Clinical Comparison of Different Humidification Strategies During Noninvasive Ventilation With Helmet

Background. Non invasive positive pressure ventilation (NIV) is among first line treatments of acute respiratory failure. Several interfaces are available for non-invasive ventilation.Despite full face and oronasal masks are more frequently used, some evidence suggests that helmets may optimize patients' comfort and NIV tolerability.

During NIV, humidification strategies (heat and moisture exchangers HME or heated humidifiers HH) may significantly affect patient's comfort and work of breathing.

Despite physiological data suggested heated humidification as the best strategy during NIV with full face masks, no differences were found in a randomized controlled study assessing the effects of HME or HH on a pragmatic clinical outcome.

However, the higher dead space (i.e. 18 L/min) and rebreathing rate observed during helmet NIV make such results not applicable to this particular setting.

The investigators designed a randomized-crossover trial to assess the effect of four humidification strategies during helmet NIV on patients with acute respiratory failure, in terms of comfort, work of breathing and patient-ventilator interaction.

Methods. All awake, collaborative, hypoxemic patients requiring mechanical ventilation will be considered for the enrollment. Hypercapnic patients (i.e.PaCO2>45 mmHg) will be excluded.

Each enrolled patient will undergo helmet NIV with all the following humidification strategies in a random order. Each period will last 60 minutes.

• Passive humidification, double tube circuit.
• Heated humification (MR 730, Fisher & Paykel, Auckland, New Zealand), humidification chamber temperature 33°C.
• Heated humification (MR 730, Fisher & Paykel, Auckland, New Zealand), humidification chamber temperature 37°C.
• Passive humidification with HME, Y-piece circuit.

Ventilatory settings (Draeger Evita xl or Evita infinity ventilators):

Pressure support ventilation; pressure support=20 cmH20; FiO2 titrated to obtain SpO2 between 92 and 98%; positive end-expiratory pressure=10 cmH2O; maximum inspiratory time 0.9 seconds; inspiratory flow trigger = 2 l/min; expiratory trigger: 30% of the maximum inspiratory flow; pressurization time=0,00 s.

Such settings will be kept unchanged during the whole study period. An oesophageal catheter will be placed and secured to measure oesophageal pressure (Pes) and gastric pressure (Pga) (Nutrivent, Italy): the reliability of the measured pressure will be confirmed with an airway occlusion test during NIV with oronasal mask. Work of breathing will be estimated with the pressure-time product (PTP) of the pleural pressure.

A pneumotachograph (KleisTek) will record flow, airway pressure, Pes and Pga on a dedicated laptop.

At the end of each cycle, the patient will be asked to rate his/her discomfort on a visual analog scale (VAS) modified for ICU patients. The level of dyspnea will be assessed with the Borg dyspnea scale.

The following parameters will be record at the end of each cycle:

Arterial pressure, heart rate, respiratory rate, SpO2, pH, PCO2, PaO2, SaO2. Airway and esophageal pressure signals will be reviewed offline to detect patient-ventilator asynchronies (ineffective efforts, double cycling, premature cycling, delayed cycling) and asynchrony index (number of asynchrony events divided by the total respiratory rate computed as the sum of the number of ventilator cycles (triggered or not) and of wasted efforts) will be computed. The trigger delay will be also measured. The pressurization and depressurization velocity will be assessed with the PTP airway index 300 and 500 (inspiratory and expiratory), as suggested by Ferrone and coworkers. The work of breathing (WOB) for each breath will be estimated by PTPes.

An hygrometer (Dimar SRL, Italy) will measure and record on a dedicated laptop Helmet temperature, relative and absolute humidity.

Primary endpoints: patient's comfort, work of breathing and asynchrony index.

Sample Sizing:

Given the physiological design of the study, the investigators did not make an a priori sample size and plan to enroll 24 patients.