https://orcid.org/0000-0002-3184-3006
Abstract
Advanced heart failure (AdHF) is characterized by variable degrees of left ventricular (LV) dysfunction, myocardial fibrosis, and raised filling pressures which lead to left atrial (LA) dilatation and cavity dysfunction. This study investigated the relationship between LA peak atrial longitudinal strain (PALS), assessed by speckle-tracking echocardiography (STE), and invasive measures of LV filling pressures and fibrosis in a group of AdHF patients undergoing heart transplantation (HTX).
We consecutively enrolled patients with AdHF who underwent HTX at our Department. Demographic and basic echocardiographic data were registered, then invasive intracardiac pressures were obtained from right heart catheterization, and STE was also performed. After HTX, biopsy specimens from explanted hearts were collected to quantify the degree of LV myocardial fibrosis. Sixty-four patients were included in the study (mean age 62.5 ± 11 years, 42% female). The mean LV ejection fraction (LVEF) was 26.7 ± 6.1%, global PALS was 9.65 ± 4.5%, and mean pulmonary capillary wedge pressure (PCWP) was 18.8 ± 4.8 mmHg. Seventy-three % of patients proved to have severe LV fibrosis. Global PALS was inversely correlated with PCWP (R = −0.83; P < 0.0001) and with LV fibrosis severity (R = −0.78; P < 0.0001) but did not correlate with LVEF (R = 0.15; P = 0.2). Among echocardiographic indices of LV filling pressures, global PALS proved the strongest [area under the curve 0.955 (95% confidence interval 0.87–0.99)] predictor of raised (>18 mmHg) PCWP.
In patients with AdHF, reduced global PALS strongly correlated with the invasively assessed LV filling pressure and degree of LV fibrosis. Such relationship could be used as non-invasive indicator for optimum patient stratification for therapeutic strategies.
See the editorial comment for this article ‘New insights into assessing severity of advanced heart failure through left atrial mechanics’, by B.L. Gerber and B. Castilho, https://doi.org/10.1093/ehjci/jeae008.
Introduction
Heart transplantation (HTX) is considered one of the few treatment options for patients with advanced heart failure (AdHF), but irreversibly raised pulmonary vascular resistances represent an important cause for high mortality after HTX.1–4 Such a rise in vascular resistances reflects either a primary circulatory problem or is secondary to long-standing raised left atrial (LA) pressure and, consequently, pulmonary capillary wedge pressure (PCWP). Irreversibly raised pulmonary vascular resistances may affect a normally transplanted heart and cause rapid right ventricular dysfunction and failure.5
Several basic and advanced echocardiographic parameters estimating left ventricular (LV) filling pressures were identified, among which LA global peak atrial longitudinal strain (PALS) using speckle-tracking echocardiography (STE), which has been shown to correlate closely with LV end-diastolic pressure, irrespective of ejection fraction (EF).6 Abnormal LA deformation in the form of impaired global PALS can explain various diastolic disturbances, e.g. abnormal relaxation and raised pressures.
Failing hearts could present a wide spectrum of systolic function disturbances, which all gradually lead to a depressed LVEF. An independent but highly valuable predictor of LV dysfunction is myocardial fibrosis, a wide substitution of muscle fibres with fibrous tissue results in poor contractile and elastic properties of the myocardium, that impact systolic and diastolic function. These abnormalities lead to increased preload and filling pressures with additional neuro-hormonal activation causing further LV fibrosis with irreversible structural damage and poor prognosis.
The aim of this study was to investigate the relationship between LA function assessed by STE and invasively estimated increased LV filling pressure and myocardial fibrosis in patients with AdHF undergoing HTX.
Methods
Consecutive Caucasian patients with AdHF and reduced EF (HFrEF) who were undergoing HTX at Department of Cardiovascular Diseases, University of Siena, from March 2012 to June 2019 were enrolled. All patients underwent serial cardiological assessments including detailed echocardiographic examination. As a routine step before acceptance and while on the transplant list, patients underwent right heart catheterization to obtain invasive measurements of PCWP, mean pulmonary artery pressure (PAP), and cardiac index (CI). Raised PCWP was defined as >18 mmHg according to the Forrester Classification of HF.7 The Minnesota Living with Heart Failure Questionnaire (MLHFQ) was also obtained from all patients. Patients who underwent cardiac transplantation more than 1 month after the last ambulatory admission had the echocardiographic examination and right heart catheterization just few hours before surgery. The mean time interval between HTX and echocardiography/right heart catheterization was 16 ± 7 days. Patients were excluded if they had sub-optimum imaging quality of the LV or LA endocardial border and/or if they had atrial fibrillation. We also excluded patients with restrictive or infiltrative cardiomyopathies. All patients gave a written informed consent to participate in the study, which complied with the 1975 Declaration of Helsinki, and was approved by the local ethics committee.
