Advertisement

ISB clinical biomechanics award winner 2021: Tibio-femoral kinematics of natural versus replaced knees – A comparison using dynamic videofluoroscopy

Open AccessPublished:May 19, 2022DOI:https://doi.org/10.1016/j.clinbiomech.2022.105667

      Highlights

      • Each total knee replacement design exhibited characteristic motion patterns.
      • GMK Sphere closest replicated the medial centre of rotation found in natural knees.
      • No implant was able to replicate all aspects of natural tibio-femoral kinematics.

      Abstract

      Background

      A comparison of natural versus replaced tibio-femoral kinematics in vivo during challenging activities of daily living can help provide a detailed understanding of the mechanisms leading to unsatisfactory results and lay the foundations for personalised implant selection and surgical implantation, but also enhance further development of implant designs towards restoring physiological knee function. The aim of this study was to directly compare in vivo tibio-femoral kinematics in natural versus replaced knees throughout complete cycles of different gait activities using dynamic videofluoroscopy.

      Methods

      Twenty-seven healthy and 30 total knee replacement subjects (GMK Sphere, GMK PS, GMK UC) were assessed during multiple complete gait cycles of level walking, downhill walking, and stair descent using dynamic videofluoroscopy. Following 2D/3D registration, tibio-femoral rotations, condylar antero-posterior translations, and the location of the centre of rotation were compared.

      Findings

      The total knee replacement groups predominantly experienced reduced tibial internal/external rotation and altered medial and lateral condylar antero-posterior translations compared to natural knees. An average medial centre of rotation was found for the natural and GMK sphere groups in all three activities, whereas the GMK PS and UC groups experienced a more central to lateral centre of rotation.

      Interpretation

      Each total knee replacement design exhibited characteristic motion patterns, with the GMK Sphere most closely replicating the medial centre of rotation found for natural knees. Despite substantial similarities between the subject groups, none of the implant geometries was able to replicate all aspects of natural tibio-femoral kinematics, indicating that different implant geometries might best address individual functional needs.

      Keywords

      1. Introduction

      Total knee replacement (TKR) has become a standard procedure for patients suffering from late-stage osteoarthritis. By replacing the joint surfaces of the femur and tibia, TKR surgeries aim to relieve pain and restore natural functionality to the knee joint. In addition, different geometrical features of the implant design have been introduced to guide joint movement and/or mimic physiological tibio-femoral kinematics.
      Despite the overall high success rates of TKR surgeries, some 15–20% of all primary TKR patients are not satisfied with the overall outcome (
      • Bourne R.B.
      • Chesworth B.M.
      • Davis A.M.
      • Mahomed N.N.
      • Charron K.D.
      Patient satisfaction after total knee arthroplasty: who is satisfied and who is not?.
      ;
      • Dunbar M.J.
      • Richardson G.
      • Robertsson O.
      I can't get no satisfaction after my total knee replacement: rhymes and reasons.
      ;
      • Robertsson O.
      • Dunbar M.
      • Pehrsson T.
      • Knutson K.
      • Lidgren L.
      Patient satisfaction after knee arthroplasty: a report on 27,372 knees operated on between 1981 and 1995 in Sweden.
      ). Moreover, up to 30% of all patients are not satisfied with the perceived implant function during specific activities of daily living, especially when high flexion angles are required (
      • Bourne R.B.
      • Chesworth B.M.
      • Davis A.M.
      • Mahomed N.N.
      • Charron K.D.
      Patient satisfaction after total knee arthroplasty: who is satisfied and who is not?.
      ).
      Although natural tibio-femoral kinematics have been investigated during various functional activities using e.g. optical motion capture techniques, such approaches are subject to soft-tissue artefact (
      • Andriacchi T.P.
      • Alexander E.J.
      • Toney M.K.
      • Dyrby C.
      • Sum J.
      A point cluster method for in vivo motion analysis: applied to a study of knee kinematics.
      ;
      • Cappozzo A.
      • Croce U.D.
      • Leardini A.
      • Chiari L.
      Human movement analysis using stereophotogrammetry. Part 1: theoretical background.
      ;
      • Taylor W.R.
      • Ehrig R.M.
      • Duda G.N.
      • Schell H.
      • Seebeck P.
      • Heller M.O.
      On the influence of soft tissue coverage in the determination of bone kinematics using skin markers.
      ), and are therefore generally limited in their ability to accurately capture all tibio-femoral rotations and condylar translations. More invasive imaging techniques now enable the measurement of skeletal movement patterns, but these generally possess only a small field of view, therefore limiting the activities that can be captured (
      • Anderst W.
      • Zauel R.
      • Bishop J.
      • Demps E.
      • Tashman S.
      Validation of three-dimensional model-based tibio-femoral tracking during running.
      ;
      • Dennis D.
      • Komistek R.
      • Scuderi G.
      • Argenson J.N.
      • Insall J.
      • Mahfouz M.
      • Aubaniac J.M.
      • Haas B.
      In vivo three-dimensional determination of kinematics for subjects with a normal knee or a unicompartmental or total knee replacement.
      ;
      • Galvin C.R.
      • Perriman D.M.
      • Lynch J.T.
      • Pickering M.R.
      • Newman P.
      • Smith P.N.
      • Scarvell J.M.
      Age has a minimal effect on knee kinematics: a cross-sectional 3D/2D image-registration study of kneeling.
      ;
      • Grieco T.F.
      • Sharma A.
      • Komistek R.D.
      • Cates H.E.
      Single versus multiple-radii cruciate-retaining total knee arthroplasty: an in vivo mobile fluoroscopy study.
      ;
      • Kozanek M.
      • Hosseini A.
      • Liu F.
      • Van de Velde S.K.
      • Gill T.J.
      • Rubash H.E.
      • Li G.
      Tibiofemoral kinematics and condylar motion during the stance phase of gait.
      ;
      • Li G.
      • Van de Velde S.K.
      • Bingham J.T.
      Validation of a non-invasive fluoroscopic imaging technique for the measurement of dynamic knee joint motion.
      ;
      • Moewis P.
      • Duda G.N.
      • Jung T.
      • Heller M.O.
      • Boeth H.
      • Kaptein B.
      • Taylor W.R.
      The restoration of passive rotational tibio-femoral laxity after anterior cruciate ligament reconstruction.
      ;
      • Moro-oka T.A.
      • Hamai S.
      • Miura H.
      • Shimoto T.
      • Higaki H.
      • Fregly B.J.
      • Iwamoto Y.
      • Banks S.A.
      Dynamic activity dependence of in vivo normal knee kinematics.
      ). More recently, however, mobile fluoroscopy systems have been introduced, hence allowing access to knee joint motion throughout full cycles of gait activities (
      • Guan S.
      • Gray H.A.
      • Keynejad F.
      • Pandy M.G.
      Mobile biplane X-ray imaging system for measuring 3D dynamic joint motion during overground gait.
      ;
      • List R.
      • Postolka B.
      • Schütz P.
      • Hitz M.
      • Schwilch P.
      • Gerber H.
      • Ferguson S.J.
      • Taylor W.R.
      A moving fluoroscope to capture tibiofemoral kinematics during complete cycles of free level and downhill walking as well as stair descent.
      ). Using such techniques, general agreement has been achieved for tibio-femoral kinematics during repetitive flexion tasks (
      • DeFrate L.E.
      • Sun H.
      • Gill T.J.
      • Rubash H.E.
      • Li G.
      In vivo tibiofemoral contact analysis using 3D MRI-based knee models.
      ;
      • Hamai S.
      • Moro-oka T.A.
      • Dunbar N.J.
      • Miura H.
      • Iwamoto Y.
      • Banks S.A.
      In vivo healthy knee kinematics during dynamic full flexion.
      ;
      • Hill P.F.
      • Vedi V.
      • Williams A.
      • Iwaki H.
      • Pinskerova V.
      • Freeman M.A.
      Tibiofemoral movement 2: the loaded and unloaded living knee studied by MRI.
      ;
      • Moro-oka T.A.
      • Hamai S.
      • Miura H.
      • Shimoto T.
      • Higaki H.
      • Fregly B.J.
      • Iwamoto Y.
      • Banks S.A.
      Dynamic activity dependence of in vivo normal knee kinematics.
      ;
      • Pinskerova V.
      • Johal P.
      • Nakagawa S.
      • Sosna A.
      • Williams A.
      • Gedroyc W.
      • Freeman M.A.
      Does the femur roll-back with flexion?.
      ;
      • Tanifuji O.
      • Sato T.
      • Kobayashi K.
      • Mochizuki T.
      • Koga Y.
      • Yamagiwa H.
      • Omori G.
      • Endo N.
      Three-dimensional in vivo motion analysis of normal knees using single-plane fluoroscopy.
      ). However, results remain controversially discussed regarding tibial internal/external rotation, condylar antero-posterior (A-P) translation, and the resultant centre of rotation (CoR) of the natural knee during dynamic activities such as walking (
      • Dennis D.
      • Komistek R.
      • Scuderi G.
      • Argenson J.N.
      • Insall J.
      • Mahfouz M.
      • Aubaniac J.M.
      • Haas B.
      In vivo three-dimensional determination of kinematics for subjects with a normal knee or a unicompartmental or total knee replacement.
      ;
      • Gray H.A.
      • Guan S.
      • Thomeer L.T.
      • Schache A.G.
      • de Steiger R.
      • Pandy M.G.
      Three-dimensional motion of the knee-joint complex during normal walking revealed by mobile biplane X-ray imaging.
      ;
      • Komistek R.D.
      • Dennis D.A.
      • Mahfouz M.
      In vivo fluoroscopic analysis of the normal human knee.
      ;
      • Koo Y.J.
      • Koo S.
      Three-dimensional kinematic coupling in the knee during normal walking.
      ;
      • Kozanek M.
      • Hosseini A.
      • Liu F.
      • Van de Velde S.K.
      • Gill T.J.
      • Rubash H.E.
      • Li G.
      Tibiofemoral kinematics and condylar motion during the stance phase of gait.
      ;
      • Liu F.
      • Kozanek M.
      • Hosseini A.
      • Van de Velde S.K.
      • Gill T.J.
      • Rubash H.E.
      • Li G.
      In vivo tibiofemoral cartilage deformation during the stance phase of gait.
      ;
      • Postolka B.
      • Schütz P.
      • Fucentese S.F.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      • Taylor W.R.
      Tibio-femoral kinematics of the healthy knee joint throughout complete cycles of gait activities.
      ). This controversy plausibly originates from different activity executions and analysis approaches, which are known to result in varying interpretations, albeit of the same joint movement patterns (
      • Postolka B.
      • Taylor W.R.
      • Dätwyler K.
      • Heller M.O.
      • List R.
      • Schütz P.
      Interpretations of tibio-femoral kinematics critically depends upon the kinematic analysis approach: A survey and comparison of methodologies.
      ).
      The kinematic behaviour of replaced knees with different TKR designs has also been studied extensively during flexion (
      • Banks S.A.
      • Markovich G.D.
      • Hodge W.A.
      In vivo kinematics of cruciate-retaining and -substituting knee arthroplasties.
      ;
      • Cates H.E.
      • Komistek R.D.
      • Mahfouz M.R.
      • Schmidt M.A.
      • Anderle M.
      In vivo comparison of knee kinematics for subjects having either a posterior stabilized or cruciate retaining high-flexion total knee arthroplasty.
      ;
      • Moro-oka T.A.
      • Muenchinger M.
      • Canciani J.P.
      • Banks S.A.
      Comparing in vivo kinematics of anterior cruciate-retaining and posterior cruciate-retaining total knee arthroplasty.
      ;
      • Pfitzner T.
      • Moewis P.
      • Stein P.
      • Boeth H.
      • Trepczynski A.
      • von Roth P.
      • Duda G.N.
      Modifications of femoral component design in multi-radius total knee arthroplasty lead to higher lateral posterior femoro-tibial translation.
      ;
      • Scott G.
      • Imam M.A.
      • Eifert A.
      • Freeman M.A.R.
      • Pinskerova V.
      • Field R.E.
      • Skinner J.
      • Banks S.A.
      Can a total knee arthroplasty be both rotationally unconstrained and anteroposteriorly stabilised?.
      ;
      • Sharma A.
      • Leszko F.
      • Komistek R.D.
      • Scuderi G.R.
      • Cates Jr., H.E.
      • Liu F.
      In vivo patellofemoral forces in high flexion total knee arthroplasty.
      ;
      • Victor J.
      • Banks S.
      • Bellemans J.
      Kinematics of posterior cruciate ligament-retaining and -substituting total knee arthroplasty.
      ) but also functional gait activities (
      • Dennis D.A.
      • Komistek R.D.
      • Mahfouz M.R.
      In vivo fluoroscopic analysis of fixed-bearing total knee replacements.
      ;
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ;
      • List R.
      • Schütz P.
      • Angst M.
      • Ellenberger L.
      • Dätwyler K.
      • Ferguson S.J.
      • Writing C.
      Videofluoroscopic evaluation of the influence of a gradually reducing femoral radius on joint kinematics during daily activities in total knee arthroplasty.
      ;
      • Schütz P.
      • Postolka B.
      • Gerber H.
      • Ferguson S.J.
      • Taylor W.R.
      • List R.
      Knee implant kinematics are task-dependent.
      ;
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ). Existing studies report that specific implant design features are able to guide joint motion (
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ;
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ;
      • Scott G.
      • Imam M.A.
      • Eifert A.
      • Freeman M.A.R.
      • Pinskerova V.
      • Field R.E.
      • Skinner J.
      • Banks S.A.
      Can a total knee arthroplasty be both rotationally unconstrained and anteroposteriorly stabilised?.
      ), with considerable inter-subject variability presented for less constrained designs (
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ), possibly due to the additional role of surgical implantation parameters and soft-tissue sufficiency. However, only a limited number of studies have directly compared tibio-femoral kinematics during functional gait activities in natural and replaced knees using a consistent measurement set-up, as well as analysis approach (
      • Dennis D.
      • Komistek R.
      • Scuderi G.
      • Argenson J.N.
      • Insall J.
      • Mahfouz M.
      • Aubaniac J.M.
      • Haas B.
      In vivo three-dimensional determination of kinematics for subjects with a normal knee or a unicompartmental or total knee replacement.
      ;
      • Ghirardelli S.
      • Asay J.L.
      • Leonardi E.A.
      • Amoroso T.
      • Andriacchi T.P.
      • Indelli P.F.
      Kinematic comparison between medially congruent and posterior-stabilized third-generation TKA designs.
      ;
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ;
      • Miura K.
      • Ohkoshi Y.
      • Ino T.
      • Ukishiro K.
      • Kawakami K.
      • Suzuki S.
      • Suzuki K.
      • Maeda T.
      Kinematics and center of axial rotation during walking after medial pivot type total knee arthroplasty.
      ). Independent of the implant design, more tibial external rotation has been observed for the TKR subjects compared to natural knees over the course of a gait cycle, especially during late stance and early swing phases (
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ;
      • Miura K.
      • Ohkoshi Y.
      • Ino T.
      • Ukishiro K.
      • Kawakami K.
      • Suzuki S.
      • Suzuki K.
      • Maeda T.
      Kinematics and center of axial rotation during walking after medial pivot type total knee arthroplasty.
      ). Abduction/adduction has been controversially reported in the literature, while equal rotations were found among natural and TKR subjects (
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ), another study found a reduced range of motion (RoM) in TKR subjects (
      • Ghirardelli S.
      • Asay J.L.
      • Leonardi E.A.
      • Amoroso T.
      • Andriacchi T.P.
      • Indelli P.F.
      Kinematic comparison between medially congruent and posterior-stabilized third-generation TKA designs.
      ). In addition, only one study has also compared the location of the CoR. While a medial-stabilised knee implant resulted in a clear medial CoR, which was comparable to a natural cohort, a posterior-stabilised and cruciate-retaining TKR resulted in equal translations for both femoral condyles and a resultant lateral CoR (
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ). However, no study has yet directly compared tibio-femoral kinematics in both natural and TKR subjects during more challenging activities such as downhill walking or stair descent and used the same analysis approach for consistent interpretation of the resulting joint kinematics.
      To increase patient satisfaction and further enhance the success rates of TKR surgeries, it is crucial to identify mechanisms leading to unsatisfactory results. Here, a direct and consistent comparison of TKR against natural tibio-femoral kinematics in vivo during challenging activities of daily living would lay the foundations for personalisation of implant selection and surgical implantation parameters, as well as targeted development of implant design features towards restoring physiological knee function. Therefore, the aim of this study was to directly compare in vivo tibio-femoral kinematics in natural knees with replaced knees including different design concepts throughout complete cycles of level walking, downhill walking and stair descent using dynamic videofluoroscopy.

