Qunli Sun, Ph.D^1,4, Fang Lin, DSc^1,2,4, Sam Al-Saeede¹, Lissette Ruberte, MS^1,3, Ellis Nam, MD⁶, Ronald Hendrix, MD⁵, Mohsen Makhsous, PhD^1,b,d

¹Sensory Motor Performance Program, Rehabilitation Institute of Chicago, Chicago, IL, USA
Departments of ²Physical Therapy & Human Movement Sciences, ³Biomedical Engineering,
⁴Physical Medicine & Rehabilitation, and ⁵Radiology, Northwestern University, Chicago, IL, USA
⁶Department of Orthopaedic Surgery, St. Joseph Hospital, Chicago, IL, USA

ABSTRACT

A 3-dimensional (3D) finite element (FE) model of buttock-thigh was developed based on the real geometry acquired using magnetic resonance imaging (MRI) of a subject in an actual joint configuration in seated posture. Non-linear FE analysis was used to predict the distributions of deformation and internal stresses. The results showed that the compressive stresses within the fat and muscle layers are substantially different from the interface pressure. The maximum compressive stress occurred in the gluteus maximus muscle at the location overlaying IT, which was about three times as much as the interface pressure. Within muscle and skin, there were significant shear stresses.

KEY WORDS: Finite Element, Wheelchair, Seating Pressure, Buttock Tissue

INTRODUCTION

Sitting for extended time periods in a wheelchair leads to prolonged high compressive stress in buttock soft tissue over the ischial tuberosities (ITs) and coccyx, which decreases the tissue perfusion and significantly increases the risk of pressure ulcer (PU). It has been demonstrated that different layer of the soft tissue has different responses to the sitting load. However, interface pressure is currently the only available clinical tool to assess sitting load, which provides little information for deep tissues.

FE method is an effective tool to model tissues with complex structure and non-linear material behaviors. Several FE models have been reported to provide useful insight into soft tissue biomechanics in the sitting area (e.g., Ref. (1)). However, some limitations are associated with them. First, all the geometry was oversimplified. The reported MRI scans for capturing the sitting anatomy was performed in supine posture that differs from an actual sitting posture in the hip joint configuration and the tension of the soft tissues. Most of these models were assigned a single material property for all soft tissues, which inevitably introduced unknown discrepancies.

The objective of this investigation was to develop a comprehensive 3D FE model to investigate how the buttock-thigh soft tissue responds to external sitting load. The FE model was developed based on MRI images of the buttock-thigh in a true joint configuration of a seated individual. It was expected that this FE model would not only be a powerful tool for better understanding of the biomechanical response of buttock tissue to sitting load, but also facilitate development of improved prevention and treatment strategies to PU formation to promote the quality of life of people with impaired mobility.

Figure 1 d: Loading apparatus (upper row) and typical saggital MRI images of buttock-thigh structure (lower row). The loading apparatus was used to image the buttock-thigh structure in real sitting configuration. The apparatus consisted of a rubber cushion, foam support, and a rectangular air bladder. The air bladder was placed between the cushion and buttocks. Two belts were used to tie the cushion and bladder tightly against the buttock-thigh. Left figures show the apparatus (upper) and a corresponding typical MRI image (lower) in “No Load” condition in which the bladder was deflated. Right figures show the apparatus (upper) and a corresponding typical MRI image (lower) when the buttock-thigh was loaded (Sitting Load) by inflating the bladder to simulate sitting load. (Click image for larger view)

METHODS

MRI provides high-resolution images for 3D bio-structures. However, the limited space of the MRI scanner makes it difficult to scan buttock-thigh structure in real sitting posture. Therefore, a special loading apparatus was developed for imaging buttock-thigh when the subject was in an actual sitting posture with sitting load applied (Fig.1, upper row). The apparatus consisted of a rubber cushion, foam support, and a rectangular air bladder between the cushion and buttocks. Two belts were used to tie the cushion and bladder tightly against the buttock-thigh. The sitting load was applied by inflating the bladder.

