DecanulAR. Predictors of Decannulation Difficulty: A Multicenter Cohort Study

Introduction: Tracheostomy (TQT) is perhaps the most common surgical intervention in the ICU. A prolonged use of a TQT cannula may subject patients to an increased risk of complications. Decannulation time in tracheostomies is becoming increasingly important during the recovery process after critical illnesses. At present, there is no prospective, multicenter study in our country that assesses tracheostomized patients as the population of the study. In addition, factors associated with decannulation difficulty are not usually analyzed. Objective: To describe the epidemiological characteristics of the study population, to report the rate of decannulation failure, to analyze the existence of independent risk factors associated with the impossibility of decannulation and to evaluate time-related mortality while achieving decannulation. Method: A prospective, multicenter cohort study that included patients who were tracheostomized at Intensive Care Units (ICUs) and patients who were admitted to Mechanical Ventilation Weaning and Rehabilitation Centers (MVWRCs) with TQTs. Epidemiological variables were recorded prior and during their hospitalization. The total duration of the study was one year. A sample size of 200 patients was calculated in order to draw a 5% rate (expected value for decannulation failure), determining the possibility to incur in a 5% alpha error and in a 20% beta error. Results: Initially, 48 centers from different cities around the country were recruited, and 36 centers contributed patients (31 from ICUs and 5 from MVWRCs). Five hundred and seventy-six patients were included, of whom 238 were removed since they could not be weaned from mechanical ventilation. The average age was 55 years (SD± 18.3), with a median of 58 years (IQR 43-70). There were more male patients (59%; 95% CI 53.8 - 64.2). One hundred and ninety-three patients who were weaned could be decannulated (57%; 95% CI 51.7-62.2). Cumulative incidence regarding decannulation failure was 3.1% in 7 months (95% CI 1.4 - 6.6). In the multivariate logistic regression analysis, the age group of patients over 70 years old (OR 3.40; 95% CI 1.51-7.66) and TQTs connected to surgical procedures (OR 1.74; 95% CI 1.08-2.79) were found as independent predictors contraindicating decannulation. Additionally, being a patient from an ICU versus being a patient from a MVWRC acted as a protective factor (OR 0.29; 95% CI 0.15-0.56). Likewise, the 90-day mortality rate was assessed using the Kaplan-M