Author + information
- Received November 28, 2008
- Revision received April 6, 2009
- Accepted April 14, 2009
- Published online August 11, 2009.
- Goo-Yeong Cho, MD, PhD⁎,⁎ (, )
- Thomas H. Marwick, MD, PhD†,
- Hyun-Sook Kim, MD, PhD‡,
- Min-Kyu Kim, MD‡,
- Kyung-Soon Hong, MD, PhD‡ and
- Dong-Jin Oh, MD, PhD‡
- ↵⁎Reprint requests and correspondence:
Dr. Goo-Yeong Cho, Department of Medicine, Seoul National University Bundang Hospital, 300 Gumi-dong, Bundang-gu, Seongnam City, Gyeonggi-do, 463-802, South Korea
Objectives We sought to evaluate whether global 2-dimensional (2D) strain offers additional benefit over left ventricular ejection fraction (LVEF) to predict clinical events in heart failure.
Background Although 2D strain based on speckle tracking has been proposed as a simple and reproducible tool to detect systolic dysfunction, the relationship of 2D strain and prognosis has not been studied.
Methods Two hundred one patients (age 63 ± 11 years, 34% female, LVEF 34 ± 13%) hospitalized for acute heart failure underwent clinical evaluation and conventional and tissue Doppler echocardiography. Using dedicated software, we measured the global longitudinal strain (GLS) in apical 4- and 2-chamber views and the global circumferential strain (GCS) in a parasternal short-axis view. Cardiac events were defined as readmission for heart failure or cardiac death.
Results There were 23.4% clinical events during 39 ± 17 months of follow-up. In univariate analysis, age, left atrial volume, left ventricular volume, LVEF, ratio of early transmitral flow to early diastolic annular velocity (E/e′), and both GLS and GCS were predictive of cardiac events. In multivariate Cox models, age (hazard ratio [HR]: 1.06, 95% confidence interval [CI]: 1.01 to 1.10, p = 0.017) and GCS (HR: 1.15, 95% CI: 1.04 to 1.28; p = 0.006) were independently associated with cardiac events. By Cox proportional hazards model, the addition of GCS markedly improved the prognostic utility of a model containing ejection fraction, E/e′, and GLS.
Conclusions GCS is a powerful predictor of cardiac events and appears to be a better parameter than ejection fraction in patients with acute heart failure.
Left ventricular ejection fraction (LVEF) is the most extensively investigated echocardiographic systolic function and has been established as a powerful predictor of mortality in patients with systolic heart failure (1,2). However, some reports have characterized the relationship between LVEF and mortality in heart failure patients with inconsistent results (3–5). Furthermore, assessment of left ventricular (LV) systolic function using an echocardiogram is often rather subjective, especially when the endocardial border cannot be clearly defined. LV systolic function is a complex, coordinated action involving longitudinal contraction, circumferential shortening, and radial thickening. Contraction of muscle fibers in the mid-wall, which is linearly related to circumferential strain (6), may better reflect intrinsic contractility than contraction of fibers in the endocardium (7).
The recently developed 2-dimensional (2D) strain based on speckle tracking is an innovative method providing multidimensional myocardial mechanics that involves rotation and longitudinal and circumferential motion. In an experimental study, both longitudinal and circumferential strain had good correlation and agreement with sonomicrometry and had real potential for quantification of LV deformation (8).
We hypothesized that global longitudinal strain (GLS) and global circumferential strain (GCS) may potentially be stronger predictors of clinical events than LVEF in patients with congestive heart failure.
This was a single-center prospective observational study. Between June 2000 and September 2004, we prospectively recruited patients who were admitted to a cardiology department with heart failure. Of these, we narrowed our study population to include only patients who survived acute decompensated heart failure in New York Heart Association functional class ≥III. Inclusion and exclusion criteria are presented in Figure 1.
Patients received standard management as recommended for heart failure, with regard to beta-blockers, angiotensin-converting enzyme inhibitors, and diuretics, including spironolactone.
The ischemic etiology of heart failure was defined by 1 of the following criteria: 1) significant epicardial coronary artery stenosis (>50%); or 2) history of myocardial infarction or coronary revascularization. Normal myocardial perfusion on radionucleotide imaging was regarded to be of nonischemic etiology. The study protocol was approved by the Institutional Review Board of Hallym University.
