1 Introduction

A precise understanding of in vivo kinematics of the knee joint is essential to reproduce healthy knee kinematics after surgical procedures such as knee arthroplasties and to evaluate their outcomes. Previous studies have reported the direct measurement of skeletal kinematics of knee joints from three-dimensional (3D) bone models and series of radiographic images.^1,2 According to these studies, dynamic knee kinematics such as tibiofemoral flexion-extension, axial rotation angles, and anterior-posterior (AP) translation, significantly varies based on the activity performed both in healthy knees and osteoarthritic (OA) knees.^1–3

Glaister et al. reported that twisting activity makes up to 50% of daily activities in home or community environments such as walking from office to a parking lot and walking through a cafeteria.⁴ Weiss et al. reported that approximately half of the patients after total knee arthroplasty (TKA) considered their turning and cutting abilities as important to their daily activities.⁵ Therefore, an evaluation of knee kinematics during twisting activity is critical to enhancing the functional activity of knee joints during frequent daily activities.

Previous studies have assessed 3D in vivo kinematics in knees with OA under weight-bearing conditions.^1,2,6–8 Severe medial OA knees were shown to have significantly less knee extension, less posterior femoral rollback, less tibial internal rotation with flexion, and larger medial shift of the femur compared with healthy knees during knee flexion-extension activities.^2,6,7 Kinematic analyses of healthy and OA knees immediately before TKA is critical to achieving a better understanding of pathological and reconstructed knees.⁷

The purpose of this current study was to compare and contrast the in vivo kinematics in healthy and OA knees during twisting activity using density-based image-matching techniques.² We sought to answer the following two questions: (1) how does the femur flex, rotate, and translate relative to the tibia during a twisting activity in both healthy and OA knees? and (2) do OA knees demonstrate any abnormal kinematic pattern compared with healthy knees during a twisting activity?

2 Materials and methods

2.2

2.2 Radiological imaging procedure

Radiographic imaging was performed as described previously by Murakami et al.¹ Briefly, a flat-panel X-ray detector (FPD; Ultimax-I, Toshiba, Tochigi, Japan), with an image area of 420 mm (H) × 420 mm (V), a resolution of 0.274 mm × 0.274 mm/pixel, 0.02 s pulse width, 80 kV, and 360 mA, recorded continuous AP radiographic images at a frame rate of 10 Hz during the twist. The procedure entailed subjects transitioning their position through four different phases-initially standing upright, followed by twisting their body maximally to the ipsilateral side of the target knee and then, maximally to the contralateral side of the target knee. Finally, they twisted back to the neutral position. While twisting activity the upper body to the most rotated position possible at a self-selected speed, both of their feet were firmly kept on the ground. Four X-ray images — the initial upright position, the maximal ipsilateral twist, the neutral position, and the maximal contralateral twist, were recorded and used for analysis (Fig. 1). Computed tomography (CT) (Aquilion, Toshiba, Tochigi, Japan) was performed using a 512 × 512 image matrix, a 0.35 mm × 0.35 mm/pixel, and 1-mm thickness. CT scan was performed over a region spanning the superior edge of the pelvis to 150 mm below the ankle joint line, as described previously.^1,2

Close

Fig. 1 Subjects performed twist activity (upper panels) while their knee was recorded using a flat-panel X-ray detector. The representative X-ray images (lower panels) show the left knee being studied. Twist activity was divided into four phases: (a) the initial position; (b) the maximal ipsilateral twist; (c) the neutral position; and (d) the maximal contralateral twist.

Export to PPT

3 Results

The healthy and OA knees exhibited the following knee flexion angles: 0.3° ± 7.6° and 8.6° ± 6.3° at the initial upright position; 5.9° ± 13.2° and 10.5° ± 7.8° at the ipsilateral twist; 3.6° ± 12.3° and 9.6° ± 7° at the neutral position; and 5.6° ± 11.8° and 11.6° ± 7.5° at the contralateral twist, respectively. The flexion angle difference observed between the healthy and OA knees at the initial upright position was statistically significant (p < 0.05) (Fig. 2).

Close

Fig. 2 Mean flexion-extension [+ flexion, - extension] and varus-valgus angles [+ varus, - valgus] of the femur relative to the tibia during twisting activity at (a) the initial position; (b) the maximal ipsilateral twist; (c) the neutral position; and (d) the maximal contralateral twist. The solid-green and dashed-red lines indicate the healthy and OA knees, respectively. *p < 0.05.

Export to PPT

In the tibiofemoral rotation, the rotational angles for the healthy and OA knees – were 2.7° ± 9.6° and −2.2° ± 7.6° at the initial position; −11° ± 9.3° and −4.4° ± 7.2° at the ipsilateral twist; −3.7° ± 8.5° and −2.7° ± 7.4° at the neutral position; and 9.5° ± 5.6° and 2.7° ± 8° at the contralateral twist, respectively. There were significant differences between healthy and OA knees in rotational angles at the ipsilateral and contralateral twists (p < 0.05). The rotational angle from the ipsilateral twist to the contralateral twist in the OA knees (7.1° ± 4.7°) was significantly smaller than the one in the healthy knees (20.4° ± 8.9°) (p < 0.05) (Fig. 3).

