Knee contact force in subjects with symmetrical OA grades: Differences between OA severities

Abstract

In using musculoskeletal models, researchers can calculate muscle forces, and subsequently joint contact forces, providing insight into joint loading and the progression of such diseases as osteoarthritis (OA). The purpose of this study was to estimate the knee contact force (KCF) in patients with varying degrees of OA severity using muscle forces and joint reaction forces derived from OpenSim. Walking data was obtained from healthy individuals (n=14) and those with moderate (n=10) and severe knee OA (n=2). For each subject, we generated 3D, muscle-actuated, forward dynamic simulations of the walking trials. Muscle forces that reproduced each subject’s gait were calculated. KCFs were then calculated using the vector sum of the muscle forces and joint reaction forces along the longitudinal axis of the femur. Moderate OA subjects exhibited a similar KCF pattern to healthy subjects, with lower second peaks (p=0.021). Although subjects with severe OA had similar initial peak KCF to healthy and moderate OA subjects (more than 4 times BW), the pattern of the KCF was very different between groups. After an initial peak, subjects with severe OA continually unloaded the joint, whereas healthy and moderate OA subjects reloaded the knee during late stance. In subjects with symmetric OA grades, there appears to be differences in loading between OA severities. Similar initial peaks of KCF imply that reduction of peak KCF may not be a compensatory strategy for OA patients; however, reducing duration of high magnitude loads may be employed.

Introduction

Osteoarthritis (OA) is the most common form of arthritis, with the knee being the most affected joint. A combination of biochemical, biomechanical, and neuromuscular factors are thought to lead to the development and progression of OA (Felson, 2000). The progression of OA is often accompanied by pain and may result in changes in gait and neuromuscular function, which may in turn lead to increased wear on the joint and further progression of the disease (Astephen et al., 2008).

Gait analysis has provided significant information about biomechanical changes in OA (Astephen et al., 2008, Hunt et al., 2006, Baliunas et al., 2002, Sharma et al., 1998, Thorp et al., 2006, Zeni and Higginson., 2009). Studies have shown that altered loads on the articular cartilage due to obesity, increased knee laxity, decreased proprioception, increased age, increased knee adduction moments, and increased knee varus/valgus increase the risk of OA (Zhao et al., 2007, Zhao et al., 2007, Mundermann et al., 2005).

Instrumented tibiofemoral implant studies provide valuable in vivo loading information (Mundermann et al., 2008, Varadarajan et al., 2008, D'Lima et al., 2008, Zhao et al., 2007, Zhao et al., 2007, Kim et al., 2009). These studies have shown peak knee contact forces (KCF) ranging from 1.6 to 3.5 times body weight (BW) for self-selected speed walking (Mundermann et al., 2008, Zhao et al., 2007, Zhao et al., 2007, Kim et al., 2009). These studies, however, provide limited joint loading data pertaining to those individuals needing total knee arthroplasty, and require surgery for implantation, even though the data can be recorded and retrieved non-invasively post-surgery.

Recent advances in musculoskeletal modeling and computation power have enabled researchers to generate gait simulations in efforts to estimate muscle forces, and subsequently estimate joint contact forces (e.g. Kim et al., 2009, Winby et al., 2009). In rare cases, predictions of contact forces were validated with instrumented prostheses data and range from 1.9 to 3.5 BW at the tibiofemoral joint (Kim et al., 2009). KCFs during walking at self-selected speeds averaged 3.9 BW for healthy females and 3.4 BW for healthy males (Kuster et al., 1997), and exceeded 4.0 BW when using EMG-driven models (Winby et al., 2009). Taylor et al. (2004) used scaled, whole body models to calculate KCF during walking at self-selected speeds and showed forces averaging 3.1 BW.

Few studies have used computational modeling to calculate KCF during walking in patients with OA. Using a statically determinant model, Henriksen et al. (2006) compared KCF estimations between OA and healthy subjects and found significant differences. The average peak KCF calculated during early single limb support was 1.8 BW for OA subjects and 2.4 BW for healthy subjects, and 1.6 and 1.9 BW during late single limb support for OA and healthy subjects, respectively. However, they grouped all patients with radiographic evidence of OA into one group and compared them to a healthy control group (Henriksen et al., 2006). Several studies have reported differences in KCF when asymmetric loading conditions exist during walking such as with unilateral hip osteoarthritis or following joint replacement (Levinger et al., 2008, Milner, 2008, Shakoor et al., 2003), however none have addressed loading conditions for subjects with symmetric OA grades who likely use unique compensatory strategies.

The purpose of this study was to estimate the net KCF in healthy adults and those with increasing severity of symmetric knee OA using 3D, subject-specific, muscle-driven gait simulations generated using OpenSim (Delp et al., 2007). We hypothesized that individuals with increased OA severity exhibit decreased net KCF consistent with slower self-selected walking speeds.