A Comparison of the Prognostic Value of Liver Fibrosis Scores in Acute Myocardial Infarction Patients With and Without Type 2 Diabetes

Abstract

Liver fibrosis scores have been demonstrated to be associated with poor prognosis after percutaneous coronary intervention (PCI). However, no studies have compared the prognostic value of these scores in acute myocardial infarction (AMI) patients with and without diabetes. We retrospectively enrolled 1576 AMI patients who underwent PCI. There were 177 all-cause deaths and 111 cardiac deaths during follow-up (median 3.8 years). The non-alcoholic fatty liver disease fibrosis score (NFS) showed a better prognostic value than the fibrosis-8 (FIB-8) score (Harrell’s C-index: 0.703 vs 0.671, P = .014) and the fibrosis-4 (FIB-4) score (Harrell’s C-index: 0.703 vs 0.648, P < .001) in the overall population. In the time-dependent receiver operating characteristic analysis, the NFS also had the highest area under the curve across all time points. Consistent results were observed in diabetic and non-diabetic populations. Adding the NFS to traditional cardiovascular risk factors significantly improved the prediction both for all-cause mortality (Harrell’s C-index: 0.806 vs 0.771, P < .001) and cardiac death (Harrell’s C-index: 0.800 vs 0.771, P = .014). The NFS showed a better prognostic value than the FIB-8 score and the FIB-4 score in patients with AMI undergoing PCI, which might be preferable for estimating the risk of mortality regardless of the presence or absence of diabetes.

Keywords

acute myocardial infarction percutaneous coronary intervention liver fibrosis score mortality prognosis

Introduction

With the widespread use of percutaneous coronary intervention (PCI) and an increased emphasis on standardized management, the mortality of patients with acute myocardial infarction (AMI) has significantly decreased.¹ However, it remains a global leading cause of death.²

Non-alcoholic fatty liver disease (NAFLD) and coronary artery disease share risk factors, including diabetes, obesity, and metabolic syndrome.³ Approximately 58.2% of patients with coronary artery disease also have NAFLD.⁴ Liver fibrosis is a progressive and irreversible pathological feature of NAFLD.⁵ Increasing evidence suggests the extent of liver fibrosis was associated with mortality and cardiovascular events in patients with NAFLD or diabetes,^6,7 even in the general population.⁸ Although the gold standard for assessing the stage of liver fibrosis is liver biopsy, the calculation of liver fibrosis scores appears to provide an alternative, non-invasive, and cost-effective approach.^9,10 Among the liver fibrosis scores, the most validated are the non-alcoholic fatty liver disease fibrosis score (NFS)¹¹ and the fibrosis-4 (FIB-4) score,¹² which were recommended as accurate tools for estimating the severity of liver fibrosis by clinical guidelines.^9,10 Besides, numerous studies have demonstrated that the NFS and the FIB-4 score could be used not only for evaluating the extent of liver fibrosis but also for predicting poor prognosis in patients with coronary artery disease undergoing PCI.^13–15 Thus, they may provide a potential approach for identifying high-risk coronary artery disease patients.

The diagnostic accuracy of the FIB-4 score for advanced liver fibrosis was higher than that of the NFS^16,17 and it was preliminary recommended in the diabetic population by guidelines.⁹ However, this proposal remains controversial.^10,18 Recently, the FIB-8 score was developed on the basis of the FIB-4 score and showed a better predictive value than the NFS and the FIB-4 score for significant fibrosis.¹⁹ However, the prognostic value of these scores has not been compared in AMI patients with and without diabetes. This study aims to examine and compare the prognostic performance of the NFS, the FIB-8, and the FIB-4 scores in AMI patients with and without diabetes who underwent PCI.

Methods

Study Design and Population

This study complies with the declaration of Helsinki and the Ethical approval was obtained from the Ethics Committee of the China–Japan Friendship Hospital (approval numbers: 2018-46-K35). The study population included patients with AMI undergoing PCI with stent implantation at China–Japan Friendship Hospital from January 2016 to December 2019. Patients older than 18 years of age with a diagnosis of AMI according to the Fourth Universal Definition of Myocardial Infarction were eligible for inclusion.²⁰ We excluded the patients: (1) with excessive alcohol consumption (>21 drinks/week for men or >14 drinks/week for women) or with etiologies of chronic liver disease including viral hepatitis, drug-induced liver disease, and autoimmune hepatitis; (2) missing covariates needed to compute the liver fibrosis scores; (3) with type 1 diabetes; (4) without follow-up data. A total of 1741 patients were initially screened for study participation. After exclusion, 1576 individuals were included in the current analysis. Before the PCI procedure, patients were prescribed a loading dose of 300 mg aspirin and 300 mg clopidogrel or 180 mg ticagrelor if they had not taken antiplatelet medications in the last 5 days. Following the procedure, patients were recommended to administer dual antiplatelet therapy including aspirin 100 mg/day and clopidogrel 75 mg/day or ticagrelor 180 mg/day for at least 1 year. Other medical treatments were left to the physician’s discretion, but were based on current guidelines.

Clinical and laboratory information was collected from electronic medical records. Diabetes was defined as self-reported history of diabetes, hemoglobin A1c ≥ 6.5%, or antidiabetic drug use. The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease (CKD) Epidemiology Collaboration equation.²¹ CKD was defined as an estimated glomerular filtration rate <60 mL/min/1.73 m².

Clinical Outcomes and Follow-Up

The primary endpoint was all-cause mortality, with cardiac death a secondary endpoint. All patients were followed up until death, loss to follow-up, or end of follow-up (November 31, 2022) through outpatient attendance or telephone interviews.

