Objective During proportional assist ventilation (PAV) the timing and frequency of inflations are controlled by the patient and the patient's work of breathing may be relieved by elastic and/or resistive unloading. It is important and the authors' objective to determine whether ventilators delivering PAV function well in situations mimicking neonatal respiratory conditions.

Design In vitro laboratory study.

Setting Tertiary neonatal ICU.

Interventions Dynamic lung models were developed which mimicked respiratory distress syndrome, bronchopulmonary dysplasia and meconium aspiration syndrome to assess the performance of the Stephanie® neonatal ventilator.

Main outcome measures The effects of elastic and resistive unloading on inflation pressures and airway pressure wave forms and whether increasing unloading was matched by an ‘inspiratory’ load reduction.

Results During unloading, delivered pressures were between 1 and 4 cm H₂O above those expected. Oscillations appeared in the airway pressure wave form when the elastic unloading was greater than 0.5 cm H₂O/ml with a low resistance model and 1.5 cm H₂O/ml with a high resistance model and when the resistive unloading was greater than 100 cm H₂O/l/s. There was a time lag in the delivery of airway pressure of at least 60 ms, but increasing unloading was matched by an inspiratory load reduction.

Conclusions During PAV, unloading does reduce inspiratory load, but there are wave form abnormalities and a time lag in delivery of the inflation pressure. The impact of these problems needs careful evaluation in the clinical setting.