Association Between Liver Fibrosis Markers and Hemorrhagic Transformation Following Endovascular Treatment for Acute Ischemic Stroke: A Chinese Stroke Center Study

Introduction

Stroke is the second leading cause of death and the third leading cause of disability worldwide, posing a significant public health challenge.¹ Among patients with acute ischemic stroke (AIS), 28–46% present with large vessel occlusion (LVO).² These patients are of particular clinical concern because of the rapid progression of their condition, severe neurological deficits, and poor prognosis.³ In recent years, endovascular therapy (EVT) has emerged as the preferred interventional strategy for LVO-AIS in the hyperacute phase, as it enables rapid recanalization of occluded vessels, salvages the ischemic penumbra, and significantly improves functional outcomes.⁴ However, approximately half of the patients who undergo EVT develop postreperfusion hemorrhagic transformation (HT), among which symptomatic intracranial hemorrhage (sICH) is closely associated with increased acute-phase mortality and a greater risk of poor long-term functional outcomes.^5–7 Therefore, systematically identifying predictive biomarkers for post-EVT HT and developing precise prevention and treatment strategies are critical issues that need to be addressed in the current stroke research field.

Previous studies have identified multiple risk factors associated with post-EVT hemorrhagic complications, including clinical characteristics (such as atrial fibrillation, high baseline National Institutes of Health Stroke Scale [NIHSS] scores, and prolonged symptom-onset-to-reperfusion time), procedural factors (including number of retrieval attempts), laboratory parameters (such as hyperglycemia), and radiological markers (such as hyperdense artery sign on baseline computed tomography [CT] and poor collateral circulation on computed tomography angiography [CTA]).^8,9 Nevertheless, existing predictive models demonstrate suboptimal performance, and our understanding of the underlying pathophysiological mechanisms remains limited. Therefore, systematically identifying predictive biomarkers for post-EVT HT and elucidating their potential mechanisms have significant clinical value for developing individualized prevention strategies and improving patient outcomes.

Metabolic dysfunction-associated steatotic liver disease (MASLD) is the most common chronic liver disease worldwide, with a prevalence of 30.1%. Owing to the increasing prominence of risk factors such as obesity and diabetes, the disease burden of MASLD is expected to continue increasing.¹⁰ The pathological spectrum of MASLD ranges from simple steatosis to metabolic dysfunction-associated steatohepatitis (MASH), and it may progress to liver fibrosis, cirrhosis, and even hepatocellular carcinoma.¹¹ During the pathological progression of MASLD, liver fibrosis is considered a key stage that may participate in the occurrence and development of stroke through various mechanisms, including endothelial dysfunction, coagulation abnormalities, dyslipidemia, the systemic inflammatory response, and oxidative stress.^12–17

A growing body of evidence suggests that the severity of liver fibrosis is closely related to poor outcomes in stroke patients, including impaired functional recovery, increased long-term mortality, and increased recurrence rates.^18–21 Liver fibrosis has been shown to promote hematoma expansion in primary intracerebral hemorrhage²⁰ and is independently associated with an increased risk of sICH after intravenous thrombolysis.^22,23 However, the relationship between liver fibrosis and post-EVT HT remains controversial.^24,25 Some studies have shown that elevated fibrosis-4 (FIB-4) scores can independently predict the risk of post-EVT HT, whereas other studies have not reported similar associations. More importantly, most existing studies rely on a single fibrosis index, which may limit the accurate assessment of liver fibrosis severity. We focused on exploring the independent predictive value of liver fibrosis indices rather than directly developing a composite scoring system, primarily based on two considerations: first, establishing the independent contribution of specific biomarkers is a prerequisite for constructing effective composite models; second, the systemic pathological state reflected by liver fibrosis may represent a unique mechanism not yet incorporated into existing predictive models, thus providing substantial complementary value.

Therefore, the present study is the first to systematically evaluate the associations between eight non-invasive liver fibrosis indices and post-EVT HT and sICH in patients with anterior circulation LVO stroke. Furthermore, we analyzed the dose‒response relationships between these indices and hemorrhagic phenotypes (hemorrhagic infarction [HI] vs parenchymal hematoma [PH]). This study aims to clarify the potential value of liver fibrosis indices in risk stratification for post-EVT HT and provide a theoretical basis for developing targeted prevention strategies.