Definitions of the Risk Scores

The liver fibrosis scores were calculated by using the following formula: NFS = −1.675 + 0.037 × age (years) + 0.094 × body mass index (BMI) (kg/m²) + 1.13 × diabetes (yes = 1, no = 0) + 0.99 × aspartate aminotransferase/alanine aminotransferase ratio –0.013 × platelet count (×10⁹/L) –0.66 × albumin (g/dL).¹¹ FIB-4 = age (years) × aspartate aminotransferase (U/L)/[platelet count (×10⁹/L) × alanine aminotransferase (U/L)^1/2].¹² FIB-8 = FIB-4 + 0.025 × BMI (kg/m²) –0.702 × (albumin/globulin ratio) + 0.004 × gamma-glutamyl transpeptidase (U/L) + 0.858 × diabetes (yes = 1, no = 0).¹⁹ Based on the cutoff values from previous literature,^14,19 patients were divided into low, intermediate, and high-risk categories (i.e., −1.455 and 0.676 for the NFS, 1.3 and 2.67 for the FIB-4 score, 0.88 and 1.77 for the FIB-8 score, respectively).

Statistical Analysis

R statistical software (R Foundation for Statistical Computing, Vienna, Austria, version 4.2.1) and SPSS statistics software (version 25.0, SPSS, Inc., Chicago, IL, USA) were used for statistical analyses. Continuous variables were expressed as means ± standard deviations when normally distributed or medians (interquartile ranges) for variables with skewed distribution. Categorical variables were expressed as frequencies (percentages). Comparisons between groups were performed using the t test, Mann–Whitney U test, chi-squared test, or Fisher’s exact test, as appropriate.

Survival curves were plotted using Kaplan–Meier methods and analyzed by log-rank test. Cox proportional hazard regression models were used to evaluate the associations of liver fibrosis scores with survival. Variables significant in the univariate analysis (P < .05) were entered into multivariate models, including gender, BMI, current smoking, hypertension, diabetes, history of MI, prior coronary artery bypass grafting, chronic heart failure, statin, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, CKD, left ventricular ejection fraction (LVEF) <40%, multivessel disease, triglyceride, and low-density lipoprotein cholesterol, except for the variables included in the score formulas. For the subgroup analyses, interaction terms between subgroups and survival were examined. To assess possible nonlinear associations between the liver fibrosis scores and survival, restricted cubic spline models with 4 knots were built.

The prognostic performance of the liver fibrosis scores for survival was determined using Harrell’s C-statistic, continuous net reclassification improvement (NRI), and integrated discrimination improvement (IDI). To evaluate the global prognostic value of the liver fibrosis scores during follow-up, a time-dependent receiver operating characteristic analysis was performed. We also assessed the improvement in prognostic prediction by adding the liver fibrosis scores to a prognostic model that incorporated traditional cardiovascular risk factors. Prognostic model calibration was evaluated by the Gronnesby and Borgan test. Statistical significance was defined as a 2-sided P < .05.

Results

Baseline Characteristics

The flowchart was shown in Supplementary Figure 1. A total of 1576 patients with a mean age of 62.6 ± 13.1 years, 368 (23.4%) female, 969 (61.5%) ST-elevation MI, 607 (38.5%) non-ST-elevation MI were included in our study. The median follow-up was 3.8 years and there were 177 (11.2%) all-cause deaths and 111 (7.0%) cardiac deaths. Compared with the survivors, the non-survivors were older, more often female, and were less likely to receive statin and angiotensin-converting enzyme inhibitor/angiotensin receptor blocker therapy than the survivors. The prevalence of hypertension, diabetes, previous MI, chronic heart failure, CKD, and multivessel disease were higher and the levels of platelet, albumin, triglyceride, low-density lipoprotein-cholesterol, and LVEF were lower in patients who died than in those who survived. In addition, the levels of liver fibrosis scores were higher in non-survivors than in survivors. Baseline characteristics were summarized in Table 1.

Table 1.

Baseline Characteristics of Study Patients.

Factor	Survivors (n = 1399)	Non-survivors (n = 177)	P
Demographics
Age, years	61.3 ± 12.7	72.9 ± 11.6	<0.001
Age >75 years, n (%)	231 (16.5)	90 (50.8)	<0.001
Female, n (%)	308 (22.0)	60 (33.9)	0.001
BMI, kg/m²	25.4 ± 3.5	24.3 ± 3.9	0.001
Risk factors
Hypertension, n (%)	877 (62.7)	130 (73.4)	0.006
Diabetes, n (%)	557 (39.8)	98 (55.4)	<0.001
Current smoking, n (%)	652 (46.6)	53 (30.8)	<0.001
History of MI, n (%)	143 (10.2)	34 (19.2)	0.001
Prior PCI, n (%)	134 (9.6)	23 (13.0)	0.195
Prior coronary artery bypass grafting, n (%)	15 (1.1)	7 (4.0)	0.006
Chronic heart failure, n (%)	46 (3.3)	25 (14.1)	<0.001
CKD, n (%)	168 (12.0)	85 (48.0)	<0.001
Procedure performed
Multivessel disease, n (%)	1097 (78.4)	157 (88.7)	0.002
Number of stents, (mean ± SD)	1.3 ± 0.8	1.3 ± 0.8	0.370
Stent length, mm	34.4 ± 23.1	33.1 ± 23.2	0.458
Medical therapy during hospitalization
Antiplatelet, n (%)	1399 (100.0)	177 (100.0)	1.00
Statin, n (%)	1376 (98.4)	165 (93.2)	<0.001
Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, n (%)	1005 (71.8)	91 (51.4)	<0.001
β-blocker, n (%)	1156 (82.6)	137 (77.4)	0.109
Laboratory measurements
Platelet count, 10⁹/L	233 ± 68	219 ± 79	0.030
Alanine aminotransferase, IU/L	26 (17–40)	18 (13–32)	<0.001
Aspartate aminotransferase, IU/L	28 (20–51)	30 (19–57)	0.808
Gamma-glutamyl transpeptidase, IU/L	29 (20–44)	24 (18–41)	0.025
Albumin, g/L	42.4 ± 4.2	39.3 ± 4.6	<0.001
Albumin/globulin ratio, (mean ± SD)	1.8 ± 0.4	1.6 ± 0.4	<0.001
Serum creatinine, mg/dL	0.8 (0.7–1.0)	1.0 (0.7–1.8)	<0.001
Triglyceride, mmol/L	1.6 (1.1–2.4)	1.4 (0.9–1.9)	<0.001
High-density lipoprotein-cholesterol, mmol/L	1.0 ± 0.3	1.0 ± 0.3	0.820
Low-density lipoprotein-cholesterol, mmol/L	3.1 ± 1.0	2.8 ± 0.9	<0.001
LVEF <40, n (%)	101 (7.8)	29 (18.4)	<0.001
Liver fibrosis scores
NFS	−0.9 (−1.9–0.2)	0.3 (−0.6–1.4)	<0.001
FIB-8	1.6 (0.9–2.7)	2.7 (1.7–4.1)	<0.001
FIB-4	1.7 (1.1–2.7)	2.5 (1.6–4.1)	<0.001

