Individual response in patient's effort and driving pressure to variations in assistance during pressure support ventilation

Background: During Pressure Support Ventilation (PSV) an inspiratory hold allows to measure plateau pressure (Pplat), driving pressure (triangle P), respiratory system compliance (Crs) and pressure-muscle-index (PMI), an index of inspiratory effort. This study aims [1] to assess systematically how patient's effort (estimated with PMI), triangle P and tidal volume (Vt) change in response to variations in PSV and [2] to confirm the robustness of Crs measurement during PSV.Methods: 18 patients recovering from acute respiratory failure and ventilated by PSV were cross-randomized to four steps of assistance above (+ 3 and + 6 cmH(2)O) and below (-3 and -6 cmH(2)O) clinically set PS. Inspiratory and expiratory holds were performed to measure Pplat, PMI, triangle P, Vt, Crs, P0.1 and occluded inspiratory airway pressure (Pocc). Electromyography of respiratory muscles was monitored noninvasively from body surface (sEMG).Results: As PSV was decreased, Pplat (from 20.5 +/- 3.3 cmH(2)O to 16.7 +/- 2.9, P < 0.001) and triangle P (from 12.5 +/- 2.3 to 8.6 +/- 2.3 cmH(2)O, P < 0.001) decreased much less than peak airway pressure did (from 21.7 +/- 3.8 to 9.7 +/- 3.8 cmH2O, P < 0.001), given the progressive increase of patient's effort (PMI from -1.2 +/- 2.3 to 6.4 +/- 3.2 cmH(2)O) in line with sEMG of the diaphragm (r = 0.614; P < 0.001). As triangle P increased linearly with Vt, Crs did not change through steps (P = 0.119).Conclusion: Patients react to a decrease in PSV by increasing inspiratory effort-as estimated by PMI-keeping Vt and triangle P on a desired value, therefore, limiting the clinician's ability to modulate them. PMI appears a valuable index to assess the point of ventilatory overassistance when patients lose control over Vt like in a pressure-control mode. The measurement of Crs in PSV is constant-likely suggesting reliability-independently from the level of assistance and patient's effort.