Echocardiography
Echocardiographic studies were performed using a high-quality echocardiograph (Vivid e9, GE, Milwaukee, WI, USA), equipped with an M5S probe with 1.5–4.5 MHz frequency, with the subject in the left lateral recumbent position. Measurements of LV and LA dimensions, LVEF, and filling velocities were made in accordance with the current recommendations of the European Association of Cardiovascular Imaging and American Society of Echocardiography (EACVI/ASE).8 LVEF, measured by Simpson’s method, was used as a standard index of global systolic function. The ratio between peak early (E) and late (A) diastolic LV filling velocities and E wave deceleration time were used as standard indexes of LV diastolic function.9 LA volumes were measured using the biplane disc summation method, from the apical four- and two-chamber views and were subsequently indexed to body surface area to obtain LA volume indexed (LAVI). LA emptying fraction (LAEF) was calculated as (maximum LAVI − minimum LAVI) × 100/maximum LAVI. The time interval between the onset of the QRS on the superimposed electrocardiogram and the aortic and mitral valve opening and closure were measured using pulsed-wave Doppler from the LV outflow and inflow, respectively.
LV longitudinal function was studied using pulsed tissue Doppler imaging (TDI), by placing the sample volume at the level of lateral mitral annulus angle from the apical four-chamber view. Peak systolic (s′), early diastolic (e′), and late diastolic (a′) annular velocities were measured. S′ was considered as a relatively load-independent index of LV longitudinal systolic function. E′ was used as relatively load-independent markers of myocardial relaxation. The E/e′ ratio was also calculated and used as an adjunctive index of LV filling pressures.9
STE analysis was made from the apical four-, two-, and three-chamber view images using conventional 2D grey-scale echocardiography, during a brief breath hold and with a stable electrocardiogram recording. Care was taken to obtain true apical images using standard anatomical landmarks in each view without foreshortening the LV and the LA, and allowing a reliable delineation of the endocardial border. Three consecutive heart cycles were recorded and averaged measurements obtained. The frame rate was set between 60 and 80 frames/s.10
The analysis of the recorded images was performed offline by a single, experienced, and independent echocardiographer, who was not directly involved in the image acquisition and who had no knowledge of the histopathological findings, using a commercially available, semi-automated, 2D strain software (EchoPac, GE, Milwaukee, WI, USA).
Speckle-tracking myocardial measurements were made using conventional methods. LV endocardial border was manually traced in the four-, two-chamber, and apical long-axis views, delineating a region of interest (ROI), which is made of six segments. After segmental tracking, quality analysis, and manual adjustment of ROI, the longitudinal strain (LS) curves were generated by the software for each LV segment. In patients in whom some segments were excluded because of a difficulty in achieving adequate myocardial tracking, global LS (GLS) was calculated by averaging values measured in the remaining segments.11 As previously described, the LA endocardial border was traced manually in both apical four- and two-chamber views to outline LA ROI, composed of six segments. After the manual correction of the ROI, the software automatically generated the curves of LS for each segment. Global PALS as an index of the atrial reservoir was calculated, using the Q-wave as a time reference.12,13 The mean PALS values from the two- and four-chamber views were calculated to obtain global PALS. The E/e′ ratio was used in conjunction with LA global PALS to derive a dimensionless estimate of LA stiffness, calculated by the formula: LA stiffness = (E/e′)/LA global PALS.
Right heart catheterization
Invasive measurements of cardiac pressures and output were obtained by an expert operator blinded to the echocardiographic findings, on the same day of the echocardiographic examination, after 5 h maximum interval. The pressure transducers were balanced before data acquisition with the zero level at the mid-axillary line. Pulmonary artery (PA) catheters were used to measure PAP, mean right atrial pressure, and mean PCWP. The wedge position was verified by fluoroscopy, phasic changes in pressure waveforms and oxygen saturation. Cardiac output and CI were derived by the thermodilution technique (average of five cardiac cycles with <10% variation) and by the Fick equation through sampling of a mixed central venous blood gas taken in the PA and of an arterial blood gas.
Myocardial histopathology
Seven histopathological samples from the LV apex, the anterior wall, inferior septum, anterior septum, inferior wall, lateral wall, and the posterior wall, measuring ∼1 × 1 cm each, were taken from the explanted heart of each patient immediately after transplantation. Samples were fixed in 10% buffered formalin, embedded in paraffin, and were cut into slices of 4 μm thickness for haematoxylin–eosin Masson’s staining. For Masson’s staining, slices were de-waxed with xylol (two steps 2 min each and soaked into a series of gradient concentrations from 99 to 95% of alcohol). All slices were washed in distilled water and put in a solution of haematoxylin for 3 min. Subsequently, colour change was carried out with lithium carbonate. Slices were washed in pure water and coloured with red panceau staining (oven at 30°C for 20 s at 45 kW). Subsequently, slices were put into acid water and phosphomolybdic acid (1 min) and then finally green light was added, and the slices were washed with acid water. The extent of fibrosis was assessed in the sections arranged. The ratio of the fibrosis area to the total surface area of each section was used to assess the degree of LV fibrosis (%) as (fibrosis area/total area) × 100. An average value of the extent of LV fibrosis was obtained from the seven samples. According to the study by Segura et al.,14 we defined fibrosis as mild when accounting for 0–15%, moderate 15–30%, and severe >30% of the analysed myocardial tissue.
