Usefulness of the Echocardiographic Multi-Parameter Score (EMPS) in Evaluating Left Ventricular Global Heart Function

Abstract

In this study, we established a new index by combining several echocardiographic parameters to quantify heart failure. We selected 233 consecutive patients who underwent both echocardiographic and plasma B-type natriuretic peptide (BNP) tests within 24 hours after referral for suspected heart failure. The echocardiographic parameters included systolic function, diastolic function, left ventricular chamber remodeling, valvular lesions, systolic pulmonary arterial pressure, and regional wall-motion abnormality. Each factor was scored from 1 to 3 points according to its severity. The total point from these 6 factors is the echocardiographic multi-parameter score (EMPS).

The EMPS for 37, 51, 77, and 38 patients from New York Heart Association (NYHA) functional classes I, II, III, and IV, respectively, were 1.24 ± 1.25, 2.98 ± 1.83, 5.96 ± 2.38, and 7.21 ± 1.99, which were significantly different from the mean score of our 30 normal patients (P <0.001). Sensitivity, specificity, and accuracy of an EMPS ≥2 for diagnosis of NYHA classes II to IV were 93%, 83%, and 89%, respectively. The area under the receiver operating characteristic curve was 0.94 (95% confidence interval, 0.92–0.98; P <0.001). There were significant correlations between logBNP and EMPS (r=0.81, P <0.001) or Tei index (r=0.48, P <0.001). In multilinear regression analysis, EMPS, early/late transmitral flow, and peak systolic velocity from tissue Doppler were entered into the model (P <0.001). The standardized regression coefficient (r=0.68) of EMPS was much higher than those of the other 2 factors, which suggests that EMPS is a powerful predictor of BNP levels.

Heart failure is a major and growing public health problem in both developed and developing countries. It is estimated that 5 million people have heart failure in the United States, and this number has increased by an estimated 550,000 per year.¹ Overall, 50% of heart-failure patients will die within 4 years. Forty percent of patients admitted to hospitals with heart failure will be readmitted within 1 year.²

The single most useful diagnostic test for evaluation of patients with heart failure is the comprehensive 2-dimensional transthoracic echocardiogram with color-flow Doppler imaging. It provides us with abundant information in regard to such factors as morphologic and structural changes, systolic or diastolic function, and hemodynamic status. Evidence has shown that the degree of left-chamber remodeling,³ left ventricular (LV) systolic and diastolic dysfunction,⁴ valvular lesions,⁵ and elevated pulmonary artery pressure⁶ (PAP) are all predictors of outcome for heart-failure prognosis. However, no single factor reveals the whole picture of heart failure. Heart-failure patients often have more than 1 structural or functional abnormality on echocardiography.⁷ Putting the multiple echocardiographic results together objectively reveals the severity of heart failure.

In this preliminary study, we created a novel index by combining several echocardiographic measures that are commonly used for the evaluation of heart failure. We have called the index the echocardiographic multi-parameter score (EMPS). We believe that the EMPS will provide new insight into the objective evaluation of LV global heart function.

Patients and Methods

Study Design and Inclusion Criteria

We conducted a retrospective study of 550 patients who had been referred to our cardiology department for suspected heart failure from June 2006 through January 2009. We first selected 233 patients in accordance with these inclusion criteria: 1) disturbance of sinus rhythm, 2) receipt of positive echocardiographic and laboratory (plasma B-type natriuretic peptides [BNP] levels) test results within 24 hours of one another; and 3) echocardiographic images of good quality. Second, in 203 of those 233 patients, we diagnosed heart failure in accordance with the diagnostic criteria of the European Society of Cardiology⁸: 1) symptoms of heart failure (typically breathlessness or fatigue) either at rest or during exertion, or ankle swelling; 2) objective evidence of cardiac dysfunction at rest (such as echocardiographic results); and 3) improved responses to treatment directed towards heart failure. The 30 patients who did not meet these diagnostic criteria were included in our “normal” or control group. We then classified the 203 remaining patients (with heart failure) into New York Heart Association (NYHA) functional classes I, II, III, and IV.

