A Comprehensive Physiological Model of Circulation Enables Automatic Piloting of Hemodynamics in Patients with Acute Heart Failure

A comprehensive physiological model of the whole circulation is mandatory to quantitatively diagnose pathophysiology and to guide an appropriate treatment. Such a model would enable automatic piloting of hemodynamics in patients with acute heart failure. By extending Guyton's model, so as to deal with heart failure predominantly affecting left heart and to quantify left atrial pressure, we constructed such a model consisting of a venous return (VR) surface and a cardiac output (CO) curve. VR surface, the integrated property of systemic and pulmonary vascular beds, relates VR and left and right atrial pressures (P(LA), P(RA)) linearly as VR = V/W - G(R)P(RA) - G(L)P(LA), given total blood volume (V). CO curve, the pumping ability of hearts, relates CO and either P(LA) or P(RA) approximately by logarithmic functions, respectively, as CO = S(L) [ln (P(LA)- B(L)) + C(L)] = S(R) [ln (P(RA) - B(R)) + C(R)]. The slopes (S(R) and S(L)) of CO curve mainly describes the pump performance. W, G(R), G(L), B(R), B(L), C(R) and C(L) are parameters. We validate the model with animal experiments. Parameters W, G's, B's and C's were relatively constant in 19 dogs. In other 8 dogs, with or without acute left heart failure, we determined V and S's from a single set of CO, P(RA), P(LA) and standard parameter values. We then predicted hemodynamics (CO, P(RA), P(LA)) for altered V from -8 to +8 ml/kg. We identified standard values of parameters as W (0.13 min), G(R) (19.6 ml/min/kg/mmHg), G(L) (3.5 ml/min/kg/mmHg). B(L) (2.1 mmHg), C(L) (1.9), B(R) (2.0 mmHg) and C(R) (0.80). Using these, we could accurately predict CO (y = 0.93x + 6.5, r2 = 0.96, Figure 2), P(RA) (y = 0.87x + 0.4, r2 = 0.91) and P(LA) (y = 0.90x + 0.48, r2 = 0.93). Our comprehensive physiological model of circulation is useful in accurately predicting hemodynamics from the measurement of a single set of CO, P(RA) and (P(LA) following blood volume changes. Therefore, this model enables continuous monitoring of blood volume and pump performance for automatic hemodynamic piloting.