system [53,56,58,59,62]. Most studiesindicated that the surgeons had experience in CN TKA prior to their study, in orderto avoid bias from the learning curve. The choice of implants and fixation techniquesvaried across studies, when reported. Table 1 presents the characteristics of theincluded studies.
4.2. Study quality
The methodological quality of the included studies was variable. The method of al-location sequence generation varied form study to study and included use of com-puters, random number tables, block randomizations, alternation, permutation
Fig. 1. PRISMA chart of the study selection process.
239T. Cheng et al. / The Knee 19 (2012) 237–245
algorithms, day of surgery, the time of hospital admission, even/odd case record num-ber, and availability of equipment. The method of randomization was unclear in theremaining three studies [26,30,63]. Randomization appeared to have been concealedin three studies, in which an envelope had been used [30,49,55]. Triple-blinding israrely used in orthopedic surgery trials, especially blinding of the surgeon. It is notclear whether patients were blinded to the procedure in the majority of the studiesand they were blinded in three studies [29,31,48]. The outcome assessors were de-scribed as blinded in most studies. Follow-up rates were high, ranging between87.3% and 100%. Table 2 presents the methodologies of the included studies.
4.3. Publication bias
The assessment of publication bias using complications indicated there isminimal ev-idence of publication bias through some mild asymmetrical funnel plot (Fig. 2). This is ascatter plot of the treatment effects estimated from individual studies, plotted on the hor-izontal axis, against the standard error of the estimate, shown on the vertical axis. Moststudies fall within the 95%CI axis for a given standard error. There may be a few studiesmissing from the lower left-hand corner of the plot.
4.4. Intraoperative outcome
The mean operating tourniquet times were reported in fourteen trials[26,29,31,48–52,55–57,59,60,62]. Meta-analysis of these trials did find that the durationof the CN operations was significantly longer (WMD=22.42 min, 95%CI=14.15 to
30.69, Pb0.00001) than conventional surgery in these studies and therewas a statisticallysignificant difference between the tourniquet time in the two types of techniques in favorof conventional TKA (WMD=15.72 min, 95%CI=13.22 to 18.22, Pb0.00001). Using arandom effects model, pooling data showed that tourniquet time in conventional TKAwas shorter than that in image-free navigated TKA (WMD=14.46 min, 95%CI=9.40 to19.51, Pb0.00001) or image-based navigated TKA (WMD=12.81 min, 95%CI=7.27 to18.35, Pb0.00001). In our study, the quantified synthesis of blood loss during surgery isdifficult because intraoperative and postoperative bleeding compounds the total bloodloss. Two studies reported that CN TKA was associated with less blood loss during opera-tion [52,60]. However, there is no difference between the two groups in two studies[50,59]. It was not clear whether there is less blood loss after CN TKA than conventionalTKA, but CN TKA is considered at least as safe as conventional TKA in view of the volumeof blood loss.
4.5. Knee score assessment
The KSS [27,49,50,52,53,63], Western Ontario and McMaster Universities Osteoar-thritis Index (WOMAC) [49],The Hospital for Special Surgery Score (HSS) [50], patello-femoral score [61], International Knee Society score [31,59], and Oxford Knee Score(OKS) [61] were used for the clinical assessment of patients. Unfortunately, summaryscore data necessary for quantitative analyses were incompletely reported. Moreover,the time points at which the outcomes were measured differed among these studies.Therefore, pooling of data was not possible except for the outcome of KSS. The availabledata from six studies were not heterogeneous and, overall, they showed no significant
Table 1Characteristics of included studies.
Author Year Country No. of patients/knee Navigation system Implant Duration of Follow-up
differences between the compared arms at the 3-month (WMD=1.11, 95% CI,−6.33 to8.56) and 6-month (WMD=2.13, 95% CI, −2.53 to 6.79) follow-up (Fig. 3). Addition-ally, most studies found no significant differences in other kinds of knee score mea-surements at the latest follow-up.
4.6. Postoperative complication
Eighteen of these studies, on a total of 2151 knees, provided information on post-operative complications. For overall complications, there was no significant difference(OR=1.00; 95%CI=0.67 to 1.50, P=1.00) between the two groups without significantheterogeneity (I2=22%). Altogether, 53 complications occurred in 1123 patients trea-ted with CN technique, and 52 complications occurred in 1028 patients treatedwith conventional technique (Fig. 4). There was no evidence of differences in deep ve-nous thrombosis, infection, delayed wound healing, knee stiffness, periprosthetic frac-ture, hematoma, pulmonary embolism, or nerve injury between the two groups (Table
Table 2Quality of included studies.
Author Year Design Randomization method Allocation concealment
Decking 2005 RCT Computer program Sealed envelopeKim 2007 RCT Random number table and
alterationNS
Lützner 2008 RCT Permutation algorithmmedical biometrician withoutstratification using software
Matziolis 2007 RCT Random-number Generator NSMizu-uchi 2008 qRCT Alternating NoMullaji 2006 qRCT The time of admission NSO'Connor 2010 RCT Block randomization Sealed envelopePerlick 2004a qRCT Day of surgery NoPerlick 2004b qRCT Day of surgery NoSparmann 2003 qRCT Availability of equipment NoSpencer 2007 RCT NS NSVictor 2004 RCT Block randomization NSWeng 2009 qRCT Even and odd hospital number NoZhang 2008 RCT Random-number NSZumstein 2006 qRCT Availability of equipment NS
NS = not specified; RCT = randomized controlled trial; qRCT = quasi-randomized control
3). There were no significant differences in treatment effect in association with thetype of navigation system (OR=1.21, 95%CI=0.52 to 2.81, P=0.65 for image-basenavigation system studies; compared with 0.97, 95%CI=0.61 to 1.53, P=0.89for image-free navigation system studies). Conversion from CAS to the conventionaltechnique was reported in two trials [27,62]. These causes of conversion includedhardware failure and tracker fixation-device loosening.
