Lung Ultrasound Shows Good Diagnostic Accuracy for ARDS in Meta-Analysis of 5888 PatientsLung ultrasound shows high accuracy for detecting severe breathing trouble in ICU patients

This systematic review and meta-analysis evaluated the diagnostic accuracy of lung ultrasound (LUS) for acute respiratory distress syndrome (ARDS) in ICU settings. The analysis included 5888 patients from multiple studies, comparing LUS against established reference standards. The primary outcome was diagnostic accuracy, measured by diagnostic odds ratio (DOR), sensitivity, specificity, positive and negative likelihood ratios, and area under the receiver operating characteristic curve (AUROC).

The pooled diagnostic odds ratio was 14.98 (95% CI, 9.81-22.88; p < 0.001), indicating a strong overall diagnostic performance. Pooled sensitivity was 0.75 (95% CI, 0.62-0.85) and specificity was 0.87 (95% CI, 0.80-0.91). The positive likelihood ratio was 4.89 (95% CI, 3.67-6.52), and the negative likelihood ratio was 0.15 (95% CI, 0.11-0.21). The AUROC was 0.91 (95% CI, 0.88-0.93), suggesting excellent discriminatory ability.

For specific LUS findings, the specificity for bilateral B-patterns with 3 or more B-lines per intercostal space was 0.92 (95% CI, 0.87-0.96). This high specificity supports the use of LUS in confirming ARDS when characteristic patterns are present.

Safety and tolerability were not reported in the included studies, which is expected as LUS is a non-invasive imaging modality. No adverse events or discontinuations were documented.

Compared to prior landmark studies, this meta-analysis confirms and extends the evidence that LUS can be a reliable diagnostic tool for ARDS, particularly in ICU settings. Previous individual studies have shown variable accuracy, and this pooled analysis provides a more precise estimate. The high AUROC of 0.91 is comparable to or better than other bedside diagnostic methods.

Key methodological limitations include substantial heterogeneity (I² = 75.2%), which was partly explained by scanning protocols and operator experience. Higher diagnostic performance was observed in studies using 8-zone or more scanning protocols and those focusing on severe ARDS. Meta-regression identified scanning zones and operator experience as key sources of heterogeneity. The analysis did not report on funding or conflicts of interest.

Clinically, LUS offers a rapid, bedside, and radiation-free diagnostic option for ARDS, with good accuracy, especially in ICU and resource-limited settings. Comprehensive scanning protocols and trained operators enhance reliability, supporting LUS integration into clinical practice where advanced imaging is unavailable. However, the moderate sensitivity (0.75) means that a negative LUS does not rule out ARDS, and confirmatory testing may still be needed.

Remaining questions include the optimal scanning protocol and training requirements, the impact of operator experience on diagnostic accuracy, and the performance of LUS in specific ARDS subphenotypes. Prospective studies with standardized protocols are needed to further validate these findings.