The success of total knee arthroplasty (TKA) depends, in part, on proper alignment of the prosthetic components.5 It has been suggested that the most common cause of revision TKA is error in surgical technique28 because small changes in component positioning can lead to substantial changes in postoperative performance. Alignment errors greater than 3° in the frontal plane are associated with component loosening.15 Rotation of the femoral component of 5° from the transepicondylar axis alters tibiofem- oral kinematics19 and increases shear forces on the patellar component.3,19,23 Small amounts of combined femoral and tibial component internal rotation (1°-4°) have been associated with lateral tracking and tilting of the patella, whereas larger amounts of internal rotation (7°-17°) have been associated with patellar dislocation and patellar pros- thesis failure.4
Several reference techniques for establishing tibial rotational alignment have been proposed, such as the posterior condylar line of the tibia plateau,20 the transcondylar line of the tibia,31 the medial ⅓of the tibial tubercle,14 or the medial border of tubercle.18 If landmarks on the anterior tibia are used to establish alignment, it is necessary to define a posterior endpoint of the alignment axis. Akagi et al2 proposed an alignment axis from the medial border of the patellar tendon attachment to the middle of the posterior cruciate ligament (PCL) attachment point on the tibia. Dalury9 proposed a line drawn 1 mm medial to the medial border to the tibial tubercle and going through the midsulcus of the tibial spines. Other techniques frequently used to establish tibial rotational alignment are the transmalleolar axis, the line of the tibial anterior crest, and the axis of the second metatarsal. Although the femoral transepicondylar axis has been shown to approximate the flexion- extension axis of the knee,8 no such alignment axis intrinsic to the tibia has been shown to correspond to knee motion or to optimize postoperative functional outcome.
Computer-assisted surgical navigation systems10,28 have been developed to align implants more accurately than traditional mechanical guides. Although surgeons using these systems have reported more accurate alignment of implants in the frontal plane,13,16,22,27 it remains controversial whether a navigation system can reduce the variability of femoral rotational alignment.26,27 To our knowledge, only one study7 has examined whether navigation can improve tibial rotational alignment.
In this study, we characterized the variability associated with tibial rotational alignment techniques and determined whether using a computer-assisted surgical navigation system that establishes alignment axes by identifying landmarks on the tibia reduces this variability.
MATERIALS AND METHODS
We assessed the variability of four computer-assisted alignment techniques and one conventional (non-computer-assisted) technique to establish tibial rotational alignment axes in a series of cadaver specimens. Based on conservative estimates from previous studies,1,2,11,26,29 we assumed that each technique would have a standard deviation of 7°. With that assumption, a total of 110 measurements per technique would be required to achieve an a priori statistical power of 0.83 to detect a 3.5° difference between techniques.
Ten embalmed cadaver lower extremities, containing all structures distal to the femoral head, were mounted on wooden platforms with the knee in approximately 90° of flexion. The capsule of each knee was opened, exposing the distal femur and proximal tibia for total knee arthroplasty. The patellar tendon was resected proximal to the tibial tubercle to facilitate exposure. A member of the research team cut the distal femur and proximal tibia to simulate the initial steps of a typical TKA.
Eleven orthopaedic surgeons (nine attending surgeons, eight of whom have a practice specializing in total joint arthroplasty, and two total joint arthroplasty fellows) participated in this study. Each surgeon used four computer-assisted techniques and one traditional technique to establish rotational alignment axes on each specimen. All alignment axes were established on one specimen before proceeding to the next one, and all techniques were implemented in the same order on each specimen. For the computer-assisted techniques, the surgeons identified the following landmarks on the tibia with an optically tracked stylus from a navigation system: the most medial border of the tibial plateau, the most lateral border of the tibial plateau, the center of the attachment point of the PCL on the tibia, the medial ⅓ of the tibial tubercle, the medial border of the tibial tubercle, and the projected endpoint of the tibial anterior crest onto the tibial plateau. These landmarks served as the endpoints for four rotational alignment axes: the line between the most medial and most lateral points of the plateau (technique referred to as medial-lateral plateau), the line between the medial ⅓ of the tubercle and the PCL attachment (medial ⅓-PCL), the line between the medial border of the tubercle and the PCL (medial border-PCL), and the line between the projection of the anterior crest and the PCL (anterior crest-PCL). Total knee arthroplasty instrumentation (ie, tibial base plate, extramedullary drop rod, etc) was available to help the surgeons establish the alignment axis with the traditional technique. When using TKA instrumentation, the surgeons were not asked to use the instrumentation to align the tibial tray to any particular landmark, but were encouraged to align the tray in such a way that, in their opinion, would result in the best postoperative outcome. After the tibial tray was set in this position, the surgeons used the optically tracked stylus to identify two points along the anterior edge of the base plate to record its orientation.
