Introduction

Poststroke cognitive impairment (PSCI) is a syndrome characterized by cognitive impairment that occurs within six months after a stroke. This condition is associated with negative functional outcomes and even mortality1. The prevalence of PSCI varies widely, ranging from 20 to 80%, due to differences in geographic regions, ethnicity, and diagnostic criteria2. Research has shown that even mild strokes can increase the risk of cognitive impairment and impact the quality of life for patients3. Early identification and intervention are crucial for delaying disease progression. Therefore, it is essential to identify PSCI patients early using biomarkers associated with neuronal damage and promptly initiate interventions.

Neurofilaments (Nfs) are essential proteins that are synthesized in neuronal cell bodies and then assembled into intermediate filaments within axons. They play a critical role in maintaining the shape and stability of axons. The three main types of neurofilaments are neurofilament light chain (NfL), medium chain (NfM), and heavy chain (NfH)4. Extensive research has been conducted on NfL, and there is evidence suggesting that phosphorylated NfL (pNfL) is an independent risk factor for post-stroke cognitive impairment (PSCI) and cognitive decline one year after stroke5,6. On the other hand, NfH phosphorylation determines the diameter of axons and protects neurofilament proteins from degradation. When there are defects in NfH phosphorylation in neurons, it can lead to blockage of axonal transport and cell death. p-NfH serves as a reliable serum biomarker for axonal injury7,8. NfH may better reflect the pathogenic rationale of PSCI compared with such previous blood markers as oxidative damage biomarkers, inflflammatory factors, growth factors, and metabolic biomarkers9. Proteins can now be tested in serum and plasma other than in CSF, thanks to the development and implementation of new detection technique. Moreover, it exhibits relative resistance to protein degradation compared to other neurofilaments10.

The expression of p-NfH, a neuronal protein, has been found to increase in various central nervous system diseases, such as acute ischemic stroke (AIS), cerebral small vessel disease, Alzheimer’s disease, frontotemporal dementia, Lewy body dementia, traumatic brain injury, multiple sclerosis, amyotrophic lateral sclerosis, and opticospinal atrophy11,12,13,14,15,16. This protein plays a critical role in distinguishing between diseases and understanding the progression of axonal damage over time. Studies have indicated that post-stroke neuronal damage leads to the release of neuronal proteins into cerebrospinal fluid (CSF) and blood, suggesting that NfH can serve as an objective marker for AIS and axonal damage4. However, studies on the relationship between p-NfH and PSCI are still limited, So far, there has been no published data on NfH in patients with PSCI.Therefor, this study aims to explore the relationship between acute phase p-NfH levels and PSCI, as well as cognitive decline one year later, and assess the potential of p-NfH as a predictive indicator for PSCI.

Materials and methods

Study participants

From July 2020 to September 2021, patients with Acute Ischemic Stroke (AIS) treated at the Department of Neurology, Second Hospital of Hebei Medical University, were included in the observation group. Among them, 3 underwent thrombolysis, while none received thrombectomy. The baseline demographic and clinical data of all participants are shown in Table 1. Additionally, 30 healthy volunteers were selected as the control group. Inclusion Criteria for the AIS Group: (1) Age between 40 to 100 years old. (2) Diagnosed with AIS according to the "Chinese Guidelines for Diagnosis and Treatment of Acute Ischemic Stroke 2018" and confirmed by head Diffusion Weighted Imaging (DWI). (3) Time from onset of symptoms to hospital admission ≤ 7 days. (4) Informed consent from the patient or their family. (5) Right-handed.

Table 1 Baseline demographics and clinical date of all participants.
Full size table

Inclusion Criteria for the Control Group: (1) Age between 40 to 100 years old. (2) No complaints of cognitive impairment. (3) Informed consent from the individual. (4) Right-handed. Exclusion Criteria: (1) History of cerebrovascular diseases and other conditions leading to cognitive impairment (e.g., Alzheimer’s disease, frontotemporal dementia, Parkinson’s disease, traumatic brain injury, etc.) (2) Hearing impairment, visual impairment, poor cooperation. (3) Weak muscle strength leading to poor cooperation. (4) Complete aphasia. (5) History of psychiatric disorders. (6) Severe neurological disorders. (7) Severe lung, heart, liver, kidney dysfunction, malignant tumors, and other diseases. (8) Refusal to participate by patients or their family.

Data collection

The collected data included gender, age, hypertension, diabetes, smoking and alcohol history, admission NIHSS score, head MRI + MRA + DWI radiological examination, TOAST classification based on vascular involvement, and white matter grading according to the Fazekas scale (0 = no white matter hyperintensities, 1 = punctate foci of Flair hyperintensity, 2 = confluent Flair hyperintensity, 3 = irregular Flair hyperintensity)17, Studies have shown that mild leukodystrophy has little impact on cognitive function18, so patients with grade 0 and 1 were grouped together for analysis, and patients with grade 2 and grade 3 were combined for analysis. Additionally, brain infarct volume was calculated using the Pullicino formula: length × width × number of DWI high signal layers/219.

