Introduction

Proper alignment (tibial alignment, femoral alignment, and overall anatomic alignment) of the prosthesis during total knee replacement is critical in maximizing implant survival[7] and to reduce polyethylene wear[1]. Poor overall anatomic alignment of a total knee replacement was associated with a 6.9 times greater risk of failure due to tibial collapse, that varus tibial alignment is associated with a 3.2 times greater risk[2] and valgus femoral alignment is associated with a 5.1 times greater risk of failure[7]. To reduce this variability intramedullary (IM) instruments have been widely used, with increased risk of the fat emboli rate to the lungs and brain during TKA[6] and possible increase of blood loss[4, 5]. Or, alternatively, navigation has been used to achieve proper alignment and to reduce morbidity[3]. Recently, for distal femoral resection, inertial sensors have been coupled to extramedullary (EM) instruments to improve TKA surgery in terms of femoral implant alignment, with respect to femoral mechanical axis, and reduced morbidity by avoidance of IM canal violation. The purpose if this study is to compare blood loss and alignment of distal femoral cut in three cohorts of patients: 1 Operated with inertial based cutting guide; 2 Operated with navigation instruments; 3 operated with conventional IM instruments.

The main purpose of the present study is to prospectively investigate whether preoperative functional flexion axis in patients with osteoarthritisand varus-alignment changes after total knee arthroplasty and whether a correlation exists both between preoperative functional flexion axis and native limb deformity. A navigated total knee arthroplasty was performed in 108 patients using a specific software to acquire passive joint kinematics before and after implant positioning. The knee was cycled through three passive range of motions, from 0 to 120. Functional flexion axis was computed using the mean helical axis algorithm. The angle between the functional flexion axis and the surgical transepicondylar axis was determined on frontal (aF) and axial (aA) plane. The pre- and postoperative hip-kneeankle angle, related to femur mechanical axis, was determined. Postoperative functional flexion axis was different from preoperative only on frontal plane, while no differences were found on axial plane. No correlation was found between preoperative aA and native limb deformity, while a poor correlation was found in frontal plane, between aF and preoperative hip-knee-ankle angle. Total knee arthroplasty affects functional flexion axis only on frontal plane while has no effect on axial plane. Preoperative functional flexion axis is in a more varus position respect to the transepicondylar axis both in pre- and postoperative conditions. Moreover, the position of the functional axis on frontal plane in preoperative conditions is dependent on native limb alignment, while on axial plane is not dependent on the amount of preoperative varus deformity.

Introduction

The purpose of this study was to examine whether three types of mobile-bearing PCL sacrificing TKA could restore the native knee translation and rotation. The primary hypothesis was that there are differences in knee kinematics and laxity between three different cruciate-substituting TKA designs: 1 with post-cam mechanism, 2 post-cam mechanism based on an inter-condylar ‘third condyle’ concept, 3 anterior stabilized with deep-dished highly congruent tibial insert; specifically, showing different femoral external rotation with flexion, different femoral translation with flexion and different laxity under stress test. The secondary hypothesis was that there is different clinical outcome between the three TKA designs at 2 years follow-up.

Background

The optimal reference for rotational positioning of femoral component in total knee replacement (TKR) is debated. Navigation has been suggested for intra-op acquisition of patient's specific kinematics and functional flexion axis (FFA).

BACKGROUND:

The optimal reference for rotational positioning of femoral component in total knee replacement (TKR) is debated. Navigation has been suggested for intra-op acquisition of patient's specific kinematics and functional flexion axis (FFA).

INTRODUCTION

This study aimed to intra-operatively quantify the improvements in knee stability given both by anatomic double-bundle (ADB) and single-bundle with additional lateral plasty (SBLP) ACL reconstruction using a navigation system.

Introduction

Several in vitro and in vivo studies have found correspondence between transepicondylar axis (TEA) and functional flexion axis (FFA) in healthy subjects. In addition some studies suggest that the use of FFA for rotational alignment of femoral implant may be more accurate than TEA. Ostheoarthritis (OA) may modify limb alignment and therefore flexion axis, introducing a bias at different flexion ranges during kinematic acquisition. In this study we want to understand whether OA affects somehow the FFA evaluation compared to TEA and whether the FFA could be considered a usable reference for implant positioning for osteoarthritic knees

Introduction

Patellar stability is an important component for a correct kinematic behaviour of the knee that depends on several factors such as joint geometry, muscles strength and soft tissues actions. Patellofemoral (PF) maltracking can results in many joint disorders which can cause pain and mobility alterations. The medial patellofemoral ligament (MPFL) is an important stabilizing structure for the patellofemoral joint. The aim of this study was to analyze patellofemoral kinematics with particular attention to the contribution of MPFL on patella stability.