Material and Methods

Study Design and Population

We conducted a single-center retrospective cohort study, consecutively enrolling patients with anterior circulation LVO-AIS who underwent EVT at a comprehensive stroke center between January 1, 2018, and December 31, 2024. The studies involving humans were approved by the Ethics Committee of Zhongshan Hospital of Xiamen University. As the study was retrospective and the data were anonymized, the ethics committee agreed to waive the requirement for informed consent.

The inclusion criteria were age ≥18 years and the presence of anterior circulation LVO (internal carotid artery or middle cerebral artery M1/M2 segment) confirmed by computed tomography angiography or digital subtraction angiography (DSA). Exclusion criteria included: (1) pre-existing intracranial hemorrhage on baseline head CT or magnetic resonance imaging; (2) pre-stroke modified Rankin Scale score >2; (3) history of severe liver disease (viral hepatitis, cirrhosis, primary liver cancer, etc.) or severe renal dysfunction; (4) current use of hepatotoxic drugs; and (5) incomplete clinical laboratory data or imaging records. To reflect real-world clinical practice, we did not exclude patients with a history of chronic alcohol consumption.²⁴

Data Collection

Baseline data, including demographic characteristics (age, sex), lifestyle factors (smoking, alcohol consumption), medical history (hypertension, diabetes, hyperlipidemia, atrial fibrillation, previous stroke or transient ischemic attack, coronary heart disease), and vital signs at admission (systolic and diastolic blood pressure), were extracted from the hospital electronic medical record system via a standardized electronic data collection form.

Neurologists who received standardized training assessed the severity of neurological deficits at admission using the National Institutes of Health Stroke Scale (NIHSS) and evaluated the level of consciousness using the Glasgow Coma Scale (GCS).^26,27 The Alberta Stroke Program Early CT Score (ASPECTS) was used to quantitatively assess the baseline infarct volume.²⁸ Stroke etiology was classified according to the Trial of Org 10172 in Acute Stroke Treatment criteria. Furthermore, detailed information regarding preoperative anticoagulation therapy was documented, including the use of vitamin K antagonists (warfarin) and non-vitamin K antagonist oral anticoagulants.

We recorded treatment-related details, including intravenous thrombolysis status, key time points (onset-to-puncture time, onset-to-reperfusion time, puncture-to-reperfusion time), and procedural details (number of mechanical thrombectomy attempts, device strategy [stent retriever, aspiration catheter, or combined use], and balloon angioplasty). The modified Thrombolysis in Cerebral Infarction (mTICI) grading system was used to evaluate post-procedural reperfusion status, based on the final DSA results, with mTICI grades 2b-3 defined as successful reperfusion.²⁹

All blood samples were collected in the fasting state within 24 hours of admission (between 6:00–8:00 am the following morning). The samples were analyzed using automated analyzers in the hospital’s central laboratory. The measured parameters included white blood cell, neutrophil count, lymphocyte count, platelet count (PLT), total protein, albumin (ALB), total bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyl transferase (GGT), alkaline phosphatase (ALP), lipid profile (total cholesterol, triglycerides, high-density lipoprotein cholesterol [HDL-C], low-density lipoprotein cholesterol [LDL-C]), fasting blood glucose, urea, creatinine, and uric acid. Thrombocytopenia was defined as a platelet count below 150×10⁹/L in accordance with established clinical criteria.

Assessment of Liver Fibrosis

Eight validated non-invasive liver fibrosis indices were calculated using laboratory parameters:²¹ FIB-4, modified FIB-4 (mFIB-4), fibrosis-5 (FIB-5), the AST-to-PLT ratio index (APRI), the Forns index, the AST/ALT ratio (AAR), the AST/ALT-PLT ratio index (AARPRI), and liver fibrosis. The specific calculation formulas are presented in Figure 1.

