Comparison of screw-home mechanism in the unloaded living knee subjected to active and passive movements

Abstract

BACKGROUND:

The screw-home mechanism (SHM) plays an important role in the stability of the knee. Accordingly, the analysis of tibial rotation patterns can be used to elucidate the effect of SHM-related factors.

OBJECTIVE:

The purpose of this study was to compare the magnitude of the angle and the pattern of SHM between passive and active movements.

METHODS:

We studied twenty healthy males, of which the angle of knee flexion-extension and tibial longitudinal rotation (TLR) during active and passive movements were measured using the inertial measurement unit. Student’s $t$ -tests were used to compare the magnitude of TLR. The waveform similarity was quantified using a coefficient of multiple correlation (CMC).

RESULTS:

Significant differences were found in the TLR between the active and passive movements ( $p<$ 0.05). The knee flexion-extension waveform similarity was excellent (CMC $=$ 0.956). However, the waveform similarity of TLR was weak (CMC $=$ 0.629).

CONCLUSION:

The SHM increased abruptly during the last 20 ${}^{\circ}$ of the active (extension) movement compared with passive extension. The SHM occurred mainly owing to the geometry and shape of the articular surfaces of the knee joint. In addition, muscle contraction was considered to be an important factor in the articulation movement.

Keywords

Screw-home movement inertial measurement unit inertial sensor knee stability tibia longitudinal rotation

1. Introduction

Biomechanical instability of the knee joint leads to tissue damage and severe pain that can be prevented via appropriate coordination of knee-related factors [1, 2]. In other words, the joint is injured by moving beyond the threshold of stability. Although the knee joint is a structurally modified hinge joint, it is accompanied by rotational motion owing to the biomechanical coordination among knee-related factors. This kinematic phenomenon is referred to as the screw-home mechanism (SHM), and plays an important role in the stability of the knee [3, 4, 5, 6, 7, 8].

Two different arguments were proposed to explain the SHM mechanism [3, 9]. The first involved knee movements along a consistent path, and the second considered various movements in an envelope. The controversy was attributed to the tibial longitudinal rotation (TLR) across a range of magnitudes and periods during the knee flexion-extension direction, although the external rotation was accompanied by extension [3]. A previous study reported that there was no difference in the absolute value of rotation, abduction/adduction, and articular movement during passive flexion of the knee compared with passive extension using unloaded postmortem human knee specimens [9]. This result suggests that the coupled movements involving knee flexion and extension follow a consistent path. However, the path and coupled movements are affected by gravity, force of muscle activation, weight-bearing, and loaded conditions [5, 9, 10, 11, 12, 13]. In particular, SHM has been defined as the tibial rotation involving the last angle of knee extension. In contrast, several published studies indicated that the TLR occurred continually over a wider range during passive knee movements [3, 9, 13, 14]. It was insufficient to interpret the influence of SHM-related factors because the results of tibial rotation of the knee joint varied according to the condition.

An initial study of the SHM was conducted using autopsy specimens [4]. Subsequently, the TLR was investigated with the actuator to reproduce muscle contraction around the knee [11]. The SHM affected the muscle action in the knee kinematics of autopsy specimens. The SHM was measured with the use of three-dimensional magnetic resonance imaging (MRI) at specific angles in live subjects that provided high-resolution outcomes that facilitated the analysis of structural elements related to SHM [15]. In particular, Heller et al. reported longitudinal rotation at the flexion angles of 0 ${}^{\circ}$ , 30 ${}^{\circ}$ , and 90 ${}^{\circ}$ subjected to both passive and active conditions using human knees in vivo [16]. However, it was difficult to analyze the trajectory of SHM using this method during dynamic movements [5]. An inertial measurement unit (IMU) was used to measure the patterns of motion during functional activity [7, 17, 18, 19, 20]. Recently, the IMU that is a nine-axis sensor that consists of an accelerometer, gyroscope, and magnetometer, was combined with microelectromechanical systems (MEMS) to increase the reliability and validity of dynamic measurement conditions [18, 19, 20]. Previous studies related to SHM focused on the differences in the absolute value of tibial rotation [3, 5, 9, 10, 11, 12, 13, 14]. Based on the results of these studies, only the absolute variances of tibial rotation at each condition – but not the movement pattern – were explained.

