Bedsores … aka pressure sores … aka pressure ulcers, continue to slow the healing process for bedridden patients. (For purposes of this article, pressure sores will be the accepted term.) And while some mattress manufacturers claim to have solved the problem, researchers into body mechanics are finding that the truth lies deeper than that.

“Just because something feels good on the surface doesn't mean it's right for the deep tissues of the human body,” says Gerhard Silber, professor of Materials Science at the Center of Biomedical Engineering (CBME) at Frankfurt (Germany) University. “Current techniques for pressure mapping of mattresses don't adequately evaluate the underlying supporting foam materials — or how the human body interacts with them.” The fact that pressure sores are an ongoing issue in hospitals and nursing homes shows that the problem has yet to be resolved.

Silber and a group of pioneering medical and engineering researchers have taken on the challenge from the inside out, using finite element analysis (FEA) technology, in conjunction with magnetic resonance imaging (MRI), to study the dynamics between cushion materials and human skin, fat, muscle, and bone. Their findings provide insight into the causes of pressure sores. Their work also holds implications for biomechanical design optimization beyond mattresses to wheelchair cushions, operating room table covers, airplane seats, saddles, and even sports shoes and helmets.

Silber's work began with a request from a mattress manufacturer looking for a foam cushion that would prevent pressure sores, which can appear anywhere on the body, though they're most often found on a person's buttocks, where up to 40% of body weight is concentrated when lying down. Pressure sores are a costly challenge to the global healthcare industry ($4 billion a year in the UK alone) and one that may be growing because of aging populations worldwide. In the U.S., avoiding pressure sores has gained greater importance for hospitals since the Centers for Medicare and Medicaid Services (CMS) announced that treatment of hospital-acquired pressure sores is no longer being reimbursed as of October 2008. The Joint Commission also made pressure-sore prevention one of its National Patient Safety Goals for 2008.

When Silber's group began evaluating mattress materials to uncover the cause of pressure sores, they found limitations with the existing rigid test form devices used to approximate a human body resting on a surface. “The experimental setup for design and material evaluation of supports produced only force-indentation displacement data,” says Silber. “That didn't provide stress evaluation at the skin and support interface, let alone deep strain evaluation of the material or inside the body itself. These mechanical stresses and strains are key issues when talking about [pressure sores].”

Researchers used Abaqus finite element analysis software from Simulia, Providence, RI, (simulia.com) to perform more realistic simulation of the mechanics of body and bed interaction. “With Abaqus FEA we can create computer models that let us look inside the mattress material and human tissues to evaluate internal stresses and strains,” Silber says. “This is important because most pressure sores develop from deep within tissue outwards to the skin.” As a result, the sores remain undetected until they break through the top dermal layer. By then, in many cases, the subdermal tissue has already suffered irreversible necrosis (tissue death).

“It's the internal stresses and foremost strains of the recumbent body on its own tissues,” explains Christophe Then, Silber's research associate, “that lie at the root of the problem.”

Body is unique challenge

Collecting the data needed to build, and then validate, a human FEA model requires a different methodology from what an automaker or cell-phone design engineer might use. In the world of product development, graphs showing close agreement between FEA simulations and prototype tests are commonplace because the verifying data can be derived from real-world physical testing of inanimate objects. But in the case of human tissue modeling, confirming FEA stress and strain predictions with direct measurements from deep within a living body is not physically possible.

So modeling mattress foam was a more straightforward process for Silber's group: the engineers obtained the data they needed for FEA through laboratory testing using a device that would load, hold, and then unload different kinds of foam samples while directly recording force and indentation displacement. This procedure ensured a distinct separation of the elastic from the inelastic material properties of the foam.

But in order to “see” the hard-to-reach human tissues they were modeling, researchers used MRI technology to provide the data they needed. First, test subjects were MRI-scanned to obtain a regular tissue configuration of the buttocks region. Next, loading was applied (with a cylindrical-shaped indenter equipped with a force transducer and displacement-measuring apparatus) during an MRI scan. Working in an inverse fashion from the MRI images of this gluteal-tissue loading, the researchers were able to derive metrics that could be used as constraints in an optimization process to reveal the distinct mechanical properties of different tissue types.