5. Discussion
Despite excellent radiographic outcome there is a lack of meta-analysis and systematic review to analyze the clinical outcomesafter CN TKA. Furthermore, whether it is advisable to implementthis procedure in daily surgical routine should be decided based on
Blinding Baselinecomparability
Similarityof careprogram
Sample-sizecalculation
Loss tofollow-up
Patient Surgeon Assessor
NS No No Yes Yes NS NSNS No No Yes Yes Yes 4.8%Yes No Yes Yes Yes NS 0%NS No Yes Yes Yes NS 0%Yes No Yes Yes Yes Yes 2.6%
Yes No Yes Yes Yes NS 0%
No No Yes Yes Yes NS 0%NS No Yes Yes Yes Yes 0%
NS No NS Yes Yes Yes 0%
NS No NS Yes Yes NS 0%NS No Yes Yes Yes NS 12.7%NS No Yes Yes Yes NS 0%NS No Yes Yes Yes Yes 0%NS No Yes Yes Yes NS NSNS No Yes Yes Yes NS 0%NS No NS Yes Yes NS 0%NS No Yes Yes Yes Yes 4.2%NS No Yes Yes Yes Yes 0%NS No Yes Yes Yes NS 0%NS No Yes Yes Yes NS 0%NS No NS Yes Yes NS 1.1/%
led trial.
Fig. 2. Funnel plot of incidence of postoperative complications to assess publication bias.
241T. Cheng et al. / The Knee 19 (2012) 237–245
clinical results. The results of this meta-analysis and systematicreview show that while there is a significant lengthening in operativetime using navigation technique as compared to traditional tech-nique, functional outcome scores (KSS) and complication rate are nodifferent. According to reported data on short-term clinical outcomes,passive navigation systems do not offer significant clinical benefitsover conventional instrumentation in TKA.
In a recent prospective comparative study, 1000 patients were op-erated on either using a CT-free navigation system or the convention-al jig-based technique [64]. Although this particular study included avery large patient sample, it suffered from the limitation that therewas likely an effect of probable, and important, selection bias. Ideally,clinical information is obtained through large, carefully controlled,and randomized prospective trials. Such studies are, however, techni-cally and logistically complex, expensive, and often impractical orimpossible. Meta-analysis of RCTs therefore represents the goldstandard in validation of interventions, overcoming the reducedstatistical power of small sample size. Two well-conducted meta-analyses focusing on the radiographic outcome provided evidence-based conclusion that significant improvements in component orien-tation and mechanical leg axis when computer-assisted system isused in TKAs [7,33]. Implant alignment is just one part of the surgicalprocedure, and the assessment of a TKA outcome should be basedespecially on the functional outcomes and postoperative complica-tions. But such factors have not been studied in the previous meta-analysis. Recently, there have been a number of RCTs comparing CN
Fig. 3. Comparison of computer-assisted and conventional total knee arthroplasties in termsstandard deviation; CI = confidence interval; df = degrees of freedom.
TKA versus conventional TKA [26,27,31,48,50,51,53–55,60,61,63].Hence, we set out to conduct this updated meta-analysis, aiming tocombine data from RCTs or qRCTs in an effort to compare the clinicaloutcomes after computer-assisted versus conventional implantationin TKAs. Eleven RCTs and ten qRCTs formed the basis of the largermeta-analysis and systematic review.
Our previous study identified improvements in component align-ment with the use of computer-assisted surgery, as noted in previousclinical studies [7,10,33,65]. The relationship between radiologicalassessment and clinical outcomes has been previously described[31,66–69]. Recently, however the value of these findings has beenquestioned. Parratte et al. [70] found that postoperative mechanicalaxis of 0°±3°did not improve the long-term implant survivorship fol-lowing modern TKAs. Various factors, such as characteristics of thepatients, design of implant, operative technique, rotationalalignment, and dynamic gait factors, may influence knee functionand implant survivorship more than static limb alignment does[23,31,36,60,70–75]. In accordance with our results, several studiesdemonstrated that there were no differences for subjective andobjective criteria (WOMAC, OKS, KSS, HSS, degree of flexion, implantsurvival) between TKAs undertaken with computer navigation assis-tance and standard instrumentation at mid-term follow-up [76–81].In a 2011 study [82] from the Norwegian Arthroplasty Registerreported a higher risk of revision in computer-assisted surgery forthe LCS complete implant compared to conventional technique,with short-term (2 years) follow-up. The reason for this may be dueto learning curve and technical failures related to the computer nav-igation systems. The success of computer navigation as a surgical in-strument may be dependent on the design of the implant. Greateruse of National Registry for joint arthroplasty should be encouragedto provide navigation-specific data on clinical outcomes.
