We hypothesized that in patients with heart failure with normal left ventricular (LV) ejection fraction (HFNEF), the same fibrotic processes that affect the subendocardial layer of the LV could also alter the subendocardial fibers of the right ventricle (RV). Consequently, these alterations and to a lesser extent chronically elevated pulmonary arterial pressures would lead to both systolic and diastolic subendocardial dysfunction of the RV (i.e., impaired RV longitudinal systolic and diastolic function) in patients with HFNEF.
Patients with HFNEF and a control group consisting of asymptomatic patients with LV diastolic dysfunction (asymptomatic LVDD) matched by age, gender, and LV ejection fraction were studied by two-dimensional speckle-tracking echocardiography.
A total of 565 patients were included (201 with HFNEF and 364 with asymptomatic LVDD). RV longitudinal diastolic (RV global longitudinal early-diastolic strain rate [RV-SRe]) and systolic (RV global longitudinal systolic strain [RV-Strain]) function were significantly more impaired in patients with HFNEF than in patients with asymptomatic LVDD (HFNEF: RV-Strain −14.41% ± 3.80% and RV-SRe 0.86 ± 0.33 s −1 ; asymptomatic LVDD: RV-Strain −16.90% ± 4.28% and RV-SRe 1.02 ± 0.34 s −1 ; all P < .0001). On multiple regression analysis, LV global longitudinal systolic strain was the most important independent predictor of RV longitudinal systolic and diastolic function, in contrast with pulmonary arterial systolic pressure, which was weakly related to these functions. Furthermore, in patients with HFNEF the subendocardial function of both the LV and RV were significantly impaired in similar proportions. In that regard, in patients with HFNEF the prevalences of RV longitudinal systolic and diastolic dysfunction were 75% and 48%, whereas the rates of LV longitudinal systolic and diastolic dysfunction were 80% and 60%, respectively. In addition, patients with both systolic and diastolic longitudinal dysfunction of the RV presented worse New York Heart Association functional class.
In patients with HFNEF, RV subendocardial systolic and diastolic dysfunction are common and possibly associated with the same fibrotic processes that affect the subendocardial layer of the LV and to a lesser extent with RV pressure overload. Furthermore, our findings suggest that RV longitudinal systolic and diastolic dysfunction could contribute to the symptomatology of patients with HFNEF.
Heart failure (HF) with normal left ventricular (LV) ejection fraction (HFNEF) is a common pathology, with a significant increase of prevalence in patients aged ≥85 years. Unlike systolic HF, HFNEF is characterized by a normal LV systolic function often evaluated by biplane Simpson’s method. However, with the development of new echocardiographic technologies (two-dimensional speckle-tracking) recent studies have shown that despite a normal LV ejection fraction (LVEF), patients with HFNEF have significantly lower values of LV longitudinal systolic function both at rest and on submaximal exercise than healthy subjects, suggesting that in these patients LV subendocardial systolic function is not preserved. Nevertheless, despite these recent studies the analyses of right ventricular (RV) subendocardial systolic and diastolic function in patients with HFNEF have not been investigated.
Left-sided HF is a known cause of pulmonary arterial hypertension (PAH). Initial studies were focused on patients with systolic HF, and more recently it has been shown that PAH can also occur in HFNEF. An increase in RV afterload through the development of PAH secondary to chronic pulmonary venous hypertension has long been considered the main mechanism underlying RV dysfunction in patients with left-sided HF. However, recent findings suggest that the degree or duration of RV pressure overload may not fully explain RV failure and that RV myocardial fibrosis may play a role in the dysfunction of the RV. It is well known that comorbid conditions such as type 2 diabetes, obesity, hypertension, and a history of coronary artery disease (CAD) primarily affect the longitudinal systolic and diastolic function of the LV as a consequence of subendocardial fibrotic changes. In that regard, we hypothesized that in patients with HFNEF as a result of elevated rates of comorbidities (i.e., type 2 diabetes, obesity, hypertension, and history of CAD), the same fibrotic processes that affect the subendocardial layer of the LV could also alter the subendocardial fibers of the RV. Consequently, these alterations and to a lesser extent chronically elevated pulmonary arterial pressures would lead to both systolic and diastolic subendocardial dysfunction of the RV (i.e., impaired RV longitudinal systolic and diastolic function) in patients with HFNEF. With the aim of validating this hypothesis and elucidating the pathophysiologic mechanisms of HFNEF in the cardiopulmonary circulatory system, we analyzed the systolic and diastolic function of the RV and LV through two-dimensional speckle-tracking echocardiography in patients with HFNEF and in a control group consisting of asymptomatic patients with LV diastolic dysfunction (asymptomatic LVDD) matched by age, gender, and LVEF.