Analysis of patterns of changes combined with tibial rotation can be used to elucidate the effect of SHM-related factors. In the current study, we compared the absolute displacement and the pattern of SHM between passive and active movements in non-weighted bearing conditions with the use of MEMS IMU. The working hypothesis is that the magnitude and timing of TLR may vary between passive and active movements.

2. Methods

2.1 Subjects

This study was performed according to the protocol approved by the Institutional Review Board at the Sun Moon University (SM-201703-011-1). Twenty healthy young adults (20 males; age, 22.43 $\pm$ 1.98 years; weight, 75.45 $\pm$ 5.21 kg; height, 174.82 $\pm$ 5.12 cm; body mass index, 24.74 $\pm$ 2.10) were recruited by advertisements. The participants understood the purpose of the study and provided written, informed consent before enrollment. None of the subjects participated in other exercise programs during the last six months prior to the study’s onset. Exclusion criteria included conditions that prevented participation in the study, such as a history of surgery and injuries in the lower extremities, musculoskeletal disorders, neurological disorders, and pain in the knee joint.

2.2 Test procedure

Each subject sat on a chair and performed active and passive knee flexion-extension-flexion motions without bearing any weight on the tested knee. In the active movement trial, subjects were asked to extend knees from the flexed to the fully extended positions for 2 s and repositioned them to the flexed position for 2 s with the assistance of a metronome. The passive movement trial was also conducted in the same procedure as the active trial, but the knee movement was attributed to the manual force applied by the physical therapist. After confirming that the subject’s muscles were relaxed, the therapist hung the sling on the heel (calcaneus) pulled it to a fully extended position, and then allowed gravity to return it to the flexed position.

2.3 Data acquisition and analysis

Data acquisition and analysis were performed based on procedures reported earlier [21]. Angles of knee flexion-extension and TLR during active and passive movements were measured using the MEMS IMU (EBIMU24GV3, E2BOX, Seoul, Republic of Korea) at 40 Hz. The MEMS IMU consisted of a three-axis accelerometer ( $\pm$ 8 g), a magnetometer ( $\pm$ 16 Gs), and a gyroscope ( $\pm$ 2000 ${}^{\circ}$ /s). The MEMS IMU collected data from the sensors and the Kalman filter fusion algorithm was used to estimate orientation. The sensor units were attached to the lateral side of the thigh and to the proximal medial surface of the tibial shaft to construct thigh and shank segments, respectively (Fig. 1).

Figure 1.

Attachment of MEMS IMU sensors and test procedure (A: flexion and B: full extension position) during the movement trial of the knee joint.

The kneejoint angle was estimated as the relative angle between the thigh and shank segments using SMULeg software (SMULeg, XENART, Daejeon, Republic of Korea). Before performing trials of knee movement, the subject was asked to perform flexion-extension motion to determine the physically meaningful axes of the knee joint in the local coordinate system of the sensors [22, 23]. The fully extended knee (achieved manually by the examiner) was considered as the zero position. The longitudinal rotation of the tibia was shifted to a zero angle during knee flexion, which was the starting position in the examination. Positive angles indicate flexion and external rotation, respectively. Both passive and active movements (flexionextensionflexion) were time normalized to the range of %100% with a visual three-dimensional (3D) software (C Motion Inc., Germantown, MD, USA).

