Laurent Heyer

Ventilatory Support: A Dynamical Systems Approach

Misunderstanding of the dynamical behavior of the ventilatory system, especially under assisted ventilation, may explain the problems encountered in ventilatory support monitoring. Proportional assist ventilation (PAV) that theoretically gives a breath by breath assistance presents instability with high levels of assistance. We have constructed a mathematical model of interactions between three objects: the central respiratory pattern generator modelled by a modified Van der Pol oscillator, the mechanical respiratory system which is the passive part of the system and a controlled ventilator that follows its own law. The dynamical study of our model shows the existence of two crucial behaviors, i.e. oscillations and damping, depending on only two parameters, namely the time constant of the mechanical respiratory system and a cumulative interaction index. The same result is observed in simulations of spontaneous breathing as well as of PAV. In this last case, increasing assistance leads first to an increase of the tidal volume (VT), a further increase in assistance inducing a decrease in VT, ending in damping of the whole system to an attractive fixed point. We conclude that instabilities observed in PAV may be explained by the different possible dynamical behaviors of the system rather than changes in mechanical characteristics of the respiratory system.

Non-invasive detection of respiratory muscles activity during assisted ventilation.

The instantaneous pressure applied by the respiratory muscles [Pmus(t)] of a patient under ventilatory support may be continuously assessed with the help of a model of the passive respiratory system updated cycle by cycle. Inspiratory activity (IA) is considered present when Pmus goes below a given threshold. In six patients, we compared IA with (i) inspiratory activity (IAref) obtained from esophageal pressure and diaphragmatic EMG and (ii) that (IAvent) detected by the ventilator. In any case, a ventilator support onset coincides with an IA onset but the opposite is not true. IA onset is always later than IAref beginning ((0.21 +/- 0.10 s) and IA end always precedes IAref end (0.46 +/- 0.16 s). These results clearly deteriorate when the model is not updated.