Patients

Echocardiography was performed in 150 patients admitted to the Department of Internal Diseases, Hypertension and Angiology with exacerbation of heart failure who fulfilled the following criteria: (1) blood samples for TB and transaminases were taken on admission or <12 hours after admission, (2) time from admission to echocardiographic study was <24 hours, (3) patients had no known active liver disease on the basis of history and the clinical judgment of the physician in charge, (4) patients were not current chronic alcohol users, and (5) patients were not on respiratory or inotropic support. All patients admitted to our department with exacerbation of heart failure who fulfilled these criteria were included in the study, except those admitted between Friday afternoon and Sunday morning, who were not eligible because of the time criterion (>24 hours from admission). The diagnosis of heart failure was based on clinical, laboratory, and echocardiographic features: typical symptoms and findings on physical examination, elevated N-terminal pro–brain natriuretic peptide (>450 pg/mL), and significant structural abnormalities on echocardiography. Ninety-one percent of patients had been previously hospitalized for heart failure or had significant cardiac disease diagnosed in the past. In 9% of patients, heart failure was diagnosed for the first time, with gradual progression of symptoms, that had started ≥4 weeks before the current admission. Therefore, we diagnosed our patients as having exacerbation of chronic heart failure. Because of the “noninvasive cardiology” profile of our department, we did not admit and include patients with cardiogenic shock or acute heart failure due to acute myocardial infarction, myocarditis, or acute valvular insufficiency. All patients underwent chest radiography on admission. Eighty-five percent of patients had radiologic signs of pulmonary congestion, and 56% of patients had signs of pleural effusion on chest radiography. All patients were in New York Heart Association functional class III or IV. There were eight in-hospital deaths (5%). The mean length of hospitalization was 12.5 days.

There were 41% women and 59% men. The mean age of the population was 75 ± 11 years. Etiologies of cardiac failure included coronary artery disease (54%), chronic valvular disease (11%), tachycardiomyopathy due to atrial fibrillation or flutter with fast ventricular rate (19%), arterial hypertension (7%), dilated cardiomyopathy (7%), hypertrophic cardiomyopathy (0.7%), restrictive cardiomyopathy (0.7%), and anthracycline-induced cardiomyopathy (0.7%). In this group, no constrictive pericarditis etiology (as a potential cause of liver dysfunction) was suspected on the basis of clinical and echocardiographic assessment. It included the presence of known causes of patients’ clinical status, dilatation of the ventricles in most patients, and lack of echocardiographic findings suggestive of constrictive pericarditis, especially abnormal ventricular septal motion, distortion of ventricular contours, and thickening of the pericardium. The clinical characteristics of the whole study group, with division into subgroups with reduced and preserved left ventricular systolic function, are presented in Table 1 . Echocardiographic and laboratory characteristics are presented in Table 2 . The study was approved by the local ethics committee, and patients gave written informed consent to participate in the study.

Biochemistry

Serum TB, AST, and ALT as well as other routine biochemistry parameters were collected on admission or <12 hours after admission and analyzed in the central laboratory of an academic hospital using an integrated chemistry system (Dimension Xpand; Siemens Healthcare, Erlangen, Germany). Serum concentration of TB ≥1.2 mg/dL, AST >40 U/L, and ALT >56 U/L were considered abnormal.

Transthoracic Echocardiography

Two-dimensional echocardiography (Vivid 9; GE Healthcare, Little Chalfont, United Kingdom) was performed <24 hours after admission by a single experienced physician. The mean time between blood testing and echocardiography was 15.3 hours. Examinations were stored on a workstation (EchoPAC; GE Healthcare) and analyzed offline by the same dedicated physician, who was blinded to patients’ laboratory results. The following right-heart functional parameters were analyzed: right ventricular end-diastolic diameter (RVEDD) measured from the apical four-chamber view, just above the level of the tricuspid annulus; right atrial area (RAA) measured from the apical four-chamber view just before the opening of the tricuspid valve; degree of TR on the basis of vena contracta diameter (>7 mm for severe TR, >3 and ≤7 mm for moderate TR); estimated right ventricular systolic pressure; tricuspid annular plane systolic excursion (TAPSE; mean value of five consecutive cycles); tricuspid lateral annular systolic velocity (RVVTI) by tissue Doppler (mean value of five consecutive cycles); estimated RAP on the basis of end-expiratory dimension and respiratory collapsibility of the inferior vena cava (IVC), according to the 2010 American Society of Echocardiography guidelines ; and portal vein pulsatility index (PVPI). The portal vein was visualized between the ribs at the right costal angle, with the patient in a supine position. First it was identified in two-dimensional mode as a large vessel entering the liver, with slightly hyperechogenic walls, then with color Doppler showing antegrade flow in red (using a 2.5-MHz echocardiographic probe with general abdominal preset) ( Figure 1 ), and then the portal vein Doppler waveform was recorded with pulsed-wave Doppler. Because of respiratory movement of the liver, only expiratory cycles, and in patients with the Cheyne-Stokes respiration pattern only apneic cycles, were analyzed. The pulsatility index was calculated as the ratio of the difference between the maximal velocity (Vmax) and minimal velocity (Vmin) to Vmax (Vmax − Vmin/Vmax). Pulsatile flow was diagnosed when PVPI was >0.5 ( Figure 2 ). Left ventricular end-diastolic diameter (LVEDD) was measured in the parasternal long-axis view. The left ventricular ejection fraction (LVEF) was measured using the biplane Simpson method from the apical four- and two-chamber views. Cardiac output was measured at the level of the left ventricular outflow tract (the mean velocity-time integral of five consecutive cardiac cycles was used for calculations). The cardiac index represented cardiac output indexed to body surface area (BSA). Technically adequate echocardiograms were acquired in all patients.

Main trunk of portal vein seen in two-dimensional (A) and color Doppler mode (B) .

(A) Example of nonpulsatile Doppler waveform of portal vein flow. (B) Example of pulsatile Doppler waveform of portal vein flow, with significant systolic decrease in portal vein flow velocity.

Statistical Analysis

Statistical analysis was performed using Statistica version 12 (Statsoft, Tulsa, Oklahoma). Continuous data are presented as mean ± SD, and numeric data are presented as percentages. For comparisons of discrete variables, χ ² tests were implemented. For comparisons of continuous variables between two groups, Student’s t tests or Mann-Whitney U tests were used, depending on the distribution of the variables. Univariate logistic regression was used for the prediction of elevated TB, elevated transaminases, and PVPI > 0.5 by different echocardiographic indices. Selected statistically significant ( P < .05) predictors were included and analyzed in the multivariate regression model. The multivariate logistic regression was performed only in the whole group of patients. Analysis in the subgroups was not performed, because of relatively small numbers of patients. For statistically significant predictors in the univariate logistic regression analysis, receiver operating characteristic curves were generated to evaluate diagnostic performance. Intraobserver and interobserver variability analysis was performed on 30 randomly selected recordings of portal vein flow, in a blinded fashion. Interobserver variability of PVPI was measured as two SDs of the mean of the differences in PVPI (in absolute values) between readings performed by two observers. Intraobserver variability of PVPI was measured as two SDs of the mean of the differences in PVPI (in absolute values) between two independent readings by the same observer. For the classification of PVPI as >0.5, κ statistics were used.