Figure 1 Calculation Formulas for Eight Non-invasive Liver Fibrosis Indices. Abbreviations: FIB-4, fibrosis-4; mFIB-4, modified fibrosis-4; FIB-5, fibrosis-5; APRI, AST to platelet ratio index; AAR, AST to ALT ratio; AARPRI, AST/ALT-platelet ratio index; AST, aspartate aminotransferase; ALT, alanine aminotransferase; PLT, platelet count; ALB, albumin; ALP, alkaline phosphatase; GGT, γ-glutamyl transferase.

Outcome Measures

All patients underwent head CT examinations at 24 and 72 hours after EVT. HT was classified into four categories according to the Heidelberg Bleeding Classification:³⁰ HI-1 (punctate hemorrhage at the infarct margin), HI-2 (confluent hemorrhage within the infarct zone without mass effect), PH-1 (hematoma volume ≤30% of the infarcted area without significant mass effect), and PH-2 (hematoma volume >30% of the infarcted area with significant mass effect).

sICH was defined as any intracranial hemorrhage associated with one or more of the following conditions in the absence of other explanations: an increase in the total NIHSS score of ≥4 points compared to baseline, an increase in any single NIHSS item score of ≥2 points compared to baseline, or neurological deterioration requiring endotracheal intubation, decompressive craniectomy, external ventricular drainage, or other significant medical interventions. All imaging assessments were independently performed by two experienced neuroradiologists and one neurologist who were blinded to the clinical data. In case of discrepancies, consensus was reached through discussion.

Statistical Analyses

Statistical analyses were performed using R software (Version 4.2.2). The Shapiro–Wilk test was used to assess the normality of continuous variables. Normally distributed continuous variables were expressed as mean ± standard deviation and compared between groups using independent samples t-tests. Non-normally distributed continuous variables were expressed as median (interquartile range) and compared using the Mann–Whitney U-test. Categorical variables were expressed as counts (percentages) and compared using the chi-square test or Fisher’s exact test. Receiver operating characteristic (ROC) curves were used to evaluate the predictive value of liver fibrosis indices for hemorrhagic complications. The area under the curve (AUC), optimal cutoff value (determined by the Youden index), sensitivity, and specificity were calculated.

Covariate selection followed a two-step process: first, variables with statistical significance (P<0.05) in univariate analysis were identified as candidate covariates; second, multicollinearity diagnostics were performed on all candidate covariates by calculating variance inflation factor (VIF) and tolerance values, excluding variables that did not meet predetermined criteria (VIF<10 and Tolerance>0.1) (Supplemental Tables 1 and 2). For HT: Model 1 was the unadjusted baseline model; Model 2 adjusted for clinical factors (baseline NIHSS score, GCS score, ASPECTS score, systolic blood pressure, puncture-to-reperfusion time, and number of thrombectomy attempts); Model 3 further adjusted for laboratory parameters (total bilirubin, AST, triglycerides, neutrophil count, lymphocyte count, and platelet count). For sICH: Model 1 was the unadjusted baseline model; Model 2 adjusted for clinical factors (baseline NIHSS score, GCS score, ASPECTS score, preoperative anticoagulation therapy, and number of thrombectomy attempts); Model 3 further adjusted for laboratory parameters (total protein, ALT, AST, total cholesterol, fasting glucose, neutrophil count, lymphocyte count, and platelet count). Restricted cubic spline regression models were used to evaluate potential non-linear associations, with knots placed at the 5th, 35th, 65th, and 95th percentiles. All statistical tests were two-sided, with P<0.05 defined as statistically significant.

Results

Patient Characteristics and Clinical Outcomes

A total of 340 patients with anterior circulation LVO-AIS who underwent EVT were included (Table 1), among whom 161 (47.4%) developed HT and 69 (20.3%) experienced sICH. The median age of the overall population was 67 years (IQR, 57–76 years), and 224 (65.9%) were male. Compared with those in the non-HT group, patients in the HT group had more severe neurological deficits at admission (median NIHSS score: 16 vs 13, P<0.001; median GCS score: 12 vs 13, P=0.047) and lower ASPECTS scores (median: 9 vs 9, P<0.001). The HT group had a longer puncture-to-reperfusion time (P=0.007) and required more mechanical thrombectomy attempts (P=0.002). The two groups were similar in terms of successful reperfusion rates (85.1% vs 86.0%, P=0.805) and proportions of intravenous thrombolysis (41.6% vs 44.1%, P=0.639). Although the HT group had a greater proportion of patients with cardioembolism (51.6% vs 43.0%) and the non-HT group had a greater proportion of patients with large artery atherosclerosis (42.9% vs 52.0%), these differences did not reach statistical significance (P=0.242).

