Relationship between non-invasively detected liver fibrosis and in-hospital outcomes in patients with acute coronary syndrome undergoing PCI

Background Patients with acute coronary syndrome (ACS) undergoing percutaneous coronary intervention (PCI) still experience a high rate of in-hospital complications. Liver fibrosis (LF) is a risk factor for mortality in the general population. We investigated whether the presence of LF detected by the validated fibrosis 4 (FIB-4) score may indicate ACS patients at higher risk of poor outcome. Methods In the prospective ongoing REAl-world observationaL rEgistry of Acute Coronary Syndrome (REALE-ACS), LF was defined by a FIB-4 score > 3.25. We repeated the analysis using an APRI score > 0.7. The primary endpoint was in-hospital adverse events (AEs) including a composite of in-hospital cardiogenic shock, PEA/asystole, acute pulmonary edema and death. Results A total of 469 consecutive ACS consecutive patients were enrolled. Overall, 21.1% of patients had a FIB-4 score > 3.25. Patients with LF were older, less frequently on P2Y12 inhibitors (p = 0.021) and admitted with higher serum levels of white blood cells (p < 0.001), neutrophils to lymphocytes ratio (p < 0.001), C-reactive protein (p = 0.013), hs-TnT (p < 0.001), creatine-kinase MB (p < 0.001), D-Dimer levels (p < 0.001). STEMI presentation and higher Killip class/GRACE score were more common in the LF group (p < 0.001). 71 patients experienced 110 AEs. At the multivariate analysis including clinical and laboratory risk factors, FIB-4 > 3.25 (OR 3.1, 95%CI 1.4–6.9), admission left ventricular ejection fraction% below median (OR 9.2, 95%CI 3.9–21.7) and Killip class ≥ II (OR 6.3, 95%CI 2.2–18.4) were the strongest independent predictors of in-hospital AEs. Similar results were obtained using the APRI score. Conclusion LF detected by FIB-4 score > 3.25 was associated with more severe ACS presentation and worse in-hospital AEs irrespective of clinical and laboratory variables. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00392-022-02078-z.


Introduction
Patients with acute coronary syndrome (ACS) still suffer from high in-hospital morbidity/mortality and long-term poor prognosis, including recurrence of myocardial infarction (MI) in about 10% of cases [1][2][3], despite advances in revascularization procedures techniques and current best medical therapy. This residual high cardiovascular risk may be driven by several factors including poor adherence to prescription [4], not reaching therapeutic target for comorbidities, or by the presence of non-cardiovascular risk factors [5], such as chronic kidney and liver disease.
In addition, a proportion of patients may experience early complications after ACS, including cardiogenic shock, arrhythmia, and in-hospital death. However, the clinical and biochemical characteristics of patients at risk of early complications after ACS are not well established, and so far, the use of common biomarkers of disease severity, such as high-sensitive troponin, did not lead to an improvement in clinical management of these patients [6].
In the past decades, given the impossibility of performing liver biopsies on large scale, different non-invasive biochemical markers have been proposed to identify patients with advanced liver damage. Among others, the AST-to-platelet ratio index (APRI) and the Fibrosis 4 (FIB-4) score, have been shown to be reliable tools for predicting liver fibrosis (LF) [7,8].
Previous evidence also showed that a high FIB-4 was associated with an increased rate of long-term cardiovascular events in patients with cardiovascular disease, such as those affected by atrial fibrillation (AF) [9]. In addition, recent studies highlighted the association of LF with the 1 3 severity of coronary artery disease (CAD) and advanced high-risk coronary plaque [10,11].
Nevertheless, the impact of LF on in-hospital and shortterm complications in patients admitted with ACS undergoing percutaneous coronary intervention (PCI) has been poorly investigated.
Our aim was to study the characteristics of patients hospitalized for ACS presenting with high FIB-4, and to investigate the relationship between LF and in-hospital outcomes.

