The purpose of this study is to determine whether deflating the tracheal cuff in tracheostomized patients, during disconnections from ventilatory support along the weaning phase shortens the weaning time.

Randomized controlled trial.

Study protocol: patients will be eligible for the study after fulfilling standard criteria for weaning initiation and stratified for risk of aspiration (drink test)into two categories:

• Low/moderate risk patients (see later).
• High risk patients (spontaneous aspiration of pharyngeal secretions), will be excluded from the study.

After this evaluation patients will be randomized (concealed allocation) stratifying with the tracheostomy indication.

Weaning and Decannulation Protocol: Attempts to discontinue mechanical ventilation (MV) will be initiated when the tracheostomized patients fulfill the criteria (recovery from the precipitating illness; respiratory criteria consisting of a PaO2/fraction of inspired oxygen ratio of >150 mm Hg with a positive end-expiratory pressure <8 cm H2O, and an arterial pH >7.32; and clinical criteria consisting of the absence of myocardial ischemia determined by ECG, no need for vasoactive drugs or dopamine ≤5 µg/kg/min, a heart rate <140 beats/min, haemoglobin concentration >8 g/dL, temperature <38°C, no need for sedative drugs, presence of respiratory stimulus and spontaneous cough deemed clinically appropriate. Patients will be screened daily for these criteria).

A clinical algorithm for progressive weaning from mechanical ventilation will be followed at the discretion of the attending physician. Patients will be weaned following one of two methods: progressive withdrawal of pressure support ventilation or T-tube trials. When patients tolerate spontaneous breathing for 12 hours on two consecutive days, they will remain connected to the T-tube continuously.

When the patient remain disconnected from mechanical ventilation for at least 24 hours, the investigators will assess his or her preparedness for decannulation. The investigators will perform the tracheostomy tube occlusion test to exclude tracheal obstruction to airflow.

At this point, the attending physician clinically will evaluate the patient's capacity to clear respiratory secretions mainly based on the frequency of the need for suctioning and the characteristics of respiratory secretions.

The nurse will suction subglottic secretions while deflating the cuff to avoid the risk of aspiration.

Criteria for decannulation: respiratory secretions management considered adequate (<2 aspiration every 8 hours) and low risk of aspiration.

End-points:

• Weaning time: starting the day of first disconnection (lasting <30 minutes)and ending the day the patient is liberated from mechanical ventilation (patients will be considered weaned after 24 consecutive hours disconnected from mechanical ventilation).
• Tracheobronchitis and pneumonia after randomization. Safety considerations: The main safety concern is the risk of aspiration. First all patients will be categorized in three steps using the drink test: low risk patients (normal drink test) with no other specific interventions; moderate risk (abnormal drink test) the risk of aspiration study will include a videolaryngoscopy; high risk patients will be excluded from the study. In all patients a high flow-humidified oxygen-air mixture system will be connected trough the external cannula.

Inclusion criteria: tracheostomized patients for any of the following reasons:

• Prolonged time under mechanical ventilation (>21 days) or prolonged weaning.
• Motor component of Glasgow Coma Score <6 points or inability to adequate manage the airway (>2 aspirations every 8 hours).

Exclusion Criteria:

• High risk of aspiration (neuromuscular disease affecting bulbar functions, tracheostomies indicated as a part of a surgical procedure involving the upper airway or high risk of aspiration observed after specific evaluation previous to randomization).
• Death before starting weaning attempts.