Left ventricular (LV) diastolic dysfunction carries a substantial risk for the subsequent development of heart failure and reduced survival, even when it is asymptomatic. Plasma brain natriuretic peptide (BNP) level and tissue Doppler imaging indexes provide powerful incremental assessment of LV diastolic function. Accordingly, the aim of this study was to clarify whether these methodologies could identify LV diastolic dysfunction without heart failure in 280 patients with preserved LV ejection fractions (≥50%) who underwent echocardiography and cardiac catheterization for the evaluation of coronary artery disease. Patients were classified into 2 groups, those with diastolic dysfunction (τ ≥48 ms; n = 91) and those with normal diastolic function (τ <48 ms; n = 189). Plasma BNP ≥22.4 pg/ml, an unexpectedly low value, had sensitivity of 74.7% and specificity of 60.8% for identifying isolated LV diastolic dysfunction; the combined use of BNP ≥22.4 pg/mL and mitral annular velocity during early diastole <7.4 cm/s had relatively low sensitivity of 44.0% but high specificity of 86.8%. In conclusion, using plasma BNP level and with the combination of BNP level and mitral annular velocity during early diastole, invasively proved isolated LV diastolic dysfunction without heart failure could be identified in patients with coronary artery disease.
Left ventricular (LV) diastolic dysfunction carries a substantial risk for the subsequent development of heart failure and reduced survival, even when it is asymptomatic. Thus, the early identification of LV diastolic dysfunction in patients without overt symptoms may provide an opportunity to manage the underlying cause and prevent progression to diastolic heart failure. Quantitative and precise elucidation of LV diastolic function requires cardiac catheterization of the left ventricle, but it is invasive and potentially complex. In contrast, elevated plasma brain natriuretic peptide (BNP) level and reduced mitral annular velocity during early diastole (Ea) are noninvasive parameters of LV diastolic dysfunction. Accordingly, we investigated whether LV diastolic dysfunction without apparent symptoms derived from heart failure in patients with preserved LV systolic function could be identified using these parameters.
Methods
The study patients were enrolled from among 411 consecutive patients who underwent comprehensive echocardiography including tissue Doppler imaging and diagnostic cardiac catheterization for the evaluation of coronary artery disease and from whom blood samples for BNP measurement were taken at cardiac catheterization. All patients had symptoms suggestive of exercise-induced angina pectoris and/or clinical signs of coronary artery disease, including positive exercise electrocardiographic changes, abnormal myocardial perfusion scintigraphic findings, and histories of myocardial infarction or coronary revascularization. No patient had apparent symptoms or signs of heart failure as shown in the Framingham criteria for the diagnosis of heart failure. A total of 280 patients with LV ejection fractions ≥50% by left ventriculography were eligible for enrollment in this study. Patients with renal insufficiency (serum creatinine ≥2.0 mg/dl), atrial fibrillation or flutter, artificial pacemakers, hemodynamically significant valvular disease, postprosthetic valve replacement conditions, idiopathic dilated or hypertrophic cardiomyopathy, or acute coronary syndromes were excluded. According to the findings of coronary angiography and left ventriculography, 125 patients had previous myocardial infarctions, 93 patients had angina pectoris without LV wall motion abnormalities, and 62 patients had atypical chest pain without significant coronary stenosis. All patients gave written informed consent for participation in the study. The study protocol was performed according to the regulations proposed by the ethical guidelines committee of Nagoya City University Graduate School of Medical Sciences.
Each patient underwent comprehensive transthoracic echocardiography (Aplio80; Toshiba Medical Company, Tokyo, Japan). Tissue Doppler imaging was performed with the sample volume placed on both the septal and lateral corners of the mitral annulus in the apical 4-chamber view. From the spectral traces, annular velocity Ea was obtained as a mean value of velocities obtained at both sides.
LV pressure waves were obtained with a catheter-tipped micromanometer (SPC-454D; Millar Instruments, Inc., Houston, Texas) and recorded on a polygraph system (RMC-3000; Nihon Kohden, Inc., Tokyo, Japan) and on a digital data recorder (NR-2000; Keyence, Osaka, Japan), as we have reported elsewhere. From the recorded pressure waves, a time constant, τ, of decrease in LV pressure was computed by applying a monoexponential fitting with zero asymptote to LV pressure decay. LV end-diastolic pressure was also determined. LV end-systolic and end-diastolic volumes were obtained from biplane left ventriculography using the method proposed by Chapman et al. These volumes in each patient were corrected by each body surface area and were expressed as LV end-systolic and end-diastolic volume indexes.
Patients were divided into 2 groups on the basis of their LV early diastolic function, that is, whether they had time constant τ values of LV relaxation ≥48 or <48 ms. This threshold value of 48 ms to distinguish LV early diastolic dysfunction was derived from the report by the European Study Group on Diastolic Heart Failure.
