Simulation of Spinal Loading: Importance of Subject-Specific Posture and Motion Patterns

Abstract

Introduction Mechanical factors play an important role in the development of spinal disorders (e.g., fracture, disc degeneration). The overall objective of our research is to develop simulation models to investigate the effects of posture and muscular strength on spinal segmental loading, and thus the associated risk of injury.5 The aims of the present study were to further investigate alignment- and kinematics-dependent loading patterns, based on literature and laboratory measures of posture and motion data from young and elderly subjects. Materials and Methods Twenty-four young and 22 elderly subjects volunteered to participate. Average group ages were 27.5 years (standard deviation [SD] = 4.0) and 68.3 years (SD = 3.9), respectively. None of the recruited participants had undergone spinal surgery and none suffered from any back condition. Ethical approval and written consent were obtained. Spinal posture was determined with a noninvasive skin-surface device (SpinalMouse, Idiag). The spinal contour was measured twice in four body positions: (1) standing neutral, (2) erect posture, (3) flexed forward position, and (4) hyperextended posture. A musculoskeletal model of the thoracolumbar spine5 was further developed in the AnyBody Modeling System (AnyBody Technology) combining properties of lumbar models previously established by deZee, 2007 and Han, 2011.6,7 The present model adds a fully articulated thoracic region and ribcage to our prior simulation model. Lumbar motion patterns were derived from fluoroscopy measurements8 during a flexion maneuver in two groups of volunteers: healthy and with low back pain. Three motion patterns were distinguished: (1) all intervertebral joints angulate simultaneously during flexion, (2) upper lumbar joints rotate first and are followed by lower joints in a sequential manner, and (3) angulation of lower levels precedes the upper ones. Segmental forces (compression and shear) were calculated for each motion pattern. Results An average spine curvature, found from the static measurements of the neutral upright posture, was characterized by thoracic kyphosis angles of 49.4 degrees (SD = 11.3) and 45.9 degrees (SD = 6.0) for the young and elderly groups, respectively. Lumbar (L1S1) lordosis angles were 27.4 degrees (SD = 10.6) and 20.4 degrees (SD = 13.1), respectively. Simulations of the sequential and simultaneous motion patterns revealed that fewer muscles and lower muscle activities were necessary to perform lumbar flexion than when the same magnitude of flexion was achieved using the generic spinal rhythm (SR) incorporated in the AnyBody base model (constant ratio of segmental motion). Consequently, compression forces were significantly reduced, by up to 1,200N or 55%, compared with the reference model (Fig. 1). Conclusion The degree of thoracic kyphosis for young subjects measured in our study (49.4 degrees) compares favorably with previous radiographic findings (47.5 degrees). The lumbar lordosis angle that we measured (27.4 degrees) is similar to that reported in other studies with skin-surface devices (23-33 degrees); however, these values are considerably lower than those obtained from methods based on medical imaging (44-63 degrees). Therefore, skin-based systems can be used for subject-specific model definition in the thoracic region, but correction is required for the lumbar region. A surprisingly high variability in the spine curvature was observed, even for young volunteers, highlighting the potential limitations of generic simulation models. Elderly subjects tended toward a lower degree of kyphosis and significantly less lordosis. We have seen in previous simulations that flattening of the back can lead to increased segmental forces during flexion. In contrast to our prior results with arbitrarily defined spinal motion patterns, we have shown that physiological alterations in the temporal sequence of segmental motion can have a substantial influence on muscle recruitment and segmental loading. Compensatory measures, for example, for easing pain, may therefore increase-or decrease-the risk of injury. These results imply that the vertebral kinematics have a profound effect on model predictions, strengthening the necessity to measure and implement realistic vertebral kinematics for spinal motion simulations. In the referenced fluoroscopic study (Okawa, 1998), only a small number of participants were measured, and the motion recording was limited to L2-L5 levels. Therefore, the next step in our investigation will be the combination of the musculoskeletal model with subject specific, whole-spine posture, and motion data from the young and elderly subjects.