Continuous positive airway pressure in new-generation mechanical ventilators: a lung model study

A number of new microprocessor-controlled mechanical ventilators have become available over the last few years. However, the ability of these ventilators to provide continuous positive airway pressure without imposing or performing work has never been evaluated.

In a spontaneously breathing lung model, the authors evaluated the Bear 1000, Drager Evita 4, Hamilton Galileo, Nellcor-Puritan-Bennett 740 and 840, Siemens Servo 300A, and Bird Products Tbird AVS at 10 cm H(2)O continuous positive airway pressure. Lung model compliance was 50 ml/cm H(2)O with a resistance of 8.2 cm H(2)O x l(-1) x s(-1), and inspiratory time was set at 1.0 s with peak inspiratory flows of 40, 60, and 80 l/min. In ventilators with both pressure and flow triggering, the response of each was evaluated.

With all ventilators, peak inspiratory flow, lung model tidal volume, and range of pressure change (below baseline to above baseline) increased as peak flow increased. Inspiratory trigger delay time, inspiratory cycle delay time, expiratory pressure time product, and total area of pressure change were not affected by peak flow, whereas pressure change to trigger inspiration, inspiratory pressure time product, and trigger pressure time product were affected by peak flow on some ventilators. There were significant differences among ventilators on all variables evaluated, but there was little difference between pressure and flow triggering in most variables on individual ventilators except for pressure to trigger. Pressure to trigger was 3.74 +/- 1.89 cm H(2)O (mean +/- SD) in flow triggering and 4.48 +/- 1.67 cm H(2)O in pressure triggering (P < 0.01) across all ventilators.

Most ventilators evaluated only imposed a small effort to trigger, but most also provided low-level pressure support and imposed an expiratory workload. Pressure triggering during continuous positive airway pressure does require a slightly greater pressure than flow triggering.