Abbreviations: BMI, body mass index; CKD, chronic kidney disease; FIB-4, fibrosis-4; FIB-8, fibrosis-8; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NFS, non-alcoholic fatty liver disease fibrosis score; PCI, percutaneous coronary intervention.

Comparison of the Prognostic Value of the Liver Fibrosis Scores

To compare the prognostic value of the liver fibrosis scores, we utilized receiver operating characteristic analysis. The NFS showed a better prognostic value than the FIB-8 score (all-cause mortality: Harrell’s C-index 0.703 vs 0.671, P = .014; cardiac death: Harrell’s C-index 0.690 vs 0.669, NRI = 0.543, P < .001, IDI = 0.014, P < .001) and the FIB-4 score (all-cause mortality: Harrell’s C-index 0.703 vs 0.648, P < .001; cardiac death: Harrell’s C-index 0.690 vs 0.644, P = .016) in the overall population ((Figures 1A and C), and Table 2).

Table 2.

Comparison of the Prognostic Value of 3 Liver Fibrosis Scores in Patients With and Without Diabetes.

	Harrell’s C-index	p	NRI	p	IDI	P
Overall
All-cause mortality
NFS vs FIB-8	0.703 vs 0.671	0.014	0.598	<0.001	0.037	<0.001
NFS vs FIB-4	0.703 vs 0.648	<0.001	0.586	<0.001	0.041	<0.001
FIB-8 vs FIB-4	0.671 vs 0.648	0.018	0.352	<0.001	0.004	<0.001
Cardiac death
NFS vs FIB-8	0.690 vs 0.669	0.199	0.543	<0.001	0.014	<0.001
NFS vs FIB-4	0.690 vs 0.644	0.016	0.518	<0.001	0.016	<0.001
FIB-8 vs FIB-4	0.669 vs 0.644	0.039	0.424	<0.001	0.002	0.003
Patients with diabetes
All-cause mortality
NFS vs FIB-8	0.664 vs 0.630	0.112	0.528	<0.001	0.029	0.002
NFS vs FIB-4	0.664 vs 0.630	0.086	0.522	<0.001	0.030	0.001
FIB-8 vs FIB-4	0.630 vs 0.630	0.976	0.368	<0.001	0.001	0.014
Cardiac death
NFS vs FIB-8	0.652 vs 0.644	0.755	0.428	<0.001	0.005	0.282
NFS vs FIB-4	0.652 vs 0.634	0.475	0.445	<0.001	0.006	0.187
FIB-8 vs FIB-4	0.644 vs 0.634	0.406	0.577	<0.001	0.000	0.044
Patients without diabetes
All-cause mortality
NFS vs FIB-8	0.720 vs 0.674	0.004	0.643	<0.001	0.033	<0.001
NFS vs FIB-4	0.720 vs 0.674	0.004	0.700	<0.001	0.034	<0.001
FIB-8 vs FIB-4	0.674 vs 0.674	0.985	0.302	0.008	0.000	0.040
Cardiac death
NFS vs FIB-8	0.706 vs 0.662	0.024	0.534	<0.001	0.021	<0.001
NFS vs FIB-4	0.706 vs 0.660	0.022	0.581	<0.001	0.022	<0.001
FIB-8 vs FIB-4	0.662 vs 0.660	0.832	0.284	0.047	0.000	0.051

Abbreviations: FIB-4, fibrosis-4; FIB-8, fibrosis-8; NFS, non-alcoholic fatty liver disease fibrosis score; NRI, net reclassification improvement; IDI, integrated discrimination improvement.

As displayed in the time-dependent receiver operating characteristic curves, the NFS also had the highest area under the curve through all time points than the FIB-8 and the FIB-4 scores (Figures 1B and D). Besides, the NFS also gave the best prognostic value in the diabetic and non-diabetic populations (Table 2).

