A mathematical model for simulating respiratory control during support ventilation modes

Abstract

Abstract:Mathematical model simulations may assist in the selection of mechanical ventilator settings. Previously, simulations have been limited to control ventilator modes, as these models lacked representation of respiratory control. This paper presents integration of a chemoreflex respiratory control model with models describing: ventilation and pulmonary gas exchange; oxygenation and acid-base status of blood; circulation; interstitial fluid and tissue buffering; and metabolism. A sensitivity analysis showed that typical response to changing ventilator settings can be described by base excess (BE), production of CO2 (V̇CO 2 ), and model parameters describing central chemoreceptor behavior. Since BE and V̇CO 2 , can be routinely measured, changes in ventilator support may therefore be used to identify patient-specific chemoreceptor drive, enabling patient-specific predictions of the response to changes in mechanical ventilation.

Determining the appropriate model complexity for patient-specific advice on mechanical ventilation

“…In this case, pH CSF may respond well to changes in arterial PCO 2 , but the brain signal response to respiratory muscles can be reduced and, as such, A may not be sufficient so as to normalize either pH CSF , or pHa [44]. Figure 3 illustrates the mathematical models of CSF (3A) and respiratory drive (3B) included in the Beacon system, as described previously [26][27][28]. The CSF model is a modified version of that of Duffin [12].…”

“…Patient-specific tuning of the model of CSF occurs on measurement of a blood gas. In this situation, Equation 8 is solved to calculate the steady-state bicarbonate concentration in the CSF ] from arterial values [26].…”

“…Simulations are then performed using this model through the unique solution of Equations 1-7 on changes in arterial PaCO 2 , with the model simulating values of the hydrogen ion concentration ([H + csf]) or pH (pH CSF ) in the CSF, according to the patient-specific SIDcsf. Figure 3B illustrates the model of respiratory control of Duffin with modifications as described previously [26]. This model includes equations describing peripheral drive (Dp) as a linear function of the arterial hydrogen ion concentration ([Ha + ]) above a threshold (Tp) (9), with the slope of this function (Sp) dependent on the value of PaO 2 along with a scaling (A) and a threshold (P 0 ) parameter.…”

“…This model includes equations describing peripheral drive (Dp) as a linear function of the arterial hydrogen ion concentration ([Ha + ]) above a threshold (Tp) (9), with the slope of this function (Sp) dependent on the value of PaO 2 along with a scaling (A) and a threshold (P 0 ) parameter. Mechanically ventilated patients are typically ventilated with an inspired oxygen fraction so as to prevent hypoxemia, so values of peripheral drive parameters (A, P 0 , Tp) are set to constants as described previously [26]. Central drive (Dc) is a linear function of the difference between the [H + csf] and the central threshold (Tc) (10), with the slope of this function being the sensitivity of central chemoreceptors (Sc).…”

“…Central drive (Dc) is a linear function of the difference between the [H + csf] and the central threshold (Tc) (10), with the slope of this function being the sensitivity of central chemoreceptors (Sc). A is then calculated as the sum of Dp and Dc and a constant value of wakefulness drive (Dw), as described previously [26]. The patient-specific parameters of this model are Tc and Sc.…”