Materials and Methods
We enrolled consecutive patients aged ≥18 years with signs or symptoms of HF with an LVEF >50% by transthoracic echocardiography (according to the diagnostic criteria of the consensus of experts in HFNEF and in LV diastolic function of the American Society of Echocardiography [ASE] and the European Association of Echocardiography [EAE]) and a control group consisting of asymptomatic patients with LV diastolic dysfunction without a history of HFNEF (in accordance with the diagnostic criteria of the ASE and the EAE, i.e., septal e’ mitral annular peak velocity <8 cm/s, or lateral e’ mitral annular peak velocity <10 cm/s, or maximal left atrial [LA] volume index ≥34 mL/m 2 ). These two groups were matched by age, gender, and LVEF (matching 1:2, i.e., 1 patient with HFNEF/2 control patients). Three conditions were necessary for the diagnosis of HFNEF: (i) presence of signs or symptoms of congestive HF (dyspnea, New York Heart Association [NYHA] class ≥ II, pulmonary rales, pulmonary edema, bilateral lower extremity edema, hepatomegaly, or fatigue); (ii) presence of normal LV systolic function (LVEF >50% by Simpson’s method); and (iii) evidence of LV diastolic dysfunction (septal e’ mitral annular peak velocity <8 cm/s, or lateral e’ mitral annular peak velocity <10 cm/s, or maximal LA volume index ≥34 mL/m 2 ). We included consecutive inpatients and outpatients admitted in the Department of Cardiology (Campus Virchow-Klinikum) of the Charité University Hospital from April 1, 2009, to April 1, 2010. The Charité Institutional Review Board approved this research project, and informed consent was obtained from all subjects.
The selection of exclusion criteria in the present study was based on the consensus of experts in HFNEF and in LV diastolic function. To avoid reversible causes of myocardial dysfunction, patients with active CAD were excluded from this study, i.e., patients with unstable angina or non–ST-segment elevation myocardial infarction without revascularization or with revascularization in the last 72 hours, patients with ST-segment elevation acute myocardial infarction in the last 30 days, subjects waiting for coronary artery bypass graft or within 90 days postoperatively, subjects with chronic stable angina, and patients with evidence of myocardial ischemia assessed by stress echocardiography. Moreover, with the purpose of excluding causes of dyspnea or myocardial dysfunction other than HFNEF, patients with the following characteristics were excluded from this study: a) primary PAH or secondary PAH of causes other than isolated LV diastolic dysfunction or HFNEF (definition and classification of PAH were according to the diagnostic criteria of the task force for the diagnosis and treatment of PAH of the European Society of Cardiology) ; b) severe pulmonary disease, defined as pulmonary pathology with supplemental oxygen requirement; c) severe kidney disease, defined as a glomerular filtration rate <30 mL/min/1.72 m 2 for at least 3 months, history of renal transplantation, or severe acute renal failure with dialysis requirement; d) severe chronic liver disease or history of liver transplantation; e) congenital heart disease; f) pericardial disease, characterized by moderate or severe pericardial effusion (echo-free space in diastole ≥10 mm) or constrictive pericarditis; g) cardiomyopathy; h) valvular heart disease, defined as mild, moderate, or severe mitral or aortic stenosis; moderate or severe non-functional mitral or tricuspid regurgitation (TR); and moderate or severe aortic regurgitation (these definitions were in accordance with the diagnostic criteria of the guidelines for the management of patients with valvular heart disease of the American College of Cardiology). Furthermore, to avoid underestimations of myocardial and tricuspid or mitral annular measurements, patients with valvular heart surgery, mitral annular calcification (≥5 mm), cardiac pacing or cardiac resynchronization therapy, and poor two-dimensional quality in one or more RV or LV myocardial segments (not suitable for analysis by two-dimensional speckle-tracking echocardiography in apical four-chamber, two-chamber, and long-axis views) were also excluded from this study.