We have shown in a previous study that patients with combined lesions of the anterior cruciate (ACL) and medial collateral ligaments (MCL) had similar anteroposterior (AP) but greater valgus laxity at 30° after reconstruction of the ACL when compared with patients who had undergone reconstruction of an isolated ACL injury. The present study investigated the same cohort of patients after a minimum of three years to evaluate whether the residual valgus laxity led to a poorer clinical outcome.

Each patient had undergone an arthroscopic double-bundle ACL reconstruction using a semitendinosus-gracilis graft. In the combined ACL/MCL injury group, the grade II medial collateral ligament injury was not treated. At follow-up, AP laxity was measured using a KT-2000 arthrometer, while valgus laxity was evaluated with Telos valgus stress radiographs and compared with the uninjured knee. We evaluated clinical outcome scores, muscle girth and time to return to activities for the two groups.

Valgus stress radiographs showed statistically significant greater mean medial joint opening in the reconstructed compared with the uninjured knees (1.7 mm (sd 0.9) versus 0.9 mm (sd 0.7), respectively, p = 0.013), while no statistically significant difference was found between the AP laxity and the other clinical parameters. Our results show that the residual valgus laxity does not affect AP laxity significantly at a minimum follow up of three years, suggesting that no additional surgical procedure is needed for the medial collateral ligament in combined lesions.

Introduction: Several in vitro and in vivo studies have found correspondence between transepicondylar axis (TEA) and mean helical axis (MHA) in healthy subjects. In addition some studies suggest that the use of MHA for rotational alignment of femoral implant may be more accurate than TEA. Ostheoarthritis (OA) may modify limb alignment and flexion axis, introducing a bias during kinematic acquisition. An in-vivo study comparing normal and osteoarthritic knees using MHA is still lacking. The purposes of this study were: to understand whether arthritis affects somehow the functional axis evaluation and then to assess whether the MHA could be considered as reference flexion axis also for osteoarthritic knees; starting from hypothesis that there is a correspondence between TEA and MHA, to evaluate whether in pathologic subjects there still is the same correspondence.

Material and Methods: We included a group of 15 OA patients undergoing TKA and, as control group, 60 patients that underwent ACL reconstruction, since in vivo studies reported small differences in kinematics between ACL reconstructed and uninjured limbs. With a surgical navigation system we recorded intraoperative kinematic data of different passive ranges of motion (PROM) and calculated the MHA applying a least square approach to the set of finite helical axes (FHA) obtained in three different ranges of motion (0°–120°; 35°–80°; 35°–120°). We compared the difference in orientation of MHA in the three ranges with respect to the TEA on frontal (XZ) and axial (XY) planes. The correlation of preoperative limb deformity with MHA-TEA angle was also performed.

Results: The results of difference of MHA-TEA angle between the OA and ACL groups for all the three ranges of flexion and in XZ and XY views showed no statistical difference (p=0.5188; p=0.7147 respectively). No statistical difference was found also about MHA-TEA angle between the three ranges in frontal and axial views (ANOVA p=0.6373; p=0.4183 respectively). There was no difference between the flexion and extension movements in the three ranges. We also found that correlation between limb alignment and MHA-TEA angle showed good correlation (r> 0.54, p< 0.001) in frontal view and fair correlation (r< 0.37, p< 0.05) in axial view for all ranges.

Conclusions: Our work has demonstrated that pathologic knees shows no differences in MHA orientation compared to nearly healthy subjects, moreover there is the same correspondence between TEA and MHA both in XZ and XY plane. We also found that preoperative limb alignment does not correlate with MHA-TEA angle. results are in agreement to studies on healthy subjects. Therefore the MHA may be considered a reliable reference for determining femoral flexion axis and a useful tool in the determination of femoral implant positioning on axial plane, even in surgical setup on osteoarthritic patients.

Introduction: Anterior Cruciate Ligament (ACL) is primary constrain to anterior displacement of tibia with respect to the femur and secondary to internal/external (IE) and varus/valgus (VV) rotations; an ACL reconstruction should thus control not only AP but also IE and VV laxities. For this reasons, more attention has given to residual rotational instability. This study aims to verify if those subjects with high of pre-op knee laxities has also high post-op laxity after an ACL reconstruction.

