Relation Between the Volume of Subglottic Secretion and Risk of Extubation Failure in ICU Patients (SEGEX)

Between 10 and 20% of patients develop extubation failure (10.7% in our ICU at 2018), which is related to an increased in-hospital death rate, infections, higher costs and longer hospital stays. Tracheal secretions, LVEF\<30%, MV \> 7d, Weak of cough are identified as risk factors of extubation failures but are not completely performants. Avoiding reintubation remains an important clinical goal. The subglottic secretion drainage above the cuff of an endotracheal tube is recognized as an effective method to prevent ventilator-associated pneumonia (VAP) in critically ill patients. The subglottic secretion drainage is used in a regular way in our ICU. Volume of subglottic secretion is not analysed in the scientific literature. Study : The investigators hypothesize that high volume of subglottic secretion before extubation is associated with high risk of extubation failure and à high risk of pneumonia post extubation. All planned extubations in the intensive care unit after 2 days minimal of mechanic ventilation are recorded on a designated form and standard variables charted. All data will be collected retrospectively after the extubation event. All included patients will be assigned a unique identification number (UID) by the investigator, who will secure the patient identifiers in an encrypted electronic file. The cause of extubation failure will be recorded (as identified by the attending physician on service, who is not involved in the study). For statistical analysis, group that fail extubation, will be compared, with the group that was an extubation success. Standard demographics (age, sex), patient disease related factors (diagnosis, duration of intubation, secretions), care factors (cuff leak test, p/f ratio prior to extubation,LVEF\<30% ), and post extubation care (post extubation respiratory support, stridor, blood gas, pneumonia) along with any complication during extubation and reintubation and reasons for reintubation will be collected and compared. As two groups are being compared, bivariate analyses utilizing Chi-square tests or univariate logistic regression for categorical variables and Student t -tests for interval variables, will be done. Investigators will perform an interim analysis at the end of 12 months and a final analysis