The Association of Liver Fibrosis Scores With Mortality

The Kaplan–Meier curves for all-cause mortality and cardiac death were shown in Supplementary Figure 2. As continuous or categorical variables, all the liver fibrosis scores were associated with all-cause mortality and cardiac death in the overall and non-diabetic populations after adjustment. However, in patients with diabetes, only the NFS was independently associated with all-cause mortality [hazard ratio (HR) = 1.241 per 1-point increment, 95% confidence interval (CI) 1.128–1.365, P < .001] and cardiac death (HR = 1.184, 95% CI 1.019–1.375, P = .027) when treated as a continuous variable. Besides, compared with the low plus intermediate levels of the NFS and the FIB-4 score, those with high levels of the NFS (HR = 2.696, 95% CI 1.701–4.272, P < .001 for all-cause mortality; HR= 2.857, 95% CI 1.544–5.285, P = .001 for cardiac death) and the FIB-4 score (HR = 1.854, 95% CI 1.185–2.900, P = .007 for all-cause mortality; HR = 1.980, 95% CI 1.099–3.569, P = .023 for cardiac death) were associated with a higher risk of mortality. However, the FIB-8 score was not significantly associated with all-cause mortality and cardiac death after multivariate adjustment (Table 3). The associations of the NFS with all-cause mortality and cardiac death were consistent in almost all the investigated subgroups (Figure 2). In the restricted cubic spline analysis, the NFS appeared to have linear associations with the risk of both all-cause mortality (p_{for non-linearity}=0.810) and cardiac death (p_{for non-linearity}=0.701) (Figures 3A and D). The risk of mortality also elevated with the increased levels of the FIB-8 and the FIB-4 scores but reached a plateau afterward (all p_{for non-linearity}<0.05) (Figures 3B, C, E, and F).

Table 3.

The association of 3 liver fibrosis scores with all-cause mortality and cardiac death in patients with and without diabetes.

Variable	Univariate analysis		Multivariate analysis*
Variable	HR (95% CI)	P	HR (95% CI)	P
All-cause mortality
Overall
Continuous (per 1-point increment)
NFS	1.309 (1.244–1.378)	<0.001	1.283 (1.195–1.378)	<0.001
FIB-8	1.065 (1.042–1.089)	<0.001	1.051 (1.019–1.084)	0.002
FIB-4	1.061 (1.036–1.087)	<0.001	1.046 (1.013–1.079)	0.006
Categorical
NFS >0.676 vs ≤0.676	3.126 (2.315–4.220)	<0.001	2.453 (1.749–3.440)	<0.001
FIB-8 >1.77 vs ≤1.77	2.482 (1.799–3.424)	<0.001	1.704 (1.190–2.438)	0.004
FIB-4 >2.67 vs ≤2.67	2.470 (1.839–3.317)	<0.001	1.986 (1.429–2.760)	<0.001
Patients with diabetes
Continuous (per 1-point increment)
NFS	1.243 (1.155–1.337)	<0.001	1.241 (1.128–1.365)	<0.001
FIB-8	1.047 (1.017–1.077)	0.002	1.041 (1.000–1.083)	0.050
FIB-4	1.045 (1.015–1.076)	0.003	1.040 (0.999–1.082)	0.058
Categorical
NFS >0.676 vs ≤0.676	2.850 (1.916–4.239)	<0.001	2.696 (1.701–4.272)	<0.001
FIB-8 >1.77 vs ≤1.77	1.587 (1.001–2.515)	0.049	1.177 (0.707–1.957)	0.531
FIB-4 >2.67 vs ≤2.67	2.332 (1.566–3.474)	<0.001	1.854 (1.185–2.900)	0.007
Patients without diabetes
Continuous (per 1-point increment)
NFS	1.493 (1.317–1.692)	<0.001	1.395 (1.190–1.636)	<0.001
FIB-8	1.112 (1.054–1.173)	<0.001	1.085 (1.011–1.166)	0.024
FIB-4	1.110 (1.051–1.171)	<0.001	1.081 (1.004–1.163)	0.039
Categorical
NFS >0.676 vs ≤0.676	2.836 (1.721–4.672)	<0.001	2.396 (1.349–4.254)	0.003
FIB-8 >1.77 vs ≤1.77	2.987 (1.890–4.721)	<0.001	2.164 (1.289–3.632)	0.004
FIB-4 >2.67 vs ≤2.67	2.727 (1.754–4.239)	<0.001	2.078 (1.256–3.440)	0.004
Cardiac death
Overall
Continuous (per 1-point increment)
NFS	1.278 (1.195–1.366)	<0.001	1.241 (1.123–1.371)	<0.001
FIB-8	1.065 (1.036–1.095)	<0.001	1.053 (1.008–1.100)	0.021
FIB-4	1.061 (1.029–1.093)	<0.001	1.049 (1.003–1.097)	0.038
Categorical
NFS >0.676 vs ≤0.676	3.056 (2.092–4.465)	<0.001	2.413 (1.546–3.767)	<0.001
FIB-8 >1.77 vs ≤1.77	2.859 (1.880–4.346)	<0.001	1.919 (1.184–3.110)	0.008
FIB-4 >2.67 vs ≤2.67	2.587 (1.783–3.754)	<0.001	2.057 (1.336–3.167)	0.001
Patients with diabetes
Continuous (per 1-point increment)
NFS	1.197 (1.079–1.328)	<0.001	1.184 (1.019–1.375)	0.027
FIB-8	1.045 (1.006–1.085)	0.024	1.042 (0.980–1.109)	0.191
FIB-4	1.042 (1.002–1.084)	0.038	1.040 (0.975–1.108)	0.232
Categorical
NFS >0.676 vs ≤0.676	2.678 (1.620–4.426)	<0.001	2.857 (1.544–5.285)	0.001
FIB-8 >1.77 vs ≤1.77	2.349 (1.223–4.512)	0.010	1.480 (0.722–3.032)	0.285
FIB-4 >2.67 vs ≤2.67	2.570 (1.555–4.249)	<0.001	1.980 (1.099–3.569)	0.023
Patients without diabetes
Continuous (per 1-point increment)
NFS	1.490 (1.274–1.741)	<0.001	1.445 (1.170–1.785)	0.001
FIB-8	1.116 (1.050–1.185)	<0.001	1.112 (1.023–1.208)	0.012
FIB-4	1.113 (1.047–1.184)	0.001	1.108 (1.018–1.206)	0.018
Categorical
NFS >0.676 vs ≤0.676	3.007 (1.622–5.576)	<0.001	2.746 (1.295–5.821)	0.008
FIB-8 >1.77 vs ≤1.77	2.821 (1.593–4.994)	<0.001	2.290 (1.157–4.533)	0.017
FIB-4 >2.67 vs ≤2.67	2.676 (1.537–4.659)	0.001	2.227 (1.145–4.332)	0.018

Abbreviations: CI, confidence interval; FIB-4, fibrosis-4; FIB-8, fibrosis-8; HR, hazard ratio; NFS, non-alcoholic fatty liver disease fibrosis score.