Statistical analysis
Data are shown as mean ± standard deviation (SD). A P-value of <0.05 was considered statistically significant. Pearson’s correlation coefficients were calculated to assess the relationship between continuous variables. The sensitivity and specificity were calculated using standard definitions. Receiver operating characteristic (ROC) curves were constructed and the area under the curve (AUC) was calculated for the prediction of PCWP > 18 mmHg and severe LV fibrosis. The Youden index was used to assess the optimal cut-off values emerging from ROC curves. Statistical analyses were performed using SPSS (Statistical Package for the Social Sciences, Chicago, IL, USA) software Release 16.0.
Results
Of 76 patients enrolled, 6 were excluded being not in sinus rhythm, 5 for a poor acoustic window, and 1 for unwillingness to sign informed consent. Thus, 64 patients were eligible to be enrolled (Figure 1), among whom 33 patients had normal PCWP (≤18 mmHg) and the remaining 31 patients had raised PCWP (>18 mmHg). General, clinical, and echocardiographic characteristics of the two study groups are shown in Table 1. Patients were predominantly males (58%) and the mean age was 62.5 ± 11 years. The sample was representative of a population of patients with HFrEF, 81% of whom with ischaemic aetiology. The mean LVEF was 26.7 ± 6.1%. Most patients were on the following conventional HF medications including beta-blockers and renin–angiotensin–aldosterone system inhibitors or angiotensin receptor-neprilysin inhibitors as well as intracardiac defibrillator (90% of patients, 70% of whom with cardiac resynchronization therapy) (Table 1). Thirty-seven patients (58%) were in New York Heart Association (NYHA) Class I and II and the remaining 27 patients (42%) in Class III and IV. The mean MLHFQ in the population was 29.5 ± 19.0.
Study enrolment flowchart.
Clinical and echocardiographic characteristics of the study population (N = 64)
| Clinical data | PCWP ≤ 18 | PCWP >18 | P-value |
|---|---|---|---|
| N = 33 | N = 31 | ||
| Age (years) | 62 ± 13 | 63 ± 9 | 0.50 |
| Gender (n female, %) | 13 (39) | 14 (45) | 0.60 |
| Body surface area (m2) | 1.9 ± 0.4 | 2.1 ± 0.9 | 0.08 |
| Hypertension (n, %) | 26 (79) | 25 (81) | 0.70 |
| Diabetes mellitus (n, %) | 8 (24) | 6 (19) | 0.30 |
| Hypercholesterolaemia (n, %) | 25 (76) | 23 (74) | 0.90 |
| Ischaemic aetiology (n, %) | 28 (85) | 24 (77) | 0.50 |
| ICD (n, %) | 4 (12) | 9 (29) | <0.001 |
| CRT-D (n, %) | 15 (45) | 30 (96) | <0.001 |
| NYHA class > II (n, %) | 8 (24) | 19 (61) | <0.001 |
| MLHFQ score | 18 ± 16 | 41 ± 22 | <0.001 |
| ACE-inhibitors or ARB (n, %) | 15 (45) | 14 (45) | 0.60 |
| ARNI (n, %) | 12 (36) | 11 (35) | 0.08 |
| Beta-blockers (n, %) | 26 (79) | 24 (77) | 0.90 |
| Spironolactone (n, %) | 27 (82) | 19 (61) | 0.05 |
| Loop diuretics (n, %) | 24 (72) | 30 (96) | 0.06 |
| Statins (n, %) | 18 (54) | 30 (97) | 0.05 |
| Left atrial area (cm2) | 25 ± 6 | 34 ± 13 | 0.004 |
| Left atrial volume indexed (mL/m2) | 44 ± 17 | 64 ± 24 | 0.001 |
| Left atrial emptying fraction (%) | 16.0 ± 3.1 | 10.4 ± 3.6 | 0.001 |
| Left atrial stiffness | 0.7 ± 0.48 | 1.7 ± 0.97 | <0.001 |
| End-diastolic LV volume (mL) | 169 ± 57 | 212 ± 96 | 0.05 |
