Outcomes and risk factors of extubation failure: A multicenter study of the THAI surgical intensive care units (SICUs)

Abstract

Objective: To identify risk factors associated with extubation failure (EF) in patients admitted to surgical ICUs (SICUs). Material and Method: Data were gathered during April 2011-January 2013 by collecting demographic, admission details, daily summary, nutritional profile, APACHE II scores, and discharge summary from patients admitted to SICUs among nine university hospitals. Exclusion criteria include pediatric patients, non-consent patients, multiple trauma, cardiovascular and thoracic, and neurosurgical patients. Data were collected to the endpoint of 28 days of admission. Morbidity and mortality were determined. Complications or adverse events that occurred during admission were detailed in separate record forms. Result: Of 4,652 patients, 2,890 were intubated. Among them, 2,749 were successfully extubated leaving 141 with extubation failure. Overall incidence of EF was 4.88% (with range from 1.41-7.33). Patient characteristics in EF groups were compared to successful groups. Advanced age, presence of congestive heart failure, vascular disease, COPD, emergency surgery, poor APACHE II and SOFA scores, and concurrent use of vasopressors, inotropes and sedatives were significant differences. The most common causes of EF were respiratory failure, inability to cough and laryngeal edema. Outcomes of EF included prolonged length of ICU stay [2 (IQR 1-5) vs. 11 (IQR 6-15) days] and hospital stay [16 (IQR 10-27) vs. 23 (IQR 15-33) days]. Patients with EF were at risk of 6-fold longer ICU stay than successful extubation. Adjusted odds ratio of age, congestive heart failure, emergency surgery, and SOFA score were identified with statistical significance to be risk factors of EF. Conclusion: EF can affect outcomes of ICU admission. Identifying the risk factors associated with EF will help reduce its incidence and improve ICU outcomes. © 2016, Medical Association of Thailand. All rights reserved.