Total knee arthroplasty (TKA) is generally considered the standard treatment for end-stage osteoarthritis of the knee when nonoperative treatments are no longer helpful. Accurate alignment may correlate with good clinical outcomes,1,2 whereas malalignment may result in early loosening, lower functional scores, abnormal stress, and a higher implant failure rate.3–5 In the major arthroplasty centers, the rate of implant malposition could reach 20% to 40% in patients who undergo TKA with conventional instrumentation (CLI).6–8 To limit implant malpositioning, smart tools such as computer-assisted surgery (CAS) and patient-specific instrumentation (PSI) have been developed. Although CAS has been reported to improve alignment and reduce variation,9 it prolongs operative time and increases cost with a risk of pin-related complications.10,11 One recent alternative has been the development of PSI, which uses anatomical data obtained primarily from preoperative computed tomography (CT) or magnetic resonance imaging (MRI) to create disposable cutting blocks to the patient’s individual anatomy. Patient-specific instrumentation has the potential to improve component accuracy, eliminate the alignment outliers, reduce variation in implant position, and save operative time for TKA instruments.
Despite its theoretical advantage, the effectiveness of PSI remains controversial. A few clinical studies reported that PSI resulted in better component alignment accuracy than conventional TKA instrumentation,12–14 whereas others reported that PSI did not surpass the accuracy of the CLI technique.15–18 The inconsistent results may be due to small sample sizes, diverse ethnicities, and different instruments.
Several reviews have reported that PSI may achieve a high degree of mechanical alignment compared with CLI.19,20 By using meta-analysis to synthesize the evidence and estimate the effect size, the current authors sought to evaluate the performance of PSI compared with CLI in patients undergoing TKA. The authors hypothesized that PSI may have better performance than CLI in TKA concerning (1) outliers, (2) operative time, (3) blood loss, and (4) length of hospital stay.
Materials and Methods
Data Sources and Searches
The electronic databases PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials were searched for articles published between 2000 to March 2014 using the following terms: patient-specific, patient-matched, custom cutting, custom fit, total knee arthroplasty, total knee replacement, TKA, and TKR. Reference lists of the relevant articles were manually searched for additional trials. Moreover, gray literature was searched using Google search engines and the registration centers of clinical trials ( www.chictr.org/cn; www.who.int/ictrp/en; www.clinicaltrials.gov; www.nrr.nhs.uk; www.actr.org.au; www.controlled-trials.com; www.trialscentral.org). No languages or publication statuses were restricted.
The following inclusion criteria were used: (1) randomized, controlled trial (RCT) or non-RCT; (2) performed in vivo; (3) adult patients underwent primary TKA; (4) compared PSI with CLI; (5) reported the outliers of at least 1 of the following outcomes: mechanical axis, coronal femoral component (CFC), coronal tibial component (CTC), sagittal femoral component (SFC), sagittal tibial component (STC), and femoral component rotation (FCR); and (6) included at least 40 patients. If multiple studies with overlapping data were identified, the authors included only the published report with the largest sample size. All studies that did not meet these criteria were excluded.
Data Extraction and Outcome Measures
Two reviewers (C.S., Z.-H.T.) independently extracted data using a standardized extraction form. Disagreements were resolved by discussion with a third reviewer (J.-Z.H.) until consensus was reached. The primary outcomes included the outliers of mechanical axis, CFC, CTC, SFC, STC, and FCR. The cutoff value for the outliers of mechanical axis, CFC, CTC, and FCR was 3°,12–18,21,22 whereas the cutoff values for the outliers of SFC and STC were between 3° and 7° according to different PSI systems.15,16,23,24 Secondary outcomes included operative time, blood loss, and length of hospital stay.
The Cochrane Collaboration’s tools were used to assess the risk of bias. The items for quality assessment involved 7 criteria: (1) sequence generation, (2) allocation concealment, (3) blinding of participants and personnel, (4) blinding of outcome assessment, (5) incomplete outcome data, (6) selective outcome reporting, and (7) other sources of bias. Each item of the included studies was classified as yes (low risk), no (high risk), or unclear (unclear risk).
For dichotomous outcomes, the odds ratio (OR) and 95% confidence interval (CI) were calculated as the summary statistics. For continuous outcomes, mean and SD were used to calculate the weighted mean difference (WMD) with 95% CI in the meta-analysis. Heterogeneity between studies was quantified using P and I2 values.25 Heterogeneity was considered significant if the P value was less than .1 or the I2 value was greater than 50%. If heterogeneity was significant, meta-analysis was conducted using the random-effects model.26 Otherwise, the fixed-effects model was used. The fixed-effects model assumes that the only source of uncertainty derives from the within-study error. The random-effects model makes the assumption that different effects have the same source of uncertainty, plus an additional between-studies variance.
