Comparison of the demographic and clinical characters is shown in Table 1. The proportion of males(84.5% vs 61.8%, p = 0.004) and smoking history(63.4% vs 43.6%, p = 0.021) in the CSF group was significantly higher than that in the NCSF group, and there was no significant difference in age, BMI, hypertension, type 2 diabetes mellitus, and medications between the two groups.
Table 1
Demographic and clinical characteristics of the study population
|
NCSF group (n = 55)
|
CSF group (n = 71)
|
p
|
Age
|
55.73 ± 11.93
|
55.85 ± 11.06
|
0.954
|
Male
|
34(61.8%)
|
60(84.5%)
|
0.004
|
BMI
|
26.29 ± 4.97
|
27.15 ± 3.41
|
0.251
|
Smoking history
|
24(43.6%)
|
45(63.4%)
|
0.021
|
Anti-platelet drugs
|
30(54.5%)
|
36(50.7%)
|
0.402
|
β-blockers
|
19(34.5%)
|
23(32.4%)
|
0.474
|
ACE inhibitors
|
6(10.9%)
|
5(7%)
|
0.326
|
ARBs
|
8(14.5%)
|
10(14.1%)
|
0.569
|
Statins
|
26(47.3%)
|
31(43.7%)
|
0.411
|
Hypertension
|
36(65.5%)
|
42(59.2%)
|
0.296
|
Diabetes
|
13(23.6%)
|
10(14.1%)
|
0.276
|
LVEF
|
64.11 ± 4.20
|
48.68 ± 4.16
|
0.264
|
LVEDd
|
48.44 ± 4.33
|
48.68 ± 4.16
|
0.753
|
Comparison of the laboratory variables between the two groups is shown in Table 2. The differences in TG, TC, HDL-C, LDL-C, HbA1C, LVEDd, LVEF, NT-proBNP and eGFR were not significant between the groups. The diameters of the LAD(6.27 ± 1.25 vs 5.52 ± 1.33, p = 0.002), LCX(5.91 ± 1.39 vs 5.05 ± 1.45, p = 0.001) and Mean coronary arteries(5.50 ± 0.85 vs 5.18 ± 0.91, p < 0.001) were significantly larger in the CSF group compared to those in the NCSF group; however, no significant difference was found between the groups in RCA diameter(6.37 ± 2.02 vs 5.91 ± 1.55, p = 0.173). The TFC in all coronary arteries(LAD:28.43 ± 8.16 vs 19.45 ± 4.81, p < 0.001; LCX:32.54 ± 11.46 vs 20.24 ± 4.41, p < 0.001; RCA:30.13 ± 12.30 vs 18.74 ± 5.02, p < 0.001), Mean TFC(30.39 ± 8.32 vs 19.41 ± 3.77, p < 0.001), NLR(2.44 ± 1.12 vs 1.89 ± 0.58, p = 0.001) and URIC(370.78 ± 109.79 vs 329.15 ± 79.71, p = 0.019) level were higher in the CSF group than in the NCSF group, while the LMR(4.81 ± 1.66 vs 5.96 ± 1.75, p < 0.001) and ALB(44.13 ± 4.10 vs 45.69 ± 4.11, p = 0.036) level were significantly lower in the CSF group compared to those in the NCSF group.
Table 2
Comparison of laboratory variables in the study population
|
NCSF group (n = 55)
|
CSF group (n = 71)
|
p
|
LAD diameter (mm)
|
5.52 ± 1.33
|
6.27 ± 1.25
|
0.002
|
LCX diameter (mm)
|
5.05 ± 1.45
|
5.91 ± 1.39
|
0.001
|
RCA diameter (mm)
|
5.91 ± 1.55
|
6.37 ± 2.02
|
0.173
|
Mean diameter of coronary arteries
|
5.18 ± 0.91
|
5.50 ± 0.85
|
< 0.001
|
LAD TIMI frame count
|
19.45 ± 4.81
|
28.43 ± 8.16
|
< 0.001
|
LCX TIMI frame count
|
20.24 ± 4.41
|
32.54 ± 11.46
|
< 0.001
|
RCA TIMI frame count
|
18.74 ± 5.02
|
30.13 ± 12.30
|
< 0.001
|
Mean TIMI frame count
|
19.41 ± 3.77
|
30.39 ± 8.32
|
< 0.001
|
WBCs
|
6.36 ± 1.61
|
6.60 ± 1.58
|
0.399
|
Neutrophils
|
3.75 ± 1.15
|
4.11 ± 1.22
|
0.091
|
Lymphocytes
|
2.08 ± 0.60
|
1.88 ± 0.69
|
0.08
|
Monocytes
|
0.37 ± 0.11
|
0.41 ± 0.13
|
0.064
|
NLR
|
1.89 ± 0.58
|
2.44 ± 1.12
|
0.001
|
LMR
|
5.96 ± 1.75
|
4.81 ± 1.66
|
< 0.001
|
hsCRP
|
1.53 ± 1.67
|
1.98 ± 2.26
|
0.218
|
CRP
|
2.65 ± 2.16
|
3.33 ± 3.34
|
0.192
|
HGB
|
147.82 ± 14.46
|
152.15 ± 14.48
|
0.098
|
ALB
|
45.69 ± 4.11
|
44.13 ± 4.10
|
0.036
|
URIC
|
329.15 ± 79.71
|
370.78 ± 109.79
|
0.019
|
HbA1C
|
6.27 ± 1.21
|
6.12 ± 1.08
|
0.469
|
Fasting Glucose
|
6.332.13
|
5.79 ± 2.03
|
0.152
|
TG
|
1.81 ± 1.12
|
1.93 ± 1.59
|
0.653
|
TC
|
4.46 ± 1.08
|
4.50 ± 1.16
|
0.820
|
HDL-C
|
1.18 ± 0.34
|
1.20 ± 0.38
|
0.704
|
LDL-C
|
2.74 ± 0.91
|
2.74 ± 0.95
|
0.970
|
eGFR
|
103.65 ± 32.28
|
102.09 ± 25.30
|
0.762
|
NT-proBNP
|
114.53 ± 143.88
|
142.57 ± 196.85
|
0.419
|
In the Spearman correlation analyses (Table 3), the LMR(r= -0.21, p = 0.026), ALB(r= -0.187, p = 0.036) level and male sex(r= -0.265,p = 0.003) were negatively correlated with CSF, whereas the diameters of the LAD(r = 0.297,p < 0.001), LCX(r = 0.218,p = 0.016) and RCA(r = 0.235,p = 0.01), as well as the mean diameter of the coronary arteries(r = 0.337,p < 0.001), NLR(r = 0.245,p = 0.009) and URIC(r = 0.218,p = 0.021) level were positively correlated with CSF. Smoking history(r = 0.2,p = 0.026) was also positively correlated with CSF.
