Experimental and theoretical analyses of compression induced muscle damage : aetiological factors in pressure ulcers

Pressure ulcers form a major problem in health care. They often occur when patients are bedridden, wheelchair bound or wearing prostheses. The ulcers can be very painful for the patient and often lead to prolonged hospitalization. In addition, the huge costs involved with treatment and prevention put a heavy burden on heath care budgets. Pressure ulcers occur often: between 14% and 33% of the patients in health care institutions develop an ulcer, ranging from discolouration of the skin to severe wounds involving necrosis of epidermis, extending to underlying bone, tendon and joints. It is clear that pressure ulcers are caused by prolonged mechanical loading, applied at the interface between skin and support surfaces. However, the aetiology of pressure ulcers is poorly understood. This forms an important obstacle in decreasing the unacceptably high prevalence figures. It is anticipated that a better understanding of the mechanobiological pathways leading to cell and tissue damage can lead to a breakthrough in reducing pressure ulcer prevalence. In addition, a solid scientific base may establish tools for objective risk assessment and judgement of preventive measures. The present study focuses on deep ulcers that initiate in skeletal muscle tissue, since deep ulcers are more extensive and often difficult to prevent. To obtain insight into the aetiology of these deep ulcers, it is necessary to understand the transfer from externally applied loads at the skin, to the local conditions that the cells experience within the tissue. In addition, the question which local conditions are harmful to the cell needs to be investigated. By combining knowledge on "what a cell feels" with knowledge on potentially harmful conditions, a better judgement of dangerous situations may be achieved. Although several causes of cell damage may play a role in the initiation of pressure ulcers, the present study focussed on the impact of cell deformations. To investigate the hypothesis that prolonged cell deformations lead to cell damage at clinically relevant strains, an experimental model system was developed. A key requirement of this experimental model is the possibility to study the role of cell deformation on cell damage independently of other possible causes of damage. To achieve this, in-vitro engineered muscle tissue constructs were developed. These constructs were compressed using a newly developed compression device. A custom made incubator system was developed to allow monitoring of the constructs for extended periods of time. In addition, a novel assay was developed to determine the viability of the cells during compression. This assay provides quantitative and spatial information on cell damage throughout a construct in a non-invasive manner, making use of fluorescent dyes which are visualized by confocal microscopy. The compression of the engineered muscle tissue constructs indicated that a significant increase in cell death occurs within 1-2 hours and that higher strain levels led to an earlier increase in damage. In addition, it was demonstrated that cell damage was uniformly distributed across the indented area of the construct, without a gradient in percentage dead cells between the centre and periphery of the constructs. The results strongly suggest that prolonged cell deformation was the predominant cause of cell damage in these experiments. This puts a new light on observations in literature which suggested that ischaemia is not the sole determinant for the onset of pressure ulcers. Nevertheless, more experiments are needed to clarify the role of prolonged cell deformations on cell damage. First, it is recommended that the actual local cell deformations are quantified during compression of the constructs. Furthermore, from the present experiments it could not be excluded that the compression of the constructs decreased the permeability of the construct and hence affected cellular metabolism. In future, measuring diffusion pathways of both small molecules and larger vital molecules, may indicate whether this change in permeability is significant. A numerical model was developed to predict local cell deformations, in response to tissue compression. Since the local cell deformations cannot be a-priori determined on the basis of homogenized tissue deformations, a multilevel finite element approach was adopted. In this approach, cell deformations are predicted from detailed nonlinear finite element analyses of the local microstructures of the tissue, which consist of an arrangement of cells embedded in a matrix material. To avoid unacceptably large computational times, the multilevel model was designed to run on a parallel computer system. Application of the multilevel model showed that the heterogeneity of the microstructure of the tissue has a profound impact on local cell deformations, which highly exceeded macroscopic tissue deformations. Moreover, microstructural heterogeneity led to complex cell shapes and caused non-uniform deformations within the cells. To investigate the evolution of compression induced damage in skeletal muscle tissue, the multilevel model was extended with a damage law, which was derived from the in-vitro experiments. With this model, the compression of muscle tissue against a bony prominence was simulated. The percentage of cell damage in the microstructure of the tissue was computed, which could be extrapolated to the bulk tissue level. In the present form, a schematic geometry was considered that intended to elucidate general patterns of tissue damage evolution. The simulations confirmed that it is not feasible to predict the onset of tissue damage on the basis of externally applied loading conditions at the skin surface alone, since these externally applied loads are not indicative of the local mechanical conditions that the cells experience within the tissue. In addition, the simulations showed that it is necessary to consider the local load history of the cells, and the tolerance of the tissue. These findings may explain why a strikingly large variability in load/time threshold values was found in animal studies, which attempted to relate external mechanical to tissue damage, thereby ignoring the local mechanical conditions within the tissue. At present, it is premature to utilize the models presented in this thesis in clinical practice, since the extrapolation towards human patients requires more research. Clearly, further extensions and validation of the numerical model with experimental animal models will be required. This should finally lead to the application in more realistic cases, involving patient data on geometry and tissue properties. Nevertheless, the present models provided an essential step towards evidence based risk assessment and prevention.