In addition, subgroup analyses (RCT vs non-RCT, MRI vs CT, different systems of PSI) were performed to look at more narrow subsets of the studies. Review Manager version 5.2 statistical software (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark) and STATA version 11.0 statistical software (Stata Corporation, College Station, Texas) were used for all analyses. Publication bias was examined by Egger’s linear regression test27 and Begg’s test.28
Of the 401 potentially relevant studies identified through the literature search (Figure 1), 28 studies were retrieved for full-text assessment. After reviewing the full text, the authors found that 4 studies had overlapped data, 8 studies did not report the outcomes of interest, and 2 studies had a sample size less than 40. Hence, those 14 studies were excluded. Therefore, 14 studies (7 RCTs and 7 non-RCTs) with 1906 patients were included for meta-analysis.12–18,21–24,29–31
Flow chart of study selection based on the inclusion and exclusion criteria.
Study Characteristics and Quality Assessment
The basic characteristics of the included studies are presented in Table 1. All of the included studies were published in English between 2012 and 2014. The sample sizes of the included studies ranged from 40 to 306. Among the 14 studies, 13 reported the mechanical axis,12,14–18,21–24,29–31 7 reported the CFC,15–17,21,23,24,31 7 reported the CTC,15–17,21,23,24,31 6 reported the SFC,15,16,21,23,24,31 6 reported the STC,15,16,21,23,24,31 4 reported the FCR,13,16,21,31 9 reported operative time,12,15,17,21–24,30,31 2 reported length of hospital stay,15,17 and 4 reported blood loss.15,17,21,22 The assessment of methodological quality is summarized in Table 2. Seven studies15–17,22,24,29 used the method of randomization in the trial design, whereas 2 studies16,29 used allocation concealment and 9 studies14–17,22–24,29,30 used blinding. A disagreement between the 2 reviewers regarding whether the outcome assessment was blinding in the study by Roh et al21 was resolved by the third reviewer.
Risk of Bias in Included Studies
Primary Outcomes: Mechanical Axis, CFC, CTC, SFC, STC, and FCR. Meta-analysis demonstrated no significant differences between the PSI and CLI groups in mechanical axis (OR, 0.99; 95% CI, 0.65 to 1.49; P=.94; I2= 56%; Figure 2), CFC (OR, 0.64; 95% CI, 0.39 to 1.06; P=.08; I2=0%; Figure 3), SFC (OR, 0.94; 95% CI, 0.67 to 1.33; P=.74; I2=15%; Figure 4), or FCR (OR, 0.84; 95% CI, 0.33 to 2.11; P=.70; I2=63%; Figure 5).
Forest plot of mechanical axis. Individual studies are listed on the left. The squares and horizontal lines correspond to the study-specific odds ratio and 95% confidence interval. The large diamond at the bottom represents the pooled treatment effect of all studies. It lies crossing the midline, representing no significant difference between the 2 groups.
Forest plot of coronal femoral component. Individual studies are listed on the left. Pooled treatment effects of each study are represented on the right. The large diamond at the bottom represents the pooled treatment effect of all studies. It lies crossing the midline, representing no significant difference between the 2 groups.
Forest plot of sagittal femoral component. Individual studies are listed on the left. Pooled treatment effects of each study are represented on the right. The large diamond at the bottom represents the pooled treatment effect of all studies. It lies crossing the midline, representing no significant difference between the 2 groups.
Forest plot of femoral component rotation. Individual studies are listed on the left. Pooled treatment effects of each study are represented on the right. The large diamond at the bottom represents the pooled treatment effect of all studies. It lies crossing the midline, representing no significant difference between the 2 groups.
Subgroup analyses were performed to identify potential sources of heterogeneity. The authors found that type of study design (RCT vs non-RCT), method of bone modeling used preoperatively (MRI vs CT), and different systems of PSI (Signature, Biomet, Warsaw, Indiana; Patient-Specific Instruments, Zimmer, Warsaw, Indiana; Visionaire, Smith & Nephew, Memphis, Tennessee; and TruMatch, DePuy, Warsaw, Indiana) had no significant effect on the outcomes (Table 3).
Meta-analysis Including Subgroup Analysis
Meta-analysis showed that the outliers of CTC (OR, 2.29; 95% CI, 1.20 to 4.35; P=.01; I2=7%; Figure 6) and STC (OR, 1.67; 95% CI, 1.16 to 2.42; P=.006; I2=14%; Figure 7) were significantly lower in the CLI group than in the PSI group. However, subgroup analyses (MRI vs CT, different systems of PSI) revealed no significant differences between the 2 groups (Table 3).
Forest plot of coronal tibial component. Individual studies are listed on the left. The squares and horizontal lines correspond to the study-specific odds ratio and 95% confidence interval. The large diamond at the bottom represents the pooled treatment effect of all studies. It lies exclusively to the right and does not cross the midline, representing a significant difference favoring conventional instrumentation.