Each angle of the knee flexionextension and TLR were calculated by subtracting the angle in the flexed position from the angle in a fully extended position. The timing of the peak angle (TPA) was defined as the instant at which the maximum knee extension and tibial external rotation angles were achieved. The coefficient of multiple correlation (CMC) was used to assess the waveform similarity in both movements [18]. The equation for obtaining CMC is as follows:

$\displaystyle\text{CMC}=\sqrt{1-\frac{\begin{array}[]{c}\sum_{k=1}^{K}[\sum_{p% =1}^{P}\sum_{f=1}^{F}(\theta_{kp}(f)\\ -\bar{\theta}_{kf})^{2}/KF_{k}(P-1)]\\ \end{array}}{\begin{array}[]{c}\sum_{k=1}^{K}[\sum_{p=1}^{P}\sum_{f=1}^{F}(% \theta_{kp}(f)\\ -\bar{\theta}_{k})^{2}/K(PF_{k}-1)]\\ \end{array}}}$ (1)

where ${\theta}_{\textit{kpf}}$ is the angle of the frame $f$ measured by the movement $p$ (active or passive) in the knee joint motion $k$ , and $\bar{\theta}_{kf}$ denotes the mean angle of both movements at frame $f$ .

$\displaystyle\bar{\theta}_{kf}=\frac{1}{p}\sum_{p=1}^{2}\theta_{\textit{kpf}}$ (2)

where $\bar{\theta}_{k}$ represents the grand mean of the knee motion $k$ of the two movements.

$\displaystyle\bar{\theta}=\frac{1}{2F}\sum_{p=1}^{2}\sum_{f=1}^{F}Y_{\textit{% kpf}}$ (3)

where $P$ and $F$ represent the number of movements ( $P=$ 2) and the total number of frames ( $F=$ 101), respectively. $K$ represents the number of knee motions in a trial, which was a single motion in all of our study trials.

2.4 Statistical analysis

The independent $t$ -test was used to compare the TPAs according to the active and passive movements of the knee joint. The similarity of the joint angle waveforms between active and passive movements was quantified using CMC. The criteria for the CMC values were as follows: excellent, $>$ 0.95; very good, 0.850.95; good, 0.750.85; moderate, 0.650.75; weak, $<$ 0.65 [24]. Statistical analysis was performed using the SPSS software for Windows (version 22.0 IMB, Armonk, NY, USA). $P$ values $<$ 0.05 were considered statistically significant.

Table 1
Comparison of the angle and TPA of knee flexion-extension and TLR according to active and passive movements of the knee

Measurement	Movement of the knee joint		$p$
	Active	Passive
Flexion-extension
Angle ( ${}^{\circ}$ )	77.21 $\pm$ 6.73	74.83 $\pm$ 6.49	0.263
TPA (%)	49.35 $\pm$ 6.27	51.6 $\pm$ 5.14	0.222
Longitudinal rotation of tibia
Angle ( ${}^{\circ}$ )	18.98 $\pm$ 7.17	13.07 $\pm$ 6.32	0.009
TPA (%)	50.4 $\pm$ 5.07	56.8 $\pm$ 6.73	0.025

All results are expressed as mean $\pm$ standard deviation. TPA: timing of peak angle, TLR: tibial longitudinal rotation.

Figure 2.

Comparison of knee flexion-extension angles according to active and passive movements of knee flexion-extension-flexion. Dashed and solid lines represent active and passive movements, respectively. The positive values represent knee flexion. Waveform similarity (coefficient of maximum correlation (CMC) $=$ 0.956) and the timing of peak angle (TPA) extension (active $=$ 49.35% and passive $=$ 51.6%) are shown in (a). Means and standard deviations for the two movements are shown in (b) and (c).

Figure 3.

Comparison of tibial longitudinal rotation (TLR) angles between active and passive movements (knee flexion-extension-flexion motion). Dashed and solid lines represent active and passive movements, respectively. The positive values represent tibial external rotation. Waveform similarity (CMC $=$ 0.629) and the timing of peak external rotation angle (active $=$ 50.4% and passive $=$ 56.8%) are shown in (a). Mean and standard deviations for the two movements are shown in (b) and (c).

Figure 4.

Movement path of TLR angle according to the knee flexion-extension angle during flexion-extension-flexion motion. Dashed and solid lines represent active and passive movements, respectively. The positive values represent knee flexion and external rotation of the tibia.