In our study, prolonged operative time is found when using navi-gation systems in TKA. However, the CN procedure is only increasedby 12 to 22 min. This additional time seems to be tolerable in clinicalroutine. In the future, it might be even further reduced by the accu-mulation of surgeon's experience and the development of specificnavigation-adapted instruments. The included studies had individu-ally evaluated blood loss by drainage, hematocrit values, hemoglobinloss, and transfusion requirements. Due to inconsistencies in the waythat blood loss was reported, the conclusions that can be drawn arelimited. Overall, computer navigation allows minor release of softtissues without violating the intramedullary cavities and may providean opportunity for reducing blood loss [29,30,58] and limiting boneinjury [83]. In a prospective randomized trial, Conteduca et al. [84]
of Knee Society Score. CAS = computer-assisted surgery; IV = inverse variance; SD =
Fig. 4. Comparison of computer-assisted and conventional total knee arthroplasties in terms of overall complications. CAS = computer-assisted surgery; M–H=Mantel–Haenszel;CI = confidence interval; df = degrees of freedom.
242 T. Cheng et al. / The Knee 19 (2012) 237–245
determined the blood loss according to the Orthopaedic SurgeryTransfusion Haemoglobin European Overview study and found thatcomputer-assisted surgery is highly recommended in TKA to reduceblood loss. The method calculated the blood loss by a formula thattakes into account several parameters such as sex, weight, height,pre-and postoperative hematocrit, homologous or autologous trans-fusions, thus avoiding the difficulties of calculating the actualbleeding during the intraoperative and postoperative period. Theintraoperative and postoperative risks can be increased by bloodloss during and after surgery. However, it is worth noting that evenif blood loss is reduced by the computer-assisted procedure, prolon-gation of the operating time, as currently occurs in computer-assisted procedure, may negate any this benefit.
Considering the low risks of adverse events related to thisoperation, we calculated the total complication rate and various com-plications in our meta-analysis, and found no significant differencesbetween these two groups. The use of CN TKA, which does notinvolve the use of intramedullary alignment rods, would produceless fat and bone-marrow-debris embolization. Transcranial Dopplerultrasonography and transesophageal echocardiography performedin three RCTs demonstrated that computer-assisted procedure signif-icantly reduced emboli in the systemic circulation, compared with theconventional technique [48,85,86]. In contrast to these data, Kim et al.[87] reported no significant difference in the prevalence of fat and/orbone-marrow-cell embolization between these groups treated bycomputer-assisted navigation or by conventional implantation
CAS= computer-assisted surgery; OR = odd ratio; WMD = weight mean deviation; CI =
techniques. The use of computer-assisted navigation did not seemto have an effect on postoperative mental status and cognitive dys-function following TKAs [88,89]. We believe that the disagreementbetween these trials might be attributable to the surgical techniqueand examination method. It is noted that even if a statistically signif-icant decrease in embolic phenomena is observed when usingcomputer-assisted TKA, this does not equate to a clinically significantdecrease [55]. Periprosthetic fracture related to the navigation systemhas only rarely been reported in the orthopedic literature [54,90–94].The disadvantage of trans-cortical drilling of the tibia and femur to fixthe trackers to the bone is of concern. Repeated drilling may be thecause of stress fracture and infection in patients following CN TKA[92,93,95–97]. The prolongation of the operative time in the CNgroup may increase the probable risk of deep infections and the prev-alence of postoperative cognitive dysfunction [88,98]. However, thesecomplications have not been shown to occur to any significant level inour meta-analysis. This may be partly explained by the fact that thegenerally low complication rates in itself makes it difficult to identifyclinically important differences between the two groups. The mainresults of our study are in accordance to the results of other studies[99,100]. Browne et al. [99] reported the use of computer-assistedTKA in a Nationwide Inpatient Sample database. This study demon-strated similar early postoperative morbidity and mortality in pa-tients undergoing TKA with or without computer navigation.Graham et al. [100] performed a retrospective review of 327 patientsundergoing TKA. They found that navigation offered no benefit in
terms of early postoperative morbidity when compared with conven-tional jig-based technique.
The aspect of costs of both procedures could not be addressed inour study because all included trials did not report on this subject atall. Slover et al. [101] argued that computer navigation is a less cost-effective investment in health-care improvement for arthroplastycenters that perform a low volume of TKA, where the incidence ofcomplications may be higher and improvements in outcomes aremost needed [102,103]. Moreover, Yau et al. [104] found that com-puter navigation system did not improve accuracy in postoperativeradiographic alignment of implants in a low-volume hospital. Thecorrelation between cost-effectiveness and clinical benefits shouldbe integral part of every surgeon's decision to opt for using an avail-able navigation system in his practice, and this requires furtherlong-term studies for a definitive answer to this question [105].
