Clinical characteristics
Totally, 441 subjects were enrolled and divided into roughly equal three groups, tertile 1 (FR ≥ 3.71), tertile 2 (3.15 ≤ FR < 3.71) and tertile 3 (FR < 3.15), from highest to lowest levels. The baseline data of the subjects are shown in Table 1. Of all the 441 subjects, 228 (51.7%) were not meet the diagnosis of systolic HF, 330 (74.8%) were men, 118 (26.8%) had atrial fibrillation (AF), 217 (49.2%) had coronary disease, 259 (58.7%) had hypertension, 43 (9.8%) had dyslipidemia, 125 (28.3%) had diabetes mellitus (DM) and 32 (7.3%) had chronic kidney disease. With the FR tertiles decreased, the incidence of AF and NYHA function class III or IV, the left atrial and LV size, E/e’ ratio, log NT-proBNP, serum phenylalanine and tyrosine levels were gradually elevated while the LVEF and BCAA (leucine, isoleucine and valine) level were gradually declined (P ≤ 0.001). The echocardiography and laboratory examination results showed that subjects in the lowest tertile of the FR index tend to have worse heart function.
Table 1
Baseline Characteristics of all subjects according to the tertile of FR
|
FR≥3.71
|
3.15≤FR<3.71
|
FR<3.15
|
P value
|
|
(n=143)
|
(n=150)
|
(n=148)
|
|
Demographic and clinical features
|
Men, n(%)
|
121(84.6)
|
98(65.3)
|
111(75.0)
|
<0.001
|
Age (years)
|
60.85±11.36
|
66.09±10.47
|
64.61±11.55
|
<0.001
|
BMI(kg/m2)
|
25.00±3.58
|
24.63±2.61
|
24.94±4.23
|
0.634
|
Systolic BP (mmHg)
|
128.34±20.77
|
131.57±21.74
|
127.41±22.67
|
0.225
|
Diastolic BP(mmHg)
|
76.65±15.36
|
73.91±14.84
|
76.16±15.35
|
0.256
|
Heart rate (bpm)
|
80.80±12.56
|
79.02±15.61
|
80.35±15.82
|
0.559
|
NYHA functional class
|
1.66±0.86
|
1.72±0.93
|
2.16±1.02
|
<0.001
|
NYHA II
|
34(23.8)
|
31(20.7)
|
35(23.6)
|
0.770
|
NYHA III
|
24(16.8)
|
28(18.7)
|
47(31.8)
|
0.004
|
NYHA IV
|
4(2.8)
|
7(4.7)
|
14(9.5)
|
0.039
|
Medical history
|
Smoking, n(%)
|
65(45.5)
|
51(34.0)
|
30(20.3)
|
<0.001
|
Drinking, n(%)
|
36(25.2)
|
33(22.0)
|
21(14.2)
|
0.056
|
AF, n(%)
|
23(16.1)
|
43(28.7)
|
52(35.1)
|
0.001
|
Coronary disease, n(%)
|
88(61.5)
|
68(45.3)
|
61(41.2)
|
0.001
|
Hypertension, n(%)
|
103(72.0)
|
88(58.7)
|
68(45.9)
|
<0.001
|
Hyperlipidemia, n(%)
|
13(9.1)
|
16(10.7)
|
14(9.5)
|
0.892
|
DM, n(%)
|
55(38.5)
|
37(24.7)
|
33(22.3)
|
0.004
|
Chronic kidney disease,n(%)
|
12(8.4)
|
11(7.3)
|
9(6.1)
|
0.749
|
Echocardiography and laboratory values
|
Left atrial diameter (mm)
|
40.69±4.76
|
42.77±6.21
|
45.29±7.17
|
<0.001
|
Left ventricular: EDD(mm)
|
53.95±8.45
|
53.89±8.51
|
58.63±10.26
|
<0.001
|
ESD(mm)
|
39.06±10.37
|
38.93±10.55
|
45.11±12.51
|
<0.001
|
EF%
|
53.07±13.93
|
52.75±14.85
|
45.49±15.70
|
<0.001
|
E/e'
|
10.33±3.68
|
12.01±4.66
|
13.86±6.91
|
<0.001
|
Log NT-proBNP
|
2.23±0.78
|
2.53±0.83
|
2.91±0.84
|
<0.001
|
Triglyceride (mmol/L)
|
1.60±0.66
|
1.45±0.59