Table 1 Baseline Clinical Characteristics of Patients with Anterior Circulation Large Vessel Occlusion Stroke Stratified by HT Status

Laboratory Characteristics of Patients with and Without HT

Laboratory tests showed that, compared to the non-HT group, patients in the HT group had significantly higher neutrophil counts (P<0.001), total bilirubin levels (P=0.040), and AST levels (P=0.011), while lymphocyte counts (P<0.001), PLT count (P=0.005) and triglyceride levels (P=0.028) were significantly lower (Table 2). Among the eight liver fibrosis indices evaluated, seven showed significant between-group differences. Notably, FIB-4 (P<0.001), modified FIB-4 (P<0.001), APRI (P<0.001), Forns index (P=0.021), AAR score (P<0.001), and AARPRI (P<0.001) were significantly higher in the HT group compared to the non-HT group. In contrast, the FIB-5 index was significantly lower in the HT group (P<0.001).

Table 2 Laboratory Characteristics of Patients with and Without HT After Endovascular Therapy

Subgroup analysis stratified by hemorrhage type demonstrated distinct distribution patterns (Figure 2). FIB-4, modified FIB-4, AAR, and AARPRI showed an increasing trend across the non-HT, HI, and PH groups, reaching the highest levels in the PH group (all P<0.05 compared to the non-HT group). Importantly, these indices were also significantly higher in the PH group compared to the HI group. The FIB-5 index exhibited the opposite trend, being lowest in the PH group (P<0.05 compared to both the non-HT and HI groups). Although APRI, Forns index, and Fibrosis index showed similar trends, the differences did not reach statistical significance.

Figure 2 Distribution of Liver Fibrosis Indices Stratified by Hemorrhage Type Box Plots. Box plots showing the distribution of eight liver fibrosis indices across non-hemorrhagic transformation (Non-HT), hemorrhagic infarction (HI), and parenchymal hematoma (PH) groups: (a) Fibrosis-4 (FIB-4); (b) modified FIB-4; (c) Fibrosis-5 (FIB-5); (d) AST to Platelet Ratio Index (APRI); (e) Forns index; (f) AST/ALT Ratio (ARR); (g) AST/ALT-Platelet Ratio Index (AARPRI); and (h) Fibrosis index. The boxes represent the interquartile range (IQR), with the horizontal line indicating the median. Whiskers extend to 1.5 times the IQR. The red lines represent the trend across hemorrhage severity groups. Statistical significance between groups is indicated as follows: *P<0.05; **P<0.01; ***P<0.001; ns, not significant.

Clinical and Laboratory Characteristics Associated with sICH

Compared to patients without symptomatic intracranial hemorrhage (n=271), those who developed sICH (n=69) had more severe neurological deficits at admission, as indicated by higher NIHSS scores (median: 17 vs 14, P=0.001) and lower GCS scores (median: 11 vs 13, P=0.006). The sICH group had lower baseline ASPECTS scores (median: 8 vs 9, P=0.044) and required more mechanical thrombectomy attempts (median: 2 vs 1, P=0.019). Despite these differences, procedural parameters such as puncture-to-reperfusion time and successful reperfusion rates were similar between the two groups (Table 3). Furthermore, patients who developed sICH had a significantly higher proportion of preoperative warfarin therapy, whereas patients without sICH were more likely to have received no anticoagulation therapy prior to intervention.