Adjusting for gender, body mass index, current smoking, hypertension, diabetes, history of myocardial infarction, prior coronary artery bypass grafting, chronic heart failure, statin, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, chronic kidney disease, left ventricular ejection fraction <40%, multivessel disease, triglyceride, and low-density lipoprotein cholesterol (NFS and FIB-8 were not adjusted for body mass index and diabetes).

The Incremental Effect of Adding the Liver Fibrosis Scores to the Model of Traditional Cardiovascular Risk Factors for Prognostic Prediction

The Gronnesby and Borgan tests showed that all the prognostic models incorporated traditional cardiovascular risk factors were well-calibrated (original model 1: Chi-Square = 9.469, P = .395 for all-cause mortality, Chi-Square = 12.701, P = .177 for cardiac death; original model 2: Chi-Square = 14.993, P = .091 for all-cause mortality, Chi-Square = 11.164, P = .265 for cardiac death).

The addition of the NFS to the model of traditional cardiovascular risk factors led to a significant improvement both for all-cause mortality (Harrell’s C-index: 0.771 vs 0.806, P < .001) and cardiac death (Harrell’s C-index: 0.771 vs 0.800, P = .014). Nevertheless, the prognostic value of the traditional cardiovascular risk factors did not improve by adding the FIB-4 score (Harrell’s C-index: 0.788 vs 0.796, P = .063 for all-cause mortality; 0.781 vs 0.790, P = .124 for cardiac death) (Table 4).

Table 4.

The Incremental Effects of Adding the NFS and the FIB-4 Score to Traditional Cardiovascular Risk Factors on Prognostic Prediction.

	Harrell’s C-index	P
All-cause mortality
Original model 1	0.771 (0.731−0.811)	Ref
Original model 1 + NFS	0.806 (0.771−0.841)	<0.001
Original model 2	0.788 (0.751−0.825)	Ref
Original model 2 + FIB-4	0.796 (0.761−0.832)	0.063
Cardiac death
Original model 1	0.771 (0.717−0.824)	Ref
Original model 1 + NFS	0.800 (0.752−0.847)	0.014
Original model 2	0.781 (0.730−0.831)	Ref
Original model 2 + FIB-4	0.790 (0.741−0.839)	0.124

Abbreviations: FIB-4, fibrosis-4; NFS, non-alcoholic fatty liver disease fibrosis score; NRI, net reclassification improvement; IDI, integrated discrimination improvement.

*The original model 1 included the following variables: gender, current smoking, hypertension, history of myocardial infarction, prior coronary artery bypass grafting, chronic heart failure, statin, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, chronic kidney disease, left ventricular ejection fraction <40%, multivessel disease, triglyceride, and low-density lipoprotein cholesterol (body mass index and diabetes, which included in the formula for calculating NFS were not adjusted). The original model 2 included the following variables: gender, body mass index, current smoking, diabetes, hypertension, history of myocardial infarction, prior coronary artery bypass grafting, chronic heart failure, statin, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, chronic kidney disease, left ventricular ejection fraction <40%, multivessel disease, triglyceride, and low-density lipoprotein cholesterol.

Discussion

In this retrospective study, we compared the predictive value of the NFS, the FIB-8 score, and the FIB-4 score for mortality. Several major findings of our study were of interest: (1) the NFS showed a better prognostic value than the FIB-8 and the FIB-4 scores in both patients with and without diabetes. (2) There were linear associations between the NFS and the risk of mortality; however, the risk of mortality reached a plateau afterward with the rising of the FIB-8 and the FIB-4 scores. (3) The NFS provided additional prognostic information beyond the traditional cardiovascular risk factors, this was not the case for the FIB-4 score. Taken together, the NFS may be preferable for predicting long-term mortality than the FIB-8 and the FIB-4 scores in patients with AMI regardless of the presence or absence of diabetes.

Patients with coronary artery disease have a higher prevalence of diabetes, obesity, and metabolic syndrome.²² In addition to the direct involvement of the cardiovascular system, patients with these risk factors have a substantially increased risk of liver diseases including NAFLD and liver fibrosis.²² Liver diseases, diabetes, and metabolic syndrome may have synergistic relationships and lead to a synergistic deterioration of patient prognosis.²² Therefore, identifying those with diabetes and liver fibrosis and giving appropriate treatment could potentially reduce the risk of cardiovascular events.²³

Recently, non-invasive elastographic parameters of liver fibrosis have been found to be associated with an increased risk of developing AMI in patients with type 2 diabetes.²⁴ Furthermore, the NFS and the FIB-4 score, calculated with clinical and laboratory parameters, have been considered promising alternatives to liver biopsy for evaluating the severity of liver fibrosis^9,10 and were associated with an increased risk of non-fatal myocardial infarction in patients with one or more cardiovascular risk factors.²⁵ Several studies significantly extend the application scope of the NFS and the FIB-4 score to prognostic prediction in patients with coronary artery disease. Tang et al. involving patients with acute coronary syndrome who underwent PCI detected associations of the NFS¹³ and the FIB-4 score¹⁵ with major adverse cardiac and cerebrovascular events, suggesting that both the NFS and the FIB-4 score could serve as prognostic tools in the acute coronary syndrome population. A prospective population-based study in patients with stable coronary artery disease also determined that patients with higher scores of NFS or FIB-4 showed a significantly increased risk of cardiovascular death, non-fatal MI, and ischemic stroke than those with lower scores.¹⁴ Besides, only the NFS but not the FIB-4 score was independently associated with the secondary endpoint (incorporated cardiovascular death, non-fatal MI, ischemic stroke, unplanned revascularization, and hospitalized unstable angina).¹⁴ However, no studies so far have directly compared the predictive utility of these scores for long-term survival in patients with AMI undergoing PCI.