|
1.25±0.56
|
<0.001
|
Total cholesterol (mmol/L)
|
3.67±1.00
|
3.88±1.05
|
3.95±1.07
|
0.071
|
HDL-c(mmol/L)
|
1.05±0.21
|
1.14±0.26
|
1.09±0.24
|
0.006
|
LDL-c(mmol/L)
|
2.07±0.88
|
2.23±0.89
|
2.37±0.91
|
0.018
|
HbA1c%
|
6.73±1.60
|
6.05±0.98
|
6.18±0.98
|
<0.001
|
ALT(IU/L)
|
24.18±14.59
|
28.24±27.89
|
29.84±66.84
|
0.511
|
eGFR(mL/min/1.73m2 )
|
77.44±24.37
|
77.90±21.40
|
76.18±22.15
|
0.797
|
Amino acids
|
|
|
|
|
Leucine (umol/ml)
|
203.70±79.42
|
176.79±52.55
|
147.74±44.77
|
<0.001
|
Isoleucine (umol/ml)
|
120.45±63.03
|
95.22±22.63
|
85.22±23.29
|
<0.001
|
Valine (umol/ml)
|
351.38±82.41
|
301.90±52.12
|
267.58±61.98
|
<0.001
|
Phenylalanine (umol/ml)
|
90.86±21.76
|
101.52±21.65
|
112.59±30.86
|
<0.001
|
Tyrosine (umol/ml)
|
60.47±15.67
|
66.17±15.50
|
79.97±26.86
|
<0.001
|
BCAA (umol/ml)
|
675.53±163.08
|
573.91±107.70
|
500.54±113.96
|
<0.001
|
Fischer's ratio
|
4.55±1.23
|
3.42±0.17
|
2.65±0.45
|
<0.001
|
Medications
|
|
|
|
|
ACEI/ARB/ARNI,n(%)
|
95(66.4)
|
94(62.7)
|
97(65.5)
|
0.778
|
Beta-blocker,n(%)
|
120(83.9)
|
98(65.3)
|
110(74.3)
|
0.001
|
Calcium antagonist,n(%)
|
27(18.9)
|
28(18.7)
|
12(8.1)
|
0.013
|
Digoxin,n(%)
|
2(1.4)
|
4(2.7)
|
11(7.4)
|
0.018
|
Diuretics,n(%)
|
40(28.0)
|
46(30.7)
|
66(44.6)
|
0.006
|
Aspirin,n(%)
|
68(47.6)
|
47(31.3)
|
42(28.4)
|
0.001
|
Statins,n(%)
|
121(84.6)
|
117(78.0)
|
100(67.6)
|
0.002
|
Hypoglycaemic drugs,n(%)
|
57(39.9)
|
40(27.4)
|
39(26.4)
|
0.022
|
FR, Fischer's ratio; BMI, body mass index; BP, blood pressure; NYHA, New York Heart Association; AF, atrial fibrillation; DM, diabetes mellitus; EDD, left ventricular end-diastolic diameter; ESD, left ventricular end-systolic diameter; EF, ejection fraction; E/e’, the ratio of early diastolic mitral inflow to mitral annular tissue velocities; NT-proBNP, N-terminal brain natriuretic peptide; HDL-c, high-density lipoprotein–cholesterol; LDL-c, low-density lipoprotein–cholesterol; HbA1c, glycated hemoglobin A1c; ALT, alanine transaminase; eGFR, estimated glomerular filtration rate; BCAA, branched chain amino acid, the sum of leucine, isoleucine, and valine; Fischer's ratio, BCAA divided by the sum of phenylalanine and tyrosine; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor neprilysin inhibitor.
To further explore the relationship between FR and HF, subjects were grouped according to the presence of HF or not. We found that the serum FR index was lower in the HF group than that in the non-HF group (Fig. 2a; 3.32 ± 0.90 vs. 3.73 ± 1.20, P < 0.001). The serum FR index also gradually decreased among the control, HF with E/e’≤14, and HF with E/e’>14 groups by one-way Avona (Fig. 2b; 3.73 ± 1.20 vs. 3.45 ± 0.94 vs. 3.18 ± 0.83, P < 0.001) and has the significant difference between the control group and other two HF groups, respectively (all LSD, P < 0.05).
