Musculoskeletal model choice influences hip joint load estimations during gait
Introduction
Hip joint loading during activities of daily living such as walking and stair climbing is essential for maintaining healthy bone structure. The relationship between inadequate lower extremity loading and poor bone density, particularly in aging females, is well established (Jamsa et al., 2006, Vainionpaa et al., 2007). However, excessive hip joint loading may increase the likelihood of developing osteoarthritis in healthy hips (Felson, 2013). The direct measurement of internal hip forces developed during human movement, however, is difficult to achieve for practical and ethical reasons. In vivo hip contact forces and stresses have been recorded via instrumented prostheses in a limited number of patients (Bergmann et al., 2001, Bergmann et al., 1993, Damm et al., 2010, Schwachmeyer et al., 2013), and in several cases can be retrieved from the public database www.OrthoLoad.com. Recently, in silico, musculoskeletal models have been employed to estimate hip contact forces (Bergmann et al., 2016, Giarmatzis et al., 2015, Heller et al., 2001, Modenese and Phillips, 2012, Shelburne et al., 2010b, Weinhandl et al., 2017). These musculoskeletal models can potentially be implemented without the need for patient medical images. The noninvasive nature of musculoskeletal modeling thereby allows for force estimation with reduced subject risk and financial cost.
In recent years, software packages such as OpenSim (Delp et al., 2007) and Anybody (Damsgaard et al., 2006) have made possible the sharing and distribution of musculoskeletal models. As the number of musculoskeletal modeling based investigations continues to increase, the uniformity in results derived from different models is becoming more critical. Predicted knee contact forces during an idealized knee-extension task have been compared between several, commonly used generic musculoskeletal models (Wagner et al., 2013). They reported that simple scaling and usage of the same objective function for muscle force prediction was not sufficient to produce consistent muscle and knee joint contact forces between models. However, the consistency of various models at predicting hip joint contact forces remains unknown and differences in model predicted hip contact forces between studies are difficult to distinguish from natural inter-individual differences. As such, with the increased prevalence of musculoskeletal modeling studies investigating hip contact forces it is vital to understand the differences between commonly employed models. While previous models have been validated via instrumented prosthesis (Heller et al., 2001, Modenese et al., 2011), these models have not been compared to one another. The purpose of this study was therefore to evaluate differences in hip joint contact forces during gait between four OpenSim models.
Section snippets
Methods
Four musculoskeletal models, freely available in OpenSim (http://opensim.stanford.edu/) were evaluated in this study. These models included two generic gait models and two hip specific models. The generic models were gait2392 (Delp et al., 1990) and the Arnold Lower Limb Model (ALLM) (Arnold et al., 2010). The first hip specific model was hip2372 (Shelburne et al., 2010a, Shelburne et al., 2010b). This model is based on gait2392 with the addition of the obturator and rectus abdominus muscles.
Results
Ensemble hip joint contact forces for each model and the HIP98 dataset for slow, normal, and fast walking are presented in Fig. 2. Individual patient ensemble hip joint contact forces for each model are presented in Fig. 3. Pearson’s correlation coefficients indicate a strong correlation between experimentally measured and simulated hip contact forces for all models regardless of speed (Table 2). During slow walking trials LLLM yielded a higher correlation coefficient than ALLM, hip2372, and
Discussion
The objective of this study was to evaluate differences in hip joint contact force during gait between four OpenSim models: gait2392 (Delp et al., 1990), ALLM (Arnold et al., 2010), hip2372 (Shelburne et al., 2010a, Shelburne et al., 2010b), and LLLM (Modenese et al., 2011). To accomplish this objective, gait of four individuals who have had a total hip replacement with an instrumented prosthesis (Bergmann et al., 2001) was simulated using each model. The simulated hip joint forces were then
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of Competing Interest
The authors have no conflict of interest related to the present work to disclose.
References (45)
- et al.
Static and dynamic optimization solutions for gait are practically equivalent
J. Biomech.
(2001) - et al.
A normative database of hip and knee joint biomechanics during dynamic tasks using anatomical regression prediction methods
J. Biomech.
(2018) - et al.
Hip contact forces and gait patterns from routine activities
J. Biomech.
(2001) - et al.
Hip joint loading during walking and running, measured in two patients
J. Biomech.
(1993) - et al.
Total hip joint prosthesis for in vivo measurement of forces and moments
Med. Eng. Phys.
(2010) - et al.
Analysis of musculoskeletal systems in the AnyBody Modeling System
Simul. Model. Pract. Theory
(2006) - et al.
Variation of rotation moment arms with hip flexion
J. Biomech.
(1999) Osteoarthritis as a disease of mechanics
Osteoarth. Cartil.
(2013)- et al.
Musculo-skeletal loading conditions at the hip during walking and stair climbing
J. Biomech.
(2001) - et al.
Effect of daily physical activity on proximal femur
Clin. Biomech. (Bristol, Avon)
(2006)