Lung Ultrasound for Early Diagnosis of Ventilator-Associated Pneumonia

Background: Lung ultrasound (LUS) has been successfully applied for monitoring aeration in ventilator-associated pneumonia (VAP) and to diagnose and monitor community-acquired pneumonia. However, no scientific evidence is yet available on whether LUS reliably improves the diagnosis of VAP.

Methods: In a multicenter prospective study of 99 patients with suspected VAP, we investigated the diagnostic performance of LUS findings of infection, subpleural consolidation, lobar consolidation, and dynamic arborescent/linear air bronchogram. We also evaluated the combination of LUS with direct microbiologic examination of endotracheal aspirates (EA). Scores for LUS findings and EA were analyzed in two ways. First, the clinical-LUS score (ventilator-associated pneumonia lung ultrasound score [VPLUS]) was calculated as follows: ≥ 2 areas with subpleural consolidations, 1 point; ≥ 1 area with dynamic arborescent/linear air bronchogram, 2 points; and purulent EA, 1 point. Second, the VPLUS-direct gram stain examination (EAgram) was scored as follows: ≥ 2 areas with subpleural consolidations, 1 point; ≥ 1 area with dynamic arborescent/linear air bronchogram, 2 points; purulent EA, 1 point; and positive direct gram stain EA examination, 2 points.

Results: For the diagnosis of VAP, subpleural consolidation and dynamic arborescent/linear air bronchogram had a positive predictive value of 86% with a positive likelihood ratio of 2.8. Two dynamic linear/arborescent air bronchograms produced a positive predictive value of 94% with a positive likelihood ratio of 7.1. The area under the curve for VPLUS-EAgram and VPLUS were 0.832 and 0.743, respectively. VPLUS-EAgram ≥ 3 had 77% (58-90) specificity and 78% (65-88) sensitivity; VPLUS ≥ 2 had 69% (50-84) specificity and 71% (58-81) sensitivity.

Conclusions: By detecting ultrasound features of infection, LUS was a reliable tool for early VAP diagnosis at the bedside.

Trial registry: ClinicalTrials.gov; No.: NCT02244723; URL: www.clinicaltrials.gov.