Methods
The REAl-world observationaL rEgistry of Acute Coronary Syndrome (REALE-ACS) is an ongoing multicentric registry collecting data on characteristics, management and outcomes of consecutive patients admitted for ACS at the Department of Clinical Internal, Anesthesiologic, and Cardiovascular Sciences, Sapienza University of Rome, Rome, Italy and Emergency Medicine Unit Department (from January 2016).
Patients with age < 18 years were excluded. Diagnosis of non-ST-elevation ACS (i.e., unstable angina [UA] and non-ST-elevation myocardial infarction [NSTEMI]) and STelevation myocardial infarction (STEMI) were made according to the latest European Society of Cardiology guidelines [6,12].
At baseline, demographic characteristics and clinical information of each study patient were recorded as following: age, sex, anthropometric data, cardiovascular risk factors (hypertension, diabetes mellitus, smoking (former + current), hypertriglyceridemia, hypercholesterolemia), previous history of CAD, heart failure, ischemic stroke, peripheral artery disease and Global Registry of Acute Coronary Events (GRACE) score. Standard 12-lead ECGs were collected at admission and during in-hospital stay.

Clinical outcomes
The primary endpoint was a composite of in-hospital adverse events (AEs) including cardiogenic shock, pulseless electrical activity (PEA)/asystole, acute pulmonary edema and death. For the analysis, only the first event occurrence was considered.
The study was performed according to the Declaration of Helsinki.

Statistical analysis
Categorical variables were reported as counts and percentage. Continuous variables were expressed as mean and standard deviation and compared by Student t test. χ 2 test was used to compare proportions. A first descriptive analysis according to presence of high FIB-4 was performed. We also described clinical and biochemical characteristics associated with inhospital AEs. A multivariable logistic regression analysis was used to calculate the relative odds ratio (OR) and 95% confidence interval (95%CI) for each factor associated with in-hospital AEs.
In the multivariable model, only clinical variables with a p value < 0.100 at univariable analysis were included. In this model, we tested for collinearity to avoid an over adjustment of the model given the limited number of events. The variables entered in the multivariable model were FIB-4 > 3.25, below median of left ventricular ejection fraction (LVEF) value obtained for entry echocardiography, Killip Class ≥ 2, MDRD < 49 mL/min (value obtained from receiver operating characteristic [ROC] curve analysis to optimize the cut-off). We also performed a subgroup analysis on patients with albumin, CRP, and D-Dimer available values.
Only p values < 0.05 were considered as statistically significant. All tests were two tailed and analyses were performed using computer software packages (SPSS-25, SPSS Inc. and MedCalc).

Clinical and laboratory findings
A total of 469 consecutive ACS patients were included. Among them, 12 patients had a history of chronic liver disease (6 patients viral hepatitis, 4 alcoholic disorder and 2 both). Clinical characteristics of patients with or without a high FIB-4 score are summarized in Table 1. Overall, 21.1% of patients had a high FIB-4 score. Patients with LF were older (p < 0.001), less frequently overweight (p = 0.026) and more frequently affected by permanent AF (p = 0.033).
Supplementary Table 1 shows similar differences between patients with or without a high APRI score.

ACS presentation and angiographic features of patients with or without LF
Presentation of ACS was STEMI in 44.3%, NSTEMI in 29% and UA in 26.7% (Table 2). UA was more common in non-LF patients (33% vs 3%, p < 0.001), whilst STEMI was more common in the LF group (71.7% vs 37%, p < 0.001). Patients with LF had more severe ACS presentation, with a higher Killip class (34.4% vs 13.3%, p < 0.001) and GRACE score upon admission (161.82 ± 40.92 vs 126.12 ± 36.55, p = 0.001; GRACE > 140 71.7% vs 63.5%, p < 0.001). No differences between the two groups were reported when analyzing angiographic data ( Table 2). ACS and angiographic characteristics between patients with high or low APRI score were consistent with those obtained by using FIB-4 score (Supplementary Table 2).

