Background

Chronic heart failure (CHF) represents the final common pathway for the development and progression of most cardiovascular diseases and the leading cause of death [1, 2]. Currently, the global prevalence of heart failure (HF) is about 2%, with an annual incidence of 1% [3]. In addition, with the global intensification of the aging process such as that in China, the number of HF patients continues to increase [4]. How to accurately diagnose and judge the prognosis in the early stage is the prerequisite and basis for effective treatment of HF patients.

Left ventricular ejection fraction (LVEF) and the left ventricular diastolic filling (LVDF) have long been used clinically to evaluate the LV systolic and diastolic functions respectively [5, 6]. However, LVEF or LVDF alone is not closely related to the clinical condition and cannot accurately reflect the overall cardiac function of HF patients. The HF echocardiography index (HFEI) is a combination of multiple parameters of echocardiography to evaluate the cardiac function in patients with CHF, which overcomes the subjective factors of New York Heart Association Classification and is more accurate and reliable than the single LVEF [7, 8].

The 2016 European Society of Cardiology (ESC) has updated its guidelines for HF. In addition to retaining HF with reduced ejection (HFrEF) and HF with preserved ejection fraction (HFpEF), a new type named HF with mid-range ejection fraction (HFmrEF) has been added [9]. Our study aims to investigate the value of HFEI in evaluating the cardiac function in patients with different types of HF and the clinical predictive value of HFEI for the prognosis of patients with CHF.

Methods

Study population

This study included 489 HF patients who were admitted to the Department of Cardiovascular Medicine of the Affiliated Hospital of Lanzhou University (Lanzhou, China) between January 2016 and December 2018. They included 268 males and 221 females, with a mean age of 64 ± 14 years. Inclusion criteria were as follows: (1) the diagnosis of heart failure is confirmed by two or more experienced cardiologists according to the 2016 ESC guidelines;(2) we can collect comprehensive clinical data; and (3) the one-year follow-up data of the patients were complete. The exclusion criteria were as follows: (1) patients with kidney disease, liver disease or severe liver and kidney dysfunction;(2) tumor, nervous system disease; All patients signed the informed consent forms before enrollment.

Clinical data collection

The demographic characteristics and comorbidities of the patients were obtained on admission. All hematological specimens including blood routine, blood biochemistry or NT-proBNP were collected within 24 h after admission and tested in the central laboratory of the hospital. All clinical data were recorded by one physician and reviewed by another.

Echocardiography

The GE Vivid 7 ultrasound system was used, with a probe frequency of 3.4–5.0 mHz. The measurement was performed by the same professional physician using the means of the results obtained from three consecutive cardiac cycles. Measurement indicators included: (1) the cardiac structure: LVEF, LV end diastolic diameter (LVEDD), interventricular septal thickness (IVST), LV posterior wall thickness (LVPWT), left atrial volume index (LAVi), diameter of left atrium (LAD) calculate relative ventricular wall thickness (RVWT) = 2 × LVPWTd/LVEDD, and LV mass index (LVMi) = LVM/body surface area; (2) cardiac function: peak velocity during early filing(E), late filling from atrial contraction (A), E/A ratio, left ventricular early diastolic flow propagation velocity (FPV), E/FPV, peak systolic mitral annular velocity(S′), E/E’S ratio, pulmonary arterial systolic pressure (PASP), early diastolic mitral flow deceleration time (DT), pulmonary venous flow (S peak systolic wave velocity and D peak diastolic wave velocity), and D/S ratio.

HFEI

According to the related indexes of echocardiography, the total score was HFEI. The HFEI scoring criteria are as follows: (1) 35 ≤ PASP< 50 mmHg, 1 point, PASP≥50 mmHg, 2 point; (2) LV systolic function: 30% ≤ LVEF< 45% or regional ventricular wall motion abnormality, 1 point, LVEF< 30%, 2 point; (3) LV diastolic function: ①E/A < 0.5, DT > 220 ms, D/S < 1 or E/A = 1–2, DT = 150-220 ms, D/S ≥ 1, 1 point; ②E/A > 2, DT < 150 ms, D/S > 1 or restrictive change, 2 point; (4) atrioventricular remodeling: ①56 mm < LVEDD< 66 mm, IVST or LVPWT≥13 mm, LAD≥45 mm, 1 point; ②LVEDD≥66 mm or right ventricular dysfunction; (5) valvular regurgitation or valvular stenosis: moderate, 1 point, serious, 2 point (Table 1) [10, 11].

Table 1 Heart failure echocardiography index score criteria
Full size table

Follow-up evaluation and study end points

All patients were followed up on the outpatient basis, telephone interviews and readmission medical records, with a mean follow-up period of 1 year. End-point events observed in the study were all-cause death, cardiovascular death and hospitalization for HF exacerbation.

Statistical analysis

All data were analyzed using SPSS 21.0 statistical software. Mean ± standard deviation (SD) was used for continuous variables which accord with the normal distribution and the homogeneity test of variance. The three groups of data were compared by the ANOVA test. Categorical variables were expressed as percentage and analyzed by the chi-square test. Univariate and multivariate logistic regression models were used to analyze the correlation between HFEI and the risk of one-year adverse events. The ROC curve was performed to evaluate the diagnostic performance of HFEI and NT-proBNP in patients with poor prognosis. A p value < 0.05 was pre-specified to indicate statistical significance.

Results

Comparison of the baseline characteristics

Four hundred and eighty-nine patients with HF were recruited and divided into three groups, including 170 in HFrEF, 171 in HFmrEF and 148 in HFpEF. There were no significant differences in age, diastolic blood pressure (DBP), blood urea nitrogen (BUN) and α-hydroxybutyrate dehydrogenase (α-HBDH) between the three groups of HF patients, while there were significant differences in HFEI and NT-proBNP between the three groups. The characteristics of the included patients are summarized in Table 2.

