Follow-up of the TIE index can predict success at prolonged weaning

Abstract

Introduction: Patients in prolonged weaning represent a major challenge in intensive care, and the timed inspiratory effort index (TIE) has been shown to be a clinically valuable tool to predict the success weaning. Objective: to evaluate whether weekly serial measurements of the TIE index can predict the success of patients in prolonged weaning. Method: A prospective observational study, without which the patients able to initiate the weaning were followed up weekly with the TIE index. The area under the ROC curve was used to evaluate the accuracy of the TIE index. Multivariate Cox regression model was designed to test association with failure at prolonged weaning with the TIE index. Results: Seventy patients were selected, the median age 72 (62-78) years, the median duration of mechanical ventilation was 17.5 (14-28) days, and the APACHE II score was 28 (24-31). Thirty-five of the 70 cases (50%) had a fatal course, 22 of them after successful weaning. Two hundred and thirty tests were performed over 56 days, the area under the ROC curve of the TIE index was 0.93 ± 0.03. In the multivariate Cox regression, the association with failure at weaning, the TIE index ≥1.0 cmH2O/s revealed an inverse and independent association with the outcome (Hazard Ratio=0.27(0.07-0.97), P= 0.04. Also showed independent association with failure: age in years (HR=1.09(1.02-1.15), P=0.01), time of mechanical ventilation until the beginning of the protocol in days (HR=1.06(1.01-1.12), P=0.03 and weaning duration in days (HR=0.80(0.72-0.89), P Conclusion: the TIE index showed an excellent performance as a predictor in prolonged weaning patients, reinforcing that it can be a valuable tool for this population.