Previous studies have compared the diagnostic accuracy of the NFS, the FIB-8, and the FIB-4 scores for liver fibrosis. Bertot et al. retrospectively enrolled 284 patients diagnosed with NAFLD and discovered that the FIB-4 score achieved better accuracy than the NFS for predicting advanced liver fibrosis in both non-diabetic and diabetic populations.¹⁶ Similarly, a cross-sectional study in patients with type 2 diabetes who screened for non-alcoholic steatohepatitis also demonstrated that the FIB-4 score performed slightly better than the NFS for diagnosing advanced liver fibrosis.¹⁷ In light of these findings, the current clinical guideline preliminarily suggested that the FIB-4 score could be preferred for evaluating the extent of liver fibrosis in the diabetic population.⁹ However, the above two studies were based on Caucasians and Americans. A Korean NAFLD cohort study reported an opposite conclusion that the NFS showed the highest predictive value for advanced liver fibrosis.¹⁸ The prevalence of diabetes, the mean age, and the BMI of the Korean cohort were lower than those of the above-mentioned Western cohort. Furthermore, patients in the Korean cohort also had a lower incidence of advanced liver fibrosis than the above Western cohort (18.3 vs 26.4%), which could affect the results.^16,18 Thus, whether the FIB-4 score is better than the NFS for evaluating liver fibrosis remains controversial.¹⁰ Besides, a recent study involving Asian patients with biopsy-proven NAFLD, who have similar age distributions as well as a similar prevalence of diabetes compared with the above Western cohort, indicated that the FIB-8 score performed better than the NFS and the FIB-4 score for predicting significant fibrosis.¹⁹

It is still unknown which liver fibrosis scores perform better for prognostic prediction in Asians with AMI. In the present study, we found that the NFS showed the highest predictive value for all-cause mortality and cardiac death than the FIB-8 and the FIB-4 scores. These findings were consistent in patients with or without diabetes. Compared with the studies based on the NAFLD population,^16,19 our study was in an AMI population, which may have a lower severity of liver fibrosis. Besides, the BMI, the proportion of diabetes, and the liver enzyme levels of our study population were lower than that of studies based on the NAFLD population,^16,19 which may decrease the association between liver fibrosis scores and advanced liver fibrosis. Despite the liver fibrosis associated with mortality and cardiovascular events,^6–8,26 multiple factors could affect the prognosis of patients with AMI undergoing PCI. Therefore, the extent of liver fibrosis might play a partial role. Variables included in the formulas of the NFS and the FIB-8 score, such as BMI, diabetes, and albumin, were associated with the prognosis in patients undergoing PCI.^27–29 Therefore, the NFS and the FIB-8 score may provide more prognostic information than the FIB-4 score, leading to a better prognostic performance in our study population. In a prospective cohort study exploring the association between liver fibrosis scores and cardiovascular outcomes in patients with previous MI, Cao et al. revealed that when adding the liver fibrosis scores to a prognostic prediction model, the NFS provided a higher increment in C-statistic, NRI, and IDI than the FIB-4 score for both all-cause mortality and cardiovascular mortality.³⁰ Another study of a heart failure with preserved ejection fraction population also demonstrated that the NFS had a greater predictive value for all-cause death, cardiovascular death, and heart failure hospitalization compared with the FIB-4 score.³¹ These also corroborated our findings that the NFS performed better than the FIB-4 score for long-term prognostic prediction. Moreover, gamma-glutamyl transpeptidase is a uniquely incorporated variable in the FIB-8 score,¹⁹ which was associated with adverse cardiovascular events and mortality.³² Nevertheless, the concentration of gamma-glutamyl transpeptidase in our cohort was considerably lower than in that study (28 vs 63 IU/L),¹⁹ which may reduce the prognostic value of the FIB-8 score in our study. Future studies are required to investigate and validate our findings.

The mechanisms underlying the relationship between liver fibrosis scores and mortality remain poorly defined. As hepatic fibrosis develops, increased levels of procoagulant factors such as factor VIII and decreased levels of anticoagulant factors such as protein C might create a hypercoagulable and atherothrombotic state, with an elevated risk of cardiovascular events and then ultimately result in death.³³ Besides, the circulating levels of systemic inflammatory biomarkers, such as interleukin-6 and tumor necrosis factor-α, were positively associated with the severity of liver fibrosis.^34,35 They exacerbate endothelial dysfunction, promote plaque formation, and aggravate plaque instability, which further leads to adverse cardiovascular outcomes.³⁶ Although liver fibrosis scores were originally developed for detecting advanced fibrosis, variables included in the formula may also reflect metabolic disorders and systemic inflammatory state, which may contribute to an increase in cardiovascular risk.³⁷ Simon et al.³⁷ discovered that patients with higher NFS were associated with a 30% increased risk of recurrent major cardiovascular events compared to those with lower NFS. Furthermore, patients with higher NFS were more likely to benefit from the combination of ezetimibe with statin.³⁷ Therefore, the NFS may serve as a valuable tool for identifying residual risk and guiding optimal treatment in clinical practice. The association between liver fibrosis, diabetes, and mortality in patients with AMI was shown in Supplementary Figure 3. Altogether, physicians always focus on the traditional cardiovascular risk factors in clinical practice, and liver health receives comparably little attention. In the present study, the NFS was capable of adding prognostic value beyond the traditional cardiovascular risk factors, suggesting that physicians are required to pay equal attention to the extent of liver fibrosis while focusing on cardiovascular risk factors. Therefore, our findings may contribute to increasing physicians’ awareness regarding liver health and provide additional evidence for improving liver function thereby improving prognosis in patients with AMI who underwent PCI.