Table 3 Baseline Clinical Characteristics of Patients with Anterior Circulation Large Vessel Occlusion Stroke Stratified by sICH Status

The results of the laboratory tests (Table 4 and Figure 3) revealed that patients in the sICH group had significantly lower total protein (P=0.037), PLT (P<0.001), and total cholesterol (P=0.042) levels and higher fasting blood glucose (P=0.002) levels. Except for the FIB-5 index, which was significantly lower (P<0.001), six liver fibrosis indices, FIB-4, modified FIB-4, APRI, Forns index, AAR score, and AARPRI (all P<0.001), were significantly greater in the sICH group. The fibrosis index did not differ significantly between the two groups (P=0.108).

Table 4 Laboratory Characteristics of Patients with and Without sICH After Endovascular Therapy

Figure 3 Comparison of Liver Fibrosis Indices Between Patients with and without Symptomatic Intracerebral Hemorrhage. Violin plots with embedded scatter plots comparing the distribution of liver fibrosis indices between patients with and without symptomatic intracerebral hemorrhage (sICH). The width of each violin represents the kernel density estimation of the data distribution. Individual dots represent actual data points. The horizontal lines within the violins indicate the median and quartiles. *P<0.05; **P<0.01; ***P<0.001; ns, not significant.

Predictive Value of Liver Fibrosis Indices

ROC curve analysis demonstrated that liver fibrosis indices had moderate predictive value for hemorrhagic complications (Table 5 and Figure 4). In predicting HT, FIB-4 (AUC=0.62, 95% CI: 0.56–0.68), modified FIB-4 (AUC=0.62, 95% CI: 0.56–0.67), and AARPRI (AUC=0.62, 95% CI: 0.57–0.68) showed similar performance. The predictive accuracy for sICH was notably higher, with AARPRI displaying the best discriminative ability (AUC=0.70, 95% CI: 0.62–0.77), followed by FIB-5 (AUC=0.69, 95% CI: 0.61–0.76), AAR score (AUC=0.68, 95% CI: 0.61–0.75), and modified FIB-4 index (AUC=0.67, 95% CI: 0.59–0.74).

Table 5 Diagnostic Performance of Liver Fibrosis Indices in Predicting Hemorrhagic Complications After Endovascular Therapy

Figure 4 Receiver Operating Characteristic Curves for Liver Fibrosis Indices in Predicting Hemorrhagic Complications. Receiver operating characteristic (ROC) curves showing the predictive performance of liver fibrosis indices for (a) hemorrhagic transformation and (b) symptomatic intracerebral hemorrhage. The area under the curve (AUC) values with 95% confidence intervals are shown for each index. The diagonal gray line represents the reference line of no discrimination (AUC = 0.5).

Multivariable Analysis

In the fully adjusted models, AARPRI demonstrated the strongest independent association with both HT and sICH (HT: adjusted OR [aOR]=1.53, 95% CI: 1.07–2.17, P=0.019; sICH: aOR=2.36, 95% CI: 1.45–3.84, P<0.001). The AAR score (HT: aOR=1.43, 95% CI: 1.03–1.96, P=0.030; sICH: aOR=1.91, 95% CI: 1.16–3.12, P=0.010), FIB-4 (HT: aOR=1.25, 95% CI: 1.01–1.56, P=0.044; sICH: aOR=1.48, 95% CI: 1.12–1.94, P=0.006), and modified FIB-4 (HT: aOR=1.07, 95% CI: 1.01–1.15, P=0.030; sICH: aOR=1.13, 95% CI: 1.04–1.23, P=0.003) also showed significant associations. Notably, FIB-5 exhibited significant negative correlations with both outcomes (HT: aOR=0.95, 95% CI: 0.91–1.00, P=0.054; sICH: aOR=0.91, 95% CI: 0.85–0.98, P=0.002). In contrast, the APRI score, Forns index, and Fibrosis index showed no significant associations with HT or sICH risk (all P>0.05) (Table 6).

Table 6 Multivariable Logistic Regression Analysis of Fibrosis Indices for Predicting HT and sICH

Non-Linear Association Analysis

RCS regression analysis revealed that liver fibrosis indices were primarily linearly related to hemorrhagic complications (Figures 5 and 6). Only modified FIB-4 showed a significant non-linear association with HT risk (P-nonlinear=0.006). In predicting sICH, all indices maintained linear associations, with no significant non-linear relationships observed.