Material and Methods: The study includes 115 patients, that underwent ACL reconstructions between January 2005 and September 2007. Patients with associated severe ligaments tears or severe chondral defects were excluded. The joint passive kinematics was intra-operatively assessed using the BLU-IGS system (Orthokey, Delaware). We evaluated, before and after the reconstruction, the manual maximum IE rotation at 30° and 90° of flexion, VV rotation at 0° and 30° of flexion and AP displacement at 30° and 90° of flexion. We used the k-means algorithm applied to pre-op values to create two groups among the patients: the GROUP H, with higher pre-op laxity and the GROUP L, with lower pre-op laxity. The pre-op groups were compared for each test using independent Student’s t-test (p=0.01) in order to assess their difference. Student’s t-test (p=0.01) was performed on the corresponding post-op values in order to verify if the difference between H and L was maintained after the reconstruction.

Results: Mean pre-op VV at 0° was 7.1±0.9° for group H and 4.7±0.8° for group L (p< 0.01), post-op was 3.2±0.8° for group H and 2.5±0.8° for group L (p< 0.01). Mean pre-op VV at 30° was 6.2±1.5° for group H and 3.4±0.7° for group L (p< 0.01), post-op was 3.4±1.3° for group H and 2.2± 0.9° for group L (p< 0.01). Mean pre-op IE at 30° was 28.3±3.5° for group H and 19.2±3.1° for group L (p< 0.01), post-op was 21.5±3.8° for group H and 14.7±3.7° for group L (p< 0.01). Mean pre-op IE at 90° was 31.3±2.8° for group H and 22.4±3.5° for group L (p< 0.01), post-op was 22.3±4.0° for group H and 17.0±4.4° for group L (p< 0.01). Mean pre-op AP at 30° was 14.5±2.1mm for group H and 8.9±1.6mm for group L (p< 0.01), post-op was 6.2±1.6mm for group H and 4.2±1.6mm for group L (p< 0.01). Mean pre-op AP at 90° was 11.2±1.7mm for group H and 6.7±1.4mm for group L (p< 0.01), post-op was 5.4±1.8mm for group H and 3.4±1.3mm for group L (p< 0.01).

Discussion: The comparison between group H and group L showed that those patients with higher pre-op laxity had maintained higher post-op values mainly for all the tests. This finding is probably correlated to the possible presence of different tears affecting soft structures of the joint and to the proper and specific anatomy of each patient.

Information on knee kinematics during surgery is currently lacking. The aim of this study is to describe intra-operative kinematics evaluations during uni-compartmental knee arthroplasty (UKA) and total knee arthroplasty (TKA) by mean of a navigation system. Anatomical and kinematic data were acquired by Kin-Nav navigation system and analysed by a dedicated elaboration software developed at our laboratory. The study was conducted on 20 patients: 10 patients undergoing mini-invasive UKA and 10 patients undergoing posterior-substituting-rotating-platform TKA. In both group of patients the surgeon performed passive knee flexion immediately before and immediately after the prosthetic implant. Pattern and amount of internal/external tibial rotation in function of flexion were computed and significant changes between before and after implant were evaluated adopting Student’s t-test (significant level p=0.05).

UKA implant did not significantly change the pattern of internal/external tibial rotation, nor the total magnitude of tibial rotation (15.75°±7.27°) during range of flexion (10°–110°), compared to pre-operative values (17.87°±7.34°, p=0.25). Magnitude of tibial rotation in TKA group before surgery (8.00°±3.67°) was significantly less compared to UKA patients and did not changed significantly after implant (5.96°±4.88°, p=0.09). Pattern of rotation before and after TKA implant were different between each other and between pattern in UKA patients both before and after implant.

Intra-operative evaluations on tibial rotation during knee flexion confirmed some assumptions on knee implants from post-operative methods and suggest a more extensive use of surgical navigation systems for kinematic studies.