A healthy subject (male, 24 yr, 165 cm, 70 kg) was posed and loaded in the apparatus at a simulated sitting posture, i.e. with hip flexed at 80º and knee at 90º. MRI scan was performed with 2 loading conditions: “No Load” (Fig.1, Upper Left) and “Sitting Load” (Fig.1 Upper Right). Before the subject was transported into the MRI tube, the load was measured by a pressure mapping system to ensure that it was comparable to the recording from an actual upright sitting individual in a wheelchair. For comparing the images under two loading conditions, MRI sensitive markers were attached on the skin. Volumetric saggital MR images of the right buttock-thigh (576x576 matrix, 35 cm FOV, 288 slices, 0.6 mm thickness) were acquired (Fig. 1, Lower row) on a 1.5T scanner in the two loading conditions.

The MRI image sequence under “No Load” condition was segmented to reconstruct the geometry for pelvis, muscles, fat and skin. Based on the segmented surfaces, the FE mesh of the buttock-thigh was created for femur, pelvis, muscles, fat and skin with tetrahedral elements using HyperMesh. A total of 453,502 4-noded modified tetrahedral elements were created to mesh all structures (Fig. 2).

Three different boundary conditions (BC’s) were considered. The femur and pelvis were assumed as rigid bodies and constrained from motion. Considering the simulation was focused on the seating area, the upper plane was assumed to be fully constrained, while the longitudinal ends of muscles and skin that connect the rests of thigh or buttock structures were constrained against longitudinal motions. The part of medial plane of the model was assumed to be constrained against the medial-lateral motion due to the symmetry of the buttocks. However, the rest of the medial plane was left free to allow the horizontal motion. The measured interface contact pressure (157-160mmHg) under loaded sitting configuration was used to apply external pressure on the skin in the sitting area.

Figure 2 d: FE model of the buttock-thigh based on MRI images. Bones and each soft tissue structure are shown in different colors, while fat is displayed transparently. (Click image for larger view)

All soft tissues were modeled as neo-Hookean material that allows large deformation. The material model employs two parameters that can be calculated from the initial Young’s modulus and Poisson’s ratio. All tissues were assumed to have a Poisson’s ratio of 0.485, representing a nearly-incompressible material response. The Young’s modulus for skin, fat and muscle was assumed to be 0.85 (2), 0.01 (3) and 0.126 MPa (4), respectively. The static FE analysis was performed using ABAQUS 6.4. The results were examined for three regions of the tissue in the sitting area. The Regions A and B were chosen such that the three layers of soft tissues below IT and femur formed a cylindrical volume with a cross-section of the bone-tissue contact area, respectively, in a direction parallel to the external interface pressure-load. The Region C was chosen such that the three layers of tissues were not below any bone and formed a cylindrical volume with a cross-section of about 4 cm 2.

RESULTS

**Table 1:** The compressive and shear stress levels in different soft tissues: The stress was examined for three regions of the tissue layers in the sitting area. The Regions A and B were chosen such that the three layers of soft tissues below IT and femur formed a cylindrical volume with a cross-section of the bone-tissue contact area, respectively, in a direction parallel to the external pressure-load. The Region C was chosen such that the three layers of tissues were not below any bone and formed a cylindrical volume with a cross-section of about 4 cm 2.
Region	Layer	Compressive Stress (mmHg)		Shear Stress (mmHg)
Region	Layer	Mean ± SD	Max	Mean ± SD	Max
Under IT (A)	Muscle	316.9 ± 99.1	568.6	144.4 ± 28.2	205.3
	Fat	182.5 ± 7.1	200.8	20.7 ± 1.3	25.3
	Skin	280.4 ± 81.8	527.9	41.3 ± 13.5	67.3
Under Femur (B)	Muscle	245.6 ± 81.8	458.7	96.7 ± 13.6	136.4
	Fat	189.1 ± 20.1	233.2	42.5 ± 7.5	56.5
	Skin	325.2 ± 91.8	486.9	159.4 ± 26.9	234.8
No Bone Above (C)	Muscle	206.3 ± 22.5	262.9	96.9 ± 7.9	117.5
	Fat	183.3 ± 21.7	252.3	24.8 ± 4.3	38.6
	Skin	271.2 ± 58.9	419.4	75.8 ± 15.2	133.9

