Introduction

While coronal mal-alignment is thought to be one of the primary causes of knee osteoarthritis progression, not much attention has been paid to the obliquity of the knee joint line after total knee arthroplasty (TKA) in the coronal plane. In one recent study, it was reported that the joint line obliquity in the coronal plane in the native knee after skeletal maturity was parallel to the floor and perpendicular to the weight-bearing axis of the body in bipedal stance, irrespective of the magnitude of varus deformity [25]. In contrast, for symptomatic arthritic patients with varus deformity, the knee joint line slanted down to the lateral side. This parallelism might be the result of adaptation to the bipedal upright walking against gravity [14] as the geometry and alignment of the human knee joint would be affected by vertical mechanical loading. This continuous loading would determine the configuration of the knee joint at the skeletal maturity by Hueter–Volkmann’s law, and it would be reasonable to attribute the obliquity of the joint line to the weight-bearing line of the limb to achieve an efficient evolutionary selection for particular biomechanical properties of knee joint loading and shear stress [3, 25]. The fact that natural knee joint line is horizontal to the floor and perpendicular to the weight-bearing axis of the body in bipedal stance also implies that the configuration is kinematically proficient [25].

Postoperative restoration of the neutral alignment has been the primary target for classic total knee arthroplasty (TKA) over the past two decades in regard to function and longevity [12, 23]. However, long-term follow-up retrospective studies have not shown any differences in survival rate between the TKA cases with neutral alignment and outliers [2, 15, 17, 21]. Moreover, the recent introduction of “kinematically aligned” TKA (KATKA) has been regarded as the new dogma opposing the traditional doctrine of seeking neutral limb alignment in TKA [1]. In the knee joint, kinematics deals with the relative relationship of the femur, tibia, and patella at various degrees of flexion and without any applied force. The normal knee kinematics is not governed by the relative position of the centre of hip, knee, and ankle joint which conventional or computer-assisted navigation instrument relies upon to align a TKA mechanically. The joint configuration, menisci, and collateral and cruciate ligaments determine the normal knee kinematics. The modification of the conventional techniques is characterized by the restoration of pre-arthritic knee articular surface, and therefore, it has become possible to recover knee joint kinematics by placing the femoral component on kinematic axes [7]. This difference, in the principle, would lead to a difference in knee joint line obliquity. By restoring the natural knee joint parallelism, clinicians can expect to obtain a more kinematically proficient knee joint. Therefore, comparing the distribution of the knee joint line after TKA between different alignment methodologies is of great interest, and to our knowledge, no related study has been published so far on the subject.

The purposes of this study were: (1) to find the distribution of the native joint line in a population of normal patients with normal knees; (2) to compare the native joint line orientation between patients receiving conventional mechanically aligned TKA (MATKA), navigated MATKA, and KATKA; and (3) to determine which of the three TKA methods aligned the postoperative knee joint perpendicular to the weight-bearing axis of the limb in bipedal stance. Our hypothesis states that KATKA would recover pre-arthritic knee articular geometry and joint line obliquity. Consequently, it is affirmed that joint line configuration after KATKA would be more similar to that of healthy knees than after MATKA.

Materials and methods

For the distribution of the native joint line, 50 knees of 50 consecutive young adults without any history of chronic knee pain were identified from the database in our institution. Anteroposterior long-leg standing radiographs of the lower extremities were taken as a part of the routine radiologic examinations. The number of men and women was identical, and the average age for the participants was 23.5 years (range 20–25). One senior surgeon performed all the TKAs for the same indications. End-stage osteoarthritis, rheumatoid arthritis, and secondary osteoarthritis were the primary diagnoses for the TKA recipients. The average age of the each treatment group was 70.3 ± 3.4 years (range 58–81) in conventional MATKA patients, 69.6 ± 5.2 (range 59–78) in navigation MATKA patients, and 71.0 ± 4.1 years (range 59–80) in KATKA patients, and these differences were not statistically significant. The percentage of female patients was higher than 90 % for all groups, and any deviations from this composition across the groups were not statistically different.

Pre-operative HKA was 8.4° ± 5.0°, 9.5° ± 5.0°, and 10.4° ± 6.2° in conventional MATKA, in navigation MATKA, and in KATKA patients, respectively, and their differences were not significant. Sixty-five consecutive patients, who were operated for conventional MATKA from February 2013 to June 2014, were included in the study. The same surgeon operated MATKA in 65 consecutive patients from October 2007 to April 2008 using imageless electromagnetic computer navigation TKA and following a fixed protocol. The surgeon initiated implantation of TKAs with kinematic alignment since October 2013, and the initial 65 patients were included in the current study.

Kinematic alignment in the coronal and sagittal plane was achieved using a unique cartilage probing technique (Fig. 1) instead of pre-operative MRI, which was used in the original technique [9, 18]. The cartilage probing with a spinal needle enabled us to measure distal and posterior femoral and tibial cartilage depth. The difference in cartilage depth measured intraoperatively would provide more accurate information regarding pre-operative knee cartilage status. The thickness of the articular surface of the implant was supposed to be the sum of the depth of resected bone, anticipated cartilage loss, and bone defect, and the kerf of the saw blade. For the procedure, if there was a certain amount of cartilage loss in the medial or lateral side, the resection was decreased by that amount so as to ensure equally measured resection in both medial and lateral components of the distal and posterior femur and proximal tibia, thus recreating the patient’s native joint orientation. The remaining procedure was completed using standard TKA instrumentation [9].