3. Results

The angles and the TPA according to active and passive movements are shown in Table 1. The magnitude of TLR was 18.98 $\pm$ 7.17 for active and 13.07 $\pm$ 6.32 for passive movements. The values of TPA were 50.4% $\pm$ 5.07% and 56.8% $\pm$ 6.73% for active and passive movements, respectively. Significant differences were found in the magnitudes of TLR ( $p<$ 0.05) and the TPAs ( $p<$ 0.05) between the two movements (Table 1). The magnitude of the knee flexionextension was 77.21 $\pm$ 6.73 for active movements and 74.83 $\pm$ 6.49 for passive movements. The TPAs were 49.35% $\pm$ 6.27% and 51.6% $\pm$ 5.14% for active and passive movements, respectively. There were no statistically significant differences in the angle of knee flexionextension ( $p>$ 0.05) and the TPA ( $p>$ 0.05) between the two movements (Table 1).

Figures 2 and 3 present the CMC results, and the mean and standard deviation of flexion-extension and TLR angle for the two movements, respectively. The knee flexionextension waveform similarity was excellent (CMC $=$ 0.956). However, the waveform similarity of TLR was weak (CMC $=$ 0.629). Figure 4 shows the movement of the TLR angle according to the knee flexionextension angle during flexionextensionflexion motions in the two movements.

4. Discussion

The present study was conducted using MEMS IMU to demonstrate the changes in TLR by active and passive movements. Recently, the nine-axis inertial sensors with a microcontroller were miniaturized and the increased reliability of the biomechanical data acquisition was confirmed [19, 25]. To analyze the SHM mechanism, we focused on the relative comparison and pattern rather than on the absolute displacement, because the angle of TLR showed a large variation in each of the studies reported previously depending on the equipment and experimental conditions [3, 5, 9, 10, 11, 12, 13, 14, 26]. Furthermore, the tibial rotation values varied for each subject in a rig design study that used postmortem specimens [9]. In this study, the similarity of the flexionextension waveforms demonstrated excellent CMC values between active and passive movements and confirmed the overall similarity of joint angle waveforms acquired at different experimental conditions [24]. Therefore, it appears that the flexionextension angle was synchronized between active and passive conditions. In contrast, the CMC value of longitudinal rotation showed poor correlation between the two conditions. The longitudinal rotation of the active motion (18.98 ${}^{\circ}$ $\pm$ 7.17 ${}^{\circ}$ ) was increased compared with the passive movement (13.07 ${}^{\circ}$ $\pm$ 6.32 ${}^{\circ}$ ) in the full extension of the knee. In particular, the SHM of the active movement was increased abruptly during the last 20 ${}^{\circ}$ of knee extension.

In this study, the TLR of the passive movement was increased gradually in both flexion and extension and ranged between 0 ${}^{\circ}$ and 70 ${}^{\circ}$ . This was in contrast with previous studies that suggested that the SHM occurred in the last angle of knee extension. However, the SHM pattern was similar to several previous studies in which the lower leg was manually moved to a non-weighted bearing or unloaded position [9, 11, 12, 27]. In particular, Johal et al. used MRI to study the change of TLR by comparing non-weight bearing and weighted bearing knees [12]. The tibial rotation of the non-weight bearing knees occurred at 14 ${}^{\circ}$ during flexion from 0 ${}^{\circ}$ to 120 ${}^{\circ}$ , mainly after 45 ${}^{\circ}$ flexion. In the three studies that involved movement acquisitions of the femoral condyles during knee flexion, the backward movement of the lateral femoral condyle was increased to compare with the medial femoral condyle during the entire knee flexion process [12, 27, 28]. Therefore, the knee joint was gradually moved into a stable position or was finally locked because of the geometry and shape of articular surfaces.