Although our aim in the present paper was to perform a meta-analysis of high-quality studies, certain limitations must be acknowl-edged. First, the relatively poor-quality trials were included in thisstudy. It is disappointing that randomization methods were notused appropriately and not described adequately in some studies.Overly broad inclusion criteria (e.g. allowing qRCTs) in the presentstudy may provide greater statistical power but are subject to criti-cism regarding potential bias in patient selection, management andreporting. Second, the statistical heterogeneity of study parameterssuch as operative time and tourniquet time may have occurred in anumber of factors including the measurement method, the definitionof end points, the surgical techniques, and the demographic charac-teristics. When there appeared to be heterogeneity that could notbe explained, we used a random effects model to incorporate thestudy results. In addition, subgroup analyses by the type of navigationsystemwere done to investigate heterogeneous results and to answerspecific clinical questions. Third, there is an obvious lack of long-termprospective RCTs. Although 2333 knees were available, most of thetrials focused on short-term outcomes. Intermediate-term and long-term outcomes have not been evaluated in this systematic review asthe durations of follow-up have been insufficient to allow a proper as-sessment of implant survival and long-term complications. However,our data demonstrated that computer-assisted navigation doses notimprove early clinical outcomes. Another limitation is that we didnot include unpublished studies, which may incur risks of publicationbias, whereby studies with statistically significant positive resultstend to be published over negative studies. It is important to pointout that unpublished studies have not passed peer-review. Therefore,inclusion of these studies would risk incorporation of results that maynot be credible. There are several strengths of this meta-analysis.First, we made no restrictions as to language in order to avoid lan-guage bias. In our meta-analysis, the included trials from differentcountries were widely collected English and non-English medicaljournals. Second, we systematically reviewed high quality RCTs orqRCTs (Level I/II) comparing CN and conventional TKAs. Third, therewas no statistical evidence of heterogeneity for our primary outcomes(KSS, complication rates). Although we found no statistical evidenceof heterogeneity between studies, we performed sensitivity analysisbased on a prior hypothesis (the type of navigation systems) for ourprimary outcomes.
In summary our analysis represents the best available evidence onCN TKA, which seems to suggest that despite longer operative time,the post-operative complication rates and KSS in patients with CNTKA appears to be no difference when compared to conventionalTKA in the short-term follow-up. Future studies investigating the re-sults such as implant survival and cost-effectiveness of CN TKA shoulduse better experimental design, including adequate allocation con-cealment, objective and independent assessment of outcomes, largerpatient samples, and longer follow-up, incorporating data from pa-tients matched not only for age and sex but also for obesity and sever-ity of joint disease.
Conflict of interest statement
The authors have declared no conflict of interest. No companyhas had any role in the study design; the data collection, analysis,and interpretation; the manuscript preparation; or the manuscriptsubmission.
Acknowledgments
The authors thank the Library staff at the Shanghai Jiao Tong Univer-sity for their assistance in gathering the articles used in this paper. Thiswork was supported by the Shanghai Municipal Health Bureau ScienceFund for Young Scholars (2010QJ036A), the Scientific Research Fund ofZhejiang Provincial Education Department (Y200804988), the ZhejiangProvincial Natural Science Foundation of China (Y2090540), and theMajor Program of Zhejiang Extremely Key Subject for Fostering Talents.
References
[1] Gandhi R, Dhotar H, Razak F, Tso P, Davey JR, Mahomed NN. Predicting the longerterm outcomes of total knee arthroplasty. Knee 2010;17:15–8.
[2] Foote JA, Smith HK, Jonas SC, Greenwood R, Weale AE. Return to work followingknee arthroplasty. Knee 2010;17:19–22.
[3] Julin J, Jämsen E, Puolakka T, Konttinen YT, Moilanen T. Younger age increasesthe risk of early prosthesis failure following primary total knee replacement forosteoarthritis. A follow-up study of 32,019 total knee replacements in the Finn-ish Arthroplasty Register. Acta Orthop 2010;81:413–9.
[4] Robertsson O, Bizjajeva S, Fenstad AM, Furnes O, Lidgren L, Mehnert F, et al. Kneearthroplasty in Denmark, Norway and Sweden. Acta Orthop 2010;81:82–9.
[5] Lygre SH, Espehaug B, Havelin LI, Vollset SE, Furnes O. Failure of total kneearthroplasty with or without patella resurfacing. Acta Orthop 2011;82:282–92.
[6] Gillam MH, Ryan P, Graves SE, Miller LN, de Steiger RN, Salter A. Competing riskssurvival analysis applied to data from the Australian Orthopaedic Association Na-tional Joint Replacement Registry. Acta Orthop 2010;81:548–55.
[7] Brin YS, Nikolaou VS, Joseph L, Zukor DJ, Antoniou J. Imageless computer assistedversus conventional total knee replacement. A Bayesian meta-analysis of 23comparative studies. Int Orthop 2011;35:331–9.
[8] Mahaluxmivala J, Bankes MJ, Nicolai P, Aldam CH, Allen PW. The effect of surgeonexperience on component positioning in 673 press fit condylar posteriorcruciate-sacrificing total knee arthroplasties. J Arthroplasty 2001;16:635–40.
[9] Lustig S, Fleury C, Goy D, Neyret P, Donell ST. The accuracy of acquisition of animageless computer-assisted system and its implication for knee arthroplasty.Knee 2011;18:15–20.
[10] Cheng T, Zhao S, Peng X, Zhang X. Does computer-assisted surgery improve post-operative leg alignment and implant positioning following total knee arthro-plasty? A meta-analysis of randomized controlled trials? Knee Surg SportsTraumatol Arthrosc Jul. 6 2011 [Epub ahead of print].
[11] Ritter MA, Faris PM, Keating EM, Meding JB. Postoperative alignment of totalknee replacement. Its effect on survival. Clin Orthop Relat Res 1994;299:153–6.
[12] Lehnen K, Giesinger K, Warschkow R, Porter M, Koch E, Kuster MS. Clinical out-come using a ligament referencing technique in CAS versus conventional tech-nique. Knee Surg Sports Traumatol Arthrosc 2011;19:887–92.
