Pressure support ventilation and NAVA during noninvasive ventilation in acute respiratory failure: a pilot study

Noninvasive machanial ventilation (NIV) is an established intervention for
selected patients with acute respiratory failure. NIV reduces work of breathing
and improves oxygenation by recruitment of alveoli and optimizing ventilation -
perfusion matching. The application of NIV prevents endotracheal intubation in
a number of patients. Established reasons for NIV include acute exacerbation of
chronic obstructive airway disease, acute heart failure and in selected cases
pneumonia and exacerbation of asthma.
Unfortunately in many patients NIV fails and subsequent endotracheal intubation
is necessary. Important reasons for failed NIV include patiënt - ventilator
asynchrony and major airleak between the patients face and the interface.
Neurally adjusted ventilatory assist (NAVA) is a relatively new ventilatory
mode. NAVA is characterized by ventilator activation based on changes in
electrical activity of the inspiratory muscles (diaphragm). Instead in the
conventional ventilatory modes (i.e. pressure support ventilation) the
ventilator is triggered by negative pressure generated by the inspiratory
muscles. An important delay occurs between activation of the inspiratory
muscles and the detection of the negative pressure in the ventilator (up to 400
ms), resulting in elevated work of breathing and patient discomfort. In NAVA
electrical activity of the diaphragm is continuously monitored with a dedicated
nasogastric tubes. Elevated electrical activity of the diaphragm is interpreted
by the ventilator as the beginning of an inspiration. The delay between patient
inspiration and ventilator inspiration is very shot (microseconds), resulting
in better patient - ventilator synchrony. Moreover, in pressure support
ventilation inspiratory efforts may not be sensed by the ventilator (wasted
efforts) due to intrinsic PEEP or respiratory muscle weakness. As electrical
activity of the diaphragm is a much more sensitive measure for inspiratory
effort, wasted efforts are very uncommon during NAVA. Indeed, recent studies in
invasive ventilated patients have shown that NAVA improves patient - ventilator
interaction compared to pressure support ventilation.

Patient discomfort during noninvasive ventilation does not only result from
asynchrony with the ventilator but also from airleak between the patients face
and the interface. Many different interfaces are available. In the current
study we will test which of the most common interfaces is the best match with
NAVA.

The aims of the current study are:
1. Demonstrate that during noninvasive ventilation a patient with acute
respiratory failure is in better synchrony with the ventilator during NAVA
compared to PSV mode.
2. Demonstrate that the Servo-i ventilator is suitable for noninvasive
ventlation in both the PSV and NAVA mode.
3. Provide insight in patient preference for type of interface (nose vs full
face) during NAVA ventilation.

Pilot study
Prospectieve cross-over design

Noninvasive ventilation with two different ventilators (BiPAP Vision and
Servo-i), two ventilator modalities (pressure support ventilation and NAVA) and
two different interfaces (nose and full face mask).

1. Nasogastric tube
For this study patients need a dedicated nasogastric feeding tube that is
suitable for NAVA ventilation (12 or 16 french NAVA catheter). This catheter is
commercially available (Maquet, Solna Sweden) and is currently used in clinical
care worldwide, including our own ICU. initially a 16 Frnech catheter is used,
as this will be most suitable for later clinical purposes, inclusing feeding.
Although the 12 french cathter can be used for nasogastric feeding, more
frequent obstrction of the catheter occurs in our experience. If a patient
included in this study does not have a nasogastric tube yet, this dedicated
catheter will be inserted. As all mechanically ventilated patients in our ICU
need a nasogastric feeding tube, this will not impose additional discomfort or
risk for the patient. If the patient already has a nasogastric feeding tube
(but not the dedicated NAVA catheter), this tube needs to be replaced by the
NAVA catheter. The insertion of a 12 french nasogastric tube is associated with
temporal mild discomfort. The insertion of a nasogastric tube is a routine
procedure in ICU patients, and will be performed by an experienced critical
care nurse. Adequate position of the nasogastric tube will be verified
according to the clinical protocols for this procedure. If high-risk patients
are excluded as in the current study (see exclusion criteria) the risk of
complications is minimal.
Again it should be stressed that all mechanically ventilated patients need a
nasogastric tube, independent of this study.

2. Blood withdrawal
Blood will be withdrawn via an indwelling arterial catheter. This catheter is
available in all mechanically ventilated patients for routine clinical care.
During the study approximately 7.0 ml of blood will be withdrawn within a time
frame of 190 minutes. This low volume does not impose an additional risk for
adult patients.

3. Switching between ventilators and interfaces.
A. Interface
During the study patients will be ventilated with two different interfaces.
Therefore, after 90 minutes the patient will be disconnected from the
ventilator to change the interface. The change in interface will take less than
30 seconds. This brief period of disconnection from the ventilator will not
impose additional risks, as patients with very severe respiratory failure (not
able to maintain adequate oxygenation when disconnected from the ventilator for
2 minutes) will not be included in this study as sated in the exclusion
criteria. The disconnection from the ventilator in order to change the
interface does not take more than 30 seconds. It should be noted that during
routine clinical care patients are disconnected from the ventilator a couple of
times per day for mouth care and sometimes feeding.
B. Ventilator
As two different ventilators are tested in this study, the patient needs to be
disconnected from the ventilator twice. This disconnection will take less than
30 seconds and as outlined above will not impose an additional risk for the
patients.

It should be noticed that with the design of the study we have specifically
taken into account patient comfort. Although from a research perspective it
might have been better if we had randomized both the order of the ventilator
mode, the type of ventilator and the interface. However, this could have
resulted in much more ventilator disconnections (5). Although not a high risk
for the patients, it may result in additional discomfort, especially in
patients with acute respiratory failure. With the current study design the
patients are disconnected from the ventilator three times within a time frame
of ± 3 hours.