Forest plot of sagittal tibial component. Individual studies are listed on the left. The squares and horizontal lines correspond to the study-specific odds ratio and 95% confidence interval. The large diamond at the bottom represents the pooled treatment effect of all studies. It lies exclusively to the right and does not cross the midline, representing a significant difference favoring conventional instrumentation.
Secondary Outcomes: Operative Time, Blood Loss, Length of Hospital Stay, Publication Bias. Regarding operative time, data from 7 studies could be pooled. Two trials provided means but not SDs.17,21 Meta-analysis showed no significant difference between the 2 groups (WMD, −1.78; 95% CI, −4.45 to 0.90; P=.19) but with high heterogeneity (I2=89%). Subgroup analysis showed similar results (Table 3).
Four studies reported blood loss.15,17,21,22 Boonen et al15 reported that blood loss was significantly less in the PSI group than in the CLI group. However, the other 3 studies detected no difference between the groups.17,21,22 The current authors could not pool data for mean differences in mean total blood loss because only means were reported in these studies.
Two studies reported length of hospital stay.15,17 Neither study reported significant differences between the 2 groups.
The authors checked American Academy of Orthopaedic Surgeons disclosures for the authors of the assessed studies and found that authors of 5 studies were sponsored by companies.13,14,18,24,30 However, no publication bias was detected by Egger’s test (P=.965) and Begg’s test (P=.891).
Although some studies have demonstrated that PSI has several advantages over CLI in component accuracy and elimination of outliers, controversy remains about their conflicting conclusions. Previously, only 1 systematic review compared the performance of PSI with CLI in patients receiving TKA.19 Recently, several RCTs have been updated.15,17,21,22,24,29 Therefore, the current authors conducted the first meta-analysis to compare the accuracy between the 2 technologies based on RCTs and non-RCTs.
The RCT is one of the most powerful tools of experimental study. People are allocated at random to receive one of several clinical interventions; the process of randomization is determined by neither the investigators nor the participants. In the current meta-analysis of 14 studies involving 7 RCTs and 7 non-RCTs with 1906 patients, there were no significant differences in the outliers of mechanical axis, CFC, SFC, FCR, operative time, blood loss, and length of hospital stay between the PSI and CLI groups. Notably, CLI performed superiorly in the outliers of CTC and STC. To verify the reliability of their conclusion, the authors also performed subgroup analysis according to the level of trial (RCT vs non-RCT), the method of bone modeling used preoperatively (MRI vs CT), and the different systems of PSI (Signature, Visionaire, Patient-Specific Instruments, and TruMatch). The results were consistent with the overall conclusion, except for CTC and STC.
The authors’ conclusion was in accordance with previous studies, which also proved that PSI did not improve accuracy in TKA.15–18,31 In recent RCTs, Roh et al21 and Chareancholvanich et al17 found no differences in the outliers of mechanical axis and femoral and tibial alignment between PSI and CLI groups. One explanation is the differences in surgical experience using these 2 instrumentations: surgeons have performed hundreds of TKAs using CLI but a far smaller number using PSI. Therefore, lack of familiarity with the PSI may affect results. Another reason is that the PSI is a pure geometric nature of the approach, not taking soft tissue status into account. With regard to the property of soft tissue, balancing a TKA is extremely important, especially in larger deformities. However, as static tests, preoperative MRI and CT for PSI do not take into account soft tissue laxity or tightness. No dynamic information is used to make the cutting blocks. In addition, many PSI techniques do not adjust well to significant flexion contractures or varus/valgus deformities due to lack of preoperative data about passive correction of the deformity. Several studies reported that the gap-balancing technique resulted in less alignment and femoral rotation outliers compared with the measured resection technique.32–34 Paternostre et al35 reported that a more concave side ligamentous release or more constraint was necessary than anticipated based on the PSI alignment result. They found that underestimation of deformity because of soft tissue results in an unsatisfactory amount of implant constraint. Computer assisted surgery, a technique for evaluation of soft tissue and bony resection in real time, could provide precise dynamic evaluation of knee alignment.36 When treating cases with significant flexion contractures or varus/valgus deformities, CAS techniques are a good choice because of their real-time evaluation.36,37
In the current study, there were more outliers of CTC and STC in the PSI group than in the CLI group. This can be explained by the nature of the technology. Chen et al23 indicated that the PSI tibial guide had a design flaw that pinned the cutting guide at a medial offset. Approaching the knee from the medial side essentially converts the cutting jig to a side cutting jig rather than the usual anterior reference jig. Failure to handle the saw tends to cause an oblique joint osteotomy, resulting in malpositioning of the tibial component. These results were confirmed by another study,16 which also found more outliers of CTC and STC in the PSI group. However, the subgroup analyses (MRI vs CT, different systems of PSI) were not significant between the 2 groups. This was likely due to the small number of studies and, hence, insufficient statistical power of the current study.
