Background: Pulmonary barotrauma is a potentially fatal complication of positive pressure ventilation. We previously found that barotrauma occurred in patients with radiographic hyperinflation, but few objective data define the relationships among hyperinflation, objective chest radiograph (CXR) measurements, ventilator parameters, and development of barotrauma.
Objectives: We sought (1) to assess the relationships among hyperinflation, objective CXR findings, mechanical ventilator parameters, and development of barotrauma. (2) To compare radiographic hyperinflation, ventilator parameters, and incidence of barotrauma in a current group of ICU patients with historical control subjects.
Setting: Medical and surgical ICU patients in a university hospital.
Design: Prospective blinded observational study; comparison of current series with historical control subjects.
Methods: One hundred two prospectively enrolled mechanically ventilated medical and surgical ICU patients each received portable supine CXRs that were reviewed independently by three radiologists who made objective measurements and subjectively determined the likelihood of hyperinflation. Ventilator parameters were recorded at the bedside at the time each CXR was obtained. CXR measurements and ventilator parameters were then related to the development of barotrauma during the course of ventilation and compared with findings of a prospective study at our institution 1 year earlier.
Results: Radiographically recognizable hyperinflation occurred in 18 of 102 mechanically ventilated ICU patients (18%) and correlated with lung length (24.7 vs 19.8 cm; p<0.05) and the anterior rib number that intersects the hemidiaphragm (5.4 vs 4.7; p<0.05). Patients with hyperinflation were ventilated at higher tidal volume per kilogram (VT/kg) (11.0 vs 9.4; p=0.0081), but peak airway pressure, plateau pressure, and positive end-expiratory pressure were similar. There were significant decreases in VT (810 vs 739 mL; p=0.015) and VT/kg (11.0 vs 10.1 mL/kg; p<0.001) in these mechanically ventilated ICU patients in comparison to hospital control subjects evaluated during the previous year. Paralleling these changes was a decrease in the frequency of CXR hyperinflation (p=0.003) and the incidence of ventilator-associated barotrauma (6.5% vs 0.98%; p=0.048).
Conclusions: Ventilation at higher VT/kg is associated with a higher incidence of CXR hyperinflation. Radiographic hyperinflation is associated with lung length ≥24.7 cm and visualization of the sixth anterior rib. Patients with hyperinflation may be at greater risk for developing barotrauma or volutrauma. Ventilatory strategies utilizing lower volumes are associated with a lower incidence of such trauma in the current sample as compared with historical control subjects.