Study Population and Design

This was a prospective observational echocardiographic study of patients admitted to the Heart Failure Unit at the Cleveland Clinic for hemodynamically tailored treatment of acute decompensated systolic heart failure. We prospectively identified patients aged ≥18 years who were admitted to the heart failure intensive care unit at the Cleveland Clinic for pulmonary catheter–based therapy for acute decompensated systolic heart failure. The decision to treat patients in the intensive care unit was made on the basis of their initial left ventricular and RV filling pressures. Subjects were included in this study if they met the following criteria: (1) impaired systolic function, defined by a left ventricular ejection fraction ≤ 35% for ≥6 months; (2) elevated filling pressures, defined as a pulmonary capillary wedge pressure > 18 mm Hg or a central venous pressure > 8 mm Hg; and (3) New York Heart Association functional class III or IV symptoms. The exclusion criteria were (1) mechanical ventilation, (2) chronic dialysis, (3) cardiac transplantation, and (4) tricuspid valve surgery. The Cleveland Clinic Institutional Review Board approved our study project, and oral and written informed consent was obtained from all subjects.

Hemodynamic data acquisition was immediately followed by echocardiographic data acquisition, within 12 hours of admission (baseline) and after 48 to 72 hours of intensive medical therapy. Central venous pressure data were collected using a standard fluid-filled balloon-tipped pulmonary artery catheter, using the average of five cycles after balancing the transducer to the zero level at the midaxillary line. Central venous pressure was assessed at end-expiration with the patient in a supine position by an observer unaware of the echocardiographic measurements. Central venous pressure was equivalent to that of RAP.

Transthoracic Echocardiography

Comprehensive two-dimensional echocardiography was performed at the bedside with a commercially available system. Standard two-dimensional and Doppler echocardiographic images were acquired with the patient in the supine position using a phased-array transducer in the parasternal, apical, and subcostal views and with the patient in the lateral decubitus position when acquiring tricuspid annular and tricuspid inflow velocities. Three consecutive cardiac cycles were recorded and stored for subsequent offline analysis by two independent investigators experienced with echocardiographic measurements. All echocardiographic data were acquired in end-expiration except for IVCd, which was also performed during sniff (IVCd _Insp ). All investigators were unaware of the time of registration and the identity of subjects.

Data Analysis

IVCd was measured from the long-axis subxiphoid view between 5 and 30 mm ⁵ from the IVC and right atrial junction during end-expiration in the supine position. IVCd measurements were repeated during sniff.

The percentage change in IVCd (IVCd%change) was calculated as IVCd _Insp /IVCd × 100%. Right atrial volume index (RAVI) and right atrial emptying fraction (RAEF) were calculated from the apical four-chamber view images at end-systole and end-diastole using Simpson’s method in the four-chamber view at both time points. RAEF was calculated by two-dimensional echocardiography as 1 − RAVI _s /RAVI _d , where subscripts s and d indicate systole and diastole, respectively.

In patients in sinus rhythm, we calculated the ratio between peak early diastolic velocities obtained from the pulsed Doppler signal of tricuspid inflow (E) and the pulsed tissue Doppler signal of the lateral tricuspid annulus imaged in the apical four-chamber view (E′ _ta ) and the ratio between the E-wave and atrial-wave velocities of the tricuspid inflow (RV E/A ratio).

The ratio between the systolic and diastolic velocities of hepatic flow (S/D _hep ) was calculated from the pulsed Doppler signal. The RV Tei index was calculated as (D − S)/S, where D is the time interval between the end and onset of tricuspid annular diastolic velocity, and S is the duration of tricuspid annular systolic velocity or RV ejection time.

RV regional isovolumic relaxation time was assessed by measuring the time interval between the end of systolic annular motion and the onset of the E′ wave in the apical four-chamber view. RVSP was determined by adding an estimate of RAP to 4 times the square of the peak tricuspid regurgitation (TR) velocity (expressed in meters per second) and the RAP estimated by method 1 (see below).

TR was graded semiquantitatively on a five-point scale (0 = none, 1 = mild, 2 = moderate, 3 = moderately severe, 4 = severe).

RV systolic function was determined by evaluating RV fractional area change (<35% represents dysfunction), peak systolic velocity of the tricuspid annulus (<10 cm/sec represents dysfunction), and tricuspid annular plane systolic excursion obtained by M-mode echocardiography (≤1.5 cm repsresents dysfunction).

Assessment of RAP

RAP was estimated by three proposed methods as follows.

Statistical Analysis

Data are expressed as mean ± SD for continuous data and as percentages for categorical data. Analyses were performed using SPSS version 10.0 or JMP version 10.0 (SAS Institute, Cary, NC). Agreement between invasively measured central venous pressure and previously proposed methods of RAP estimation was assessed by Bland-Altman analysis. For statistical analysis, ranges provided by methods 1 and 2 were substituted by corresponding specific values (2.5, 7.5, 12.5, 17.5, and 22.5 mm Hg for method 1 and 2.5, 7.5, 12.5, and 17.5 mm Hg for method 2).

Univariate analysis of potential RAP predictors was performed by linear regression. Patient ID numbers were used as dummy variables in regression analysis. A forward stepwise multivariate linear regression analysis was then performed using parameters with significant associations on univariate analysis. Classification and regression tree (CART) analysis, run as an automated (i.e., operator-independent) statistical procedure, was then performed to assess the incremental value of parameter combinations as potential predictors of central venous pressure. CART analysis creates a decision-making tree that, unlike in methods 1 and 2 (derived from expert experience), is based on statistical parameters for the selection of both predictors and their cut values. The following parameters were entered as potential predictors: maximal IVCd, minimal IVCd, percentage IVC collapse, E/E′ _ta ratio, S/D _hep ratio, RAVI, RAEF, regional isovolumic relaxation time, RV E/A ratio, RV Tei index, right atrial area, and RVSP. The CART algorithm was stopped if the final cell had a sample size of <5% or if the log worth of the parameter was <1. An arbitrary value of 10 mm Hg was used as a cutoff separating low from elevated RAP.