Echocardiographic Evaluation

Comprehensive transthoracic echocardiography was performed using commercially available Hewlett-Packard Sonos 5500® and Sonos 7500® machines (Hewlett-Packard Company; Palo Alto, Calif). Two-dimensional and color-flow Doppler images were obtained in standard parasternal and apical views. Diastolic indices were acquired at 5 consecutive beats using sweep speeds of 50 cm/s. With pulsed-wave Doppler, we acquired transmitral flow images using a 1- to 2-mm sample volume placed at the mitral leaflet tips in the apical 4-chamber view, and we acquired pulmonary venous flow using a 4-mm sample volume placed in the right upper pulmonary vein. The tissue-Doppler image was acquired with standard presets optimized to eliminate background noise and enhance tissue signals; we used a 5- to 10-mm sample volume placed at the lateral and septal mitral annular margins in the 4-chamber view.

All images were stored onto magneto-optical discs and were measured later by another investigator who had more than 7 years of experience and was blinded to the clinical and BNP data. Measurements were averaged over 3 cycles. Left ventricular end-systolic and end-diastolic dimensions, LV wall thickness, and left atrial dimensions were measured in accordance with American Society of Echocardiography recommendations.⁹ For evaluation of LV diastolic function, we measured peak transmitral early (E) and late (A) velocities; early deceleration time (DT); pulmonary vein systolic (S) and diastolic (D) flows; atrial reversal (AR) flows by pulsed-wave Doppler; and peak systolic (Sa) and early (Ea) and late (Aa) diastolic velocities by tissue-Doppler interrogation of the septal and lateral margins. The overall diastolic stages, determined from the pattern of transmitral and pulmonary venous flows, were defined as abnormal relaxation (stage I: transmitral E/A, <1; DT, >220 ms; pulmonary venous D/S, <1; and AR, <35 cm/s), as pseudonormalization (stage II: transmitral E/A, 1–2; DT, 150–220 ms; pulmonary venous D/S, >1; and AR, >35 cm/s), or as restrictive (stage III: transmitral E/A, >2; DT, <150 ms; pulmonary venous D/S, >1; and AR, >35 cm/s).¹⁰

Valvular regurgitation or stenosis was evaluated by incorporating several qualitative variables and was categorized as mildly, moderately, or severely abnormal, in accordance with recently published guidelines and recommendations.^11,12 Systolic PAP was recorded as reported by the reading echocardiographer, who measured the peak tricuspid regurgitant jet velocity and added an estimated right atrial pressure.¹³ The Tei index consists of the ratio of the isovolumic contraction and isovolumic relaxation times to the ejection time—all of which can be obtained by Doppler interrogation.¹⁴ Left ventricular end-diastolic volume, LV end-systolic volume, and LV ejection fraction (LVEF) were measured using the Simpson biplane method.

Description of Scoring System

A scoring system designed by us was used in accordance with the multiple indices mentioned above. The scoring system appears in Table I.

TABLE I. Indices for Echocardiographic Multi-Parameter Score

The score value for LVEF is 1, 2, or 3 points according to severity. Score values for left-chamber remodeling and pulmonary arterial pressure are just 1 point for mild or moderate abnormalities and 2 points for severe abnormalities. Score values for valvular lesions and diastolic function are 1 point for moderate abnormalities and 2 points for severe abnormalities, because we think that mildly abnormal parameters appear in most normal persons.

If 2 parameters of a patient carry different scores in the same category, the higher score should be used in the calculation. For example, if the interventicular septal thickness (IVST) and left ventricular end-diastolic dimension (LVEDD) of a female patient are 13 mm and 63 mm, the score for this category is 2 points. The totals, when scores for all factors are added, constitute the EMPS.

Plasma B-Type Natriuretic Peptide Assays

Venous blood samples were collected within 24 hours of the echocardiography. Plasma BNP was measured using the Triage® BNP Test (Biosite Incorporated; San Diego, Calif), as described elsewhere.¹⁵ The Triage BNP test is a fluorescence immunoassay for quantification of biologically active BNP-32 in whole blood or plasma samples.

