Enough data to conclude

Knowledge-based systems were initially developped to automatically adapt pressure support settings during invasive ventilation, and proved to be at least as efficient as experienced clinicians.

Non-invasive ventilation has become the standard of care for patients suffering from acute hypercapnic respiratory failure (ARF)and has reduced the need for endotracheal intubation in these patients, thus reducing their hospital mortality.

NIV success or failure is closely related to the tolerance of NIV treatment, which is tightly correlated to patient-ventilator synchrony. As severe asynchronies frequently occurs during NIV (namely in more than 40% of patients) and as the occurence of asynchronies is related to the use of high pressure support levels, to the presence of leaks and/or to non optimal expiratory trigger settings, very frequent ventilator settings adaptations should allow reducing patient-ventilator asynchronies but require the presence of an experienced clinician at the bedside during NIV treatment.

A computer-driven ventilator settings adaptation has the adavantage of permitting very frequent ventilator settings adaptation whithout requiring the presence of an experienced clinician at the bedside and could possibly improve patient-ventilator interaction.

The aim of the present study is to test the faisability of using the Smartcare NIV computer-driven system to automatically adapt ventilator settings during non invasive ventilation delivered because of acute respiratory failure.

Inclusion Criteria:

• Acute respiratory failure requiring non invasive ventilation
• ICU patient equipped with an arterial line.

Exclusion Criteria:

• Contra-indications to NIV:
• impaired consciousness
• absence of patient cooperation
• severe hemodynamic instability
• vomiting
• facial lesions
• indication for immediate intubation
• Acute neurological problem
• Poor short term prognosis