3.1 Baseline characteristics
There were 59 patients in total and 57 healthy controls in this study. Based on the presence or absence of heart failure symptoms or signs, BNP levels, and structural alterations on echocardiography, patients were separated into a heart failure group and a non-heart failure group. Baseline characteristics are summarized in Table 1. Age, gender, BMI, and the presence of underlying diseases (hypertension, diabetes, stroke) were comparable between heart failure patients and healthy controls. Compared with the healthy controls, heart failure patients had significantly higher levels of NTproBNP (P < 0.001), LVEF (P < 0.001), IL-6 (P < 0.001), IL-8 (P < 0.001), IL-10 (P < 0.001), IL-17 (P = 0.007), TNF-α (P < 0.001), hsCRP (P = 0.009), ESR (P = 0.048), Albumin (P < 0.001), N# (P = 0.004), and M# (P = 0.007), whereas the difference in the level of TG, TC, BG, HbA1C, globulin, and C4 was comparable. In contrast, heart failure patients had significantly lower levels of Lp-PLA2 (P = 0.025), C3 (P = 0.005), Hb (P = 0.001), and L# (P = 0.001).
Table 1
Demographic and clinical data of patients with heart failure.
|
heart failure group(n = 59)
|
non-heart failure group(n = 57)
|
p
|
Age,Mean ± SD
|
69.29 ± 12.08
|
66.70 ± 12.01
|
0.250
|
sex,n(%)
|
|
|
NS
|
male
|
34(57.6)
|
27(47.4)
|
|
felmale
|
25(42.4)
|
30(52.6)
|
|
NYHA class,n(%)
|
|
|
NS
|
II
|
21(35.6)
|
|
|
III
|
22(37.3)
|
|
|
IV
|
16(27.1)
|
|
|
Etiology,n(%)
|
|
|
NS
|
CAD
|
26(44.1)
|
|
|
Valve disease
|
7(11.9)
|
|
|
Arrhythmias
|
21(35.6)
|
|
|
CMPs
|
6(10.2)
|
|
|
Others
|
3(5.1)
|
|
|
underlying diseases,n(%)
|
|
|
|
Diabetes
|
18(30.5)
|
18(31.6)
|
0.376
|
Hypertension
|
39(66.1)
|
42(73.7)
|
0.901
|
Cerebral infarction
|
2(3.4)
|
3(5.3)
|
0.635
|
myocardial infarction
|
8(13.6)
|
1
|
0.017
|
atrial fivrillation
|
23(39.0)
|
0
|
0.000
|
BMI,Mean ± SD
|
25.03 ± 3.71
|
24.76 ± 3.34
|
0.649
|
TG,M(IQR)
|
1.01(0.77,1.53)
|
1.28(0.80,1.95)
|
0.085
|
TC,Mean ± SD
|
3.59 ± 1.12
|
3.93 ± 1.21
|
0.093
|
BG,M(IQR)
|
5.09(4.36,6.31)
|
5.27(4.71,5.94)
|
0.513
|
HbA1C,M(IQR)
|
6.30(5.80,7.00)
|
6.15(5.70,6.85)
|
0.623
|
NTproBNP,M(IQR)
|
3628.00(2315.00,6158.00)
|
71.65(19.43,235.83)
|
0.000
|
LVEF,Mean ± SD
|
45.53 ± 12.92
|
63.75 ± 4.11
|
0.000
|
Interleukin 6(IL-6),M(IQR)
|
32.91(21.38,94.27)
|
8.97(6.26,12.16)
|
0.000
|
Interleukin 8(IL-8),M(IQR)
|
72.50(33.13,170.28)
|
23.86(17.99,31.39)
|
0.000
|
Interleukin 10(IL-10),M(IQR)
|
7.71(4.80,15.33)
|
4.63(3.23,6.14)
|
0.000
|
Interleukin 17(IL-17),M(IQR)
|
17.66(7.33,23.96)
|
11.72(6.52,17.26)
|
0.007
|
Tumor Necrosis Factor alpha(TNF-α),M(IQR)
|
9.59(6.57,15.26)
|
7.19(4.22,10.11)
|
0.000
|
Lp-PLA2,M(IQR)
|
219.00(175.00,303.00)
|
237.00(219.00,298.25)
|
0.025
|
hsCRP,M(IQR)
|
6.98(0.84,15.16)
|
3.13(1.02,5.05)
|
0.009
|
ESR,M(IQR)
|
10.00(6.00–18.00)
|
7.00(3.00-13.75)
|
0.048
|
Albumin,Mean ± SD
|
34.73 ± 2.23
|
37.76 ± 2.56
|
0.000
|
Globulin,Mean ± SD
|
26.50 ± 4.47
|
27.79 ± 4.65
|
0.130
|
Serum complement C3,Mean ± SD
|
0.82 ± 0.23
|
0.94 ± 0.22
|
0.005
|
Serum complement C4,M(IQR)
|
0.26(0.21,0.37)
|
0.27(0.21,0.37)
|
0.840
|
Hemoglobin,Mean ± SD
|
127.08 ± 16.05