      2. Methods

      Twenty-seven healthy subjects were recruited into this study to undergo dynamic videofluoroscopic assessment of the knee for comparison against 30 subjects with a unilateral TKR, measured previously (
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ). Both studies were approved by the local ethics committee (KEK-ZH-Nr. 2015-00140 & 2016-00410) and all subjects provided written informed consent prior to participation.

      2.1 Subjects

      All 27 natural subjects were in good health condition and showed no signs of knee disorders (Table 1). Each subject underwent a clinical knee examination by an experienced orthopaedic surgeon, including testing of soft-tissue sufficiency and measurement of the individual hip-knee-ankle (HKA) angle (EOS imagine, Paris, France) (
      • Folinais D.
      • Thelen P.
      • Delin C.
      • Radier C.
      • Catonne Y.
      • Lazennec J.Y.
      Measuring femoral and rotational alignment: EOS system versus computed tomography.
      ;
      • Postolka B.
      • Schütz P.
      • Fucentese S.F.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      • Taylor W.R.
      Tibio-femoral kinematics of the healthy knee joint throughout complete cycles of gait activities.
      ), which varied from 8° valgus to 9° varus. To gain information of each subject's knee geometry, a CT scan (~20 cm proximal/distal of the joint line, 0.5 × 0.5 mm resolution, 1 mm slice thickness) was acquired for generation of subject-specific volumetric bone models using the open-source software MITK-GEM (
      • Pauchard Y.
      • Fitze T.
      • Browarnik D.
      • Eskandari A.
      • Pauchard I.
      • Enns-Bray W.
      • Palsson H.
      • Sigurdsson S.
      • Ferguson S.J.
      • Harris T.B.
      • Gudnason V.
      • Helgason B.
      Interactive graph-cut segmentation for fast creation of finite element models from clinical ct data for hip fracture prediction.
      ).
      Table 1Group characteristics in terms of gender (female (f) / male (m)), age, body mass index (BMI), time post-op, and hip-knee-ankle (HKA) angle. Mean ± standard deviations are presented for each group.
      NaturalGMK SphereGMK PSGMK UC
      (n = 27)(n = 10)(n = 10)(n = 10)
      gender (f/m)14 f / 13 m8 f / 2 m5 f / 5 m7 f / 3 m
      age [years]27.1 ± 10.668.8 ± 9.969.0 ± 6.575.0 ± 5.1
      BMI [kg/m2]21.3 ± 2.225.4 ± 3.727.6 ± 3.525.9 ± 3.2
      post-op [years]1.7 ± 0.73.1 ± 1.63.9 ± 1.5
      HKA [°]1.3 ± 4.40.3 ± 1.9−0.1 ± 1.90.7 ± 0.9
      While detailed information about the TKR subject groups is provided in the original publication (
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ), a brief overview is presented here: Each TKA subject had been implanted previously with either a GMK Sphere (medial-congruent, fixed-bearing; n = 10), a GMK Primary PS (posterior-stabilised, fixed-bearing; n = 10), or a GMK Primary UC (ultra-congruent, mobile-bearing; n = 10) prothesis (all Medacta International, Castel San Pietro, Switzerland) (Table 1). Only subjects with a good clinical outcome (Western Ontario and McMaster Universities Arthritis Index (WOMAC) 0–28, pain Visual Analogue Scale (VAS) ≤2) and with aligned implant components (−3° < HKA < 3°, assessed using frontal full limb x-rays) were included in this study.

      2.2 Videofluoroscopic assessment

      Tibio-femoral kinematics of each subject were assessed during two-legged standing (feet hip-width apart), as well as during at least five complete gait cycles of level walking, downhill walking (10° declined slope), and stair descent (three 0.18 m height steps) using a moving videofluoroscope operated at 25/30 Hz (1 ms shutter time, 1000 × 1000 pixels image resolution) (
      • List R.
      • Postolka B.
      • Schütz P.
      • Hitz M.
      • Schwilch P.
      • Gerber H.
      • Ferguson S.J.
      • Taylor W.R.
      A moving fluoroscope to capture tibiofemoral kinematics during complete cycles of free level and downhill walking as well as stair descent.
      ). All fluoroscopic images were distortion corrected and the optical projection parameters were calculated for each individual session (
      • Foresti M.
      In vivo measurement of total knee joint replacement kinematics and kinetics during stair descent, D-MAVT.
      ).
      The segmented natural bone models or the 3D geometries of the implant components were then registered to the 2D fluoroscopic images using previously published 2D/3D registration software (
      • Burckhardt K.
      • Szekely G.
      • Notzli H.
      • Hodler J.
      • Gerber C.
      Submillimeter measurement of cup migration in clinical standard radiographs.
      ;
      • Postolka B.
      • List R.
      • Thelen B.
      • Schütz P.
      • Taylor W.R.
      • Zheng G.
      Evaluation of an intensity-based algorithm for 2D/3D registration of natural knee videofluoroscopy data.
      ). For the natural knee data, the reported registration accuracy was <1° for all rotations, <0.6 mm for in-plane, and <7.1 mm for out-of-plane translations (
      • Postolka B.
      • List R.
      • Thelen B.
      • Schütz P.
      • Taylor W.R.
      • Zheng G.
      Evaluation of an intensity-based algorithm for 2D/3D registration of natural knee videofluoroscopy data.
      ). For the registration of the metal TKR components, the reported registration errors were <0.25° for all rotations, <0.3 mm for in-plane, and <1.0 mm for out-of-plane translations for a comparable TKR design (
      • Foresti M.
      In vivo measurement of total knee joint replacement kinematics and kinetics during stair descent, D-MAVT.
      ;
      • List R.
      • Foresti M.
      • Gerber H.
      • Goldhahn J.
      • Rippstein P.
      • Stüssi E.
      Three-dimensional kinematics of an unconstrained ankle arthroplasty: a preliminary in vivo videofluoroscopic feasibility study.
      ).

      2.3 Kinematic evaluation

      Ground reaction forces (GRFs) were acquired using five force plates embedded in the floor and two mobile force plates mounted on the ramp and stair (sampling frequency 2000 Hz, Kistler, Winterthur, Switzerland). In addition, a 22 infrared camera system was used to capture the movement of markers attached to the ipsilateral heel and sacrum (sampling frequency 100 Hz, Vicon MX system, Oxford, United Kingdom). All heel-strike and toe-off events were determined using the GRFs with a threshold of 25 N, except for the second heel-strike of downhill walking, which was defined based on the trajectories of the heel marker. Gait velocity was calculated for each gait cycle based on the trajectories of the sacrum marker.
      Local anatomical coordinate systems were established for the natural femur and tibia. The femoral coordinate system used the mean symmetrical axis of rotation approach (SARA) femoral functional flexion axis (
      • Ehrig R.M.
      • Taylor W.R.
      • Duda G.N.
      • Heller M.O.
      A survey of formal methods for determining functional joint axes.
      ) as the primary axis, calculated from three deep knee bending trials covering 15 to 90° of flexion (
      • Postolka B.
      • Schütz P.
      • Fucentese S.F.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      • Taylor W.R.
      Tibio-femoral kinematics of the healthy knee joint throughout complete cycles of gait activities.
      ). The tibial coordinate system used the CT shaft axis as the leading axis, defined using a cylinder fit to the periosteal surface of the proximal tibia. For both components of all three TKR implant geometries, the coordinate systems were defined based on the implant geometry, where the medio-lateral femoral axis was aligned with the geometric centre axis (
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ).
      Relative rotations according to the joint coordinate system approach (
      • Grood E.S.
      • Suntay W.J.
      A joint coordinate system for the clinical description of three-dimensional motions: application to the knee.
      ) were then determined based on the local femoral and tibial coordinate systems. Condylar A-P translations were calculated using the medial and lateral points on the functional flexion axis (natural subjects) or the geometric centre axis (TKR subjects) that intersected with the mid-sagittal plane of each condyle. Based on the femoral medio-lateral axis, a mean CoR in the transverse plane was calculated over the complete gait cycle using the symmetrical centre of rotation estimation (SCoRE) (
      • Ehrig R.M.
      • Taylor W.R.
      • Duda G.N.
      • Heller M.O.
      A survey of formal methods for determining the centre of rotation of ball joints.
      ). To minimise the effect of the individual knee size or implant component size, A-P translation as well as the location of the CoR were normalised to a medium sized femoral implant with a distance of 42.76 mm between the two condylar mid-sagittal planes. All tibio-femoral rotations and condylar translations were normalised to a gait cycle and linearly interpolated to 101 data points to allow interpretation over complete gait cycles.

