Introduction

Previous fluoroscopic studies of total knee arthroplasty (TKA) have revealed significant kinematic differences compared to the normal knee. Often, subjects having a TKA experienced kinematic patterns opposite of the normal knee. Therefore, the objective of this study was to determine the in vivo kinematics of subjects implanted with either a customized-individual-made (CIM) or the traditional (OTS) PS TKA to determine if customization offers a distinct advantage to the patient.

Introduction

Historically, knee implants have been designed using average patient anatomy and despite excellent implant survivorship, patient satisfaction is not consistently achieved. One possibility for this dissatisfaction relates to the individual patient anatomic variability. To reduce this inter-patient variability, recent advances in imaging and manufacturing have allowed for the implementation of patient specific posterior cruciate retaining (PCR) total knee arthroplasty (TKA). These implants are individually made based on a patient's femoral and tibial anatomy determined from a pre-operative CT scan. Although in-vitro studies have demonstrated promising results, there are few studies evaluating these implants in vivo. The objective of this study was to determine the in vivo kinematics for subjects having a customized, individually made(CIM) knee implant or one of several traditional, off-the-shelf (OTS) TKA designs.

Introduction

Previous fluoroscopic studies compared total knee arthroplasty (TKA) kinematics to normal knees. It was our hypothesis that comparing TKA directly to its non-replaced controlateral knee may provide more realistic kinematics information. Using fluoroscopic analysis, we aimed to compare knee flexion angles, femoral roll-back, patellar tracking and internal and external rotation of the tibia.

Subjects having a posterior cruciate ligament sacrificing (PCLS) mobile bearing TKA seem to experience less translation during gait, but often achieve less weight-bearing flexion. More recently, posterior stabilisation has been added to PCLS mobile bearing TKA, hoping to increase flexion. Therefore, the objective of this multi-center study was to determine the in vivo kinematics for subjects implanted with a mobile bearing PS TKA that attempts to maintain high contact area.

Subjects with 10 TKA from 2 surgeons were asked to perform maximum weight-bearing flexion (deep knee bend (DKB)) and gait while under fluoroscopic surveillance. During weight bearing flexion, the 3-D kinematics of the TKA were determined by analyzing fluoroscopic images in the sagittal plane at 30 degree increments. Fluoroscopic images taken in the frontal plane from four increments during the stance phase of gait were analyzed.

The average weight-bearing flexion was 116 degrees and the average medial and lateral anteriorposterior (AP) translation was posterior with −1.9 mm and −5.4 mm, respectively, from full extension to maximum weight-bearing flexion.

The average femorotibial axial rotation from full extension to maximum weight-bearing flexion was 3.9 degrees. During the stance phase of treadmill gait, patients experienced 0.8 mm (0.1 mm to 2.3 mm, SD=0.8 mm) of “pure” mediolateral translation of the femur relative to the tibia. The femorotibial axial rotation was 4.6 degrees from heel-strike to toe-off (Table 3).

The posterior femoral rollback and axial rotation patterns were similar to the normal knee, albeit experiencing less overall motion. More noticeably, subjects in this study experienced a significantly greater weight-bearing flexion than previous subjects analyzed with a mobile bearing PCLS TKA and more reproducible “fan-like” patterns, where the lateral condyle rolled greater posteriorly than the medial condyle.

All over the world, obesity rates are on the rise. Medical complications and increased health risks are often associated with being overweight or obese, but a thorough understanding of in vivo motions for obese, overweight and normal weight subjects does not exist. Therefore, the objective of this study was to compare knee kinematics in TKA subjects by body mass index (BMI).

In vivo knee kinematics were determined for 253 TKA subjects during a Deep Knee Bend (DKB) from full extension to maximum flexion using a 3D to 2D image registration technique. Each of these subjects was then classified into one of three BMI categories: obese (BMI greater than or equal to 30), overweight (BMI greater than or equal to 25 and less than 30) and normal weight (BMI less than 25 and greater than or equal to 18.5). Subjects were provided by 11 surgeons using ten different TKA devices. All subjects were deemed clinically successful.

On average, weight bearing range of motion (ROM) for the obese (n=79), overweight (n=113) and normal weight (n=61) groups were 107.7° (range: 74° to 136°, standard deviation (σ) =14.9°), 109.6° (60° to 150°, σ=17.5°) and 114.1° (72° to 147°, σ=14.4), respectively. ROM of 90° or less was seen in 16.5% of the obese subjects, 14.2% of the overweigh subjects and 6.6% of the normal weight subjects. ROM of 125° or more was seen in 15.2% of the obese subjects, 16.8% of the overweight subjects and 23.0% of the normal weight subjects.