Some studies have proven that coronal malalignment can lead to a higher implant failure rate.3,38 This may result from asymmetric tibiofemoral tracking and subsequent abnormal stresses at the weight-bearing surfaces. A retrospective study showed that the subsequent loosening of implants in patients with coronal malalignment exceeding 3° occurred at a rate of 24%, compared with a rate of 3% in patients within 3° of neutral mechanical axis.39 To avoid these complications, some surgeons recommend a postoperative alignment within the range of 0°±3° of the mechanical axis.39,40 In the current study, the cutoff value for the outliers of mechanical axis, coronal component CTC and CFC alignment, and FCR was 3°. The pooled results showed that the use of PSI did not reduce the outliers of mechanical axis, coronal component CTC and CFC alignment, and FCR in TKA.
Component alignment in the sagittal plane was less reliable than that in the coronal plane.41 Optimal prosthetic alignment for TKA in the sagittal plane is uniform.41–43 For example, the target of the Signature was 3° for the femoral and tibial components, whereas the target of the Visionaire was 4° for the femoral components and 3° for the tibial components. Interestingly, subgroup analyses yielded similar results when using different PSI systems. Moreover, some authors claimed that bone models generated from MRI scans were dimensionally less accurate than those generated from CT scans.44 The current authors performed subgroup analysis based on the preoperative MRI or CT used for the planning of the cutting jigs. The results showed no differences in the outliers of mechanical axis, CFC, CTC, SFC, and STC between the MRI and CT subgroups.
In addition, the authors’ results showed that operative time was not decreased when using PSI in TKA. Manufacturers often claim that a reduction in operative time is an important superiority of PSI. The manufacturers surmised that after eliminating the bulk and complexity of CLI, operating room efficiency was enhanced, and thus intraoperative time was reduced. The opinion should be challenged while taking into account the time spent for preoperative planning and the additional cost for the MRI or CT scan. Moreover, some authors have raised concerns about the complications of CAS in TKA, which introduces pin-site loosening or fracture.11 However, none of studies included in the current meta-analysis reported the related complications.
It is a challenge to restore function and alignment when treating knee arthritis with CLI in patients who have an extra-articular deformity from a malunion or with retained hardware. A multicenter retrospective study by Thienpont et al45 demonstrated that the use of PSI systems to perform TKA in patients without access to the intramedullary canal because of extra-articular deformity or fixation devices improved function and restored limb alignment.
This meta-analysis has several potential limitations. First, the authors included studies with different levels of evidence (RCTs and non-RCTs). As a result of study design limitations, using non-RCTs may introduce a bias. Thus, they performed subgroup analyses for the RCTs and non-RCTs and found similar results. Second, allocation concealment was not used in most studies, which may introduce selection bias. Third, although subgroup analyses yielded similar results in different PSI systems, the nonuniform systems may bring bias. Fourth, 5 studies were sponsored by companies, and their results should be interpreted with caution.13,14,18,24,30 Funding source should be seriously assessed in systematic reviews. It has been reported that results are more likely to favor a product when an investigator has a financial interest in or funding from the product’s manufacturer.46 Due to these limitations, further studies are warranted.
The authors believe that PSI may have no advantage over CLI for patients undergoing TKA. It is likely that the use of PSI neither improves the alignment accuracy nor benefits the clinical outcomes of TKA, including operative time, blood loss, and length of hospital stay. Surgery with PSI as an alternative to conventional TKA should be used with caution. Perhaps PSI can be used where conventional instrumentation cannot, such as in previous femoral fractures with deformities or retained hardware. Further well-designed, large-sample RCTs with long follow-up are warranted.
- Choong PF, Dowsey MM, Stoney JD. Does accurate anatomical alignment result in better function and quality of life? Comparing conventional and computer-assisted total knee arthroplasty. J Arthroplasty. 2009; 24(4):560–569. doi:10.1016/j.arth.2008.02.018 [CrossRef]
- Longstaff LM, Sloan K, Stamp N, Scaddan M, Beaver R. Good alignment after total knee arthroplasty leads to faster rehabilitation and better function. J Arthroplasty. 2009; 24(4):570–578. doi:10.1016/j.arth.2008.03.002 [CrossRef]
- Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM. Insall Award paper: why are total knee arthroplasties failing today?Clin Orthop Relat Res. 2002; (404):7–13. doi:10.1097/00003086-200211000-00003 [CrossRef]
- Bargren JH, Blaha JD, Freeman MA. Alignment in total knee arthroplasty: correlated biomechanical and clinical observations. Clin Orthop Relat Res. 1983; (173):178–183.
- Lotke PA, Ecker ML. Influence of positioning of prosthesis in total knee replacement. J Bone Joint Surg Am. 1977; 59(1):77–79.
- Bathis H, Perlick L, Tingart M, Luring C, Zurakowski D, Grifka J. Alignment in total knee arthroplasty: a comparison of computer-assisted surgery with the conventional technique. J Bone Joint Surg Br. 2004; 86(5):682–687. doi:10.1302/0301-620X.86B5.14927 [CrossRef]
- Iorio R, Bolle G, Conteduca F, et al. Accuracy of manual instrumentation of tibial cutting guide in total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc. 2013; 21(10):2296–2300. doi:10.1007/s00167-012-2005-7 [CrossRef]
- Mahaluxmivala J, Bankes MJ, Nicolai P, Aldam CH, Allen PW. The effect of surgeon experience on component positioning in 673 Press Fit Condylar posterior cruciate-sacrificing total knee arthroplasties. J Arthroplasty. 2001; 16(5):635–640. doi:10.1054/arth.2001.23569 [CrossRef]
- Cheng T, Zhao S, Peng X, Zhang X. Does computer-assisted surgery improve postoperative leg alignment and implant positioning following total knee arthroplasty? A meta-analysis of randomized controlled trials. Knee Surg Sports Traumatol Arthrosc. 2012; 20(7):1307–1322. doi:10.1007/s00167-011-1588-8 [CrossRef]
- Quack VM, Kathrein S, Rath B, Tingart M, Luring C. Computer-assisted navigation in total knee arthroplasty: a review of literature. Biomed Tech (Berl). 2012; 57(4):269–275. doi:10.1515/bmt-2011-0096 [CrossRef]
- Beldame J, Boisrenoult P, Beaufils P. Pin track induced fractures around computer-assisted TKA. Orthop Traumatol Surg Res. 2010; 96(3):249–255. doi:10.1016/j.otsr.2009.12.005 [CrossRef]
- Daniilidis K, Tibesku CO. A comparison of conventional and patient-specific instruments in total knee arthroplasty. Int Orthop. 2014; 38(3):503–508. doi:10.1007/s00264-013-2028-9 [CrossRef]
- Heyse TJ, Tibesku CO. Improved femoral component rotation in TKA using patient-specific instrumentation. Knee. 2014; 21(1):268–271. doi:10.1016/j.knee.2012.10.009 [CrossRef]
- Ng VY, DeClaire JH, Berend KR, Gulick BC, Lombardi AV Jr, . Improved accuracy of alignment with patient-specific positioning guides compared with manual instrumentation in TKA. Clin Orthop Relat Res. 2012; 470(1):99–107. doi:10.1007/s11999-011-1996-6 [CrossRef]
- Boonen B, Schotanus MG, Kerens B, van der Weegen W, van Drumpt RA, Kort NP. Intraoperative results and radiological outcome of conventional and patient-specific surgery in total knee arthroplasty: a multicentre, randomised controlled trial. Knee Surg Sports Traumatol Arthrosc. 2013; 21(10):2206–2212. doi:10.1007/s00167-013-2620-y [CrossRef]
- Victor J, Dujardin J, Vandenneucker H, Arnout N, Bellemans J. Patient-specific guides do not improve accuracy in total knee arthroplasty: a prospective randomized controlled trial. Clin Orthop Relat Res. 2014; 472(1):263–271. doi:10.1007/s11999-013-2997-4 [CrossRef]
- Chareancholvanich K, Narkbunnam R, Pornrattanamaneewong C. A prospective randomised controlled study of patient-specific cutting guides compared with conventional instrumentation in total knee replacement. Bone Joint J. 2013; 95(3):354–359. doi:10.1302/0301-620X.95B3.29903 [CrossRef]
- Barrack RL, Ruh EL, Williams BM, Ford AD, Foreman K, Nunley RM. Patient specific cutting blocks are currently of no proven value. J Bone Joint Surg Br. 2012; 94(11 suppl):95–99. doi:10.1302/0301-620X.94B11.30834 [CrossRef]
- Thienpont E, Fennema P, Price A. Can technology improve alignment during knee arthroplasty. Knee. 2013; 20(suppl 1):S21–S28. doi:10.1016/S0968-0160(13)70005-X [CrossRef]
- Ast MP, Nam D, Haas SB. Patient-specific instrumentation for total knee arthroplasty: a review. Orthop Clin North Am. 2012; 43(5):e17–e22. doi:10.1016/j.ocl.2012.07.004 [CrossRef]
- Roh YW, Kim TW, Lee S, Seong SC, Lee MC. Is TKA using patient-specific instruments comparable to conventional TKA? A randomized controlled study of one system. Clin Orthop Relat Res. 2013; 471(12):3988–3995. doi:10.1007/s11999-013-3206-1 [CrossRef]
- Chotanaphuti T, Wangwittayakul V, Khuangsirikul S, Foojareonyos T. The accuracy of component alignment in custom cutting blocks compared with conventional total knee arthroplasty instrumentation: prospective control trial. Knee. 2014; 21(1):185–188. doi:10.1016/j.knee.2013.08.003 [CrossRef]
- Chen JY, Yeo SJ, Yew AK, et al. The radiological outcomes of patient-specific instrumentation versus conventional total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc. 2014; 22(3):630–635. doi:10.1007/s00167-013-2638-1 [CrossRef]
- Hamilton WG, Parks NL, Saxena A. Patient-specific instrumentation does not shorten surgical time: a prospective, randomized trial. J Arthroplasty. 2013; 28(8 suppl):96–100. doi:10.1016/j.arth.2013.04.049 [CrossRef]
- Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003; 327(7414):557–560. doi:10.1136/bmj.327.7414.557 [CrossRef]
- DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7(3):177–188. doi:10.1016/0197-2456(86)90046-2 [CrossRef]
- Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997; 315(7109):629–634. doi:10.1136/bmj.315.7109.629 [CrossRef]
- Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994; 50(4):1088–1101. doi:10.2307/2533446 [CrossRef]
- Parratte S, Blanc G, Boussemart T, Ollivier M, Le Corroller T, Argenson JN. Rotation in total knee arthroplasty: no difference between patient-specific and conventional instrumentation. Knee Surg Sports Traumatol Arthrosc. 2013; 21(10):2213–2219. doi:10.1007/s00167-013-2623-8 [CrossRef]
- Barrett W, Hoeffel D, Dalury D, Mason JB, Murphy J, Himden S. In-vivo alignment comparing patient specific instrumentation with both conventional and computer assisted surgery (CAS) instrumentation in total knee arthroplasty. J Arthroplasty. 2014; 29(2):343–347. doi:10.1016/j.arth.2013.06.029 [CrossRef]
- Marimuthu K, Chen DB, Harris IA, Wheatley E, Bryant CJ, Macdessi SJ. A multi-planar CT-based comparative analysis of patient-specific cutting guides with conventional instrumentation in total knee arthroplasty. J Arthroplasty. 2014; 29(6):1138–1142. doi:10.1016/j.arth.2013.12.019 [CrossRef]
- Lee DH, Park JH, Song DI, Padhy D, Jeong WK, Han SB. Accuracy of soft tissue balancing in TKA: comparison between navigation-assisted gap balancing and conventional measured resection. Knee Surg Sports Traumatol Arthrosc. 2010; 18(3):381–387. doi:10.1007/s00167-009-0983-x [CrossRef]
- Pang HN, Yeo SJ, Chong HC, Chin PL, Ong J, Lo NN. Computer-assisted gap balancing technique improves outcome in total knee arthroplasty, compared with conventional measured resection technique. Knee Surg Sports Traumatol Arthrosc. 2011; 19(9):1496–1503. doi:10.1007/s00167-011-1483-3 [CrossRef]
- Fehring TK. Rotational malalignment of the femoral component in total knee arthroplasty. Clin Orthop Relat Res. 2000; (380):72–79. doi:10.1097/00003086-200011000-00010 [CrossRef]
- Paternostre F, Schwab PE, Thienpont E. The difference between weight-bearing and non-weight-bearing alignment in patient-specific instrumentation planning. Knee Surg Sports Traumatol Arthrosc. 2014; 22(3):674–679. doi:10.1007/s00167-013-2687-5 [CrossRef]
- Schwarzkopf R, Hadley S, Abbasi M, Meere PA. Computer-assisted surgery patterns of ligamentous deformity of the knee: a clinical and cadaveric study. J Knee Surg. 2013; 26(4):233–238. doi:10.1055/s-0032-1329716 [CrossRef]
- Hsu WH, Hsu RW, Weng YJ. Effect of preoperative deformity on postoperative leg axis in total knee arthroplasty: a prospective randomized study. Knee Surg Sports Traumatol Arthrosc. 2010; 18(10):1323–1327. doi:10.1007/s00167-010-1146-9 [CrossRef]
- Berend ME, Ritter MA, Meding JB, et al. Tibial component failure mechanisms in total knee arthroplasty. Clin Orthop Relat Res. 2004; (428):26–34. doi:10.1097/01.blo.0000148578.22729.0e [CrossRef]
- Jeffery RS, Morris RW, Denham RA. Coronal alignment after total knee replacement. J Bone Joint Surg Br. 1991; 73(5):709–714.
- Ritter MA, Faris PM, Keating EM, Meding JB. Postoperative alignment of total knee replacement: its effect on survival. Clin Orthop Relat Res. 1994; (299):153–156.
- Minoda Y, Kobayashi A, Iwaki H, Ohashi H, Takaoka K. TKA sagittal alignment with navigation systems and conventional techniques vary only a few degrees. Clin Orthop Relat Res. 2009; 467(4):1000–1006. doi:10.1007/s11999-008-0449-3 [CrossRef]
- Collier MB, Engh CA Jr, McAuley JP, Ginn SD, Engh GA. Osteolysis after total knee arthroplasty: influence of tibial baseplate surface finish and sterilization of polyethylene insert: findings at five to ten years postoperatively. J Bone Joint Surg Am. 2005; 87(12):2702–2708. doi:10.2106/JBJS.E.00074 [CrossRef]
- Massin P, Gournay A. Optimization of the posterior condylar offset, tibial slope, and condylar roll-back in total knee arthroplasty. J Arthroplasty. 2006; 21(6):889–896. doi:10.1016/j.arth.2005.10.019 [CrossRef]
- White D, Chelule KL, Seedhom BB. Accuracy of MRI vs CT imaging with particular reference to patient specific templates for total knee replacement surgery. Int J Med Robot. 2008; 4(3):224–231. doi:10.1002/rcs.201 [CrossRef]
- Thienpont E, Paternostre F, Pietsch M, Hafez M, Howell S. Total knee arthroplasty with patient-specific instruments improves function and restores limb alignment in patients with extra-articular deformity. Knee. 2013; 20(6):407–411. doi:10.1016/j.knee.2013.07.001 [CrossRef]
- Stelfox HT, Chua G, O’Rourke K, Detsky AS. Conflict of interest in the debate over calcium-channel antagonists. N Engl J Med. 1998; 338(2):101–106. doi:10.1056/NEJM199801083380206 [CrossRef]
|Study||Country||Level of Evidence||Study Design||No. of Knees||Males, %||Mean Age (Range), y||Mean BMI (Range), kg/m2||PSI System||Outcomes|
|Daniilidis & Tibesku12||Germany||III||Retrospective cohort||150||156||63.3||50.6||66.1±8.8||65.0±9.2||31.7±4.6||30.3±5.3||Visionairea||MA, OT|
|Heyse & Tibesku13||Germany||III||Retrospective cohort||46||Visionaire||FCR|
|Ng et al14||United States||III||Retrospective cohort||105||55||Signatureb||MA|
|Boonen et al15||Netherlands||I||RCT||90||90||38||44||69±8.0||65±8.8||30.3||29.5||Signature||MA, CFC, CTC, SFC, STC, OT, MTBL, LHS|
|Victor et al16||Belgium||I||RCT||64||64||33||33||67 (52–87)||66 (36–92)||Signature, TruMatch,c Visionaire, Patient-Specific Instruments||MA, CFC, CTC, SFC, STC, FCR|
|Chareancholvanich et al17||Thailand||II||RCT||40||40||15||10||69.5 (55–84)||70.3 (53–85)||27.7 (20.2–44.15)||28.0 (22–39.6)||Patient-Specific Instrumentsd||MA, CFC, CTC, OT, MTBL, LHS|
|Barrack et al18||United States||III||Retrospective cohort||100||100||40||43||64.8 (31.6–90.1)||65.6 (42.8–91.0)||Signature||MA|
|Roh et al21||Korea||II||RCT||42||48||7||10||70±7.2||70±5.1||27±4.2||27±2.7||Signature||MA, CFC, CTC, SFC, STC, FCR, OT, MTBL|
|Chotanaphuti et al22||Thailand||II||RCT||40||40||69.7±5.5||69.3±5.5||25.0±2.4||25.0±2.1||TruMatch||MA, OT, MTBL|
|Chen et al23||Singapore||II||Prospective cohort||29||30||31||17||65±8||65±8||29.4±6.5||29.1±5.8||Patient-Specific Instruments||MA, CFC, CTC, SFC, STC, OT|
|Hamilton et al24||United States||II||RCT||26||26||54||27||68.1 (52–86)||67.6 (51–88)||30.9 (21.5–39.6)||31.1 (22–38.4)||TruMatch||MA, CFC, CTC, SFC, STC, OT|
|Parratte et al29||France||II||RCT||20||20||Patient-Specific Instruments||MA|
|Barrett et al30||United States||II||Prospective cohort||66||86||38||34||66.4±8.4||64.6±7.6||33.2±7.4||31.8±6.0||TruMatch||MA, OT|
|Marimuthu et al31||Australia||III||Retrospective cohort||115||185||68.3±8.8||67.6±9.7||30.8±8.6||30.6±5.4||Visionaire||MA, CFC, CTC, SFC, STC, OT|
Risk of Bias in Included Studies
|Study||Random Sequence Generation||Allocation Concealment||Blinding of Participants||Blinding of Outcome Assessment||Incomplete Outcome Data||Selective Reporting||Other Bias|
|Daniilidis & Tibesku12||No||No||No||Unclear||No||No||Unclear|
