Pressure Support Ventilation in Patients with Acute Lung Injury

Objectives: To assess the success rate of pressure support ventilation (PSV) in acute lung injury patients undergoing continuous positive pressure ventilation (CPPV), to study physiologic changes after the transition from CPPV to PSV, and to investigate differences between patients who succeed and patients who fail PSV according to predetermined criteria.

Design: Observational study.

Setting: General intensive care unit in a teaching hospital.

Subjects: We studied 48 patients having acute lung injury, as defined by a PaO2/F(IO2) <300 mm Hg and the presence of bilateral infiltrates on chest radiograph, and ventilated with CPPV. We included patients with PaO2 >80 mm Hg, at positive end-expiratory pressure of <15 cm H2O and with F(IO2) up to 1.0.

Interventions: After enrollment, PSV was instituted and patients were strictly monitored during the following 48 hrs. Subjects who met any of the predefined PSV failure criteria during this period were returned to CPPV (Group F). PSV was continued in the remaining patients (Group S).

Measurements And Main Results: Gas exchange, respiratory mechanics, and hemodynamics measurements were collected before switching from CPPV to PSV and were repeated at 24 hrs after beginning PSV, or immediately before return to CPPV in Group F patients. The physiologic deadspace volume to tidal volume ratio (V(D)/V(T)) was obtained by the Enghoff's equation from the measurement of the mixed expired CO2 fraction. PSV resulted in a significant PaCO2 decrease (49.2+/-10.9 mm Hg to 44.4+/-7.2 mm Hg) and significant increases in minute volume (V(E))(9.0+/-2.3 L/min to 12.0+/-4.0 L/min) and arterial blood pH (7.405+/-0.054 to 7.435+/-0.064), with stable oxygenation and hemodynamics. In patients who were hypercapnic (PaCO2 >50 mm Hg) during CPPV, the V(E) increase was higher than in normocapnic patients. In the latter patients, PaCO2 and pH did not change significantly going from CPPV to PSV. A total of 38 patients (79%) were allocated to Group S and the remaining 10 patients were included in Group F. In Group S, positive endexpiratory pressure of 9.4+/-2.9 cm H2O (range, 3-14 cm H2O) and a PSV level of 14.9+/-3.8 cm H2O (range, 9-22 cm H2O) were applied. In Group F, positive end-expiratory pressure of 8.9+/-3.1 cm H2O (range, 5-15 cm H2O) and a PSV level of 21.6+/-4.6 cm H2O (range, 16-31 cm H2O) were adopted. Compared with Group S, Group F had a longer duration of intubation (20.2+/-19.2 days vs. 9.2+/-13.5 days), a lower static compliance of the respiratory system (30.4+/-16.5 mL/cm H2O vs. 41.7+/-15.0 mL/cm H2O), and a higher V(D)/V(T) (0.70+/-0.09 vs. 0.52+/-0.10), but similar oxygenation and positive end-expiratory pressure. V(E) was higher in Group F during both CPPV and PSV.

Conclusions: In a relatively high proportion of the investigated patients, PSV was successful. The institution of PSV led to no major changes in oxygenation or in hemodynamics. PSV was associated with increases in V(E) and respiratory frequency. In patients who had been hypercapnic during CPPV, PaCO2 decreased despite a compensated pH. Compared with PSV success patients, patients who failed PSV appeared to be sicker, as shown by the higher duration of respiratory support, increased ventilatory needs, and decreased respiratory system compliance, despite similar arterial oxygenation and positive end-expiratory pressure.