Statistical Analysis

All continuous variables are reported as mean ± SD or as percentages. Analysis was done with SPSS 16.0 (SPSS, an IBM company; Chicago, Ill). Because BNP was not distributed normally, the common logarithm of BNP (logBNP) levels was used for correlation and regression analysis, with echocardiographic parameters. One-way analysis of variance was used to compare differences in echocardiographic parameters between NYHA functional class groups. Receiver operating characteristic (ROC) curves were used to evaluate the sensitivity and specificity of EMPS for distinguishing between 1) normal function or NYHA class I heart failure and 2) NYHA class II to IV heart failure. Associations of logBNP with echocardiographic index or EMPS were evaluated by use of the Spearman correlation coefficient. Independent predictors of BNP were determined by means of multiple regression analysis, with candidate variables added to a model containing logBNP as the dependent variable and transmitral E/A and lateral annulus Sa as covariates. Standardized β regression coefficients and their significance as determined by multilinear regression analysis were reported.

Results

Baseline characteristics of the 233 patients included in the study appear in Table II. The eligible patients were 23 to 90 years of age; 150 patients were men and 83 were women. Most subjects were elderly men. The predominant causes of the heart failure were hypertension or hypertensive heart disease (35%), ischemic cardiomyopathy (24%), and dilated cardiomyopathy (18%). Patients who presented with NYHA class I, II, III, and IV numbered 37, 51, 77, and 38, respectively. Another 30 patients were eventually recognized to be normal. Most patients had NYHA class III symptoms.

TABLE II. Baseline Characteristics of the 233 Patients Included in the Study

Echocardiographic findings are summarized in Table II. The pattern of diastolic filling was restrictive in 57 patients and pseudonormal in 34. Mitral or aortic stenosis (or regurgitation) was moderate in 53 patients and severe in 21. Sixty-eight and 21 patients had, respectively, moderate and severe systolic PAP. There were 53 patients with abnormal regional wall motion. The mean EMPS scores for patients from NYHA functional classes I, II, III and IV were, respectively, 1.24 ± 1.25, 2.98 ± 1.83, 5.96 ± 2.38, and 7.21 ± 1.99; these scores were significantly different from those of the normal group (0.27 ± 0.35, P <0.001). The EMPS scores of patients in NYHA functional class III were also different from those in NYHA class IV (P=0.04) (Fig. 1). The ability of EMPS to detect patients with heart failure was evaluated by means of ROC analysis (Fig. 2). The area under the ROC curve was 0.94 (95% confidence interval, 0.92–0.98; P <0.001). An EMPS value ≥2 had a sensitivity of 93%, a specificity of 83%, and an accuracy of 89% for detecting NYHA class II to IV heart failure. Table III shows the univariate and multivariate relationships of EMPS and echocardiographic parameters to logBNP. In Spearman analysis, statistically significant correlations were found between logBNP and EMPS and other echocardiographic parameters. Figure 3 shows that the correlation coefficient of EMPS was 0.81 (P <0.001), greater than that of the Tei index (r = 0.48, P <0.001) and other echocardiographic parameters. In multiple linear regression analysis, EMPS, transmitral flow E/A, and Sa by tissue-Doppler interrogation of the lateral margins were added to the model (P <0.001); no other parameters were added. The standardized regression coefficient (r = 0.68) of EMPS was much higher than those of transmitral flow E/A and lateral Sa (−0.149 and −0.265, respectively), which suggested that EMPS was a powerful independent predictor of logBNP levels.

Fig. 1 Histographic depiction of echocardiographic multi-parameter score (EMPS) in accordance with New York Heart Association (NYHA) functional class. Values are expressed as mean ± SD. Means not sharing a letter (a, b, c, d) are significantly different (P <0.001) between groups, but the difference between NYHA classes III and IV is P =0.04 (1-way analysis of variance with Tukey's post hoc test used for all between-group analysis).

