4307Lung ultrasound in stable ambulatory chronic heart failure patients

Abstract

Abstract Background The role of lung ultrasound (LUS) in diagnosis and response to diuretic treatment of patients with acute HF has been widely studied, but less is known about its value in chronic HF. Purpose To assess the prognostic value of LUS in a cohort of chronic HF stable ambulatory patients and to explore the relationship of LUS findings with clinical data, such as NYHA functional class, left ventricular ejection fraction (LVEF) and NTproBNP. Methods Consecutive stable ambulatory patients who attended a scheduled follow-up visit in a HF clinic were included. LUS were performed with a pocket device and examined 4 chest areas per side (two anterior and two lateral). Scans were analysed offline by two investigators blinded to clinical data, who evaluated the number of B-lines of each area. The addition number of B-lines of each area and the quartiles of such addition were used for the analyses. The primary outcome end-point was the composite of all-cause death or hospitalization due to HF at one year. Linear regression and Cox regression analyses were performed. Results Five-hundred seventy-seven patients were included between July 2016 and July 2017 (age 69±12 years, 72% men). The main HF aetiology was ischemic heart disease (43%) followed by dilated cardiomyopathy (20%). Median HF duration was 79 months (Q1-Q3 38–144). Mean LVEF was 45%±13 (mean LVEF when admitted at the Unit 34%±13). Most patients were in NYHA functional class II (70%), 13% were in class I and 17% in class III. Median NTproBNP was 722 ng/L (Q1-Q3 262–1760). Mean number of B-lines was 5±6 (Q1, 0; Q2, 1–3; Q3, 4–7; Q4, ≥8). The number of B-lines was associated with age (beta-coefficient 0.11, p<0.001), NYHA functional class (beta-coefficient 1.75, p<0.001), and logNTproBNP (beta-coefficient 1.40, p<0.001). Mean number of B-lines according to NYHA functional class was: class I, 3.5±6; class II, 4.9±6; and class III, 7.1±7. During the one year follow-up 47 patients suffered the primary end-point. In total there were 24 HF related hospitalizations and 26 deaths. In Cox regression analysis, Q4 of B-lines showed a double risk of suffering the primary end-point (HR 2.13 [95% CI 1.18–3.84], p=0.01). However, statistically significance was not maintained for LUS results in the multivariable analysis when age, NYHA functional class and logNTproBNP were included in the model, although a 38% increase in the risk of suffering the primary end-point for Q4 was observed (HR 1.38 [95% CI 0.75–2.54], p=0.31). Conclusion In outpatients with stable chronic HF, the number of B-lines detected in LUS was associated with age, NYHA functional class and NTproBNP. Patients having ≥8 B-lines had a significant double risk of HF related hospitalization or all-cause death at one year. However, when strongly powerful prognostic variables such as NYHA class and NTproBNP were included in the model LUS did not retain an independent prognostic role.