[13] Lee DH, Padhy D, Lee SH, Nha KW, Park JH, Han SB. Osteoporosis affects compo-nent positioning in computer navigation-assisted total knee arthroplasty. KneeApr. 27 2011 [Epub ahead of print].
[15] Hasegawa M, Yoshida K, Wakabayashi H, Sudo A. Minimally invasive total kneearthroplasty: comparison of jig-based technique versus computer navigationfor clinical and alignment outcome. Knee Surg Sports Traumatol Arthrosc2011;19:904–10.
[16] Bejek Z, Paróczai R, Szendröi M, Kiss RM. Gait analysis following TKA: compari-son of conventional technique, computer-assisted navigation and minimally in-vasive technique combined with computer-assisted navigation. Knee SurgSports Traumatol Arthrosc 2011;19:285–91.
[17] Millar NL, Deakin AH, Millar LL, Kinnimonth AW, Picard F. Blood loss followingtotal knee replacement in the morbidly obese: effects of computer navigation.Knee 2011;18:108–12.
[18] Lützner J, Günther KP, Kirschner S. Functional outcome after computer-assistedversus conventional total knee arthroplasty: a randomized controlled study.Knee Surg Sports Traumatol Arthrosc 2010;18:1339–44.
[19] Walde TA, Bussert J, Sehmisch S, Balcarek P, Stürmer KM, Walde HJ, et al. Opti-mized functional femoral rotation in navigated total knee arthroplasty consider-ing ligament tension. Knee 2010;17:381–6.
[20] Lee DH, Choi J, Nha KW, Kim HJ, Han SB. No difference in early functional out-comes for mini-midvastus and limited medial parapatellar approaches innavigation-assisted total knee arthroplasty: a prospective randomized clinicaltrial. Knee Surg Sports Traumatol Arthrosc 2011;19:66–73.
244 T. Cheng et al. / The Knee 19 (2012) 237–245
[21] Shetty GM,Mullaji A, Bhayde S, Chandra Vadapalli R, Desai D. Simultaneous bilateralversus unilateral computer-assisted total knee arthroplasty: a prospective compari-son of early postoperative pain and functional recovery. Knee 2010;17:191–5.
[22] Tigani D, Sabbioni G, Ben Ayad R, Filanti M, Rani N, Del Piccolo N. Comparison be-tween two computer-assisted total knee arthroplasty: gap-balancing versusmeasured resection technique. Knee Surg Sports Traumatol Arthrosc 2010;18:1304–10.
[23] Pang HN, Yeo SJ, Chong HC, Chin PL, Ong J, Lo NN. Computer-assisted gap balancingtechnique improves outcome in total knee arthroplasty, compared with conven-tional measured resection technique. Knee Surg Sports Traumatol Arthrosc2011;19:1496–503.
[24] Ensini A, Catani F, Biasca N, Belvedere C, Giannini S, Leardini A. Joint line is wellrestored when navigation surgery is performed for total knee arthroplasty. KneeSurg Sports Traumatol Arthrosc May 28 2011 [Epub ahead of print].
[25] Bäthis H, Perlick L, Tingart M, Lüring C, Zurakowski D, Grifka J. Alignment in totalknee arthroplasty. A comparison of computer-assisted surgery with the conven-tional technique. J Bone Joint Surg Br 2004;86:682–7.
[26] Bejek Z, Sólyom L, Szendrõi M. Experiences with computer navigated total kneearthroplasty. Int Orthop 2007;31:617–22.
[27] Bertsch C, Holz U, Konrad G, Vakili A, Oberst M. Early clinical outcome after nav-igated total knee arthroplasty. Comparison with conventional implantation inTKA: a controlled and prospective analysis. Orthopade 2007;36:739–45.
[28] Böhling U, Schamberger H, Grittner U, Scholz J. Computerised and technical nav-igation in total knee-arthroplasty. J Orthopaed Traumatol 2005;6:69–75.
[29] Chauhan SK, Scott RG, Breidahl W, Beaver RJ. Computer-assisted knee arthro-plasty versus a conventional jig-based technique. A randomised, prospectivetrial. J Bone Joint Surg Br 2004;86:372–7.
[30] Chin PL, Yang KY, Yeo SJ, Lo NN. Randomized control trial comparing radiograph-ic total knee arthroplasty implant placement using computer navigation versusconventional technique. J Arthroplasty 2005;20:618–26.
[31] Choong PF, Dowsey MM, Stoney JD. Does accurate anatomical alignment resultin better function and quality of life? A prospective randomized controlledtrial comparing conventional and computer-assisted total knee arthroplasty.J Arthroplasty 2009;24:560–9.
[32] Ensini A, Catani F, Leardini A, Romagnoli M, Giannini S. Alignments and clinicalresults in conventional and navigated total knee arthroplasty. Clin Orthop RelatRes 2007;457:156–62.
[33] Mason JB, Fehring TK, Estok R, Banel D, Fahrbach K. Meta-analysis of alignmentoutcomes in computer-assisted total knee arthroplasty surgery. J Arthroplasty2007;22:1097–106.
[34] Klein GR, Austin MS, Smith EB, Hozack WJ. Total knee arthroplasty usingcomputer-assisted navigation in patients with deformities of the femur andtibia. J Arthroplasty 2006;21:284–8.
[35] Bottros J, Klika AK, Lee HH, Polousky J, Barsoum WK. The use of navigation intotal knee arthroplasty for patients with extraarticular deformity. J Arthroplasty2008;23:74–8.