Figure 5 Restricted Cubic Spline Analysis of the Association Between Liver Fibrosis Indices and Hemorrhagic Transformation Risk. Restricted cubic spline curves showing the dose-response relationship between liver fibrosis indices and the risk of hemorrhagic transformation. The solid blue lines represent the estimated odds ratios (ORs), and the shaded areas represent the 95% confidence intervals. Histograms show the distribution of each index in the study population. The reference line (OR = 1) is indicated by the horizontal dotted line. P-overall indicates the significance of the overall association, and P-nonlinear indicates the significance of non-linear components.

Figure 6 Restricted Cubic Spline Analysis of the Association Between Liver Fibrosis Indices and Symptomatic Intracerebral Hemorrhage Risk. Restricted cubic spline curves showing the dose-response relationship between liver fibrosis indices and the risk of hemorrhagic transformation. The solid blue lines represent the estimated odds ratios (ORs), and the shaded areas represent the 95% confidence intervals. Histograms show the distribution of each index in the study population. The reference line (OR = 1) is indicated by the horizontal dotted line. P-overall indicates the significance of the overall association, and P-nonlinear indicates the significance of non-linear components.

Discussion

In this retrospective cohort study, we systematically evaluated the associations between eight non-invasive liver fibrosis indices and post-EVT hemorrhagic complications. Our findings provide several important insights. First, even after adjusting for multiple confounding factors, FIB-4, modified FIB-4, FIB-5, AAR, and AARPRI remained significantly associated with the risk of both HT and sICH. Second, the levels of liver fibrosis indices showed a progressive increase across the non-HT, HI, and PH groups (PH vs HI: P<0.05), suggesting that the severity of liver fibrosis may drive the worsening of hemorrhagic phenotypes. Furthermore, although the predictive values of these liver fibrosis indices were similar, the aforementioned indices demonstrated significantly better predictive ability for sICH than for HT.

Growing evidence suggests a strong association between liver fibrosis and cerebrovascular diseases. Numerous studies have shown that the degree of liver fibrosis is closely associated with poor outcomes in AIS patients, including increased short-term mortality, delayed long-term functional recovery, and increased risk of stroke recurrence.^18–21 Moreover, research based on non-invasive liver fibrosis indices has found that the severity of liver fibrosis is significantly correlated with the incidence of HT and sICH after intravenous thrombolysis.^22,23 In intracerebral hemorrhage studies, liver fibrosis has been proven to be closely related to hematoma volume at admission and its subsequent expansion.³¹ However, the association between liver fibrosis and post-EVT hemorrhagic complications remains controversial.^24,25 Xu et al²⁴ based on a cohort study of Asian populations, demonstrated that elevated FIB-4 is an independent risk factor for post-EVT sICH (OR=1.306, 95% CI: 1.127–1.512, P=0.001). In contrast, Höfler et al²⁵ in a European cohort study, found that although FIB-4 was associated with poor neurological functional outcomes and mortality 3 months after AIS, it was not related to the risk of HT or sICH. Our study supports the notion that liver fibrosis is an independent influence on post-EVT hemorrhagic risk, consistent with evidence from Asian populations. The reasons for the discrepancies in research results may include the following: First, there may be population differences between Asian and European cohorts, but the specific mechanisms still require further in-depth exploration in future studies. Second, the high incidence of HT after EVT may weaken the statistical association between the two. Furthermore, liver fibrosis indices were only significantly associated with the risk of PH (HI vs non-HT group: P>0.05), and previous controversies may have arisen from the failure to differentiate hemorrhage severity.