      2.4 Statistics

      For all parameters, group means (mean across all subjects in each group) as well as subject means (mean across all gait cycles acquired - at least 5) and standard deviations were calculated. Three mixed-model analysis of variance (ANOVA) were performed to test the effect of the subject group on the ranges of tibio-femoral rotation. The rotational RoM was set as the dependent variable while group (four levels: natural, GMK Sphere, GMK PS, GMK UC) and activity (three levels: level walking, downhill walking, stair descent) were set as fixed effects, while subjects were considered a random effect. An additional ANOVA was performed to compare the ranges of condylar A-P translation. The range of A-P translation was set as the dependent variable. The group, activity, and condylar side (two levels: medial, lateral) were used as fixed effects, and the individual subjects as a random effect. Post-hoc comparisons were conducted using a least significant difference approach, with significance levels adjusted for multiple comparisons using Bonferroni correction. All ANOVA's were conducted using the SPSS software suite (SPSS v26, IBM, USA).
      In order to analyse the effect of the groups on the characteristics of the rotational and translational behaviour over the time-series of a complete gait cycle, one-dimensional statistical parametric mapping (SPM) was performed (
      • Pataky T.C.
      • Robinson M.A.
      • Vanrenterghem J.
      Region-of-interest analyses of one-dimensional biomechanical trajectories: bridging 0D and 1D theory, augmenting statistical power.
      ). A total of 15 one-way ANOVAs were performed. The rotation or translation respectively, was set as the dependent variable, while the group was used as the independent variable. If the ANOVA revealed significant differences between the groups, a post-hoc two-sample t-test was performed with significance levels adjusted for multiple comparisons using Bonferroni correction.

      3. Results

      3.1 Two-legged standing

      Both the natural and TKR groups showed an overall neutral abduction/adduction angle during standing (range 0.7° abduction to 0.2° adduction, Table 2). While the TKR groups showed consistent abduction/adduction among all subjects, the natural subjects showed a large inter-subject variability with individual values ranging from 9.6° abduction to 8.7° adduction. The natural group showed a slight tibial external rotation across all subjects, but also high inter-subject variability, with subjects exhibiting a clear tibial externally rotated position (up to 12.8°) while other subjects showed an internally rotated tibia (up to −8.7°). For the GMK Sphere, a consistent location of the medial condyle was found among all subjects, but with both internally as well as externally rotated tibia for different subjects. For the GMK PS and GMK UC however, the location of the medial as well as lateral condyle varied slightly among the individual subjects, but overall the tibia showed a slight internally rotated position (Fig. 1, Table 2).
      Table 2Tibio-femoral rotations and antero-posterior locations of the medial and lateral femoral condyle during standing, as well as the ranges of tibio-femoral rotations and condylar antero-posterior (A-P) translations during level walking, downhill walking and stair descent, each for the complete gait cycle (GC), the loaded stance phase and the unloaded swing phase.
      NaturalGMK SphereGMK PSGMK UC
      Standingflexion/extension [°]−1.8 ± 5.4−10.2 ± 8.4−11.5 ± 8−6.4 ± 5.3
      tibial internal/external rotation [°]2.3 ± 5.9−2.5 ± 5.6−3.1 ± 4.0−4.3 ± 3.6
      abduction/adduction [°]−0.7 ± 3.80.2 ± 0.40.0 ± 0.30.1 ± 0.2
      medial antero (+) / posterior (−) location [mm]−7.2 ± 3.5−4.7 ± 1.3−4.5 ± 2.0−5.5 ± 1.4
      lateral antero (+) / posterior (−) location [mm]−5.5 ± 3.0−6.4 ± 5.0−6.9 ± 1.9−8.7 ± 2.6
      Level walkingflexion/extension [°]complete GC64.6 ± 4.562.7 ± 4.963.5 ± 4.757.2 ± 4.8
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance51.7 ± 4.834.2 ± 6.233.0 ± 7.230.2 ± 4.2
      swing64.2 ± 4.461.9 ± 5.461.8 ± 5.756.4 ± 5.5
      tibial internal/external rotation [°]complete GC14.3 ± 2.111.9 ± 4.2
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      10.5 ± 1.9
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      8.1 ± 2.5
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance12.0 ± 1.97.3 ± 2.87.9 ± 1.36.2 ± 2.8
      swing10.0 ± 2.69.2 ± 3.28.3 ± 1.55.9 ± 1.6
      abduction/adduction [°]complete GC5.8 ± 1.22.8 ± 0.8
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      2.9 ± 0.8
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      2.3 ± 0.6
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance4.8 ± 1.01.8 ± 0.42.2 ± 0.61.6 ± 0.3
      swing4.9 ± 1.22.5 ± 0.72.3 ± 0.72.0 ± 0.7
      medial A-P translation [mm]complete GC7.6 ± 1.5
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      3.5 ± 1.0
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      11.1 ± 2.4
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      7.6 ± 1.4
      stance6.9 ± 1.41.6 ± 0.36.7 ± 2.05.4 ± 1.6
      swing5.4 ± 1.33.3 ± 1.110.3 ± 2.66.7 ± 1.2
      lateral A-P translation [mm]complete GC9.4 ± 2.0
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      9.7 ± 4.0
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      10.6 ± 1.97.8 ± 1.7
      stance7.0 ± 1.95.5 ± 2.57.3 ± 1.65.1 ± 1.6
      swing7.0 ± 1.98.1 ± 2.98.3 ± 2.27.1 ± 1.7
      Downhill walkingflexion/extension [°]complete GC70.4 ± 4.270.0 ± 4.569.9 ± 5.366.1 ± 3.4
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance60.5 ± 4.651.5 ± 6.050.1 ± 8.347.3 ± 4.8
      swing70.3 ± 4.269.9 ± 4.668.7 ± 6.365.6 ± 3.8
      tibial internal/external rotation [°]complete GC14.2 ± 2.611.5 ± 2.7
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      8.9 ± 2.1
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      7.9 ± 2.3
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance11.9 ± 2.26.3 ± 1.76.3 ± 2.05.6 ± 1.9
      swing9.1 ± 2.510.1 ± 2.57.3 ± 1.86.1 ± 1.9
      abduction/adduction [°]complete GC5.7 ± 1.22.6 ± 1.0
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      2.6 ± 0.7
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      2.2 ± 0.7
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance4.7 ± 1.01.5 ± 0.22.1 ± 0.71.5 ± 0.2
      swing4.5 ± 1.32.3 ± 0.92.0 ± 0.61.9 ± 0.7
      medial A-P translation [mm]complete GC7.2 ± 1.6
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      3.0 ± 0.9
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      11.2 ± 2.6
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      8.6 ± 1.0
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      stance6.1 ± 1.71.5 ± 0.36.3 ± 1.84.7 ± 1.3
      swing5.3 ± 1.12.8 ± 1.110.0 ± 3.18.2 ± 1.1
      lateral A-P translation [mm]complete GC9.4 ± 1.8
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      9.0 ± 2.7
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      9.9 ± 2.6
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      7.3 ± 1.3
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      stance7.2 ± 1.84.4 ± 1.45.4 ± 1.33.8 ± 1.2
      swing6.6 ± 1.48.2 ± 2.58.5 ± 2.17.1 ± 1.3
      Stair descentflexion/extension [°]complete GC92.0 ± 4.889.5 ± 5.590.2 ± 5.587.5 ± 4.4
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance83.5 ± 5.377.5 ± 6.478.8 ± 9.276.6 ± 5.6
      swing91.8 ± 4.889.1 ± 5.389.9 ± 5.687.3 ± 4.5
      tibial internal/external rotation [°]complete GC14.5 ± 2.913.2 ± 2.29.0 ± 2.5
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      8.4 ± 3.3
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance10.9 ± 2.56.6 ± 2.46.6 ± 3.16.0 ± 2.5
      swing9.9 ± 3.012.8 ± 2.47.6 ± 1.97.3 ± 3.3
      abduction/adduction [°]complete GC6.5 ± 1.43.0 ± 1.2
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      2.9 ± 0.7
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      2.3 ± 0.7
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      stance5.7 ± 1.52.1 ± 0.82.5 ± 0.72.0 ± 0.5
      swing4.5 ± 1.22.3 ± 1.02.0 ± 0.61.7 ± 0.5
      medial A-P translation [mm]complete GC7.1 ± 1.6
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      3.7 ± 1.5
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      12.5 ± 1.6
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      8.7 ± 1.3
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      stance6.1 ± 1.32.2 ± 1.14.3 ± 1.13.8 ± 1.3
      swing5.9 ± 1.63.4 ± 1.612.1 ± 1.78.3 ± 1.4
      lateral A-P translation [mm]complete GC10.9 ± 2.5
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      11.0 ± 2.1
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      12.1 ± 2.87.6 ± 1.6
      Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      stance7.5 ± 1.44.9 ± 2.05.2 ± 1.23.6 ± 0.7
      swing8.6 ± 2.810.8 ± 2.211.9 ± 2.97.3 ± 1.5
      Mean ± standard deviations are presented for each subject group.
      low asterisk Significant differences to the natural group based on an adjusted level of significance of α = 0.0056.
      Significant difference of medial and lateral condylar translation on an adjusted level of significance of α = 0.0042.
      Fig. 1
      Fig. 1Location of the medial and lateral femoral condyle during standing. Mean locations for each subject group (thick line) as well as individual subject locations (thin line) are indicated.

