Pulmonary Ultrasound and Diaphragmatic Shortening Fraction Combined Analysis for Extubation-Failure-Prediction in Critical Care Patients

Introduction

Invasive respiratory support is a cornerstone of Critical Care Medicine, however, protocols for withdrawal of mechanical ventilation are still far from perfect. Failure to extubation occurs in up to 20% of patients, despite a successful spontaneous breathing trial (SBT).

Methods

We prospectively included ventilated patients admitted to medical and surgical intensive care unit in a university hospital in northern Mexico. At the end of a successful SBT, we measured diaphragmatic shortening fraction (DSF) by the formula: diaphragmatic thickness at the end of inspiration – diaphragmatic thickness at the end of expiration/diaphragmatic thickness at the end of expiration × 100, and the presence of B-lines in five regions of the right and left lung. The primary objective was to determine whether analysis of DSF combined with pulmonary ultrasound improves prediction of extubation failure.

Results

Eighty-two patients were included, 24 (29.2%) failed to extubation. At univariate analysis, DSF (Youden's J: >30% [sensibility and specificity 62 and 50%, respectively]) and number of B-lines regions (Youden's J: >1 zone [sensibility and specificity 66 and 92%, respectively]) were significant related to extubation failure (area under the curve 0.66 [0.52–0.80] and 0.81 [0.70–0.93], respectively). At the binomial logistic regression, only the number of B-lines regions remains significantly related to extubation failure (OR 5.91 [2.33–14.98], P < .001).

Conclusion

In patients with a successfully SBT, the absence of B-lines significantly decreases the probability of extubation failure. Diaphragmatic shortening fraction analysis does not add predictive power over the use of pulmonary ultrasound.