Making a weaning decision for a patient on a mechanical ventilator is an important clinical issue. The most common index to predict successful weaning is the rapid shallow breathing index (RSBI), however, the accuracy of RSBI to predict successful weaning have been questioned.

The investigators proposed a new mathematical model and algorithm, called WIN, which capture the essential feature of the variability ruling the physiological dynamics to provides better perdition to wean than RSBI.

Making a weaning decision for a patient on a mechanical ventilator is an important clinical issue.

It is thus important to decide accurately when patients can be weaned from the ventilator. To increase the weaning success, the present common practice is to conduct spontaneous breathing trials to get physiological signals that may provide the information about capacity of successful weaning. The most common index is the rapid shallow breathing index (RSBI), however, the accuracy of RSBI to predict successful weaning have been questioned. Weaning failure usually results from a complex interplay of multiple factors. Thus, predictors targeting a single pathophysiologic mechanism tend to be unreliable for heterogeneous abnormalities.

The investigators proposed a new mathematical model and algorithm, which capture the essential feature of the variability ruling the physiological dynamics. Through the modern adaptive signal processing techniques, the investigators develop an index called WIN, which is evaluated from the 5 minutes continuous physiological signal and provides better perdition to wean than RSBI in a retrospective analysis. In this study, the investigators evaluate the predictive power of WIN and RSBI prospectively in patients undergoing weaning prospectively.

Inclusion Criteria:

1. Patients with mechanical ventilation via an endotracheal tube (oral endotracheal tube or tracheostomy tube) for >24 hours; 2. Patients are concomitant with presence of the following criteria of ready be weaned, a spontaneous breathing trial (SBT) will then be evaluated by 120-min T-piece:

1. clear improvement of the condition that led to mechanical ventilation;
2. no acute pulmonary or neuromuscular disease or signs of increased intracranial pressure;
3. conscious and lying on a bed with the upper body elevated to a 30◦ angle
4. adequate oxygenation (PaO2 ≥ 60mmHg and fraction of inspired oxygen inspired oxygen fraction (FiO2) ≤ 40% with positive end expiratory pressure (PEEP) ≤ 8cm H2O, or PaO2 /FiO2 >150 mmHg);
5. no significant respiratory acidosis, PaCO2<50mmHg, or increasing <10% for patients with chronic CO2 retention.
6. stable cardiovascular status (Heat beat ≤140/min, systolic blood pressure 90-160mmHg);
7. no requirement for vasopressive or inotropic dugs≥ 8 hours;
8. no intravenous sedatives during the previous 24 hours;
9. ability to cough while suction;
10. afebrile with ≤ 38◦ C temperature.
11. negative cuff leakage test: >110ml or >12%

Exclusion Criteria:

1. Presence of tracheostomy
2. Home ventilation prior to ICU admission
3. Decision not to re-intubate or withdrawal of care anticipated
4. Further surgery requiring sedation planned next 48 hours

1. Patients with mechanical ventilation via an endotracheal tube (oral endotracheal tube or tracheostomy tube) for >24 hours;
2. Patients are ready be weaned to performed a spontaneous breathing trial by 120-min T-piece.