Fibrosis-4 index reflects right ventricular function and prognosis in heart failure with preserved ejection fraction

Study patients

From February 2012 to December 2015, 165 patients admitted to the National Hospital Organization Iwakuni Clinical Center for ADHF and diagnosed as HFpEF were prospectively enrolled. A primary diagnosis of ADHF included signs and symptoms of fluid overload. Patients were eligible for inclusion in the trial if they had a left ventricular ejection fraction (LVEF) of 50% or more by echocardiography performed on admission. Patients who died before discharge (n = 3), with chronic liver diseases (hepatitis B, n = 4; hepatitis C, n = 12; primary biliary cirrhosis, n = 1; alcoholic liver disease, n = 3; non-alcoholic fatty liver disease, n = 1, history of hospitalization due to liver disorder of unknown cause, n = 1), and who underwent planned thoracic surgery and/or percutaneous coronary intervention within a half year after the onset of ADHF (n = 24) were excluded. Chronic liver diseases were diagnosed by a hepatologist along with blood examination, pre-existing liver disease, and history of treatment. Therefore, 49 patients were excluded from the study, and the remaining 116 patients were analysed. The investigation conforms to the principles outlined in the Declaration of Helsinki. This study was approved by our institution's human research committee, and all patients enrolled in the study provided written informed consent.

Fibrosis-4 index

Data of blood examinations on admission and before discharge were used. The FIB-4 index was calculated as follows: FIB-4 index = age (years) × AST (U/L)/[ALT (U/L)^1/2 × platelet count (10⁹/L)]. In addition, total bilirubin and serum albumin to evaluate liver function, haemoglobin, creatinine, HgbA1c, and NT-pro-BNP were measured because these parameters are thought to affect cardiovascular prognosis independently.

Echocardiography

A full echocardiographic assessment of cardiac functions was performed by experienced technicians using an IE33 echo machine (Philips Japan, Ltd., Tokyo, Japan) before discharge. The technicians who performed echocardiography were blinded to the FIB-4 index. From the apical approach, we measured the tricuspid annular plane systolic excursion (TAPSE) and tricuspid lateral annular systolic velocity (S′) to assess RV function. TAPSE was measured by the distance of the systolic excursion of the RV annulus along its longitudinal plane using M-mode presentation in an RV-focused apical 4-chamber view. S′ was measured by the velocity of tricuspid lateral annular using Doppler tissue imaging in an RV-focused apical 4-chamber view. The LVEF was measured by the modified Simpson technique using B-mode presentation in apical-2-chamber view and apical-4-chamber view. We also measured the peak early diastolic velocities (E), and the early diastolic myocardial velocities (e′) using general methods. The ratio of E and e′ (E/e′) was calculated to estimate LV filling pressures. The inter-observer variability of LVEF, TAPSE, and S′ was evaluated for 21 patients with heart failure independent of this study.

Endpoints and follow-up

All clinical events after discharge were obtained by medical records retrospectively until November 2016. The endpoint was defined as MACE, which is the composite outcome of death from cardiovascular causes, readmission for heart failure, nonfatal non-ST elevation/ST elevation, myocardial infarction, and ischaemic/non-ischaemic stroke.¹²

Statistical analysis

Categorical variables are shown as numbers (%). Continuous variables that were normally distributed are shown as the means ± standard deviation and were compared between groups using the Student's t-test. Continuous variables that were not normally distributed are shown as medians with interquartile ranges (IQR) and were compared using the Mann–Whitney U-test. The association of TAPSE and S′ with clinical variables including the FIB-4 index was examined using univariate and multivariate linear regression analyses. Continuous variables that were not normally distributed were turned into natural logarithmic values for linear regression analysis. In multivariate regression analysis, components of the FIB-4 index including age, AST, ALT, and platelet count were not included. Pearson linear regression analysis and Bland–Altman plots were performed to evaluate the inter-observer variability of the LVEF, TAPSE, and S′. Prognostic factors including the FIB-4 index, sex, serum creatinine, and haemoglobin were evaluated by multivariate Cox regression analysis. We evaluated the association with NT-pro-BNP instead of the FIB-4 index using the same model. To assess MACE, receiver operating characteristic (ROC) curve analysis was performed. The optimal cut-off value was defined as the point maximizing the Youden index (=max[sensitivity + specificity − 1]). To assess the prognostic value, the optimal cut-off values of the FIB-4 index were used to divide patients into two groups for Kaplan–Meier analysis, with event-free survival compared using a two-sided log-rank test. Statistically significant values were defined when P < 0.05. These analyses were performed using SPSS statistical software (version 25; IBM Corp., Armonk, NY, USA).