Close

Fig. 3 Mean rotation angle [+ external rotation, - internal rotation] and anterior-posterior (AP) translation of the femur relative to the tibia [+ anterior, - posterior] during twisting activity at (a) the initial position; (b) the maximal ipsilateral twist; (c) the neutral position; and (d) the maximal contralateral twist. The solid-green and dashed-red lines indicate the healthy and OA knees, respectively. *p < 0.05.

Export to PPT

Varus-valgus angle measurements showed the following angles for the healthy and OA knees: 0.9° ± 3.6° and 6.4° ± 5.2° at the initial position; 0° ± 4.6° and 6.2° ± 5.3° at the ipsilateral twist; 1.1° ± 3.4° and 6.4° ± 5.2° at the neutral position; and 0.9° ± 3.4° and 6.5° ± 5.1° at the contralateral twist, respectively (Fig. 2). OA knees showed significantly larger varus angles in all positions compared to healthy knees (p < 0.05).

The analysis of the AP position revealed the following for the healthy and OA knees: 5 mm ± 6 mm and 3.2 mm ± 6.8 mm at the initial position; 0.7 mm ± 7.6 mm and 3.1 mm ± 6 mm at the ipsilateral twist; 3.6 mm ± 7.4 mm and 2.6 mm ± 6.4 mm at the neutral position; and 5.5 mm ± 7.4 mm and 3.6 mm ± 6.1 mm at the contralateral twist, respectively. The OA knees showed a tendency for decreased posterior movement in ipsilateral twist compared to the healthy knees without a significant difference (p > 0.05) (Fig. 3). With respect to the medial/lateral femoral condyles, the absolute values of the AP translations in healthy and OA knees were observed to be 8.6 mm ± 4.9 mm/15.7 mm ± 5.4 mm and 4.9 mm ± 3.5 mm/5.0 mm ± 3.4 mm, respectively (Fig. 4). There were significant differences in the absolute values of the AP translations in both the medial and the lateral condyles (p < 0.05).

Close

Fig. 4 Mean movement of the femoral trans-epicondylar axis (TEA) projected onto the axial plane of the tibia during twisting activity. The blue, green, yellow, and red lines indicate the initial position, the maximal ipsilateral twist, the neutral position, and the maximal contralateral twist, respectively. OA, osteoarthritis.

Export to PPT

Tibiofemoral rotational angle in the ACL-intact/attenuated and -deficient OA knees was −3° ± 6.4° and −0.8° ± 9.4° at the initial position; −5.5° ± 7.1° and −2.5° ± 7.5° at the ipsilateral twist; −3.9° ± 6.4° and −0.8° ± 8.8° at the neutral position; and 1.6° ± 7.3° and 4.5° ± 9.1° at the contralateral twist, respectively. The observed differences in rotational angles for intact/attenuated vs. deficient ACL in OA knees lacked statistical significance (Fig. 5). Data obtained for AP position in ACL-intact/attenuated and ACL-deficient OA knees showed 4.2 mm ± 6.6 mm and 1.6 mm ± 7.3 mm at the initial position; 4.7 mm ± 5.7 mm and 0.6 mm ± 5.8 mm at the ipsilateral twist; 4.3 mm ± 6.1 mm and −0.2 mm ± 6.1 mm at the neutral position; and 4.9 mm ± 6.2 mm and 1.5 mm ± 5.7 mm at the contralateral twist, respectively. The differences in AP translations between the intact/attenuated and deficient ACL in OA knees were statistically significant at the neutral phase (p < 0.05).

Close

Fig. 5 Mean rotation angle [+ external rotation, - internal rotation] and anterior-posterior (AP) translation of the femur relative to the tibia [+ anterior, - posterior] during twisting activity at (a) the initial position; (b) the maximal ipsilateral twist; (c) the neutral position; and (d) the maximal contralateral twist. The solid-green, dashed-blue, and dotted-orange lines indicate the healthy, ACL intact/attenuated, and ACL-deficient knees, respectively.

Export to PPT

4 Discussion

The most remarkable findings of this study evaluating knee kinematics during twisting activity are that OA knees showed a significant decrease in – a) tibiofemoral rotations at both the ipsilateral and the contralateral twists, and b) the rotational angle from the ipsilateral twist to the contralateral twist – compared to healthy knees. A 40% decrease in tibiofemoral rotation at the ipsilateral angle (−4.4° vs. −11°), a 28% decrease in the contralateral angle (2.7° vs. 9.5°), and a 35% decrease in ipsilateral-to-contralateral rotational angle (7.1° vs. 20.4°) was observed. The mean flexion/varus angle was 3.8°/0.7° and 10.1°/6.4° in healthy and OA knees, respectively, with a statistically significant difference. The medial and lateral femoral condyles demonstrated significantly smaller AP translations (57% and 32%) relative to the tibia in the OA knees (4.9 mm and 5.0 mm) than those in the healthy knees (8.6 mm and 15.7 mm).

