Bone Anatomy and Rotational Alignment in Total Knee Arthroplasty

Abstract

The purpose of the current study was to investigate the bone anatomy in determining the rotational alignment in total knee arthroplasty using computed tomography. Axial images of 109 knees in 83 patients with varus osteoarthritis who had total knee arthroplasty were analyzed. On the images of the distal femur and the proximal tibia, a baseline for the anteroposterior axis of each component was drawn based on the epicondylar axis for the femur and the medial ⅓ of the tibial tuberosity for the tibia. The angle between these two lines was analyzed as the rotational mismatch between the components when they were aligned to the anatomic landmarks of each bone. Fifty-four knees (49.5%) had an angle of 5° or greater and 13 knees (11.9%) had an angle of 10° or greater. There was a tendency to align the tibial component in external rotation relative to the femoral component. The results indicated that the landmarks of each bone were the intrinsic cause of the rotational mismatch in knees with varus osteoarthritis. Surgeons doing total knee arthroplasties should be aware of this and check the rotational mismatch between the components. When it is present, the tibial component should be realigned to match the femoral component rotation to minimize problems caused by the mismatch.

Improper rotation of the femoral and tibial components in total knee arthroplasty may lead to various patellofemoral complications such as subluxation, dislocation, and wear. Despite the current high success rate of total knee arthroplasty, patellofemoral complications remain the most common cause of revision surgeries. ¹ One difficulty is that the rotational alignment of the total knee arthroplasty prostheses cannot be evaluated accurately in routine radiographs and quantitative analysis only is possible with three-dimensional imaging such as computed tomography (CT). ³

Although the importance of proper rotational alignment in total knee arthroplasty surgery is recognized, determination of rotational alignment is controversial when compared with the determination of axial alignment. Traditionally, rotational alignment of the femoral and tibial components has been determined separately based on the bony landmarks of each bone. As for the femoral component, aligning it to the epicondylar axis of the femur has been a widely used method. ^2,4,5 The tibial component traditionally has been aligned to the medial ⅓ of the tibial tuberosity. ⁶ Therefore, it is possible that the rotational alignments of the components of the prosthesis are not in complete accordance with each other. This discord in alignment can cause problems especially in knee prostheses with more conforming articular surfaces, which currently are popular.

In the current study, CT scans of patients with varus osteoarthritic knees who had total knee arthroplasty were analyzed for possible rotational malalignment between the femoral and tibial components when they were aligned to the corresponding landmarks of each bone.

MATERIALS AND METHODS

The current study was based on 109 knees in 83 patients who had total knee arthroplasty because of varus osteoarthritis of the knee. Patients with valgus deformity and rheumatoid arthritis were not included. There were 19 men and 64 women, and the average age was 70.9 years (range, 57–83 years). Preoperative CT scans were obtained to observe the epicondylar axis of the femur and the resection level of the proximal tibia. The patient was placed in the supine position on the CT scanning table with the involved extremity stabilized in extension in a specially designed leg holder (Fig 1). Using the anteroposterior (AP) and lateral scout views, the scan direction was aligned to be in the plane perpendicular to the mechanical axis of the tibia in the frontal plane with a 3° posterior tilt. Computed tomography scans of the epicondyles of the femur (3 mm thickness) and the proximal tibia (1 mm thickness) were obtained (Fig 2). In each scan sequence, baselines for the rotational positions of the femoral and tibial components were determined as described below.

Fig 1 A–B.: (A) For CT imaging, a device is used to hold the lower extremity in the desired rotational position. (B) The patient was placed in the supine position with the involved extremity in maximum extension.

Fig 2.: The scan direction was determined in a CT scout view and scans of 3-mm and 1-mm thicknesses were obtained in the epicondyles of the femur and the resection level in the proximal tibia.

The Baseline for the Femoral Component (The epicondylar axis of the femur)

A slice was selected in which the lateral and medial epicondylar prominences could be detected with certainty. In the chosen slice, the epicondylar axis was identified as follows: the surgical epicondylar axis, which connects the lateral epicondylar prominence and the medial sulcus of the medial epicondyle, ² could be drawn in 80 knees (73.4%) (Fig 3A). In the remaining 29 knees, the medial sulcus was not identified clearly and the clinical epicondylar axis which used the most prominent point of the medial epicondyle, was used for these knees instead (Fig 3B). ²

Fig 3 A–B.: The baseline for the femoral component is shown. (A) The surgical epicondylar axis was drawn when the medial sulcus was identified clearly. (B) The clinical epicondylar axis which used the most prominent point of the medial epicondyle, was used when the medial sulcus was not identified clearly.

The Baselines of the Tibial Component

A slice that traversed the optimum resection level in the proximal tibia (usually 6 to 8 mm below the intact lateral tibial articular surface) was selected. For the chosen slice, a line paralleling the epicondylar axis and traversing the longest distance on the tibia was drawn (Fig 4, Line A). The first line was drawn to be perpendicular and through the midpoint of Line A (Fig 4, Line AC). The second line was drawn connecting the midpoint of the Line A and the medial ⅓ of the patella tendon (Fig 4, Line BC). Theoretically, Line BC represents the rotational position of the tibial component determined based on the medial ⅓ of the tibial tuberosity. When the rotational positions of the femoral component were determined based on the epicondylar axis, the angle between these two lines (Angle ACB) represented the possible rotational mismatch between the components in extension. When Line BC was rotated externally relative to Line AC, the angle was expressed as a plus value.

Fig 4.: The angle ACB was measured based on the following baselines: Line A, a line paralleling the epicondylar axis and traversing the longest distance on the tibia was drawn, Line AC, a line perpendicular and through the midpoint of the tibial projection of the epicondylar axis, and Line BC, the line connecting the midpoint of the tibial projection of the epicondylar axis and the medial ⅓ of the patella tendon. The angle was expressed as a plus value when Line BC was externally rotated relative to Line AC.

RESULTS

The average of the angle ACB was 2.6° ± 5.4° (mean ± standard deviation; range, 16° to −10°) indicating that the rotational alignment of the femoral and tibial components was in good accordance as a whole, but there was a tendency for the tibial baseline to be slightly externally rotated relative to the femoral baseline. No statistically significant difference was observed between the knees in which the surgical and clinical epicondylar axes were used as the reference (p = 0.55 unpaired Student’s t test).

The distribution of the angle for each knee is shown in Figure 5. When the knees were examined, 54 knees (49.5%) had an angle of 5° or greater. Among these knees, 40 knees had a positive value, and 14 knees had a negative value. Thirteen knees (11.9%) had an angle of 10° or greater and all knees except one had a positive value.

Fig 5.: Distribution of the angle ACB is shown.

DISCUSSION

Rotational malalignment of total knee arthroplasty prostheses has a significant influence on patellofemoral complications and is the most common cause of revision surgeries. ¹ One possible cause is that the rotational alignment of the femoral and tibial components traditionally has been determined separately based on the bony landmarks of each bone. The current study was designed to quantitatively analyze the possible mismatch derived from the reference method based on the landmarks of each bone.

At least four methods have been proposed to determine the rotational alignment of the femoral components. ^8,9 As the rotational position of the femoral components influences the flexion space, all methods should be checked with reference to the configuration of the flexion space. Among the methods, aligning the components according to the surgical epicondylar axis provided less patellofemoral problems and a balanced flexion space. ² However, the practical usefulness of this axis is limited because the authors found that 27% of the knees (29 of 109) did not have an identifiable medial sulcus on CT images. On the tibial side, however, the rotational position of the component does not influence the size of the extension space but is limited in view of the coverage or the overhang on the tibial resection or both. There have been two techniques to determine the rotational position of the tibial component. One technique aligns the tibial components to the medial ⅓ of the tibial tubercle as described by Insall. ⁶ However, the current authors are unaware of any published literature on the theoretical background of this technique. The other technique aligns the tibial component according to the femoral component rotation. In this method, a conforming and mobile tibial-bearing surface is placed and the knee was flexed several times. The rotational position of the mobile-bearing surface, which was self-aligned to the femoral component rotation, is marked on the tibia and the tibial component is inserted accordingly. The two baselines used in this study represented the AP axis of the femoral and tibial components when they were aligned perfectly to the anatomic landmarks of each bone. The patella tendon was used instead of the tibial tuberosity because it was more accurately identifiable on the CT images.

The overall results showed that the mean value of the angle ACB was less than 2° and the femoral and tibial baselines were in good accordance as a whole. However, when the knees were evaluated individually, a mismatch of 10° or more was observed in 11.9% of the knees and ½ of the knees had a 5° or greater mismatch. This amount of difference could be tolerated in the prostheses with less conforming articular surfaces used in the 1980s. However, recent attention to wear problems has led to more widespread use of prostheses with conforming, deep-dished articular surfaces. When these conforming articular surfaces are used, a possible mismatch of 5° to 10° has become increasingly important in view of the rotational stress applied to the bone-implant interface and wear. Another design concept to obviate this problem would be mobile-bearing knee prostheses in which the femorotibial rotation is self-aligning and therefore could compensate the rotational mismatch.

There was a tendency to put the tibial component in an externally rotated position with reference to the femoral component. This was in contrast to the common belief that the possible error was to put the tibial component in an internally rotated position. Theoretically, internal rotation of the tibial component leads to patellofemoral complications whereas an externally rotated tibial component is advantageous in this respect. ² However, excessive external rotation potentially causes posterolateral overhang of the tibial component, aggravation of tibial internal torsion, and subsequent toe-in gait already present in patients with varus osteoarthritis of the knee. ¹⁰ There was a significant correlation between the external rotation of the tibia and femorotibial angle (data not shown) indicating that this mismatch is increasingly important in advanced varus knees.

A possible criticism of the current study would be that soft tissue release affects the rotational position of the tibia and thereby influences the results. To date, however, no data are available on the effect of soft tissue release on the relative rotational positions of the femur and the tibia. Considering that the mismatches shown in the current study were internal and external in direction, they could not be eliminated by medial soft tissue release alone. The two methods were tested intraoperatively to determine tibial component rotation after the soft tissue release, and it was found that a frequent substantial discrepancy was to put the tibial component in external rotation when it was aligned to the medial ⅓ of the tuberosity. This observation prompted the current authors to quantitatively analyze the mismatches and the same tendency was observed in the current study. The effect of soft tissue release probably does not change the main conclusion of this study and the finding is of clinical importance in alerting the possible cause of mismatch in knees with advanced varus osteoarthritis. Another consideration is the ethnic difference. It was reported that the varus knee in Japanese individuals has more proximal tibia vara and more bowing of the femoral shaft, ⁷ and the results shown in the current study may not be directly extrapolated to knees in Caucasian individuals.

The current study showed a possible cause of mismatch in rotational alignment in total knee arthroplasty. Fifty-four knees (49.5%) had an angle of 5° or greater and 13 knees (11.9%) had an angle of 10° or greater. There was a tendency to put the tibial component in external rotation relative to the femoral component when they were aligned to the medial ⅓ of the tibial tuberosity. Surgeons doing total knee arthroplasties should be aware of this and should check the rotational mismatch between the components. When it is present, the tibial component should be realigned according to the femoral component rotation to minimize the problems caused by the rotational mismatch.