Fig. 1
figure 1

Cartilage probing technique with a spinal needle in which fine lines were laser engraved 1 mm apart to measure femoral and tibial cartilage depth

Full size image

All measurements of coronal orientation were taken using long-leg standing radiographs, and the radiographs were taken for every patient as routine. Drawings and measurements were performed using the embedded functions of the PACS system (PiViewSTAR, Infinitt, Seoul, Korea). Pre-operative hip–knee–ankle (HKA) angle was measured in all TKA patients. The HKA angle describes the relation between the mechanical axis of the femur and tibia. The femoral axis was defined as a line connecting the centre of the femoral head and the mid-point of the knee. A line connecting the centre of the knee and the centre of the ankle defined the axis of the tibia. A best-fit circle was drawn around the perimeter of the femoral head to determine the centre. The centre of the knee was defined by the mid-point of the femoral condyles at the level of the top of the intercondylar notch. The mid-point of the width of the talar dome determined the centre of the ankle. The HKA angle was described as a deviation from 180°. Varus overall limb alignment was expressed as a positive value. Pre-operative and postoperative joint line orientation angles (JLOAs) were measured. The definition of JLOA is the angle formed between the tibial joint line and a line parallel to the floor [10, 25]. The pre-operative tibial joint line was defined as a line tangential to the medial and lateral tibial plateau. The postoperative tibial joint line was determined as a line drawn along the tibial plate. The tibial joint line that slanted down to the medial side was expressed as positive value.

The study protocol was approved by the institutional review board (IRB) of our institution (Ajou University Hospital, MED-MDB-14-172).

Statistical analysis

All numbers were calculated to the second decimal place and presented to the first after raising the second. The distribution of the data was examined with the Kolmogorov–Smirnov (K–S) test, and all measurements showed a normal distribution. Subsequently, the differences were compared using the one-way ANOVA to determine whether there were any significant differences between the means of the three TKA groups. Tukey’s test was used as a post hoc test. To decide the best way to align postoperative knee joint line to horizontal, the percentage of a patient whose JLOA remained within one and two standard deviations of the average JLOA of the native knee was evaluated for respective three treatment groups and compared using Chi-square test. If the percentage of such patient was higher in one group than in the other, the alignment principle employed in the group was to align postoperative knee joint line similar to the natural knee. p values smaller than 0.05 were considered statistically significant. All analyses were performed using SPSS software, version 16.0 (SPSS Inc., Chicago, IL, USA). A power analysis revealed that 65 patients would be required to detect a difference in measurements of 1.6° with a power of 80 % and a significance level of 0.05 [10].

Results

Average JLOA in 50 knees in 50 young normal controls was 0.2° ± 1.1° and also parallel to the floor. The average pre-operative JLOA for 65 KATKAs, 65 mechanical MATKAs, and 65 navigation MATKs was −1.7° ± 1.9°, −2.5° ± 2.6°, and −2.3° ± 2.9°, respectively, and they were not statistically different. The postoperative JLOA slanted down to the lateral side in conventional MATKA (−3.3° ± 2.2°) and in navigation MATKA (−2.6° ± 1.8°), while it was parallel in KATKA (0.6° ± 1.7°). The distribution of JLOA of young adults, postoperative JLOA of conventional MATKA, navigation MATKA, and KATKA is shown in Fig. 2a, b. Postoperative JLOAs of the native knee and that of three treatment groups were compared, and they were significantly different (p < 0.001). The JLOA of natural knee and KATKA was significantly larger than that of conventional MATKA and navigation MATKA (Fig. 2a, b). The percentage of patients whose JLOA remained within one standard deviation of the average JLOA of the young adults was 6.9 % for conventional MATKA, 16.9 % for navigation MATKA, and 50.8 % for KATKA. The proportion of patients whose JLOA was within two standard deviations was 20.7, 38.5, and 81.5 %, respectively. The percentage of the MATKA groups was significantly smaller than that of KATKA with respect to both the comparisons.

Fig. 2
figure 2

a Histogram depicts knee joint line orientation angle (JLOA) in healthy adults, conventional MATKA, navigation MATKA, and KATKA. The JLOAs of the MATKAs slanted down to the lateral side and were broadly distributed, while they were horizontal to the floor and perpendicular to the weight-bearing axis in the bipedal stance in normal healthy adults and patients receiving KATKA. b One-way diamond mean plots with 95 % confidence interval of the postoperative JLOA in healthy adults, conventional MATKA, navigation MATKA, and KATKA. The JLOA of the groups with a single asterisk mark is significantly lower than of the other groups without asterisk mark (p < 0.001)

Full size image

Discussion

The notable finding of this study is that the postoperative knee JLOA in standing posture was horizontal to the floor and perpendicular to the weight-bearing line in most of the cases in KATKA patients. More than 80 % of the KATKA patients had postoperative JLOA within two standard deviations of the average knee JLOA of young adults. However, the postoperative JLOA in MATKA patients was relatively slanted down to the lateral side in both conventional and navigation MATKA.

