Non-invasive scoring systems of liver fibrosis predict prognosis in the cohort with myocardial infarction

Introduction

Acute myocardial infarction (AMI) is responsible for more than one-third of all deaths in developed nations^1,2. AMI incidence is predicted to see a three-fold rise by 2050³, and will become the leading driver of cardiovascular-related mortality^4,5. Conventional risk stratification models such as the Global Registry of Acute Coronary Events (GRACE) Risk Score (GRS) have been used to estimate AMI prognosis and guide management⁶. However, current evidence suggests that risk stratification using GRS in AMI patients did not improve cardiovascular disease (CVD) events at 12 months compared to standard care⁷, and GRS predicts outcomes more poorly in patients with type 2 diabetes mellitus (T2DM)⁶. In the era of personalised medicine, enhanced risk stratification tools following AMI are needed⁸.

Previous studies have demonstrated that the presence of liver fibrosis is associated with increased liver and cardiovascular-related outcomes in the general population⁹, and is a key predictor of long-term mortality in the AMI population^10–14. However, the gold standard for the diagnosis of hepatic fibrosis requires liver biopsy¹⁵, which is limited by procedural invasiveness and risk of bleeding, especially in the population with CVD requiring anti-platelet therapy^16,17. As such, serum-based non-invasive tests (NITs) have been recommended in large-scale populations for the screening of hepatic fibrosis¹⁸, which include Fibrosis-4 Index (FIB-4), Aspartate Aminotransferase (AST) to Platelet Ratio Index (APRI) and Non-alcoholic Fatty Liver Disease Fibrosis Score (NFS)¹⁹. These NITs take advantage of readily available, low-cost serum-based biomarkers, and have been shown to add to the risk stratification of CVD risk in individuals with metabolic dysfunction-associated liver steatotic disease (MASLD). Studies have shown that FIB-4 ≥ 2.67 is associated with a 40% increase in cardiovascular mortality, and predicts major adverse cardiovascular events (MACE)^17,20,21. Previous studies have shown that FIB-4 score is correlated with an increased risk of MACE and cerebrovascular events in patients following AMI, with or without the presence of T2DM^19,22. However, other NITs such as APRI and NFS have not been investigated in a secondary AMI cohort.

The incorporation of NITs in conventional CVD risk prediction equations for the AMI population will require the examination of the performance of these liver indices in a secondary prevention population^19,23. Given that more than one-third of people with concurrent T2DM and MASLD are affected by liver fibrosis²⁴, there is a need to compare the utility of NITs in the population with and without T2DM. As such, this present study aims to fill the gap in the literature by performing head-to-head comparisons of NITs in the AMI cohort, stratified by the presence of T2DM^25–30.

Patient and methods

In this retrospective observational cohort, we studied patients who presented with AMI between 1 January 2011 and 31 March 2021 at a tertiary hospital in Singapore. The academic centre is part of the Western network, consisting of one hub and two spoke hospitals, that provides percutaneous coronary intervention services to the western region of Singapore³¹. All patients were aged 18 years and older, presenting with either ST-elevation myocardial infarction (STEMI) or non-ST elevation myocardial infarction (NSTEMI), via the Emergency Department at the hub hospital or interhospital transfers from the two spoke hospitals. Patients with AMI were allocated into high, intermediate and low-risk groups using established NITs (FIB-4, APRI, and NFS) to evaluate liver fibrosis^32,33. They were further stratified based on the presence of T2DM.

AMI diagnosis by attending cardiologists were based on clinical evidence of acute myocardial ischaemia, determined by a rise and/or fall of troponin values with at least one value above the 99th percentile upper reference limit coupled with at least one of the following: (1) symptoms of myocardial ischaemia, (2) new ischaemic electrocardiographic changes, (3) development of pathological Q waves, (4) imaging evidence of new loss of viable myocardium or new regional wall motion abnormality in a pattern consistent with an ischaemic aetiology, or (5) identification of a coronary thrombus by angiography. These diagnoses fit the current universal definition of type 1 AMI^34–36. Baseline demographic, past medical histories, clinical characteristics, angiographic and procedure characteristics and clinical outcomes were retrieved retrospectively from the electronic clinical records.

Results

Discussion

Previous studies have shown that FIB-4 score is correlated with an increased risk of MACE and cerebrovascular events in patients following AMI, with or without the presence of T2DM^19,22. This study adds to the growing evidence that individuals at high-risk of liver fibrosis, scored based on NITs, had unfavourable survival following AMI. The key findings highlighted that (1) there was stepwise incremental risk in AMI-related complications including, cardiac arrest, heart failure, cardiogenic shock and stroke, with increasing risk scores for FIB-4, APRI and NFS, independent of T2DM status; (2) Cox regression showed that high FIB-4, intermediate and high APRI, and intermediate and high NFS risk groups were independent predictors of long-term mortality following AMI; (3) Although FIB-4 and APRI high-risk groups independently predicted mortality regardless of T2DM status, NFS only predicted mortality in the T2DM group; (4) A positive linear relationship with all three risk scores (as a continuous variable) and mortality risk following AMI was observed (Graphical Abstract).

Liver disease exists as a continuous spectrum from simple hepatic steatosis to steatohepatitis, that may progress to advanced fibrosis and cirrhosis¹³. Increasing liver fibrosis severity has been associated with higher incidence of cardiometabolic diseases and greater CVD severity^51–55. Several studies have shown the increasingly adverse metabolic-inflammatory burden with higher degrees of liver fibrosis, including increased hepatic insulin resistance, dysregulated lipid and glucose metabolism, altered secretion of coagulation factors, systemic low-grade inflammation, and endothelial dysfunction^2,27,56–59. Furthermore, this dose-response relationship between the severity of liver disease and adverse cardiovascular outcomes has been described in the AMI population, with individuals with hepatic fibrosis at significantly higher mortality risk in contrast to those without^12,13. Our study highlights the prognostic value of the three liver indices given the stepwise incremental associations with AMI-related outcomes such as cardiac arrest, heart failure, cardiogenic shock, stroke, and mortality rates, from the low, intermediate to high-risk groups. As such, the use of liver indices for the assessment of hepatic fibrosis and its degree of severity holds promising potential in the role as both predictive and prognostic biomarkers following the AMI event^12,23,59.

Several studies have examined each of the liver indices in silos, and have shown that high FIB-4⁶⁰, NFS⁶¹ and APRI⁶² score were significantly associated with elevated risk of MACE and all-cause mortality in patients with AMI. Our head-to-head comparison of the various NITs highlight that the high-risk category across all 3 liver NITs have important long-term prognostic value in risk stratifying the population presenting with AMI. Furthermore, this dose-response relationship from the low to intermediate and high-risk categories, associated with incremental rise in mortality risk, was evident in the APRI and NFS risk stratification, but not in the FIB-4. The current practice guidelines have identified FIB-4 as an effective tool for stratification of liver fibrosis risk, and our study extends the literature demonstrating that a high FIB-4 is also an effective predictor of all-cause mortality in the AMI cohort^19,33,41. However, as intermediate FIB-4 did not predict a higher mortality risk compared to the low FIB-4 score, this highlights the need for a nuanced understanding of its use in mortality risk stratification following AMI^63–65.

Our study also compared the use of different cut-offs of FIB-4 in predicting outcomes following AMI. The higher cut-off of 3.25 has been well validated in studies for ruling in severe fibrosis with a high specificity of 92.6% but a lower sensitivity of 41.4%.^66,67 In an effort to improve sensitivity, recent studies have also proposed a lower cut-off of 2.67 to identify those at high risk for advanced fibrosis, with positive predictive values ranging from 60 to 80%, and a sensitivity of 70.7% and specificity of 79.1%.^40,41,68,69 As such, the lower FIB-4 score of 2.67 has been recommended for initial screening while the higher FIB-4 score of 3.25 has been recommended for ruling in severe fibrosis^66,69. Our sensitivity analysis demonstrated that both FIB-4 score of ≥ 2.67 and ≥ 3.25 independently predicted poorer long-term survival in AMI patients.

T2DM is a key driver of MASLD progression, with more than a third of individuals with concurrent T2DM and MASLD progressing to hepatic fibrosis²⁴. Moreover, the associations between hepatic fibrosis and cardiometabolic risk factors are often bidirectional^70,71, with the presence of fibrosis conferring an increased risk of incident T2DM^72–74. Notably, prevalence of T2DM was higher in low and intermediate-risk versus high-risk groups when stratified by FIB-4 and APRI. Although more studies are needed to examine this finding, existing evidence suggests that MASLD without elevated serum aminotransferase is prevalent in individuals with concurrent T2DM, leading to falsely low FIB-4 and APRI scores^75,76. While previous studies have examined the performance of serum-based scoring systems in patients with T2DM⁷⁷, these scoring systems have yet to be evaluated in an AMI cohort stratified by T2DM^12,22. Our study highlights the key finding that the high-risk groups determined by all three liver indices were predictive of mortality in the AMI population with T2DM. However, NFS was not predictive of mortality in patients without T2DM, in line with previous evidence which suggests limited sensitivity of NFS in this subgroup⁷⁸. As the presence of hyperglycemia is one of the key components within the NFS formula, the risk of hepatic fibrosis will be underestimated in patients without T2DM^78,79. While existing CVD risk-prediction models may underestimate the cardiovascular risk in the population with liver fibrosis – given that these models do not incorporate fibrosis-related risk factors such as insulin resistance, obesity, and hypertriglyceridemia^74,80–85– the goal of improving the discriminant value in these models must be met with cautious integration of NITs with the understanding of the limitations of specific NITs in certain population groups (in particular, the NFS in the non-diabetic cohort).

The study incorporated continuous FIB-4, APRI and NFS scores into the restricted cubic splines to examine the correlation between these indices and cardiovascular events. The findings demonstrate a positive linear relationship with mortality risk for FIB-4, and a significant non-linear relationship for APRI and NFS following AMI. However, NFS did not demonstrate a significant relationship with mortality risk for non-T2DM patients. Notably, many of these liver biomarkers have been implemented in relation to a cross-sectional, binary assessment of categorical risk scores, which do not reflect the continuous and dynamic nature of liver fibrogenesis⁸⁶. The integration of the levels of liver biomarkers as a continuum should be considered when developing future risk stratification tools.

Limitation

This study has its limitations. First, the gold standard liver biopsy for the identification and staging of liver fibrosis and ultrasound-based techniques such as elastography were not used. Nonetheless, the use of serum-based liver indices have been shown to be well-validated and feasible in the larger scale-screening of hepatic fibrosis, especially in an ‘all-comers’ AMI cohort^12,33,70. In addition, as both imaging and biopsy are not feasible in a large-scale population screening, the use of readily available, low-cost serum based indicators may be important in identifying high risk patients⁸⁷. Second, serum transaminases may be deranged during the acute phase of AMI. Most NITs used in hepatology practice incorporate AST. However, AST serves as a biomarker not only for hepatic conditions but also for cardiac and muscular disorders. In acute clinical settings, reliance on these indicators may lead to misinterpretation, as cardiac injuries disproportionately elevate AST levels compared to ALT. Therefore, it is essential to consider these limitations when interpreting results to avoid diagnostic confusion. Nevertheless, this confounder was mitigated by using serum transaminases measured at discharge as a more precise indicator of liver enzyme derangement⁸⁸. Third, the study is limited by its retrospective, monocentric design, which may affect the generalizability of the findings. Given the varied prevalence of T2DM and MASLD across different populations, the generalisability and interpretation of the findings from this multi-ethnic Asian cohort should be made with caution, especially when applying these findings to other demographic cohorts⁸⁹.

Conclusion

In conclusion, individuals at high risk of advanced hepatic fibrosis, staged by high FIB-4, high APRI, and high NFS scores, could predict all-cause mortality following AMI. Both FIB-4 and APRI were helpful risk stratification tools, irrespective of T2DM status; but this discrimination value was not observed with NFS in the non-T2DM cohort. The stepwise incremental risk of mortality and AMI-related complications, as well as the positive relationship with the liver risk scores and AMI-related mortality risk, suggest that these readily available tools can be beneficial for risk stratification of patients presenting with AMI, irrespective of the presence of T2DM. This may guide timely identification and management of high risk patients with AMI, with the concerted goal of improving cardiovascular prognosis. The study findings should be validated against future large prospective, non-selective AMI cohorts with systematic concomitant evaluation of hepatic fibrosis, preferably by liver elastography (short of biopsy descriptors), to better understand and assess the reproducibility of hepatic fibrosis biomarkers as a predictor of mortality and cardiovascular-related outcomes.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

G.K, J.C and N.W.S.C conceived and designed the study. J.C was involved in data acquisition and statistical analyses. E.W, Y.C, and G.K were involved in interpretation of data. E.W, Y.C, G.K, J.C, A.Z, A.M, J.J, B.C, G.C, B.N, M.Y.C, M.M, P.L, M.A.M and N.W.S.C drafted the manuscript. N.W.S.C provided supervision and mentorship. All authors reviewed and revised the manuscript, gave final approval, and agreed to be accountable for all aspects of the work.

Funding

This was supported by the CSDU Clinician-Scientist Grant, and the National Medical Research Council Transition Award (TA24jul-0008), and the CArdiovascular DiseasE National Collaborative Enterprise (CADENCE) National Clinical Translational Program (MOH-001277-01). N.W.S.C has received research grant support from NUHS Seed Fund (NUHSRO/2022/RO5 + 6/Seed-Mar/03), National Medical Research Council Research Training Fellowship (MH 095:003/008-303), National University of Singapore Yong Loo Lin School of Medicine’s Academic Fellowship Scheme, the NUHS Clinician Scientist Program (NCSP2.0/2024/NUHS/NCWS), CSDU Clinician-Scientist Grant, and the National Medical Research Council Transition Award (TA24jul-0008). M.Y.C. receives speaker’s fees and research grants from Astra Zeneca, Abbott Technologies, and Boston Scientific.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations