Analyse biomécanique de la réductibilité de la scoliose idiopathique par rééducation posturale globale en physiothérapie

Idiopathic scoliosis (IS) is a complex and evolutive three-dimensional (SD) deformation of the spine of unknown cause. It leads to a visible postural asymmetry, which may affect self-esteem and self-confidence, limit recreational activities and impact respiratory capacity. Up to 3 % of the general population suffers from a form of scoliosis; 80 % of the scoliosis cases are idiopathic and the evolutive cases affect primarily young girls during their growth period. The severity of the deformation, which is measured using the Cobb angle, is used to adapt the prescribed treatment. Patients diagnosed with mild scoliosis (Cobb < 25°) are kept under observation while orthopaedic bracing may be prescribed for moderate scoliosis (25° - 45°). Physiotherapy specific scoliosis exercises are considered an alternative or complementary treatment to manage mild to moderate scoliosis. Surgery is recommended for severe deformations (Cobb > 45°). Global postural re-education (GPR) is a physiotherapy approach developed in the 80's in France to treat musculoskeletal disorders, and is part of the specific exercises for scoliosis. This treatment approach is based on active stretching postures and task-oriented exercises. Main treatment objectives are to decrease curve progression, improve posture and eliminate back pain. The postural evaluation in GPR to orient the choice of exercises is based on three steps: general photography, muscular retraction evaluation and re-equilibration. The goal of re-equilibration is to determine the correction capacity and trunk stiffness. More specifically, stand-up manual correction and self-correction postures aim to evaluate patient curve reducibility potential. Manual correction aims to qualitatively verify trunk stiffness and to evaluate the impact of this correction on the overall posture. Self-correction aims to evaluate patient's ability to integrate the corrections by soliciting the muscles of the trunk. GPR applied to scoliosis is poorly documented in scientific literature. To our knowledge, there currently is no tool developed to quantitatively measure trunk stiffness during a physiotherapy treatment for scoliosis. A finite element model (FEM) representing the human trunk anatomical structures was previously developed to simulate immediate scoliosis correction in bracing. This model has the potential to be adapted and modified to simulate other conservative treatments such as physiotherapy manoeuvers. The objective of this study was to evaluate, using finite elements modeling, the curve reducibility of manual correction and self-correction of GPR physiotherapy approach in idiopathic scoliosis. A clinical study with 16 patients presenting with IS (2 juvenile IS and 14 adolescent IS) allowed to experimentally measure GPR re-equilibration postures. For each patient, a FEM was built from surface topography and biplanar low-dose radiographs of a standing reference posture. Gloves equipped with force sensors worn by the therapist allowed to measure the force applied at the hand-torso interface during manual correction. The manual correction was then simulated by reproducing the force measured under the right hand with an equivalent force vector on a node of the ribs in the FEM. A stiffness index was defined as the ratio between the force applied at thoracic apex over the thoracic Cobb angle reduction (forceΔ Cobb angle). In order to be compared to the standing reference posture, the self-correction was clinically measured through surface topography and biplannar radiographs. To simulate the self-correction, the positions of the first thoracic vertebra (T1) and the thoracic and lumbar apical vertebrae were imposed as measured on the self-corrected coronal and sagittal radiographs. The self-correction stiffness index was defined as the reaction force computed at thoracic apex over the thoracic curve reduction (forceΔ Cobb angle). The statistical analysis consisted of an ANOVA with Tukey post-hoc test to compare the Cobb angles in different postures. Correlations were evaluated using the Pearson coefficient. During manual correction, an average corrective force of 31 ± 6 N [23 N – 55 N] was measured under the therapist right hand. The simulation of this posture resulted in an average correction of 26 % (p < 0.05) of thoracic curvature with a non-significant clinical increase of the counter curvature. The thoracic kyphosis increased on average by 3˚ (p < 0.05) while the lumbar lordosis remained unchanged. Manual correction stiffness index varied from 2 to 10 N per degree of correction. Self-correction reduced on average by 33 % the thoracic curvature (p < 0.05) without affecting the lumbar counter curvature, while the sagittal curvatures flattened on average by 6˚ for the kyphosis and 5˚ for the lordosis (p < 0.05). Good correlation (r = 0.9) was obtained between the actual and simulated self-correction, despite the fact that the simulation tends to under-estimate the actual correction by 3˚ on average (p < 0.05). Average reaction force computed at thoracic apical vertebra was 45 N, which resulted in a stiffness index ranging between 0 and 21 N per degree of correction. There was no correlation between curve reduction obtained with manual correction and self-correction. The ratio between the force applied and the thoracic curve reduction in self-correction allowed to define a stiffness index for a relative ranking of patient trunk flexibility. Self-correction stiffness index informs on patient's ability to correct by himself/herself the spine deformation. The lack of correlation between the corrections obtained in the two postures suggests that different correction mechanisms are at work: a passive correction (manual correction) and an active correction through muscular solicitation (self-correction). The FEM computational model is a simplification of reality that allows testing of the effects of controlled parameters. As boundary conditions, the pelvis orientation was fixed in space and a predetermined position of T1 had to be established in the transverse plane. Manual correction could not be verified because it was ethically impossible to take a radiography while maintaining the posture, however the future exploitation of surface topography could eventually lead to a preliminary verification. Vertebral rotation was not measured nor implemented in the simulation. This modeling presents the general force required to maintain the posture of self-correction and manual correction; the different combinations of muscular recruitment to retain these postures are not explicitly represented in the model. The limited number of patients did not allow to establish relationship between the prescribed treatment (bracing, physiotherapy prior to this study or in progress) and the curve reductions measured and simulated in the GPR postures. This study is one of the first to measure corrective force applied by physiotherapist hands during scoliosis GPR treatment and allowed to compute a stiffness index. More studies are required to complete the full validation of the model to simulate manual correction and self-correction. This would ultimately allow to quantify trunk stiffness without the necessity of ionised flexibility test. The definition of a stiffness index to quantify trunk resistance to correction would contribute to treatment planning and physiotherapy follow-up. In addition, the evaluation of self-correction potential finds its relevance in determining the effectiveness of the treatment regarding the integration of the corrected posture. Further research into the biomechanics of GPR correction mechanisms would support the contribution of this approach to treatment.