      3.2 Dynamic gait activities

      The mean gait velocities of the natural knee subjects over complete gait cycles were 0.82 ± 0.07 m/s, 0.78 ± 0.08 m/s and 0.57 ± 0.05 m/s for level walking, downhill walking, and stair descent, respectively. For all three gait activities, the TKR subjects showed comparable mean gait velocities ranging from 0.87–0.92 m/s for level walking, 0.77–0.79 m/s for downhill walking and 0.55–0.58 m/s for stair descent (Table A1).
      Compared to the natural knees, the GMK UC showed a reduced range of flexion for all three gait activities (Table 2), as well as significantly different flexion patterns around toe-off during level walking and in the second half of the stance phase for stair descent (Fig. 2). For the GMK Sphere and GMK PS groups significantly different flexion patterns were observed for all three gait activities during the unloaded swing phase compared to the subjects with natural knees (Fig. 2). Over the complete gait cycle of all three activities, the TKR groups generally exhibited smaller ranges of abduction/adduction as well as tibial internal/external rotation compared to the natural group (Table 2). While the natural tibiae progressively rotated internally during the first half of the stance phase, the TKR subjects showed almost no tibial internal/external rotation, especially for downhill walking and stair descent (Fig. 2). During the unloaded swing phase, the GMK Sphere group showed a characteristic mean peak rotation which was not evident in the natural group (Fig. 2). For abduction/adduction all groups (natural and TKR) showed almost no movement during the loaded stance phase, but a trend towards different movement patterns in the unloaded swing phase. Here, the natural subjects showed substantially larger standard deviations (0.70–2.30°) than the TKR subjects (GMK Sphere: 0.25–1.35°, GMK PS: 0.22–0.80°, GMK UC: 0.17–0.73°) (Fig. 2).
      Fig. 2
      Fig. 2Tibio-femoral flexion/extension (flex/ext), tibial internal/external rotation (tib int/ext rot), and abduction/adduction (abd/add) throughout complete gait cycles of level walking (left column), downhill walking (middle column), and stair descent (right column). All rotations are presented relative to the mean position during standing. The mean (thick line) and standard deviations (dashed line) across each group are presented. Significant differences of the GMK Sphere, GMK PS and GMK UC group compared to the natural group are indicated with bars in the colours of the respective group with an adjusted level of significance of α = 0.0167. The average instance of toe-off across all subjects of each group is indicated as a vertical line.
      The GMK Sphere subjects showed significantly less medial but equal lateral ranges of condylar translations compared to the natural subjects for all three activities (Table 2). Although the GMK PS also exhibited equal lateral condylar translations compared to the natural subjects, the ranges of medial condylar A-P translations were significantly larger. Unlike the other two TKR designs, the GMK UC presented equal medial translations but significantly smaller ranges of translation for the lateral condyle during downhill walking and stair descent (Table 2). Overall, differences in the A-P translational patterns of the medial condyle of the TKR compared to the natural subjects occurred predominantly in late stance and early to mid-swing phase. However, differences over a larger region of the complete gait cycle occurred for the lateral condylar A-P translation. While the natural subjects showed an initial posterior translation after heel-strike and a fast anterior translation around toe-off, the TKR groups showed a clear anterior peak during the second half of the swing phase (Fig. 3).
      Fig. 3
      Fig. 3Antero (ant) – posterior (post) translation of the medial and lateral femoral condyle throughout complete gait cycles of level walking (left column), downhill walking (middle column), and stair descent (right column). All translations are presented relative to the mean position during standing. The mean (thick line) and standard deviations (dashed line) across each group are presented. Significant differences of the GMK Sphere, GMK PS and GMK UC group compared to the natural group are indicated with bars in the colours of the respective group with an adjusted level of significance of α = 0.0167. The average instance of toe-off across all subjects of each group is indicated as a vertical line.
      While the natural and GMK Sphere subjects exhibited significantly less medial than lateral condylar A-P translation for level walking, downhill walking and stair descent, the GMK PS and GMK UC subjects showed equal or more lateral condylar translations (Table 2, Table A2). As a result, the natural and GMK Sphere subjects demonstrated on average a medial CoR for all three gait activities whereas the GMK PS and GMK UC showed larger variability between subjects and a more central to lateral average CoR location (Fig. 4).
      Fig. 4
      Fig. 4Location of the centre of rotation over the complete gait cycle of level walking (top), downhill walking (middle), and stair descent (bottom). Mean locations and standard deviations for each subject group (thick line) as well as individual subject locations (stars) are indicated.

      4. Discussion

      Knowledge of natural tibio-femoral movement patterns is critical for a detailed understanding of kinematic deficits and the underlying mechanisms that result in poor outcomes, but also lays the foundations for targeted implant selection towards restoration of physiological knee joint function. This study compared in vivo tibio-femoral kinematics in natural versus replaced knees throughout complete gait cycles of not only level walking, but also more challenging activities such as downhill walking and stair descent using a consistent dynamic videofluoroscopic measurement set-up and corresponding analysis approach for both natural and TKR subjects.
      In general, each TKR design exhibited distinct motion patterns, clearly driven by its specific design features, with tibio-femoral rotation and condylar translation characteristics varying from the movement observed for the natural knee for certain phases of the gait cycle (Fig. 2, Fig. 3, Fig. 4, Table 2). While only small changes to the range of abduction/adduction patterns were evident during the swing phase for the TKR groups (compared to natural knee motion) for all three activities, the most prominent differences occurred for tibial internal/external rotations and the resultant medial and lateral condylar A-P translations (Fig. 2, Fig. 3, Figs. A1 and A2). Here, the TKR groups showed comparable tibial internal/external rotation during standing and at heel strike, but the natural group showed an overall externally rotated tibia during standing but a more internally rotated position at heel strike for all three activities (Fig. 1, Fig. 2). In agreement with previous studies (
      • Gray H.A.
      • Guan S.
      • Young T.J.
      • Dowsey M.M.
      • Choong P.F.
      • Pandy M.G.
      Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
      ;
      • Miura K.
      • Ohkoshi Y.
      • Ino T.
      • Ukishiro K.
      • Kawakami K.
      • Suzuki S.
      • Suzuki K.
      • Maeda T.
      Kinematics and center of axial rotation during walking after medial pivot type total knee arthroplasty.
      ), more tibial internal rotation over the course of the gait cycles as well as a reduced range of tibial internal/external rotation was found for the natural compared to the TKR groups (Fig. 2 and Fig. A1, Table 2).
      Among the three TKR implant geometries analysed in this study, the GMK Sphere most closely replicated the overall medial CoR observed for the natural group (Fig. 4). However, the highly congruent medial compartment resulted in a significantly reduced range of medial A-P translation for all three gait activities compared to the natural group. The flat unconstrained geometry of the lateral compartment provided rotational freedom, especially during the unloaded swing phase, and resulted in highly subject-specific lateral translational characteristics, almost certainly driven by the surrounding soft-tissues. The result is an overall lateral range of condylar translation comparable to the natural group (Table 2). Furthermore, in all activities, the GMK Sphere showed a distinct peak rotation during the swing phase, which was not present in the natural group, plausibly due to the medial congruency limiting simultaneous anterior motion of both condyles. Interestingly, the smallest range of tibial rotation among the three TKR designs was observed for the GMK UC despite its mobile bearing platform, possibly due to the highly constrained medial and lateral geometry of the tibial inlay (
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ). It therefore seems that the functional potential of the rotating platform is generally not fully used during normal gait activities. Here, while the ranges of medial condylar A-P translation best approximated the range found for the natural group, the lateral condyle A-P translation was especially restricted during downhill walking and stair descent (Table 2). The GMK PS exhibited a similar translational RoM in the medial and lateral compartments, but reduced ranges of tibial rotation and substantially more medial translation compared to the natural knees (Table 2).
      Within the GMK PS and UC groups, some subjects exhibited almost pure A-P translation with minimal internal/external rotations. For such parallel movements, the calculation of the CoR is not robust (
      • Ehrig R.M.
      • Taylor W.R.
      • Duda G.N.
      • Heller M.O.
      A survey of formal methods for determining functional joint axes.
      ) and therefore results in large standard deviations for the CoR location (up to 31.7 mm for a single subject). As a result, the mean location of the CoR in these two implants should be interpreted with caution.
      While all TKR designs have deficits in replicating the entire spectrum of characteristics and RoMs of abduction/adduction and tibial internal/external rotation, as well as both condylar translations, they are able to mimic certain aspects of natural knee kinematics. The challenge for TKA designs to provide a balance between guiding the motion but also allowing freedom for individual kinematic characteristics is clear, especially considering the role of anatomical differences between the subjects. All TKR subjects within this study were satisfied with their outcome and demonstrated the ability to handle different implant constraints, possibly due to their individual soft-tissue and muscular sufficiency. While further development of TKR designs could help to even closer replicate all aspects of natural tibio-femoral kinematics, implant selection based on the patients' conditions and activity levels seems to be critical for allowing the necessary rotational and translational ranges of motion but also providing sufficient guidance for joint stability, while simultaneously not overloading the surrounding tissues (
      • Hosseini Nasab S.H.
      • Smith C.
      • Schütz P.
      • Postolka B.
      • Ferguson S.
      • Taylor W.R.
      • List R.
      Elongation patterns of the posterior cruciate ligament after total knee arthroplasty.
      ;
      • Nasab S.H.H.
      • Smith C.R.
      • Postolka B.
      • Schütz P.
      • List R.
      • Taylor W.R.
      In vivo elongation patterns of the collateral ligaments in healthy knees during functional activities.
      ).
      The TKR subjects examined in this study were mechanically aligned (range of HKA angle during standing: 3° valgus to 3° varus, Table 1) and only showed minimal dynamic inter-subject variability for abduction/adduction (Fig. 2). On the other hand, the healthy subjects covered a wide range of individual limb alignment (8° valgus to 9° varus), hence providing a valuable representation of the healthy population (
      • Bellemans J.
      • Colyn W.
      • Vandenneucker H.
      • Victor J.
      The Chitranjan Ranawat award: is neutral mechanical alignment normal for all patients? The concept of constitutional varus.
      ), resulting in a large inter-subject variability for abduction/adduction over the complete gait cycles (Fig. A1). Despite the large variation in limb alignment, however, a relatively stable medial CoR in the transverse plane was observed. Nevertheless, the role of limb alignment on tibio-femoral kinematics was not investigated further here and will be analysed in greater detail in follow-up studies.
      Within the current study, both the natural subjects as well as all TKR subjects were assessed using the same videofluoroscopy set-up. Although all subjects generally walked slower compared to self-selected walking speeds (
      • Hitz M.
      • Schütz P.
      • Angst M.
      • Taylor W.R.
      • List R.
      Influence of the moving fluoroscope on gait patterns.
      ;
      • Stacoff A.
      • Diezi C.
      • Luder G.
      • Stüssi E.
      • Kramers-de Quervain I.A.
      Ground reaction forces on stairs: effects of stair inclination and age.
      ), the mean gait velocities among the four subject groups were comparable (Table A1) and allowed direct comparison of the natural and TKR groups for all investigated activities. However, differences in the orientation of the femur relative to the tibia, as well as the joint geometry and implantation of the components, are all likely to cause some of the observed differences between the natural and TKR groups. While the medio-lateral axis of the TKR subjects was defined based on the implant geometry, closely approximating the geometric centre axis, the corresponding axis of the natural knees was defined functionally based on three deep knee bending trials (
      • Ehrig R.M.
      • Taylor W.R.
      • Duda G.N.
      • Heller M.O.
      A survey of formal methods for determining functional joint axes.
      ;
      • Postolka B.
      • Schütz P.
      • Fucentese S.F.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      • Taylor W.R.
      Tibio-femoral kinematics of the healthy knee joint throughout complete cycles of gait activities.
      ;
      • Postolka B.
      • Taylor W.R.
      • Dätwyler K.
      • Heller M.O.
      • List R.
      • Schütz P.
      Interpretations of tibio-femoral kinematics critically depends upon the kinematic analysis approach: A survey and comparison of methodologies.
      ;
      • Schütz P.
      • Taylor W.R.
      • Postolka B.
      • Fucentese S.F.
      • Koch P.P.
      • Freeman M.A.R.
      • Pinskerova V.
      • List R.
      Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
      ). Although the geometric centre axis and the functional flexion axis were shown to closely approximate one another (
      • Postolka B.
      • Taylor W.R.
      • Dätwyler K.
      • Heller M.O.
      • List R.
      • Schütz P.
      Interpretations of tibio-femoral kinematics critically depends upon the kinematic analysis approach: A survey and comparison of methodologies.
      ), differences occurring from the altered coordinate system definitions are inherently plausible. Here, presentation of the kinematics relative to the standing trials were chosen to minimise this effect.
      A few other limitations of this study should be acknowledged. Here, the single-plane moving fluoroscope system, in combination with 2D/3D registration procedures, is able to reconstruct in-plane tibio-femoral rotations and condylar translations with high accuracy but is limited by larger out-of-plane registration errors (
      • Foresti M.
      In vivo measurement of total knee joint replacement kinematics and kinetics during stair descent, D-MAVT.
      ;
      • Postolka B.
      • List R.
      • Thelen B.
      • Schütz P.
      • Taylor W.R.
      • Zheng G.
      Evaluation of an intensity-based algorithm for 2D/3D registration of natural knee videofluoroscopy data.
      ). As a result, the focus of this study was on tibio-femoral rotations and the condylar A-P translation, while medio-lateral tibio-femoral translations were not considered. In addition, it must be mentioned that the subjects of the natural group were not age-matched to the implant groups, which was specifically chosen in this study to provide a comparison against subjects with truly healthy kinematics. Here, variability in soft-tissue sufficiency, including the effects of subject-specific ligament balancing (which was not standardised intra-operatively), could exacerbate these differences. However, the subjects measured in this study were all able to adapt to the kinematic changes observed and were therefore classified as good outcome and satisfied subjects and measured in a standardised and consistent manner.

      5. Conclusions

      The TKR designs analysed within this study use different geometric features to guide tibio-femoral kinematics. Despite substantial similarities between the TKR and natural tibio-femoral kinematics, none of the TKR designs was fully able to replicate the natural tibio-femoral rotations and condylar A-P translations. The GMK Sphere best mimicked the medial CoR found in the natural knees across a wide range of limb alignments but showed some differences in the rotational behaviour and less translation in the medial compartment compared to the natural group. The GMK PS and GMK UC, however, showed equal ranges of translation for the lateral and medial condyle. When compared to the natural knee, the GMK PS subjects experienced more A-P translation for the medial condyle but a comparable RoM for the lateral condyle. For the GMK UC, on the other hand, equal ranges of A-P translation found for the medial condyle compared to the natural knee but significantly reduced ranges of lateral A-P translation. Tibio-femoral kinematic variability among the healthy subjects indicates the necessity for different implant designs to address subject-specific requirements in restoring physiological knee joint function and stability.

      Funding

      This study was financially supported by the Commission for Technology and Innovation (Bern, Switzerland, Project Number: 17078.1 PFLS-LS) and Medacta International SA (Castel San Pietro, Switzerland).