Total knee arthroplasty (TKA) is actually a satisfactory technique to reduce pain and enhance mobility in osteoartritic pathologies (OA) of the knee. However, life of the implant is strictly dependent on restoration of correct knee kinematics, as alteration of motion pattern could led to abnormal wear in prosthetic components and also damage soft tissues. The aim of our study was to evaluate new kinematic tests to be performed during surgery in order to improve the standard intra-operative evaluation of the outcome on the individual case. We used Kin-Nav navigation system to acquire anatomic and kinematic data, which were analysed by a dedicated elaboration software developed at our laboratory. Ten patients undergoing rotating platform cruciate substituting TKA were considered for this study. Immediately before the implant and immediately after component positioning, the surgeon performed 3 complete knee flexion imposing internal tibial rotation (IPROM) and 3 complete knee flexion imposing external tibial rotation (EPROM). Tibial rotation during IPROM and EPROM tests was plotted in function of flexion (in the range 10°–110°). Repeatability of IPROM and EPROM was tested by calculating ICC (Intra-class Correlation Coefficient) between 3 repeated curves. Distance between IPROM curve and EPROM curve was computed at various degree of flexion. Maximum distance obtained during all range of flexion before and after the implant were compared by Student’s t-test (significant level p=0.05).

ICC for repeated motions were 0.99 for IPROM and 0.98 for EPROM. Maximum distance between tibial rotation in IPROM and EPROM was 27.82±6.98 before implant and significantly increased (p=0.001) to 40.09±6.92 after TKA. In one case we observed that the value remained similar before and after implant (from 33.11 to 33.98) while in one case we observed very large increase of rotation (from 30.56 to 50.01).

The proposed kinematic tests were able to quantify the increase of tibial rotation after TKA implant. Future development of the study are encouraging and will include a larger sample and reflections on individual findings.

Aims: This work describes a new intraoperative computer-assisted method for the evaluation of joint kinematics in both total (TKA) and uni-compartmental (UKA) knee arthroplasty. We report schematically the protocol and the preliminary in-vivo results we obtained on 11 patients (9 UKA – 2 TKA).

Methods: The system consists of an optoelectronic localizer, 2 reference arrays and a dedicated acquisition software, that permits the real-time control of limb position and allows the acquisition of joint motions. After a first phase of registration (anatomical landmarks identification) the surgeon executes, both before and after the reconstruction, a series of passive tests: range of motion (PROM) evaluation, varus-valgus (VV) stress at 0°, and VV at 30°. Furthermore the surgeon can acquire also anatomical surfaces (tibial plateaus, femoral condyles, prosthetic components, etc.). The 3D kinematic evaluations and anatomical data are recorded before and after the joint reconstruction. This new methodology has been used during 11 interventions fulfilled at our institute. We compare the PROM results with literature, and we also analyzed the interoperator repeatability in the execution of the tests (3 repetitions performed by a senior surgeon).

Results: The kinematic analysis of the PROM showed that there were no significant differences between per-operative and post-operative in all UKA cases. In the 2 TKR cases internal-external (IE) rotations appeared reduced after the implant, but further data are necessary to have a statistical evidence. The extension was improved both in UKA and TKA. The VV laxity at 0 ° was significantly reduced (p < 0.001), while at 30 ° stayed constant (p = 0.010). In all the TKR cases the evaluation of contact areas between femoral and tibial components showed normal pattern, and in UKA the contacts remain inside the prosthesis areas. Measured kinematic parameters (knee rotations, screw-home mechanism and alignment) were comparable with literature and manual estimation at surgical time.

Conclusions: The proposed protocol optimizes surgical times and minimizes invasiveness. The preliminary results showed that the system is able to quantify new kinematic parameters during intraoperative evaluations, provides data about alignments, gaps, stability and 3D motions of the individual knee and therefore can allow an accurate and real-time estimation of the passive knee function. Moreover the new 3D anatomical and kinematic data can improve the biomechanical understanding of the pathological and prosthetic knees.

Aims: In Revision Total Knee Arthroplasty (RTKA), bone deficiencies and lack of anatomical references make it difficult to understand the normal knee kinematic and adequately plan the intervention. To our knowledge there are no data about computer assisted navigation system specifically developed for RTKA in the literature and existing navigated techniques for RTKA use navigation systems developed for primary TKA. A new computer assisted technique for RTKA is presented.

Methods: This technique is based on the use of a navigation system, RTKANav consisting of an optical localizer, a dedicated software and some navigated tools specifically done for RTKA. The system doesn’t use medical images, and patient anatomy model is represented with dots and lines corresponding to acquired landmarks, providing the surgeon with the main references for the intervention monitored in real-time. During the most critical steps of the intervention (soft tissue balancing and the consequent choice of implant size, and joint line height restoration), the system provide the surgeon with graphical and numerical tools to improve the surgical outcome. Several criteria to set each degree of freedom of prosthetic components are considered and compared, and even if some required landmarks can not be identified, the system is always able suggest an intervention plan. The surgeon is provided with tools to analyze and modify the proposed plan, and to reproduce it on the patient.