Fig. 2 Receiver operating characteristic (ROC) curve comparing sensitivity and specificity of echocardiographic multi-parameter score diagnosis of New York Heart Association functional class II through IV heart failure. Area under curve is 0.94 (P <0.001).

TABLE III. Univariate and Multivariate Relationships of EMPS and Echocardiographic Indices to logBNP

Fig. 3 Correlation of echocardiographic multi-parameter score (EMPS) and common logarithm of BNP (logBNP) (r = 0.81, P <0.001). * = mean of values

Discussion

Heart failure has a progressive course, which usually starts with normal LV function in the presence of high risk factors, such as hypertension, atherosclerotic disease, or diabetes mellitus; then it develops into LV hypertrophy or dilation with reserved heart function, progresses to LV systolic and diastolic dysfunction, and culminates in end-stage heart failure.¹⁶ Routine measurement of echocardiographic parameters basically reflects this process. Functional parameters provided by echocardiography include 1) LV systolic function parameters, of which LVEF is the hallmark index; and 2) LV diastolic function parameters established via study of Doppler mitral inflow variables, pulmonary venous Doppler flow indices, and tissue-Doppler imaging. The structural and hemodynamic factors include 1) left-heart remodeling, specifically of left atrial and LV dimensions and LV wall thickness, 2) valvular lesions, specifically mitral and aortic regurgitation or stenosis, and 3) PAP. In addition to describing a result as normal or abnormal, echocardiographers quantify the degree of abnormality as mild, moderate, or severe to better understand the condition. Moreover, regional wall-motion abnormalities imply reduced LV function. In consideration of this, we have attempted to measure each echocardiographic parameter to arrive at a score that identifies the presence and the severity of heart failure. The maximum score attainable by EMPS is 12 points.

Our results have shown that the EMPS derived from multiple echocardiographic indices can reflect heart-failure progression. In a normal (control) group, the EMPS rarely exceeds 1 point. With worsening of NYHA functional class, the severity of the change in chamber size and structure not only affects systolic and diastolic performance but indicates regurgitant blood flow through the valves. The rise in EMPS correlates with the worsening of NYHA functional class. The EMPS in NYHA functional class IV patients often exceeds 6 points. An EMPS of 2 points or above is considered to indicate the onset of NYHA class II to IV heart failure. Correlations were statistically significant between logBNP and both EMPS (r = 0.81, P <0.001) and the Tei index (r = 0.48, P <0.001). In multilinear regression analysis, we found that EMPS, transmitral flow E/A, and lateral Sa all entered into the model. Transmitral flow E/A, used to indicate LV diastolic function, is one of the chief components in classifying the severity of diastolic dysfunction. We have taken transmitral flow E/A into consideration in calculating the score for diastolic function, which is part of EMPS. Although septal Sa reflects LV systolic function, its value was not added to the scoring system of EMPS. Nevertheless, correlation analysis showed that EMPS correlated highly both with transmitral flow E/A and with lateral Sa (r was 0.701 and –0.730, respectively; P <0.001). In addition, the standardized β regression coefficient of EMPS was much higher than those of the other 2 factors. These analyses suggested that EMPS was a powerful predictor of logBNP. Therefore, it might be feasible to use this novel index of EMPS to diagnose heart failure and to evaluate its severity.

The severity of heart failure is most often categorized by NYHA functional classification. A more recent classification system has its basis in the structure of the heart and in patients' symptoms.¹⁷ It has been recognized for many years that the NYHA functional classification reflects a subjective evaluation by a healthcare provider and changes frequently over a short period of time.¹⁸ It has also been recognized that heart-failure treatment might not differ greatly among the functional classes. Another study showed that no consistent method of determining NYHA class is in use and that interoperator classification of patients into NYHA functional classes II and III yields a result little better than chance.¹⁹ All of the parameters that we chose for the calculation of EMPS involve objective quantitative or semiquantitative data that overcome the deficiency of subjective assessment. The EMPS consists of data on LV remodeling, valvular lesions, PAP, and abnormal regional wall motion, which provides objective measurement of impaired heart function. This study shows that this objective index, the EMPS—derived as it is from functional, structural, and hemodynamic indicators—correlates well with the progression of NYHA functional class.

