Objective evaluation of cough and diaphragmatic ultrasound improve the predictive capacity of spontaneous breathing test for extubation success - COBRE-US trial

Abstract

Background Use of clinical tests and ventilator weaning tests in conjunction with the spontaneous breathing test (SBT) can increase the predictive capacity of extubation success. Therefore, we assessed the predictive ability of SBT, cuff leak test, rapid-shallow-breathing-index, cough intensity, and velocity of diaphragmatic contraction (VDC) in successful weaning of mechanically ventilated patients. Methods A prospective multicenter observational study from in which consecutive adult patients admitted to 4 intensive care units. The performance of tests in discriminating between success and failure of the weaning process was assessed at different time points using ROC-curve in which the dependent variable was successful extubation. Results A total of 367 subjects who received invasive mechanical ventilation, 456 SBT were performed, with a success rate of 76.5%. The ROC curve of this model (0.52xCough)-(0.12xVDC) + 0.3 to predict a successful SBT was 0.63 (95%CI:0.60–0.65), a cut-off point > 0.84 had a sensitivity of 90.1% and a specificity of 25.3%. The ROC-curve of the model (5.7xSBT)+(0.75xCough)–(0.25xVDC)–4.5 to predict a successful extubation was 0.89 (95%CI:0.85–0.94), a cut-off point > 5.97 had a sensitivity of 98.9% and a specificity of 73.2%. The Hosmer-Lemeshow test showed no significant differences between observed and predicted values (p = 0.92). Conclusion Objective measures of cough and diaphragmatic function demonstrate excellent discriminatory ability to predict successful extubation with a cutoff of 5.97. The probability of having a successful extubation could be evaluated using the proposed model.