Introduction

The Global Burden of Disease Study 2019 showed a significant increase in annual stroke incidence and deaths from 1990 to 2019.¹ Stroke is the leading cause of death in China, with the overall prevalence, morbidity, and mortality of stroke among adults aged 40 years and older estimated to be 2.6%, 505.2/100,000 person-years, and 343.4/100,000 person-years in 2020, respectively.² Acute ischemic stroke (AIS), which accounts for more than 86.9% of stroke, has a high recurrence rate, with 3.6% and 5.6% recurrence rates at 3 months and 12 months, respectively.³ Epidemiological findings suggest that the 1-year risk of recurrence of AIS is 7–13%, and its 10-year risk of recurrence approaches 40%.^4,5 A number of biomarkers have been found to predict recurrence of AIS, such as cyclophilin A, cystatin C, lipoprotein-associated phospholipase A2, and homocysteine.^6–9 However, some biomarkers have certain limitations in clinical application or controversial results in some studies. Therefore, accurate prediction of recurrence of AIS remains challenging.

The occurrence of AIS induces inflammatory responses and oxidative stress, which directly or indirectly result in neuronal cell damage or death, thereby further exacerbating the condition.^10–12 In addition, it has been found that inflammatory responses and oxidative stress stimulate the release of high mobility group box 1 (HMGB1) from damaged or necrotic cells.^13,14 HMGB1 normally acts as a damage-associated molecular pattern molecule in the extracellular space, modulating the inflammatory response and oxidative stress through different pathways.^14,15 HMGB1 released after AIS was found to play a crucial role in both the early pathological damage process and the late promotion of brain tissue repair and remodeling.¹⁶

A systematic review and meta-analysis including 28 studies found that patients with AIS exhibit elevated plasma HMGB1 levels, which are positively correlated with disease severity and infarct volume.¹⁷ Furthermore, a study involving 132 patients with AIS after intravenous thrombolysis found that plasma HMGB1 levels peak 6 hours after thrombolytic therapy and were associated with a poor prognosis at 3 months.¹⁸ Elevated baseline plasma HMGB1 levels were also an independent predictor of poor prognosis at 1 year after AIS.^19,20 These findings strongly support the importance of monitoring baseline HMGB1 levels in patients with AIS, especially for the assessment of disease severity and prognosis.

Genetic analysis and follow-up of 871 patients with AIS revealed that HMGB1 polymorphisms were significantly associated with a high risk of ischemic stroke, in particular the HMGB1 rs1412125 polymorphism, which may be a new genetic biomarker and potential target for detection of susceptibility and recurrence of ischemic stroke.²¹ However, there are relatively few studies of baseline serum HMGB1 levels and AIS on a more long-term basis, and whether they are associated with ischemic stroke recurrence. Therefore, this study investigated the correlation between baseline serum HMGB1 levels and AIS recurrence through a long-term study in patients with AIS.

Material and Methods

Study Population

The study was a prospective study. A total of 544 patients with AIS treated at Jiangyin People’s Hospital from January 2020 to September 2020 were included in this study. The study was approved by the Ethics Committee of Jiangyin People’s Hospital (approval number: 2020009) in accordance with the Declaration of Helsinki. All enrolled patients met the definition of Chinese AIS diagnostic and therapeutic guidelines, were given medication according to the guidelines, and signed a written informed consent before enrolment.²² Patients will be excluded if they present with the following diseases: recurrent ischemic stroke or cerebral haemorrhage with neurodegenerative diseases such as Parkinson’s disease, Alzheimer’s disease or vascular dementia, transient ischemic attack, haematological diseases, autoimmune diseases, co-infections, inability to follow up.

Clinical and Laboratory Assessments

All enrolled patients underwent a neurological physical examination by a specialist with extensive clinical experience, and a detailed history of current and past medical conditions, particularly any previous history of hypertension, diabetes mellitus, or atrial fibrillation, was taken and recorded. Elbow venous blood was drawn from fasting enrolled patients on the second day of hospitalisation.

A portion of the blood samples were sent to the central laboratory for the measurement of low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) using Roche e602 modules, as in our previously published article.²³ A portion of the blood sample was centrifuged at 3000r/min for 15 minutes and the upper serum layer was extracted into an Eppendorf tube. Then the Eppendorf tubes were frozen at −80°C for examination. Serum HMGB1 levels were measured by human ELISA kit (Elabscience, China) according to the instruction. The level of HMGB1 was calculated by reading the absorbance at 450 nm with a spectrophotometer.