From full extension to maximum flexion the obese, overweight and normal weight groups averaged 8.65° (−5.14° to 22.51°, σ=6.22°), 7.58° (−2.85° to 24.72°, σ=5.71°) and 5.72° (−4.84° to 19.43°, σ=5.65°) of axial rotation. Axial rotation of 3° or less was seen in 20.25% of the obese subjects, 23.01% of the overweight subjects and 39.34% of the normal weight subjects. Axial rotation of greater than 9° was seen in 51.90% of the obese subjects, 35.40% of the overweight subjects and 26.23% of the normal weight subjects. Opposite axial rotation was seen in 8.86% of the subjects in the obese group, 9.73% of the overweight group and 9.84% of the normal weight group.

On average, from full extension to maximum flexion, the medial condyle for the obese, overweight and normal weight groups experienced −5.44mm (−22.20mm to 8.04mm, σ=7.9mm), −6.30mm (−25.22mm to 5.35mm, σ=7.36mm) and −4.78mm (−20.79mm to 5.49mm, σ=6.68mm) of posterior femoral rollback (PFR), respectively. The obese, overweight and normal weight groups averaged −12.66 mm (−34.57mm to 0.34mm, σ=9.32mm), −12.38mm (−36.72mm to 1.83mm, σ=10.33mm) and −9.39 mm (−34.55mm to 0.35mm, σ=8.98mm) of lateral PFR, respectively.

Condylar lift-off of greater than 1mm was seen in 16.46% of obese subjects, 10.62% of overweight subjects and 11.48% of normal weight subjects.

Various statistical differences were seen across the groups. The normal weight subjects had significantly higher ROM that the obese subjects (p=0.0184), while there was no difference seen between the normal weight and overweight groups or the overweight and obese groups. The obese and the overweight groups had significantly more axial rotation than the normal weight group from 0° to 90°, 0° to maximum flexion, 30° to 90°, 30° to maximum flexion and 60° to 90°. There were a significantly higher number of cases of condylar lift-off for obese subjects when compared to both normal weight and overweight groups.

It can be concluded that body mass index does play a factor in TKA kinematics.

Many nonoperative techniques exist to alleviate pain in unicompartmental osteoarthritic knees including physical therapy, heel wedges and off-loading knee braces [1]. Arthritic knee braces are particularly effective since they can be used on a regular basis at home, work, etc. Previous knee brace studies focused on their ability to stabilize anterior cruciate ligament (ACL) deficient knees. A standard technique for analyzing brace effectiveness is the use of an athrometer to look at the range-of-motion. Although this is helpful, it is more useful to use X-ray or fluoroscopy techniques to analyze the in vivo 3-D conditions of the femur and tibia. One method for doing this is Roentgen Steroephotogrammetric Analysis, which uses a calibration object and two static X-rays to perform 3-D registration of the femur and tibia. This technique is limited to static and typically non-weight bearing analysis.

We have analyzed five patients with moderate to severe osteoarthritis in both step up and step down activities with two different knee braces and also without a knee brace. Fluoroscopy of the five patients performing these activities was obtained as well as a CT scan of the knee joint for each patient. 3-D models of the femur and tibia were obtained from manual segmentation and overlaid to the fluoroscopy images using a novel 3-D to 2-D registration method [2]. This allowed analysis of 3-D in vivo weight bearing conditions. This work builds off of an analysis where 15 patients were analyzed in vivo during gait with and without knee braces [3].

All five patients experienced substantially less pain when performing the step up and step down activities with a knee brace versus without a knee brace. It should be noted that none of the five patients were obese, which can limit brace effectiveness. Preliminary results show that medial condyle separation was increased by 1.4–1.6 mm when using a knee brace versus not using a knee brace during the heel-strike and 33% phases of step up and step down activities. Also, the condylar separation angle was reduced by an average of 1.5–2.5°. Finally, consistently less condylar separation was seen during step down versus step up activities (0.5–1 mm), which can be attributed to a greater initial impact force on the knee joint during step down versus step up activities.

Previous fluoroscopic analyses of Total Hip Arthroplasty (THA) determined that the femoral head slides within the acetabular cup, leading to separation of certain aspects of the articular geometries. Although separation has been well documented, it has not been correlated to clinical complications or a more indepth understanding of the cause and effect. Surgical technique is one of the important clinical factors when considering THA procedures, and it is hypothesized, that it could affect the magnitude and occurrence of femoral head separation (sliding) in THAs. Hence, the objective of this study was to determine and compare in-vivo THA kinematics for subjects implanted with a THA using two different surgical approaches.

