Assessment of intracranial dynamics in hydrocephalus: effects of viscoelasticity on the outcome of infusion tests

Abstract

Object The treatment of hydrocephalus requires insight into the intracranial dynamics in the patient. Resistance to CSF outflow (R0) is a clinically obtainable parameter of intracranial fluid dynamics that quantifies the apparent resistance to CSF absorption. It is used as a criterion for the selection of shunt candidates and serves as an indicator of shunt performance. The R0 is obtained clinically by performing 1 of 3 infusion tests: constant flow, constant pressure, or bolus infusion. Among these, the bolus infusion method has the shortest examination times and provides the shortest time of exposure of patients to artificially increased intracranial pressure (ICP) levels. However, for unknown reasons, the bolus infusion method systematically underestimates the R0. Here, the authors have tested and verified the hypothesis that this underestimation is due to lack of accounting for viscoelasticity of the craniospinal space in the calculation of the R0. Methods The authors developed a phantom model of the human craniospinal space in order to reproduce in vivo pressure-volume (PV) relationships during infusion testing. The phantom model followed the Marmarou exponential PV equation and also included a viscoelastic response to volume changes. Parameters of intracranial fluid dynamics, such as the R0, could be controlled and set independently. In addition to the phantom model, the authors designed a computational framework for virtual infusion testing in which viscoelasticity can be turned on or off in a controlled manner. Constant flow, constant pressure, and bolus infusion tests were performed on the phantom model, as well as on the virtual computational platform, using standard clinical protocols. Values for the R0 were derived from each infusion test by using both a standard method based on the Marmarou PV equation and a novel method based on a system identification approach that takes into account viscoelastic behavior. Results Experiments with the phantom model confirmed clinical observations that both the constant flow and constant pressure infusion tests, but not the bolus infusion test, yield correct R0 values when they are determined with the standard method according to Marmarou. Equivalent results were obtained using the computational framework. When the novel system identification approach was used to determine the R0, all of the 3 infusion tests yielded correct values for the R0. Conclusions The authors' investigations demonstrate that intracranial dynamics have a substantial viscoelastic component. When this viscoelastic component is taken into account in calculations, the R0, is no longer underestimated in the bolus infusion test.