Rational. The mismatch between the activity of the respiratory muscles and the assistance delivered by the ventilator results in patient-ventilator disharmony, which is commonly observed in ICU patients and is associated with dyspnea and patient-ventilator asynchrony. Both dyspnea and asynchrony are in turn associated with a worse prognosis. Unlike conventional modes of mechanical ventilation, such as pressure support ventilation (PSV) that deliver a constant level of assistance regardless of the patient effort, Proportional Assisted Ventilation (PAV) adjusts the level of ventilator assistance to the activity of respiratory muscles. To date, data on the impact of PAV on dyspnea and patient ventilator asynchrony are scarce and most studies have been conducted in healthy subjects or in ICU patients who had no severe dyspnea nor severe asynchrony. To our knowledge, there are no data in patients with severe patient-ventilator dysharmony.

Study Aim. To evaluate the impact of PAV on dyspnea and patient-ventilator asynchrony in ICU mechanically ventilated patients in intensive care with severe patient-ventilator disharmony defined as either severe dyspnea or severe patient-ventilator asynchrony.

Patients and Methods. Will be included 24 ICU mechanically ventilated patient exhibiting severe patient-ventilator dysharmony with PSV. The intensity of dyspnea will be assessed by the VAS, the ICRDOSS and by the electromyogram of extradiaphragmatic inspiratory muscles and pre inspiratory potential collected from the electroencephalogram. The prevalence of patient-ventilator asynchrony will be quantified.

Expected results. It is anticipated that the switch from PSV to PAV will decrease the prevalence and severity of dyspnea and the prevalence of patient-ventilator asynchrony.

Rational As opposed to controlled mechanical ventilation, partial modes of assisted ventilation maintains a certain level spontaneous activity of respiratory muscles. As a consequence, assisted ventilation may contribute to prevents ventilator induced diaphragm dysfunction (1-3), improves gas exchanges (4), reduces the use of sedative agents, which can ultimately shorten weaning from mechanical ventilation (5).

The most widely used partial ventilatory assistance mode is pressure support ventilation (PSV) (6), in which a constant preset level of pressure assists each inspiration regardless of the patient's inspiratory effort. Mismatching between patient demand and level of assistance, which the investigators will term patient-ventilator dysharmony in the present project is therefore possible and can be potentially harmful. On the one hand, underassistance may induce respiratory discomfort and dyspnea (7), which is an immediate cause of suffering, generates anxiety and is a source of delayed neuropsychological sequelae such as dark respiratory recollections and post-traumatic stress disorders(8-12). One the other hand, overassistance may cause lung overdistension and volutrauma (13). Finally, both underassistance and overassistance may generate patient-ventilator asynchrony that is associated with poorer clinical outcomes (14). Of notice, underassistance is likely to be associated with an asynchrony named double-triggering while over assistance is more commonly associated with ineffective efforts(15).

Proportional modes of mechanical ventilation have been designed to overcome this weakness of (PSV). Indeed, as opposed to PSV that delivers a constant level of assistance regardless of the patient inspiratory effort, proportional modes of ventilation adjust the amount of assistance delivered with respect to the patient's efforts. Proportional Assisted Ventilation (PAV) is one of these modes and adjusts ventilator assistance to the activity of respiratory muscles estimated by an algorithm (16-23). Previous studies have shown the potential benefits of PAV to prevent the risk of overassistance(24) and in turn to reduce the prevalence of ineffective effort (25-28). In addition, PAV increases the variability of the breathing pattern (17, 20-23, 29-32). To date, data on the impact of PAV on dyspnea and patient ventilator asynchrony are scarce (24-28, 33). Most of these works have been conducted in healthy subjects or in ICU patients with no severe dyspnea nor severe asynchrony (24-28, 33). To our knowledge, there are no data in patients with severe patient-ventilator dysharmony.

Because PAV adjusts the level of assistance to the activity of respiratory muscles, a surrogate of the respiratory drive, it is licit to hypothesize that PAV should prevent severe patient-ventilator dysharmony, defined as either severe dyspnea or severe patient-ventilator asynchrony.

The objective of the present research proposal is to evaluate the impact of PAV on dyspnea and patient-ventilator asynchrony in ICU mechanically ventilated patients in intensive care with severe patient-ventilator disharmony defined as either severe dyspnea or severe patient-ventilator asynchrony.

The specific objectives are to compare in these patients the impact of a switch of the ventilator mode from PSV to PAV in terms of:

1. The intensity of dyspnea quantified by a self-assessment visual analogic scale and by two electrophysiological tools such as the electromyogram of extradiaphragmatic inspiratory muscles and the pre inspiratory potentials on the electroencephalogram (see below, Patients and Methods).
2. The prevalence of two major patient-ventilator asynchronies that are ineffective efforts and double triggering (see below, Patients and Methods).