Transthoracic Echocardiography Measurements
All patients were examined at rest in the left lateral decubitus position using a Vivid-7 (GE Healthcare, Horten, Norway) ultrasound system followed by an offline analysis using an EchoPac 6.1 workstation (namely, measurements by two-dimensional speckle-tracking echocardiography). The echocardiographic measurements and analyses were performed by experienced echocardiographers blinded to each other’s results. RV and LV diameters, LV volumes, LV mass and grading of LV hypertrophy, LVEF (Simpson’s method), RV systolic parameters (i.e., RV fractional area change, lateral tricuspid annular plane systolic excursion [TAPSE], and lateral tricuspid annular systolic peak velocity [s’] using spectral tissue Doppler imaging), LA volume, LV diastolic function, and the noninvasive estimation of LV filling pressures (i.e., ratio of early-diastolic mitral inflow peak velocity by pulsed-wave Doppler to e’ mitral annular [average septal-lateral] peak velocity using tissue Doppler imaging) were assessed as recommended by the ASE and the EAE. The peak velocity of the TR jet by continuous-wave Doppler together with right atrial pressure (RAP) (using the modified Bernoulli equation and with fixed values of RAP at 10 mm Hg) were used to determine pulmonary arterial systolic pressure (PASP echo ). In addition, we have also assessed pulmonary vascular resistance (PVR), which was derived using a simple ratio of peak TR velocity (in meters per second) to the time-velocity integral (TVI) in the RV outflow tract (RVOT) (in centimeters), namely, PVR = [TR/TVI RVOT × 10 + 0.16]. All echocardiographic measurements (heart rate = 71 ± 11 bpm) were the average of three consecutive cycles and five cycles, if atrial fibrillation (AF) was present.
Two-Dimensional Speckle-Tracking Echocardiography
The analyses by two-dimensional speckle-tracking echocardiography were performed offline (frame rate: 68.8 ± 5.2 frames/s) and blinded to the clinical characteristics of the patients. The measurements of LV longitudinal systolic strain and LV longitudinal early-diastolic strain rate (SRe) were performed in the apical four-chamber, two-chamber, and long-axis views. The average value of peak longitudinal systolic strain and peak longitudinal early-diastolic SRe, obtained of all segments of the LV, was denominated as LV global longitudinal systolic strain (LV-Strain) and LV global longitudinal early-diastolic SRe (LV-SRe), respectively (see Figures I and II in Supplementary Data ). The measurements of RV longitudinal systolic strain and RV longitudinal early-diastolic SRe were performed in the apical four-chamber view. The average value of peak longitudinal systolic strain and peak longitudinal early-diastolic SRe, obtained of all segments of the free and septal wall of the RV, was defined as RV global longitudinal systolic strain (RV-Strain) and RV global longitudinal early-diastolic SRe (RV-SRe), respectively (see Figures III and IV in Supplementary Data ). All measurements by two-dimensional speckle-tracking echocardiography were the average of three consecutive cycles and five cycles, if atrial fibrillation was present.
The criteria of LV and RV longitudinal systolic and diastolic dysfunction were based on previously validated studies and cohorts of healthy subjects (i.e., values <95% confidence interval). Myocardial dysfunctions were defined as: LV longitudinal systolic dysfunction = LV-Strain >−16%, LV longitudinal diastolic dysfunction = LV-SRe < 0.80 s −1 , RV longitudinal systolic dysfunction = RV-Strain >−16%, RV longitudinal diastolic dysfunction = RV-SRe < 0.80 s −1 . In addition, noteworthy that LV and RV longitudinal systolic and diastolic function were considered as parameters of the systolic and diastolic function of the subendocardial fibers of the LV and the RV, respectively. Furthermore, the criteria used to define RV systolic dysfunction (i.e., TAPSE < 16 mm and fractional area change <35%), RV remodeling (i.e., RV diameter >42 mm at the base or >35 mm at the mid-level, or RV longitudinal dimension >86 mm), and RV hypertrophy (i.e., RV free wall >5 mm) were according to the recent guidelines for the echocardiographic assessment of the right heart developed by the ASE and the EAE. In addition, the definition of RV pressure overload or increased RV afterload (i.e., PASP echo >41 mm Hg or peak TR velocity >2.8 m/s or PVR >3 Wood units) were defined in accordance to the task force for the diagnosis and treatment of PAH of the European Society of Cardiology and to the recent guidelines for the echocardiographic assessment of the right heart developed by the ASE and EAE.
