OBJECTIVE: Inspiratory pressure (Pi) support delivered by a bilevel device has become the technique of choice for noninvasive home ventilation. Considerable progress has been made in the performance and functionality of these devices. The present bench study was designed to compare the various characteristics of 10 recently developed bilevel Pi devices under different conditions of respiratory mechanics. DESIGN: Bench model study. SETTING: Research laboratory, university hospital. MEASUREMENTS: Ventilators were connected to a lung model, the mechanics of which were set to normal, restrictive, and obstructive, that was driven by an ICU ventilator to mimic patient effort. Pressure support levels of 10 and 15 cm H(2)O, and maximum were tested, with "patient" inspiratory efforts of 5, 10, 15, 20, and 25 cm H(2)O. Tests were conducted in the absence and presence of leaks in the system. Trigger delay, trigger-associated inspiratory workload, pressurization capabilities, and cycling were analyzed. RESULTS: All devices had very short trigger delays and triggering workload. Pressurization capability varied widely among the machines, with some bilevel devices lagging behind when faced with a high inspiratory demand. Cycling was usually not synchronous with patient inspiratory time when the default settings were used, but was considerably improved by modifying cycling settings, when that option was available. CONCLUSIONS: A better knowledge of the technical performance of bilevel devices (ie, pressurization capabilities and cycling profile) may prove to be useful in choosing the machine that is best suited for a patient's respiratory mechanics and inspiratory demand. Clinical algorithms to help set cycling criteria for improving patient-ventilator synchrony and patient comfort should now be developed.
OBJECTIVE: Inspiratory pressure (Pi) support delivered by a bilevel device has become the technique of choice for noninvasive home ventilation. Considerable progress has been made in the performance and functionality of these devices. The present bench study was designed to compare the various characteristics of 10 recently developed bilevel Pi devices under different conditions of respiratory mechanics. DESIGN: Bench model study. SETTING: Research laboratory, university hospital. MEASUREMENTS: Ventilators were connected to a lung model, the mechanics of which were set to normal, restrictive, and obstructive, that was driven by an ICU ventilator to mimic patient effort. Pressure support levels of 10 and 15 cm H(2)O, and maximum were tested, with "patient" inspiratory efforts of 5, 10, 15, 20, and 25 cm H(2)O. Tests were conducted in the absence and presence of leaks in the system. Trigger delay, trigger-associated inspiratory workload, pressurization capabilities, and cycling were analyzed. RESULTS: All devices had very short trigger delays and triggering workload. Pressurization capability varied widely among the machines, with some bilevel devices lagging behind when faced with a high inspiratory demand. Cycling was usually not synchronous with patient inspiratory time when the default settings were used, but was considerably improved by modifying cycling settings, when that option was available. CONCLUSIONS: A better knowledge of the technical performance of bilevel devices (ie, pressurization capabilities and cycling profile) may prove to be useful in choosing the machine that is best suited for a patient's respiratory mechanics and inspiratory demand. Clinical algorithms to help set cycling criteria for improving patient-ventilator synchrony and patient comfort should now be developed.