Logistic regression and ROC analysis
After full adjustment for cardiovascular risk factors (gender, age, BMI, smoking, drinking, DM, hyperlipidemia, hypertension, coronary disease, AF and chronic kidney disease), the tertiles of low FR were correlated to the presence of systolic HF in all subjects (OR: 2.124; 95% CI: 1.595–2.829; P < 0.001). Compared with the highest tertile, the risk of HF in the lowest tertile increased by 337.4% (OR: 4.374; 95% CI: 2.464–7.765; P < 0.001) (Table 2). The ROC plot was calculated to test the predictive values of the low FR tertiles carried out by logistic regression for the occurrence of systolic HF in all subjects. The optimal value of the cut-off point for the probability was 0.223 (sensitivity 45.07%, specificity 77.19%). The area under the curve (AUC) was 0.618 (95% CI, 0.565–0.670; P < 0.001) (Fig. 3a). Then, another ROC plot was calculated to test the prediction probabilities with the above full adjustment and low FR tertiles carried out by logistic regression for the occurrence of systolic HF in all subjects. The additional inclusion of low FR tertiles provided a significantly improved predictive value (AUC, 0.772 vs. 0.652; P < 0.001). The optimal value of the cut-off point for the probability was 0.419 (sensitivity 62.91%, specificity 78.95%). The area under the curve (AUC) was 0.722 (95% CI, 0.675–0.770; P < 0.001) (Fig. 3b).
Table 2
Low FR value was associated with the presence of chronic HF in all subjects
|
Unadjusted OR
|
P value
|
Adjusted for Model 1 OR
|
P value
|
Adjusted for Model 2 OR
|
P value
|
FR tertiles
|
1.723(1.358-2.188)
|
<0.001
|
1.751(1.369-2.238)
|
<0.001
|
2.124(1.595-2.829)
|
<0.001
|
Tertile 1
|
1(ref)
|
|
1(ref)
|
|
1(ref)
|
|
Tertile 2
|
1.127(0.706-1.800)
|
0.616
|
1.166(0.714-1.906)
|
0.540
|
1.353(0.793-2.308)
|
0.267
|
Tertile 3
|
2.954(1.834-4.759)
|
<0.001
|
2.984(1.826-4.876)
|
<0.001
|
4.374(2.464-7.765)
|
<0.001
|
Model 1: Adjusted for age and gender.
Model 2: Adjusted for gender, age, BMI, smoking, drinking, DM, hyperlipidemia, hypertension, coronary disease, AF and chronic kidney disease.
FR, Fischer’s ratio; HF, heart failure; OR, odds Ratio; Other abbreviations are as in Table 1
Furthermore, in systolic HF subjects, the univariate and multivariate regression analyses showed that BMI, DM, low eGFR, LVEF and low FR tertiles were associated with elevated LV filling pressure (E/e’>14, all P < 0.05). Subjects in the lowest tertile had a 134.5% increased risk of abnormal LV filling pressure compared with those in the highest tertile (OR: 2.345; 95% CI: 1.113–4.940; P = 0.025) (Table 3). The prediction probability by this multivariate regression analysis was also verified by the ROC. The optimal value of the cut-off point for the probability was 0.371 (sensitivity 61.39%, specificity 75.68%). The AUC was 0.732 (95% CI, 0.665–0.799; P < 0.001) (Fig. 4).
