ABC | Volume 110, Nº6, June 2018

Review Article Muniz et al Pulmonary ultrasound in patients with heart failure Arq Bras Cardiol. 2018; 110(6):577-584 Table 1 – Summary of the articles selected and their results. Diagnostic assessment of dyspnea in prehospital settings (AHF or DCHF) PU was useful for the diagnosis in 68% of dyspneic patients in the prehospital setting with no delay in treatment and/or transportation, PE being present in 100% of those with decompensated HF, in 17% of patients with ACS, and in 20% of patients with COPD (p < 0.01), PE thus being a diagnostic marker in patients with decompensated HF. 13 In the diagnosis of HF on PU, the S = 100% and E = 95% were comparable to those of NT-proBNP (> 1.000 pg/mL), S = 92% and E = 89%, and superior to those of the modified Boston criteria, S = 85% and E = 86%. The combination of PU and NT-proBNP showed S and E of 100%. 18 Diagnostic assessment of dyspnea in emergency settings (AHF or DCHF) Studies reported S ranging from 70% to 96.2% and E from 54% to 75%, 23-25,27,29,31 diagnostic reclassification ranging from 19% to 47%, 23,24 with change in treatment in 43% of the cases, 24 figures comparable to those of BNP > 500 (S = 75% and E = 83%). 27 PU accuracy of 90% versus 67% (p = 0.0001) for clinical examination, and 81% (p = 0.04) for the combination of clinical examination + NT-proBNP + X-ray. 25 PU was better for the diagnosis of DCHF (S = 100%) and of PNM (S = 75%) as compared to stethoscope auscultation (S = 89% and S = 73%, respectively). 26 Interobserver agreement was better in the anterior/superior thoracic zones for both pairs expert/expert and expert/beginner, 16 and the PU performed by beginners versus experts had S and E of 79-85% and 84-88%, respectively, 17,37 and PPV of 64-75% and NPV of 90.9-94%. 17,29 Global agreement with the gold-standard method for pulmonary edema interpretation on PU was 74%, higher than that with X-ray (58%, p< 0.0001). 28 A combination of PU and US of IVC had S = 94.3%, E = 91.9%, NPV = 91.9% and PPV = 94.3% to differentiate AHF from pulmonary disease, 29 and JVD-US is a sensitive test (S = 98.2%) to identify pulmonary edema in dyspneic patients with suspicion of congestive AHF. 30 Studies have shown an LR(+) of PU of 3.88-4.8% and an LR(-) of PU of 0.20-0.50% 24,31 for the diagnosis of AHF or DCHF, being higher than the LR(+) of NT‑proBNP [= 2.3] and similar to the LR(-) of NT-proBNP [= 0.24]. 31 Diagnostic assessment in intensive care settings (AHF or DCHF) Agreement of PU with the final diagnosis was 84%, with S = 86% and E = 87% for cardiac pulmonary edema, 32 and IVC values > 9 mm on B mode had S = 84.4% and E = 92.9% [LR(+) = 11.8, LR(-) = 0.16] for the diagnosis of cardiac dyspnea. 33 Diagnostic assessment in outpatient settings Primary outcome (hospitalization due to DCHF and all-cause death) was 4x more frequent in patients of the third tertile than in patients of the first tertile with B-lines ≥ 3 (p < 0.001), whose time alive or outside the hospital was shorter (p< 0.001). 36 The finding of B-lines or PE or both increased the risk of death or hospitalization (p< 0.05) 19 and correlated in a paired way with the estimates of PCWP (p < 0.001) and with the fluid impedance index (p < 0.001); the impedance monitoring alert detected clinical deterioration of HF with S = 92%, while B-lines ≥ 5 showed S = 83%. 35 HF decompensation was present in 68% of the patients when the number of B-lines ≥ 15, and correlated with NT-proBNP > 1000 (p < 0.0001) and with an E/e’ ratio > 15 (p < 0.0001). 