Limitations

First, this study is a retrospective cohort study with inherent limitations, such as cause-and-effect inferences that cannot be made. Additional prospective, large sample-size studies to validate our findings are warranted. Second, there is a lack of liver biopsy, transient elastography, and magnetic resonance elastography, which could help confirm the extent of liver fibrosis. Although the gold standard for assessing the stage of liver fibrosis is a liver biopsy, this procedure is invasive, expensive, and associated with potential complications making it impossible for routine screening. Other noninvasive methods such as transient elastography and magnetic resonance elastography can also be used to assess the severity of liver fibrosis.³⁸ However, these assays are not routinely used and are not appropriate for screening in clinical practice due to the feasibility and cost limitations. Third, although patients with excessive alcohol consumption and chronic liver diseases, such as viral hepatitis and drug-induced liver injury, have been excluded from our analysis, undiagnosed liver diseases might introduce inevitable confounding effects. Besides, we were unable to report the incidence of NAFLD in our study population because abdominal imaging examinations such as abdominal ultrasonography and magnetic resonance were not performed among most of the patients in our study. Lastly, due to the retrospective nature of our study, we did not collect data on liver fibrosis scores during the follow-up period and only calculated their baseline levels. The prognostic impact of dynamic changes on liver fibrosis scores was not evaluated. However, to our knowledge, there are no studies evaluating the prognostic impact of changes in liver fibrosis scores among patients with AMI who underwent PCI. Future prospective studies with serial liver fibrosis scores can validate our findings.

Conclusions

Among patients with AMI who underwent PCI, the NFS showed a better prognostic value than the FIB-8 and the FIB-4 scores. This finding was consistent for both diabetic and non-diabetic populations. Moreover, only the NFS but not the FIB-4 score provided added prognostic value beyond the traditional cardiovascular risk factors, which might be more appropriate for evaluating the risk of mortality in patients with AMI.

Supplemental Material

Supplemental Material - A Comparison of the Prognostic Value of Liver Fibrosis Scores in Acute Myocardial Infarction Patients With and Without Type 2 Diabetes

Supplemental Material for A Comparison of the Prognostic Value of Liver Fibrosis Scores in Acute Myocardial Infarction Patients With and Without Type 2 Diabetes by Hao-ming He, MD, Shu-wen Zheng, MD, Yi-nong Chen, MD, Long-yang Zhu, MD, Zhe Wang, MD, Si-qi Jiao, MD, Fu-rong Yang, MD, and Yi-hong Sun, MD, PhD in Angiology

Footnotes

Acknowledgments

The authors are grateful to the free drawing support platform “Figdraw” ().

Author Contributions

All authors contributed to: (1) substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, and, (3) final approval of the version to be published.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Capital Health Research and Development of Special Fund (2020-2-4065) and the National High-Level Hospital Clinical Research Fund (2022-NHLHCRF-PY-19). The funding sources had no role in the study design, data collection, or writing of this manuscript.

ORCID iD

Yi-hong Sun

Data Availability Statement

The generated and analyzed during the current study are not publicly available due to confidentiality reasons but are available from the corresponding author upon scientific review and approval of the request by the Study’s Scientific Committee.

Supplemental Material

Supplemental material for this article is available online.

References

Alfonso

Gonzalo

Rivero

Escaned

. The year in cardiovascular medicine 2020: interventional cardiology. Eur Heart J. 2021; 42: 985–1003.

Reed

Rossi

Cannon

. Acute myocardial infarction. Lancet. 2017; 389: 197–210.

Rinella

. Nonalcoholic fatty liver disease. JAMA. 2015; 313: 73–226.

Wong

Yeung

Fung Chan Chang , et al. Long‐term clinical outcomes after fatty liver screening in patients undergoing coronary angiogram: a prospective cohort study. Hepatology. 2016; 63: 754–763.

Cohen

Horton

Hobbs

. Human fatty liver disease: old questions and new insights. Science. 2011;332:1519–1523.

Dulai

Singh

Patel

Soni Prokop Younossi , et al. Increased risk of mortality by fibrosis stage in nonalcoholic fatty liver disease: systematic review and meta‐analysis. Hepatology. 2017; 65: 1557–1565.

Cardoso

CRL

Villela-Nogueira

Leite

Salles

. Prognostic impact of liver fibrosis and steatosis by transient elastography for cardiovascular and mortality outcomes in individuals with nonalcoholic fatty liver disease and type 2 diabetes: the Rio de Janeiro Cohort Study. Cardiovasc Diabetol. 2021; 20: 193.

Ostovaneh

Ambale-Venkatesh

Fuji

Bakhshi Shah Murthy , et al. Association of liver fibrosis with cardiovascular diseases in the general population: the multi-ethnic study of atherosclerosis (MESA). Circulation. Cardiovascular imaging. 2018; 11: e007241.

EASL Clinical Practice Guidelines on non-invasive tests for evaluation of liver disease severity and prognosis - 2021 update. J Hepatol. 2021; 75(3): 659–689. doi:10.1016/j.jhep.2021.05.025.

10.

Kang

Lee

Yoo

Cho Kim Lee , et al. KASL clinical practice guidelines: management of nonalcoholic fatty liver disease. Clin Mol Hepatol. 2021; 27: 363–401.

11.

Angulo

Hui

Marchesini

Bugianesi George Farrell , et al The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology. 2007; 45: 846–854.

12.

Sterling

Lissen

Clumeck

Sola Correa Montaner , et al. Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology. 2006; 43: 1317–1325.

13.

Liu

Song

Wang

Zhang Yang Wen , et al. Association between liver fibrosis and thrombotic or bleeding events in acute coronary syndrome patients. Thromb J. 2022; 20: 82.

14.