Although LVEF (normal, >0.45–0.50) is the most practical means to distinguish patients with systolic dysfunction from those with preserved systolic function, LVEF is not equivalent to contractility because it depends heavily on volume, preload, afterload, heart rate, and valvular function.²⁰ Diastolic heart failure currently accounts for more than 50% of all heart failure and is referred to as heart failure with normal LVEF. Heart failure with normal LVEF is considered the precursor of heart failure with reduced LVEF.²¹ The stratification of cardiac function on the basis of echocardiographic or clinical markers of LV filling pressures is a better predictor of death and hospitalization than is stratification by ventricular systolic function alone.²² Evaluation of heart function by any single criterion is unreliable. The EMPS contains multiple indices, which include reduced systolic and diastolic function (not just one of these, but both). By revealing the combined effect of various abnormalities, EMPS overcomes the weakness of unilateral evaluation of heart function.

Left ventricular remodeling, valvular lesions, and systolic PAP have also been taken into consideration in calculating the EMPS, which helps us to classify that state during which patients have structural abnormality without impaired LV function. At that stage, the EMPS hardly exceeds 2 points. As the heart-failure patient's condition deteriorates, progressive LV dilation or hypertrophy causes mitral regurgitation, and the impairment of right ventricular function is often accompanied by pulmonary hypertension. The EMPS provides an objective indication of the effect of valvular lesions and systolic PAP on heart failure, and increases in the EMPS reflect the deterioration of heart failure.

The plasma concentration of natriuretic peptide is a useful biomarker in the diagnosis of heart failure and in the management of patients with established chronic heart failure.²³ Evidence supports its use for diagnosing, staging, hospitalization, and discharge decisions. In this study, the EMPS showed a positive correlation with logBNP value (r = 0.81, P <0.001). The correlation curve rose steeply when the EMPS ranged from 0 to 5 points, and the curve flattened when the EMPS was over 5 points. Five points, then, might be a cutoff value for the severity of heart failure. The Tei index, compared with the EMPS, showed less of a trend toward positive correlation with logBNP. In multilinear regression analysis, the EMPS was added to the model while the Tei index was excluded, which indicates that the EMPS is more accurate in measuring the severity of heart failure than is the Tei index. It also suggests that EMPS is a powerful independent predictor of logBNP levels.

Limitations of the Study

The study population was small, and the study was retrospective.

In this retrospective study, we retained—in our normal (or control) group—30 patients who met our inclusion criteria but failed to meet the diagnostic criteria for heart failure. Although our application of EMPS to normal patients did support the positive correlation between EMPS and severity of heart failure, we found that such application could lead to the inadvertent conflation of “normal” with the early stages of heart failure (that is, mildly abnormal parameters appear in most normal persons). Therefore, we believe that EMPS is more appropriately applied to patients in advanced heart failure (the equivalent of NYHA III and NYHA IV).

In addition, some factors affecting the EMPS—like LVEF, size of chamber, and thickness of interventricular septum—are continuous data, which may lose statistical efficiency if we treat them as categorical data. The E/Ea, a tissue-Doppler parameter for the evaluation of LV diastolic function, has not been incorporated into the EMPS system. Moreover, we have not determined whether the same scores in different EMPS categories carry equivalent value. Although the right heart is involved in heart failure, enlargement of the right atrium and ventricle occurs before systolic PAP occurs. These factors have not been considered in determining the EMPS.

Conclusion

The EMPS as an index for heart failure has such advantages as these: 1) the EMPS, derived from routine echocardiographic examination, is an echocardiographic index independent of other detection methods; 2) the EMPS is an objective measurement; and 3) the EMPS comprehensively reflects LV global function and the severity of heart failure. It warrants our further research.