Table 2 Baseline characteristics of the study population
Full size table

Comparison of the echocardiographic indexes

Analysis of the echocardiographic parameters in the three groups of patients showed that there were significant differences in the indicators that reflect the ventricular structure including LVEDD, LAD, LVMI and RWT, and in the indicators that reflect the ventricular function including E/FPV and E/E’S. No variances were found in the remaining indicators (Table 3).

Table 3 Echocardiographic indexes of the study population
Full size table

Logistic regression analysis

Multiple logistic regression analysis was performed on variables that may affect the adverse prognosis of HF patients. It was found that NT-proBNP and HFEI were related to the risk of one-year adverse events. However, there was no correlation between LAVi, LVMi, α-HBDH and endpoint events (Table 4).

Table 4 Logistic regression analysis of one-year adverse event risk in patients with heart failure
Full size table

The value of HFEI and NT-proBNP in predicting the one-year risk of adverse events in HF patients

The sensitivity and specificity of NT-proBNP and HFEI to predict the risk of one-year adverse events in patients with HF were analyzed by plotting the ROC curve, and the Jordon index was calculated. When NT-proBNP was 3000 pg/ml, the area under the ROC curve was 0.698(sensitivity 68%, specificity 65%). The greatest area under the ROC curve of HFEI was 0.712, and the optimal cut-off value was 3.5(sensitivity 64%, specificity 78%) (Fig. 1). According to the cutoff value of HFEI and NT-proBNP, HF patients were divided into the following four groups: group A: HFEI<3.5 and NT-proBNP<3000 pg/ml; group B: HFEI<3.5 and NT-proBNP>3000 pg/ml; group C: HFEI>3.5 and NT-proBNP<3000 pg/ml; group D: HFEI>3.5 and NT-proBNP>3000 pg/ml. The one-year risk of adverse events in HF patients between groups was analyzed using Partitions of χ2 method. The incidence of one-year adverse events in group A was lower than that in group C and D. Group D had a higher incidence of one-year adverse events than group D (Table 5).

Fig. 1
figure 1

Receiver operating characteristic (ROC) curves of heart failure echocardiography index (HFEI) and NT-proBNP for predicting adverse events within 1 year in patients with heart failure. The maximum area under the ROC curve of heart failure echocardiography index and NT-proBNP was 0.712 and 0.698, respectively

Full size image
Table 5 Partitions of χ2 method analysis of one-year adverse event risk in patients with heart failure
Full size table

Discussion

This observational study yielded the following results: (1) there were significant differences in HFEI between HFrEF, HFmrEF and HFpEF patients; (2) multivariate regression analysis indicated that HFEI and NT-proBNP were independent risk factors for the prognosis of one-year adverse events in HF patients; and (3) HFEI and NT-proBNP had a good value in predicting the short-term prognosis of HF patients.

Clinically, LVEF and LV filling indicators are generally used to evaluate the cardiac function in HF patients, but single indexes cannot reflect the overall cardiac function, especially in patients with HFpEF and HFrEF, in whom LVEF does not always reflect the severity of HF. HFEI is a comprehensive ultrasound index for evaluating the cardiac function by taking into full account the changes in LV systolic and diastolic function, valve regurgitation and atrioventricular remodeling [12, 13]. In addition, HFEI is not affected by the heart rate and ventricular geometry, and is highly correlated with the LV function indicators measured by cardiac catheterization, which can fully reflect the overall function of the heart. As the the present study has demonstrated significant differences in HFEI between the three types of HF patients, the evaluation of patients could be conducted objectively through the comprehensive echocardiographic parameters.

NT-proBNP is a cardiac neurohormone secreted by the ventricles, playing an important role in the diagnosis and prognosis of HF [14, 15]. However, studies have shown that plasma NT-proBNP level in patients with chronic kidney disease and cor pulmonale are also significantly increased, so NT-proBNP is not the specific diagnostic criterion for HF [16]. Several studies have shown a correlation between the multiple echocardiographic evaluation indicators (E/A and LVEF) included in HFEI and plasma NT-proBNP levels [8, 17]. Chen et al. suggested that HFEI could judge the degree of HF in HFmrEF patients and had a good correlation with NT-proBNP [18]. In this study, all the HF patients were followed up for 1 year to record the occurrence of adverse events and evaluate the prognosis. The result of multivariate logistic regression analysis also showed a significant correlation between NT-proBNP and short-term prognosis of HF patients. We also compared the value of HFEI and NT-proBNP in assessing short-term outcomes in HF patients. By plotting the ROC curve, it was found that the area under the curve for HFEI and NT-proBNP differed by 0.712 and 0.698, respectively. Therefore, NT-proBNP alone may not be sufficient to predict the risk of one-year adverse events in HF patients. Combination of a high plasma level of NT-proBNP and HFEI may improve the accuracy of prognostic prediction.

There are some shortcomings in this study. On the one hand, the study population was from a single center with a smaller sample size. On the other hand, the follow-up period was relatively short. Therefore, more studies with larger sample sizes and longer follow-up time are required to fully demonstrate the value of HFEI in the diagnosis and prognostic prediction of HF patients.

Conclusion

In summary, the preliminary data obtained in this study suggest that HFEI could objectively predict the prognosis of patients with HF. HFEI and plasma NT-proBNP level are well correlated with the risk of adverse events, and combination of the two can improve the accuracy of prediction.