Simplified lung ultrasound protocol shows excellent prediction of extravascular lung water in ventilated intensive care patients

Abstract

Introduction: Ultrasound of the lung and quantification of B lines was recently introduced as a novel tool to detect overhydration. In the present study, we aimed to evaluate a four-region protocol of lung ultrasound to determine the pulmonary fluid status in ventilated patients in the intensive care unit.

Methods: Fifty patients underwent both lung ultrasound and transpulmonary thermodilution measurement with the PiCCO system. An ultrasound score based on number of single and confluent B lines per intercostal space was used to quantify pulmonary overhydration. To check for reproducibility, two different intensivists who were blinded as to the ultrasound pictures reassessed and classified them using the same scoring system. The results were compared with those obtained using other methods of evaluating hydration status, including extravascular lung water index (EVLWI) and intrathoracic blood volume index calculated with data from transpulmonary thermodilution measurements. Moreover, chest radiographs were assessed regarding signs of pulmonary overhydration and categorized based on a numeric rating scale.

Results: Lung water assessment by ultrasound using a simplified protocol showed excellent correlation with EVLWI over a broad range of lung hydration grades and ventilator settings. Correlation of chest radiography and EVLWI was less accurate. No correlation whatsoever was found with central venous pressure measurement.

Conclusion: Lung ultrasound is a useful, non-invasive tool in predicting hydration status in mechanically ventilated patients. The four-region protocol that we used is time-saving, correlates well with transpulmonary thermodilution measurements and performs markedly better than chest radiography.

Figure 1

Scheme of the four parasternal views corresponding to the intercostal spaces between the third and fourth ribs and between the sixth and seventh ribs used to calculate the ultrasound score.

Figure 2

Chest radiographs (left) and corresponding ultrasound screenshots (right) of two study patients. (A) Dry lung with a normal extravascular lung water index (EVLWI) and predominant A lines. (B) Severe, non-cardiac pulmonary edema with a high EVLWI and confluent B lines.

Figure 3

Correlation of the extravascular lung water index with the ultrasound score. (A) We found a close correlation of the ultrasound (US) score with the extravascular lung water index (EVLWI) (Spearman’s r = 0.91, P < 0.0001). (B) Correlation of the blinded US score as a mean of two independent examiners is shown (Spearman’s r = 0.72, P < 0.0001). (C) Bland-Altman plot comparing the difference (EVLWI − US score) with the average (of EVLWI and US score). Additionally, a linear regression (difference = 7.62 − 0.46 × average) and the 95% confident intervals (linear regression ± 1.96 × 3.6) are plotted. (D) Receiver operating characteristic curves of the US score obtained to identify patients with EVLWIs >7 and >15 show excellent diagnostic performance, as indicated by the areas under the curve of 0.9419 and 0.9636.

Figure 4

Comparison of pulmonary fluid status evaluated by chest radiography and ultrasound. US, Ultrasound.

Figure 5

Correlation of extravascular lung water index to chest radiography (Spearman’s r  = 0.33, and P  = 0.03) and central venous pressure (Spearman’s r  = 0.24, P  = 0.11). CVP, Central venous pressure; EVLWI, Extravascular lung water index; Rx, Chest radiography.