|
136.33 ± 12.39
|
0.001
|
Neutrophil count 109/L,M(IQR)
|
4.80(3.30,7.70)
|
3.80(3.10,4.38)
|
0.004
|
Lymphocyte count 109/L,M(IQR)
|
1.20(0.90,1.70)
|
1.55(1.20,1.90)
|
0.001
|
Monocyte count 109/L,M(IQR)
|
0.51(0.35,0.62)
|
0.38(0.31,0.50)
|
0.007
|
mean (SD), mean (standard deviation); median (IQR), median (interquartile range);NS, not significant.
3.2 Cytokines were significantly elevated in patients with heart failure
Patients with heart failure frequently have high cytokine levels, which is a phenomenon known as cytokine storm. More specifically, the serum level of IL-6, IL-8, IL-10, IL-17, and TNF-α is markedly elevated, with some patients having cytokine levels up to 210 times higher than those in healthy control groups.(Fig. 1) Previous researchs have established that TNF-α and IL-6 levels are significantly elevated in heart failure patients, and the increase in their homologous receptor levels is correlated with the risk of mortality. More significantly, sTNFR2 is a major indicator of mortality in people with heart failure.[11] Indeed, among numerous cytokines, the change in the level of IL-6 was most pronounced, and consequently, IL-6 has been identified as a biomarker of HFpEF. It may participate in the progression from asymptomatic or subclinical left ventricular dysfunction to symptomatic left ventricular dysfunction and clinical heart failure.[12]
3.4 Increased hs-CRP release and decreased Lp-PLA2 level in patients with heart failure
The plaque screening test determined that compared to patients in the healthy control group, heart failure patients had significantly elevated levels of hs-CRP in their blood.(Fig. 2) CRP is principally synthesized by hepatocytes and can also be produced by damaged heart muscle tissue. Elevated CRP levels generally indicate a strong inflammatory response, and thus, elevated CRP levels conduce to heart failure.[14]
An enzyme called lipoprotein-associated phospholipase A2 (Lp-PLA2), which can hydrolyze platelet-activating substances and is thought to have anti-inflammatory properties, is produced by macrophages and activated platelets. However, according to some studies, Lp-PLA2 is related to low-density lipoprotein (LDL).[15] 80% of Lp-PLA2 can bind to LDL in the blood and hydrolyzes oxidized phospholipids; hence, it is speculated to exert pro-inflammatory effects. At present, compelling evidence suggests that Lp-PLA2 is an independent predictor of heart failure[16] and is a double-edged sword in atherosclerosis development. In other words, it can not only promote atherosclerosis by driving the production of pro-inflammatory factors such as oxidized fatty acids (OX-FA) but also inhibit atherosclerosis via the aforementioned pathway. In the present study, our results uncovered that Lp-PLA2 mediates anti-inflammatory effects in heart failure patients.(Fig. 2) Moreover, Lp-PLA2 has the ability to hydrolyze inflammatory phospholipids and inhibit prothrombotic factors. Lp-PLA2 levels in heart failure patients' blood are decreased, which inhibits the body's capacity to control inflammatory responses, causing the inflammatory response to worsen.
3.5 Increased neutrophil and monocyte counts, decreased lymphocyte counts, decreased hemoglobin levels
Blood routine tests of heart failure patients exposed that compared to the healthy control group, heart failure patients had increased neutrophil and monocyte counts and decreased lymphocyte counts.(Fig. 3) Neutrophils are involved in numerous acute inflammatory response processes. A previous study pointed out that in a mouse model of heart failure induced by acute myocardial infarction, neutrophils were abundantly distributed in the infarction border zone, whilst depletion of neutrophils significantly alleviated myocardial fibrosis and enhanced left ventricular function.[17] According to previous studies, the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) can serve as novel predictors of heart failure. A study found that NLR, in conjunction with PLR, is closely correlated with heart failure-related mortality and can be a valuable predictor of death.[18]
Monocytes can phagocytize antigens and subsequently present them to lymphocytes, inducing lymphocytes to synthesize antigen-specific antibodies. They can also differentiate into macrophages by crossing the vascular wall. Prior studies have postulated that the CCr2 + subset of macrophages has the ability to create inflammatory chemokines, cytokines, and oxidants, coordinating the recruitment of neutrophils and monocytes, and participating in ventricular remodeling.[19–21] In addition, blood routine tests revealed a connection between low hemoglobin levels and heart failure, which may be explained by the direct inhibition of hematological function by the inflammatory response as well as decreased cardiac output that leading to renal anemia.
3.6 Analysis of risk factors in heart failure patients
Based on the above-mentioned research, individuals with heart failure and the control group showed substantial disparities in a number of indices. After adjusting for confounding factors, logistic regression analysis was done on these indicators, and the results exposed that IL-6, IL-8, and IL-17 were independent risk factors for heart failure, whereas Lp-PLA2 and albumin were protective factors.(Table 2)
Table 2
Univariate and multivariate logistic regression analysis.
Variable
|
Univariate analysis
|
Multivariate analysis
|
OR
|
OR(95%CI)
|
P
|
OR
|
OR(95%CI)
|
P
|
Interleukin 6(IL-6)
|
1.212
|
1.121–1.309
|
0.000
|
1.269
|
1.094–1.472
|
0.002
|
Interleukin 8(IL-8)
|
1.089
|
0.049–1.130
|
0.000
|
1.071
|
1.012–1.134
|
0.018
|
Interleukin 10(IL-10)
|
1.326
|
1.158–1.518
|
0.000
|
0.803
|
0.544–1.184
|
NS
|
Interleukin 17(IL-17)
|
1.052
|
1.017–1.089
|
0.004
|
1.180
|
1.010–1.378
|
0.037
|
Tumor Necrosis Factor alpha(TNF-α)
|
1.196
|
1.082–1.322
|
0.000
|
0.859
|
0.758–0.974
|
0.018
|
Lp-PLA2
|
0.995
|
0.990-1.000
|
0.041
|
0.986
|
0.972–0.999
|
0.036
|
hsCRP
|
1.155
|
1.070–1.248
|
0.000
|
0.772
|
0.578–1.031
|
NS
|
ESR
|
1.061
|
1.003–1.121
|
0.039
|
1.045
|
0.901–1.213
|
NS
|
albumin
|
0.579
|
0.446–0.719
|
0.000
|
0.476
|
0.267–0.848
|
0.012
|
Serum complement C3
|
0.093
|
0.017–0.517
|
0.007
|
0.047
|
0.001–3.889
|
NS
|
Hemoglobin
|
0.957
|
0.931–0.983
|
0.001
|
0.984
|
0.921–1.051
|
NS
|
Neutrophil count 109/L
|
1.419
|
1.138–1.770
|
0.002
|
1.440
|
0.755–2.746
|
NS
|
Lymphocyte count 109/L
|
0.365
|
0.183–0.728
|
0.004
|
1.415
|
0.239–8.385
|
NS
|
Monocyte count 109/L
|
28.185
|
2.843-279.391
|
0.004
|
0.016
|
0.000-18.667
|
NS
|
OR, odds ratio; CI, confidence interval; NS, not significant.
