Lung Ultrasound Effectively Detects HIV-Associated Interstitial Pulmonary Disease

Objectives To prospectively evaluate lung ultrasound in comparison with radiography and computed tomography (CT) for detecting HIV-related lung diseases.Methods Ultrasound examinations in HIV-positive patients were evaluated by three raters; available conventional imaging was evaluated by another ra...

Full description

Bibliographic Details
Main Authors: Daniel T. Marggrander, Sinem Koç-Günel, Nesrin Tekeli-Camcı, Simon Martin, Rejane Golbach, Timo Wolf
Format: Article
Language:English
Published: Elsevier 2021-10-01
Series:International Journal of Infectious Diseases
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1201971221006639
Description
Summary:Objectives To prospectively evaluate lung ultrasound in comparison with radiography and computed tomography (CT) for detecting HIV-related lung diseases.Methods Ultrasound examinations in HIV-positive patients were evaluated by three raters; available conventional imaging was evaluated by another rater. Results were compared with each other and the definite diagnosis. Interrater reliability was calculated for each finding.Results Eighty HIV-positive patients received lung ultrasound examinations; 74 received conventional imaging. The overall sensitivity was 97.5% for CT, 90.7% for ultrasound and 78.1% for radiography. The most common diagnoses were Pneumocystis jirovecii pneumonia (21 cases) and bacterial pneumonia (17 cases). The most frequent and sensitive ultrasonographic findings were interstitial abnormalities indicated by B-lines, independent of the aetiology. Interrater reliability was high for interstitial abnormalities (ICC=0.82). The interrater reliability for consolidations and effusion increased during the study (r=0.88 and r=0.37, respectively).Conclusions Ultrasound is a fast, reliable and sensitive point-of-care tool, particularly in detecting interstitial lung disease, which is common in HIV-associated illness. It does not effectively discriminate between different aetiologies. A longer learning period might be required to reliably identify consolidations and effusions.
ISSN:1201-9712