The alignment error was defined as the angle between the axes established by the surgeons and a reference axis on each specimen (Fig 1). After all surgeons participated in the study, a member of the research team used the optical stylus to digitize the most medial and the lateral aspects of the malleoli and 25 points along the medial and lateral edges of the tibial plateau on each specimen. The points along the plateau were used to find the most medial and lateral points on the tibial plateau and establish a reference transverse axis.31 We formed a superior- inferior axis between the midpoint of the calculated medial and lateral points of the plateau and the center of the ankle, the midpoint of the most medial and lateral aspects of the malleoli.24 The cross product of the superior-inferior axis with the trans- verse axis formed a reference anteroposterior axis. Because two of the axes identified by the surgeons were primarily orientated in the mediolateral direction and three axes were directed primarily in the anteroposterior direction, we used different references axes for different techniques to avoid offsets of approximately 90° with some techniques. We defined the error associated with the medial-lateral plateau and TKA instrumentation techniques as the angle between the axes established by the surgeon and the transverse reference axis. For the remaining three techniques, the error was defined as the angle between the surgeon-defined axis and the anteroposterior reference axis.
We assessed the precision of each method by investigating its standard deviation. We used the Bartlett test to evaluate homogeneity of variance among the methods and to detect significant differences in the standard deviations between methods. After identifying unequal variances of the techniques, we used the Kruskal-Wallis test to assess accuracy and identify statistically significant differences in the mean errors of the methods, and the Tukey-Kramer method was used to further investigate significant results. The level of statistical significance was set at α = 0.05.
There was high variability in the tibial rotational alignment associated with all techniques. We found 13.1% (72 of 550) of the axes identified in this study were rotated less than 5° from the reference axis. The rotational alignment with respect to the reference axis ranged from 44° of internal rotation (with the medial border-PCL technique) to 46° of external rotation (with medial 1/3-PCL) (Fig 2). The two techniques that referenced the tubercle had the largest standard deviations of 27.4° and 28.1° and were significantly less precise (p < 0.001) than the two other computer-assisted techniques (medial-lateral plateau and anterior crest-PCL) and the traditional technique (TKA instrumentation), which had standard deviations of 9.9°, 10.8°, and 12.1°, respectively.
The medial border-PCL technique produced greater (p < 0.001) errors than the other four techniques (14.4° ± 27.4° internal rotation versus a mean error of less than 1° from the reference axis, respectively) (Fig 3).
We assessed the variability of techniques that establish tibial rotational alignment and determined whether a navigation system reduces this variability. Although the mean errors were less than 1° for four of the five techniques, we found that all of the techniques in this study were highly variable.
We note several limitations. Several factors could explain the high variability of tibial rotational alignment axes in the current study. Previous work in rotational alignment has used computed tomography (CT) scans,1,2,29 and the high variability observed in the current study suggests that establishing alignment axes may be more difficult from a small number of points on physical specimens than from CT images. The errors in the current study also are larger and more variable than what has been reported in single- operator studies with cadavers11 or patients with TKA,9 so it is possible that the range of errors from eleven surgeons is larger than for a single surgeon. In the current study, the range of errors for any one surgeon was smaller than the range of errors for all 11 surgeons. In some instances, the range of errors for one surgeon with a particular technique was less than half of the range of errors for that technique across all surgeons.
The lack of the femoral component removed certain visual cues that surgeons may use in surgery to test component alignment. If the femoral component were present, the surgeons may have been able to check for potential rotational alignment mismatch between the femoral and tibial components, which may have reduced the large range of observed alignment errors. Additionally, we were unable to test tibial alignment techniques in which the rotation of the tibial component is set relative to a fixed femoral component. For example, Dalury9 and Eckhoff et al11 have described a technique in which the orientation of the tibial tray was determined by allowing it to float into position with respect to the femoral component while the knee was placed through a full arc of motion. However, if the tibial component from a fixed-bearing system is aligned with respect to a fixed femoral component, the surgeon must decide whether to align the components when the knee is in flexion or extension. Although the prosthesis is subject to large forces and torques when the knee is near full extension during gait,12 Nagura et al21 have shown that high forces also are present in deep flexion. The decision as to whether to align a fixed tibial component with respect to the femoral component in flexion or extension warrants future investigation.