|Heyse & Tibesku13||No||No||No||No||Unclear||No||Unclear|
|Ng et al14||No||No||No||Yes (2 independent blinded investigators)||Unclear||No||Unclear|
|Boonen et al15||Yes (random number generator)||Unclear||Unclear||Yes (2 independent reviewers)||No||No||Unclear|
|Victor et al16||Yes (not reported)||Yes (sealed envelopes)||Yes||Yes (a blinded investigator)||No||No||Unclear|
|Chareancholvanich et al17||Yes (blocks-of-four method)||Unclear||Unclear||Yes (2 independent blinded investigators)||No||No||Unclear|
|Barrack et al18||No||No||No||No||Unclear||No||Unclear|
|Roh et al21||Yes (permuted block randomization program)||Unclear||Unclear||Unclear||No||No||Unclear|
|Chotanaphuti et al22||Yes (not reported)||No||No||Yes (two independent blinded investigators)||Unclear||No||Unclear|
|Chen et al23||No||No||No||Yes (2 independent blinded investigators)||Unclear||No||Unclear|
|Hamilton et al24||Yes (not reported)||No||No||Yes (a blinded investigator)||Unclear||No||Unclear|
|Parratte et al29||Yes (systematic sampling method)||Yes (sealed envelopes)||Yes||Yes (2 independent observers)||No||No||Unclear|
|Barrett et al30||No||No||No||Yes (a blinded investigator)||Unclear||No||Unclear|
|Marimuthu et al31||No||No||No||No||No||No||Unclear|
Meta-analysis Including Subgroup Analysis
|Subgroup and Outcome||No. of Studies||No. of Patients||OR (95% CI)||P||I2|
| Overall results||13||1812||0.99 (0.65 to 1.49)||.94||56%|
| RCT||7||635||1.17 (0.76 to 1.78)||.48||1%|
| Non-RCT||6||1177||0.94 (0.49 to 1.82)||.86||75%|
| PSI based on MRI||7||1313||0.96 (0.53 to 1.76)||.90||70%|
| PSI based on CT||4||374||1.25 (0.70 to 2.25)||.46||0%|
| Signaturea||4||618||1.08 (0.51 to 2.28)||.85||68%|
| Patient-Specific Instrumentsb||3||179||2.07 (0.75 to 5.70)||.16||41%|
| TruMatchc||3||284||1.27 (0.66 to 2.45)||.47||27%|
| Visionaired||2||606||0.56 (0.27 to 1.15)||.11||58%|
| Overall results||7||874||0.64 (0.39 to 1.06)||.08||0%|
| RCT||5||515||0.67 (0.32 to 1.41)||.29||22%|
| Non-RCT||2||359||0.76 (0.27 to 2.11)||.59||0%|
| PSI based on MRI||4||607||0.63 (0.34 to 1.17)||.14||18%|
| PSI based on CT||2||142||1.04 (0.31 to 3.50)||.95||0%|
| Signature||2||258||1.06 (0.45 to 2.45)||.89||0%|
| Patient-Specific Instruments||139||0.24 (0.02 to 3.03)||.27||55%|
| Overall results||7||874||2.29 (1.20 to 4.35)||.01||7%|
| RCT||5||515||2.62 (1.19 to 5.79)||.02||27%|
| Non-RCT||2||359||1.69 (0.55 to 5.22)||.36||0%|
| PSI based on MRI||4||607||2.00 (0.87 to 4.62)||.10||0%|
| PSI based on CT||2||142||1.20 (0.06 to 23.12)||.91||59%|
| Signature||2||258||1.28 (0.08 to 21.66)||.86||64%|
| Patient-Specific Instruments||2||139||1.38 (0.15 to 12.70)||.66||24%|
| Overall results||6||796||0.94 (0.67 to 1.33)||.74||15%|
| RCT||4||437||0.87 (0.58 to 1.31)||.51||41%|
| Non-RCT||2||359||1.16 (0.60 to 2.24)||.66||0%|
| PSI based on MRI||3||529||0.76 (0.48 to 1.19)||.22||39%|
| PSI based on CT||2||142||1.62 (0.65 to 4.08)||.30||0%|
| Signature||2||260||0.71 (0.27 to 1.89)||.09||40%|
| Overall results||6||796||1.67 (1.16 to 2.42)||.006||14%|
| RCT||4||437||1.80 (1.10 to 2.93)||.02||47%|
| Non-RCT||2||359||1.51 (0.86 to 2.66)||.15||0%|
| PSI based on MRI||3||529||1.42 (0.93 to 2.18)||.11||0%|
| PSI based on CT||2||142||1.47 (0.57 to 3.75)||.43||0%|
| Signature||2||260||1.23 (0.66 to 2.27)||.52||0%|
| Overall results||4||609||0.84 (0.33 to 2.11)||.70||63%|
| RCT||2||215||1.17 (0.56 to 2.43)||.67||0%|
| Non-RCT||2||394||0.39 (0.02 to 7.97)||.54||87%|
|Visionaire||2||394||0.39 (0.02 to 7.97)||.54||87%|
| Overall results||7||1117||−1.78 (−4.45 to 0.90)||.19||89%|
| RCT||3||300||−1.66 (−4.75 to 1.44)||.45||94%|
| Non-RCT||4||817||−0.81 (−2.25 to 0.62)||.27||38%|
| TruMatch||3||284||−1.66 (−8.23 to 4.92)||.62||94%|
| Visionaire||2||606||−0.91 (−3.51 to 1.69)||.49||55%|