Even though “looking back” movements are frequently used in daily activities and have the potential to help patients avoid an injury, they have largely not been investigated. The present study is noteworthy because the number of studies^8,12 reporting a decrease in active, rotating movement is more limited than the studies that have identified a decrease in passive or unconscious rotating movement occurring in response to the bending of the knee. With respect to the flexion angle, medial OA knees tended to have larger angles in all phases than those of healthy knees. Fiacchi et al. also reported that medial OA knees, with a KL grade of III or IV, could not reach full extension during activities such as rising from a chair and stair climbing.¹³ Osteophytes located in the femoral intercondylar notch and the anterior tibial articulation, and ligament stiffness in the OA knees could affect extension mechanism leading to a limited range of extension angles under weight-bearing conditions.⁷

The significantly smaller rotational angle from the ipsilateral twist to the contralateral twist in the OA knees is a noteworthy observation. Several studies have described decreased rotation in knees with medial OA during knee flexion-extension activities.^1,7,14,15 Hamai et al. reported that medial OA knees showed significantly smaller rotational angles (11° and 10°) than healthy knees (29° and 22°) for squatting and kneeling, respectively.^7,14 The smaller rotational angle in medial OA knees (6°) compared to healthy knees (10°) was also revealed in studies by Kuroyanagi et al. during stair-up-and-down activity ranging from 10° to 70° knee flexion.¹⁵ The present study is the first to evaluate the tibiofemoral rotation during rotational activity and demonstrated a decrease in the active range of both external and internal rotations.

In the medial and lateral femoral condyles, our findings of significantly smaller AP translations in OA knees compared with the healthy knees, corroborate several previously reported kinematic studies.^7,13,16,17 Varus OA knees were shown to have restricted rotational tibial mobility and AP translation with a limited range of extension during movements concerning daily activities, such as squatting, sitting and standing on the chair due to contraction of ligaments and osteophyte formation.^16,17 Matsuki et al. also reported that the mean AP translations of both the medial and lateral condyles in early OA knees were already significantly smaller than healthy knees during pivot activity.⁸

In this study, the ACL-deficient OA knees showed significantly more posterior translation at the neutral phase and a tendency for more posterior translation at other phases, in contrast to the ACL-intact/attenuated OA knees. Meanwhile, late degenerative ACL deficiency with the progression of varus knee OA demonstrated no axial rotational instability with a small amount of rotation during twisting activity. The results of this present study are supported by those in a previous study by Yamaguchi et al.¹² but differ from those reported by Nakamura et al.¹⁸ Yamaguchi et al. reported that the femur of the ACL-deficient knee was significantly more posterior than the contralateral healthy knee during tibial neutral to internal rotation in pivot activity.¹² Nakamura et al. reported that rotational instability was observed during “giving way position”.¹⁸ However, these previous studies examined non-OA knees in young athletic patients. Severe deformity of the tibial surface, osteophytes, cartilage-bone erosion in the medial compartment, flexion contracture, and the stiffness of soft tissue such as ligament, capsule, and muscle due to OA might decrease rotational instability and restrict AP translation of the femur relative to the tibia.⁸

The outcomes of this study are impacted by certain limitations. First, the healthy knees cohort includes only young male subjects. Age and gender may affect knee kinematics^19,20; therefore, the results in healthy knees may not represent those from elder and female subjects. A follow-up study evaluating the effects of age and gender is warranted. Second, we didn't evaluate the effect of status of spine or hip health on the results of the present study. Twist activity represents the combinational movement of lower extremities and body trunk; hence spinal and hip conditions could affect the twisting function of the knee.²¹ However, none of the patients in the study cohort had symptomatic lumbar or hip diseases.

5 Conclusion

The in vivo kinematic analysis of OA knees during twisting activity revealed significantly smaller rotation and AP movement compared to those of healthy knees. These findings can be useful for evaluating surgical procedures that allow a more physiological restoration of knee joint function.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author contribution

Masato Kiyohara: Data curation, Writing - original draft, Formal analysis, Writing - review & editing. Satoshi Hamai: Data curation, Writing - original draft, Formal analysis, Writing - review & editing. Hirotaka Gond: Data curation, Formal analysis, Writing - review & editing. Hidehiko Higaki: Data curation, Formal analysis, Writing - review & editing. Satoru Ikebe: Data curation, Formal analysis, Writing - review & editing. Tetsuro Ush: Data curation, Formal analysis, Writing - review & editing. Koji Murakami: Data curation, Formal analysis, Writing - review & editing. Yasuharu Nakashima: Formal analysis, Writing - review & editing.