Results: Till now the presented technique was used on five patients by an expert surgeon. Qualitative results, collected after the intervention through a questionnaire on surgeon feelings, in order to assess the functionality, user friendliness and the data visualization criteria implemented were very satisfying. System reliability was assessed intraoperatively analyzing joint line height, limb alignment and knee stability using trial components: based on his experience, the surgeon checked some acceptable components combination and compared the corresponding outcome with the one provided by the implant planned by the system. In three out five cases the suggested implant was considered the best by the surgeon, while in one case he decided to change the tibial insert of one size because of knee instability and in another case he changed the tibial component of one size because the planned one was too small. Final limb alignment evaluated with postoperative x-rays, was satisfactory in all cases.

Conclusions: Presented navigation system showed early promising results providing the surgeon with intraoperative quantitative and qualitative information on the main surgical parameters, useful to achieve a satisfactory prosthesis implant. Moreover this system use anatomical patient specific landmarks acquired after prosthesys removal, while navigation systems developed for primary TKA use both reference taken from preoperative x-rays and anatomical references acquired on metal component to be removed. Therefore in this case the operation planning is based on rough anatomical landmarks that do not reflect patients original anatomy.

The Vector Vision CT free navigation system of Brain-Lab (Heimstetten, Germany) for total knee arthroplasty incorporates a ligament balancing feature which allows recording of flexion extension gaps and clinical alignment [3]. Since routine spacer blocks do not necessarily load the joint space symmetrically, if either the bone cuts are asymmetric or the ligaments are not evenly balanced, a tensioning device that applies a constant load to the medial and lateral joint space separately and which can collapse or expand on each side independently should be able to provide a better evaluation of ligament tension and allow the computer software to better plan the appropriate bone cuts or ligament release. The tensioning device comprises 2 linked plates contacting the femur and tibia separated by two independent springs in the medial and lateral compartments. It can be positioned precisely in the joint with the navigation system and, with respect to spacer blocks, this device was designed in order to allow a dynamic evaluation of joint stiffness during all range of motion and, thanks to it’s reduced dimensions, with patella in situ. The springs apply a consistent known force on the compartments, while at the same time the gap produced by the applied forces is measured by the navigation. This study integrates previous article [1] on the validation of the tensioning device and reports the first phase of the clinical validation of the tensioning device, including first qualitative comparison with standard navigated technique and consideration on the use of the device.

A spring loaded mechanical device was designed with a constant 6kg load in the springs for each compartment. For the clinical evaluation, the device was inserted into flexion and extension spaces after the tibial cut was performed in routine computer assisted total knee arthroplasty. The gap produced by the applied forces is measured, by the navigation system, as the distance between tibial cut and most distal points on condyles in extension or most posterior points on condyles in flexion. Under the same conditions a set of solid spacer blocks were inserted to obtain a gap able to balance the knee according to surgeon’s sensitivity. This gap was used as reference and compared with the gap obtained with the spring device. The clinical evaluation was performed in order to determine whether there was a difference between the gaps as indicated by both the tensioning device and the spacer blocks. Five experienced orthopaedic surgeons were involved in a randomized study producing 58 complete data sets. Eight measurements (medial and lateral gap, in flexion and extension for tensioning device and spacer blocks) were taken intra-operatively using the ligament balancing features of the VectorVision system. Because many of the measurements were not normally distributed, nonparametric statistical tests (Mann-Whitney and Wilcoxon) were chosen to look for statistical differences, looking for differences in medians and ranking of data instead of differences in averages and distributions. Repeatability of measurements performed with the spring device was defined as the occurrence of same values obtained with the device and by surgeons with spacer blocks. Since the uncertainty in the measurements, due to optical navigation system is near, differences within a range of ±1mm were considered the same value, moreover data within a range of ±2mm were considered a positive result. This data was analyzed with Minitab software version 13.31.

Results are reported in Tab.1.