| End-systolic LV volume (mL) | 112 ± 47 | 153 ± 80 | 0.05 |
| LV mass index (g/m2) | 117 ± 27 | 118 ± 34 | 0.30 |
| LV ejection fraction (%) | 27.1 ± 2.4 | 26.3 ± 9.8 | 0.20 |
| Mitral E (cm/s) | 0.7 ± 0.2 | 0.9 ± 0.3 | 0.002 |
| Mitral E/A ratio | 1.8 ± 2.0 | 3.3 ± 1.6 | 0.008 |
| e′mit (cm/s) | 7.1 ± 0.38 | 8.2 ± 0.12 | 0.50 |
| E/e′ (cm/s) | 13.6 ± 5.0 | 15.7 ± 6.6 | 0.03 |
| GLS (%) | −11.1 ± 3.9 | −8.5 ± 4.1 | 0.002 |
| Global PALS (%) | 11.8 ± 5.2 | 7.9 ± 3.7 | <0.001 |
| Four-chamber PALS (%) | 11 ± 6.4 | 7.5 ± 3.4 | <0.001 |
| Two-chamber PALS (%) | 12.6 ± 4.0 | 8.2 ± 4.0 | <0.001 |
| Global TPLS (ms) | 410 ± 61 | 430 ± 95 | 0.001 |
| Clinical data | PCWP ≤ 18 | PCWP >18 | P-value |
|---|---|---|---|
| N = 33 | N = 31 | ||
| Age (years) | 62 ± 13 | 63 ± 9 | 0.50 |
| Gender (n female, %) | 13 (39) | 14 (45) | 0.60 |
| Body surface area (m2) | 1.9 ± 0.4 | 2.1 ± 0.9 | 0.08 |
| Hypertension (n, %) | 26 (79) | 25 (81) | 0.70 |
| Diabetes mellitus (n, %) | 8 (24) | 6 (19) | 0.30 |
| Hypercholesterolaemia (n, %) | 25 (76) | 23 (74) | 0.90 |
| Ischaemic aetiology (n, %) | 28 (85) | 24 (77) | 0.50 |
| ICD (n, %) | 4 (12) | 9 (29) | <0.001 |
| CRT-D (n, %) | 15 (45) | 30 (96) | <0.001 |
| NYHA class > II (n, %) | 8 (24) | 19 (61) | <0.001 |
| MLHFQ score | 18 ± 16 | 41 ± 22 | <0.001 |
| ACE-inhibitors or ARB (n, %) | 15 (45) | 14 (45) | 0.60 |
| ARNI (n, %) | 12 (36) | 11 (35) | 0.08 |
| Beta-blockers (n, %) | 26 (79) | 24 (77) | 0.90 |
| Spironolactone (n, %) | 27 (82) | 19 (61) | 0.05 |
| Loop diuretics (n, %) | 24 (72) | 30 (96) | 0.06 |
| Statins (n, %) | 18 (54) | 30 (97) | 0.05 |
| Left atrial area (cm2) | 25 ± 6 | 34 ± 13 | 0.004 |
| Left atrial volume indexed (mL/m2) | 44 ± 17 | 64 ± 24 | 0.001 |
| Left atrial emptying fraction (%) | 16.0 ± 3.1 | 10.4 ± 3.6 | 0.001 |
| Left atrial stiffness | 0.7 ± 0.48 | 1.7 ± 0.97 | <0.001 |
| End-diastolic LV volume (mL) | 169 ± 57 | 212 ± 96 | 0.05 |
| End-systolic LV volume (mL) | 112 ± 47 | 153 ± 80 | 0.05 |
| LV mass index (g/m2) | 117 ± 27 | 118 ± 34 | 0.30 |
| LV ejection fraction (%) | 27.1 ± 2.4 | 26.3 ± 9.8 | 0.20 |
| Mitral E (cm/s) | 0.7 ± 0.2 | 0.9 ± 0.3 | 0.002 |
| Mitral E/A ratio | 1.8 ± 2.0 | 3.3 ± 1.6 | 0.008 |
| e′mit (cm/s) | 7.1 ± 0.38 | 8.2 ± 0.12 | 0.50 |
| E/e′ (cm/s) | 13.6 ± 5.0 | 15.7 ± 6.6 | 0.03 |
| GLS (%) | −11.1 ± 3.9 | −8.5 ± 4.1 | 0.002 |
| Global PALS (%) | 11.8 ± 5.2 | 7.9 ± 3.7 | <0.001 |
| Four-chamber PALS (%) | 11 ± 6.4 | 7.5 ± 3.4 | <0.001 |
| Two-chamber PALS (%) | 12.6 ± 4.0 | 8.2 ± 4.0 | <0.001 |
| Global TPLS (ms) | 410 ± 61 | 430 ± 95 | 0.001 |
The bold values are those statistically significant.
A, atrial transmitral flow velocity; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor, CRT-D, previous cardiac resynchronization therapy with defibrillation implantation; E, early transmitral flow velocity; e′, early diastolic mitral annular velocity; GLS, global longitudinal strain; ICD, implantable cardioverter defibrillator; LV, left ventricular; MLHF, Minnesota living with heart failure; n.s., non-significant; PALS, peak atrial longitudinal strain; PCWP, pulmonary capillary wedge pressure; TPLS, time-to-peak atrial longitudinal strain; PAP, pulmonary arterial pressure.
Clinical and echocardiographic characteristics of the study population (N = 64)
| Clinical data | PCWP ≤ 18 | PCWP >18 | P-value |
|---|---|---|---|
| N = 33 | N = 31 | ||
| Age (years) | 62 ± 13 | 63 ± 9 | 0.50 |
| Gender (n female, %) | 13 (39) | 14 (45) | 0.60 |
| Body surface area (m2) | 1.9 ± 0.4 | 2.1 ± 0.9 | 0.08 |
| Hypertension (n, %) | 26 (79) | 25 (81) | 0.70 |
| Diabetes mellitus (n, %) | 8 (24) | 6 (19) | 0.30 |
| Hypercholesterolaemia (n, %) | 25 (76) | 23 (74) | 0.90 |
| Ischaemic aetiology (n, %) | 28 (85) | 24 (77) | 0.50 |
| ICD (n, %) | 4 (12) | 9 (29) | <0.001 |
| CRT-D (n, %) | 15 (45) | 30 (96) | <0.001 |
| NYHA class > II (n, %) | 8 (24) | 19 (61) | <0.001 |
| MLHFQ score | 18 ± 16 | 41 ± 22 | <0.001 |
| ACE-inhibitors or ARB (n, %) | 15 (45) | 14 (45) | 0.60 |
| ARNI (n, %) | 12 (36) | 11 (35) | 0.08 |
| Beta-blockers (n, %) | 26 (79) | 24 (77) | 0.90 |
| Spironolactone (n, %) | 27 (82) | 19 (61) | 0.05 |
| Loop diuretics (n, %) | 24 (72) | 30 (96) | 0.06 |
| Statins (n, %) | 18 (54) | 30 (97) | 0.05 |
| Left atrial area (cm2) | 25 ± 6 | 34 ± 13 | 0.004 |
| Left atrial volume indexed (mL/m2) | 44 ± 17 | 64 ± 24 | 0.001 |
| Left atrial emptying fraction (%) | 16.0 ± 3.1 | 10.4 ± 3.6 | 0.001 |
| Left atrial stiffness | 0.7 ± 0.48 | 1.7 ± 0.97 | <0.001 |
| End-diastolic LV volume (mL) | 169 ± 57 | 212 ± 96 | 0.05 |
| End-systolic LV volume (mL) | 112 ± 47 | 153 ± 80 | 0.05 |
| LV mass index (g/m2) | 117 ± 27 | 118 ± 34 | 0.30 |
| LV ejection fraction (%) | 27.1 ± 2.4 | 26.3 ± 9.8 | 0.20 |
| Mitral E (cm/s) | 0.7 ± 0.2 | 0.9 ± 0.3 | 0.002 |
| Mitral E/A ratio | 1.8 ± 2.0 | 3.3 ± 1.6 | 0.008 |
| e′mit (cm/s) | 7.1 ± 0.38 | 8.2 ± 0.12 | 0.50 |
| E/e′ (cm/s) | 13.6 ± 5.0 | 15.7 ± 6.6 | 0.03 |
| GLS (%) | −11.1 ± 3.9 | −8.5 ± 4.1 | 0.002 |
| Global PALS (%) | 11.8 ± 5.2 | 7.9 ± 3.7 | <0.001 |
| Four-chamber PALS (%) | 11 ± 6.4 | 7.5 ± 3.4 | <0.001 |
| Two-chamber PALS (%) | 12.6 ± 4.0 | 8.2 ± 4.0 | <0.001 |
| Global TPLS (ms) | 410 ± 61 | 430 ± 95 | 0.001 |
| Clinical data | PCWP ≤ 18 | PCWP >18 | P-value |
|---|---|---|---|
| N = 33 | N = 31 | ||
| Age (years) | 62 ± 13 | 63 ± 9 | 0.50 |
| Gender (n female, %) | 13 (39) | 14 (45) | 0.60 |
| Body surface area (m2) | 1.9 ± 0.4 | 2.1 ± 0.9 | 0.08 |
| Hypertension (n, %) | 26 (79) | 25 (81) | 0.70 |
| Diabetes mellitus (n, %) | 8 (24) | 6 (19) | 0.30 |
| Hypercholesterolaemia (n, %) | 25 (76) | 23 (74) | 0.90 |
| Ischaemic aetiology (n, %) | 28 (85) | 24 (77) | 0.50 |
| ICD (n, %) | 4 (12) | 9 (29) | <0.001 |
| CRT-D (n, %) | 15 (45) | 30 (96) | <0.001 |
| NYHA class > II (n, %) | 8 (24) | 19 (61) | <0.001 |
| MLHFQ score | 18 ± 16 | 41 ± 22 | <0.001 |
| ACE-inhibitors or ARB (n, %) | 15 (45) | 14 (45) | 0.60 |
| ARNI (n, %) | 12 (36) | 11 (35) | 0.08 |
| Beta-blockers (n, %) | 26 (79) | 24 (77) | 0.90 |
| Spironolactone (n, %) | 27 (82) | 19 (61) | 0.05 |
| Loop diuretics (n, %) | 24 (72) | 30 (96) | 0.06 |
| Statins (n, %) | 18 (54) | 30 (97) | 0.05 |
| Left atrial area (cm2) | 25 ± 6 | 34 ± 13 | 0.004 |
| Left atrial volume indexed (mL/m2) | 44 ± 17 | 64 ± 24 | 0.001 |
| Left atrial emptying fraction (%) | 16.0 ± 3.1 | 10.4 ± 3.6 | 0.001 |
| Left atrial stiffness | 0.7 ± 0.48 | 1.7 ± 0.97 | <0.001 |
| End-diastolic LV volume (mL) | 169 ± 57 | 212 ± 96 | 0.05 |
| End-systolic LV volume (mL) | 112 ± 47 | 153 ± 80 | 0.05 |
| LV mass index (g/m2) | 117 ± 27 | 118 ± 34 | 0.30 |
| LV ejection fraction (%) | 27.1 ± 2.4 | 26.3 ± 9.8 | 0.20 |
| Mitral E (cm/s) | 0.7 ± 0.2 | 0.9 ± 0.3 | 0.002 |
| Mitral E/A ratio | 1.8 ± 2.0 | 3.3 ± 1.6 | 0.008 |
| e′mit (cm/s) | 7.1 ± 0.38 | 8.2 ± 0.12 | 0.50 |
| E/e′ (cm/s) | 13.6 ± 5.0 | 15.7 ± 6.6 | 0.03 |
| GLS (%) | −11.1 ± 3.9 | −8.5 ± 4.1 | 0.002 |
| Global PALS (%) | 11.8 ± 5.2 | 7.9 ± 3.7 | <0.001 |
| Four-chamber PALS (%) | 11 ± 6.4 | 7.5 ± 3.4 | <0.001 |
| Two-chamber PALS (%) | 12.6 ± 4.0 | 8.2 ± 4.0 | <0.001 |
| Global TPLS (ms) | 410 ± 61 | 430 ± 95 | 0.001 |
The bold values are those statistically significant.