Liu

Cao

Jin

Hua Li Guo , et al. Liver fibrosis scoring systems as novel tools for predicting cardiovascular outcomes in patients following elective percutaneous coronary intervention. J Am Heart Assoc. 2021; 10: e018869.

15.

Song

Liu

Chen

Yang Wen Gao , et al. Fibrosis-4 stage of liver fibrosis predicts cardiovascular outcomes in acute coronary syndrome patients with and without type 2 diabetes mellitus. Diabetes Res Clin Pract. 2023; 195: 110206.

16.

Bertot

Jeffrey

de Boer

MacQuillan Garas Chin , et al. Diabetes impacts prediction of cirrhosis and prognosis by non-invasive fibrosis models in non-alcoholic fatty liver disease. Liver Int. 2018; 38: 1793–1802.

17.

Bril

McPhaul

Caulfield

Clark Soldevilla-Pico Firpi-Morell , et al. Performance of plasma biomarkers and diagnostic panels for nonalcoholic steatohepatitis and advanced fibrosis in patients with type 2 diabetes. Diabetes Care. 2020; 43: 290–297.

18.

Jun

Kim

Park

Jin Lee Lee , et al. External validation of the non-alcoholic fatty liver disease fibrosis score for assessing advanced fibrosis in Korean patients. J Gastroenterol Hepatol. 2017; 32: 1094–1099.

19.

Prasoppokakorn

Chan

Wong

Pitisuttithum Mahadeva Nik Mustapha , et al. Validation model of fibrosis-8 index score to predict significant fibrosis among patients with nonalcoholic fatty liver disease. World J Gastroenterol. 2022; 28: 1563–1573.

20.

Thygesen

Alpert

Jaffe

Chaitman Bax Morrow , et al. Fourth universal definition of myocardial infarction (2018). J Am Coll Cardiol. 2018; 72: 2231–2264.

21.

Inker

Schmid

Tighiouart

Eckfeldt Feldman Greene , et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012; 367: 20–29.

22.

Said

Mukherjee

Whayne

. Interrelationships with metabolic syndrome, obesity and cardiovascular risk. Curr Vasc Pharmacol. 2016; 14: 415–425.

23.

Paolisso

Bergamaschi

Gragnano

Gallinoro Cesaro Sardu , et al. Outcomes in diabetic patients treated with SGLT2-Inhibitors with acute myocardial infarction undergoing PCI: The SGLT2-I AMI PROTECT Registry. Pharmacol Res. 2023; 187: 106597.

24.

Mikolasevic

Domislovic

Ruzic

Hauser Rahelic Klobucar-Majanovic , et al. Elastographic parameters of liver steatosis and fibrosis predict independently the risk of incident chronic kidney disease and acute myocardial infarction in patients with type 2 diabetes mellitus. J Diabetes Complicat. 2022; 36: 108226.

25.

Riccio

Mazzanti

Vero

, et al. Liver fibrosis is associated with an increased risk of non-fatal myocardial infarction. Eur J Clin Invest. 2023: e14061.

26.

Taylor

Bayliss

Hagström Nasr Schattenberg , et al. Association between fibrosis stage and outcomes of patients with nonalcoholic fatty liver disease: a systematic review and meta-analysis. Gastroenterology. 2020; 158: 1611–1625.

27.

Mei

Zhou

Chen

. Body mass index and all-cause mortality in patients with percutaneous coronary intervention: a dose-response meta-analysis of obesity paradox. Obes Rev. 2021; 22: e13107.

28.

Wilson

Vakili

Sherman

Sanborn

Brown

. Effect of diabetes on long-term mortality following contemporary percutaneous coronary intervention. Diabetes Care. 2004; 27: 1137–1142.

29.

Manolis

Melita

Mikhailidis

Manolis

. Low serum albumin: a neglected predictor in patients with cardiovascular disease. Eur J Intern Med. 2022; 102: 24–39.

30.

Cao

Zhang

Jin Liu Zhang , et al. Impact of liver fibrosis score on prognosis in patients with previous myocardial infarction: A prospective cohort study. Liver Int. 2021; 41: 1294–1304.

31.

Cao

Guo

Dong

Liu

Zhu

. Comparison of liver fibrosis scores for predicting mortality and morbidity in heart failure with preserved ejection fraction. ESC Heart Failure. 2023; 10: 1771–1780.

32.

Ruttmann

Brant

Concin

Diem

Rapp

Ulmer

. γ-Glutamyltransferase as a risk factor for cardiovascular disease mortality. Circulation. 2005; 112: 2130–2137.

33.

Tripodi

Fracanzani

Primignani

Chantarangkul Clerici Mannucci , et al. Procoagulant imbalance in patients with non-alcoholic fatty liver disease. J Hepatol. 2014; 61: 148–154.

34.

Wieckowska

Papouchado

Lopez

Zein

Feldstein

. Increased hepatic and circulating interleukin-6 levels in human nonalcoholic steatohepatitis. Am J Gastroenterol. 2008; 103: 1372–1379.

35.

du Plessis

van Pelt

Korf

Mathieu van der Schueren Lannoo , et al. Association of adipose tissue inflammation with histologic severity of nonalcoholic fatty liver disease. Gastroenterology. 2015; 149: 635–648.

36.

Kofler

Nickel

Weis

. Role of cytokines in cardiovascular diseases: a focus on endothelial responses to inflammation. Clin Sci 2005; 108: 205–213.

37.

Simon

Corey

Cannon

Blazing Park O'Donoghue , et al. The nonalcoholic fatty liver disease (NAFLD) fibrosis score, cardiovascular risk stratification and a strategy for secondary prevention with ezetimibe. Int J Cardiol. 2018; 270: 245–252.

38.

Imajo

Honda

Kobayashi

Nagai Ozaki Iwaki , et al. Direct Comparison of US and MR Elastography for staging liver fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol. 2022; 20: 908–917.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB