The Morphology of a Kinematically Aligned Distal Femoral Osteotomy Is Different from That Obtained with Mechanical Alignment and Could Have Implications for the Design of Total Knee Arthroplasty

Background: Kinematically aligned total knee arthroplasty (KA-TKA) may lead to a different pattern of osteotomy from mechanically aligned total knee arthroplasty (MA-TKA). This paper aims to analyze the effects of KA and MA on the morphology of the distal femoral osteotomy surface. Methods: Computed tomography scans of 80 TKA candidates were reconstructed into 3D models. The measurement of bone morphology was performed after the distal femur cut according to two different alignment techniques. The aspect ratio, trapezoidicity ratio, and asymmetry ratio of the distal femur were assessed. Results: The aspect ratio and the asymmetry ratio in the KA group was significantly lower than that in the MA group in the general population (p < 0.001). The trapezoidicity ratio in the KA group was significantly higher than that in the MA group in the general population (p < 0.001). Conclusions: It was found that KA-TKA and MA-TKA presented different morphologies of the distal femoral osteotomy surface, and this difference was also influenced by gender. The surgery pattern of KA-TKA and MA-TKA and gender should be considered when surgeons choose femoral prostheses.


Introduction
Total knee arthroplasty (TKA) has been proven successful in improving knee pain and functions and the quality of life of patients with advanced knee arthritis [1]. The choice of prosthesis that matches the bone surface morphology after cutting is of high importance, otherwise mismatching may influence pain, function, and flexion after TKA [2,3]. It has been reported that bone-implant mismatch can occur in TKA due to variation in the aspect ratio (ML/AP; ML = medial-lateral width, AP = anteroposterior length) and the asymmetry ratio (LAP/MAP; LAP = lateral anteroposterior, MAP = medial anteroposterior) [4]. Marmor et al. reported that over-voluming occurred in 24% of TKA pre-operative planning [5]. In the past decade, studies on distal femur morphology have primarily focused on the aspect ratio of the distal femur [6][7][8][9][10]. Based on current findings, manufacturers tend to produce narrow femoral prosthesis to improve match with bone surface morphology. Mahfouz et al. described the complex variation of femur morphology, considering it to be too simplistic to divide distal femur morphology into wide and narrow [2]. Trapezoidicity is a recently proposed measure for describing distal femur morphology. Distal femurs are reported to be more trapezoidal than the common femoral implants in Caucasians [11]. Although some researchers have reported ethnic and gender differences in distal femur morphology [11][12][13], there are no such differences in prostheses used for TKA.
Mechanical alignment (MA) TKA is often considered the gold standard alignment technique for TKA. MA-TKA aims to restore mechanical alignment and achieve a joint 2 of 8 line perpendicular to the mechanical axis of the lower extremities. In recent studies, kinematically aligned (KA) TKA has been found to achieve good outcomes in TKA [14][15][16]. KA-TKA is differentiated from MA-TKA in that the former aims to restore the thickness of the bones and cartilages removed, with the external rotation of the femur set at 0 • to the anatomical joint line [17]. However, to the best of our knowledge, no study has compared the morphology of the distal femoral osteotomy surface using KA-TKA and MA-TKA techniques.
Therefore, the purpose of this study was to compare the aspect ratio, trapezoidicity ratio, and asymmetry ratio of the distal femur in KA and MA groups. We hypothesized that the morphology of a kinematically aligned distal femoral osteotomy would be different from that obtained with mechanical alignment.

Materials and Methods
The present study was approved by the institutional review board of our institution. Measurement was performed on the preoperative CT scan images of the lower limbs in 86 patients scheduled for primary TKA from 2014 to 2015. There were 31 males and 55 females. Six patients were excluded due to a history of surgery or trauma, or unclear CT images caused by artifacts related to metal or contrast agents. Finally, 30 males and 50 females were included, who were aged 72 ± 2.6 years old on average, with an average body mass index of 26.8 ± 3.5 kg/m 2 ( Table 1). All the patients included had knee osteoarthritis with varus deformity before surgery. The CT images were acquired in a supine position using a 64-slice multi-detector CT unit (Siemens Sensation, Munich, Germany) for all patients. The CT scan was performed in full knee extension, with two legs immobilized in neutral rotation.

Virtual Surgery
KA-TKA was performed as follows: The distal femoral cutting was simulated on a plane parallel to the line facing the distal surfaces of both condyles ( Figure 1A). The thickness of the distal femoral cutting was 7 mm considering a mean articular cartilage thickness of 2 mm. In other words, 9 mm thickness resected was to substitute the corresponding region of the femoral component. The maximum medial AP perpendicular to the line connecting the medial condyle and the posterior margin of the medial and lateral condyles was taken as medial anteroposterior (MAP). The maximum lateral AP was lateral anteroposterior (LAP). The larger of the two values was set to the AP value ( Figure 2). The medial-lateral widths were measured at three positions: from the outmost points of the medial and lateral posterior condyles, the posterior medial-lateral (PML) width was measured at 7 mm anteriorly, the central medial-lateral (CML) width was measured at 50% AP, and the anterior medial-lateral (AML) width was measured at 75% AP ( Figure 3). Narrowing angles were measured between the line perpendicular to the PML line and the cortex at 75% (angle α or 'anterior narrowing angle') and at 50% (angle β or 'central narrowing angle') of the AP. Angles were measured both on the medial (αM and βM) and on the lateral side (αL and βL) ( Figure 3). and the anterior medial-lateral (AML) width was measured at 75% AP ( Figure 3). Narrowing angles were measured between the line perpendicular to the PML line and the cortex at 75% (angle α or 'anterior narrowing angle') and at 50% (angle β or 'central narrowing angle') of the AP. Angles were measured both on the medial (αM and βM) and on the lateral side (αL and βL) ( Figure 3).    and the anterior medial-lateral (AML) width was measured at 75% AP ( Figure 3). Narrowing angles were measured between the line perpendicular to the PML line and the cortex at 75% (angle α or 'anterior narrowing angle') and at 50% (angle β or 'central narrowing angle') of the AP. Angles were measured both on the medial (αM and βM) and on the lateral side (αL and βL) ( Figure 3).    The posterior medial-lateral (PML) width was the posterior femoral condylar cutting width and the anterior medial-lateral (AML) width was measured at 75% AP. and the anterior medial-lateral (AML) width was measured at 75% AP (Figure 3). Narrowing angles were measured between the line perpendicular to the PML line and the cortex at 75% (angle α or 'anterior narrowing angle') and at 50% (angle β or 'central narrowing angle') of the AP. Angles were measured both on the medial (αM and βM) and on the lateral side (αL and βL) (Figure 3).    MA-TKA was performed as follows: The distal femoral cutting was simulated on a plane perpendicular to the femoral mechanical axis ( Figure 1B). The thickness of the distal femoral cutting was 7 mm. The maximum medial AP perpendicular to the line connecting the medial condyle and the posterior margin of the lateral condyle was taken as MAP. The maximum lateral AP was LAP. The larger of the two values was set to the AP value (Figure 2). The medial-lateral widths were measured at three positions: parallel with surgical epicondylar axis and from the outmost point of the posterior femoral condyle, PML was measured at 7 mm anteriorly, CML was measured at 50% AP, and AML was measured at 75% AP. Narrowing angles were measured between the line perpendicular to the PML line and the cortex at 75% (angle α or 'anterior narrowing angle') and at 50% (angle ß or 'central narrowing angle') of the AP. Angles were measured both on the medial (αM and βM) and on the lateral side (αL and βL) (Figure 3). This simulation has previously been validated in research by Kim et al. [18]. The aspect ratio (ML/AP), asymmetry ratio (LAP/MAP), and trapezoidicity ratio (PML/AML) of the distal femur were assessed (Figure 2). For each ratio, we classified the shape relative to the median value: femurs were considered wide if the aspect ratio was above the median; femurs were considered narrow if the aspect ratio was below the median; femurs were trapezoidal if the trapezoidicity ratio was above the median; femurs were rectangular if the trapezoidicity ratio was below the median.

Statistical Analysis
Statistical analysis was performed using the IBM SPSS Statistics software package (IBM SPSS Statistics 21, SPSS IBM, NY, USA). Descriptive statistics were used to summarize the data. Data were presented as mean ± SD. Shapiro-Wilk tests were used to assess the normality of distributions. An independent sample t test was used to test statistical significance. p values < 0.05 were considered statistically significant.
All the measurements on the analysis were made by one orthopedic surgeon. Another orthopedic surgeon measured 30 randomly selected patients to assess interobserver reliability of measurements. The intraclass correlation coefficient of interobserver reliability was 0.88 which was indicative of excellent agreement. The intraobserver agreement rate was 0.94.

Results
The aspect ratio (ML/AP) in the KA group was significantly lower than that in the MA group in the general population (p < 0.001, Table 2), males (p < 0.001, Table 3), and females (p < 0.001, Table 4).     The asymmetry ratio (LAP/MAP) in the KA group was significantly lower than that in the MA group in the general population (p < 0.001, Table 2), males (p = 0.01, Table 3), and females (p < 0.001, Table 4).
The trapezoidicity ratio (PML/AML) in the KA group was significantly higher than that in the MA group in the general population (p < 0.001, Table 2), males (p < 0.001, Table 3), and females (p < 0.001, Table 4).
Overall αL was significantly smaller in the KA group than in the MA group (p < 0.001). Overall βL was significantly smaller in the KA group than in the MA group (p = 0.002). There was no significant difference in overall αM between KA and MA groups (p = 0.189).
Overall βM was significantly larger in the KA group than in the MA group (p < 0.001) ( Table 2).
All of the indicators above of females were lower than those of males in the KA group. In the MA group, there was no gender difference in LAP/MAP and βM. All of the remaining indicators were lower in females than in males (Table 5).