One recent study showed that the postoperative JLOA in KATKA is horizontal to the floor [10]. However, the present study also had JLOA of MATKA measured by the same operator using the conventional and navigation techniques, while the previous study did not. Postoperative JLOA was 1.0° ± 1.9° in 52 patients [10], which was comparable to that of the current study (0.6° ± 1.7°). The current study also showed similarity in JLOA between KATKA patients and normal healthy knees (0.2° ± 1.1°). A past study performed in asymptomatic Caucasian reported a 0.2°–0.3° (SD 1.9°–2.2°) JLOA, while the current study, which was performed in Korean adults, reported comparable JLOA.

The clinical significance of this study lies in the resemblance of JLOA between KATKA patients and asymptomatic adults. KATKA may theoretically recover pre-arthritic knee joint cartilage thickness while minimizing soft tissue release [7, 9], and as such, the joint line configuration and knee joint kinematics should be restored. The current study showed proper restoration of coronal joint line configuration relative to the floor or the vertical axis when compared with healthy individuals even though no attention was paid to the postoperative knee alignment intraoperatively.

This joint line parallelism to the floor can be found only in humans when compared with anthropoids and monkeys [14] and can be explained as the result of evolution. Bipedalism might have resulted in efficient coronal joint line configuration, which is parallel to the floor. All human beings go through the same developmental pathways until adulthood in every part of the body including the knee joint. Initial genu varum in newborn leads to neutral alignment during the toddler period, which is followed by physiologic genu valgum [27]. During adulthood, most of the asymptomatic knees would possess coronal joint line parallelism regardless of the amount of HKA varus [25]. The knees that fail to sustain this parallelism lead to the knee joint line slanting down to the lateral side eventually and arthritic symptoms may ensue. Postoperative MATKA showed similar coronal joint line configuration because the proximal tibia is always resected perpendicular to neutral HKA, which is about 3° valgus to the vertical axis. This difference in coronal joint line configuration found in standing radiographs in this study might be the reason for the clinical and biomechanical superiority of KATKA when compared with MATKA, as reported in recent studies [4, 8, 11].

Despite the clinical and biomechanical significance, there is still concern for long-term survivorship of the KATKA. One biomechanical study on MATKA predicted significant load variations in outliers [26]. Higher failure rates in TKAs with excessive joint line orientation, especially in tibial component varus, have been reported [5, 20, 22]. Still others have challenged this long-held notion that coronal alignment is crucial for the long-term survival of the implant [2, 17, 21]. One recent 6-year follow-up of 214 consecutive KATKA showed that the amount of varus alignment of the tibial component played no adverse effect on the implant survival or function [8]. The proposed reason for this finding is that recovered pre-arthritic knee joint line in KATKA would restore pre-arthritic collateral ligament and enable a more physiologic strain than in MATKA [8]. However, one recent computer simulation study showed higher contact stress in medial tibiofemoral and patellofemoral joints despite near normal knee kinematics [11]. Another computer simulation study also showed that neutral alignment did not always lead to the equal load distribution in standing position with respect to both medial and lateral components in MATKA [19]. This contradicting evidence warrants further studies regarding the outcome and biomechanics of KATKA.

The present study had several limitations. The numbers of included patients were relatively small, and as a result, the patient group might not have thoroughly represented the variation in patients undergoing TKA. Additionally, it can be argued that a single coronal measurement is not sufficient to properly assess biomechanical effects, and evidence on the functional importance of tibial component orientation in the sagittal plane is currently inconsistent [13, 24]. Finally, the clinical significance of standing knee joint line configuration is still questionable. Recent studies have shown that static postoperative tibiofemoral alignment alone does not predict dynamic tibial loading during gait [6, 16]. Therefore, joint line parallelism to the floor in the standing position does not guarantee that the parallelism is maintained during the gait cycle.

Despite these limitations, clinicians can expect the natural joint line obliquity to be restored postoperatively by KATKA. Even though long-term survival data of KATKA are unavailable, the current study showed KATKA can be a valid alternative to replace an arthritic knee joint because the coronal joint line configuration is similar to the natural knee joint that is parallel to the floor and vertical to the weight-bearing axis. This structural resemblance may ensure clinical superiority that was shown in a previous study [4] with short-term follow-up. KATKA may be a better option of TKA in elderly patients in whom longevity of the implant matters less than in younger patients.

Conclusion

This study showed that postoperative JLOA after KATKA was more similar to that of healthy knees than that of MATKA. JLOA was restored postoperatively in most of the patients receiving KATKA. It is concluded that KATKA can restore pre-arthritic knee joint line orientation, while MATKA is inefficient in achieving the same if navigation TKA is employed. The restoration of the JLOA might explain the clinical superiority of KATKA while further studies are required to document its effect on implant longevity.