VENTILATORY MUSCLE LOADS AND THE FREQUENCY-TIDAL VOLUME PATTERN DURING INSPIRATORY PRESSURE ASSISTED (PRESSURE-SUPPORTED) VENTILATION

Pressure support ventilation (PSV) is a new form of mechanical ventilatory support that assists a patient's spontaneous ventilatory effort with a clinician-selected amount of inspiratory pressure. In order to assess the muscle unloading effect and the ventilatory pattern response to increasing levels of this inspiratory pressure assist, we first utilized a computer respiratory system model with variable alveolar ventilation demands and impedances. From this model, we calculated ventilatory muscle loads (expressed either as the work/min or as the pressure time index) during simulated, unassisted breathing and during simulated breathing with levels of inspiratory pressure assist up to that which resulted in a VT of 800 ml and no work being performed by the muscles (defined as PSVmax for the model conditions being studied). The optimal ventilatory pattern (i.e., frequency-tidal volume) under each ventilation and impedance condition was defined as that which resulted in minimal muscle load. Under these model conditions, we found that PSVmax ranged from 5 to 41 cm H2O and that as the level of inspiratory pressure assist was increased from zero to PSVmax, there was a biphasic response of both the ventilatory muscle loading and the ventilatory pattern. Specifically, at low levels of inspiratory pressure assist, the model predicted that the applied pressure would only partially unload the ventilatory muscles. Continued muscle energy expenditure would thus still be required, whereas the ventilatory pattern would change little. Conversely, at higher levels of inspiratory pressure assist, the model predicted that the applied pressure would be sufficient to completely unload the ventilatory muscles. Under these conditions, no muscle energy would be required, and further increases in inspiratory pressure assist would result primarily in larger VT (and consequently smaller frequency). The crossover from partially loaded to unloaded conditions was found to be variable (i.e., 30 to 90% of PSVmax) and depended upon both the ventilation and the impedance conditions studied. We then reanalyzed, within the context of this model, data from a previous study of 15 patients receiving various levels of inspiratory pressure assist sufficient to prevent clinical evidence of ventilatory muscle fatigue. From this analysis, both predicted and observed estimates of load at PSVmax were near zero. Moreover, in accordance with model predictions, the observed ventilatory pattern appeared biphasic, with the crossover between partially loaded and unloaded conditions occurring at 82 to 90% of PSVmax.