“We needed to find the appropriate material parameters for in vivo fat and muscle tissue that would reflect the test conditions of tissue indentation,” says Then. “So we parameterized the material constants and simulated the models iteratively until the force-displacement and simulation output coincided.” Using the Abaqus FEA software, the researchers were able to accurately describe the fat, muscle, and connecting tissue parameters, build their FEA models describing body-support interactions, and simulate shear effects between the different tissues and bone.

Capturing complex behavior

The group converted graphic information from the MRI, a Siemens Sonata system (siemens.com), using MIMICS from Materialise (materialize.com), into a neutral IGES format for direct importation into an Altair (altair.com) Hypermesh pre-processor. For material parameter optimization they ran their own custom routines interacting with Abaqus FEA results depending on the particular algorithm being employed.

When creating their FEA models, the team captured the irregular and complex anatomy of human tissue by using second-order tetrahedral (TET10) elements for fat and muscle. TET10 elements are used to model complex sides because they are not restricted to straight-line edges. Bone was modeled with triangular shell elements and was considered to be a rigid body (at least relative to tissue) for the analysis. The fat model contained about 6,000 elements with 10,000 nodes; the muscle/bone model about 14,000 elements with 20,000 nodes.

To model the behavior of human soft tissue and account for large tissue deformations, the researchers used the Ogden equation for isotropic, nonlinear, hyperelastic, and slightly compressible materials. “In effect, biological tissue is viscoelastic, nonlinear and anisotropic,” explains Then, “but for our purposes we assumed it to be hyperelastic and isotropic.”

With both tissue and foam materials characterized, the group could combine the two into simulations of a human posterior resting on foam and run simulations using different cushion designs and materials to approximate the effects of someone lying in bed.

“Since pressure sores in deep tissue regions have been shown to develop over minutes or hours, as opposed to the seconds it takes for short-term relaxation of stressed human tissue, only the long-term tissue response was taken into account in our models,” says Then. Built with half-symmetry to facilitate analysis, a typical body-foam-interaction model contained about 132,000 elements with more than 200,000 nodes. To run the models, researchers employed an Intel (Intel.com) Core (2 Quad 9550), using four CPUs with 2.8GHz each and 16GB RAM, on a Windows XP operating system.

To validate FEA models of body and foam interaction, the researchers used MRI. “By superimposing a simulation result over the corresponding MRI image — both of them at the same deformed state — we were able to compare the boundaries of the human tissue and the outer surface of the foam we were testing,” says Silber. “Using imaging techniques in this way is essential for biomechanical modeling; it provides key information for validation.”

Silber's results clearly supported clinical observations of where pressure sores arise. The FEA images showed highest stress and strain concentration near the bones of the lower back and pelvis — the ischial tuberosity, the posterior superior iliac spine, and the sacral and tail bones — exactly below where visible pressure sores are clinically observed to occur most frequently on the skin surface.

More important than the location of the sores was their origin within the body. “FEA showed areas of greatest stress and strain at the deep interface between muscle and bone, not in the surface skin and foam support interface,” says Then. The researchers theorize that this is due to the normal ‘irregularities’ of the human skeletal structure. “Tissue movement is restricted at the relatively small, prominent surface of a bone,” explains Then. “As loading causes the tissue to displace ‘around’ a bone prominence, stress and strain increase particularly in the immediate neighborhood of that prominence.” These results are consistent with surgical findings that show cone-shaped necroses, with the base located near the bone surface, in the majority of cases of severe deep-tissue pressure sores.

“Clearly, healthcare products require better design to effectively reduce or eliminate [pressure sores] and improve the quality of life for patients,” points out Silber. “Our research is providing data that can be a foundation for that kind of design.” With continued funding from foam and healthcare companies, the team has expanded the initial scope of their work to model many different mattress configurations and materials to analyze and compare their impact on human tissue models. They are also studying the effects of biological variability of mechanical human soft tissue characteristics — taking into account gender, age and physical condition — on tissue displacement under loading.

The researchers are extending their scope beyond the gluteal area to larger “BOSS (Body-Optimized-Simulations-Systems) Models” in seated and recumbent postures, with the addition of leg and spine FEA. “Our BOSS-Models let us explore such areas as mattress and heel impact and car seat vibration,” says Then. “The kind of research methodology we have developed could be applied to products interacting with any part of the body such as feet and running shoes, or heads and helmets.”