RAP estimates are useful if they decrease the initial uncertainty about individual RAP values. Because this initial uncertainty is equal to the standard deviation of RAP, we quantified the uncertainty decrease by precision gain (i.e., by how much the standard error of the estimate [SEE] of the particular method used to estimate RAP is smaller than the initial standard deviation). The SEE for both regression analysis and CART algorithm was calculated as

Patient Characteristics

A total of 73 patients met the eligibility criteria during the study period. One patient was excluded because of poor echocardiographic windows, and another patient was excluded because of inadequate central venous pressure measurements. Overall, 123 hemodynamic measurements were performed (71 measurements at baseline and 52 measurements 48–72 hours after admission to the unit; in 19 patients, measurements were not repeated, because their clinical status improved to the point that the invasive hemodynamic monitoring had been discontinued). Baseline characteristics of the final 71 patients are summarized in Table 1 . Of note, 67 patients (94%) had mean pulmonary artery pressures ≥ 25 mm Hg, and 51 (72%) had pulmonary capillary wedge pressures ≥ 15 mm Hg at baseline. Measurements of RAP < 10 mm Hg were present in 36 patients (29%). Only 8% had greater than moderate TR. Six patients (8%) had TR grades ≥ 3. Mean RV fractional area change was 30 ± 7%, with 51 patients (71%) having fractional area change < 35%. Mean peak systolic velocity of the tricuspid annulus was 8 ± 3 cm/sec, with 15 patients (22%) having s-wave velocities < 10 cm/sec. Mean tricuspid annular plane systolic excursion was 1.2 ± 0.3 cm, with 50 patients (96%) having ≤ 1.5-cm excursion. These results revealed that most patients had concomitant RV systolic dysfunction in addition to severely decreased left ventricular ejection fractions.

Echocardiographic and Hemodynamic Measurements

The average RAP value obtained invasively was 12.8 ± 6.3 mm Hg, and RAP was ≥10 mm Hg in 87 of 123 measurements. Figure 1 shows examples of the echocardiographic data used to calculate RAP estimates obtained in a single patient. Figures 2 A to 2C represent Bland-Altman plots that show the agreement between invasively measured RAP and RAP estimated by methods 1, 2, and 3, respectively. Although all methods showed significant correlations with invasively measured RAP ( r = 0.40, r = 0.28, and r = 0.19, respectively, P < .05 for all vs invasive RAP), all methods showed wide 95% limits of agreement (−12.6 to 11.4 mm Hg for method 1, −11.8 to 13.6 mm Hg for method 2, and −15.8 to 16.6 mm Hg for method 3). Of note, the use of more sophisticated echocardiographic parameters such as hepatic vein flow or tissue Doppler of the tricuspid annulus in methods 2 and 3 did not improve the reliability of estimation. As validation analysis showed relatively poor performance of previously proposed methods of RAP estimation, we tried to identify potential correlates of RAP pressure.

An example of echocardiographic data used to calculate RAP estimates in a single patient.

Bland-Altman agreement analysis of predictors of RAP by (A) method 1 using IVCd at baseline and during inspiration ( Figure 1 A), by (B) method 2 using inspiratory collapse of the IVC and the profile of hepatic vein flow ( Figure 1 B), and by (C) method 3 using the ratio of the peak early diastolic velocities of tricuspid inflow and of tricuspid annular motion ( Figure 1 C). See text for details.

Tables 2 and 3 show correlations between various echocardiographic parameters and invasively measured RAP by linear regression. As shown, the correlation with invasively measured RAP was moderate at best, with the strongest shown with IVCd _Insp , followed by percentage IVC change (collapse) during inspiration, and then resting IVCd. Importantly, the maximum precision gain was only 13%, meaning that initial standard deviation of 6.3 mm Hg decreased to 5.4 mm Hg. Multiple stepwise regression analysis failed to detect any additional independent predictor of RAP. Figures 3 A to 3D show the correlation between measured RAP and the four main echocardiographic parameters used to estimate RAP by methods 1 to 3. Furthermore, given the uncertainty of how the heart failure therapy has affected the measured parameters, we correlated the change of RAP during therapy with the changes in these four parameters ( Figure 3 E to 3H). Because E/E′ _ta underperformed compared with IVC measurements, we tested the mechanistic assumptions on which E/E′ _ta estimation of RAP is based. E′ _ta showed no correlation with markers of RV relaxation, while E-wave velocity showed only a weak correlation with RAP after controlling for RV relaxation (see Appendix ; available at www.onlinejase.com ).

Table 2

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Diastolic Stress Echocardiography in the Young: A Study in Nonathletic and Endurance-Trained Healthy Subjects
2. Coronary Flow Velocity Reserve during Pharmacologic Stress Echocardiography with Normal Contractility Adds Important Prognostic Value in Diabetic and Nondiabetic Patients
3. Association between Doppler Flow of Atrial Fibrillatory Contraction and Recurrence of Atrial Fibrillation after Electrical Cardioversion
4. Sonographer Travel Grants: Fueling Passion for Echocardiography

Stay updated, free articles. Join our Telegram channel

Join