      Declaration of interest

      RL has received speaker's fee from Medacta. SFF and PPK receive advisory fees from Medacta as part of their consultancy work.

      Acknowledgements

      The authors like to thank Nathalie Kühne and Michel Schläppi for recruiting the total knee arthroplasty subjects as well as organising the clinical assessment sessions of the natural subjects. Furthermore, we are grateful to Prof. Dr. Michael A.R. Freeman and Prof. Dr. Vera Pinskerova for the fruitful discussions over the course of the whole project. The help of Michael Angst, Michael Plüss and all the students during the measurement sessions and the post-processing of the data is greatly appreciated.

      Appendix A. Supplementary data

      References

        • Anderst W.
        • Zauel R.
        • Bishop J.
        • Demps E.
        • Tashman S.
        Validation of three-dimensional model-based tibio-femoral tracking during running.
        Med. Eng. Phys. 2009; 31: 10-16
        • Andriacchi T.P.
        • Alexander E.J.
        • Toney M.K.
        • Dyrby C.
        • Sum J.
        A point cluster method for in vivo motion analysis: applied to a study of knee kinematics.
        J. Biomech. Eng. 1998; 120: 743-749
        • Banks S.A.
        • Markovich G.D.
        • Hodge W.A.
        In vivo kinematics of cruciate-retaining and -substituting knee arthroplasties.
        J. Arthroplast. 1997; 12: 297-304
        • Bellemans J.
        • Colyn W.
        • Vandenneucker H.
        • Victor J.
        The Chitranjan Ranawat award: is neutral mechanical alignment normal for all patients? The concept of constitutional varus.
        Clin. Orthop. Relat. Res. 2012; 470: 45-53
        • Bourne R.B.
        • Chesworth B.M.
        • Davis A.M.
        • Mahomed N.N.
        • Charron K.D.
        Patient satisfaction after total knee arthroplasty: who is satisfied and who is not?.
        Clin. Orthop. Relat. Res. 2010; 468: 57-63
        • Burckhardt K.
        • Szekely G.
        • Notzli H.
        • Hodler J.
        • Gerber C.
        Submillimeter measurement of cup migration in clinical standard radiographs.
        IEEE Trans. Med. Imaging. 2005; 24: 676-688
        • Cappozzo A.
        • Croce U.D.
        • Leardini A.
        • Chiari L.
        Human movement analysis using stereophotogrammetry. Part 1: theoretical background.
        Gait Posture. 2005; 21: 186-196
        • Cates H.E.
        • Komistek R.D.
        • Mahfouz M.R.
        • Schmidt M.A.
        • Anderle M.
        In vivo comparison of knee kinematics for subjects having either a posterior stabilized or cruciate retaining high-flexion total knee arthroplasty.
        J. Arthroplast. 2008; 23: 1057-1067
        • DeFrate L.E.
        • Sun H.
        • Gill T.J.
        • Rubash H.E.
        • Li G.
        In vivo tibiofemoral contact analysis using 3D MRI-based knee models.
        J. Biomech. 2004; 37: 1499-1504
        • Dennis D.
        • Komistek R.
        • Scuderi G.
        • Argenson J.N.
        • Insall J.
        • Mahfouz M.
        • Aubaniac J.M.
        • Haas B.
        In vivo three-dimensional determination of kinematics for subjects with a normal knee or a unicompartmental or total knee replacement.
        J. Bone Joint Surg. Am. 2001; 83-A: 104-115
        • Dennis D.A.
        • Komistek R.D.
        • Mahfouz M.R.
        In vivo fluoroscopic analysis of fixed-bearing total knee replacements.
        Clin. Orthop. Relat. Res. 2003; : 114-130
        • Dunbar M.J.
        • Richardson G.
        • Robertsson O.
        I can't get no satisfaction after my total knee replacement: rhymes and reasons.
        Bone Joint J. 2013; 95-B: 148-152
        • Ehrig R.M.
        • Taylor W.R.
        • Duda G.N.
        • Heller M.O.
        A survey of formal methods for determining the centre of rotation of ball joints.
        J. Biomech. 2006; 39: 2798-2809
        • Ehrig R.M.
        • Taylor W.R.
        • Duda G.N.
        • Heller M.O.
        A survey of formal methods for determining functional joint axes.
        J. Biomech. 2007; 40: 2150-2157
        • Folinais D.
        • Thelen P.
        • Delin C.
        • Radier C.
        • Catonne Y.
        • Lazennec J.Y.
        Measuring femoral and rotational alignment: EOS system versus computed tomography.
        Orthop. Traumatol. Surg. Res. 2013; 99: 509-516
        • Foresti M.
        In vivo measurement of total knee joint replacement kinematics and kinetics during stair descent, D-MAVT.
        ETH Zürich, Zürich2009 (p. 77)
        • Galvin C.R.
        • Perriman D.M.
        • Lynch J.T.
        • Pickering M.R.
        • Newman P.
        • Smith P.N.
        • Scarvell J.M.
        Age has a minimal effect on knee kinematics: a cross-sectional 3D/2D image-registration study of kneeling.
        Knee. 2019; 26: 988-1002
        • Ghirardelli S.
        • Asay J.L.
        • Leonardi E.A.
        • Amoroso T.
        • Andriacchi T.P.
        • Indelli P.F.
        Kinematic comparison between medially congruent and posterior-stabilized third-generation TKA designs.
        J. Funct. Morphol. Kinesiol. 2021; 6
        • Gray H.A.
        • Guan S.
        • Thomeer L.T.
        • Schache A.G.
        • de Steiger R.
        • Pandy M.G.
        Three-dimensional motion of the knee-joint complex during normal walking revealed by mobile biplane X-ray imaging.
        J. Orthop. Res. 2019; 37: 615-630
        • Gray H.A.
        • Guan S.
        • Young T.J.
        • Dowsey M.M.
        • Choong P.F.
        • Pandy M.G.
        Comparison of posterior-stabilized, cruciate-retaining, and medial-stabilized knee implant motion during gait.
        J. Orthop. Res. 2020; 38: 1753-1768
        • Grieco T.F.
        • Sharma A.
        • Komistek R.D.
        • Cates H.E.
        Single versus multiple-radii cruciate-retaining total knee arthroplasty: an in vivo mobile fluoroscopy study.
        J. Arthroplast. 2016; 31: 694-701
        • Grood E.S.
        • Suntay W.J.
        A joint coordinate system for the clinical description of three-dimensional motions: application to the knee.
        J. Biomech. Eng. 1983; 105: 136-144
        • Guan S.
        • Gray H.A.
        • Keynejad F.
        • Pandy M.G.
        Mobile biplane X-ray imaging system for measuring 3D dynamic joint motion during overground gait.
        IEEE Trans. Med. Imaging. 2016; 35: 326-336
        • Hamai S.
        • Moro-oka T.A.
        • Dunbar N.J.
        • Miura H.
        • Iwamoto Y.
        • Banks S.A.
        In vivo healthy knee kinematics during dynamic full flexion.
        Biomed. Res. Int. 2013; 2013717546
        • Hill P.F.
        • Vedi V.
        • Williams A.
        • Iwaki H.
        • Pinskerova V.
        • Freeman M.A.
        Tibiofemoral movement 2: the loaded and unloaded living knee studied by MRI.
        J. Bone Joint Surg. (Br.). 2000; 82: 1196-1198
        • Hitz M.
        • Schütz P.
        • Angst M.
        • Taylor W.R.
        • List R.
        Influence of the moving fluoroscope on gait patterns.
        PLoS One. 2018; 13e0200608
        • Hosseini Nasab S.H.
        • Smith C.
        • Schütz P.
        • Postolka B.
        • Ferguson S.
        • Taylor W.R.
        • List R.
        Elongation patterns of the posterior cruciate ligament after total knee arthroplasty.
        J. Clin. Med. 2020; 9: 2078
        • Komistek R.D.
        • Dennis D.A.
        • Mahfouz M.
        In vivo fluoroscopic analysis of the normal human knee.
        Clin. Orthop. Relat. Res. 2003; : 69-81
        • Koo Y.J.
        • Koo S.
        Three-dimensional kinematic coupling in the knee during normal walking.
        J. Biomech. Eng. 2019; 141081012
        • Kozanek M.
        • Hosseini A.
        • Liu F.
        • Van de Velde S.K.