In the current study, the TLR of the active movement showed a statistically significant difference from the passive movement. The magnitude of TLR increased, and the timing of TPA was rapid during active knee extension. The dominant muscle in our experiment was the quadriceps during the knee movement because of the force and direction of gravity in the sitting position. In the two previous studies, the force related to knee muscles was reproduced with the use of an actuator on the knee specimens from postmortem humans, and the TLR was reduced relative to the loading of the knee [11, 29]. The load of anterior cruciate ligament (ACL) was increased in the active trial [16] because the force of quadriceps muscle pulled the tibia forward and ensured that the angular orientation ranged between 0 ${}^{\circ}$ and 60 ${}^{\circ}$ [11]. The biomechanical function of the ACL limited the excessive anterior tibial translation and TLR during knee movements [30]. In the clinical evaluation (anterior drawer and Lachman tests), the tibia moved anteriorly but also rotated internally [19]. Therefore, the increased SHM during the active movement might be insufficient to directly explain the effect of muscle contraction and increased ACL tension.

Two previous studies reported that the in situ force in the ACL decreased at 0 ${}^{\circ}$ after reached the maximum intensity during knee flexion for angular changes from 30 ${}^{\circ}$ to 15 ${}^{\circ}$ [30, 31]. The in situ force was defined as the force transmitted by the ligament in response to the load applied to the knee. Therefore, the diminished in situ force in the last extension period was attributed to the articulation movement. The increased TLR during active movement shown in this study can be explained by the following two processes. First, the ascended ACL tension increases the anterior tibial displacement and limits the external rotation up to approximately 20 ${}^{\circ}$ . Second, the knee joint reaches a stable or final locked position under the influence of the geometry and shape of the articular surface in the last extension period. In this process, the tension of ACL decreases owing to the diminished muscle activity and the result of articulation movement. Finally, the TLR sharply increases in the last 20 ${}^{\circ}$ of knee extension.

In conclusion, we demonstrated the changes in magnitude and pattern of SHM between passive and active movements in unconstrained knees in vivo using MEMS IMU. The TLR rapidly increased from the last 20 ${}^{\circ}$ of the knee extension during active movement. According to a previous study, the magnitude of tibial rotation was influenced by the direction of the surrounding soft tissues, including the collateral ligaments and capsules, directions of the cruciate ligaments, the geometry of the condyles, and the shape of the menisci [32]. Since the 1970s, the SHM has been attributed to the larger medial femoral condyle and medial meniscus compared with the lateral condyle and meniscus [33]. We believe that the SHM occurred mainly owing to the geometry and shape of the articular surfaces of the knee joint. In addition, muscle contraction of the knee joint was considered to be an important factor in the articulation movement.

The following study limitations should be considered during the interpretation of the results. First, this study investigated biomechanical data based on in vivo IMU. Although biomechanical data using skin markers were affected by inaccuracies of skin movement, this study minimized the effect of bias by changing only the conditions of movement. Second, to minimize the diversity of dependent variables, we measured the range within 90 ${}^{\circ}$ in the sitting position because flexions greater than 90 ${}^{\circ}$ might have altered the direction of resistance based on gravity. Third, this study was a relatively small sample size. Additional studies are needed in a larger sample size to delineate the biomechanical changes associated with knee injury. We found that the magnitude and timing of TLR differs between passive and active movements. Clinically, the results of this study provide a scientific basis for the accompanying gradual TLR from 70 ${}^{\circ}$ to fully extended position when physical therapists passively apply stretching to the knee. In active muscle strengthening exercises, it would be necessary to develop an exercise that considers the accompanying TLR in the last range.

Footnotes

Acknowledgments

This work was supported by the Sun Moon University Research Grant of 2019.

Conflict of interest

The authors have no conflict of interest to report.

References

Madeti

Chalamalasetti

. Biomechanics of knee joint – a review. Frontiers of Mechanical Engineering. 2015; 10(2): 176–186. doi: 10.1007/s11465-014-0306-x.

Woo

SL-Y

Abramowitch

Kilger

Liang

. Biomechanics of knee ligaments: injury, healing, and repair. Journal of Biomechanics. 2006; 39(1): 1–20. doi: 10.1016/j.jbiomech.2004.10.025.