A, atrial transmitral flow velocity; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor, CRT-D, previous cardiac resynchronization therapy with defibrillation implantation; E, early transmitral flow velocity; e′, early diastolic mitral annular velocity; GLS, global longitudinal strain; ICD, implantable cardioverter defibrillator; LV, left ventricular; MLHF, Minnesota living with heart failure; n.s., non-significant; PALS, peak atrial longitudinal strain; PCWP, pulmonary capillary wedge pressure; TPLS, time-to-peak atrial longitudinal strain; PAP, pulmonary arterial pressure.
There was a tendency for the cohort of patients to have left heart chambers remodelling with not only LV dilatation (Table 1), but also LA enlargement (mean LAVI 54.0 ± 20.5 mL/m2; Supplementary data online, Figure S1), which was more evident in those with PCWP >18 mmHg. All patients had some degree of diastolic dysfunction with the mean E/A ratio 2.6 ± 1.8 and E/e′ ratio 14.6 ± 5.9, that was significantly higher when PCWP was >18 mmHg. Both LV and LA STE analysis showed impaired intrinsic myocardial function, as shown by reduced LV GLS (overall −9.8 ± 4.0%) and reduced global PALS (overall 9.65 ± 4.5%). Patients with PCWP >18 mmHg had significantly worse LV and LA strain values (all P < 0.005, Table 1). The median and medium number of segments excluded in the analysis were 0 and 0.23 for LV GLS and 0 and 0.34, respectively, for PALS.
Right heart catheterization data are shown in Table 2. The mean PCWP in the population was 18.8 ± 4.8 mmHg and the mean PAP was 27.0 ± 9.9 mmHg, both slightly increased but not against the recommendation for HTX. The mean CI values were reduced. LV fibrosis was present in all biopsy specimens; 73% of the patients had severe LV fibrosis, with the whole population having at least moderate-to-severe LV fibrosis (>25%).
Catheterization data
| PALS ≤ 13% | PALS > 13% | P-value | |
|---|---|---|---|
| Heart rate (bpm) | 69 ± 15 | 70 ± 16 | 0.80 |
| Systolic blood pressure (mmHg) | 114 ± 16 | 116 ± 28 | 0.60 |
| Diastolic blood pressure (mmHg) | 73 ± 8 | 72 ± 10 | 0.70 |
| Mean PAP (mmHg) | 32 ± 9 | 22 ± 11 | 0.003 |
| PCWP (mmHg) | 24.3 ± 5.6 | 12.9 ± 4.0 | <0.001 |
| CI therm (mL/min/m2) | 2.0 ± 1.0 | 2.2 ± 0.7 | 0.09 |
| CI fick (mL/min/m2) | 1.9 ± 0.7 | 2.0 ± 0.3 | 0.70 |
| PALS ≤ 13% | PALS > 13% | P-value | |
|---|---|---|---|
| Heart rate (bpm) | 69 ± 15 | 70 ± 16 | 0.80 |
| Systolic blood pressure (mmHg) | 114 ± 16 | 116 ± 28 | 0.60 |
| Diastolic blood pressure (mmHg) | 73 ± 8 | 72 ± 10 | 0.70 |
| Mean PAP (mmHg) | 32 ± 9 | 22 ± 11 | 0.003 |
| PCWP (mmHg) | 24.3 ± 5.6 | 12.9 ± 4.0 | <0.001 |
| CI therm (mL/min/m2) | 2.0 ± 1.0 | 2.2 ± 0.7 | 0.09 |
| CI fick (mL/min/m2) | 1.9 ± 0.7 | 2.0 ± 0.3 | 0.70 |
The bold values are those statistically significant.
CI, cardiac index; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure.