[36] Hernández-Vaquero D, Suarez-VazquezA, Sandoval-GarciaMA, Noriega-FernandezA. Computer assistance increases precision of component placement in total kneearthroplasty with articular deformity. Clin Orthop Relat Res 2010;468:1237–41.
[37] Catani F, Digennaro V, Ensini A, Leardini A, Giannini S. Navigation-assisted totalknee arthroplasty in knees with osteoarthritis due to extra-articular deformity.Knee Surg Sports Traumatol Arthrosc Jul. 29 2011 [Epub ahead of print].
[38] Huang TW, Hsu WH, Peng KT, Wen-Wei Hsu R, Weng YJ, Shen WJ. Total kneearthroplasty with use of computer-assisted navigation compared with conven-tional guiding systems in the same patient: radiographic results in Asian pa-tients. J Bone Joint Surg Am 2011;93:1197–202.
[39] Shao JJ, Zhang XL, Wang Q, Chen YS, Shen H, Jiang Y. Total knee arthroplastyusing computer assisted navigation in patients with severe valgus deformity ofthe knee. Chin Med J (Engl) 2010;123:2666–70.
[40] Hsu WH, Hsu RW, Weng YJ. Effect of preoperative deformity on postoperativeleg axis in total knee arthroplasty: a prospective randomized study. Knee SurgSports Traumatol Arthrosc 2010;18:1323–7.
[41] Takasaki M, Matsuda S, Fukagawa S, Mitsuyasu H, Miura H, Iwamoto Y. Accuracyof image-free navigation for severely deformed knees. Knee Surg Sports Trauma-tol Arthrosc 2010;18:763–8.
[42] Schmitt J, Hauk C, Kienapfel H, Pfeiffer M, Efe T, Fuchs-Winkelmann S, et al. Nav-igation of total knee arthroplasty: rotation of components and clinical results in aprospectively randomized study. BMC Musculoskelet Disord 2011;12:16.
[43] Yang JH, Yoon JR, Pandher DS, Oh KJ. Clinical and radiologic outcomes of contem-porary 3 techniques of TKA. Orthopedics 2010;33(10 Suppl):76–81.
[44] Hiscox CM, Bohm ER, Turgeon TR, Hedden DR, Burnell CD. Randomized trial ofcomputer-assisted knee arthroplasty: impact on clinical and radiographic out-comes. J Arthroplasty May 16 2011 [Epub ahead of print].
[45] Cheng T, Zhang G, Zhang X. Clinical and radiographic outcomes of image-basedcomputer-assisted total knee arthroplasty: an evidence-based evaluation. SurgInnov 2011;18:15–20.
[46] Higgins J, Green S. Cochrane Handbook for Systematic Reviews of InterventionsVersion 5.1.0. http://www.cochrane.org/training/cochrane-handbook.
[47] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematicreviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097.
[48] Church JS, Scadden JE, Gupta RR, Cokis C, Williams KA, Janes GC. Embolic phe-nomena during computer-assisted and conventional total knee replacement.J Bone Joint Surg Br 2007;89:481–5.
[49] Decking R, Markmann Y, Fuchs J, Puhl W, Scharf HP. Leg axis after computer nav-igated total knee arthroplasty: a prospective randomized trial comparing com-puter navigated and manual implantation. J Arthroplasty 2005;20:282–8.
[50] Kim YH, Kim JS, Yoon SH. Alignment and orientation of the components in totalknee replacement with and without navigation support: a prospective, rando-mised study. J Bone Joint Surg Br 2007;89:471–6.
[51] Lützner J, Krummenauer F,Wolf C, Günther KP, Kirschner S. Computer-assisted andconventional total knee replacement: a comparative, prospective, randomisedstudy with radiological and CT evaluation. J Bone Joint Surg Br 2008;90:1039–44.
[52] Matziolis G, Krocker D, Weiss U, Tohtz S, Perka C. A prospective, randomizedstudy of computer-assisted and conventional total knee arthroplasty. Three-dimensional evaluation of implant alignment and rotation. J Bone Joint SurgAm 2007;89:236–43.
[53] Mizu-uchi H, Matsuda S, Miura H, Okazaki K, Akasaki Y, Iwamoto Y. The evalua-tion of post-operative alignment in total knee replacement using a CT-basednavigation system. J Bone Joint Surg Br 2008;90:1025–31.
[54] Mullaji A, Kanna R, Marawar S, Kohli A, Sharma A. Comparison of limb and com-ponent alignment using computer-assisted navigation versus image intensifier-guided conventional total knee arthroplasty: a prospective, randomized,single-surgeon study of 467 knees. J Arthroplasty 2007;22:953–9.
[55] O'ConnorMI, BrodersenMP, Feinglass NG, Leone B, Crook JE, Switzer BE. Fat emboliin total knee arthroplasty a prospective randomized study of computer-assistednavigation vs standard surgical technique. J Arthroplasty 2010;25:1034–40.
[56] Perlick L, Bäthis H, Tingart M, Perlick C, Grifka J. Navigation in total-knee arthro-plasty: CT-based implantation compared with the conventional technique. ActaOrthop Scand 2004;75:464–70.
[57] Perlick L, Bäthis H, Lerch K, Lüring C, Tingart M, Grifka J. Navigated implantationof total knee endoprostheses in secondary knee osteoarthritis of rheumatoid ar-thritis patients as compared to conventional technique. Z Rheumatol 2004;63:140–6.