Blankevoort

Huiskes

de Lange

. The envelope of passive knee joint motion. Journal of Biomechanics. 1988; 21(9): 705–720. doi: 10.1016/0021-9290(88)90280-1.

Hallen

Lindahl

. The “screw-home” movement in the knee-joint. Acta Orthopaedica Scandinavica. 1966; 37(1): 97–106. doi: 10.3109/17453676608989407.

Kim

Yang

Jeung

Choi

Choy

. Screw-home movement of the tibiofemoral joint during normal gait: three-dimensional analysis. Clinics in Orthopedic Surgery. 2015; 7(3): 303–309. doi: 10.4055/cios.2015.7.3.303.

Moglo

Shirazi-Adl

. Cruciate coupling and screw-home mechanism in passive knee joint during extension-flexion. Journal of Biomechanics. 2005; 38(5): 1075–1083. doi: 10.1016/j.jbiomech.2004.05.033.

Piazza

Cavanagh

. Measurement of the screw-home motion of the knee is sensitive to errors in axis alignment. Journal of Biomechanics. 2000; 33(8): 1029–1034. doi: 10.1016/s0021-9290(00)00056-7.

Zhang

Wang

Niu

Liu

Wang

. Relationship between patellar tracking and the “screw-home” mechanism of tibiofemoral joint. Orthopaedic Surgery. 2016; 8(4): 490–495. doi: 10.1111/os.12295.

Wilson

Feikes

Zavatsky

O’Connor

. The components of passive knee movement are coupled to flexion angle. Journal of Biomechanics. 2000; 33(4): 465–473. doi: 10.1016/s0021-9290(99)00206-7.

10.

Kaiser

Bradford

Johnson

Wieben

Thelen

. Measurement of tibiofemoral kinematics using highly accelerated 3D radial sampling. Magnetic Resonance in Medicine. 2013; 69(5): 1310–1316. doi: 10.1002/mrm.24362.

11.

Victor

Labey

Wong

Innocenti

Bellemans

. The influence of muscle load on tibiofemoral knee kinematics. Journal of Orthopaedic Research. 2010; 28(4): 419–428. doi: 10.1002/jor.21019.

12.

Johal

Williams

Wragg

Hunt

Gedroyc

. Tibio-femoral movement in the living knee. A study of weight bearing and non-weight bearing knee kinematics using ‘interventional’ MRI. Journal of Biomechanics. 2005; 38(2): 269–276. doi: 10.1016/j.jbiomech.2004.02.008.

13.

Tsai

Kuo

Hsu

Chen

. In vivo three-dimensional kinematics of the normal knee during active extension under unloaded and loaded conditions using single-plane fluoroscopy. Medical Engineering & Physics. 2008; 30(8): 1004–1012. doi: 10.1016/j.medengphy.2008.03.001.

14.

Hacker

Ignatius

Durselen

. The influence of the test setup on knee joint kinematics – a meta-analysis of tibial rotation. Journal of Biomechanics. 2016; 49(13): 2982–2988. doi: 10.1016/j.jbiomech.2016.07.025.

15.

Samukawa

Yamamoto

Miyamoto

Yamaguchi

Katayose

. Analysis of tibial rotation using magnetic resonance imaging. Manual Therapy. 2009; 14(6): 712–715. doi: 10.1016/j.math.2009.01.003.

16.

Heller

Konig

Graichen

Hinterwimmer

Ehrig

Duda

, et al. A new model to predict in vivo human knee kinematics under physiological-like muscle activation. Journal of Biomechanics. 2007; 40(Suppl 1): S45–53. doi: 10.1016/j.jbiomech.2007.03.005.

17.

Brennan

Deluzio

. Assessment of anatomical frame variation effect on joint angles: a linear perturbation approach. Journal of Biomechanics. 2011; 44(16): 2838–2842. doi: 10.1016/j.jbiomech.2011.09.006.

18.