Catheterization data
| PALS ≤ 13% | PALS > 13% | P-value | |
|---|---|---|---|
| Heart rate (bpm) | 69 ± 15 | 70 ± 16 | 0.80 |
| Systolic blood pressure (mmHg) | 114 ± 16 | 116 ± 28 | 0.60 |
| Diastolic blood pressure (mmHg) | 73 ± 8 | 72 ± 10 | 0.70 |
| Mean PAP (mmHg) | 32 ± 9 | 22 ± 11 | 0.003 |
| PCWP (mmHg) | 24.3 ± 5.6 | 12.9 ± 4.0 | <0.001 |
| CI therm (mL/min/m2) | 2.0 ± 1.0 | 2.2 ± 0.7 | 0.09 |
| CI fick (mL/min/m2) | 1.9 ± 0.7 | 2.0 ± 0.3 | 0.70 |
| PALS ≤ 13% | PALS > 13% | P-value | |
|---|---|---|---|
| Heart rate (bpm) | 69 ± 15 | 70 ± 16 | 0.80 |
| Systolic blood pressure (mmHg) | 114 ± 16 | 116 ± 28 | 0.60 |
| Diastolic blood pressure (mmHg) | 73 ± 8 | 72 ± 10 | 0.70 |
| Mean PAP (mmHg) | 32 ± 9 | 22 ± 11 | 0.003 |
| PCWP (mmHg) | 24.3 ± 5.6 | 12.9 ± 4.0 | <0.001 |
| CI therm (mL/min/m2) | 2.0 ± 1.0 | 2.2 ± 0.7 | 0.09 |
| CI fick (mL/min/m2) | 1.9 ± 0.7 | 2.0 ± 0.3 | 0.70 |
The bold values are those statistically significant.
CI, cardiac index; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure.
Global PALS was strongly inversely correlated with PCWP (R −0.83; P < 0.001) (Figure 2) and with LV fibrosis (R −0.78; P < 0.001) (Figure 3), but not with LV dimensions or LVEF. Global PALS correlated modestly with LA antero-posterior diameter (R −0.35; P = 0.05), LA area (R −0.37; P = 0.05), LA indexed volume (R −0.39; P = 0.01), and LAEF (R −0.41; P = 0.001). There was also a modest correlation between LV fibrosis and both LVEF and LV GLS (R −0.55 for the correlation of LV fibrosis with LV GLS; P = 0.005).
Global PALS and PCWP. Correlation between global PALS and PCWP assessed by right atrial catheterization.
Global PALS and LV fibrosis. Correlation between global PALS and LV fibrosis detected by myocardial biopsy.
Among the analysed echocardiographic indices (global PALS, LAVI, E/e′ ratio, LA stiffness, LAEF, LVEF, and LV GLS), a global PALS < 13% was the strongest predictor of raised PCWP >18 mmHg, with high sensitivity and specificity on ROC curve analysis [AUC 0.955 (95% confidence interval, CI 0.87–0.99) vs. 0.74, 0.54, 0.78, 0.49, 0.59, respectively] (Figure 4 and Supplementary data online, Figure S2). Moreover, global PALS proved to be an accurate marker of severe (>30%) LV fibrosis on the ROC curve [AUC = 0.90 (95% CI 0.82–0.97); P < 0.001] (see Supplementary data online, Figure S3).
ROC curves for raised PCWP. ROC curves for the overall performance of basic and advanced echocardiographic parameters in the estimation of raised (>18 mmHg) PCWP. Early diastolic filling wave by power Doppler/early diastolic mitral annulus velocity by tissue Doppler imaging (E/E′). GLS, global longitudinal strain; LA, left atrial; LAEF, left atrial emptying fraction; PALS, peak atrial longitudinal strain.
Discussion
Our study shows that in patients with AdHF, LA myocardial dysfunction is strongly correlated with PCWP, confirming previous results that showed it as added value in the non-invasive assessment of filling pressure in patients, irrespective of EF.15 Its assessment by STE proved to be superior to other basic and advanced echocardiographic parameters for predicting raised LV filling pressures. Moreover, we have shown a strong correlation between LV fibrosis, histopathologically quantified from myocardial samples, and LA strain. The respective relationship with LV myocardial function, e.g. LVEF and GLS was only modest.
The poor correlation we found between LA strain, GLS and LVEF, conflicts with previous results.16,17 This could probably be explained by the nature of the patients included in this study, with severely reduced LA function in the presence of severely reduced LVEF and LV GLS.