Follow-Up

After patients were discharged from the hospital, they were followed up with regular monthly outpatient interviews or telephone interviews with patients’ families. The study was followed until the end of December 2023, with a median follow-up time of 43 months (interquartile range, 39–45 months). The endpoint event in this study was out-of-hospital recurrent AIS. A total of 62 enrolled patients with recurrent AIS were recorded in this study (Figure 1).

Figure 1 Flow chart of the study. AIS, acute ischemic stroke.

Statistical Analysis

The statistical analysis was conducted using the SPSS 25.0 statistical package. Quantitative variables were presented as means ± standard deviations and compared using Student’s t-test. Categorical variables were presented as absolute numbers (percentages) and compared using the chi-square test and Mann–Whitney-Wilcoxon test. The receiver operating characteristic curve (ROC) was used to determine the best cut-off value for each variable to predict recurrence of AIS. Kaplan-Meier survival curves were used to conduct survival analysis which were then compared using the Log rank test. To analyze the relationships between HMGB1 levels and AIS recurrence, univariate and multivariate Cox proportional hazards models were employed, calculating hazard ratios (HRs) and 95% confidence intervals (CIs). All variables that reached statistical significance in univariate Cox analysis were included in the multivariate Cox analysis. A p value<0.05 was considered statistically significant (two-tailed).

Results

The Comparison of Baseline Clinical Data

During the follow-up interval with a median time of 43 months, 62 recurrences were recorded in 544 patients with AIS. Patients were divided into an AIS recurrence group and a no AIS recurrence group based on whether or not they had recurrence of AIS. The comparison of baseline clinical data between these two groups was detailed in Table 1. The prevalence of hypertension, diabetes mellitus and atrial fibrillation was significantly higher in the AIS recurrence group than in the no AIS recurrence group (p<0.05). LDL-C levels, HMGB1 levels and national institute of health stroke scale (NIHSS) were significantly higher in the AIS recurrence group than in the no-AIS recurrence group, while HDL-C levels were lower than in the no-AIS recurrence group (p<0.05). There was no statistical difference in gender and age between the two groups.

Table 1 The Demographic and Baseline Clinical Characteristics of the Patients

Assessment of the Predictive Value of AIS Recurrence

Serum HMGB1 levels and NIHSS used to assess the severity of AIS were used for ROC curve analysis (Figure 2). Based on ROC curve analysis, HMGB1 was a significant predictor of AIS recurrence with an area under curve (AUC) of 0.745 (95% CI: 0.663–0.826). The AUC of NIHSS for predicting AIS recurrence was 0.822 (95% CI: 0.758–0.886). The combined HMGB1 and NIHSS had an AUC of 0.855 (95% CI: 0.800–0.911) for predicting AIS recurrence, which has a higher predictive value than either factor (p<0.05).

Figure 2 Receiver operating characteristic curve for predicting AIS recurrence. Receiver operating characteristic curves were used to determine the area under the curve for each variable and z-test comparisons were performed. Abbreviations: AIS, acute ischemic stroke; HMGB1, high mobility group protein box 1; NIHSS, National Institute of Health stroke scale.

Survival Analysis of Factors Affecting AIS Recurrence

The optimal cut-off value for HMGB1 (13.62µg/L) was then calculated by the Jordon’s index, based on which the patients were divided into a high HMGB1 level group and a low HMGB1 level group for survival analysis. Kaplan-Meier analysis revealed a significant difference between the cumulative survival rates without AIS recurrence between the two groups (Figure 3). The cumulative survival free of AIS recurrence was significantly lower in patients in the high HMGB1 level group than in the low HMGB1 level group (p<0.05).

Figure 3 Kaplan-Meier curves showing the probability of no recurrence in patients with AIS at different HMGB1 levels. Kaplan-Meier survival curves were used to conduct survival analysis which were then compared using the Log rank test. Abbreviations: AIS, acute ischemic stroke; HMGB1, high mobility group protein box 1.

The univariate Cox analysis found that elevated baseline serum HMGB1 levels were as much a predictor of AIS recurrence as a history of previous hypertension, diabetes mellitus, and atrial fibrillation, as well as elevated LDL-C levels and high NIHSS (Table 2). Subsequent multifactorial cox analyses of statistically different metrics in the univariate cox found that elevated baseline HMGB1 (HR:7.489, 95% CI:4.383–12.795) and LDL-C (HR:2.433, 95% CI:1.435–4.124) levels and high NIHSS (HR:7.904, 95% CI:4.508–13.858), and a history of atrial fibrillation (HR:2.383, 95% CI:1.384–4.103) remained predictive factors for recurrence of AIS.

