The Knee 22 (2015) 47–50
Contents lists available at ScienceDirect
The Knee
Intra-operative deviation in limb alignment occurring at implantation in total knee arthroplasty D.F. Howie ⁎, G.J. Love 1, A.H. Deakin 1, A.W.G. Kinninmonth 1 Golden Jubilee National Hospital, Agamemnon Street, Clydebank G81 4DY, United Kingdom
a r t i c l e
i n f o
Article history: Received 24 November 2013 Received in revised form 7 November 2014 Accepted 10 November 2014 Keywords: TKR Alignment Deviation Cement
a b s t r a c t Background: Long-term survival of knee replacement depends on accurate alignment. Despite improvements in cut accuracy mal-alignment of 3° or more is still seen. All methods share common implantation techniques. This study examines the effect of implantation on overall limb alignment relating it to cut alignment and trial alignment. Methods: A retrospective review of navigated primary knee replacements was undertaken (n = 113). Overall coronal limb alignments for the aggregated cuts, trial and final implanted components were examined. Results: All 113 knees had coronal aggregated cut alignment within 2° of neutral (range: 2° varus to 2° valgus). With trial components 99 knees (88%) had an overall coronal limb alignment within 2° of neutral (range: 3° varus to 4° valgus). After final implantation 106 knees (94%) were within 2° of neutral (range: 4° varus to 4° valgus). Forty eight knees (42%) showed no alignment deviation occurring between trial and the final implanted prostheses and 16 knees (14%) shoed a deviation of 2° or more. There was a correlation of both aggregated cut (r = 0.284, p = 0.002) and trial (r = 0.794, p b 0.001) with final alignment. There was no significant difference between the final alignment and the aggregated cut alignment(mean difference = −0.15°, p = 0.254) or trial alignment (mean difference −0.13°, p = 0.155). Conclusions: Even when the aggregated alignment produced by the bone cuts is accurate, inaccuracy in final alignment can result from the implantation process. It may be productive for surgeons to concentrate on the implantation process to improve alignment and reduce outliers. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The long term survival of a knee replacement is known to depend on accurate alignment at the time of primary implantation [1–3]. Much attention has been focussed on technique to improve the accuracy of bony cut alignment with reference to the mechanical axis [4–6]. These include recent innovations such as refined manual instrumentation, computer navigation, robotics and shape matching. Despite these efforts malalignment of 3° or more is still seen to occur [4]. Although the methods guiding the bone cuts vary, all the techniques share a common implantation process. Deviation of component position at implantation was originally described by Sambatakakis et al. [7] in 1991. They noted on postoperative radiographs that the tibial component did not always sit parallel to the tibial cut, which they called the ‘cement wedge sign’. They found that this ‘cement wedge sign’ was strongly associated with the subsequent development of radio-lucent lines beneath the tibial component and
⁎ Corresponding author. Tel.: +44 141 951 5570. E-mail address: [email protected] (D.F. Howie). 1 Tel.: +44 141 951 5570.
http://dx.doi.org/10.1016/j.knee.2014.11.005 0968-0160/© 2014 Elsevier B.V. All rights reserved.
was thought to be due to residual ligament imbalance when reducing the knee with soft cement. More recently Catani et al. [8] used a computer navigation system to demonstrate individual component deviation of over a degree occurring at implantation. Catani et al. also observed an apparent tendency to correct alignment to neutral during implantation. This occurs when the final components are placed, the limb is in extension, and the cement is curing, the real time limb alignment figure can be observed and an appropriate varus/valgus force is applied to obtain zero degrees. This could compensate for possible cut inaccuracies, but may alter the cement mantle and could be detrimental to the final fixation of the implant. Overall limb alignment has been compared to cut alignment using long leg films which have their own limitations [9,10]. Overall limb alignment has also been compared to aggregated tibial and femoral cuts which have been used to calculate overall alignment [8]. There is no published literature where the specific effect of implantation on overall limb alignment and its relationship to both cut alignment and trial alignment have been observed using the same system. The aim of this study was to examine the effects of implantation on overall limb alignment using alignment recorded by a computer navigation system and to relate final alignment to both the aggregated cut alignment and the trial alignment. The null hypothesis was that there
48
D.F. Howie et al. / The Knee 22 (2015) 47–50
would be no significant difference between overall coronal limb alignment with the trial components and the overall coronal limb alignment following final cemented implantation. 