All images were obtained with a standard ultrasound machine (System 5 or Vivid 7, GE Vingmed, Horten, Norway) with a 2.5-MHz probe. Standard techniques were used to obtain M-mode, 2D, and Doppler measurement in accordance with American Society of Echocardiography guidelines. Tissue Doppler-derived peak systolic (s′), early (e′), and late diastolic (a′) velocities were derived from the septal mitral annulus. LV end-systolic and end-diastolic volumes along with the LVEF were calculated by the biplane Simpson's method from apical 4- and 2-chamber views. The percentage of LV fractional shortening was also calculated as follows: ([LV end-diastolic dimension – end systolic dimension]/LV end diastolic dimension) × 100. Mitral regurgitation was characterized as follows: mild (regurgitation orifice area ≤0.2 cm2), moderate (0.2 to 0.39 cm2), and severe (≥0.4 cm2).
For global 2D strain analysis, a digital loop was acquired from a parasternal short-axis view at the mid-papillary level, apical 4-chamber, and apical-2 chamber views. All images were stored digitally on a magnetic optical disc and analyzed off-line. We traced along the LV endocardial border at the end-systolic frame. The strain curve was extracted from the gray-scale images by using dedicated software (EchoPac, GE Vingmed). Peak strain was defined as the peak negative value on the strain curve during the entire cardiac cycle. Peak GLS and GCS were calculated for the entire U-shaped (GLS) and circular-shaped (GCS) LV myocardium as: global strain (%) = (L[end-systole] – L[end-diastole])/L(end-diastole) × 100 (9), where the global strain is the change of the whole myocardium, not an averaged value of each segmental strain, and L is the whole LV myocardium as one big segment. GLS was averaged from the apical 4- and 2-chamber views. All echocardiographic data were obtained at the time of discharge.
Follow-up and end points
All patients were seen within 4 weeks post-discharge and every 2 months thereafter for the duration of the follow-up period. No patient was lost to follow-up. The cardiac events were defined as readmission for worsening of heart failure and cardiac death. In case of death, the reasons for death were verified from hospital records or a death certificate from the Korea National Statistical Office. The cause of death was identified in all patients.
Data are expressed as mean ± SD for continuous variables and as proportions for categorical variables. The comparison of continuous variables was performed by an independent sample ttest. For categorical variables, the chi-square test was used. A Cox proportional hazards model was used for multivariate analysis (forward stepwise) to investigate which prognostic factors identified using univariate analysis were significantly associated with cardiac events. Receiver-operating characteristic (ROC) curve analysis was used to determine optimal cutoff values of continuous variable. The best cutoff value was defined as the point with the highest sum of sensitivity and specificity. The overall event-free survival rates were calculated using the Kaplan-Meier analysis, and the event rates were compared using the log-rank test. The 2-tailed p values of <0.05 were considered statistically significant. Statistical analysis was performed using SPSS version 13.0 (SPSS Inc., Chicago, Illinois).
A total of 201 patients fulfilled all inclusion and exclusion criteria. During the mean follow-up duration of 39 ± 17 months, cardiac events occurred in 23.4% (47 of 201 patients), including cardiac death in 13%. The clinical and echocardiographic parameters are summarized in Table 1.As compared with patients who did not develop cardiac events, patients who developed cardiac events were older, and had a greater left atrium and LV volume, a higher E/e′, a lower LVEF, and a lower GLS and GCS, but were similar with respect to the etiology of heart failure, LV mass index, and transmitral flow.
Predictors of cardiac events
By univariate survival using a Cox proportional hazards model, age, left atrial volume, LV volume, LVFS, LVEF, systolic and early diastolic mitral annular velocities, E/e′, and both GLS and GCS were significantly associated with cardiac events. The hazard ratio (HR) of each variable is shown in Table 2.Using multivariate Cox analysis, age (HR: 1.09; p = 0.012) and GCS (HR: 1.15; p = 0.007) were independent predictors of cardiac events in patients with heart failure. Importantly, reduced GCS was related to an increased risk of cardiac events in patients with heart failure of both an ischemic and nonischemic origin, but GLS was associated with patients with heart failure of an ischemic origin only (Fig. 2).
The additional benefit of global strain to predict cardiac events is shown in Figure 3.With regard to the incremental value, GLS offers a small additional benefit over conventional parameters (LVEF and E/e′). However, the addition of GCS markedly improved the prognostic utility of the model containing LVEF, E/e′, and GLS.
ROC curves to predict cardiac events
By analyzing the ROC curve, the area under the ROC curve of GCS was the greatest, with a best cutoff point of −10.7% (Fig. 4).However, the areas under the ROC curve between each variable are not statistically different.
The mean event-free survival evaluated by Kaplan-Meier analysis was longer in patients with GCS ≤−10.7% (59.0 ± 2.0 months) than in those with GCS >−10.7% (47.1 ± 3.4 months) (Fig. 5).
Feasibility and reproducibility
We were able to measure ejection fraction (EF) by biplane Simpson's rule in 198 patients (98.5%). We obtained global strain only in the case of adequate tracking quality ≥5 of the 6 segments per view. Of the 201 patients, 88% of mean longitudinal strain and 92% of circumferential strain could be measured. The reproducibility was performed on a single set of recordings. Variability in the measurement of global strain was evaluated in 20 randomly selected patients. For intraobserver variability, the same observer measured global strain for each of the selected patients again 15 days later. The coefficients of variation of intraobserver variability for GLS and GCS were 3.5% and 4.9%, respectively. For the interobserver variability, a second independent observer repeated the analysis. The coefficients of variation of interobserver variability for GLS and GCS were 3.6% and 6.3%, respectively.
This study is the first to demonstrate the prognostic significance of global 2D strain in patients hospitalized with acute heart failure. GCS is an independent prognosticator to predict cardiac events in heart failure regardless age, LVEF, and E/e′ and has greater prognostic power than GLS or LVEF. Furthermore, the measurement of global 2D strain is simple and highly feasible and shows excellent reproducibility.
Endocardial versus mid-wall function
Although LVEF is an accepted prognostic indicator in heart failure, the relationship between EF and adverse outcome in the full spectrum of heart failure is less well documented. The prognosis in patients with most forms of heart failure is inversely related to LVEF, at least for an LVEF <45% (1,2,10). However, recent evidence suggests that the survival of patients with heart failure with preserved LV systolic function was similar to that of patients with a reduced EF (11).
LV systolic function is a complex, coordinated action by longitudinal and circumferential fiber contraction. The subepicardial and subendocardial fiber layers are longitudinally oriented, with a significant contribution to long-axis function, whereas the middle layer is circumferentially arranged and contributes to thickening and short-axis function. The circumferential fibers in the mid-wall play an important role in extensive endocardial thickening (12). Nonuniformity of fiber orientation was thus matched by a nonuniform susceptibility of the various layers to injury. Therefore, the measurement of both longitudinal and circumferential contraction is of great help to understand LV contractile function. Before the development of 2D strain, only magnetic resonance tagging had been used to compute circumferential strain. Mid-wall shortening can now be directly measured by 2D strain and might conceivably identify patients at risk for a poor outcome. The correlation between circumferential strain with mid-wall shortening is significant (6). In this group of patients, the association between circumferential strain and the cardiac events was present even after adjusting for the clinical factors most associated with prognosis: age, etiology of heart failure, chamber dimension, ejection fraction, and diastolic function.
In a prospective epidemiologic study, LV long-axis function rather than short-axis function independently predicted survival in chronic heart failure (13). This result is in contrast to the findings reported herein. In that study, short-axis function was assessed by endocardial fractional shortening. However, conventional parameters, such as LVEF or endocardial fractional shortening, reflect the geometric change of LV rather than the contractile function of the myocardium. Because LV geometry and wall thickening are not always uniform, mid-wall function has been proposed for evaluation of systolic function in conditions with altered LV geometry, especially LV hypertrophy (14). Furthermore, mid-wall shortening serves as an independent prediction of adverse outcome in arterial hypertension (15).
Ischemic versus nonischemic origin
In this study, decreased circumferential strain rather than longitudinal strain was associated with increased cardiac events of both ischemic and nonischemic origin.
Because subendocardial fibers, which are mainly longitudinally oriented, are more susceptible to ischemia, it might be expected that the longitudinal function is altered earlier than the mid-wall function (8). Experimental studies have shown that ischemia-induced myocardial necrosis is characterized by extension from the endocardial layer to the subepicardium (16). Therefore, in early ischemic conditions, subendocardial dysfunction is likely to selectively affect the longitudinally directed fibers and manifest itself as decreased LVEF and longitudinal strain, whereas mid-wall function is relatively preserved. In contrast, transmural infarct is associated with a reduction of both long-axis and short-axis function (17).
