• Keyvan Razazi, Jean-François Deux, Nicolas de Prost, Florence Boissier, Elise Cuquemelle, Frédéric Galactéros, Alain Rahmouni, Bernard Maître, Christian Brun-Buisson, and Armand Mekontso Dessap.
• From the Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Henri Mondor, DHU A-TVB, Service de Réanimation Médicale (KR, NdP, FB, EC, CB-B, AMD); UPEC, Faculté de Médecine de Créteil, IMRB, GRC CARMAS (KR, NdP, BM, CB-B, AMD); UPEC, Faculté de Médecine de Créteil (J-FD, FG, AR); Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Henri Mondor, Service de Radiologie (J-FD, AR); Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Henri Mondor, Unité des Maladies Génétiques du Globule Rouge (FG); and Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Henri Mondor, DHU A-TVB, Antenne de Pneumologie, Service de Réanimation Médicale (BM), Créteil, France.
• Medicine (Baltimore). 2016 Feb 1; 95 (7): e2553.

AbstractLung ultrasound (LU) is increasingly used to assess pleural and lung disease in intensive care unit (ICU) and emergency unit at the bedside. We assessed the performance of bedside chest radiograph (CR) and LU during severe acute chest syndrome (ACS), using computed tomography (CT) as the reference standard. We prospectively explored 44 ACS episodes (in 41 patients) admitted to the medical ICU. Three imaging findings were evaluated (consolidation, ground-glass opacities, and pleural effusion). A score was used to quantify and compare loss of lung aeration with each technique and assess its association with outcome. A total number of 496, 507, and 519 lung regions could be assessed by CT scan, bedside CR, and bedside LU, respectively. Consolidations were the most common pattern and prevailed in lung bases (especially postero-inferior regions). The agreement with CT scan patterns was significantly higher for LU as compared to CR (κ coefficients of 0.45 ± 0.03 vs 0.30 ± 0.03, P < 0.01 for the parenchyma, and 0.73 ± 0.08 vs 0.06 ± 0.09, P < 0.001 for pleural effusion). The Bland and Altman analysis showed a nonfixed bias of -1.0 (P = 0.12) between LU score and CT score whereas CR score underestimated CT score with a fixed bias of -5.8 (P < 0.001). The specificity for the detection of consolidated regions or pleural effusion (using CT scan as the reference standard) was high for LU and CR, whereas the sensitivity was high for LU but low for CR. As compared to others, ACS patients with an LU score above the median value of 11 had a larger volume of transfused and exsanguinated blood, greater oxygen requirements, more need for mechanical ventilation, and a longer ICU length of stay. LU outperformed CR for the diagnosis of consolidations and pleural effusion during ACS. Higher values of LU score identified patients at risk of worse outcome.

19 keywords...

hide…