        • Gill T.J.
        • Rubash H.E.
        • Li G.
        Tibiofemoral kinematics and condylar motion during the stance phase of gait.
        J. Biomech. 2009; 42: 1877-1884
        • Li G.
        • Van de Velde S.K.
        • Bingham J.T.
        Validation of a non-invasive fluoroscopic imaging technique for the measurement of dynamic knee joint motion.
        J. Biomech. 2008; 41: 1616-1622
        • List R.
        • Foresti M.
        • Gerber H.
        • Goldhahn J.
        • Rippstein P.
        • Stüssi E.
        Three-dimensional kinematics of an unconstrained ankle arthroplasty: a preliminary in vivo videofluoroscopic feasibility study.
        Foot Ankle Int./Am. Orthop. Foot Ankle Soc. [and] Swiss Foot Ankle Soc. 2012; 33: 883-892
        • List R.
        • Postolka B.
        • Schütz P.
        • Hitz M.
        • Schwilch P.
        • Gerber H.
        • Ferguson S.J.
        • Taylor W.R.
        A moving fluoroscope to capture tibiofemoral kinematics during complete cycles of free level and downhill walking as well as stair descent.
        PLoS One. 2017; 12e0185952
        • List R.
        • Schütz P.
        • Angst M.
        • Ellenberger L.
        • Dätwyler K.
        • Ferguson S.J.
        • Writing C.
        Videofluoroscopic evaluation of the influence of a gradually reducing femoral radius on joint kinematics during daily activities in total knee arthroplasty.
        J. Arthroplast. 2020; 35: 3010-3030
        • Liu F.
        • Kozanek M.
        • Hosseini A.
        • Van de Velde S.K.
        • Gill T.J.
        • Rubash H.E.
        • Li G.
        In vivo tibiofemoral cartilage deformation during the stance phase of gait.
        J. Biomech. 2010; 43: 658-665
        • Miura K.
        • Ohkoshi Y.
        • Ino T.
        • Ukishiro K.
        • Kawakami K.
        • Suzuki S.
        • Suzuki K.
        • Maeda T.
        Kinematics and center of axial rotation during walking after medial pivot type total knee arthroplasty.
        J. Exp. Orthop. 2020; 7: 72
        • Moewis P.
        • Duda G.N.
        • Jung T.
        • Heller M.O.
        • Boeth H.
        • Kaptein B.
        • Taylor W.R.
        The restoration of passive rotational tibio-femoral laxity after anterior cruciate ligament reconstruction.
        PLoS One. 2016; 11e0159600
        • Moro-oka T.A.
        • Hamai S.
        • Miura H.
        • Shimoto T.
        • Higaki H.
        • Fregly B.J.
        • Iwamoto Y.
        • Banks S.A.
        Dynamic activity dependence of in vivo normal knee kinematics.
        J. Orthop. Res. 2008; 26: 428-434
        • Moro-oka T.A.
        • Muenchinger M.
        • Canciani J.P.
        • Banks S.A.
        Comparing in vivo kinematics of anterior cruciate-retaining and posterior cruciate-retaining total knee arthroplasty.
        Knee Surg. Sports Traumatol. Arthrosc. 2007; 15: 93-99
        • Nasab S.H.H.
        • Smith C.R.
        • Postolka B.
        • Schütz P.
        • List R.
        • Taylor W.R.
        In vivo elongation patterns of the collateral ligaments in healthy knees during functional activities.
        J. Bone Joint Surg. Am. 2021; 103: 1620-1627
        • Pataky T.C.
        • Robinson M.A.
        • Vanrenterghem J.
        Region-of-interest analyses of one-dimensional biomechanical trajectories: bridging 0D and 1D theory, augmenting statistical power.
        PeerJ. 2016; 4e2652
        • Pauchard Y.
        • Fitze T.
        • Browarnik D.
        • Eskandari A.
        • Pauchard I.
        • Enns-Bray W.
        • Palsson H.
        • Sigurdsson S.
        • Ferguson S.J.
        • Harris T.B.
        • Gudnason V.
        • Helgason B.
        Interactive graph-cut segmentation for fast creation of finite element models from clinical ct data for hip fracture prediction.
        Comput. Methods Biomech. Biomed. Eng. 2016; 19: 1693-1703
        • Pfitzner T.
        • Moewis P.
        • Stein P.
        • Boeth H.
        • Trepczynski A.
        • von Roth P.
        • Duda G.N.
        Modifications of femoral component design in multi-radius total knee arthroplasty lead to higher lateral posterior femoro-tibial translation.
        Knee Surg. Sports Traumatol. Arthrosc. 2018; 26: 1645-1655
        • Pinskerova V.
        • Johal P.
        • Nakagawa S.
        • Sosna A.
        • Williams A.
        • Gedroyc W.
        • Freeman M.A.
        Does the femur roll-back with flexion?.
        J. Bone Joint Surg. (Br.). 2004; 86: 925-931
        • Postolka B.
        • List R.
        • Thelen B.
        • Schütz P.
        • Taylor W.R.
        • Zheng G.
        Evaluation of an intensity-based algorithm for 2D/3D registration of natural knee videofluoroscopy data.
        Med. Eng. Phys. 2020; 77: 107-113
        • Postolka B.
        • Schütz P.
        • Fucentese S.F.
        • Freeman M.A.R.
        • Pinskerova V.
        • List R.
        • Taylor W.R.
        Tibio-femoral kinematics of the healthy knee joint throughout complete cycles of gait activities.
        J. Biomech. 2020; : 110
        • Postolka B.
        • Taylor W.R.
        • Dätwyler K.
        • Heller M.O.
        • List R.
        • Schütz P.
        Interpretations of tibio-femoral kinematics critically depends upon the kinematic analysis approach: A survey and comparison of methodologies.
        J. Biomech. 2022; (in review)
        • Robertsson O.
        • Dunbar M.
        • Pehrsson T.
        • Knutson K.
        • Lidgren L.
        Patient satisfaction after knee arthroplasty: a report on 27,372 knees operated on between 1981 and 1995 in Sweden.
        Acta Orthop. Scand. 2000; 71: 262-267
        • Schütz P.
        • Postolka B.
        • Gerber H.
        • Ferguson S.J.
        • Taylor W.R.
        • List R.
        Knee implant kinematics are task-dependent.
        J. R. Soc. Interface. 2019; 16: 20180678
        • Schütz P.
        • Taylor W.R.
        • Postolka B.
        • Fucentese S.F.
        • Koch P.P.
        • Freeman M.A.R.
        • Pinskerova V.
        • List R.
        Kinematic evaluation of the GMK sphere implant during gait activities: a dynamic videofluoroscopy study.
        J. Orthop. Res. 2019; 37: 2337-2347
        • Scott G.
        • Imam M.A.
        • Eifert A.
        • Freeman M.A.R.
        • Pinskerova V.
        • Field R.E.
        • Skinner J.
        • Banks S.A.
        Can a total knee arthroplasty be both rotationally unconstrained and anteroposteriorly stabilised?.
        Bone Joint Res. 2016; 5: 80-86
        • Sharma A.
        • Leszko F.
        • Komistek R.D.
        • Scuderi G.R.
        • Cates Jr., H.E.
        • Liu F.
        In vivo patellofemoral forces in high flexion total knee arthroplasty.
        J. Biomech. 2008; 41: 642-648
        • Stacoff A.
        • Diezi C.
        • Luder G.
        • Stüssi E.
        • Kramers-de Quervain I.A.
        Ground reaction forces on stairs: effects of stair inclination and age.
        Gait Posture. 2005; 21: 24-38
        • Tanifuji O.
        • Sato T.
        • Kobayashi K.
        • Mochizuki T.
        • Koga Y.
        • Yamagiwa H.
        • Omori G.
        • Endo N.
        Three-dimensional in vivo motion analysis of normal knees using single-plane fluoroscopy.
        J. Orthop. Sci. 2011; 16: 710-718
        • Taylor W.R.
        • Ehrig R.M.
        • Duda G.N.
        • Schell H.
        • Seebeck P.
        • Heller M.O.
        On the influence of soft tissue coverage in the determination of bone kinematics using skin markers.
        J. Orthop. Res. 2005; 23: 726-734
        • Victor J.
        • Banks S.
        • Bellemans J.
        Kinematics of posterior cruciate ligament-retaining and -substituting total knee arthroplasty.
        J. Bone Joint Surg. (Br.). 2005; 87B: 646-655