3.7 Combination of multiple inflammatory cytokines could elevate the accuracy of the diagnosis of heart failure and inflammatory reaction.
Based on our research data, it can be deduced that the AUC for the diagnosis of heart failure using a single cytokine, namely IL-6, was 0.9120 (95% CI: 0.8579–0.9661, P < 0.0001), with a Youden index of 0.7598, sensitivity of 84.75%, and specificity of 91.23%. Regarding IL-8, the AUC was 0.8663 (95% CI: 0.7988–0.9338, P < 0.0001), with a Youden index of 0.66, sensitivity of 66.1%, and specificity of 100%. Concerning IL-17, the AUC was 0.6447 (95% CI: 0.5441–0.7452, P = 0.0072), with a Youden index of 0.2804, sensitivity of 50.85%, and specificity of 77.19%. On the other hand, the AUC of the combination of two cytokines, namely IL-6 + IL-8, for heart failure diagnosis was 0.9251 (95% CI: 0.8740–0.9761, P < 0.0001), with a Youden index of 0.7767, sensitivity of 86.44%, and specificity of 91.23%. Likewise, the AUC was 0.9129 (95% CI: 0.8581–0.9677, P < 0.0001) for IL-6 + IL-17, with a Youden index of 0.7598, sensitivity of 84.75%, and specificity of 91.23%. For IL-8 + IL-17, the AUC was 0.8748 (95% CI: 0.8102–0.9394, P < 0.0001), with a Youden index of 0.678, sensitivity of 67.8%, and specificity of 100%. Finally, the AUC for the combination of IL-6 + IL-8 + IL-17 was 0.9277 (95% CI: 0.8788–0.9767, P < 0.0001), with a Youden index of 0.7779, sensitivity of 83.05%, and specificity of 93.74%. As anticipated, the use of multiple cytokines was associated with increased AUC and Youden index, signifying that their accuracy for diagnosing heart failure was higher than that of a single cytokine or the combination of two cytokines. However, compared with NT-proBNP, which had an AUC of 0.9914 (95% CI: 0.9805-1.000, P < 0.0001), a Youden index of 0.9914, sensitivity of 94.74%, and specificity of 98.25%, the diagnosis of heart failure using cytokines still face some limitations. The sensitivity of diagnosis using cytokines is generally lower than that of NT-proBNP, implying that the actual positive rate for heart failure is low. Although the specificity of IL-8 or IL-8 + IL-17 may be higher than NT-proBNP, using NT-proBNP for heart failure diagnosis remains more reliable and practical than using cytokines.(Fig. 4,Table 3)
Table 3
ROC Curve Analysis of Inflammatory Cytokines in Diagnosis of Heart failure and inflammation
Indicators
|
AUC (95% CI)
|
P value
|
Youden Index
|
Cutoff
|
Sensitivity (%)
|
Specificity (%)
|
IL-8
|
0.8663(0.7988–0.9338)
|
< 0.0001
|
0.66
|
46.71
|
66.1
|
100
|
IL-17
|
0.6447(0.5441–0.7452)
|
0.0072
|
0.2804
|
17.6
|
50.85
|
77.19
|
IL-6 + IL-8
|
0.9251(0.8740–0.9761)
|
< 0.0001
|
0.7767
|
0.4367
|
86.44
|
91.23
|
IL-6 + IL-17
|
0.9129(0.8581–0.9677)
|
< 0.0001
|
0.7598
|
0.4005
|
84.75
|
91.23
|
IL-8 + IL-17
|
0.8748(0.8102–0.9394)
|
< 0.0001
|
0.678
|
0.677
|
67.8
|
100
|
IL-6 + IL-8 + IL-17
|
0.9277(0.8788–0.9767)
|
< 0.0001
|
