Airway leak size in neonates and autocycling of three flow-triggered ventilators.

OBJECTIVES: To define the spectrum of airway leak in the neonatal population and examine the occurrence rate of autocycling of three flow-triggered ventilators within the defined spectrum of airleak.

DESIGN: Prospective study of pulmonary function tests of intubated infants and performance of ventilators on a mechanical lung model under simulated clinical conditions.

SETTING: An intensive care nursery and research laboratory at a university medical center.

INTERVENTIONS: Analysis of pulmonary function tests of 50 infants from our intensive care nursery, selected at random, to determine size of airleak around the endotracheal tube. The rate of autocycling of ventilators due to airleak of variable size, while connected to a test lung was subsequently studied. Ventilators were set on the assist-control mode with the control rate set at 0 breath/min. Each ventilator was studied at the maximum sensitivity setting, which was 1, 2.5, and 3.3 mL/sec for each ventilator, respectively, and also at decreased sensitivity settings to 10 mL/sec. Airleak size was varied (10% to 45%) by increasing the orifice size within the endotracheal tube adapter/connector sideport and/or the positive end-expiratory pressure level (2 to 8 cm H2O).

MEASUREMENTS AND MAIN RESULTS: In the infants, airleak size was calculated during synchronous ventilator breaths as (inspiratory minus expiratory) tidal volume/expiratory tidal volume x 100% (n = 25 +/- 11 breaths/patient). Mean +/- SD leak size in the infants was 15.6 +/- 11%. A minimal leak size of 0 to 10% was present in 15 (30%) infants, leak size of 10% to 20% in 24 (48%), leak size of 20% to 30% in seven (14%), and leak size > 30% in four (8%) infants. The relative tendency of the three ventilators to autocycle is a function of the maximum sensitivity setting, which varies with each ventilator. The ventilator with the maximum sensitivity set at 1 mL/sec autocycled rapidly (> or = 40 breaths/min) at leak size of > 10%; the ventilator set at 2.5 mL/sec autocycled rapidly at leak size of > or = 20%; and the ventilator set at 3.3 mL/sec autocycled rapidly at leak size of > or = 30%. In all ventilators, the rate of autocycling increased with increased leak size, and decreased with decreased sensitivity setting.

CONCLUSIONS: Flow-triggered ventilators are susceptible to autocycling due to flow compensation to maintain positive end-expiratory pressure levels in the presence of an airway leak. The difference in autocycling is due to the maximum sensitivity setting of each ventilator, and not to intrinsic ventilator flowsensing or other software mechanisms. The 3.3-mL/sec setting was the least prone to autocycling and seems appropriate. The ventilator set at 2.5 mL/sec at the time of this study has been released instead at 4 mL/sec, due to these findings. The ventilator with the maximum setting at 1 mL/sec autocycled readily at leak size of > or = 10%. Since such a leak size was present in 70% of infants, this setting should be used with caution. Using these guidelines, autocycling of all three ventilators is likely to occur mainly in 8% of infants with leak size of > 30%. In these cases, lowering the sensitivity setting and/or positive end-expiratory pressure level may decrease autocycling, or may necessitate reintubation with a larger endotracheal tube.