Hemodynamics and Autonomic and Cognitive Performance After Carotid Revascularization Procedures

Hypothesis and Significance: Late clinical outcome and prognosis after CAS may be inferior to that after CEA in terms of autonomic modulation, hemodynamic remodeling, and cognitive function.

Specific Aim: 1) To compare the impact of CEA and CAS on long-term post-operative baroreceptor function and on cognitive brain function, and analyze their influence on clinical outcome. The specific goal is to assess the potential correlation between post-operative autonomic and cognitive function. 2) To assess the solicitation on the carotid wall due to CAS as compared to CEA through structural analysis and mechanical modeling. The specific goal is to assess the potential correlation between stenting, wall damage, baroceptor impairment, and late neurological sequelae. 3) To assess the post-operative carotid hemodynamics combining medical image analysis, clinical data, and computer simulations. The specific goal aims at correlating both local (e.g., wall stress stress) and global phenomena (controlateral flow, arterial stiffening) with baroreflex function and post-operative neurological outcomes.

Experimental Design Aim 1: A computerized method requiring small operator interaction will be used to assess indices of autonomic sympathovagal balance directed to the sinoatrial node, sympathetic vasomotor modulation, and baroreflex gain, all from spontaneous beat-by-beat variations of the R-R interval and systolic arterial pressure (SAP) variability, considering only sinus rhythm conditions. After electrode and sensor positioning, patients will be maintained for 10 minute in resting supine position, necessary for stabilization, subsequently blood pressure waveforms, electrocardiogram and respiratory activity will be continuously recorded during a nominal 5-minute baseline and then for subsequent 5-minute period of active standing.

Cardiovascular signals will be acquired by a 4 channel digital polygraph. The electrocardiogram will be recorded with two electrodes placed on the patient's chest, breathing pattern will be recorded by a piezoelectric belt and finger arterial blood pressure will be continuously monitored by a CNAP 500 HD continuous noninvasive hemodynamics monitor (CNSystems Medizintechnik AG, Austria). As described previously, (8) a series of indexes indirectly reflecting autonomic cardiovascular modulation will be derived from the spectral analysis of R-R interval and SAP variability. Postoperative cardiovascular autonomic control will be correlated to clinical outcome and measures of cognitive performance. Enrolled patients will be submitted to Mini-Mental State Examination for general cognitive impairment screening. Cognitive P300 evoked potentials will be then recorded, before and after treatment, with Ag/AgCl electrodes with a Brain Vision Recorder (Brain Products GmbH, Gilching, Germany). P300 evoked potentials will be generated after a binaurally presented tone discrimination paradigm (odd-ball paradigm) with frequent (80%) tones of 1000 Hz and rare (20%) target tones of 2000 Hz at 75 dB HL. Filter bandpass will be 0.01 to 30 Hz. Active electrodes will be placed at Cz (vertex) and Fz (frontal), respectively, and referenced to linked earlobe A1/2 electrodes (10/20 international system). During the paradigm, the patients will be instructed to keep a running mental count of the rare 2000-Hz target tones. To verify attention, P300 recordings with a discrepancy of>10% between the actual number of stimuli and the number counted by the patients will be rejected and repeated. P300 evoked potential recording will result in a stable sequence of positive and negative peaks. Latencies in milliseconds (ms) of the cognitive P300 peak will be assessed. To confirm reproducibility, two sets of P300 measurements will be recorded for all patients.

Experimental Design Aim 2: The computer-based simulation of CAS is performed exploiting a computational framework, which can be used to analyze both stent apposition and vessel wall stress in a virtual manner. The framework accounts for two main parts: the vessel model and stent model. Preoperative and postoperative medical images (including high resolution Contrast Enhanced (CE)-MRI and Computed Tomography Angiography (CTA)) will represent the input to build a patient specific carotid model. The 3D lumen profile of the vessel is reconstructed through the segmentation of DICOM image stack using tools like ITK-SNAP (www.itksnap.org) or VMTK (www.vmtk.org). The computational domain (the so-called mesh), used to solve the equilibrium equations governing the structural stent-vessel interaction, is created by an in-house developed procedure, coded in Matlab (The Mathworks Inc., Natick,MA, USA). The non-linear mechanical response of the arterial tissue is reproduced adopting an anisotropic hyperelastic strain-energy function, accounting for two families of fibers, oriented along a preferred direction with a certain degree of dispersion. The model parameters will be calibrated with respect to experimental tensile tests of the carotid tissue. The arterial model is then assembled in the simulation environment with a given stent model, picked from a predefined library of stent designs (the stent mesh generation is based on geometrical measurements performed on high-resolution micro-CT of stent samples). The CAS simulation is performed through structural finite element analysis (FEA); the commercial FEA solver Abaqus (Simulia, Dassault Systemes, FR), is adopted to run the simulations. The engineering outcomes of the simulations, (i.e. the nodal displacement field, stress tensor and strain at each integration points of the mesh), are elaborated to assess clinically relevant parameters of stenting performance (e.g. lumen gain, vessel straightening, stent cell size). The output will be used as an input of the Computational Fluid Dynamics analysis to evaluate the impact of the implanted design on the local haemodynamics (e.g., wall shear stress, oscillatory shear index, etc). Similarly, the structural analysis of CEA will be performed through the virtual pressurization of the post-operative arterial geometry.

Experimental design 3: Moving from Computational Fluid Dynamics and Fluid-Structure Interaction analysis, the investigators plan to introduce a specific model to describe the baroreflex function and this may be affected by the two different types of treatment (CEA and CAS). This requires the set up of a so-called "geometrical multiscale" model. With this, the investigators mean a numerical model coupling a local description of the hemodynamics (the one developed in Aim 2) with a more systemic representation. The latter consists of two components:

1. a 1D network mathematically described by a system of partial differential equations representing the propagation of the pressure wave along the arterial tree; each segment of the network is given by a hyperbolic system called "Euler equations" b) a compartment model for representing the peripheral microcirculation and for including the feedback mechanisms induced by the baroreflex function. Following, this will be represented by a system of ordinary differential equations where resistances properly depend on the baroreflex function.

In this aim the investigators plan two sub-aims:

1. Set up of a computational multiscale model within the framework of the LifeV solver, a finite element general purpose C++ Object Oriented library, developed by A. Veneziani and his collaborators (in Milan Politecnico and Lausanne EPFL) www.lifev.org and openly downloadable. At this stage, the investigators will reproduce the general model of Blanco et al. In particular, the identification of the parameters for the two different levels of models (1D and Lumped Parameters) will be carried out following the procedure suggested. Validation of the solver will also take advantage of the benchmarks presented in this paper.
2. Adoption of the solver previously developed for the test cases considered in Aim 2. This means that the 3D model developed in Aim 2 will be adopted for the 3D part of the geometrical multiscale model. This will allow to provide a quantitative analysis of the different impact of the two treatments and ultimately to assess in a virtual scenario how the changes of carotid compliance may impair the baroreflex function. All the CAS and CEA cases considered in Aim 2 will be equipped of this multiscale framework. The expected deliverables of this aim are therefore: a) a validated open source geometrical multiscale solver including the baroreflex function to be used systematically in patient-specific settings. b) extensive comparison of the performances of the different options.