34 Prognostic assessment Event-free survival (all-cause death and re-hospitalization) of patients with HF and B-lines ≥ 30 was shorter than that of patients with B-lines < 30 (p < 0.0001) in 3 months 10 and of patients with B-lines ≥ 15 in 6 months, 11 and the presence of B-lines ≥ 30 was a predictor of death with BNP > 700 (p = 0.002). 10 Therapeutic assessment The number of B-lines reduced with treatment (p < 0.05), and the PU score showed a linear correlation with the radiologic (p < 0.05) and clinical scores (p < 0.05) and with BNP levels (p < 0.05). 8 Assessment of PU as compared to other diagnostic methods An increase in the number of B-lines correlated with LVEDV (p = 0.036); 20 LV end-systolic diameter (p = 0.026); 20 PW (p = 0.009); 20 LV mass index (p = 0.001); 20 RA volume index (p = 0.005); 20 TR velocity (p = 0.005); 20 measures of RA, DPAP, MPAP, PVR, all p < 0,005, 21 and SPAP (p = 0.003-0,005), 20-21 and, for each B-line, there was an increase of 1 mm Hg in SPAP and of 0.1 Woods units in RVP. 21 In the analysis of the number of B-lines, the US device types used did not statistically differ (4 or 8 zones assessed; p= 0.67), 22 but the clip duration did differ: 4 versus 2 seconds (p < 0.001 for 4 and 8 zones) and 6 versus 4 seconds (p = 0.057 for 4 zones; and p = 0.018 for 8 zones). 22 AHF: acute heart failure; DCHF: decompensated chronic heart failure; HF: heart failure; PU: pulmonary ultrasound; COPD: chronic obstructive pulmonary disease; PE: pleural effusion;ACS: acute coronary syndrome; S: sensitivity; E: specificity; NPV: negative predictive value; PPV: positive predictive value; NT-proBNP: N-terminal pro-brain natriuretic peptide; LR(+): positive likelihood ratio; LR(-): negative likelihood ratio; US: ultrasound; X-ray: chest X-ray; PNM: pneumonia; IVC: inferior vena cava; JVD-US: jugular vein distension on ultrasound; PCWP: pulmonary capillary wedge pressure; BNP: brain natriuretic peptide; LVEDV: left ventricular end-diastolic volume; PW: posterior wall; LV: left ventricular; LA: left atrium; TR: tricuspid regurgitation; RA: right atrium; DPAP: diastolic pulmonary artery pressure; MPAP: mean pulmonary artery pressure; PVR: pulmonary vascular resistance; SPAP: systolic pulmonary artery pressure. In the intensive care setting, Dexheimer Neto et al., 32 using the BLUE protocol in dyspneic patients, have found an 84% agreement between PU and the final diagnosis of pneumonia or acute pulmonary edema (total kappa=0.81). Yamanoglu et al. 33 have detected the cardiac origin of dyspnea by using the caval index (sensitivity= 84.4% and specificity= 92.9%). In our clinical practice, we observed that PU increases the diagnostic accuracy of pulmonary congestion, being better than the stethoscope auscultation in both the emergency and the cardiac intensive care unit settings. In the outpatient care setting, Miglioranza et al. 34 have reported that a number of B-lines ≥15 correlates with NT‑proBNP > 1000 (p < 0.001), E/e’ ratio >15 (p = 0.001) and clinical assessment (p < 0.001), with sensitivity of 85% and specificity of 83%, for the risk of decompensated HF. Maines et al. 35 have reported a correlation between the presence of B-lines and the impedance fluid index (p < 0.001) of patients with HF at regular outpatient follow-up. Pulmonary ultrasound and prognostic assessment In the outpatient clinic context, Platz et al. 36 have identified that patients with more than three B-lines had a four-fold increase in the chance of hospitalization due to HF or of all‑cause death, being worth noting that 81% of those 580