Ultrasonographic Assessment of Extravascular Lung Water in Hospitalized Patients Requiring Hemodialysis: A Prospective Observational Study

Abstract

<b><i>Introduction:</i></b> Sonographic technologies can estimate extravascular lung water (EVLW) in hemodialysis (HD) patients. This study investigated the suitability of a handheld scanner in contrast to a portable scanner for quantifying EVLW in hospitalized patients requiring HD. <b><i>Methods:</i></b> In this prospective study, 54 hospitalized HD patients were enrolled. Bedside lung ultrasound was performed within 30 min before and after dialysis using handheld (phased array transducer, 1.7–3.8 MHz) and portable (curved probe, 5–2 MHz) ultrasound devices. Eight lung zones were scanned for total B-lines number (TBLN). The maximum diameter of inferior vena cava (IVC) was measured. We performed Passing-Bablok regression, Deming regression, Bland-Altman, and logistic regression analysis. <b><i>Results:</i></b> The 2 devices did not differ in measuring TBLN and IVC (<i>p</i> &#x3e; 0.05), showing a high correlation (<i>r</i> = 0.92 and <i>r</i> = 0.51, respectively). Passing-Bablok regression had a slope of 1.11 and an intercept of 0 for TBLN, and the slope of Deming regression was 1.02 within the CI bands of 0.94 and 1.11 in the full cohort. TBLN was logarithmically transformed for Bland-Altman analysis, showing a bias of 0.06 (TBLN = 1.2) between devices. The slope and intercept of the Deming regression in IVC measurements were 0.77 and 0.46, respectively; Bland-Altman plot showed a bias of −0.07. Compared with predialysis, TBLN significantly (<i>p</i> &#x3c; 0.001) decreased after dialysis, while IVC was unchanged (<i>p</i> = 0.16). Univariate analysis showed that cardiovascular disease (odds ratio [OR] 8.94 [2.13–61.96], <i>p</i> = 0.002), smoking history (OR 5.75 [1.8–20.46], <i>p</i> = 0.003), and right pleural effusion (OR 5.0 [1.2–25.99], <i>p</i> = 0.03) were strong predictors of EVLW indicated by TBLN ≥ 4. <b><i>Conclusion:</i></b> The lung and IVC findings obtained from handheld and portable ultrasound scanners are comparable and concordant. Cardiovascular disease and smoking history were strong predictors of EVLW. The use of TBLN to assess EVLW in hospitalized HD patients is feasible. Further studies are needed to determine if TBLN can help guide volume removal in HD patients.