Results: extension Flexion compartment Medial lateral medial lateral average difference 0.2 (±2.5 mm) 0.0 (±3.6 mm) 1.3 (±2.1 mm) 1.4 (±2.5 mm) number of cases 57 100 57 100 57 100 57 100 median ±1mm 30 52.6 24 42.1 34 59.6 23 40.4 median ±2mm 43 75.4 36 63.2 47 82.5 36 63.2 Table.1. Average difference between gaps obtained with tensioning device and spacer blocks, and related occurrences in medial\lateral compartments inflexion and extension. Difference between the tensioning device and the spacer blocks in flexion is 0.2mm medially and 0.0mm laterally, while in flexion it is 1.3mm medially and 1.6mm laterally, moreover the alignments of resulting femoral cuts obtained with spring device can be considered the same of the alignments obtained with spacer blocks (difference is < 1°). Data, summarized in Tab.1, highlight that knee has a different behaviour in flexion and extension. Applying the same force with the tensioning device the resulting gap in extension (10mm medial, 10.5mm lateral) is lower than the one flexion (10.5mm medial, 12.5 lateral). The percentage of values around the average in the range of 1mm is 52.6% – 59.6% medially and 40.4% – 42.1% laterally, showing a higher variability on lateral compartment, while the percentage of values in a range of 2mm is75.4% – 82.5% medially and 63.2% laterally, confirming the variability.

Determining the soft tissue balance in total knee replacement is important to proper reconstruction. Traditional spacer blocks are unable to load the medial and lateral joint space independently which may compromise the surgical plan. A tensioning device which loads the-joint space independently with a constant load should theoretically allow proper planning of the bone resections and ligament releases during reconstructive surgery. The spring loaded tensor device coupled with image navigation and compared to independent spacer blocks performed by different surgeons revealed that there is no statistical difference between the gaps obtained with spacer blocks and tensioning device in extension, while in flexion there is an average difference 1.4mm, it revealed also that there was greater surgeon variability in the use of spacer blocks compared to the tensioning device. Furthermore, the device produced results that were similar to the results obtained by the spacer blocks especially when surgeon’s variation in technique was taken into account. The use of a joint tensioning device, coupled with computer assisted surgery, allows planning appropriate bone resection and ligament releases to produce matched medial and lateral joint spaces in flexion and extension. There were no reports, either inter-operatively or post-operatively, of any complications or adverse events nor any malfunctions of the device. On a number of occasions it was felt necessary to perform additional bone resections to allow the insertion of the spring device. However this should be considered as a normal part of TKA and of the inter-operative decision making process. This data also revealed, as would be expected clinically that the joint space is less in extension than in flexion after the tibial cut is performed [3]; surgeons with the help of spacer blocks apply less force in flexion in order to obtain the same gap during range of motion, the spring device, applying always the same force, is opening the joint more in flexion. Furthermore, when evaluating the lateral joint space, the tensioning device has a greater variability than the spacer blocks [4]. In this series of patients, a spring loaded tension device decreased surgeon variability in the assessment of ligament tension and, when coupled with computer navigation, allowed the surgeons to appropriately plan the femoral resection to create balanced flexion and extension spaces.

The kinematic effect of tunnel orientation and position, during ACL reconstruction, has been only recently related to the control of rotational instability.

This paper presents a detailed computer-assisted in vitro evaluation of two different femoral tunnel orientations with the same tunnel position, at 10.30 ‘o clock, during the intervention of ACL reconstruction with double bundle technique. Results highlighted better kinematic performances of the horizontal tunnel, with respect to the vertical one, in controlling antero-posterior (AP) laxities at 30°, and internal-external (IE) laxities.

Elongations of anterior and posterior bundles of reconstructed ACL, for both reconstruction, decreased during PROM respectively by 20% and 40%. Total length of the graft varied during PROM, mainly due to graft elongation during tests, graft length on horizontal tunnel varied from 237 to 213mm while graft length on vertical tunnel varied from 257 to 233mm. Kinematic tests showed a better performance of horizontal tunnel in the control of IE rotations at 30° and 90° and of the Lachman test with respect to the vertical one. Stability was restored with both reconstructions.

This study identifies parameters that allow to foresee the necessity of soft tissue release (STR) before surgery. Femoral and tibial morphotype were defined evaluating several radiological parameters. Intra-operative STR during surgery was correlated to radiographic parameters identified. 33 cases were analysed and divided in 2 groups, release (6) no release (27), statistical evaluation has been performed using Mann-Whitney test and contingency tables for most relevant parameters. Three parameters were measured on femur and four on tibia.

The results confirmed the usability of angle between femoral anatomical axis and transepicondylar axis ATA (p< 0.001) and between femoral mechanical axis and tangent to distal condyles MCA (p< 0.001 ) as predictors, among tibial parameters angle between mechanical axis and tangent to tibial plateaux gives good results (p=0.028).The use of contingency tables highlighted that the combined use of ATA and MCA, gives better specificity than the use of a single angle.