The present study will use a new type of respirator in premature babies who need help with their breathing. This new respirator uses signals from the baby's diaphragm - the most important breathing muscle - to control the timing and the amount of air that the baby needs. The goal of the study is to demonstrate that this new respirator can synchronize delivery of air to the baby's efforts, and that synchrony is maintained regardless of whether the baby is breathing with a tube or a mask.

There is an abundance of evidence in the literature suggesting that maintenance of spontaneous breathing with a synchronized mode of ventilatory assist, and the use of non-invasive interface to deliver the assist, has the potential to significantly improve neonatal respiratory care. Conventional modes of mechanical ventilation use pneumatic signals such as airway pressure, flow, or volume, which are dampened by respiratory muscle weakness, increased load (impaired respiratory mechanics), and leaks. In order to improve patient ventilator synchrony, further development over current technology is required.

This study deals with the implementation and clinical evaluation of neural control of mechanical ventilation in the neonatal intensive care unit. The goal is to demonstrate, in pre-term newborns with extremely low birth weight, that neural control of mechanical ventilation, using the electrical activity of the diaphragm (EAdi), can synchronize delivery of assist to the patient's inspiratory drive, and that synchrony is maintained regardless of the interface used. This proposal will introduce for the first time technology for neural triggering and cycling-off as well as neurally adjusted ventilatory assist (NAVA) in the treatment of pre-term infants.

In Project 1, the aim is to demonstrate that neural triggering and cycling-off (i.e. initiation and termination of ventilatory assist using EAdi) improve infant-ventilator synchrony, compared to conventional pneumatic trigger systems in pre-term infants with extremely low birth weight. It is hypothesized that neural triggering and cycling-off of mechanical ventilation improves infant-ventilator synchrony. This will be evaluated by comparing the infant's neural timings (inspiratory and expiratory) to ventilator timings, during conventional pressure support ventilation and during neural triggering and cycling-off. We expect that patient-ventilator synchrony will be improved in the neural mode, and that comfort will be lowest with increased asynchrony (conventional modes) and highest with improved infant-ventilator synchrony (neural triggering and cycling-off).

In Project 2, the aim is to demonstrate that administration of NAVA with invasive (endotracheal intubation) or non-invasive interface (nasal prongs) is equally efficient in terms of triggering and cycling-off. The hypothesis is that with NAVA, non-invasive ventilation with nasal prong is equally efficient as invasive ventilation. In premature infants deemed ready for extubation, NAVA will be implemented prior to and post-extubation (with single nasal prong). We anticipate that ventilatory assist will be delivered with full synchrony regardless of invasive or non-invasive delivery of assist, and that there should be no difference in the delays between the onset of EAdi and ventilatory assist and in the delays between peak of EAdi and cycling-off. We also expect that due to less airway resistance during non-invasive ventilation, peak applied pressures and diaphragm activation levels will be lower.

By improving patient-ventilator interaction and allowing use of a non-invasive patient-ventilator interface, neural control of mechanical ventilators has the potential to significantly reduce ventilator-related complications, reduce the incidence of lung injury, facilitate weaning from mechanical ventilation, and decrease the duration of stay in the intensive care unit and overall hospitalization. These issues can be addressed in future randomized clinical trials in the case that the present short-term work has a positive outcome.

Inclusion Criteria:

Part 1: Preterm newborns (>1 week postnatal age, gestational age> 23 weeks, birth weight <1000g), intubated and mechanically ventilated, and breathing spontaneously. Infants should be breathing on a triggered mode of ventilation (SIMV, assist control, or pressure support) with the following ventilator parameters: assist control/pressure support < 20 cm H2O, PEEP4-7 cm H2O, SIMV rate< 30, FIO2<40%/.

Part 2:Newborns recovering from respiratory illness and deemed to be ready for elective extubation by their attending physician, intubated and mechanically ventilated, and breathing spontaneously.

Exclusion Criteria:

Protocols 1 and 2: Pneumothorax, degenerative neuromuscular disease, bleeding disorders, cardiovascular instability (defined as hypotension or need for treatment with volume expansion or inotropes in the previous 24 hours), cyanotic congenital cardiovascular disease, treatment with narcotics, phrenic nerve damage/diaphragm paralysis, esophageal perforation, use of high frequency oscillatory ventilation