Association of Subclinical Liver Fibrosis With Death in Patients With Coronary Artery Disease: A Post Hoc Analysis of the ISCHEMIA Trial

Abstract

Background

The fibrosis‐4 index (FIB‐4) score, a noninvasive marker of subclinical liver fibrosis, has shown prognostic utility in general surgical populations. Current risk assessment models for patients with coronary artery disease undergoing percutaneous coronary intervention or coronary artery bypass grafting do not account for liver dysfunction apart from overt liver cirrhosis. We analyzed the distribution of the baseline FIB‐4 score and its association with all‐cause death in patients with coronary artery disease using data from the International Study of Comparative Health Effectiveness With Medical and Invasive Approaches (ISCHEMIA) trial.

Methods

The baseline FIB‐4 score was calculated for all ISCHEMIA randomized participants with laboratory data (platelet count, aspartate aminotransferase, and alanine aminotransferase). The primary outcome was the association between baseline FIB‐4 and all‐cause death. Secondary outcomes were cardiovascular death, heart failure, myocardial infarction, and stroke. Multivariable Cox regression was performed adjusting for key risk factors.

Results

The FIB‐4 score was calculated for 3735 participants. Baseline FIB‐4 score was significantly associated with an increased risk of all‐cause (hazard ratio [HR], 1.19 [95% CI, 1.07–1.32]; P=0.001) and cardiovascular death (HR, 1.19 [95% CI, 1.04–1.36]; P=0.011). This association was consistent across the overall population and within subgroups of patients treated with percutaneous coronary intervention, coronary artery bypass grafting, and medical therapy. There was no significant association regarding heart failure, myocardial infarction, and stroke.

Conclusions

The FIB‐4 score may be a significant predictor of death in patients with coronary artery disease. Preprocedural hepatic assessment should be considered to stratify risk in patients undergoing invasive cardiac procedures.

Nonstandard Abbreviations and Acronyms

FIB‐4

fibrosis‐4 index

ISCHEMIA

International Study of Comparative Health Effectiveness With Medical and Invasive Approaches

Clinical Perspective.

What Is New?

• The fibrosis‐4 index score, a marker of subclinical liver fibrosis, is significantly associated with all‐cause and cardiovascular death in patients with coronary artery disease enrolled in the ISCHEMIA trial.

• This association persists across treatment strategies, including percutaneous coronary intervention, coronary artery bypass grafting, and medical therapy.

What Are the Clinical Implications?

• Incorporating hepatic function assessment into preprocedural risk stratification may improve prognostication and guide clinical decision‐making in coronary artery disease.

Patients with liver cirrhosis, representing advanced fibrosis, are at elevated risk ¹ of adverse clinical events following cardiac interventions. Current risk assessment tools, such as the Society of Thoracic Surgeons score and the EuroScore, do not incorporate liver fibrosis as a risk factor, except in the case of overt liver cirrhosis. However, recent retrospective data from a single‐center study of over 5000 patients suggest that even subclinical liver fibrosis may be associated with increased surgical risk in patients undergoing coronary artery bypass grafting (CABG). ² This raises the possibility that milder degrees of liver fibrosis, which often go undetected, could also have prognostic implications in this setting.

In the context of percutaneous coronary intervention (PCI), not just end‐stage liver disease ³ but also subclinical signs of hepatic fibrosis ⁴ are associated with poorer outcomes. Even without overt liver failure, hepatic fibrosis can contribute to systemic inflammation, endothelial dysfunction, and altered coagulation, all of which increase the risks of thrombosis, bleeding, and complications during PCI. ⁵ , ⁶

The fibrosis‐4 index (FIB‐4) score is a noninvasive score to assess liver fibrosis, which is calculated using 4 variables: age, platelet count, aspartate aminotransferase, and alanine aminotransferase levels. ⁷ Initially developed to predict the necessity of liver biopsy in patients with hepatitis, ⁷ the FIB‐4 score is now widely used to guide clinical decision‐making and monitor disease progression in various liver disorders. ⁸ Moreover, the FIB‐4 score has demonstrated prognostic utility in populations without overt liver diseases. ⁹

We hypothesize that patients with coronary artery disease (CAD) may have varying degrees of subclinical liver fibrosis that could affect their clinical outcomes. Therefore, leveraging data from the ISCHEMIA (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches) trial population, we sought to explore the association between the degree of liver fibrosis, as indicated by the baseline FIB‐4 score, and clinical outcomes.