Thirty seven subjects, each implanted with one of two types of THA were analysed under in vivo, weight-bearing conditions using video fluoroscopy while performing a sit-to-stand activity. Ten subjects were implanted by Surgeon 1 using a long incision postero-lateral approach (G1); while a further 10 subjects were implanted by the same surgeon using a short incision posterolateral approach (G2). The remaining 17 subjects were implanted using the anterolateral approach; 10 by Surgeon 2 (G3) and seven by Surgeon 3 (G4). All patients with excellent clinical results, without pain or functional deficits were invited to participate in the study (HHS > 90). 3D kinematics of the hip joint was determined, with the help of a previously published 2D-to-3D registration technique. From a completely seated position to the standing position, four frames of the fluoroscopy video were analysed.

Subjects in all groups experienced some degree of femoral head separation at all increments of the sit-to-stand activity that were analysed. The magnitude and frequency of separation greater than 1.0mm varied between each surgeon group, between incision types, between incision lengths and between the two types of THA that were analysed. The average maximum separation was 1.3, 1.1, 1.3 and 1.4mm for G1, G2, G3 and G4 respectively. Though there was no difference in the average maximum separation values for the 4 groups, the maimum separation varied significantly. While the maximum separation in G2 was 1.8mm, the maximum separation in G4 was 3.0mm. G1 and G3 had maximum separation values of 2.3mm and 2.4mm respectively.

This study suggests that there may be a correlation between incision lengths and surgical approach with femoral head separation in THAs. The maximum separation that was seen among all groups was a subject with a traditional long incision, while the short incision group had less incidence of separation. Results from this study may give researchers and implant developers a better understanding of kinematics around the hip joint and how they vary with respect to different surgical techniques. Further analysis is being conducted on the subjects before definitive conclusions can be made.

Previous in vivo studies have not documented if ethnicity or gender influence knee kinematics for the healthy knee joint. Other measurements, such as hip-knee-ankle alignment have been previously shown to be significantly different between females and males, as well as Japanese and Caucasian populations in the young healthy knee [1]. Differences in knee kinematics in high flexion positions may relate to both etiology of osteoarthritis and success in knee replacement designs. Although differences in knee anatomy have been identified, their significance in knee function has not yet been clarified. Therefore, the objective of this study was to determine the 3D, in vivo normal knee kinematics for various subjects from different gender and ethnic backgrounds, and to identify significant differences, if any, between populations.

The 3D, in vivo, weight bearing normal knee kinematics was determined for 79 healthy subjects, including 48 Caucasians, 24 Japanese, 42 males, and 37 females. Each participant performed deep knee bend activity from a standing (full extension) to squatting to a lunge motion, until maximum knee flexion was reached. The study was approved by the Institutional Review Board and informed consent form was obtained from all subjects. The 3D bone models, created by segmentation from MR images, were used to recreate the 3D knee kinematics using the previously described fluoroscopic and 3D-to-2D registration techniques (Fig. 1) [2,3]. Tibiofemoral rotations were described using the ISB recommended Grood and Suntay convention [4,5]. Anterior-posterior translations of the centers of the posterior femoral condyles were normalized due to significantly different anthropometry in the subjects. Anterior cruciate ligament (ACL) laxity was also measured using a KT-1000 device for 72 of these subjects. Statistical analysis was performed using the Student’s t-test, set at the 95% confidence interval.

Most subjects achieved very high flexion, however substantial variability occurred in all groups. Range of motion (ROM) varied from 117° to 177°, while average external rotation was 31°± 9.9° for all subjects. Japanese and female subjects achieved greater ROM than Caucasian (p=0.048) and male (p=0.014) subjects. From full extension to 140° of flexion (which 87% of subjects achieved), few significant differences between any of the populations were observed. At deeper flexion, the external rotation was higher for female than for male subjects, however not statistically significant (p=0.0564 at 155°). Also at deep flexion, the adduction was significantly higher for female subjects. The translations of the lateral condyle were very similar between respective groups, but at deep flexion, the medial condyle remained significantly more anterior for females, leading to greater axial rotation and ROM. As ACL laxity increased, flexion/extension ROM significantly increased (r2=0.184, p< 0.001). In addition, ACL laxity was also higher for females (6.8 mm) compared to males (5.6 mm, p=0.011), as well as Japanese (7.5 mm) compared to Caucasian (5.6 mm, p=0.0002) subjects.

High variability and ROM in knee kinematics were similar to those seen in previous studies of healthy subjects during a deep knee bending activity [6]. Subjects in this study achieved much greater axial rotation and ROM than previously analyzed TKA patients. A relationship was found between greater axial rotation and increased ROM, and may be related in part to increased ACL laxity in the knee. Significant differences in ROM and laxity were identified between genders and ethnic groups. Also the medial condyle remaining significantly more anterior for females than for males in deep flexion may explain higher external rotation and consequently higher flexion experienced by women. However, understanding the causes for variability within each group may be the key to improved implant design.