Zhang

J-T

Novak

Brouwer

. Concurrent validation of Xsens MVN measurement of lower limb joint angular kinematics. Physiological Measurement. 2013; 34(8): N63. doi: 10.1088/0967-3334/34/8/N63.

19.

Al-Amri

Nicholas

Button

Sparkes

Sheeran

Davies

. Inertial measurement units for clinical movement analysis: reliability and concurrent validity. Sensors (Basel). 2018; 18(3). doi: 10.3390/s18030719.

20.

Robert-Lachaine

Mecheri

Larue

Plamondon

. Validation of inertial measurement units with an optoelectronic system for whole-body motion analysis. Medical & Biological Engineering & Computing. 2017; 55(4): 609–619. doi: 10.1007/s11517-016-1537-2.

21.

Jeon

Lee

Kim

Hong

. Detection and quantification of screw-home movement using nine-axis inertial sensors. The Journal of Korean Physical Therapy. 2019; 31(6): 333–338. doi: 10.18857/jkpt.2019.31.6.333.

22.

Seel

Raisch

Schauer

. IMU-based joint angle measurement for gait analysis. Sensors (Basel). 2014; 14(4): 6891–6909. doi: 10.3390/s140406891.

23.

Takeda

Tadano

Natorigawa

Todoh

Yoshinari

. Gait posture estimation using wearable acceleration and gyro sensors. Journal of Biomechanics. 2009; 42(15): 2486–2494. doi: 10.1016/j.jbiomech.2009.07.016.

24.

Ferrari

Cutti

Cappello

. A new formulation of the coefficient of multiple correlation to assess the similarity of waveforms measured synchronously by different motion analysis protocols. Gait Posture. 2010; 31(4): 540–542. doi: 10.1016/j.gaitpost.2010.02.009.

25.

Brigante

Abbate

Basile

Faulisi

Sessa

. Towards miniaturization of a MEMS-based wearable motion capture system. IEEE Transactions on Industrial Electronics. 2011; 58(8): 3234–3241. doi: 10.1109/TIE.2011.2148671.

26.

Smith

Refshauge

Scarvell

. Development of the concepts of knee kinematics. Archives of Physical Medicine and Rehabilitation. 2003; 84(12): 1895–1902. doi: 10.1016/S0003-9993(03)00281-8.

27.

Most

Zayontz

Otterberg

Sabbag

Rubash

. Femoral rollback after cruciate-retaining and stabilizing total knee arthroplasty. Clinical Orthopaedics and Related Research. 2003; (410): 101–113. doi: 10.1097/01.blo.0000062380.79828.2e.

28.

Iwaki

Pinskerova

Freeman

. Tibiofemoral movement 1: the shapes and relative movements of the femur and tibia in the unloaded cadaver knee. The Journal of Bone and Joint Surgery British Volume. 2000; 82(8): 1189–1195. doi: 10.1302/0301-620X.82B8.0821189.

29.

Ferrari

Wilson

Hayes

. The effect of release of the popliteus and quadriceps force on rotation of the knee. Clinical Orthopaedics and Related Research. 2003; (412): 225–233. doi: 10.1097/01.blo.0000071752.41516.6e.

30.

Yagi

Wong

Kanamori

Debski

Woo

. Biomechanical analysis of an anatomic anterior cruciate ligament reconstruction. The American Journal of Sports Medicine. 2002; 30(5): 660–666. doi: 10.1177/03635465020300050501.

31.

Woo

Fox

Sakane

Livesay

Rudy

. Biomechanics of the ACL: measurements of in situ force in the ACL and knee kinematics. The Knee. 1998; 5(4): 267–288. doi: 10.1016/S0968-0160(98)00014-3.

32.

Matsumoto

Seedhom

Suda

Otani

Fujikawa

. Axis location of tibial rotation and its change with flexion angle. Clinical Orthopaedics and Related Research. 2000; (371): 178–182. doi: 10.1097/00003086-200002000-00022.

33.

Goodfellow