In patients with coronary artery stenosis, circumferential myocardial shortening deteriorated to a greater extent during stress echocardiography among 3-directional deformations. The level of contribution of circumferential fiber to LV myocardial thickening is greater than that of longitudinal fibers (18). Therefore, a decrease in circumferential strain, which reflects more advanced myocardial ischemia, is more closely related with prognosis than longitudinal strain in ischemic heart failure.
Although there are no published reports of the prognostic implication of circumferential strain in nonischemic heart failure, the major abnormality in idiopathic dilated cardiomyopathy is the marked reduction in fiber shortening (19). In nonischemic dilated cardiomyopathy, approximately 30% of patients have mid-wall fibrosis as determined by cardiovascular magnetic resonance, which is a predictor of adverse outcome regardless of LVEF (20,21). This could explain why decreased GCS was associated with an increased risk of cardiac events.
Although circumferential shortening currently cannot be considered part of routine clinical practice, this new technology provides important prognostic information in heart failure. Because estimation of global strain is reasonable, even for images with poor tracking quality in 1 segment of the 6, global strain has a theoretical advantage over other echocardiographic-based methods (9).
The importance of this study seems limited because 2D strain can be measured using only dedicated software. The feasibility of GLS and GSC was only 88% and 92%, respectively. The 2D strain based on speckle tracking is highly dependent on image quality and the inherent limitation therein. We used 2 kinds of echocardiographs (Vivid 5 and 7). Although the 2D strain from Vivid 5 was less validated, it should work if the frame rate and images are adequate. For longitudinal strain, we analyzed data from the apical 4- and 2-chamber views. Additional analysis of the apical long-axis view provides optimal information for longitudinal strain. Although the relation between GCS and cardiac events remained significant after correction for classic risk factors, we did not obtain important factors in heart failure, such as brain natriuretic peptide, degree of pulmonary hypertension, or right ventricular function. Furthermore, the number of hard events was relatively low in identifying various predictors of heart failure. Finally, our conclusions apply to patients who were hospitalized for newly developed acute heart failure and should not be extrapolated to all heart failure patients.
In our single-center study, we demonstrated that GCS is an independent predictor of cardiac events and appears to be a better parameter than LVEF or GLS for prognostic stratification in patients with acute heart failure.
The authors greatly thank O. K. Kim and M. J. Kum for their excellent contribution of 2D strain measurements.
Dr. Marwick has had research collaborations/grants with GE Medical Systems and Philips.
- Abbreviations and Acronyms
- confidence interval
- ratio of early transmitral flow to early diastolic annular velocity
- ejection fraction
- global circumferential strain
- global longitudinal strain
- hazard ratio
- left ventricle/ventricular
- left ventricular ejection fraction
- receiver-operating characteristic
- Received November 28, 2008.
- Revision received April 6, 2009.
- Accepted April 14, 2009.
- American College of Cardiology Foundation
- St John S.M.,
- Pfeffer M.A.,
- Moye L.,
- et al.
- Curtis J.P.,
- Sokol S.I.,
- Wang Y.,
- et al.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Reant P.,
- Labrousse L.,
- Lafitte S.,
- et al.
- Gustafsson F.,
- Torp-Pedersen C.,
- Brendorp B.,
- Seibaek M.,
- Burchardt H.,
- Kober L.
- Rademakers F.E.,
- Rogers W.J.,
- Guier W.H.,
- et al.
- Svealv B.G.,
- Olofsson E.L.,
- Andersson B.
- de Simone G.,
- Devereux R.B.,
- Koren M.J.,
- Mensah G.A.,
- Casale P.N.,
- Laragh J.H.
- Reimer K.A.,
- Lowe J.E.,
- Rasmussen M.M.,
- Jennings R.B.
- Chan J.,
- Hanekom L.,
- Wong C.,
- Leano R.,
- Cho G.Y.,
- Marwick T.H.
- Tanaka H.,
- Oishi Y.,
- Mizuguchi Y.,
- et al.
- MacGowan G.A.,
- Shapiro E.P.,
- Azhari H.,
- et al.
- Assomull R.G.,
- Prasad S.K.,
- Lyne J.,
- et al.
- McCrohon J.A.,
- Moon J.C.,
- Prasad S.K.,
- et al.