Patient characteristics and laboratory data

Table 1 shows the clinical characteristics of all patients. Study patients included 50 men (43%) with a mean age of 79 years. The median duration of hospitalization was 19 days (12.0–30.5 days), and the median follow-up period was 736 days (328–1103 days). Table 2 shows laboratory data from the day of admission and before discharge. The FIB-4 index before discharge was significantly lower than at admission (2.62 [1.92–3.46] and 3.03 [2.05–4.67], P < 0.001). AST, ALT, total bilirubin, and NT-proBNP before discharge were also lower than at admission (P < 0.001, P = 0.002, P < 0.001, and P < 0.001). Haemoglobin, platelet count, albumin, and serum creatinine did not differ between admission and discharge.

Association of the fibrosis-4 index with tricuspid annular plane systolic excursion and S′

LVEF, TAPSE, S′, and E/e′ before discharge were 62.7 (55.9–68.6) %, 17.5 ± 4.6 mm, 10.0 (8.0–12.0) cm/s, and 13.6 (10.0–17.7), respectively. In terms of inter-observer variability, the measurements of LVEF, TAPSE, and S′ correlated well between the two technicians (r = 0.965 [P < 0.001], r = 0.899 [P < 0.001], and r = 0.826 [P < 0.001], respectively). The Bland–Altman analyses for LVEF, TAPSE, and S′ are shown in Figure 1(A–C).

Table 3 shows the univariate and multivariate analyses of factors related to TAPSE and S′. TAPSE was significantly related to age (P = 0.003), systolic blood pressure (P = 0.004), platelet count (P = 0.011), and FIB-4 index (P = 0.001) by univariate analysis. Multivariate regression analysis revealed that TAPSE was significantly related to the FIB-4 index (P = 0.014). In addition, S′ was also significantly related to systolic blood pressure (P = 0.004), platelet count (P < 0.001), and FIB-4 index (P < 0.001) by univariate analysis. Multivariate regression analysis revealed that S′ was significantly related to the FIB-4 index (P = 0.009).

Relationship between the fibrosis-4 index and major adverse cardiovascular events

During the median follow-up of 736 days, 37 MACE occurred (two cardiovascular deaths, 30 readmissions for heart failure, and five strokes). As shown in Table 4, multivariate Cox regression analysis showed that the FIB-4 index on admission was significantly associated with the incidence of MACE (hazard ratio [HR] 1.124, 95% confidence interval [CI] 1.008–1.253, P = 0.035) after adjustment for male, serum creatinine, and haemoglobin. In addition, as shown in Table 5, the FIB-4 index before discharge was significantly associated with the incidence of MACE (HR 1.270, 95% CI 1.052–1.532, P = 0.013) after adjustment for male, serum creatinine, and haemoglobin. However, when NT-proBNP before discharge was used instead of FIB-4 in the same model, there was no significant association between NT-proBNP and the incidence of MACE (HR 1.000, 95% CI 0.999–1.000, P = 0.458).

The ROC curve analysis showed that C-statics of the FIB-4 index before discharge to predict MACE were 0.573 (95% CI 0.462–0.685, P = 0.204), and the value maximizing the Youden index was 3.11. Figure 2 shows the Kaplan–Meier analysis of patients stratified according to the optimal cut-off value of the FIB-4 index before discharge. Patients with a higher FIB-4 index before discharge had a significantly poorer prognosis than patients with a lower FIB-4 index before discharge (P = 0.029). Patients with an FIB-4 index >3.11 had a 2.202-fold (95% CI 1.110–4.368) increased risk of MACE after adjustment for male, serum creatinine, and haemoglobin.

Association of fibrosis-4 index and liver fibrosis

Liver biopsy remains the reference standard to assess the severity of liver disease. Currently, several non-invasive methods including serum biomarkers and imaging techniques are under investigation for the quantification of liver fibrosis. Among these methods, the FIB-4 index has been used as a surrogate marker of liver stiffness.⁷ A recent comparative study including chronic liver diseases such as non-alcoholic fatty liver disease and chronic hepatitis C demonstrated that the diagnostic performance of gadoxetate-enhanced MRI, transient elastography, and FIB-4 index to predict significant fibrosis was sufficiently accurate (area under the ROC: 0.78, 0.80, and 0.67, respectively).¹³ Thus, the FIB-4 index might be a marker of liver fibrosis in chronic liver disease.

Correlation of fibrosis-4 index with echocardiographic parameters of right ventricular function

The FIB-4 index is considered a simple biomarker of liver fibrosis in patients with liver diseases. However, the liver becomes stiff because of long-term hepatic congestion and accompanying fibrosis in patients with HF. Furthermore, liver stiffness measured by transient elastography increases as the central venous blood pressure increases.¹⁴ Pulmonary hypertension and right-sided heart failure occur more often in HFpEF patients than in heart failure with reduced ejection fraction (HFrEF) patients. RV dysfunction exacerbates right-sided heart failure and TAPSE and S′ are indexes of RV contraction in the longitudinal plane. The American Society of Echocardiography recommends using TAPSE and S′ to assess RV function because they show a correlation with other measures of RV global systolic function.¹⁵ Another study demonstrated that liver stiffness, measured by transient elastography, increased with a decrease in TAPSE.⁶ In this study, we demonstrated that a high FIB-4 index was associated with lower TAPSE and S′. Therefore, the FIB-4 index can be regarded as a biomarker for RV dysfunction in patients with HFpEF.

Prognostic impact of fibrosis-4 index in heart failure with preserved ejection fraction

The prognosis of patients with HF deteriorates as the RV function deteriorates and right heart failure progresses. Several studies indicated that RV dysfunction is an independent prognostic factor for HFpEF patients.^16-18 Furthermore, a high FIB-4 index was associated with an increase in all-cause mortality in patients hospitalized for ADHF.¹¹ In this study, we focused on patients with HFpEF and ADHF and also found that a high FIB-4 index was associated with an increase in the incidence of cardiovascular events. Our results indicate that the FIB-4 index is both a biomarker of RV function and a predictor of cardiovascular events in HFpEF patients. Thus, the FIB-4 index may provide important information on pathological changes in the liver in addition to RV dysfunction. There is a pathological link between the heart and the liver.¹ HF causes liver congestion and hypoperfusion, and if it continues for a long time, liver fibrosis can occur, which reduces natriuresis and enhances fluid retention¹⁹ and arterial sclerosis, which impairs left ventricular diastolic function.^{20, 21}

The cut-off value of the FIB-4 index to predict the prognosis of patients with HF has not been fully clarified. Sato et al. reported that in patients with decompensated HF, all-cause mortality in the second tertile (1.72 ≤ FIB-4 index < 3.01) and third tertile (3.01 ≤ FIB-4 index) groups was significantly greater than that in the first tertile (FIB-4 index < 1.72) group.¹¹ Daichi et al. reported that an FIB-4 index of 2.15 was a cut-off value to predict the composite clinical outcome of all-cause death, readmission for HF, or left ventricular assist device implantation in patients with HF.¹⁰ This study showed the optimal cut-off value was 3.11 to predict MACE in patients with HFpEF. Larger prospective studies are needed to evaluate the most appropriate cut-off value of the FIB-4 index as a risk predictor of cardiovascular prognosis in HFpEF patients.

Interestingly, NT-pro-BNP was not an independent predictor of MACE in our patients with HFpEF, and there was no significant difference in NT-pro-BNP between the high and low FIB-4 index groups. These observations contradict those of previous studies.^{22, 23} Other studies showed that an increase in BNP or NT-pro-BNP was associated with a worse clinical outcome in patients with HFpEF, but that the level of BNP or NT-pro-BNP was significantly lower than that in patients with HFrEF.^{24, 25} Another study indicated that BNP or NT-pro-BNP was not likely to be increased in right-sided heart failure alone.²⁶ Therefore, BNP or NT-pro-BNP may be insufficient to predict the prognosis of patients with HFpEF and a history of hospitalized for ADHF. In this situation, measurement of the FIB-4 index in addition to NT-pro-BNP may add independent and valuable information regarding future MACE in patients with HFpEF.

This study had several limitations. First, the study had a relatively small sample size and was a prospective cohort study at a single centre, which may be limited and underpowered. Second, chronic liver diseases were not completely ruled out because we did not perform additional examinations such as liver biopsy or computed tomography scans for evaluation. Third, we excluded patients who underwent thoracic surgery and/or percutaneous coronary intervention within half a year of hospitalization for ADHF, because these interventions might influence MACE. Consequently, some ischaemic cardiomyopathies or valvar diseases were excluded. Finally, we evaluated the FIB-4 index before discharge in this study; therefore, changes in the FIB-4 index during hospitalization or after discharge remain unclear.

The FIB-4 index, a marker of liver fibrosis, is associated with RV dysfunction and a high risk of a future cardiovascular event in patients with HFpEF.