Table 2 Univariate and Multivariate COX Analysis for Recurrence of AIS

Discussion

Previous studies have found that plasma HMGB1 levels are significantly higher and positively correlated with NIHSS in patients with AIS.^18,24 Another study found HMGB1 to be the best biomarker for diagnosing and predicting AIS severity.²⁵ This is similar to our study. In the present study, serum HMGB1 levels were found to be significantly elevated in AIS recurrence. HMGB1 had the same value as NIHSS in predicting AIS recurrence. The predictive value of HMGB1 was even greater when combined with NIHSS. HMGB1 was similar to LDL-C, NIHSS, and history of atrial fibrillation as predictors of AIS recurrence.

In AIS patients, large amounts of HMGB1 are secreted by ischemic brain tissue and released extracellularly to promote an inflammatory response to exacerbate brain damage, and large amounts of HMGB1 can also be detected in ischemic brain tissue in the chronic phase.^26,27 Elevated HMGB1 can further activate platelets and induce thrombosis, exacerbating brain tissue injury.²⁸

The release of HMGB1 from ischemic brain tissue may be the main reason for the elevation of HMGB1 after ischemia.²⁹ Moreover, it remains elevated during the subacute phase, making it a time-sensitive marker of brain injury and recovery.³⁰ Factors responsive to systemic inflammatory responses such as CRP, IL-6 and TNF-a have been found to be associated with long-term prognosis and recurrent events in AIS.^31,32 However, the temporal dynamics and specificity of HMGB1 provide additional prognostic value, which makes HMGB1 a more relevant predictor of AIS recurrence.

Methods commonly used as clinical treatments for AIS such as intravenous thrombolysis and endovascular thrombectomy affect HMGB1 levels. Although intravenous thrombolysis can initiate reperfusion, it may also lead to HMGB1 release caused by reperfusion injury, elevating HMGB1 levels.¹⁸ In contrast, patients with successful reperfusion via thrombectomy typically exhibit lower HMGB1 levels.³³ However, in cases of incomplete reperfusion, HMGB1 levels may remain elevated, indicating continued tissue damage.

HMGB1 is not only a prognostic marker for AIS, but is also considered an important therapeutic target.³⁰ In a rat model of cerebral ischemia-reperfusion, HMGB1 was massively translocated to the cytoplasm and its plasma level was significantly elevated, whereas inhibition of HMGB1 reduced the volume of cerebral infarcts and ameliorated neurological deficits.³⁴ In addition, inhibition of HMGB1 effectively attenuated neuroinflammation, reduced infarct volume, and promoted motor function recovery.³⁵ In a rat model of ischemic stroke, HMGB1 inhibitor significantly reduced mortality, attenuated cerebral hemorrhage, brain swelling, blood-brain barrier injury and neuronal apoptosis, and improved the prognosis of neurological function.³⁶ A large number of studies have found that some clinically used drugs such as atorvastatin, berberine, minocycline and diosgenin can inhibit the expression and transport of HMGB1, thereby suppressing inflammation, reducing neuronal damage, decreasing infarct size and improving neurological function.^37–40

This study has a few limitations. Firstly, it was a single-centre study. Case ascertainment was carried out in only one hospital, which may have introduced some selection bias. Secondly, serum HMGB1 levels were only measured at admission, not at the time of the patient’s AIS relapse, which did not allow for a before and after control study. In addition, this study only investigated the relationship between serum HMGB1 levels and AIS recurrence. However, the association between elevated serum HMGB1 levels and poor functional prognosis and death in AIS was not investigated, as previous studies have shown that elevated HMGB1 levels were associated with poor prognosis in patients with AIS.^41,42 Therefore, large, multi-centre, long-term clinical trials are needed to confirm the association between HMGB1 and AIS recurrence.

Conclusion

In the present study, we found that AIS relapsers had higher serum HMGB1 levels. Elevated serum HMGB1 levels at baseline had a high predictive value for AIS recurrence, which was even higher when combined with NIHSS. HMGB1 along with multiple factors LDL-C, NIHSS, and history of atrial fibrillation were predictors of AIS recurrence. This study provides a foundation for the clinical application of serum HMGB1 levels in AIS management, particularly in risk stratification, early warning, individualized treatment, and prognosis. Future studies could focus on validating the effectiveness of HMGB1 as a predictive biomarker, particularly a diagnostic tool combining HMGB1 with other established predictors, and exploring its applicability in different patient populations.