Liver fibrosis can contribute to the occurrence of post-EVT HT and sICH through multiple pathological mechanisms, including endothelial dysfunction, coagulation-fibrinolysis abnormalities, metabolic disorders, systemic inflammatory responses, and oxidative stress. First, circulating extracellular vesicles released by hepatocytes in MASLD patients can activate vascular endothelial inflammatory responses through signaling pathways such as miRNA-novel-7 and microRNA-1, promoting atherosclerotic plaque instability and microvascular leakage.^12,32–34 The chronic pro-inflammatory microenvironment leads to increased permeability of cerebral microvascular endothelium, resulting in blood-brain barrier (BBB) dysfunction and exacerbating post-ischemic neuronal injury. This BBB disruption-related cerebrovascular endothelial dysfunction has been proven to be a key factor in post-stroke HT.³⁵ Second, liver fibrosis can also increase the risk of hemorrhage by affecting coagulation function and metabolic status. Most coagulation factors, such as fibrinogen, thrombin, and coagulation factors V, VII, IX, and X, are synthesized by the liver and undergo post-translational modifications. Hepatocellular injury can lead to abnormal synthesis and function of coagulation factors.^13,36 Furthermore, the progression of liver fibrosis can inhibit the production of thrombopoietin, leading to thrombocytopenia, which is consistent with the decreased platelet count observed in the hemorrhage group in this study.³⁷ These factors can collectively increase the risk of hemorrhage. Additionally, MASLD is often accompanied by atherogenic dyslipidemia, manifested as elevated triglycerides, increased small, dense low-density lipoprotein particles, decreased high-density lipoprotein cholesterol, and type 2 diabetes, all of which contribute to increased vascular wall fragility.^14,38

Liver fibrosis can also participate in the pathological process of HT by inducing systemic inflammatory responses. du Plessis et al³⁹ found that the severity of MASLD histology positively correlated with the levels of inflammatory mediators. In particular, serum MMP-9 levels were significantly higher in MASH patients compared to controls (12.6±6.5 vs 8.1±3.5, P<0.01).¹⁵ MMP-9 is closely related to HT in AIS patients.⁴⁰ Previous studies have found that after the occurrence of intracerebral hemorrhage, leukocyte infiltration can activate microglia and recruit peripheral immune cells, accelerating hematoma expansion.^41,42 MASLD can exacerbate this cascade reaction by mobilizing peripheral immune cells.⁴³ This study found that patients in the PH group had significantly higher liver fibrosis indices compared to the HI and non-HT groups, suggesting that the degree of liver fibrosis may be related to the severity of hemorrhagic lesions. Furthermore, oxidative stress, as a common pathological basis for liver diseases and cardiovascular diseases, can trigger the occurrence of HT by degrading the extracellular matrix and tight junction proteins of the BBB, thereby disrupting BBB integrity.^16,44

This study has several limitations. First, as a single-center retrospective study, there may be selection bias, and the generalizability of the results needs to be validated. The characteristics of retrospective studies also limit the acquisition and analysis of certain potentially relevant variables (such as detailed medication history and liver imaging data). Furthermore, our dataset lacked body mass index measurements, which represents a significant limitation given the established association between BMI and AIS outcomes. The absence of this anthropometric parameter may have precluded a more comprehensive analysis of potential confounding factors. Second, although the correlation between liver fibrosis and post-EVT hemorrhage has been confirmed, the guiding significance for clinical practice still needs to be further explored. Additionally, the recent emergence of therapeutic interventions for MASLD, including FDA-approved resmetirom and investigational agents, may potentially alter the natural history and severity of liver fibrosis in future patient populations, which could affect the generalizability of our findings to prospective cohorts treated with these novel therapies.Third, this study employed multiple validated non-invasive liver fibrosis indices, but lacked the corroboration of the diagnostic gold standard—liver biopsy. Future studies could combine imaging examinations (such as liver elastography) and serological indicators to improve the accuracy of liver fibrosis diagnosis. Finally, given the heterogeneity of existing research results, the robustness of the association between post-EVT hemorrhage and liver fibrosis needs to be verified, and differences in the included populations may be a potential reason. The relationship between the two needs to be further clarified through multicenter, prospective studies.

Conclusion

This study confirmed the independent association between non-invasive liver fibrosis indices and post-EVT hemorrhagic complications (HT and sICH) in patients with anterior circulation stroke. Multiple liver fibrosis indices had similar predictive values for both hemorrhagic events. Notably, the severity of liver fibrosis may be related to the type of hemorrhagic transformation (PH vs HI). These findings reveal the potential value of liver fibrosis indices in risk stratification for post-EVT HT. In the future, multicenter prospective studies should be conducted to further validate the robustness of this association and explore hemorrhage prevention strategies for high-risk patients.