A Bench Evaluation of Ineffective Efforts Related to the Use of Leak Compensation Algorithm During Pressure Support Ventilation

Abstract

Background: Leak compensation algorithm is a key success element in patient-triggered ventilation. However, the effect of leak compensation on ventilator-triggered performance in the absence of leak has not been fully investigated. This bench study aimed to test the hypothesis that a change in the breathing pattern of certain lung mechanics incorrectly compensates for trigger sensitivity, causing ineffective efforts (IEs).

Methods: Two ICU ventilators (PB840 and PB980, Covidien) with activated leak compensation were compared in a leak-free setting using a lung simulator (ASL 5000, IngMar Medical). The default disconnect sensitivity was set to 40 L/min. Three adult lung models with different lung mechanics (normal, ARDS, and COPD) were tested using a combination of low and high muscle pressures (P_mus) and breathing frequency. The ventilators in a pressure support mode were connected to the ASL 5000. Pressure support was set at 5 or 10 cm H₂O; PEEP was set at 5 cm H₂O; apnea time was set at 20 s; flow trigger sensitivity was set at 3 L/min; and termination criteria were set at 5%, 25%, and 40% for ARDS, normal, and COPD mechanics, respectively. All combinations of frequency and P_mus were sequentially changed in the lung model, and the incidence of IEs diagnosed on the ASL 5000 waveforms and activation of apnea backup ventilation was observed for 60 s after each change. Three trials were run for each model. The influence of lung mechanics, breathing pattern, and the direction of change in the breathing pattern was examined on the outcomes.

Results: The incidence of IEs was significantly higher with PB840 than with PB980 for pressure supports of 5 and 10 cm H₂O (21% vs. 2% and 7% vs. 2%, respectively; P < .05 for both comparisons). All IEs resulted in apnea backup ventilation with PB840 but not with PB980. With PB840, 35% and 65% of the IEs were observed in ARDS and COPD, respectively. Conversely, with PB980, all IEs were observed in COPD. Although most IEs (91%) occurred during reduced frequency, the incidence of IEs was not associated with the magnitude of P_mus or the direction of change in P_mus or frequency.

Conclusions: The activation of leak compensation in a leak-free setting resulted in IEs and the eventual backup ventilation with both PB840 and PB980. This was suggested to be accompanied by high airway resistance and low frequency. Thus, PB980 was superior to PB840 in ameliorating this problem. Further clinical investigation is needed.