INTRODUCTION

Pulmonary ultrasound (PU) has become an essential tool for rapidly identifying the etiology of acute respiratory failure (ARF), following treatment progress, and clarifying non-specific chest radiograph abnormalities amongst critically ill patients1-5, and with test characteristics better than the clinical and6,7. When used in combination with cardiac and vascular ultrasound, it can enhance understanding of etiology3 and may reduce the need for or CT8-10.
Acquisition, interpretation, and integration of PU findings at an isolated point in time is essential to prompt an accurate diagnosis. Tracking PU changes over time is equally important in confirming a diagnosis and adjusting treatment. To do so requires a standardized approach to PU such that providers can document and agree not only with themselves, but also with each other over time11-14.
PU scoring models have been developed to meet the need for standardization and have been shown to correlate with various metrics inspecific patient populations5, 11, 15-17. Scoring systems correlate with mortality in patients with acute respiratory distress syndrome (ARDS)17,18, on hemodialysis19, with dyspnea and/or chest pain15, and congestive heart failure (CHF)20, and PU can be used to quantify extravascular lung water16, 21–24, which predicts outcomes in critically ill patients4, 15, 16, 25–35.
An ideal examination and scoring model would: 1) have good intra- and inter-rater agreement, 2) be rooted in a routine clinical PU workflow, 3) be technically feasible, 4) be quick to perform and document, but 5) be sensitive enough to detect significant but subtle pathology, and 6) yield a score that has clinical predictive value.
The primary objective of this study was to establish a PU workflow and scoring system in a large, diagnostically diverse group of intubated patients that met the criteria above, and was associated with clinical outcomes.