Materials and methods used and statistical methods

This observational, single-centre prospective study will be performed in the Medical intensive care unit (ICU) of the Respiratory and ICU Division of Pitié-Salpêtrière hospital, Paris, France.

1. Population, sampling The inclusion of patients will be done after informing patients and obtaining their informed consent.

1.1 Inclusion criteria Patients will be included as soon as the meet the following criteria.

• Intubation and mechanical ventilation for a respiratory cause with severe hypoxemia defined as a PaO2 to FiO2 ratio <300 recorded at least once during the present ICU stay.

• PSV ventilation for > 6 hours.

• Severe patient-ventilator disharmony

• Decision of the physician in charge of the patient to switch mechanical ventilation from PSV mode to PAV.

• Remaining duration of mechanical ventilation estimated ≥ 24 hours. 1.2 Exclusion criteria Exclusion criteria will be as follows.

• Severe hypoxemia defined as a PaO2 to FiO2 ratio <150 mmHg.

• Delirium according to the CAM-ICU (1)

• Hemodynamic instability defined by the need for intravenous fluids or catecholamine during the previous 24 hours.

• Age <18 years; pregnant woman. 1.3 sample size Our objective is to study a convenient sample of 24 patients. Given the recruitment unit, the duration of the study should be 6 months.

  2 Study design A first 10-minutes recording in PSV will be performed. Dyspnea-VAS, IC-RDOS will be measured at the beginning and at the end of this period. EMG and EEG will be recorded continuously. Patients will be subsequently switched to PAV.

The PAV mode will be delivered by Puritan Bennett 980 ventilator (Covidien, Boulder, USA). Levels of PEEP and FiO2 will be kept constant. The level of assistance in PAV, named %-assistance will be set in order to keep the patient in a respiratory effort zone corresponding to a respiratory muscles pressure time product (PTPmus) between 50 and 150 cm H2O • s / min (8). As it is not possible to calculate directly the PTPmus at bedside, the investigators will use as a substitute its main component, the pressure peak muscle of the airways according to the previous report from Carteaux et al.(8). This setting has been described extensively and its use has been the subject of a feasibility study in 50 patients. After a 10-minutes stabilization period, a 10-minutes recording will be performed. Dyspnea-VAS, IC-RDOS will be measured at the beginning and at the end of this period. EMG and EEG will be recorded continuously.

During the whole procedure, the usual hemodynamic and respiratory variables - non-invasive blood pressure or invasive if any, pulse oximetry, respiratory rate, - will be monitored continuously.

4 Statistical analysis Statistical analysis will be conducted with the Prism 5.0 software (GraphPad Software, USA). The distribution followed by the analysed data will be evaluated by the normality test of Kolmogorov-Smirnov. Probability of Type I error p less than or equal to 0.05 will be considered statistically significant. To investigate the effects of ventilation mode, the descriptors of dyspnea, the amplitude of the EMG as well as the PPI will be compared using a Mann-Whitney test. The prevalence of main asynchronies will be compared with a

Inclusion Criteria:

Patients will be included as soon as the meet the following criteria.

• Intubation and mechanical ventilation for a respiratory cause with severe hypoxemia defined as a PaO2 to FiO2 ratio <300 recorded at least once during the present ICU stay.

• PSV ventilation for > 6 hours.

• Severe patient-ventilator disharmony defined by either

  • a dyspnea ≥ 4 on a visual analogic scale (VAS) from 0 to 10 with respiratory rate ≥ 24 /minute and a drawing of neck muscles,
  • or by an asynchrony index (IA) ≥ 10%, defined as = number of asynchrony events/total respiratory rate (ventilator cycles +wasted efforts) × 100

• No improvement of disharmony despite an optimization of ventilator setting defined as follows.

  • No improvement of dyspnea or double triggering despite an increase of the level of pressure support that should not generate a tidal volume > 10 ml/kg
  • No improvement of ineffective efforts despite a decrease of the level of pressure support or generation of a dyspnea (defined as VAS>4) in response of the decrease of the level of pressure support.

• Decision of the physician in charge of the patient to switch mechanical ventilation from PSV mode to PAV.

• Remaining duration of mechanical ventilation estimated ≥ 24 hours.

• Patient able to communicate (Richmond Agitation and Sedation Scale between -1 and +1).

Exclusion Criteria:

• Exclusion criteria will be as follows.

  • Severe hypoxemia defined as a PaO2 to FiO2 ratio <150 mmHg.
  • Delirium according to the CAM-ICU (1)
  • Hemodynamic instability defined by the need for intravenous fluids or catecholamine during the previous 24 hours.
  • Age <18 years; pregnant woman.