0.7779
|
0.4757
|
83.05
|
94.74
|
NTpro-BNP
|
0.9914(0.9805-1.000)
|
< 0.0001
|
0.9299
|
1124
|
94.74
|
98.25
|
hs-CRP
|
0.6409(0.5361–0.7458)
|
0.0089
|
0.4032
|
6.385
|
50.85
|
89.47
|
IL-6 + hs-CRP
|
0.9197(0.8691–0.9703)
|
< 0.0001
|
0.7598
|
0.4512
|
84.75
|
91.23
|
IL-8 + hs-CRP
|
0.8638(0.7932–0.9345)
|
< 0.0001
|
0.7113
|
0.5856
|
72.88
|
98.25
|
IL-17 + hs-CRP
|
0.7217(0.6268–0.8166)
|
< 0.0001
|
0.454
|
0.5749
|
55.93
|
89.47
|
IL-6 + IL-8 + hs-CRP
|
0.9352(0.8906–0.9797)
|
< 0.0001
|
0.7773
|
0.5321
|
74.75
|
92.98
|
IL-6 + IL-17 + hs-CRP
|
0.9215(0.8709–0.9721)
|
< 0.0001
|
0.7761
|
0.3773
|
88.14
|
89.47
|
IL-8 + IL-17 + hs-CRP
|
0.8742(0.8074–0.9410)
|
< 0.0001
|
0.7119
|
0.6750
|
71.19
|
100
|
IL-6 + IL-8 + IL-17 + hs-CRP
|
0.9358(0.8919–0.9797)
|
< 0.0001
|
0.7942
|
0.4740
|
86.44
|
92.98
|
Abbreviations: ROC, receiver operating characteristic; AUC, area under curve; CI, confidence interval; IL-6, interleukin-6; IL-8, interleukin-8; IL-17, interleukin-17; TNF-α, tumor necrosis factor-α; NTpro-BNP, N-terminal forebrain natriuretic peptide antigen; hs-CRP, hypersensitive-C-reactive-protein |
Notably, the AUC of hs-CRP was 0.6409 (95% CI: 0.5361-0.7458, P=0.0089), with a Youden index of 0.4032, sensitivity of 50.85%, and specificity of 89.47%. The AUC for IL-6+hs-CRP was 0.9197 (95% CI: 0.8691-0.9703, P<0.0001), with a Youden index of 0.7598, sensitivity of 85.75%, and specificity of 91.23%. Additionally, the AUC for IL-8+hs-CRP was 0.8638 (95% CI: 0.7932-0.9345, P<0.0001), with a Youden index of 0.7113, sensitivity of 72.88%, and specificity of 98.25%. The AUC for IL-17+hs-CRP was 0.7217 (95% CI: 0.6268-0.8166, P<0.0001), with a Youden index of 0.454, sensitivity of 55.93% and specificity of 89.47%. The AUC using a combination of two cytokines with hs-CRP for diagnosis, namely IL-6+IL-8+hs-CRP, was 0.9352 (95% CI: 0.8906-0.9797, P<0.0001), with a Youden index of 0.7773, sensitivity of 74.75%, and specificity of 92.98%. The AUC for IL-6+IL-17+hs-CRP was 0.9215 (95% CI: 0.8709-0.9721, P<0.0001), with a Youden index of 0.7761, sensitivity of 88.14%, and specificity of 89.47%. The AUC for IL-8+IL-17+hs-CRP was 0.8742 (95% CI: 0.8074-0.9410, P<0.0001), with a Youden index of 0.7119, sensitivity of 71.19%, and specificity of 100%. The AUC for the combination of three cytokines with CRP (IL-6+IL-8+IL-17+hs-CRP) was 0.9358 (95% CI: 0.8919-0.9797, P<0.0001), with a Youden index of 0.7942, sensitivity of 86.44%, and specificity of 92.78%. Compared with hs-CRP, cytokines have a stronger ability to diagnose vascular inflammation. For instance, sensitivity and specificity were higher using IL-6+IL-8+IL-17 combined with hs-CRP for evaluating intracellular inflammatory responses, indicating higher diagnostic efficacy. Thus, our results insinuate that combining these three cytokines with hs-CRP can more accurately reflect ongoing inflammatory responses in the body.(Figure 5,Table 3)