Overview

Computational Models of Cardiovascular Function for Simulation, Data Integration and Clinical Decision Support

Principal Investigator:
Roger G. Mark, M.D., Ph.D.

Organization:
Massachusetts Institute of Technology
Harvard-MIT Division of Health Sciences and Technology

NASA Taskbook Entry

---

Technical Summary

One of the highest priority problems in the current manned space program is orthostatic intolerance (OI) experienced by astronauts upon their return to the normal gravitational environment. This problem has been well known since the earliest days of manned spaceflight, and has been intensely studied in flight and in many ground-based (bed-rest) studies. A number of countermeasures have been proposed and evaluated, but to date no effective and practical countermeasure has been developed. While the effectiveness of countermeasures can only be explored in ground-based experiments and ultimately be proven successful when used by astronauts, the small number of subjects studied in both situations makes the interpretation of limited experimental data for countermeasure development and refinement difficult.

Mathematical models that represent the essential functional components of the cardiovascular system provide a powerful means to extract physiologically relevant information from multiparameter experimental data.

Specific Aims
We will extend progress already made, with the following aims:

1. Develop procedures and algorithms for automated model-based extraction of physiologically relevant information from multivariate data streams, particularly those relevant to countermeasure development for OI. We will expand our unified data archive of hemodynamic signals during gravitational stress.
2. Develop strategies for systematic and effective model identification, i.e., for the adaptation of model structures and parameters to the characteristics of the available physiological data. These strategies will be based on analytical and computational studies of candidate models and their sensitivities to parameter changes.
3. Apply our computational modeling and parameter identification technology to the challenge of cardiovascular monitoring of critically ill patients. Hemodynamic measurements will be transformed into physiologically meaningful parameters through model-based processing.
4. Develop and thoroughly test a graphical user interface for our current cardiovascular model. The resultant cardiovascular simulator will be open-source and capable of running on all major operating systems. Furthermore, the entire model will be made publicly available to researchers and educators.

Key Findings and Deliverables

1. Post-spaceflight orthostatic intolerance (OI): Our simulations indicate that post-spaceflight OI can be explained either on the basis of a sufficiently severe degree of hypovolemia or by a moderate (and by itself benign) degree of hypovolemia coupled to alterations in 'secondary' cardiovascular parameters. Our simulations further indicate that alterations in the vasoconstriction loop of the arterial baroreflex can lead to a particularly dramatic collapse of arterial blood pressure homeostasis in the upright posture. These results are corroborated by ground-based studies on bed-rest subjects and inflight experiments on astronauts that suggest that midodrine (an alpha1-adrenergic agonist acting to enhance peripheral vasoconstriction) is effective in preventing OI in otherwise susceptible individuals. Our results suggest a countermeasure strategy aimed at two targets: the underlying hypovolemia or the secondary mechanisms. Midodrine affects the secondary mechanism of arterial baroreflex dysfunction.
2. Model-based patient monitoring: We have developed estimation algorithms that have the potential to improve significantly the way in which hemodynamic variables are being monitored in data-rich clinical environments such as the intensive care unit. Using a single, continuously available arterial blood pressure waveform signal, we estimate continuously cardiac output, left ventricular ejection fraction and left ventricular end-diastolic volume. These algorithms have been tested against both animal (canine and porcine hemorrhage models) and human data (taken from intensive care patients). Two reference measurements for each algorithm allow for absolute tracking of the estimated variables. While primarily intended for use in ground-based clinical environments, these algorithms can also be used to track astronaut hemodynamic performance while in space or on the lunar surface.
3. Platform-independent, user friendly simulator: We have developed and tested a platform-independent, user friendly version of both a teaching version and a research version of our main cardiovascular model. The teaching version has been tested to great reviews with students at one of MIT's quantitative physiology courses. It has been made available to the general population through the PhysioNet internet portal and can be accessed at http://physionet.org/physiotools/cvsim. This site has already been visited over 1,200 times over the past five months.

The results obtained during the current funding cycle demonstrate impressively the utility of the models in integrating and interpreting hemodynamic data from astronauts and intensive care patients alike.

---

Earth Applications

Outside the scope of modeling the effects of microgravity, numerical cardiovascular simulations:

1. Provide a basis for interpretation of clinical data in the context of intensive care. Simulations are basic components of an intelligent patient monitoring system, which can couple real-time physiologic data to cardiovascular pathology.
2. Provide a powerful environment for teaching of normal and abnormal cardiovascular physiology.

---

This project's funding ended in 2007