Continuous data are presented as mean ± SD and dichotomous data in percentage. Differences in continuous variables between groups (comparisons of two groups) were assessed by an unpaired Student t test only, because all data were normally distributed (the Kolmogorov–Smirnov test was used to test for normal distribution). Categoric variables were compared by chi-square test and Fisher exact test when appropriate. Comparisons among three or more groups were assessed by one-way analysis of variance. The relationships between continuous variables were analyzed using simple linear regression analysis. Selection of independent variables for prediction of RV and LV global longitudinal systolic strain and early-diastolic SRe were performed using forward stepwise multivariate analysis. With the purpose of determining the intra- and interobserver variability, we analyzed the mean absolute differences and interclass correlation coefficient of measurements of RV and LV global longitudinal systolic strain and early-diastolic SRe in 22 randomly selected patients. All statistical analyses were performed with SAS 9 (SAS Institute, Inc., Cary, NC). Differences were considered statistically significant with a P value < .05.
Patient Clinical Characteristics
A total of 654 patients met the eligibility criteria during the study period (218 with HFNEF and 436 with asymptomatic LVDD). However, 89 patients (17 with HFNEF and 72 with asymptomatic LVDD) could not be enrolled because of a poor two-dimensional quality in ≥1 segments of the RV and the LV not suitable for analysis by two-dimensional speckle-tracking echocardiography ( n = 24), severe kidney disease ( n = 12), cardiac pacing ( n = 8), severe chronic liver disease ( n = 8), non–ST-segment elevation acute myocardial infarction in the last 72 hours ( n = 7), ST-segment elevation acute myocardial infarction in the last 30 days ( n = 15), coronary artery bypass graft in the last 90 days ( n = 4), evidence of myocardial ischemia assessed by stress echocardiography ( n = 5), mild aortic stenosis ( n = 4), mild mitral stenosis ( n = 1), and moderate pericardial effusion ( n = 1). Thus, 565 patients were ultimately studied and analyzed (201 with HFNEF and 364 with asymptomatic LVDD). Clinical characteristics and echocardiographic measurements of the LV of these patients are summarized in Table 1 .
( n = 201)
|Asymptomatic LVDD |
( n = 364)
|Age, y||71.2 ± 10.1||70 ± 9.1||.0940|
|Women, n (%)||87 (43.2)||143 (39.2)||.3524|
|Body mass index, kg/m 2||29.3 ± 5.3||27.3 ± 4.1||.0002|
|Hemoglobin, g/dL||13.4 ± 1.6||13.5 ± 1.6||.2788|
|GFR, mL/min/1.73 m 2||67.9 ± 23||76 ± 20.6||<.0001|
|Hypertension, n (%)||199 (99)||300 (82.4)||<.0001|
|Type 2 diabetes, n (%)||79 (39.3)||75 (20.6)||<.0001|
|Obesity, n (%)||78 (38.8)||48 (13.1)||<.0001|
|History of CAD, n (%)||110 (54.7)||120 (32.9)||<.0001|
|Atrial fibrillation, n (%)||56 (27.8)||52 (14.2)||<.0001|
|Systolic blood pressure, mm Hg||137.1 ± 22.3||132.9 ± 20.7||.0653|
|Diastolic blood pressure, mm Hg||79.4 ± 11.9||78.6 ± 12.1||.5508|
|LV conventional measurements|
|LV ejection fraction, %||59.8 ± 7.2||60.2 ± 6.2||.0822|
|LVEDVI, mL/m 2||42.9 ± 12.5||42.9 ± 11.5||.9663|
|LV mass index, g/m 2||123.4 ± 30.9||105.1 ± 26||<.0001|
|Septal e’ mitral annular peak velocity, cm/s||4.8 ± 2.3||6 ± 1.4||<.0001|
|Lateral e’ mitral annular peak velocity, cm/s||6.3 ± 1.4||8 ± 1.5||<.0001|
|Mitral E/e’(average septal-lateral) ratio||17.4 ± 6||10.6 ± 3.8||<.0001|
|LV measurements by speckle-tracking|
|LV global longitudinal systolic strain, %||−13.63 ± 3.02||−18.87 ± 2.76||<.0001|
|LV longitudinal systolic dysfunction (strain >−16%), n (%)||162 (80.5)||56 (15.3)||<.0001|
|LV global longitudinal diastolic SRe, s −1||0.84 ± 0.28||1.05 ± 0.31||<.0001|
|LV longitudinal diastolic dysfunction (SRe < 0.80 s −1 ), n (%)||122 (60.6)||82 (22.5)||<.0001|
Left Ventricular Myocardial Systolic and Diastolic Function in HFNEF
LV global longitudinal systolic strain and early-diastolic SRe were significantly more impaired in patients with HFNEF than in patients with asymptomatic LVDD ( Table 1 ). Furthermore, in subgroup analyses, patients with both systolic and diastolic longitudinal dysfunction of the LV showed significantly higher values of LV filling pressures (i.e., mitral E/e’average septal-lateral ratio) and worse NYHA functional class compared with those with preserved LV longitudinal function ( Table I in Supplementary Data ). In addition, on multiple regression analysis, LV global longitudinal early-diastolic SRe was significantly related to LV global longitudinal systolic strain (r 0.59, P < .0001; Table II in Supplementary Data ).
Right Ventricular Systolic and Diastolic Function in HFNEF
The analyses of the cardiopulmonary circulatory system showed that patients with HFNEF had significantly more impaired RV longitudinal systolic and diastolic function, as well as higher PASP echo and PVR, than patients with asymptomatic LVDD ( Tables 2 and 3 , Figures 1 and 2 ), these differences also being significant between patients with or without atrial fibrillation (HFNEF without AF [ n = 145] = RV-Strain −14.92% ± 3.89% and RV-SRe 0.87 ± 0.31 s −1 vs asymptomatic LVDD without AF [ n = 312] = RV-Strain −17.14% ± 4.33% and RV-SRe 1.02 ± 0.34 s −1 , all P < .001; HFNEF with AF [ n = 56] = RV-Strain −13.48% ± 3.44% and RV-SRe 0.85 ± 0.37 s −1 vs asymptomatic LVDD with AF [ n = 52] = RV-Strain −15.32% ± 3.67% and RV-SRe 1.04 ± 0.34 s −1 , all P < .01). Moreover, patients in NYHA functional class II, III, and IV had significantly lower values of RV longitudinal systolic and diastolic function than patients in NYHA functional class I ( Table 4 ). Furthermore, on multiple regression analysis, LV-Strain, LV-SRe, and RV wall thickness were the most important independent predictors of RV longitudinal systolic and diastolic function, in contrast with PASP echo and PVR, which were weakly related to these functions ( Table 5 , Figures 3 and 4 ).
( n = 201)
|Asymptomatic LVDD |
( n = 364)
|RV systolic function|
|RV global longitudinal systolic strain, %||−14.41 ± 3.80||−16.90 ± 4.28||<.0001|
|TAPSE, mm||16.5 ± 3.5||18.7 ± 4.0||<.0001|
|RV fractional area change, %||40.1 ± 9.2||44.0 ± 8.7||<.0001|
|Lateral s’ tricuspid annular peak velocity by TDI, cm/s †||11.0 ± 1.9||13.0 ± 2.8||.0005|
|RV diastolic function|
|RV global longitudinal diastolic SRe, s −1||0.86 ± 0.33||1.02 ± 0.34||<.0001|
|Lateral e’ tricuspid annular peak velocity by TDI, cm/s †||9.3 ± 2.8||13.2 ± 4.2||<.0001|
|RV wall thickness|
|RV free wall, ∗ mm||5.5 ± 0.9||4.9 ± 0.9||<.0001|
|RV chamber dimensions|
|RV basal diameter, mm||30.2 ± 5||27.1 ± 4||<.0001|
|RV mid cavity diameter, mm||24.5 ± 6||23.2 ± 5||<.0001|
|RV longitudinal dimension, mm||69.2 ± 10||65.5 ± 8||<.0001|
( n = 201)
|Asymptomatic LVDD |
( n = 364)
|RV pressure overload|
|PASP echo , mm Hg||40.9 ± 10.5||32.2 ± 6.8||<.0001|
|Peak TR velocity, m/s||2.78 ± 0.48||2.36 ± 0.38||<.0001|
|PASP echo > 41 mm Hg or peak TR velocity > 2.8 m/s||52.7%||12.0%||<.0001|
|PASP echo > 45 mm Hg or peak TR velocity > 3 m/s||24.9%||1.6%||<.0001|
|PASP echo > 50 mm Hg or peak TR velocity > 3.2 m/s||15.9%||0.0%||<.0001|
|PASP echo > 60 mm Hg or peak TR velocity > 3.5 m/s||2.9%||0.0%||.0009|
|RV afterload ∗|
|Pulmonary vascular resistance (PVR), Wood units||1.45 ± 0.29||1.26 ± 0.24||<.0001|
|High PVR (>3 Wood units)||0.0%||0.0%||ns|
|Border-Line PVR (1.5 to 3 Wood units)||41.8%||12.3%||<.0001|
|Normal PVR (<1.5 Wood units)||58.2%||87.7%||<.0001|
∗ Measurements of PVR performed only in the subgroup of patients in whom it was possible to perform TVI in the RVOT by pulsed-Doppler at an angle < 30 degrees with respect to RVOT (i.e., 141 patients with HFNEF and 253 patients with asymptomatic LVDD). PVR = [peak velocity of the TR (in meters)/TVI RVOT (in centimeters) × 10 + 0.16].
|NYHA functional class|
|Class I |
( n = 364)
|Class II |
( n = 140)
|Class III |
( n = 36)
|Class IV |
( n = 25)
|P analysis of variance|
|Parameters of RV systolic function|
|RV global longitudinal systolic strain, %||−16.90 ± 4.28||−14.58 ± 3.72 ∗||−14.06 ± 4.02 †||−13.94 ± 4.03 ‡||<.0001|
|TAPSE, mm||18.7 ± 4.0||16.8 ± 2.9 ∗||16.6 ± 4.5 †||14.1 ± 3.8 ‡||<.0001|
|RV fractional area change, %||44.0 ± 8.7||40.3 ± 8.3 ∗||40.3 ± 9.2 †||34.0 ± 10.3 ‡||<.0001|
|Parameter of RV diastolic function|
|RV global longitudinal diastolic SRe, s −1||1.02 ± 0.34||0.83 ± 0.29 ∗||0.95 ± 0.42||0.91 ± 0.39||<.0001|
|Variables||RV longitudinal systolic strain||RV longitudinal diastolic SRe|
|R Value||P Value||R Value||P value|
|LV global longitudinal systolic strain, %||0.44||<.0001 †||0.29||<.0001 ∗|
|LV global longitudinal diastolic SRe, s −1||0.24||<.0001 †||0.26||<.0001 ∗|
|RV free wall thickness, mm||0.33||<.0001 †||0.28||<.0001 ∗|
|LV septal thickness, mm||0.30||<.0001 †||0.25||<.0001 ∗|
|PASP, ‡ mm Hg||0.10||.0369||0.11||.0141|
|Pulmonary vascular resistance, § Wood units||0.18||.0006||0.04||.4180|
|LV filling pressures (E/e’ septal-lateral mitral ratio)||0.15||.0035||0.14||.0047|
|TAPSE, mm||0.40||<.0001 †||0.28||<.0001 ∗|
|RV fractional area change, %||0.30||<.0001 †||0.24||<.0001|
|RV basal diameter, mm||0.24||<.0001||0.21||<.0001|
|LV ejection fraction, %||0.20||<.0001||0.13||.0037|
|LV mass, g||0.20||<.0001||0.21||<.0001|
|Pulse pressure, mm Hg||0.09||.0732||0.01||.9008|
|Body mass index, kg/m 2||0.21||.0002||0.19||.0007|