Figure 3 d: Compressive stress distribution in each soft tissue structure from FE analysis. The left, middle, and right panels show the compressive stress distributions in muscle, fat, and skin, respectively. Different compressive stress levels are displayed in color contours that are numerically indicated in the legend on the left. High compressive stresses are shown in red spots. Regions A and B represent the soft tissues in muscle, fat, and skin below IT and femur, respectively. Region C represents the soft tissues in muscle, fat, and skin without any bone above them. Note that the red spots near constraints should not be taken into account because the constraints may not be the real constraints and it may not represent the real condition. (Click image for larger view)

The results of the simulations showed that the internal compressive and shear stress distributions were substantially different among muscles, fat, and skin. The maximum compressive stress occurred in the muscle under IT (in Region A; Table 1 and Fig. 3), which is consistent with the experimental observation from other researchers (5). The compressive stress in fat layer was distributed relatively uniform, while in muscle and skin layers, the soft tissue under bony prominences bore substantially higher internal compressive stress (Table 1 and Fig. 3). The maximum shear stress occurred within muscle “Under IT” and skin “Under Femur”. The shear stress within muscle and skin were significantly higher than that in fat. This supports the conclusion that the shear stress is one of factors to form the PU’s in the skin by other researchers (6).

The simulation also showed that the layered soft tissues were deformed significantly and the deformation was not uniform among the soft-tissue layers of the sitting area: The maximum deformation involved the deep layer, i.e. muscular layer, which was consistent to the observation obtained from the unloaded and loaded MRI images (Fig. 1, Lower Row).

DISCUSSION

The study indicates that tissue necrosis may first occur in the deep buttock-thigh soft tissue of wheelchair users because the higher internal pressure would clog the vascular system and thus reduce or terminate the supply of blood. The high shear stress within the skin and muscle may contribute to the formation of PU’s in skin and muscle. In the study, the tetrahedrons were used to handle the extremely complex structures. The mesh density analysis showed that at least two layers of elements along thickness direction should be used for the skin to avoid shear locking and inaccurate solution. For other layers, appropriately fine mesh should be used. Several BCs were tested to mimic the actual conditions of different layers as much as possible. Unreasonable BCs would produce irrational results in the sitting area. For instance, a free BC at the two longitudinal edges of skin resulted in unreasonable high shear stresses within skin because inappropriately-large sliding motion was allowed between skin and fat, which did not represent the real sitting condition. The FE model will be further improved by (i) involving the cushion and contact analysis in the model; (ii) providing substantial validation of the model through comparing the spatial deformations of layered soft tissue between FE results and MRI images; (iii) incorporating more realistic material properties in the model. The improved FE model is expected to provide a reliable tool to predict the normal biomechanical response, the outcomes from potential interventions, and the interaction between the buttock-thigh and cushion.

REFERENCES

Todd, B. A. & Thacker, J. G. (1994). Three-dimensional computer model of the human buttocks, in vivo. J Rehabil Res Dev, 31(2), 111-119.
Agache, P. G., Monncur, C., Leveque, J. I., & De Rigal, J. (1980). Mechanical properties and Young's modulus of human skin in vivo. Arch Dermatol Res, 269, 221-232.
Maaß, H. & Kühnapfel, U. (1999). Noninvasive Measurement of Elastic Properties of Living Tissue. Proc 13th Internat. Congress on Computer Assisted Radiology and Surgery, Paris, France. Elsevier Science Amsterdam.
Levinson, S. F., Shinagawa, M., & Sato, T. (1995) Sonoelastic determination of human skeletal muscle elasticity. J Biomech, 28(10), 1145-1154.
Linder-Ganz, E. & Gefen, A. (2004). Mechanical compression-induced pressure sores in rat hindlimb: muscle stiffness, histology, and computational models. J Appl Physiol, 96(6), 2034-2049.
Bouten, C.V., Oomens, C. W., Baaijens F. P., & Bader D. L. (2003). The etiology of pressure ulcers: skin deep or muscle bound? Arch Phys Med Rehabil, 84, 616-619

ACKNOWLEDGEMENT

The project was supported in part by PVA Award #2321-01, R24 Rehab Network, NIH Award #R21 HD046844-01A1, NIOSH Award #R21 OH007737, Medical Technology Systems and Falk Medical Research Trust.

Correspondence Author:

Mohsen Makhsous
Rehabilitation Institute of Chicago
345 E. Superior St., Suite 1406
Chicago, IL 60611
Phone 312-503-0073
Email: m-makhsous2@northwestern.edu