In-hospital adverse events
Overall, 71 patients experienced 110 AEs. The number and type of AEs according to presence or absence of a high FIB-4 score are reported in Table 3. Table 4 reports clinical events occurring during the inhospital staying according to the presence or absence of LF. Noteworthy, AEs were significantly more frequent in patients with LF than in those without (26.3% vs 12.2%, respectively, p = 0.001), with this difference being more evident in STEMI patients ( Supplementary Fig. 1). In particular, cardiogenic shock (17.2% vs 7.3%, p = 0.006) and PEA/ asystole (8.1% vs 3.0%, p = 0.039) were more frequent in LF patients, while cardiac deaths and acute pulmonary edema did not differ between the two groups. Table 4 shows characteristics of patients experiencing or not AEs during hospitalization. Patients suffering from in-hospital AEs were characterized by an older age, longer in-hospital stay, lower values of LVEF, higher GRACE score and had more frequently STEMI and cardiogenic shock presentation. At laboratory analysis on admission, they also showed more frequently hypoalbuminemia and higher serum values of CRP, AST, WBC count, neutrophils/lymphocytes ratio, D-Dimer, glucose, hs-TnT and CK-MB (Table 4). Table 5 shows univariable and multivariable odds ratio (OR) and 95% confidence interval (95% CI) of factors associated with in-hospital AEs.
When we repeated the analysis using the negative value of FIB-4 < 1.45 (rule out criterion for the presence of LF), we found an inverse association between low FIB-4 score and AEs at univariable OR: 0.45, 95%CI 0.24-0.86, p = 0.015, a marginally significant value at multivariable analysis (using Model A variables) with OR 0.55, 95%CI 0.30-1.01, p = 0.055).

Discussion
The main findings of our prospective cohort study were that non-invasive diagnosis of LF was achieved in one out of five ACS patients, identifying patients with severe MI and at higher risk of in-hospital adverse events independent of traditional clinical and laboratory risk factors.
In our study on 469 patients with ACS, we found a 21.1% of LF prevalence as defined by a FIB-4 score > 3. 25. In a study including 3263 patients with CAD in China, 1035 (31.7%) had a high FIB-4 defined by a value > 2.67 [14].
We found that patients with LF had a more severe ACS presentation. First, they suffered more frequently from hypoalbuminemia, thrombocytopenia and highest levels of inflammatory/severity markers such as hs-TnT, CRP,   [11,20]. We found a similar proportion of multivessel disease between patients with and without high FIB-4, that may be probably due to the high percentage of STEMI in the former group.
Interestingly, LF patients were less frequently on DAPT therapy, especially P2Y12 inhibitors, on admission. The association between antiplatelet therapy and LF has been investigated in the past and a protective association between the use of antiplatelet agents and the occurrence of liver fibrosis has been demonstrated in a prospective cohort study of patients at high risk of cardiovascular events [21]. Previous data have shown a pathogenic role of platelets in the pathogenesis of LF, and antiplatelet agents was found to be effective in attenuating liver steatosis in preclinical studies [22,23].
We also investigated the association between LF and inhospital outcomes. We found that LF as detected by an easy inexpensive non-invasive score calculated with routine laboratory data was independently associated with in-hospital AEs independently from traditional clinical and laboratory risk factors. The association between LF and mortality has been investigated only in one previous study performed in China using a lower FIB-4 cut-off, in which Chen et al. found that higher LF scores were associated with increased risks of all-cause and cardiovascular mortality among CAD patients followed for 7.5 years [14].
The pathophysiology linking between LF and CAD is unclear but may rely on different mechanisms. LF and eventually liver steatosis have been associated with increased levels of pro-inflammatory cytokines, such as tumor necrosis factor α, interleukin 6, and CRP, biomarkers well known to have an impact on CAD and cardiovascular prognosis [24]. As well, advanced LF has been shown to be associated with endothelial dysfunction increasing atherosclerosis [25]. Previous studies have shown an association between LF assessed with non-invasive markers and imaging indicators of atherosclerosis such as carotid mean intimal thickness and coronary calcium score [26,27].
Novel pathogenetic insights come from the study of gut and gut-derived products. Thus, an increased liver localization of gut-derived lipopolysaccharides (LPS) has been found in patients with liver disease, associated with increased liver inflammation [28].
Additionally, a growing body of evidence suggest that LPS may contribute to increase cardiovascular risk favoring platelet activation [29,30] and that LPS may found into human atherosclerotic plaque potentially contributing to plaque vulnerability and in turn favoring plaque rupture [31,32]. Indeed, in ACS patients with LF, LPS may contribute to develop a more severe coronary disease. This aspect deserves further study.
Our study had some limitations. First, this was a singlecenter study and results must be confirmed in larger studies. Although liver biopsy is the best standard to confirm/exclude LF, it is invasive, expensive, prone to sampling error, and can cause rare but potentially life-threatening complications. Nowadays, it has been almost entirely replaced by non-invasive methods that measure liver stiffness, such as transient elastography, or biochemical markers and scoring systems. In terms of cost-effectiveness, considering the high prevalence of LF and CAD, biochemical markers, such as the APRI score and the FIB-4 score, recently shown to be