Author + information
- Received July 15, 2004
- Accepted September 29, 2004
- Published online January 18, 2005.
- Mei Wang, MD, PhD,
- Gabriel Yip, MD,
- Cheuk-Man Yu, MD, FRACP, FRCP,
- Qing Zhang, MD,
- Yan Zhang, MD,
- Deko Tse, EN,
- Shun-Ling Kong, RN and
- John E. Sanderson, MD, FRCP, FACC* ()
- ↵*Reprint requests and correspondence:
Dr. John E. Sanderson, 9/F, Clinical Sciences Building, Division of Cardiology, Department of Medicine and Therapeutics, The Chinese University Hong Kong, Prince of Wales Hospital, Hong Kong SAR, China
Objectives This study sought to investigate the incremental prognostic value of non-invasive measures of early myocardial relaxation and left ventricular diastolic pressure (LVDP) in patients with impaired left ventricular (LV) systolic function.
Background The early diastolic mitral annulus velocity (Em) reflects myocardial relaxation, and the combined ratio of the early transmitral flow velocity (E) to Em (E/Em) >15 correlates well with elevated mean LVDP. It is unknown if these new indexes will predict poorer survival in patients with LV systolic dysfunction.
Methods Echocardiograms were prospectively obtained in 182 patients with impaired LV systolic function, defined as an LV ejection fraction <0.50. The end point was cardiac mortality. The majority of this patient sample (80%) has been reported on in a previous publication.
Results After a median 48 months' follow-up, Em emerged as an independent predictor of survival (hazard ratio 0.61, 95% confidence interval 0.45 to 0.82). An Em <3 cm/s was associated with a significantly excess mortality (log-rank statistic 9.36, p = 0.002), and this measurement added incremental prognostic value to standard indexes of systolic or diastolic function, including a deceleration time <140 ms and an E/Em >15 (p = 0.038).
Conclusions Early diastolic mitral annulus velocity is a powerful predictor of cardiac mortality in patients with LV systolic impairment; Em <3 cm/s emerged as the best prognosticator in long-term follow-up, incremental to other clinical or echocardiographic variables, including the ratio E/Em.
In the past decade, various echocardiographic indexes have been used to predict prognosis in patients with left ventricular (LV) dysfunction and chronic heart failure, including LV dimensions, ejection fraction, and Doppler mitral inflow velocities (1–4). In particular, the deceleration time (DT) of early transmitral flow is the simplest echocardiographic diastolic parameter that independently predicts poor prognosis (4–6) and appears to correlate with LV filling pressure in patients with systolic dysfunction (7).
Using tissue Doppler imaging (TDI), myocardial velocities at the mitral annulus have been shown to be useful for quantifying global systolic and diastolic LV function (8). The early diastolic myocardial velocity (Em), in particular, reflects myocardial relaxation and is relatively load-independent (9). In addition, the ratio of early transmitral flow velocity (E) to early diastolic septal mitral annulus velocity (E/Em) correlates well with mean LV diastolic pressure (10) and has recently been shown to be a powerful independent predictor of all-cause mortality after acute myocardial infarction (11). In a previous study, we found that an Em value <3 cm/s provides incremental prognostic value for predicting cardiac mortality in patients with a variety of cardiovascular diseases (12). However, it is uncertain if Em alone has prognostic value in patients with systolic dysfunction or if the ratio E/Em is superior. To test this hypothesis, we evaluated the prognostic value of Em in a large series of unselected patients with impaired LV systolic function and compared this with other established clinical, systolic, and diastolic parameters, including the ratio E/Em.
The study population consisted of 182 consecutive cardiovascular patients who had a clinically indicated transthoracic echocardiogram that showed impaired systolic function with a left ventricular ejection fraction (LVEF) <50% by two-dimensional echocardiography using Simpson's method (range 12.12% to 49.98%, 57% patients with LVEF <40%; 43% with LVEF 40% to 50%). Patients with aortic or mitral stenosis, prosthetic valves, or severe mitral annular calcification were excluded.
Echocardiograms were obtained using General Electric VingMed Sound AB (Horten, Norway) with a 3.5-MHz transducer. Methods of acquiring two-dimensional Doppler, TDI, and their measurements were as previously described (12). All recordings were performed by one investigator (M.W.). Left atrial maximal volume (LAVmax) was traced at the end of systole from apical four- and two-chamber views. Abnormal diastolic function was categorized into abnormal relaxation filling pattern, pseudonormal filling pattern, and restrictive filling pattern according to Garcia et al. (13). We used average value of myocardial velocities measured at septal, lateral, inferior, and anterior aspect as previously described. Peak velocities during systole (Sm), early diastole (Em), and late diastole (Am) were measured.
Death and mode of death were identified from hospital records or telephone interviews with relatives. Cardiac death was defined as death caused by heart disease, including sudden death.