Table 3
Low FR was correlated with abnormal left ventricular filling pressure (E/e’>14) in Systolic HF
|
Univariate
|
|
Multivariate 1
|
|
Multivariate 2
|
OR(95% CI)
|
P value
|
|
OR(95% CI)
|
P value
|
|
OR(95% CI)
|
P value
|
Gender(men)
|
0.666(0.347–1.277)
|
0.221
|
|
—
|
—
|
|
—
|
—
|
Age
|
1.284(0.749–2.202)
|
0.363
|
|
—
|
—
|
|
—
|
—
|
BMI
|
1.511(1.047–2.182)
|
0.028
|
|
1.631(1.088–2.445)
|
0.018
|
|
1.662(1.104–2.502)
|
0.015
|
Smoking
|
0.840(0.477–1.481)
|
0.547
|
|
—
|
—
|
|
—
|
—
|
Drinking
|
0.852(0.458–1.584)
|
0.613
|
|
—
|
—
|
|
—
|
—
|
DM
|
2.081(1.149–3.769)
|
0.016
|
|
2.142(1.112–4.127)
|
0.023
|
|
2.199(1.135–4.260)
|
0.020
|
Hyperlipidemia
|
1.352(0.400-4.571)
|
0.628
|
|
—
|
—
|
|
—
|
—
|
Hypertension
|
0.777(0.449–1.345)
|
0.368
|
|
—
|
—
|
|
—
|
—
|
Coronary disease
|
0.592(0.344–1.020)
|
0.059
|
|
—
|
—
|
|
—
|
—
|
AF
|
1.040(0.561–1.929)
|
0.901
|
|
—
|
—
|
|
—
|
—
|
Low eGFR
|
1.642(1.176–2.293)
|
0.004
|
|
1.581(1.102–2.268)
|
0.013
|
|
1.578(1.098–2.267)
|
0.014
|
LVEF
|
2.995(1.714–5.235)
|
< 0.001
|
|
2.612(1.445–4.722)
|
0.001
|
|
2.666(1.470–4.836)
|
0.001
|
FR tertiles
|
1.493(1.067–2.089)
|
0.019
|
|
1.525(1.053–2.209)
|
0.025
|
|
—
|
—
|
FR tertile 1
|
1(ref)
|
—
|
|
—
|
—
|
|
1(ref)
|
—
|
FR tertile 2
|
1.683(0.861–3.290)
|
0.128
|
|
—
|
—
|
|
2.048(0.977–4.294)
|
0.058
|
FR tertile 3
|
2.233(1.139–4.378)
|
0.019
|
|
—
|
—
|
|
2.345(1.113–4.940)
|
0.025
|
Multivariate model 1 or 2, adjusted for BMI, DM, low eGFR, left ventricular EF and with the addition of the FR tertiles. |
FR, Fischer’s ratio; HF, heart failure; OR, odds ratio; 95% CI, 95% confidence Internal; Other abbreviations are as in Table 1. |
Correlation between FR value and other cardiac index
The FR value was correlated negatively not only with E/e’ value (r= -0.230, P = 0.001), but also with left atrial diameters (r= -0.242, P < 0.001) and log NT-proBNP value (r= -0.243, P < 0.001), respectively (Fig. 5).
Prognosis study of FR in systolic HF patients
To explore the prognostic value of FR in chronic HF, 188 subjects were followed for a mean year (11.74 ± 1.44 months) after their discharges and 25 subjects were lost. Then, 188 subjects were reclassification into three groups, tertile 1 (FR ≥ 3.60), tertile 2 (2.87 ≤ FR < 3.60) and tertile 3 (FR < 2.87), from highest to lowest, according to the tertiles of FR levels. The composite of cardiovascular death and the first HF readmission were chosen as HF endpoint events. Of our 188 HF patients, 26 met the HF endpoint events, including 24 HF readmission and 7 cardiovascular death. The HF subjects who met the HF endpoint or HF readmission or only met cardiovascular death had significantly lower FR at baseline than those who did not (all P༜0.05) (Fig. 6a). The Kaplan–Meier survival curves to visualize the relationship between FR and outcomes. When grouped by FR tertiles, subjects in the lowest tertile were more likely to meet the HF endpoint events or HF readmission compared with subjects in the higher tertile (all Log-rank P < 0.01) (Fig. 6b). Following multivariable cox regression analyses adjusting for conventional risk factors including gender, age, BMI, smoking, drinking, DM, hyperlipidemia, hypertension, coronary disease, AF, chronic kidney disease and LVEF, low FR tertiles remained correlated to the risk of HF endpoint (HR: 1.838, 95% CI: 1.035–3.263, P = 0.038) (Fig. 6c).