Methods

Data Availability Statement

Data were obtained through the National Heart, Lung, and Blood Institute Biologic Specimen and Data 215 Repositories Information Coordinating Center. The need for review was waived by the Cornell Medicine Institutional Review Board.

The ISCHEMIA Trial

The ISCHEMIA trial design and results have been previously published. ¹⁰ , ¹¹ In brief, ISCHEMIA was a multicenter, randomized, controlled trial that randomly assigned patients with moderate or severe ischemia on noninvasive testing and acceptable or absent levels of angina to an initial invasive strategy (coronary angiography and coronary revascularization if appropriate) plus medical therapy, or to an initial conservative strategy with medical therapy alone, with angiography/revascularization reserved for medical therapy failure (refractory angina or a primary end point event). In the invasive arm, the mode of revascularization (CABG or PCI) was at the discretion of the local heart team. The primary outcome was the composite of death from cardiovascular causes, myocardial infarction (MI), or hospitalization for unstable angina, heart failure, or resuscitated cardiac arrest.

Study Design and Outcomes

The study population consisted of participants randomized in the ISCHEMIA trial. Participants without complete baseline laboratory data (platelet count, aspartate aminotransferase, and alanine aminotransferase) were excluded from the analysis. The primary outcome was all‐cause death. The secondary outcomes were cardiovascular death, heart failure, MI (procedural and spontaneous), and stroke.

Statistical Analysis

Categorical variables were described with percentages. Continuous variables (age, ejection fraction, and estimated glomerular filtration rate) were described with median and interquartile range (IQR). Multivariable Cox proportional hazards regression was used to examine the association between baseline FIB‐4 and the outcomes. The model was adjusted for the following clinical and angiographic covariates: sex, prior MI, smoking status, diabetes, left ventricular ejection fraction <45%, prior CABG, prior heart failure hospitalization, chronic lung disease, prior stroke, known peripheral vascular disease, White versus non‐White race, and estimated glomerular filtration rate lower versus higher than 60 mL/min.

Hazard ratio (HR) and its 95% CI were calculated to address the difference between groups. In the Cox regression models, the HR reflects the risk associated with a 1‐unit increase in the baseline FIB‐4 score. Futhermore, a subgroup analysis addressing the association of HR and patients' subsets of the trial (CABG, PCI, and conservative therapy) was performed.

The “rms” package in R (R Foundation for Statistical Computing, Vienna, Austria) was used to model the relationship between baseline FIB‐4 and the relative hazard of all‐cause death. First, FIB‐4 was labeled appropriately, and the variable distribution was defined using the “datadist” function to enable proper handling within the “rms” package. A Cox proportional hazards model was then fitted using restricted cubic splines with 4 knots to allow for a flexible, nonlinear relationship between FIB‐4 and mortality risk. The statistical significance of FIB‐4 in the model was assessed using ANOVA, and its distribution was summarized. A 2×2 plotting grid was set up to accommodate potential subgroup analyses. The final plot was generated using the “Predict” function, which computed HRs on the exponential scale, and the resulting relationship was visualized with appropriate axis labels and titles. This approach follows the methodology outlined by Harrell (2015) in Regression Modeling Strategies, which describes the use of restricted cubic splines for flexible modeling in survival analysis. A P value <0.05 was considered statistically significant. All statistical analyses were performed using R version 4.3.1 within RStudio.

Results

Study Population

A total of 3735 patients were included in the analysis. Baseline patient characteristics are presented in Table 1. The median FIB‐4 in the cohort was 1.32 (IQR, 0.99–1.77). Most of the patients were men (77.6%) with a median age of 64 years (IQR, 57–70). These patients presented mainly with an adequate left ventricular ejection fraction (60 [IQR, 55–65]) and renal function (estimated glomerular filtration rate: 82.2 [IQR, 67.6–98.1]). Hypertension was the most frequent comorbidity (73.5%), and almost one‐fifth had a previous MI (19.1%) and underwent prior PCI (20.6%) instead of CABG (3.8%). Ultimately, the vast majority presented with a history of angina (90.2%).

Table 1.

Baseline Patient Characteristics of the Included Population

	n=3735
FIB‐4, median (IQR)	1.32 (0.99–1.77)
Age, median (IQR)	64.0 (57.0–70.0)
Female sex, n (%)	838 (22.4)
Race, n (%)
Black	160 (4.3)
Other	1192 (32.1)
White	2358 (63.6)
Continuous ejection fraction, %, median (IQR)	60.0 (55.0–65.0)
MDRD calculated eGFR at enrollment, median (IQR)	82.2 (67.6–98.1)
Hypertension, n (%)	2739 (73.5)
Diabetes, n (%)	1611 (43.1)
Smoking status, n (%)
Current smoker	479 (12.8)
Former smoker	1668 (44.7)
Never smoked	1585 (42.5)
Family history of CAD, n (%)	803 (24.9)
Prior MI, n (%)	713 (19.1)
Prior PCI, n (%)	768 (20.6)
Prior CABG, n (%)	142 (3.8)
Prior HF hospitalization, n (%)	43 (1.2)
Atrial fibrillation/atrial flutter, n (%)	158 (4.2)
Prior stroke (unknowns set to missing), n (%)	106 (2.8)
Prior cerebrovascular disease, n (%)	262 (7.0)
Prior peripheral artery disease, n (%)	146 (3.9)
History of angina, n (%)	3369 (90.2)

CABG indicates coronary artery bypass grafting; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; FIB‐4, fibrosis‐4 index; HF, heart failure; IQR, interquartile range; LVEF, left ventricular ejection fraction; MDRD, modification of diet in renal disease; MI, myocardial infarction; and PCI, percutaneous coronary intervention.

Association Between FIB‐4 and Clinical Outcomes

Baseline FIB‐4 score was significantly associated with an increased risk of all‐cause death (HR, 1.19 [95% CI, 1.07–1.32]; P=0.001). This association was present in the overall population and in all patient subsets (Figure).

Figure 1. Hazard of all‐cause death in relation to baseline FIB‐4 level among entire cohort and patients receiving CABG, PCI, or medical therapy, respectively.

CABG indicates coronary artery bypass grafting; and PCI, percutaneous coronary intervention.

Moreover, baseline FIB‐4 score was significantly associated with an increased risk of cardiovascular death (HR, 1.19 [95% CI, 1.04–1.36]; P=0.011). This association was consistent across the overall population and within subgroups of patients treated with PCI, CABG, and medical therapy. There was no significant difference between groups with respect to the incidence of heart failure (HR, 0.87 [95% CI, 0.58–1.29]; P=0.480), MI (HR, 0.95 [95% CI, 0.81–1.11]; P=0.508) and stroke (HR, 1.09 [95% CI, 0.82–1.45]; P=0.554; Table 2).

Table 2.

Outcomes of Included Patients and Corresponding Unadjusted and Adjusted HR for Baseline FIB‐4

Outcome	Overall (n=3735)	Unadjusted HR of FIB‐4 (95% CI), P value	Adjusted HR of FIB‐4 (95% CI), P value*
All‐cause death	217 (5.8)	1.29 (1.19–1.41), <0.001	1.19 (1.07–1.32), 0.001
Cardiovascular death	153 (4.1)	1.26 (1.13–1.41), <0.001	1.19 (1.04–1.36), 0.011
Heart failure	49 (2.9)	1.21 (0.96–1.5), 0.104	0.87 (0.58–1.29), 0.480
Myocardial infarction	431 (25.2)	1.07 (0.95–1.21), 0.240	0.95 (0.81–1.11), 0.508
Stroke	65 (1.7)	1.22 (1.02–1.47), 0.034	1.09 (0.82–1.45), 0.554

FIB‐4 indicates fibrosis‐4 index; and HR, hazard ratio.

^*The model was adjusted for the following clinical and angiographic covariates: sex, prior myocardial infarction, smoking status, diabetes, left ventricular ejection fraction <45%, prior coronary artery bypass grafting, prior heart failure hospitalization, chronic lung disease, prior stroke, known peripheral vascular disease, White vs non‐White race, and estimated glomerular filtration rate lower vs higher than 60 mL/min.