Continuous data are expressed as mean values ± SD, and comparisons between groups were tested by unpaired ttest. Categorical data are presented as absolute values and percentages, and comparisons were tested by Fisher exact test. Estimations of risk were performed using Cox proportional hazard models by univariate and multivariate analysis. Patients who died from other causes or were lost to follow-up were censored.
The incremental value of TDI over clinical data and transmitral inflow variables was assessed in four modeling steps in the same order as in clinical practice. The first step consisted of clinical data used as baseline risk factors. Transmitral inflow variable was then added in the next block. The third block was E/Em >15. In the final step, TDI variables were added. A significant improvement in model prediction was based on the likelihood ratio statistic, which follows a chi-square distribution, and the p value was based on the incremental value compared with the previous model.
Cumulative survival curve was performed by Kaplan-Meier method, and mortality rates were compared using the log-rank test. SPSS version 11.5 (SPSS Inc., Chicago, Illinois) was used for all analyses.
The study population consisted of 124 male (68%) and 58 female (32%) patients with a mean age of 63 ± 15 years. These patients had high prevalence of hypertension, a history of ischemic heart disease and/or diabetes mellitus, and were more likely to present with clinical evidence of LV heart failure (Table 1).Two patients had atrial fibrillation. The echocardiographic characteristics of those patients with Em ≥3 and Em <3 cm/s are shown in Table 2.
Sixty-nine patients (38%) had Em <3. These patients were older and more often had a history of ischemic heart disease than those with Em ≥3. Patients with Em <3 had worse LVEF, lower E, reduced Sm, and higher E/Em. In addition, of those with Em <3, 94% had an E/Em >15. However, there was no difference in LAVmax and the prevalence of moderate or severe mitral regurgitation between the two groups.
Predictors and outcome
All patients were followed for a median of 48 months (range 0.3 to 68 months); 36 patients (20%) died of a cardiac cause. A history of ischemic heart disease, valvular heart disease, or heart failure were the most common baseline diagnoses in those who died (Table 3).In addition, 12 patients died due to other causes such as bacterial septicemia, cancer, or unknown.
The majority of cardiac deaths (22 of 36 [61%]) occurred in patients with Em <3. Figure 1shows the increased mortality associated with Em <3 (log-rank = 9.36, p = 0.002). Similar analyses were performed after stratifying patients on the basis of their DT; Em <3 was a strong predictor of cardiac mortality in patients with a DT >140 ms (n = 132, log-rank = 7.20, p = 0.007). In patients with DT ≤140 ms, the prognostic utility of an Em <3 did not attain statistical significance (n = 44, log-rank = 1.48, p = 0.22), which was likely due to the insufficient numbers of events. However, there was a trend showing more patients with Em <3 died after one-year of follow-up. When patients were stratified on the basis of their E/Em, Em <3 still predicted increased cardiac mortality even in patients with an E/Em ≥15 (n = 141, log-rank = 5.15, p = 0.02). Of 32 patients with an E/Em <15, only one patient died who had an Em >3. Thus, the combination of an E/Em <15 and Em >3 cm/s predicted a more favorable outcome.
The association between the clinical and echocardiographic variables and the end point was assessed by univariate analysis as shown in Table 3. Clinical risk factors such as age, ischemic heart disease, valvular heart disease, and heart failure were associated with an adverse outcome. Systolic mitral annular velocity, Em, Am, E/Em, and DT ≤140 were associated with cardiac death. In multivariate Cox regression, backward stepwise analysis adjusted by age, Sm, Em, Am, E/Em, and DT, Em remained the strongest predictor for cardiac death (0.61, 95% confidence interval: 0.45 to 0.82).
The incremental value of an Em <3 is shown in Figure 2;E/Em >15 improved the prognostic value of a model containing clinical risk factors and DT ≤140 (p = 0.05). However, Em <3 significantly improved the outcome of a mode that contained clinical risk factors, DT ≤140 and E/Em >15 (p = 0.038).
Medication and outcome
At baseline, there was no difference in the use of angiotensin-converting enzyme inhibitors/angiotensin II antagonists (63% vs. 59%, p = 0.62), calcium-channel blockade (21% vs. 27%, p = 0.45), diuretics (57% vs. 52%, p = 0.63), and statin therapy (27% vs. 22%, p = 0.46) in patients with an Em <3 and ≥3, respectively. In contrast, patients with Em <3 tended to receive more aspirin (67% vs. 49%, p = 0.03) and beta-blockers (52% vs. 36%, p = 0.05). In our study group, medication was not predictive of subsequent mortality.
The principal finding of the current study was that Em was an independent powerful predictor of survival in patients with impaired LV systolic function (LVEF <50%). The addition of Em <3 cm/s further improved the prognostic utility of a model that included DT ≤140 ms and E/Em >15.