Venous blood samples (6 ml) for the assay of plasma BNP concentration were collected from the right femoral veins of all patients at cardiac catheterization. Blood samples were centrifuged and then stored at −70°C. Plasma BNP concentrations were measured with an immunoradiometric assay specific for human BNP (Shionoria; Shionogi Company, Ltd., Osaka, Japan). With this method, the minimal detectable quantity of human BNP is 2 pg/ml.
SPSS version 17.0 (SPSS, Inc., Chicago, Illinois) was used for statistical analysis. BNP levels are summarized as medians and interquartile ranges. Other data are presented as mean ± SD or frequency (percentage). Parameters were compared between the 2 groups using unpaired Student’s t tests. Differences in prevalence between 2 groups were also compared using the chi-square test. Relations between 2 parameters were evaluated by univariate linear regression analysis. Because BNP levels were not normally distributed, a logarithmic transformation was applied. The ability of BNP and tissue Doppler imaging indexes to identify LV diastolic dysfunction was evaluated using receiver-operating characteristic curve analysis; areas under the curve and 95% confidence intervals are indicated. Differences with p values <0.05 were considered statistically significant.
Results
Table 1 lists the baseline characteristics of the patients. No significant difference was found in age, heart rate, body mass index, or mean blood pressure between patients with isolated LV diastolic dysfunction and those with normal LV diastolic function. LV ejection fractions were significantly higher in patients with normal LV diastolic function than in those with isolated LV diastolic dysfunction. LV end-systolic volume index, end-diastolic volume index, and end-diastolic pressure were significantly greater in patients with isolated LV diastolic dysfunction than in the group with normal LV diastolic function. Slight LV remodeling was observed in patients with isolated LV diastolic dysfunction, even if they had preserved LV systolic function.
| Characteristic | Normal Diastolic Function (τ <48 ms) | Isolated Diastolic Dysfunction (τ ≥48 ms) | p Value |
|---|---|---|---|
| (n = 189) | (n = 91) | ||
| Men/women | 131/58 | 74/17 | 0.02 |
| Age (years) | 66.3 ± 8.5 | 67.4 ± 8.2 | 0.30 |
| Body mass index (kg/m 2 ) | 24.1 ± 3.3 | 24.8 ± 3.0 | 0.14 |
| Heart rate (beats/min) | 65.0 ± 9.7 | 63.1 ± 9.5 | 0.14 |
| Mean blood pressure (mm Hg) | 95.3 ± 11.0 | 94.4 ± 11.4 | 0.53 |
| τ (ms) | 39.6 ± 5.2 | 54.0 ± 6.1 | <0.001 |
| LV ejection fraction (%) | 68.8 ± 8.5 | 64.8 ± 8.6 | <0.001 |
| LV end-diastolic volume index (ml/m 2 ) | 76.8 ± 15.7 | 86.0 ± 17.9 | <0.001 |
| LV end-systolic volume index (ml/m 2 ) | 24.2 ± 9.4 | 30.5 ± 11.2 | <0.001 |
| LV end-diastolic pressure (mm Hg) | 12.3 ± 4.2 | 17.1 ± 5.1 | <0.001 |
| Hypertension | 41.9% | 42.2% | 0.50 |
| Diabetes mellitus | 29.8% | 28.9% | 0.76 |
Figure 1 shows a comparison of BNP level and Ea between the groups. Plasma BNP level was significantly higher in patients with isolated LV diastolic dysfunction than in those with normal diastolic function (median 39.2 pg/ml, interquartile range 21.2 to 75.5, vs median 16.9 pg/ml, interquartile range 6.9 to 32.8; p <0.001). Ea was significantly less (7.1 ± 2.1 vs 8.6 ± 2.6 cm/s, p <0.001) in patients with isolated LV diastolic dysfunction than in those with normal diastolic function. Logarithmically transformed plasma BNP level was weakly but significantly correlated with the time constant τ of LV relaxation (r = 0.35, p <0.001), and Ea was also weakly but significantly correlated with τ (r = −0.23, p <0.001) in this patient cohort.
Receiver-operating characteristic analysis indicated that plasma BNP level ( Figure 2 ) and Ea had significant potential for identifying isolated LV diastolic dysfunction. The areas under the curves for BNP level and Ea to detect the patients with τ ≥48 ms were 0.71 (95% confidence interval 0.64 to 0.78, p <0.001) and 0.67 (95% confidence interval 0.60 to 0.74, p <0.001), respectively. From these receiver-operating characteristic curves, a BNP level of 22.4 pg/ml and an Ea of 7.4 cm/s were determined as the cut-off values for identifying isolated LV diastolic dysfunction most precisely using each parameter. A BNP level ≥22.4 pg/ml was most sensitive, and the combination of BNP ≥22.4 pg/ml and Ea <7.4 cm/s was most specific for identifying isolated LV diastolic dysfunction ( Table 2 ).
