Acute ischemic stroke (AIS) is one of the leading causes of death and long-term disability worldwide [1]. Intravenous thrombolytic therapy is a key ultra-early treatment for AIS, and the application of recombinant tissue plasminogen activator has significantly improved the prognostic outcomes [2]. Despite these advances, early neurological deterioration (END) after thrombolysis remains a major clinical concern, occurring in about 10%–30% of patients, and is strongly associated with poor functional prognosis and increased mortality [3, 4]. The mechanisms underlying END pathogenesis are complex and involve a variety of pathophysiological processes, including reperfusion injury, recurrent thrombosis, hemorrhagic transformation, and the inflammatory responses [5, 6].

In recent years, the role of neuroelectrophysiological measures and immune-inflammatory indicators has received increasing interest in stroke evaluation. Amplitude-integrated electroencephalography (aEEG), a convenient approach for continuous monitoring of brain function, can sensitively reflect the functional status of the cortex and is particularly valuable for evaluating disorders of consciousness and predicting neurological prognosis [7]. Additionally, the CD4+/CD8+ T-cell ratio, a crucial indicator of immune status, has been linked to post-stroke inflammatory responses and recovery of neurological function [8, 9].

Studies indicate that abnormal aEEG patterns are closely associated with the immune-inflammatory responses, contributing to the pathology of END [10, 11]. However, most existing predictive models for END rely heavily on clinical scores (such as the National Institutes of Health Stroke Scale [NIHSS]) and neuroimaging findings [12, 13]. While valuable, these approaches may not fully capture the dynamic interplay between neuroelectrophysiological dysfunction and systemic immune inflammation. Although inflammatory markers such as the Systemic Inflammation Response Index (SIRI) and neutrophil-to-lymphocyte ratio have been associated with stroke prognosis [14], their integration with neurofunctional monitoring tools like aEEG remains underexplored. Integrating aEEG, which directly and continuously reflects cortical functional integrity, with the CD4+/CD8+ ratio, a key indicator of immune homeostasis and T-cell-mediated regulation, may provide a more holistic and pathophysiologically informed framework for risk stratification in END.

We hypothesized that this integrated approach would capture complementary aspects of END pathogenesis, such as real-time neuronal excitability and the systemic inflammatory-immune balance. Therefore, the purpose of this study was to evaluate whether integrating aEEG scores with the CD4+/CD8+ ratio offers incremental diagnostic value beyond established risk factors for END after thrombolysis, thereby supporting earlier and more accurate identification of high-risk patients for individualized interventions.

The END after thrombolytic therapy for AIS is a major unresolved concern in clinical practice. At present, the clinical approaches to predict END are still suboptimal. Although the established predictors, such as baseline NIHSS score and imaging findings, have certain predictive applications, their sensitivity and specificity are limited. Furthermore, the pathophysiology of END is highly complex, involving interaction among multiple factors, including ischemia-reperfusion injury, thrombosis, hemorrhagic transformation, and inflammatory response [20, 21]. In particular, the inflammatory response, which serves as a crucial interplay between ischemic injury and subsequent neurological deterioration, has not been fully elucidated in the context of END. Traditional inflammatory markers, such as C-reactive protein and white blood cell count, reflect only a single dimension of the inflammatory response and are inadequate to assess the complex inflammatory-immune regulatory network of the host.

In clinical practice, there is an urgent need for a simple, economical, and efficient biomarker that can accurately identify patients at high risk of END immediately after thrombolysis, thereby facilitating timely intervention. Therefore, exploring the diagnostic value of novel inflammation-related indicators for END is not only theoretically essential but also clinically relevant. Given this unmet clinical need, this study was designed to systematically evaluate the predictive efficacy of aEEG score and the CD4+/CD8+ ratio for END after thrombolysis in AIS patients, aiming to provide new insights and evidence for early risk stratification and the early warning of END.

Amplitude-integrated EEG (aEEG), a simple approach for continuous monitoring of brain function, can sensitively reflect the functional status of the cortex. The higher aEEG score indicate more severe cerebral injury. The correlation between aEEG score and END likely reflects its ability to quantitatively evaluate the degree of inhibition of neuronal electrical activity after cerebral ischemia. Cerebral ischemia-induced impairments in neuronal energy metabolism destabilize the cell membrane potential and lead to weakening or loss of electrical activity. aEEG can identify neuronal dysfunction at an early stage by monitoring integrated changes in EEG amplitude [7].