2. Methods Formal ethical consent was deemed unnecessary by the local research ethics service and institutional clinical governance approval was obtained. A retrospective review of a single surgeon (AWGK) consecutive series of navigated primary knee replacements was undertaken using data collected prospectively during patient care from July 2006 to July 2011 (n = 235). Data was not saved centrally for analysis in 39 patients. Patients with incomplete intra-operative data were excluded, n = 83 (34 had incomplete cut data recorded, 32 had no trial alignment recorded, seven patients had no final alignment recorded and 10 patients had missing demographic data). Therefore 113 computer-assisted primary total knee replacements with dissimilar base plate designs, but similar femoral geometries were included, Scorpio (n = 18), Scorpio NRG (n = 75) and Triathlon (n = 19) all manufactured by Stryker. The patient demographics are displayed in Table 1. All surgeries were carried out using the image free eNlite navigation system (Stryker, Kalamazoo, Michigan, USA) through a medial parapatellar approach following the same surgical sequence; bicortical femoral and tibial trackers were placed, the navigation system was calibrated, and anatomical landmarks and joint centres were acquired in the usual manner. Using the navigation system the distal femoral and tibial cutting blocks were aligned at 90° to the overall mechanical axis in the coronal plane and cuts made. The remaining femoral cuts were then made using standard cutting blocks. The final distal femoral cut, posterior condylar cuts and tibial cut were recorded by the system using an instrumented probe. The final trial alignment, as balanced with the chosen insert thickness, was recorded throughout the knee range of motion. The senior author ensured that during both trial and final implant data recording there was no soft tissue interposed between trials/bone/implant. Care was also taken to ensure that the limb was in neutral rotation and visual confirmation was made that each of the femoral condyles were seated against the tibial articular surface. Manual impaction of the definitive components was undertaken separately using a single mix of cement, with care taken to seat both femoral and tibial components flush with the bone cuts, using the final polyethylene bearing, and putting the knee into extension whilst the cement cured. Final alignment of the knee throughout its range of motion was then recorded before closure. The overall coronal limb alignment data for both the trial components and the final implanted components at 0° flexion were examined. The orientations of the distal femoral and proximal tibial cuts, recorded by the navigation system, were used to determine an overall aggregated cut alignment, a method used by Catani [8].
For the purposes of mathematical analysis varus was defined as a negative angle and valgus as a positive angle. Analysis of relationships between parameters was performed using Pearson's correlation coefficient. Student's T test (paired two sample means) was used to look for significant differences between the data sets. 3. Results Following bone resection all 113 knees had coronal aggregated cut alignment in the range of 2° varus to 2° valgus (Fig. 1). When the trial components were placed overall coronal limb alignment ranged from 3° varus to 4° valgus (Fig. 2). After final implantation overall coronal limb alignment ranged from 4° varus to 4° valgus (Fig. 3). Statistically there was no significant difference found between the final alignment and the aggregated cut alignment (mean difference = −0.15°, p = 0.254) or trial alignment (mean difference = −0.13°, p = 0.155). The deviation in coronal plane limb alignment (towards and away from neutral) occurring at implantation is represented in Fig. 4. This shows the change in alignment occurring between trial prosthesis and the final implanted prosthesis. Forty eight knees (42%) showed no change at implantation whilst 16 (14%) showed overall limb alignment change of 2° or more occurring at implantation, 14 of these were improvements in alignment, and two were worsening alignment. The deviation in coronal plane limb alignment occurring between aggregated cut alignment and trial alignment is seen in Fig. 5. 32 (28%) and showed no change at implantation whilst 32 knees (28%) showed overall limb alignment change of 2° or more occurring at implantation. A weak correlation was seen between final alignment and aggregated cut alignment (Pearson's correlation r = 0.284, p = 0.002). A strong correlation was seen between final alignment and trial alignment (Pearson's correlation r = 0.794, p b 0.001). Analysing tibial and femoral cuts individually; the tibial cut had a stronger correlation than distal femoral cut to final limb alignment (r = 0.222, p = 0.18 vs r = 0.138, p = 0.16). A weak negative correlation (Pearson's correlation r = −0.270, p = 0.004) between the aggregated cut alignment and the alignment deviation occurring between the trial and final implantation was found — i.e. an improvement in alignment following implantation.
4. Discussion A number of studies have supported the use of navigation as a means to reduce outliers and improve the accuracy of bone resection [1,4–6]. The intra-operative data from this study showing final limb alignment obtained following implantation is in keeping with previously published work on the alignment achievable using navigation. The given accuracy of computer navigation systems is usually quoted at 1° and 1 mm, although some authors have reported higher accuracies than this [11]. We therefore took a difference of 2° or more between measurements as a real change in lower limb alignment as 1° could be measurement error. For the assessment as to whether or not overall limb alignment was acceptable we took the widely accepted limits of ±3° from neutral (0°).
Aggregated Cut Alignment 43 39
Data were analysed using IBM Statistics (IBM New York, USA), and graphs were produced in Excel 2007 (Microsoft, Washington, USA).
Number
2.1. Data analysis
18
Table 1 Demographic data of study cohort, mean (SD), is presented. Age (years) Gender (M/F) Body mass index (kg/m2) Pre-implant deformity a
67.7 (8.5) 43 to 86 49/64 31.7 (5.3) 18.3 to 45.7 2.4° varus (±5.7°, range 14° varus to 16° valgus)
a Patient's pre-implant deformity was recorded by the navigation system after initial approach to the knee was made, trackers were placed and landmarks were acquired with the limb in maximum extension.
