Author + information
- Received January 8, 2013
- Revision received April 12, 2013
- Accepted April 15, 2013
- Published online July 1, 2013.
- Masaru Obokata, MD∗,
- Kazuaki Negishi, MD, PhD∗∗ (, )
- Koji Kurosawa, MD, PhD∗,
- Hitomi Arima, RMS†,
- Rieko Tateno, MD∗,
- Goro Ui, MD†,
- Shoichi Tange, MD†,
- Masashi Arai, MD, PhD∗ and
- Masahiko Kurabayashi, MD, PhD∗
- ∗Department of Medicine and Biological Science, Gunma University Graduate School of Medicine, Maebashi, Japan
- †Department of Cardiovascular Medicine, Maebashi Red Cross Hospital, Maebashi, Japan
- ↵∗Reprint requests and correspondence to:
Dr. Kazuaki Negishi, Department of Medicine and Biological Science, Gunma University Graduate School of Medicine, 3-39-22 Showa-machi, Maebashi, Gunma 371-8511, Japan.
Objectives The purposes of this study were to examine left atrial (LA) functional reserve in patients with heart failure (HF) with preserved ejection fraction (HFpEF) and to determine whether LA strain has an incremental diagnostic value over clinical and conventional echocardiographic parameters.
Background Patients with HFpEF have multiple cardiovascular reserve abnormalities. Although the LA is dysfunctional in HFpEF, the diagnostic value of LA strain remains unknown.
Methods The LA at rest and during passive leg lifts was echocardiographically assessed in 40 patients with HFpEF and in 46 patients with hypertension without HF (HT controls). Global peak atrial longitudinal strain during ventricular systole (global LAS) and booster strain during atrial contraction (global LAB) were assessed using speckle tracking.
Results Patients with HFpEF had an enlarged LA and reduced LA emptying fraction compared with HT controls at rest, while LA stroke volume (SV) was similar between the groups. During leg lifts, increases in LA reservoir and contractile function (i.e., global LAS and LAB) were blunted in HFpEF patients compared with HT controls, resulting in impaired LASV responses. Global LAS and LAB during leg lifts accurately differentiated HFpEF from HT controls (areas under the curve: 0.95 and 0.92, respectively). Resting global LAS had a significant incremental diagnostic value over clinical (age and sex) and conventional echocardiographic parameters (E/E′ ratio, left ventricular mass index, and maximum LA volume index) (global chi-square: 49.6 vs. 30.8; p < 0.0001). The diagnostic value was further improved by adding global LAS during leg lifts (global chi-square: 72.2 vs. 49.6; p < 0.0001).
Conclusions An enlarged LA compensates for LA dysfunction and maintains LASV at rest in patients with HFpEF. However, depressed LA reserve affects LA performance during leg lifts. Evaluation of LA function, including LA strain using leg lifts, might provide incremental diagnostic value for HFpEF.
Understanding about the pathophysiological mechanisms of heart failure (HF) with preserved ejection fraction (HFpEF) has recently improved. This condition is not simply the result of a single impairment in left ventricular (LV) diastolic function, but rather an accumulation of multiple domains of cardiovascular dysfunction (1). One of the most distinct features of HFpEF is increased LV and vascular stiffness, which contributes to pathophysiology and cardiovascular reserve dysfunction in such patients.
Left atrial (LA) function in close interdependence with LV function plays an important role in maintaining optimal cardiac performance by modulating LV filling. Several investigations have found that as diastolic dysfunction progresses, the relative contribution of LA function to LV filling increases at early stage and later declines, at least partly because of increasing LA stiffness (2,3). Those findings imply that a decrease in LA compliance plays a key role in the pathophysiology of HFpEF. However, the mechanisms through which LA stiffness contributes to the development of HFpEF remain unknown. We hypothesized that LA functional reserve would be impaired in patients with HFpEF.
Speckle tracking echocardiography enables a precise assessment of the function of the LA, including its stiffness, and the feasibility and reproducibility of speckle tracking have been established (4). The purposes of this study were to examine LA functional reserve in patients with HFpEF and to determine whether LA strain has an incremental diagnostic value over clinical and conventional echocardiographic parameters.
We studied 48 consecutive patients who were diagnosed with HFpEF based on the guidelines of the European Society of Cardiology (5) and referred to our hospital for pulmonary edema due to HF. Patients with acute coronary syndrome, atrial fibrillation, more than mild valvular heart diseases, cardiomyopathies, pulmonary disease and inadequate images were excluded. We included 48 consecutive individuals with hypertension (HT) but without a history of HF (HT controls) during the same period. All patients were assessed by echocardiography while in a compensated state at discharge. The institutional medical ethics committee of Maebashi Red Cross Hospital approved the study protocol, and all participants provided written informed consent.
An experienced sonographer performed all echocardiographic studies using a portable device (Vivid i, GE Medical, Milwaukee, Wisconsin) equipped with a 2.5-MHz transducer. All patients underwent echocardiography in a supine or slightly left decubitus position.
Both LV and LA volumes were quantified according to the recommendations of the American Society of Echocardiography (6). LA volumes were calculated using the biapical area-length method and subsequently indexed by body surface area (7) to measure maximum, pre-A, and minimum LA volume indices (LAVImax, LAVIpre-A, and LAVImin). Total, passive, and active LA stroke volume (SV) and LA emptying fraction were assessed as previously described (4,8).
Transmitral flow (E-wave, A-wave, and deceleration time [DecT]) and mitral annular tissue-Doppler (E′, A′, and S′) velocities were measured, and the E/E′ ratio was used to estimate pulmonary capillary wedge pressure (PCWP).
Echocardiography during leg lifts
Measurements were repeated during passive leg lifts and care was taken to obtain the same imaging planes of the LA as in the baseline. An assistant carefully held the patient's legs at 60° to maintain the position of the patient's thorax. During leg lifts, the sonographer focused solely on capturing the same imaging planes as those in the baseline images. After 3 min of passive leg lifts, data were acquired during leg raises (9). LA lengths at rest and during leg lifts were 52 ± 6 mm and 53 ± 6 mm, respectively (ρ = 0.84; p < 0.001; difference = 1.3 ± 3.4 mm).
Speckle tracking echocardiography
Three consecutive cardiac cycles were recorded and averaged. The frame rate was set between 60 and 80 frames/s. Data were analyzed off-line using a customized software package (EchoPAC, GE, Milwaukee, Wisconsin). We used the onset of the R-wave as the reference point to calculate LA strain (10). The LA endocardial border was manually traced in 4- and 2-chamber views. After manual adjustment of the region of interest to cover the full thickness of the myocardium, the software divided it into 6 segments and automatically scored segmental tracking quality. Segments in which image quality was inadequate were rejected by the software and excluded from analysis. Longitudinal strain curves were generated for each of 12 atrial segments in 4- and 2-chamber views. Because longitudinal strain curves were lower in 2 segments of the LA roof than in the other 4 segments, they were excluded from the 4- and 2-chamber views (Fig. 1) (11). Global peak atrial longitudinal strain during ventricular systole (global LAS) and booster strain during atrial contraction (global LAB) were then measured by averaging values obtained in the other eight LA segments. Because the initial strain was set at the ventricular end-diastole, these strains are presented as positive values in this method (10).
The LV longitudinal strain was assessed using speckle tracking. The LV endocardium in end-diastole was traced in 4-, 2-, and 3-chamber views, and the thickness of region of interest was adjusted to include the entire myocardium. The software automatically tracks myocardial motion and results were displayed. LV global strain was defined as the average of values for longitudinal global strain obtained in 4-, 2-, and 3-chamber images. We also measured LV basal strain by averaging the strain values of four LV bases (anterior, lateral, septal, and inferior). The same examiner performed all echocardiographic measurements.
Estimation of LA stiffness
Continuous variables are shown as mean ± SD unless otherwise specified. Categorical variables were analyzed using Fisher's exact test. Normality was evaluated by the Shapiro-Wilk W test. Comparisons between groups were assessed using Student's t test and the Mann-Whitney U test for normally and non-normally distributed data, respectively. Normally and non-normally intragroup distributed data were compared using a paired t test and the Wilcoxon signed rank test, respectively. Baseline group differences were adjusted using general linear models. The diagnostic ability of LA strain to distinguish HFpEF from HT controls was determined using receiver-operating characteristic curves. The area under the curve (AUC) for each parameter was compared using paired analyses (13) with Bonferroni correction. The optimal cutoff was taken when the sum of sensitivity and specificity was the highest. Multivariate models were created to assess the independent diagnostic value of global LAS adjusted for clinical data and echocardiographic findings. The incremental value was assessed by comparing the global chi-square values for each model. The reproducibility of LA strain was assessed in 20 randomly selected individuals. Intraobserver and interobserver variability were evaluated by having the same observer and another experienced reader repeat the analysis and they are reported as interclass correlation coefficients (ICCs). A value of p < 0.05 was considered to indicate statistical significance. We compared receiver-operating characteristic curves using MedCalc version 12.1.4 (MedCalc Software, Mariakerke, Belgium), and all other data were statistically analyzed using SPSS version 19.0 (SPSS Inc., Chicago, Illinois).
Baseline clinical characteristics
Of the 96 enrolled subjects, 8 patients with HFpEF and 2 HT controls were excluded due to poor-quality echocardiographic findings (n = 5), atrial fibrillation (n = 3), and moderate mitral valvular disease (n = 2). Therefore, data were analyzed from 40 patients with HFpEF and 46 HT controls. Table 1 shows the baseline clinical characteristics. Age, sex, body mass index, and systolic blood pressure were similar between groups. Chronic kidney disease was more prevalent among the patients with HFpEF, but the prevalences of HT, diabetes mellitus, coronary artery disease, dyslipidemia, and current smoking were similar between the groups. The brain natriuretic peptide concentration was significantly higher and hemoglobin level was lower in patients with HFpEF, who were also more likely to be receiving loop diuretics. Other medications did not significantly differ between groups.
Resting LV structure and function
The LV mass index (LVMI) and septal wall thickness were significantly higher in patients with HFpEF than in HT controls (Table 1). At rest, LV end-diastolic volume, LV end-systolic volume, LVEF, and LVSV were similar between groups (Table 2). However, patients with HFpEF had significantly lower S′, LV global, and basal strain values than did the HT controls. These results indicated that patients with HFpEF had pathological LV hypertrophy with depressed longitudinal systolic function despite having a preserved EF.
Resting E/A ratio, DecT, and isovolumic relaxation time did not differ between the groups. However, patients with HFpEF had significantly lower E′ and A′, and higher E/E′ ratio, than did HT controls.
LA volumes, function, strain, and stiffness
The total and active LA emptying fractions were significantly lower in patients with HFpEF at rest, whereas the passive fraction was not (Table 3). On the other hand, LA volume indices, such as LAVImax, LAVIpre-A, and LAVImin, were significantly higher in patients with HFpEF than in HT controls. Thus, LASV values were similar between the HFpEF patients and HT controls. These results indicated that LA enlargement compensated for the reduced LA emptying fraction to maintain LASV. Global LAS and LAB strain were significantly lower in patients with HFpEF, whereas LA stiffness was significantly higher in patients with HFpEF. Overall differences in resting LV/LA volumes/mechanics remained significant after adjustment for estimated glomerular filtration rate.
Effects of leg lifts on LV function
During leg lifts, the increase in LVSV was blunted in patients with HFpEF (Fig. 2A). Reflecting this, increased preload significantly raised E-wave, which resulted in elevated E/A ratio in those patients. In contrast, although E′ was significantly increased with leg lifts in HT controls, it remained unchanged in the patients with HFpEF (Table 2). The baseline E/E′ ratio remained significantly higher in the patients with HFpEF than in the HT controls during leg lifts. These results suggest that leg lifts caused a more restrictive filling pattern in the patients with HFpEF.
LA reserve responses during leg lifts
The increase in LAVmax during leg lifts was lower in the patients than in HT controls (Fig. 2B), and the increase in global LAS was blunted in the patients with HFpEF (Fig. 2C). These findings suggest that the patients with HFpEF had depressed reserve capacity for LA reservoir function. The increase in LA contractility assessed by A′ and global LAB was lower in the patients with HFpEF than in the HT controls during leg lifts (Figs. 2D and 2E). Consequently, augmentation in total LASV was significantly reduced in patients with HFpEF (Fig. 2F). All differences in LV and LA reserve remained significant after adjustment for estimated glomerular filtration rate.
Discrimination of patients with HFpEF from HT controls
Both global LAS and LAB at rest could distinguish patients with HFpEF from HT controls, with AUCs of 0.88 and 0.81, respectively (Fig. 3). Importantly, the diagnostic accuracy was significantly higher for both global LAS and LAB during leg lifts than for resting global LAS and LAB, respectively (AUC: 0.95 and 0.92, respectively; Bonferroni-adjusted p < 0.05). Global LAS and global LAB of <31.2% and <15.2%, respectively, during leg lifts discriminated HFpEF from HT controls, with sensitivities of 90% and 89%, respectively, and specificities of 83% and 94%.
Independent and incremental diagnostic value of LA strain over clinical characteristics and conventional echocardiographic measurements
Multivariate analysis adjusted for age, sex, conventional echocardiographic values (E/E′ ratio, LVMI, and LAVImax), and log-transformed brain natriuretic peptide revealed global LAS during leg lifts as an independent discriminator of HFpEF (odds ratio: 0.77; 95% CI: 0.66 to 0.89). Furthermore, adding resting global LAS improved the model based on clinical data and conventional echocardiographic values (Fig. 4). The diagnostic value was further improved by adding global LAS during leg lifts.
Intraobserver and interobserver variability
The ICC values for intraobserver variability for global LAS and LAB determined at rest were 0.94 and 0.95, respectively, and during leg lifts were 0.96 and 0.95. The ICC values for interobserver variability at rest were 0.96 and 0.95, respectively, and during leg lifts were 0.94 and 0.84.
We comprehensively assessed LA strain at rest and during passive leg lifts in patients with HFpEF and in HT controls. We found that a larger LA volume in patients at rest was not associated with reductions in LASV and LVSV despite increased LA stiffness and a reduced LA emptying fraction. Increases in LA reservoir and contractile function during leg lifts were blunted in the patients with HFpEF, resulting in a depressed LASV response. Global LAS during leg lifts distinguished patients with HFpEF from HT controls with excellent diagnostic accuracy and provided independent and incremental diagnostic value over clinical and conventional echocardiographic parameters.
Compensatory LA dilation in HFpEF
We found that patients with HFpEF had a dilated LA and reduced LA strain at rest, which was consistent with previous findings (2,3). Dilation of the LA is likely to be an adaptive change in HFpEF to compensate for the increase in LV filling pressure that occurs with increasing severity of diastolic dysfunction (14). With the increase in LV filling pressure, LA dilation works as a compensatory mechanism via the LA Frank-Starling mechanism (8). Chronically elevated LV filling pressure and comorbidities such as HT, diabetes mellitus, chronic kidney disease, and atrial fibrillation promote further LA dilation, as well as the hypertrophy and interstitial fibrosis that characterize LA remodeling (15). Consequently, these structural changes might decrease LA compliance, reservoir, and contractile function (16). Our finding that an increase in LA volume was not associated with the reductions in LASV and LVSV in resting patients with HFpEF suggests that the enlarged LA helps to preserve LV filling despite LA dysfunction.
Reduced LA reserve in HFpEF
The LA reserve function was blunted in response to increased preload in patients with HFpEF. Recent studies have focused on the importance of abnormalities in multiple cardiovascular reserve function in HFpEF (1). Our findings are consistent with those from a previous study showing a blunted A′ response in patients with HFpEF during handgrip exercise (3). Because a stiffer LA has reduced reservoir function in patients with HFpEF, increased preload from the pulmonary vein cannot enlarge the volume of the LA, resulting in a limited ability to use the LA Frank-Starling mechanism. Consequently, augmentation during LV filling might be inadequate.
Value of LA strain assessment by passive leg lifts in HFpEF
This study showed that global LAS during leg lifts can be an incremental diagnostic parameter for HFpEF. Whereas invasive measurements of LV pressure by cardiac catheterization remain the gold standard for a diagnosis of HFpEF and should be considered in patients with unexplained HF symptoms (17), noninvasive evaluation using echocardiography plays an important role in routine clinical practice. The American Society of Echocardiography and the European Association of Echocardiography recently released recommendations for the classification of diastolic dysfunction, which incorporate several parameters reflecting various LV diastolic properties (18). However, Unzek et al. (19) showed that remaining ambiguity was sufficient to lead to suboptimal interobserver agreement regarding diagnoses. Thus, new markers of diastolic dysfunction are required to improve the accuracy of HFpEF diagnosis.
Although global LAS and LAB at rest have excellent discrimination between HFpEF and HT controls (2), diagnostic accuracy of these parameters obtained during leg lifts was significantly improved. Recent studies have demonstrated that passive leg lifts can detect concealed abnormalities in LV diastolic dysfunction (17,20). Because patients with HFpEF have abnormal cardiovascular reserve during exercise stress (1), evaluating HFpEF during stress might provide important information. We found that disparities in LA reservoir and contractile function (that is, global LAS and LAB) between HFpEF and HT controls became more pronounced during leg lifts and that the 2 conditions could be clearly discriminated.
We also demonstrated that global LAS during leg lifts had independent and incremental diagnostic value over clinical and conventional echocardiographic data that comprise the diagnostic algorithms of the European Society of Cardiology (5).
Although HFpEF is mostly a disease of the elderly (21), and exercise testing should be considered when values determined at rest are unremarkable (17), maximal exercise is sometimes unfeasible in such patients. However, diagnostic stress testing might still be required. All such patients could perform passive leg lifts regardless of health status, and additional cost is not imposed. Values obtained in this manner added incremental information to routine bedside echocardiographic findings. From this viewpoint, assessing LA strain during leg lifts might be used as a diagnostic test for HFpEF, especially among elderly patients.
Pressures were not directly measured by invasive cardiac catheterization but were estimated from noninvasive echocardiographic surrogates. In this study, PCWP was estimated using E/E′ ratio, which may be less accurate in normal individuals with normal LV relaxation (22). However, PCWP can be estimated in patients with cardiac disease using the E/E′ ratio (18). We believe that this method was valid for our control subjects who had hypertension with impaired LV relaxation, and the E/E′ ratio has proven reliable in HFpEF (23). Although care was taken to acquire the same image planes as in the baseline, views might have differed somewhat during leg lifts. However, such differences could similarly occur in both groups, and variable changes in atrial volume between groups were unlikely to be due to different views. Pulmonary HT is often associated with HFpEF, and estimated pulmonary artery systolic pressure using Doppler echocardiography might provide noninvasive diagnostic information about HFpEF (17). However, because patients with HFpEF have a higher prevalence of chronic obstructive pulmonary disease (24), pulmonary artery pressure could be overestimated in this population. Patients with paroxysmal atrial fibrillation might not be completely excluded due to its paroxysmal nature. However, we believe that this is very rare because all patients with HF were carefully monitored by continuous ECG during at least 71.0% of hospitalization (94.6 ± 7.3% of hospital stay). Finally, a larger prospective study is warranted as our patient cohort was relatively small.
An enlarged LA in HFpEF patients compensated for LA dysfunction and maintains LASV at rest. However, depressed LA reserve affected LA performance during leg lifts. Evaluation of LA function including LA strain using passive leg lifts might provide incremental diagnostic value for HFpEF.
The authors are very grateful to Michiko Imai, Norihiro Kobayashi, Yohei Ono, and Shinya Kogure for excellent technical assistance, and to Tatsuhiro Saito for advice regarding the statistical analysis.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- deceleration time
- heart failure with preserved ejection fraction
- peak atrial longitudinal booster strain during atrial contraction
- peak atrial longitudinal strain during ventricular systole
- left atrial volume index
- left ventricular mass index
- pulmonary capillary wedge pressure
- stroke volume
- Received January 8, 2013.
- Revision received April 12, 2013.
- Accepted April 15, 2013.
- American College of Cardiology Foundation
- Borlaug B.A.,
- Olson T.P.,
- Lam C.S.,
- et al.
- Kurt M.,
- Wang J.,
- Torre-Amione G.,
- Nagueh S.F.
- Melenovsky V.,
- Borlaug B.A.,
- Rosen B.,
- et al.
- Paulus W.J.,
- Tschöpe C.,
- Sanderson J.E.,
- et al.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Anwar A.M.,
- Geleijnse M.L.,
- Soliman O.I.,
- Nemes A.,
- ten Cate F.J.
- Pozzoli M.,
- Traversi E.,
- Cioffi G.,
- Stenner R.,
- Sanarico M.,
- Tavazzi L.
- Hoit B.D.,
- Walsh R.A.
- Daccarett M.,
- Badger T.J.,
- Akoum N.,
- et al.
- Borlaug B.A.,
- Nishimura R.A.,
- Sorajja P.,
- Lam C.S.,
- Redfield M.M.
- Firstenberg M.S.,
- Levine B.D.,
- Garcia M.J.,
- et al.
- Ather S.,
- Chan W.,
- Bozkurt B.,
- et al.