Randomized crossover comparison of proportional assist ventilation and patient-triggered ventilation in extremely low birth weight infants with evolving chronic lung disease.

BACKGROUND Refinement of ventilatory techniques remains a challenge given the persistence of chronic lung disease of preterm infants. METHODS To test the hypothesis that proportional assist ventilation (PAV) will allow to lower the ventilator pressure at equivalent fractions of inspiratory oxygen (FiO(2)) and arterial hemoglobin oxygen saturation in ventilator-dependent extremely low birth weight infants in comparison with standard patient-triggered ventilation (PTV). DESIGN Randomized crossover design. SETTING Two level-3 university perinatal centers. PATIENTS 22 infants (mean (SD): birth weight, 705 g (215); gestational age, 25.6 weeks (2.0); age at study, 22.9 days (15.6)). INTERVENTIONS One 4-hour period of PAV was applied on each of 2 consecutive days and compared with epochs of standard PTV. RESULTS Mean airway pressure was 5.64 (SD, 0.81) cm H(2)O during PAV and 6.59 (SD, 1.26) cm H(2)O during PTV (p < 0.0001), the mean peak inspiratory pressure was 10.3 (SD, 2.48) cm H(2)O and 15.1 (SD, 3.64) cm H(2)O (p < 0.001), respectively. The FiO(2) (0.34 (0.13) vs. 0.34 (0.14)) and pulse oximetry readings were not significantly different. The incidence of arterial oxygen desaturations was not different (3.48 (3.2) vs. 3.34 (3.0) episodes/h) but desaturations lasted longer during PAV (2.60 (2.8) vs. 1.85 (2.2) min of desaturation/h, p = 0.049). PaCO(2) measured transcutaneously in a subgroup of 12 infants was similar. One infant met prespecified PAV failure criteria. No adverse events occurred during the 164 cumulative hours of PAV application. CONCLUSIONS PAV safely maintains gas exchange at lower mean airway pressures compared with PTV without adverse effects in this population. Backup conventional ventilation breaths must be provided to prevent apnea-related desaturations.