Discussion

In this analysis of the ISCHEMIA trial population, we found that elevated baseline FIB‐4 scores were associated with a significantly increased risk of all‐cause and cardiovascular death in patients with CAD. Importantly, this relationship persisted even after adjustment for a comprehensive set of clinical and demographic covariates, including age, which is a component of the FIB‐4 score. These findings suggest that the increased risk associated with a higher baseline FIB‐4 score is not merely a reflection of advanced age but rather reflects a more complex interplay between hepatic and cardiovascular pathophysiology. ⁵ , ⁶ The pathophysiological mechanisms linking subclinical liver fibrosis to cardiovascular outcomes are not entirely understood but may involve systemic inflammation, endothelial dysfunction, and metabolic dysregulation, all of which are known to contribute to adverse cardiovascular events. ⁶ , ¹² , ¹³ , ¹⁴ Subclinical liver fibrosis is associated with subclinical atherosclerosis, highlighting a significant connection between liver disease and cardiovascular conditions, even in the absence of clinical symptoms. ¹⁵ , ¹⁶ , ¹⁷

A previous study reported a direct correlation between hepatic fibrosis and coronary microvascular dysfunction measured by myocardial perfusion reserve index in magnetic resonance imaging. After adjusting for cardiometabolic risk factors, it was found that a higher baseline FIB‐4 score was directly associated with a decrease in the myocardial perfusion reserve index (β, −1.12 [SE, 0.46]; P=0.02). ¹⁸ Additionally, research suggests that chronic inflammation and metabolic processes involved in liver fibrosis may contribute to vascular dysfunction and endothelial inflammation, both key factors in the development of atherosclerosis. A Japanese study investigated the relationship between nonalcoholic fatty liver disease severity and subclinical atherosclerosis, focusing on carotid intima‐media thickness. The prevalence of maximum carotid intima‐media thickness ≥1.2 mm was significantly higher in patients with advanced fibrosis compared with those without (75.4% versus 44.0%; P<0.01). Noninvasive liver fibrosis markers, such as the FIB‐4 index and nonalcoholic fatty liver disease fibrosis score, demonstrated diagnostic accuracy for maximum carotid intima‐media thickness ≥1.2 mm comparable with biopsy‐based fibrosis staging. ¹⁹

The FIB‐4 score was initially developed to use routine laboratory tests for estimating liver fibrosis in patients with HIV/hepatitis C virus coinfection. It has since emerged as a valuable tool for identifying liver fibrosis across various clinical settings.

Despite its widespread use, there is limited evidence linking FIB‐4 levels to death or other clinical outcomes in patients who underwent cardiac procedures. A recent investigation from Richter et al ² found that in a large cohort of patients undergoing elective CABG, elevated baseline FIB‐4 scores were associated with increased surgical risk, and some studies ²⁰ , ²¹ found also that elevated baseline FIB‐4 levels were associated with higher mortality risk in patients with acute coronary syndrome (adjusted HR, 2.8 [95% CI, 1.07–7.86]; P=0.03). ²¹

Our study shows that even subclinical liver fibrosis, as measured by baseline FIB‐4, may have significant prognostic implications in patients with CAD and underscores the need for a comprehensive approach to risk stratification in these patients. In clinical practice, the baseline FIB‐4 score could be used as an additional tool to identify high‐risk patients who might benefit from more intensive monitoring and therapeutic interventions. Future studies are needed to validate these findings in larger, more diverse populations and to explore whether interventions aimed at reducing hepatic fibrosis can improve outcomes in patients with CAD.

Study Limitations

Our study has several limitations. The mere identification of a documented risk factor frequently fails to encompass the severity of a specific comorbidity (eg, initial diabetes and diabetes with other organ failures). Additionally, an important fraction of patients was excluded from the analysis, which may limit the generalizability of the findings. The study's relatively short follow‐up period may not capture long‐term outcomes or delayed complications.

Conclusions

This post hoc analysis of the ISCHEMIA trial found that baseline FIB‐4 is associated with death in patients with CAD.

Sources of Funding

T.C. is supported by the Deutsche Forschungsgemeinschaft (German Research Foundation) Clinician Scientist Program OrganAge funding number 413668513, by the Deutsche Herzstiftung (German Heart Foundation) funding number S/03/23, and by the Interdisciplinary Center of Clinical Research of the Medical Faculty Jena.

Disclosures

B.R. receives grant support from the European Research Council, the Swedish Scientific Council, Swedish Heart and Lung Foundation, and the Swedish Society for Medical Research, and consulting payments from Pfizer and Boehringer Ingelheim. M.F.L.G. receives grant support from the National Institutes of Health, the Canadian Institutes of Health and Research, the Patient‐Centered Outcomes Research Institute, Abbott Vascular, and the Starr Foundation. The remaining authors have nothing to disclose.