LV filling pressure and prognosis
Elevated LV filling pressure is associated with increased mortality in dilated cardiomyopathy (3,14) and after acute myocardial infarction (15,16). A short DT correlates well with elevated left atrial pressure, especially in patients with impaired LV systolic function (7,10) and is an independent predictor of adverse outcome in patients with acute myocardial infarction (17). Our study reaffirms the prognostic value of a short DT in patients with subnormal LVEF.
Incremental value of E/Em
However, an elevated E/Em is a more powerful prognostic indicator than transmitral flow variables. Our calculated E/Em appears to be higher than previously reported. In our study, Em was a mean value taken from four different LV annular sites and was measured after post-processing of the color Doppler data that contained mean velocities. This is in contrast with the velocities recorded by pulsed Doppler where peak instantaneous velocities are measured in the sample volume and are usually recorded only at the septal mitral annulus. The off-line values are generally about 15% to 20% lower than those obtained by pulsed Doppler (18) and would explain why the ratio E/Em is higher.
Incremental value of Em
The current study shows that, after a median follow-up of 48 months, Em gradually emerged as the best predictor of cardiac survival with Em <3 cm/s adding incremental prognostic data to clinical and standard echocardiographic parameters. This long-term study further supports our previous work in a more general group of cardiac patients that Em, among other TDI variables, was the strongest predictor of cardiac mortality (12); Em correlates closely with the time constant of isovolumic relaxation (tau) and is relatively insensitive to preload changes in the presence of LV systolic dysfunction (9), and this may explain why Em is a good indicator of function and outcome. Furthermore, abnormalities of diastolic function are often more sensitive indicators of overall ventricular function (19). Therefore, measurements of E/Em and Em should be routinely performed in patients with reduced LV systolic function, not merely for estimating the LV filling pressure and the degree of impairment of myocardial relaxation but, more importantly, for assessing prognosis and risk stratification.
This cohort has a relatively high prevalence of patients suffering from hypertension (51%) with approximately equal prevalence of coronary artery disease (38%) and diabetes mellitus (35%). However, it represents a consecutive series of patients and may simply confirm our previous data that hypertension was the most common associated risk factor for congestive heart failure in Hong Kong (20) and that itself alone or in combination with coronary artery disease was the most common condition predating the onset of heart failure (21).
Our study end point was cardiac mortality including sudden death. It is assumed that this is due to a cardiac arrhythmia, although this cannot be proved conclusively.
The current data demonstrate that Em is a powerful predictor of survival in patients with LV systolic impairment; Em <3 cm/s emerges as the best prognosticator in long-term follow-up, incremental to other clinical or echocardiographic variables, including the E/Em. This easily acquired diastolic index assists in stratifying cardiovascular risk in this group of patients.
- Abbreviations and acronyms
- late diastolic mitral annular velocity
- mitral deceleration time
- early diastolic transmitral flow velocity
- early diastolic mitral annular velocity
- left atrial volume at the end of left ventricular systole
- left ventricle/ventricular
- left ventricular ejection fraction
- systolic mitral annular velocity
- tissue Doppler imaging
- Received July 15, 2004.
- Accepted September 29, 2004.
- American College of Cardiology Foundation
- St. John Sutton M.,
- Pfeffer M.,
- Plappert T.,
- et al.
- Bolognese L.,
- Neskovic A.N.,
- Parodi G.,
- et al.
- Pinamonti B.,
- Zecchin M.,
- Di Lenarda A.,
- Gregori D.,
- Sinagra G.,
- Camerini F.
- Xie G.Y.,
- Berk M.R.,
- Smith M.D.,
- Gurley J.C.,
- DeMaria A.N.
- Giannuzzi P.,
- Temporelli P.L.,
- Bosimini E.,
- et al.
- Nishimura R.A.,
- Appleton C.P.,
- Redfield M.M.,
- Ilstrup D.M.,
- Holmes D.R. Jr.,
- Tajik A.J.
- Nagueh S.F.,
- Middleton K.J.,
- Kopelen H.A.,
- Zoghbi W.A.,
- Quinones M.A.
- Ommen S.R.,
- Nishimura R.A.,
- Appleton C.P.,
- et al.
- Hillis G.S.,
- Moller J.E.,
- Pellikka P.A.,
- et al.
- Wang M.,
- Yip G.W.K.,
- Wang A.Y.M.,
- et al.
- Garcia M.J.,
- Thomas J.D.,
- Klein A.L.
- Temporelli P.L.,
- Giannuzzi P.,
- Nicolosi G.L.,
- et al.
- Yip G.,
- Wang M.,
- Zhang Y.,
- Fung J.W.,
- Ho P.Y.,
- Sanderson J.E.