A rigid thorax assumption affects model loading predictions at the upper but not lower lumbar levels

A number of musculoskeletal models of the human spine have been used for predictions of lumbar and muscle forces. However, the predictive power of these models might be limited by a commonly made assumption; thoracic region is represented as a single lumped rigid body. This study hence aims to investigate the impact of such assumption on the predictions of spinal and muscle forces. A validated thoracolumbar spine model was used with a flexible thorax (T1-T12), a completely rigid one or rigid with thoracic posture updated at each analysis step. The simulations of isometric forward flexion up to 80°, with and without a 20 kg hand load, were performed, based on the previously measured kinematics. Depending on the simulated task, the rigid model predicted slightly or moderately lower compressive loading than the flexible one. The differences were relatively greater at the upper lumbar levels (average underestimation of 14% at the T12L1 for flexion tasks and of 18% for flexion tasks with hand load) as compared to the lower levels (3% and 8% at the L5S1 for unloaded and loaded tasks, respectively). The rigid model with updated thoracic posture predicted compressive forces similar to those of the rigid model. Predicted muscle forces were, however, very different between the three models. This study indicates that the lumbar spine models with a rigid thorax definition can be used for loading investigations at the lowermost spinal levels. For predictions of upper lumbar spine loading, using models with an articulated thorax is advised.