[58] Sparmann M, Wolke B, Czupalla H, Banzer D, Zink A. Positioning of total kneearthroplasty with and without navigation support. A prospective, randomizedstudy. J Bone Joint Surg Br 2003;85:830–5.
[59] Victor J, Hoste D. Image-based computer-assisted total knee arthroplasty leads tolower variability in coronal alignment. Clin Orthop Relat Res 2004;428:131–9.
[60] Weng YJ, Hsu RW, Hsu WH. Comparison of computer-assisted navigation andconventional instrumentation for bilateral total knee arthroplasty. J Arthroplasty2009;24:668–73.
[61] Zhang W, Shao JJ, Zhang XL. The effect of computer assisted navigation on axialalignment of lower extremity in total knee replacement. Zhonghua gu ke zazhi 2008;28:819–23.
[62] Zumstein MA, Frauchiger L, Wyss D, Hess R, Ballmer PM. Is restricted femoralnavigation sufficient for accuracy of total knee arthroplasty? Clin Orthop RelatRes 2006;451:80–6.
[63] Spencer JM, Chauhan SK, Sloan K, Taylor A, Beaver RJ. Computer navigation ver-sus conventional total knee replacement: no difference in functional results attwo years. J Bone Joint Surg Br 2007;89:477–80.
[64] Tingart M, Lüring C, Bäthis H, Beckmann J, Grifka J, Perlick L. Computer-assistedtotal knee arthroplasty versus the conventional technique: how precise is navi-gation in clinical routine? Knee Surg Sports Traumatol Arthrosc 2008;16:44–50.
[65] Bauwens K,Matthes G,WichM, Gebhard F, Hanson B, Ekkernkamp A, et al. Navigat-ed total knee replacement. A meta-analysis. J Bone Joint Surg Am 2007;89:261–9.
[66] Bach CM, Mayr E, Liebensteiner M, Gstöttner M, Nogler M, Thaler M. Correlationbetween radiographic assessment and quality of life after total knee arthro-plasty. Knee 2009;16:207–10.
[67] Fornasier VL, Goodman SB, Protzner K, Kamel M, Song Y, Shojaci A. The role of im-plant alignment on stability and particles on periprosthetic osteolysis—a rabbitmodel of implant failure. J Biomed Mater Res B Appl Biomater 2004;70:179–86.
[68] Czurda T, Fennema P, Baumgartner M, Ritschl P. The association between com-ponent malalignment and post-operative pain following navigation-assistedtotal knee arthroplasty: results of a cohort/nested case–control study. KneeSurg Sports Traumatol Arthrosc 2010;18:863–9.
[69] Longstaff LM, Sloan K, Stamp N, Scaddan M, Beaver R. Good alignment after totalknee arthroplasty leads to faster rehabilitation and better function. J Arthro-plasty 2009;24:570–8.
[70] Parratte S, PagnanoMW, Trousdale RT, Berry DJ. Effect of postoperative mechanicalaxis alignment on thefifteen-year survival ofmodern, cemented total knee replace-ments. J Bone Joint Surg Am 2010;92:2143–9.
[71] Incavo SJ, Wild JJ, Coughlin KM, Beynnon BD. Early revision for component mal-rotation in total knee arthroplasty. Clin Orthop Relat Res 2007;458:131–6.
[72] Forster MC. Survival analysis of primary cemented total knee arthroplasty:which designs last? J Arthroplasty 2003;18:265–70.
[73] Cheng T, Liu T, Zhang G, Peng X, Zhang X. Does minimally invasive surgery im-prove short-term recovery in total knee arthroplasty? Clin Orthop Relat Res2010;468:1635–48.
[74] Cheng T, Zhang G, Zhang X. Metal-backed versus all-polyethylene tibial compo-nents in total knee arthroplasty: a meta-analysis and systematic review of ran-domized controlled trials. Acta Orthop 2011;82(5):589–95.
[75] Nicoll D, Rowley DI. Internal rotational error of the tibial component is a majorcause of pain after total knee replacement. J Bone Joint Surg Br 2010;92:1238–44.
[76] Lüring C, Kauper M, Bäthis H, Perlick L, Beckmann J, Grifka J, et al. A five to sevenyear follow-up comparing computer-assisted vs freehand TKR with regard toclinical parameters. Int Orthop Jun. 15 2011 [Epub ahead of print].
[77] Seon JK, Song EK, Park SJ, Lee DS. The use of navigation to obtain rectangularflexion and extension gaps during primary total knee arthroplasty and midtermclinical results. J Arthroplasty 2011;26:582–90.
[78] Ishida K, Matsumoto T, Tsumura N, Kubo S, Kitagawa A, Chin T, et al. Mid-term out-comes of computer-assisted total knee arthroplasty. Knee Surg Sports TraumatolArthrosc 2011;19:1107–12.
[79] Molfetta L, Caldo D. Computer navigation versus conventional implantation forvarus knee total arthroplasty: a case–control study at 5 years follow-up. Knee2008;15:75–9.
[80] Kamat YD, Aurakzai KM, Adhikari AR, Matthews D, Kalairajah Y, Field RE. Doescomputer navigation in total knee arthroplasty improve patient outcome at mid-term follow-up? Int Orthop 2009;33:1567–70.