In this study, patients in the END group showed significantly higher aEEG scores than those in the non-END group, and multivariate logistic regression analysis revealed that a higher aEEG score was an independent risk factor for END. These observations suggest that aEEG is crucial for evaluating neurological impairment and predicting the occurrence of END. The decreased CD4+/CD8+ ratio, as a key predictor of immune status, may indicate impaired immune regulatory function, aggravate inflammatory responses, and consequent deterioration of neurological function. A reduction in this ratio reflects a shift in the T helper/T cytotoxic (Th/Tc) cell balance toward a pro-inflammatory state. The relative increase in CD8+ cytotoxic T-cells may aggravate neural injury by directly targeting neurons within the ischemic penumbra or by promoting the release of inflammatory mediators [8, 9].

In our study, the CD4+/CD8+ ratio was significantly lower in the END group than in the non-END group, and this higher ratio emerged as a protective factor against END. Furthermore, the combined application of aEEG score and CD4+/CD8+ ratio further improved the predictive efficacy for END (AUC = 0.862), indicating that combining EEG neurophysiological evaluation with immune-inflammatory indicators provides more comprehensive and informative risk stratification for early END identification.

The mechanisms underlying END in AIS patients after thrombolysis are complicated and likely reflect the synergistic effect of multiple factors. A study by Zhang et al. [22] showed that the time from onset to thrombolysis and white blood cell count were linked to the occurrence of END after thrombolysis in AIS patients. The present study similarly confirmed that prolonged time from onset to thrombolysis and increased white blood cell count at admission were independent risk factors for END after thrombolysis in AIS patients. The impact of delayed thrombolysis on END is primarily attributed to irreversible damage within the ischemic penumbra and the aggravation of reperfusion injury [23, 24]. With prolonged ischemia time, the core area of infarction expands, metabolic impairments of penumbral neurons aggravate, and mitochondrial dysfunction leads to excessive accumulation of reactive oxygen species. When the blood flow is restored after thrombolysis, reperfusion triggers a sudden increase in oxidative stress. Reactive oxygen species activate the nuclear factor kappa B (NF-$\kappa$B) pathway, promoting the release of inflammatory factors such as interleukin-6 (IL-6) and tumor necrosis factor $\alpha$ (TNF-$\alpha$), further compromising the integrity of the blood-brain barrier.

The study also found a high NIHSS score at admission as a risk factor for END after thrombolysis in AIS patients, consistent with the results of Liu et al. [25]. An increased NIHSS score at admission suggests more severe neurological deficits, and the mechanism underlying END may involve the following aspects [26, 27]: (1) Larger infarct size; high NIHSS scores often reflect extensive infarction or involvement of key functional areas such as cortex and basal ganglia, which are more prone to neurological deterioration due to edema or microcirculatory dysfunction after reperfusion. (2) Insufficient collateral circulation compensation: patients with severe defects show poor collateral blood flow and are more prone to ineffective reperfusion, for example, persistent microvascular occlusion, after thrombolysis, which aggravates ischemia-reperfusion injury. (3) Higher susceptibility to secondary injury: patients with high NIHSS scores have compromised cerebral metabolic reserve, which reduces tolerance to secondary injury, such as inflammatory factors (e.g., TNF-$\alpha$) and oxidative stress.

SIRI is a comprehensive inflammatory marker which reflects the inflammatory state by integrating neutrophil, monocyte, and lymphocyte counts [28]. Li et al. [29] reported that high SIRI was an independent risk factor for END in patients with minor ischemic stroke. Consistent with these observations, our study also confirmed that SIRI levels were substantially higher in the END group than in the non-END group, and multivariate logistic regression analysis showed increased SIRI as an independent risk factor for END. The diagnostic value of SIRI for END may be due to its involvement in multiple aspects of underlying pathophysiological processes.

Neutrophils, as the first-line effector cells in the inflammatory response, rapidly infiltrate brain tissue after ischemia-reperfusion, where they release reactive oxygen species and proteases, directly destroying the blood-brain barrier and exacerbating brain edema. After ischemia, monocytes differentiate into macrophages and amplify the inflammatory cascade by secreting pro-inflammatory factors such as IL-1$\beta$ and TNF-$\alpha$. In parallel, lymphocyte decline reflects immunosuppression and decreased anti-inflammatory capability [14, 30]. By integrating these three cell types, SIRI more sensitively captures the pathological state characterized by the coexistence of an inflammatory storm and immune paralysis.