The LV longitudinal function has a major impact on LA function when LV function is preserved or moderately reduced. We suppose that in our population, the capability of LV to pull the mitral valve annulus towards the apex, promoting the LA reservoir function and filling, is so reduced that other determinants, mostly related to diastolic dysfunction and increased left heart filling pressure, play the major role. Previous studies have shown that LVEF, LV GLS, or E/e′ ratio poorly correlate with LV filling pressures in severely reduced LVEF, both in chronic15 and acute18 settings. Impaired LA function rather than size/volume has been identified as a predictor of adverse outcomes in the setting of HF, irrespective of other parameters of cardiac function19 including LVEF and GLS.20,21 Both EF and GLS represent two important parameters for the initial evaluation of patients with new-onset HF; however, their prognostic power falls in patients with AdHF, since LV intrinsic function is severely compromised and cardiac output depends on raised filling pressures, hence the poor risk stratification.22,23 In our patients, E/e′ ratio proved to have limited value in identifying most patients with raised PCWP. Although it has previously been regarded as a reliable surrogate of invasively assessed intracardiac pressures, E/e′ has several limitations. It is accurate in identifying patients with normal or high LV end-diastolic pressures but puts those with intermediate values in a ‘grey zone’.24 Also, E/e′ normally increases with age and changes proportionally with afterload, so its reflection of LV filling pressures cannot be consistently clinically reliable. Finally, TDI, and to some extent, spectral Doppler velocity recordings could be influenced by the acquisition angles.23
LA represents a connecting channel between the LV and pulmonary circulation. In patients with HF, the LA exhausts its anatomical and function properties to maintain physiological filling of the LV without impacting the pulmonary vascular resistance.25 LA myocardial contraction exaggerates, then the cavity enlarges and perseveres the raised pressures to maintain adequate LV filling and cardiac output. We have previously demonstrated how PALS is strongly correlated with the extent of LA fibrosis in patients with AdHF.26
In this study, we have illustrated that, in addition to the above, the reduced LA myocardial strain was related to the raised PCWP in end-stage HF, thus suggesting a coupling relationship between LA and LV. PALS also correlated, although to a lesser degree, with the degree of the reduced LV compliance but not with the magnitude of LV systolic dysfunction itself, expressed as LVEF. This relationship was not surprising since the strong relationship was between two LA components, myocardial function, and cavity pressure whereas LV parameters are anatomically remote to the LA.
The second important finding was the relationship between LA dysfunction and LV fibrosis, which is poorly established. One study investigated such relationship in patients with LVEF > 35% and concluded that LV fibrosis, defined by magnetic resonance imaging, does not affect LA function.27 To our knowledge, our study is the first to demonstrate biopsy-based LV fibrosis to correlate with LA dysfunction in HF patients with EF < 35%. In severe HF, LA is characterized by increased wall stiffness and progressive loss of elastic properties which, in turn, results in reduced compliance with its additional impact on HF symptoms and functional capacity.28 The inability of the atrial cavity to accommodate pressure fluctuations results in raised pulmonary venous pressures and capillary wedge pressures. While the pulmonary circulation reserve, on its turn, attempts to accommodate the pressure changes, the chronicity of such disturbed physiology causes raised pulmonary vascular resistance that paves the way to end-stage HF. This relationship is clearly shown by the correlation, we found, between PALS and extent of LV myocardial fibrosis. Thus, the availability and ability of a non-invasive, reproducible, and quick tool that can predict PCWP and extent of LV fibrosis puts PALS among the most useful parameters for such clinical scenarios. We have also previously29,30 shown that PALS is a reliable marker of LA fibrosis and is superior to E/e′ in predicting the invasively assessed LV filling pressures.26 It should be remembered that the attenuated PALS could also be influenced by both ultrastructural changes due to LA fibrosis and the high LV filling pressures itself or both concomitantly.
Clinical implications
In the present study, PALS emerged not only as the best echocardiographic parameter for the assessment of LV filling pressures but also as a good correlate with myocardial fibrosis and as a major predictor of severe myocardial fibrosis. The study population includes a selected group of patients with end-stage HF; however, the current percentage of patients reaching the advanced stage of the disease is increasing, thanks to new therapies and better diagnostic capabilities. The presence of extensive fibrosis in the LV is associated with a lower response to pharmacological and also to electrical therapies (e.g. cardiac resynchronization) in addition to a higher arrhythmic burden. The availability of parameters able to non-invasively predict the extent of fibrosis should help in the identification of patients who are near reaching end-stage disease in order to optimize the timing of referral to a centre for advanced therapies (HTX and left ventricular assist devices) and to tailor follow-up management strategy, thus avoiding unplanned hospitalizations or visits for worsening signs and symptoms.
Therefore, PALS, also characterized by good reproducibility and high feasibility, could be a useful tool for the selection of the best treatment strategy for AdHF patients.
Limitations
Despite its promising results, some limitations of this study should be mentioned. First, a single-centre study is known for its limitation, with small number of patients, however, presenting a heart-transplanted cohort would not be expected to have large cohort. Secondly, we are not able to provide analysis of clinical outcome after HTX for our patients, as this has still been collected. Finally, Cardiac Magnetic Resonance (CMR) data were not available for all our patients because of limited availability, a well-known limitation of CMR. Our biopsy data, on the other hand, provided a direct assessment of the presence and extent of myocardial fibrosis, which should be seen as more powerful evidence than what is obtained from imaging modalities, particularly in showing coexisting focal and diffuse patterns of fibrosis. We tried to be accurate in obtaining samples from seven different areas of the LV myocardium; however, it is probable that some of the samples taken and studied did not represent the overall pattern of fibrosis in the ventricle. Finally, data regarding the amount of LA fibrosis in the same patients were not available for a direct comparison with LV fibrosis.
Conclusion
In patients with end-stage HF, abnormal LA function expressed as reduced global PALS is a strong predictor of raised PCWP and is associated with the extent of LV myocardial fibrosis.
Supplementary data
Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.
Funding
None declared.
Data availability
Data available on request.
References
Author notes
Conflict of interest: None declared.