Table 3
Spearman’s rho correlation analysis between the CSF with risk factors
Variable
|
r
|
p
|
Gender
|
-0.265
|
0.003
|
Smoking history
|
0.2
|
0.026
|
LAD diameter (mm)
|
0.297
|
< 0.001
|
LCX diameter (mm)
|
0.218
|
0.016
|
RCA diameter (mm)
|
0.235
|
0.01
|
Mean D (mm)
|
0.337
|
< 0.001
|
NLR
|
0.245
|
0.009
|
LMR
|
-0.21
|
0.026
|
URIC
|
0.218
|
0.021
|
ALB
|
-0.187
|
0.036
|
To further explore the independent predictor(s) of CSF, univariable and multivariable logistic regression model analyses were performed based on the correlation analysis results (Table 4). Because the diameters of the LAD and LCX were components of the Mean D, we thought that it might negatively affect the regression analysis results. Therefore, we didn’t evaluate the diameters of the LAD and LCX in the multivariable regression analysis. Collinearity diagnostics showed that the VIF of males, smoking history, mean diameter of coronary arteries, NLR, LMR, ALB level and URIC level were less than 10. We performed the multivariate analysis by using the backward LR method. Male sex(odds ratio(OR):1.601,95%CI:0.443–5.781,p = 0.437), smoking history (OR:1.478, 95%CI: 0.505–4.324, p = 0.476), NLR(OR:1.741,95%CI:0.854–3.549,p = 0.127) and ALB(OR:0.917,95%CI:0.826–1.019, p = 0.107) level were not found to be independent predictors for CSF phenomenon. Decreased LMR (OR: 0.614; 95% CI: 0.464–0.814; p = 0.001), increased mean diameter of coronary arteries (OR: 2.634; 95% CI: 1.54–4.51; p < 0.001) and URIC level (OR: 1.006; 95% CI: 1.001–1.012; p = 0.018) were found to be independent predictors of CSF development.
Table 4
Univariable and multivariable logistic regression analyses on the presence of slow coronary flow
Variables
|
Univariable
|
Multivariable
|
OR (95% CI) p
|
OR (95% CI) p
|
Male
|
3.369(1.451–7.820) 0.005
|
1.601(0.443–5.781) 0.437
|
Smoking history
|
2.236(0.218–0.918) 0.028
|
1.478(0.505–4.324) 0.476
|
LAD diameter
|
1.584(1.169–2.147) 0.003
|
|
LCX diameter
|
1.531(1.169–2.005) 0.002
|
|
Mean diameter of coronary arteries
|
2.62(1.543–4.117) <0.001
|
2.634(1.54–4.51) <0.001
|
URIC
|
1.005(1.001–1.009) 0.023
|
1.006(1.001–1.012) 0.018
|
NLR
|
2.18(1.321–3.599) 0.002
|
1.741(0.854–3.549) 0.127
|
LMR
|
0.67(0.530–0.847) 0.001
|
0.614(0.464–0.814) 0.001
|
ALB
|
0.911(0.833–0.995) 0.039
|
0.917(0.826–1.019) 0.107
|
OR: odds ratio
|
The ROC curve analysis demonstrated that the specificity of the Mean D > 5.76 mm in predicting CSF was 70.59% and the sensitivity was 64.81% (AUC = 0.696, p < 0.001). It was revealed that using a cutoff level of LMR < 4.11 predicted CSF with a sensitivity of 90.9% and specificity of 40.85% (AUC = 0.698, p < 0.001). In addition, the URIC level was found to have the area under the curve (AUC = 0.616, p = 0.026) with an optimal URIC cutoff value of 314.39 µmol/L (sensitivity 74.65%, specificity 49.09%). In addition, the combination revealed a cutoff level of ≤-4.08 with a sensitivity of 86.76% and specificity of 61.11% (AUC = 0.772, p < 0.001). Combine Mean D,URIC and LMR to produce a new predictor, the combination(AUC = 0.772, p < 0.001) with an optimal cutoff value of less than − 4.08 with a sentivity of 86.76% and speciticity of 61.11%. Comparison of four ROC curves by using the method of DeLong et al. was also performed. The predictive performance of the combination was significantly superior to that of Mean D(p = 0.0053) and URIC level ( p = 0.0266); however, the difference between the areas of combination and LMR was not significant (p = 0.1061).