10 0
0
-4
-3
3 -2
-1
0
1
2
0
0
3
4
Degrees Fig. 1. Histogram of aggregated (tibial and femoral) cut alignment, rounded to the nearest degree. (Positive values: valgus, negative values: varus).
D.F. Howie et al. / The Knee 22 (2015) 47–50
49
Trial Alignment
Deviation Trial - Final
35 48
22 Frequency
Number
21 13 9
8
32
17
4
-4
12
1
0 -3
-2
-1
0
1
2
3
4
Degrees
0
0
-4
-3
2 -2
-1
0
Fig. 2. Histogram of trial alignment, rounded to the nearest degree. (Positive values: valgus, negative values: varus).
Catani et al. [8] have demonstrated the effects of individual component deviation as a result of implantation technique but did not report overall final limb alignment, instead aggregating their tibial and femoral alignment figures. We used a navigation system to examine the differences between the overall limb alignment with the final implanted components, the trial components and the aggregated cut alignment. There is individual variation in limb alignment between the aggregated bony resections and the final implanted limb alignment, although these were not statistically significant. We were unable to reject our null hypothesis that the alignment with the trial components would be no different from the alignment with the final implanted components. This may be because the influence of implantation is insignificant, or it may be that the care taken to ensure accurate final insertion by individual component impaction rather than combined positioning by extending the knee whilst the cement was soft was effective. However it can be seen that the tibial and distal femoral cuts in this series had an acceptable level of accuracy (combined orientation all within 2° of neutral), but the implanted alignment varied more widely (4° varus to 4° valgus), demonstrating that cut accuracy in this series does not account for all of the variation observed in final limb alignment. The data in this study suggests that the outliers were predicted by the alignment of the trial and that both the trial and final implanted alignments are influenced by factor(s) extrinsic to the distal femoral and tibial cuts. This may be attributable to patient factors such as soft tissues — as originally suggested by Sambatakakis et al. [7], or surgeon factors such as inaccurate femoral chamfer cuts or high points/ridges
1
Deviation Cuts-Trial 32
22 Frequency
Number
27
17
10 6
1 -4
-3
3
-2
-1
0
4
on the cut tibial surface dictating the position of the implant and not allowing the component to seat accurately on the cut surface. The anterior and posterior femoral cuts could also influence the femoral component position, with divergence away from the central axis of the femur causing the femoral component not to seat on the distal femoral cut. This may be particularly relevant in sclerotic bone, and in surgeon's attempts to avoid femoral notching with the anterior cut. Unfortunately these cuts were not recorded, and we cannot comment on their influence in this study. Tibial rotation could account for some of the apparent deviation in coronal plane alignment, as the tibial slope rotates away from neutral it can influence coronal plane alignment. Rotation was set using the navigation system to replicate the vector recorded in the initial registration process. This uses an instrumented probe, placed after exposure with the tip in the centre of the tibial spines and the long axis of the probe is used to record rotation, as viewed from above, to reflect native anatomy, taking into account the shape of the tibial condyles and the position of the tibial tuberosity. Planned tibial cuts were for 3° of posterior slope (all in this series were within 2° of this plan) which according to Tsukeoka et al. [12] can account for up to 1.2 ° coronal
12
2
3
Fig. 4. Histogram of deviation between trial and implanted alignments, rounded to the nearest degree. Positive values are towards neutral (better alignment), negative values are away from neutral (worse alignment).
30
18
1
Degrees
Final Alignment 29
2
1
1
2
3
3
1 4
4
5 3 1
0 -5
-4
-3
-2
-1
0
1
2
3
4
5
Degrees
Degrees Fig. 3. Histogram of final implanted coronal alignment, rounded to the nearest degree. (Positive values: valgus, negative values: varus).
Fig. 5. Histogram of deviation between aggregated cuts and trial alignment, rounded to the nearest degree. Positive values are towards neutral (better alignment), and negative values are away from neutral (worse alignment).
50
D.F. Howie et al. / The Knee 22 (2015) 47–50
plane malalignment if the tibial component is malrotated 25° away from neutral. Despite malalignment occurring we did not see the ‘cement wedge sign’ (as described by Sambatakakis et al. [7]) in any postoperative radiographs in this series. This may be because the ‘wedge sign’ only appears with greater deviation errors than we have observed, the radiographs were taken at an inclination to obscure tibial cement mantle, or because the deviation we have observed in this series was not caused by uneven cement mantles. The weak negative correlation in our data between the aggregated cut alignment and the alignment deviation occurring at implantation (r =− 0.24) was seen in Catani's study to a greater degree (r = −0.497) who suggests it was because of implicit tendency of the surgeon to correct alignment. This would occur once the components have been implanted and the knee has been put into extension, as the cement is curing. At this stage the surgeon can use the navigation system to see and potentially improve, the overall alignment using the real time figure on the screen. Such attempts may result in better final alignment, but one has to assume that alterations to the cement mantle have occurred. Uneven cement pressurisation may adversely affect the penetration of cement into cancellous bone, which is known to increase tensile and shear stresses at the cement–bone interface [13–15]. These would pre-dispose to inadequate implant fixation and subsequent micromotion which may lead to early failure [7,14]. It has been the senior author's practice not to look at the navigation system during this phase, accepting the position of the implanted components whilst taking care to hold the leg neutral in extension and ensure the components are symmetrically pressurised. We therefore infer that if the weak negative correlation found between cut alignment and implantation deviation is to be believed, it is likely as a result of the knee and its ligaments inherently tending towards neutral alignment, in contrast to Sambatakakis's interpretation of their data. This study examined only coronal plane limb alignment at 0° of flexion, and can therefore not comment on deviation in other planes. The absence of recorded femoral anterior, posterior or chamfer cuts precludes any further comment on their influence. The aggregated cut alignment figure (as used by Catani et al. [8]) is conceptual, and derived from the recorded cut data, and perhaps a more representative way of assessing overall alignment obtained from the cuts would be to use a rectangular spacer block and record the alignment with the system. This study uses data from a single experienced high volume surgeon who is not on his learning curve, paying close attention to seating the implants correctly in each case. 5. Conclusions We found that although implantation did alter overall coronal limb alignment, it was not common in this series, and was not statistically
significant. Therefore on the basis of this series we were unable to reject our null hypothesis. This study suggests that the “extreme” cut accuracy may not have as significant an impact on final coronal alignment as other factors. The results of this study support previous work in this area, care needs to be taken during implantation in order that the accuracy of the cuts is translated into accurate overall limb alignment. More focus on implantation may help to improve alignment and minimise outliers.
References [1] Lutzner J, Krummenauer F, Wolf C, Gunther KP, Kirschner S. Computer-assisted and conventional total knee replacement: a comparative, prospective, randomised study with radiological and CT evaluation. J Bone Joint Surg 2008;90B:1039–44. [2] Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM. Why are total knee arthroplasties failing today? Clin Orthop 2002;404:7. [3] Fehring TK, Odum S, Griffin WL, Mason JB, Nadaud M. Early failures in total knee arthroplasty. Clin Orthop 2001;392:315. [4] Victor J, Hoste D. Image-based computer-assisted total knee arthroplasty leads to lower variability in coronal alignment. Clin Orthop Relat Res 2004;428:131–9. [5] Decking R, Markmann Y, Fuchs J, Puhl W, Scharf H-P. Leg axis after computernavigated total knee arthroplasty. A prospective randomised trial comparing computer-navigated and manual implantation. J Arthroplasty 2005;20:282–8. [6] Chin PL, Yang KY, Yeo SJ, Lo NN. Randomised control trial comparing radiographic total knee arthroplasty implant placement using computer navigation versus conventional technique. J Arthroplasty 2005;20:618–26. [7] Sambatakakis A, Wilton TJ, Newton G. Radiographic sign of persistent soft-tissue imbalance after knee replacement. J Bone Joint Surg Br Sep 1991;73(5):751–6. [8] Catani F, Biasca N, Ensini A, Leardini A, Bianchi L, Digennaro V, et al. Alignment deviation between bone resection and final implant positioning in computernavigated total knee arthroplasty. J Bone Joint Surg 2008;90A:765–71. [9] Yaffe MA, Koo SS, Stulberg SD. Radiographic and navigation measurements of TKA limb alignment do not correlate. Clin Orthop Relat Res 2008;466:2736–44. http:// dx.doi.org/10.1007/s11999-008-0427-9. [10] Hauschild O, Konstantinidis L, Baumann T, Niemeyer P, Suedkamp NP, Helwig P. Correlation of radiographic and navigated measurements of TKA limb alignment: a matter of time? Knee Surg Sports Traumatol Arthrosc Oct 2010;18(10):1317–22 [Epub 2010 Apr 21]. [11] Lustig S, Fleury C, Goy D, Neyret P, Donell ST. The accuracy of acquisition of an imageless computer-assisted system and its implication for knee arthroplasty. Knee Jan 2011;18(1):15–20. [12] Tsukeoka T, Tsuneizumi Y, Lee TH. The effect of the posterior slope of the tibial plateau osteotomy with a rotational error on tibial component malalignment in total knee replacement. Bone Joint J September 2013;95-B:1201–3. [13] Miskovsky C, Whiteside LA, White SE. The cemented unicondylar knee arthroplasty. An in-vitro comparison of three cement techniques. Clin Orthop Relat Res 1992;215. [14] Ritter MA, Herbst SA, Keating EM, Faris PM. Radiolucency at the bone–cement interface in total knee replacement — the effects of bone surface preparation and cement technique. J Bone Joint Surg 1994;76A:60–5. [15] Bauze AJ, Costi JJ, Stavrou P, Rankin WA, Hearn TC, Krishnan J, et al. Cement penetration and stiffness of the cement–bone composite in the proximal tibia in a porcine model. J Orthop Surg 2004;12(2):194–8.
Contents lists available at ScienceDirect
The Knee
Intra-operative deviation in limb alignment occurring at implantation in total knee arthroplasty D.F. Howie ⁎, G.J. Love 1, A.H. Deakin 1, A.W.G. Kinninmonth 1 Golden Jubilee National Hospital, Agamemnon Street, Clydebank G81 4DY, United Kingdom
a r t i c l e
i n f o
Article history: Received 24 November 2013 Received in revised form 7 November 2014 Accepted 10 November 2014 Keywords: TKR Alignment Deviation Cement
a b s t r a c t Background: Long-term survival of knee replacement depends on accurate alignment. Despite improvements in cut accuracy mal-alignment of 3° or more is still seen. All methods share common implantation techniques. This study examines the effect of implantation on overall limb alignment relating it to cut alignment and trial alignment. Methods: A retrospective review of navigated primary knee replacements was undertaken (n = 113). Overall coronal limb alignments for the aggregated cuts, trial and final implanted components were examined. Results: All 113 knees had coronal aggregated cut alignment within 2° of neutral (range: 2° varus to 2° valgus). With trial components 99 knees (88%) had an overall coronal limb alignment within 2° of neutral (range: 3° varus to 4° valgus). After final implantation 106 knees (94%) were within 2° of neutral (range: 4° varus to 4° valgus). Forty eight knees (42%) showed no alignment deviation occurring between trial and the final implanted prostheses and 16 knees (14%) shoed a deviation of 2° or more. There was a correlation of both aggregated cut (r = 0.284, p = 0.002) and trial (r = 0.794, p b 0.001) with final alignment. There was no significant difference between the final alignment and the aggregated cut alignment(mean difference = −0.15°, p = 0.254) or trial alignment (mean difference −0.13°, p = 0.155). Conclusions: Even when the aggregated alignment produced by the bone cuts is accurate, inaccuracy in final alignment can result from the implantation process. It may be productive for surgeons to concentrate on the implantation process to improve alignment and reduce outliers. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The long term survival of a knee replacement is known to depend on accurate alignment at the time of primary implantation [1–3]. Much attention has been focussed on technique to improve the accuracy of bony cut alignment with reference to the mechanical axis [4–6]. These include recent innovations such as refined manual instrumentation, computer navigation, robotics and shape matching. Despite these efforts malalignment of 3° or more is still seen to occur [4]. Although the methods guiding the bone cuts vary, all the techniques share a common implantation process. Deviation of component position at implantation was originally described by Sambatakakis et al. [7] in 1991. They noted on postoperative radiographs that the tibial component did not always sit parallel to the tibial cut, which they called the ‘cement wedge sign’. They found that this ‘cement wedge sign’ was strongly associated with the subsequent development of radio-lucent lines beneath the tibial component and
⁎ Corresponding author. Tel.: +44 141 951 5570. E-mail address: [email protected] (D.F. Howie). 1 Tel.: +44 141 951 5570.
http://dx.doi.org/10.1016/j.knee.2014.11.005 0968-0160/© 2014 Elsevier B.V. All rights reserved.
was thought to be due to residual ligament imbalance when reducing the knee with soft cement. More recently Catani et al. [8] used a computer navigation system to demonstrate individual component deviation of over a degree occurring at implantation. Catani et al. also observed an apparent tendency to correct alignment to neutral during implantation. This occurs when the final components are placed, the limb is in extension, and the cement is curing, the real time limb alignment figure can be observed and an appropriate varus/valgus force is applied to obtain zero degrees. This could compensate for possible cut inaccuracies, but may alter the cement mantle and could be detrimental to the final fixation of the implant. Overall limb alignment has been compared to cut alignment using long leg films which have their own limitations [9,10]. Overall limb alignment has also been compared to aggregated tibial and femoral cuts which have been used to calculate overall alignment [8]. There is no published literature where the specific effect of implantation on overall limb alignment and its relationship to both cut alignment and trial alignment have been observed using the same system. The aim of this study was to examine the effects of implantation on overall limb alignment using alignment recorded by a computer navigation system and to relate final alignment to both the aggregated cut alignment and the trial alignment. The null hypothesis was that there
48
D.F. Howie et al. / The Knee 22 (2015) 47–50
would be no significant difference between overall coronal limb alignment with the trial components and the overall coronal limb alignment following final cemented implantation. 2. Methods Formal ethical consent was deemed unnecessary by the local research ethics service and institutional clinical governance approval was obtained. A retrospective review of a single surgeon (AWGK) consecutive series of navigated primary knee replacements was undertaken using data collected prospectively during patient care from July 2006 to July 2011 (n = 235). Data was not saved centrally for analysis in 39 patients. Patients with incomplete intra-operative data were excluded, n = 83 (34 had incomplete cut data recorded, 32 had no trial alignment recorded, seven patients had no final alignment recorded and 10 patients had missing demographic data). Therefore 113 computer-assisted primary total knee replacements with dissimilar base plate designs, but similar femoral geometries were included, Scorpio (n = 18), Scorpio NRG (n = 75) and Triathlon (n = 19) all manufactured by Stryker. The patient demographics are displayed in Table 1. All surgeries were carried out using the image free eNlite navigation system (Stryker, Kalamazoo, Michigan, USA) through a medial parapatellar approach following the same surgical sequence; bicortical femoral and tibial trackers were placed, the navigation system was calibrated, and anatomical landmarks and joint centres were acquired in the usual manner. Using the navigation system the distal femoral and tibial cutting blocks were aligned at 90° to the overall mechanical axis in the coronal plane and cuts made. The remaining femoral cuts were then made using standard cutting blocks. The final distal femoral cut, posterior condylar cuts and tibial cut were recorded by the system using an instrumented probe. The final trial alignment, as balanced with the chosen insert thickness, was recorded throughout the knee range of motion. The senior author ensured that during both trial and final implant data recording there was no soft tissue interposed between trials/bone/implant. Care was also taken to ensure that the limb was in neutral rotation and visual confirmation was made that each of the femoral condyles were seated against the tibial articular surface. Manual impaction of the definitive components was undertaken separately using a single mix of cement, with care taken to seat both femoral and tibial components flush with the bone cuts, using the final polyethylene bearing, and putting the knee into extension whilst the cement cured. Final alignment of the knee throughout its range of motion was then recorded before closure. The overall coronal limb alignment data for both the trial components and the final implanted components at 0° flexion were examined. The orientations of the distal femoral and proximal tibial cuts, recorded by the navigation system, were used to determine an overall aggregated cut alignment, a method used by Catani [8].
For the purposes of mathematical analysis varus was defined as a negative angle and valgus as a positive angle. Analysis of relationships between parameters was performed using Pearson's correlation coefficient. Student's T test (paired two sample means) was used to look for significant differences between the data sets. 3. Results Following bone resection all 113 knees had coronal aggregated cut alignment in the range of 2° varus to 2° valgus (Fig. 1). When the trial components were placed overall coronal limb alignment ranged from 3° varus to 4° valgus (Fig. 2). After final implantation overall coronal limb alignment ranged from 4° varus to 4° valgus (Fig. 3). Statistically there was no significant difference found between the final alignment and the aggregated cut alignment (mean difference = −0.15°, p = 0.254) or trial alignment (mean difference = −0.13°, p = 0.155). The deviation in coronal plane limb alignment (towards and away from neutral) occurring at implantation is represented in Fig. 4. This shows the change in alignment occurring between trial prosthesis and the final implanted prosthesis. Forty eight knees (42%) showed no change at implantation whilst 16 (14%) showed overall limb alignment change of 2° or more occurring at implantation, 14 of these were improvements in alignment, and two were worsening alignment. The deviation in coronal plane limb alignment occurring between aggregated cut alignment and trial alignment is seen in Fig. 5. 32 (28%) and showed no change at implantation whilst 32 knees (28%) showed overall limb alignment change of 2° or more occurring at implantation. A weak correlation was seen between final alignment and aggregated cut alignment (Pearson's correlation r = 0.284, p = 0.002). A strong correlation was seen between final alignment and trial alignment (Pearson's correlation r = 0.794, p b 0.001). Analysing tibial and femoral cuts individually; the tibial cut had a stronger correlation than distal femoral cut to final limb alignment (r = 0.222, p = 0.18 vs r = 0.138, p = 0.16). A weak negative correlation (Pearson's correlation r = −0.270, p = 0.004) between the aggregated cut alignment and the alignment deviation occurring between the trial and final implantation was found — i.e. an improvement in alignment following implantation.
4. Discussion A number of studies have supported the use of navigation as a means to reduce outliers and improve the accuracy of bone resection [1,4–6]. The intra-operative data from this study showing final limb alignment obtained following implantation is in keeping with previously published work on the alignment achievable using navigation. The given accuracy of computer navigation systems is usually quoted at 1° and 1 mm, although some authors have reported higher accuracies than this [11]. We therefore took a difference of 2° or more between measurements as a real change in lower limb alignment as 1° could be measurement error. For the assessment as to whether or not overall limb alignment was acceptable we took the widely accepted limits of ±3° from neutral (0°).
Aggregated Cut Alignment 43 39
Data were analysed using IBM Statistics (IBM New York, USA), and graphs were produced in Excel 2007 (Microsoft, Washington, USA).
Number
2.1. Data analysis
18
Table 1 Demographic data of study cohort, mean (SD), is presented. Age (years) Gender (M/F) Body mass index (kg/m2) Pre-implant deformity a
67.7 (8.5) 43 to 86 49/64 31.7 (5.3) 18.3 to 45.7 2.4° varus (±5.7°, range 14° varus to 16° valgus)
a Patient's pre-implant deformity was recorded by the navigation system after initial approach to the knee was made, trackers were placed and landmarks were acquired with the limb in maximum extension.
10 0
0
-4
-3
3 -2
-1
0
1
2
0
0
3
4
Degrees Fig. 1. Histogram of aggregated (tibial and femoral) cut alignment, rounded to the nearest degree. (Positive values: valgus, negative values: varus).
D.F. Howie et al. / The Knee 22 (2015) 47–50
49
Trial Alignment
Deviation Trial - Final
35 48
22 Frequency
Number
21 13 9
8
32
17
4
-4
12
1
0 -3
-2
-1
0
1
2
3
4
Degrees
0
0
-4
-3
2 -2
-1
0
Fig. 2. Histogram of trial alignment, rounded to the nearest degree. (Positive values: valgus, negative values: varus).
Catani et al. [8] have demonstrated the effects of individual component deviation as a result of implantation technique but did not report overall final limb alignment, instead aggregating their tibial and femoral alignment figures. We used a navigation system to examine the differences between the overall limb alignment with the final implanted components, the trial components and the aggregated cut alignment. There is individual variation in limb alignment between the aggregated bony resections and the final implanted limb alignment, although these were not statistically significant. We were unable to reject our null hypothesis that the alignment with the trial components would be no different from the alignment with the final implanted components. This may be because the influence of implantation is insignificant, or it may be that the care taken to ensure accurate final insertion by individual component impaction rather than combined positioning by extending the knee whilst the cement was soft was effective. However it can be seen that the tibial and distal femoral cuts in this series had an acceptable level of accuracy (combined orientation all within 2° of neutral), but the implanted alignment varied more widely (4° varus to 4° valgus), demonstrating that cut accuracy in this series does not account for all of the variation observed in final limb alignment. The data in this study suggests that the outliers were predicted by the alignment of the trial and that both the trial and final implanted alignments are influenced by factor(s) extrinsic to the distal femoral and tibial cuts. This may be attributable to patient factors such as soft tissues — as originally suggested by Sambatakakis et al. [7], or surgeon factors such as inaccurate femoral chamfer cuts or high points/ridges
1
Deviation Cuts-Trial 32
22 Frequency
Number
27
17
10 6
1 -4
-3
3
-2
-1
0
4
on the cut tibial surface dictating the position of the implant and not allowing the component to seat accurately on the cut surface. The anterior and posterior femoral cuts could also influence the femoral component position, with divergence away from the central axis of the femur causing the femoral component not to seat on the distal femoral cut. This may be particularly relevant in sclerotic bone, and in surgeon's attempts to avoid femoral notching with the anterior cut. Unfortunately these cuts were not recorded, and we cannot comment on their influence in this study. Tibial rotation could account for some of the apparent deviation in coronal plane alignment, as the tibial slope rotates away from neutral it can influence coronal plane alignment. Rotation was set using the navigation system to replicate the vector recorded in the initial registration process. This uses an instrumented probe, placed after exposure with the tip in the centre of the tibial spines and the long axis of the probe is used to record rotation, as viewed from above, to reflect native anatomy, taking into account the shape of the tibial condyles and the position of the tibial tuberosity. Planned tibial cuts were for 3° of posterior slope (all in this series were within 2° of this plan) which according to Tsukeoka et al. [12] can account for up to 1.2 ° coronal
12
2
3
Fig. 4. Histogram of deviation between trial and implanted alignments, rounded to the nearest degree. Positive values are towards neutral (better alignment), negative values are away from neutral (worse alignment).
30
18
1
Degrees
Final Alignment 29
2
1
1
2
3
3
1 4
4
5 3 1
0 -5
-4
-3
-2
-1
0
1
2
3
4
5
Degrees
Degrees Fig. 3. Histogram of final implanted coronal alignment, rounded to the nearest degree. (Positive values: valgus, negative values: varus).
Fig. 5. Histogram of deviation between aggregated cuts and trial alignment, rounded to the nearest degree. Positive values are towards neutral (better alignment), and negative values are away from neutral (worse alignment).
50
D.F. Howie et al. / The Knee 22 (2015) 47–50
plane malalignment if the tibial component is malrotated 25° away from neutral. Despite malalignment occurring we did not see the ‘cement wedge sign’ (as described by Sambatakakis et al. [7]) in any postoperative radiographs in this series. This may be because the ‘wedge sign’ only appears with greater deviation errors than we have observed, the radiographs were taken at an inclination to obscure tibial cement mantle, or because the deviation we have observed in this series was not caused by uneven cement mantles. The weak negative correlation in our data between the aggregated cut alignment and the alignment deviation occurring at implantation (r =− 0.24) was seen in Catani's study to a greater degree (r = −0.497) who suggests it was because of implicit tendency of the surgeon to correct alignment. This would occur once the components have been implanted and the knee has been put into extension, as the cement is curing. At this stage the surgeon can use the navigation system to see and potentially improve, the overall alignment using the real time figure on the screen. Such attempts may result in better final alignment, but one has to assume that alterations to the cement mantle have occurred. Uneven cement pressurisation may adversely affect the penetration of cement into cancellous bone, which is known to increase tensile and shear stresses at the cement–bone interface [13–15]. These would pre-dispose to inadequate implant fixation and subsequent micromotion which may lead to early failure [7,14]. It has been the senior author's practice not to look at the navigation system during this phase, accepting the position of the implanted components whilst taking care to hold the leg neutral in extension and ensure the components are symmetrically pressurised. We therefore infer that if the weak negative correlation found between cut alignment and implantation deviation is to be believed, it is likely as a result of the knee and its ligaments inherently tending towards neutral alignment, in contrast to Sambatakakis's interpretation of their data. This study examined only coronal plane limb alignment at 0° of flexion, and can therefore not comment on deviation in other planes. The absence of recorded femoral anterior, posterior or chamfer cuts precludes any further comment on their influence. The aggregated cut alignment figure (as used by Catani et al. [8]) is conceptual, and derived from the recorded cut data, and perhaps a more representative way of assessing overall alignment obtained from the cuts would be to use a rectangular spacer block and record the alignment with the system. This study uses data from a single experienced high volume surgeon who is not on his learning curve, paying close attention to seating the implants correctly in each case. 5. Conclusions We found that although implantation did alter overall coronal limb alignment, it was not common in this series, and was not statistically
significant. Therefore on the basis of this series we were unable to reject our null hypothesis. This study suggests that the “extreme” cut accuracy may not have as significant an impact on final coronal alignment as other factors. The results of this study support previous work in this area, care needs to be taken during implantation in order that the accuracy of the cuts is translated into accurate overall limb alignment. More focus on implantation may help to improve alignment and minimise outliers.
References [1] Lutzner J, Krummenauer F, Wolf C, Gunther KP, Kirschner S. Computer-assisted and conventional total knee replacement: a comparative, prospective, randomised study with radiological and CT evaluation. J Bone Joint Surg 2008;90B:1039–44. [2] Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM. Why are total knee arthroplasties failing today? Clin Orthop 2002;404:7. [3] Fehring TK, Odum S, Griffin WL, Mason JB, Nadaud M. Early failures in total knee arthroplasty. Clin Orthop 2001;392:315. [4] Victor J, Hoste D. Image-based computer-assisted total knee arthroplasty leads to lower variability in coronal alignment. Clin Orthop Relat Res 2004;428:131–9. [5] Decking R, Markmann Y, Fuchs J, Puhl W, Scharf H-P. Leg axis after computernavigated total knee arthroplasty. A prospective randomised trial comparing computer-navigated and manual implantation. J Arthroplasty 2005;20:282–8. [6] Chin PL, Yang KY, Yeo SJ, Lo NN. Randomised control trial comparing radiographic total knee arthroplasty implant placement using computer navigation versus conventional technique. J Arthroplasty 2005;20:618–26. [7] Sambatakakis A, Wilton TJ, Newton G. Radiographic sign of persistent soft-tissue imbalance after knee replacement. J Bone Joint Surg Br Sep 1991;73(5):751–6. [8] Catani F, Biasca N, Ensini A, Leardini A, Bianchi L, Digennaro V, et al. Alignment deviation between bone resection and final implant positioning in computernavigated total knee arthroplasty. J Bone Joint Surg 2008;90A:765–71. [9] Yaffe MA, Koo SS, Stulberg SD. Radiographic and navigation measurements of TKA limb alignment do not correlate. Clin Orthop Relat Res 2008;466:2736–44. http:// dx.doi.org/10.1007/s11999-008-0427-9. [10] Hauschild O, Konstantinidis L, Baumann T, Niemeyer P, Suedkamp NP, Helwig P. Correlation of radiographic and navigated measurements of TKA limb alignment: a matter of time? Knee Surg Sports Traumatol Arthrosc Oct 2010;18(10):1317–22 [Epub 2010 Apr 21]. [11] Lustig S, Fleury C, Goy D, Neyret P, Donell ST. The accuracy of acquisition of an imageless computer-assisted system and its implication for knee arthroplasty. Knee Jan 2011;18(1):15–20. [12] Tsukeoka T, Tsuneizumi Y, Lee TH. The effect of the posterior slope of the tibial plateau osteotomy with a rotational error on tibial component malalignment in total knee replacement. Bone Joint J September 2013;95-B:1201–3. [13] Miskovsky C, Whiteside LA, White SE. The cemented unicondylar knee arthroplasty. An in-vitro comparison of three cement techniques. Clin Orthop Relat Res 1992;215. [14] Ritter MA, Herbst SA, Keating EM, Faris PM. Radiolucency at the bone–cement interface in total knee replacement — the effects of bone surface preparation and cement technique. J Bone Joint Surg 1994;76A:60–5. [15] Bauze AJ, Costi JJ, Stavrou P, Rankin WA, Hearn TC, Krishnan J, et al. Cement penetration and stiffness of the cement–bone composite in the proximal tibia in a porcine model. J Orthop Surg 2004;12(2):194–8.