To further assess the incremental diagnostic value of aEEG scores and the CD4+/CD8+ ratio, two predictive models were constructed in this study. Model 1, which excluded aEEG score and CD4+/CD8+, demonstrated an AUC of 0.944, whereas the AUC of Model 2, which combined both aEEG score and CD4+/CD8+, increased to 0.991. Comparison of these models showed that integrating aEEG score and CD4+/CD8+ improved predictive performance, with sensitivity increasing from 87.21% to 91.34% and specificity from 87.40% to 100%. These observations indicate that aEEG scores and CD4+/CD8+ ratios provide predictive information beyond traditional risk factors, and their inclusion improves model discrimination and clinical applicability. However, the ROC curves of the two models showed relatively fewer inflection points, suggesting certain limitations in risk stratification, which may be related to the predominant use of continuous variables. Future studies should optimize the model’s stratification capability by introducing categorical variables or applying more detailed categorization of continuous variables.

This study has several limitations that should be acknowledged before interpreting these findings. First, its single-center, retrospective design may introduce potential selection bias and limit the generalizability of the findings, as the characteristics of our patient population and local clinical practices may not fully represent those at other institutions or regions. Second, although internal validation techniques (e.g., bootstrapping and cross-validation) were used, the model’s exceptional performance (e.g., AUC of 0.991) should be interpreted with caution. In the absence of a truly independent, external validation cohort, the risk of overfitting cannot be excluded, and the general applicability of the model requires further confirmation. Furthermore, the aEEG scoring system (1–5 scale) used in this study was adapted from scales originally validated in neonatal encephalopathy. While it provides a structured framework for quantifying background EEG suppression, its application and interpretation in adult AIS patients require further validation. The underlying pathophysiology of brain injury in adult stroke differs from that in neonates, which may affect the clinical interpretation of these scores. Therefore, our findings regarding the aEEG score should be interpreted with caution. Future prospective, multi-center studies with larger sample sizes are required to validate the robustness and clinical utility of the model and to further elucidate the predictive contribution of aEEG and immune markers in END.

In summary, this study confirms that both the aEEG score and the CD4+/CD8+ ratio are independent predictors of END in AIS patients after thrombolysis, and combining these two significantly improves predictive efficacy. The multifactorial prediction model incorporating these variables showed excellent discriminative ability (AUC = 0.991) and holds potential clinical transformation value. The implementation of individualized monitoring and early intervention is advisable for patients with these high-risk factors, including improved neurological assessment, optimized blood pressure management, and rational use of anti-inflammatory and neuroprotective treatments. Future multi-center and large-sample studies are needed to further validate the clinical applicability of this model and to explore targeted intervention strategies based on aEEG and immune markers to improve neurological prognosis after thrombolysis in AIS.

• This study identifies the aEEG score and the CD4+/CD8+ ratio as significant and independent predictors for END following thrombolysis in patients with AIS.

• The combination of the aEEG score, which reflects cortical functional integrity, and the CD4+/CD8+ ratio, which indicates systemic immune-inflammatory status, provides superior predictive performance for END than either marker alone, with a combined area under the curve of 0.862.

• A diagnostic model incorporating these novel biomarkers alongside established risk factors demonstrated outstanding discriminative ability (AUC = 0.991), highlighting its potential for clinical translation to identify high-risk patients.

• These findings underscore the critical interplay between neuroelectrophysiological dysfunction and the immune response in the pathogenesis of END, suggesting that integrated monitoring strategies could enable earlier intervention and improved prognostic outcomes.

The data used to support the findings of this study are available from the corresponding author upon request.

QR designed the research study. HW performed the research. LY and SZ collected the data. QR and JZ analyzed the data and drafted the manuscript. All authors contributed to the important editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

This study adhered to the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Shangyu People’s Hospital of Shaoxing (SYRY-REC-2024-A010). All patients provided written informed consent.

Not applicable.

This work was supported by the Shaoxing Health Commission Science and Technology Plan Project [Grant No. 2024SKY110] and the 2024 District-Level Social Development Science and Technology Project of Shangyu District, Shaoxing City [Grant No. 2024-10].

The authors declare no conflicts of interest.