The surgeons were not experienced with the navigation system. Each surgeon was given an orientation using the system before the start of the experiment, but only one surgeon had extensive experience with any navigation system. A learning curve is present with navigation systems,6,28 and the surgeons may have performed differently if they had more experience with the equipment. Additionally, the axes were established in the same order on each specimen, and this lack of randomization may have influenced the results. However, we could not identify a trend in the alignment errors as the surgeons progressed through the techniques on the specimens, therefore we do not feel that our results were influenced by either a learning curve or surgeon fatigue throughout the experiment. Additionally, we could not determine any significant difference between the fellows' and attending surgeons' performances, nor could we detect differences among the attending surgeons based on experience.
The range of errors in the current study is greater than in previous studies. Eckhoff et al11 reported that using the tubercle resulted in tibial rotational alignment that was 19° ± 3° external to a femoral component. In a study using CT scans, Uehara et al29 reported that an axis using the medial ⅓ of the tubercle was rotated 2.6° ± 5.4° external to a projected femoral epicondylar axis, with a range of errors from 16° external to 10° internal. Akagi et al1 reported that their anteroposterior axis was more reliably identified on CT images than the tibial transmalleolar axis or an alignment axis that references the second metatarsal. One of the techniques in our study (medial border-PCL) is similar to this proposed anteroposterior axis, but we found that me- dial border-PCL technique was more variable than what was reported by Akagi et al.1,2
Few studies have investigated the clinical significance of the variability of tibial rotational alignment. Berger et al4 found that suboptimal postoperative functional outcomes, such as patellar maltracking or dislocation, were associated with combined femoral and tibial component internal rotation but not with external rotation. Retrieval studies have documented that different patterns of implant wear occur with internal tibial rotation than with external rotation,17,30 and Dalury9 reported that internal or external tibial rotation could lead to impingement on the polyethylene, which could lead to loss of motion and possibly wear. The clinical relevance of tibial rotational alignment is related to the design of the prosthesis. Errors in tibial rotational alignment are more critical in constrained articulations than in posterior cruciate ligament retaining and posterior-stabilized designs that allow for greater amounts of internal and external rotation; errors in tibial rotational alignment are even less critical for rotating platform designs. Ultimately, how much variability of tibial alignment that can be tolerated is unknown, and future studies relating rotational alignment to complications, failures, and revision for different component designs are needed to address this question.
As navigation systems become more widely used, it is important to evaluate all aspects of their performance. The results of the current study suggest that a navigation system that relies on digitization of landmarks to establish a rotational alignment axis does not provide a more reliable means of rotational alignment than using traditional TKA instrumentation. When the tibial tubercle is referenced by the navigation system, the resultant alignment axes are significantly less reliable than traditional instrumentation. These results contrast the demonstrated ability of navigation systems to improve alignment in the frontal plane.13,16,22,27 In the frontal plane, the landmarks that serve as the endpoints of alignment axes (eg, the center of the ankle) may have an error of up to 6 mm and correspond to an alignment error of approximately 1°.25 However, in the transverse plane, an anteroposterior error of 6 mm corresponds to an error of approximately 4.5°. The presence of osteophytes on the periphery of the knee during a TKA may distort the normal anatomy and the relatively subjective nature of landmarks that are used to determine rotational alignment (eg, the medial ⅓ of the tibial tubercle) also may contribute to the greater variability in the transverse plane.
The findings of the present study are not meant to discourage the use of navigation systems. The surgeons involved in the study are using navigation as a research tool in the operating room and to verify frontal and sagittal plane alignment. They continue to guide all of the cuts with standard instrumentation but are working to develop new techniques that will make tibial rotational alignment more accurate and precise. We hope the results of our study will illustrate the challenges in establishing component rotational alignment and emphasize the need for the development of the next generation of navigation technology that leads to a reduction in alignment outliers in the frontal and transverse planes.
We thank the surgeons who participated in this study.
1. Akagi M, Mori S, Nishimura S, Nishimura A, Asano T, Hamanishi C. Variability of extraarticular tibial rotation references for total knee arthroplasty. Clin Orthop Relat Res
2. Akagi M, Oh M, Nonaka T, Tsujimoto H, Asano T, Hamanishi C. An anteroposterior axis of the tibia for total knee arthroplasty. Clin Orthop Relat Res
3. Anouchi YS, Whiteside LA, Kaiser AD, Milliano MT. The effects of axial rotational alignment of the femoral component on knee stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens. Clin Orthop Relat Res
. 1993;287: 170-177.
4. Berger RA, Crossett LS, Jacobs JJ, Rubash HE. Malrotation causing patellofemoral complications after total knee arthroplasty. Clin Orthop Relat Res
5. Berger RA, Rubash HE, Seel MJ, Thompson WH, Crossett LS. Determining the rotational alignment of the femoral component in total knee arthroplasty using the epicondylar axis. Clin Orthop Relat Res
6. Bolognesi M, Hofmann A. Computer navigation versus standard instrumentation for TKA: a single-surgeon experience. Clin Orthop Relat Res
7. Chauhan SK, Scott RG, Breidahl W, Beaver RJ. Computer-assisted knee arthroplasty versus a conventional jig-based technique. A randomised, prospective trial. J Bone Joint Surg Br
8. Churchill DL, Incavo SJ, Johnson CC, Beynnon BD. The transepicondylar axis approximates the optimal flexion axis of the knee. Clin Orthop Relat Res
9. Dalury DF. Observations of the proximal tibia in total knee arthroplasty. Clin Orthop Relat Res
10. Delp SL, Stulberg SD, Davies B, Picard F, Leitner F. Computer assisted knee replacement. Clin Orthop Relat Res
11. Eckhoff DG, Metzger RG, Vandewalle MV. Malrotation associated with implant alignment technique in total knee arthroplasty. Clin Orthop Relat Res
12. Godest AC, Beaugonin M, Haug E, Taylor M, Gregson PJ. Simulation of a knee joint replacement during a gait cycle using explicit finite element analysis. J Biomech
13. Hart R, Janecek M, Chaker A, Bucek P. Total knee arthroplasty implanted with and without kinematic navigation. Int Orthop
. 2003; 27:366-369.
14. Insall JN. Surgical Techniques and Instrumentation in Total Knee Arthroplasty. In Insall JN, Windsor RE, Scott WN, Kelly M, Aglietti P, eds. Surgery of the Knee Ed. 2
. Second Edition. New York: Churchill-Livingstone; 1993:739-804.
15. Jeffery RS, Morris RW, Denham RA. Coronal alignment after total knee replacement. J Bone Joint Surg Br
16. Jenny JY, Boeri C. Computer-assisted implantation of total knee prostheses: a case-control comparative study with classical instrumentation. Comput Aided Surg
17. Lewis P, Rorabeck CH, Bourne RB, Devane P. Posteromedial tibial polyethylene failure in total knee replacements. Clin Orthop Relat Res
18. Merkow RL, Soudry M, Insall JN. Patellar dislocation following total knee replacement. J Bone Joint Surg Am
19. Miller MC, Berger RA, Petrella AJ, Karmas A, Rubash HE. Optimizing femoral component rotation in total knee arthroplasty. Clin Orthop Relat Res
20. Moreland JR. Mechanisms of failure in total knee arthroplasty. Clin Orthop Relat Res
21. Nagura T, Dyrby CO, Alexander EJ, Andriacchi TP. Mechanical loads at the knee joint during deep flexion. J Orthop Res
. 2002;20: 881-886.
22. Saragaglia D, Picard F, Chaussard C, Montbarbon E, Leitner F, Cinquin P. Computer-assisted knee arthroplasty: comparison with a conventional procedure: results of 50 cases in a prospective randomized study. Rev Chir Orthop Reparatrice Appar Mot
. 2001;87: 18-28.
23. Singerman R, Pagan HD, Peyser AB, Goldberg VM. Effect of femoral component rotation and patellar design on patellar forces. Clin Orthop Relat Res
24. Siston RA, Daub AC, Giori NJ, Goodman SB, Delp SL. Evaluation of methods that locate the center of the ankle for computer-assisted total knee arthroplasty. Clin Orthop Relat Res
25. Siston RA, Daub AC, Giori NJ, Goodman SB, Delp SL. Evaluation of methods that locate the center of the ankle for computer-assisted total knee arthroplasty. Clin Orthop Relat Res
26. Siston RA, Patel JJ, Goodman SB, Delp SL, Giori NJ. The variability of femoral rotational alignment in total knee arthroplasty. J Bone Joint Surg Am
27. Stöckl B, Nogler M, Rosiek R, Fischer M, Krismer M, Kessler O. Navigation improves accuracy of rotational alignment in total knee arthroplasty. Clin Orthop Relat Res
28. Stulberg SD, Loan P, Sarin V. Computer-assisted navigation in total knee replacement: results of an initial experience in thirty-five patients. J Bone Joint Surg Am
. 2002;84(Suppl 2):90-98.
29. Uehara K, Kadoya Y, Kobayashi A, Ohashi H, Yamano Y. Bone anatomy and rotational alignment in total knee arthroplasty. Clin Orthop Relat Res
30. Wasielewski RC, Galante JO, Leighty RM, Natarajan RN, Rosenberg AG. Wear patterns on retrieved polyethylene tibial inserts and their relationship to technical considerations during total knee arthroplasty. Clin Orthop Relat Res
31. Yoshioka Y, Siu DW, Scudamore RA, Cooke TD. Tibial anatomy and functional axes. J Orthop Res