[81] Hernández-Vaquero D, Suarez-Vazquez A, Iglesias-Fernandez S. Can computerassistance improve the clinical and functional scores in total knee arthroplasty?Clin Orthop Relat Res 2011;469:3436–42.
[82] Gøthesen O, Espehaug B, Havelin L, Petursson G, Furnes O. Short-term outcomeof 1,465 computer-navigated primary total knee replacements 2005–2008.Acta Orthop 2011;82:293–300.
[83] Shen H, Zhang N, Zhang X, Ji W. C-reactive protein levels after 4 types of arthro-plasty. Acta Orthop 2009;80:330–3.
[84] Conteduca F, Massai F, Iorio R, Zanzotto E, Luzon D, Ferretti A. Blood loss incomputer-assisted mobile bearing total knee arthroplasty. A comparison ofcomputer-assisted surgery with a conventional technique. Int Orthop 2009;33:1609–13.
[85] Kalairajah Y, Cossey A, Verrall GM, Ludbrook G, Spriggins AJ. Are systemic embolireduced in computer-assisted knee surgery?: a prospective, randomised, clinicaltrial. J Bone Joint Surg Br 2006;88:198–202.
[86] Ooi LH, Lo NN, Yeo SJ, Ong BC, Ding ZP, Lefi A. Does computer-assisted surgicalnavigation total knee arthroplasty reduce venous thromboembolism com-pared with conventional total knee arthroplasty? Singapore Med J 2008;49:610–4.
[87] Kim YH, Kim JS, Hong KS, Kim YJ, Kim JH. Prevalence of fat embolism after totalknee arthroplasty performed with or without computer navigation. J Bone JointSurg Am 2008;90:123–8.
[88] DeoH,West G, Butcher C, Lewis P. Theprevalence of cognitive dysfunction after con-ventional and computer-assisted total knee replacement. Knee 2011;18:117–20.
[89] Haytmanek CT, Pour AE, Restrepo C, Nikhil J, Parvizi J, Hozack WJ. Cognition fol-lowing computer-assisted total knee arthroplasty: a prospective cohort study.J Bone Joint Surg Am 2010;92:92–7.
[90] Park SE, Lee CT. Comparison of robotic-assisted and conventional manual im-plantation of a primary total knee arthroplasty. J Arthroplasty 2007;22:1054–9.
[91] Jung HJ, Jung YB, Song KS, Park SJ, Lee JS. Fractures associated with computer-navigated total knee arthroplasty. A report of two cases. J Bone Joint Surg Am2007;89:2280–4.
[92] Ossendorf C, Fuchs B, Koch P. Femoral stress fracture after computer navigatedtotal knee arthroplasty. Knee 2006;13:397–9.
[93] Beldame J, Boisrenoult P, Beaufils P. Pin track induced fractures aroundcomputer-assisted TKA. Orthop Traumatol Surg Res 2010;96(3):249–55.
[94] Jung KA, Lee SC, Ahn NK, Song MB, Nam CH, Shon OJ. Delayed femoral fracturethrough a tracker pin site after navigated total knee arthroplasty. J Arthroplasty2011;26:505.e9–505.e11.
[95] Hoke D, Jafari SM, Orozco F, Ong A. Tibial shaft stress fractures resulting fromplacement of navigation tracker pins. J Arthroplasty 2011;26:504.e5–8.
[96] Berning ET, Fowler RM. Thermal damage and tracker-pin track infection in computer-navigated total knee arthroplasty. J Arthroplasty 2011;26:977.e21–4.
[97] Wysocki RW, Sheinkop MB, Virkus WW, Della Valle CJ. Femoral fracture througha previous pin site after computer-assisted total knee arthroplasty. J Arthroplasty2008;23:462–5.
[98] Kurtz SM, Ong KL, Lau E, Bozic KJ, Berry D, Parvizi J. Prosthetic joint infection riskafter TKA in the Medicare population. Clin Orthop Relat Res 2010;468:52–6.
[99] Browne JA, Cook C, Hofmann AA, Bolognesi MP. Postoperative morbidity andmortality following total knee arthroplasty with computer navigation. Knee2010;17:152–6.
[100] Graham DJ, Harvie P, Sloan K, Beaver RJ. Morbidity of navigated vs conventionaltotal knee arthroplasty a retrospective review of 327 cases. J Arthroplasty Apr. 62011 [Epub ahead of print].
[101] Slover JD, Tosteson AN, Bozic KJ, Rubash HE, Malchau H. Impact of hospital vol-ume on the economic value of computer navigation for total knee replacement.J Bone Joint Surg Am 2008;90:1492–500.
[102] Soohoo NF, Zingmond D, Lieberman JR, Ko CY. Primary total knee arthroplasty inCalifornia 1991 to 2001: does hospital volume affect outcomes? J Arthroplasty2006;21:199–205.
[103] Katz JN, Barrett J, Mahomed NN, Baron JA, Wright RJ, Losina E. Association be-tween hospital and surgeon procedure volume and the outcomes of total kneereplacement. J Bone Joint Surg Am 2004;86:1909–16.
[104] Yau WP, Chiu KY, Zuo JL, Tang WM, Ng TP. Computer navigation did not improvealignment in a lower-volume total knee practice. Clin Orthop Relat Res 2008;466:935–45.
[105] Beringer DC, Patel JJ, Bozic KJ. An overview of economic issues in computer-assisted total joint arthroplasty. Clin Orthop Relat Res 2007;463:26–30.