Cognitive status before stroke was assessed using the Short Form of the Informant Questionnaire on Cognitive Decline in the Elderly (Short IQCODE)20. The average IQCODE score < 3.19 is considered as cognitive normal21. The patients who underwent neuropsychological assessments were followed up at 6 and 12 months post-stroke. At 6 months, a comprehensive cognitive assessment was conducted using the Montreal Cognitive Assessment (MoCA) to assess overall cognitive function, Digit Span to evaluate attention, Auditory Verbal Learning Test to assess patient memory, ROCF (Rey-Osterrieth Complex Figure) to evaluate visual-spatial function, semantic fluency (fruit, vegetable, animal) to assess language, executive function, and memory, and the Stroop Color-Word Test to assess attention and executive function. The Boston Naming Test was used to evaluate language and visual-spatial abilities. Cognitive impairment was defined as results falling 1.5 standard deviations below or above the standardized mean. Based on the neuropsychological battery, impairment in at least one domain was diagnosed as PSCI22. At 12 months, the MoCA was used to assess overall cognitive function, with a decrease in MoCA score indicating progression, while a stable or increased MoCA score indicating stability.

Measurement of p-NfH

Fasting venous blood (5 mL) was collected from all AIS patients and healthy volunteers in the early morning. The blood was centrifuged at 1500 g for 10 min to separate the serum, which was then stored at − 80 °C for future testing. p-NfH levels were measured using an ELISA kit according to the manufacturer’s instructions.

Statistical analysis

G*Power software was used to calculate the sample size with a test efficiency of 0.95 and a significance level of 0.05. Statistical analysis was performed using SPSS 24.0. Continuous variables were expressed as mean ± SD or median (interquartile range), and normality was tested using the Kolmogorov–Smirnov test. Normally distributed data were analyzed with independent two-sample t-tests, while skewed data were analyzed using the Mann–Whitney U test. Categorical data were presented as frequencies (proportions) and analyzed with the chi-square test. Binary logistic regression was applied to compare clinical characteristics and p-NfH levels between the PSCI and NPSCI groups, as well as between the progression and stable groups. Spearman correlation analysis was used to assess the correlation between p-NfH, NIHSS, and infarct volume. A two-tailed p-value of < 0.05 was considered statistically significant.

Results

Comparison between the PSCI and NPSCI group

Based on a comprehensive assessment across multiple cognitive domains, 31 patients were diagnosed with PSCI (53.4%). PSCI patients were older, had lower educational levels, larger stroke volumes, higher NIHSS scores, and more severe white matter lesions. A comparison of clinical characteristics between the PSCI and NPSCI groups is shown in Table 1.

Comparison of p-NfH Levels between AIS and control group

All 72 participants completed initial clinical and neuropsychological assessments, along with a brain MRI, within 7 days of stroke onset. Fourteen patients were lost to follow-up, but no significant differences were observed between the baseline characteristics of participants who completed the study and those lost to follow-up. Data from 58 patients were included in the final analysis (Fig. 1). Additionally, 30 healthy volunteers were enrolled as controls. p-NfH levels in the AIS group were significantly higher than those in the control group (p < 0.01) (Table 2).

Fig. 1
Fig. 1
Full size image

Flowchart of patient enrollment.

Table 2 Baseline demographics and clinical date between the stroke and control group.
Full size table

Correlation analysis of p-NfH with stroke characteristics in the AIS group

The correlation between p-NfH levels in patients and infarction volume, NIHSS scores, and age was examined. The results showed that p-NfH levels were positively correlated with NIHSS scores (Fig. 2) and infarction volume (Fig. 3). There was a trend towards correlation with age, although it did not reach statistical significance (r = 0.249, p = 0.06).

Fig. 2
Fig. 2
Full size image

Correlation between p-NfH and NIHSS Scores.

Fig. 3
Fig. 3
Full size image

Correlation between p-NfH and infarct Volume.

Comparison of p-NfH levels between patients with Fazekas grade 2 and 3 white matter lesions and those with grade 0 and 1 lesions

The results (Table 3) indicated that patients with moderate to severe white matter lesions had significantly higher p-NfH levels compared to those with no or mild white matter lesions.

Table 3 Comparison of p-NfH Levels between the No or Mild White Matter (Grade 0 and 1) and the Moderate to severe group (Grade 2 and 3).
Full size table

Binary logistic regression for PSCI

Factors with a p-value less than 0.1 were included in the binary logistic regression model. The results showed that elevated p-NfH levels were independently associated with the occurrence of PSCI at 6 months, with an adjusted odds ratio (OR) of 1.06 (95% CI 1.004–1.11, p = 0.033). Additionally, age remained statistically significant (p = 0.038) (Table 4).

Table 4 Logistic regression analysis of risk factors for PSCI at 6 months.
Full size table

Comparison between the progression and stability group

A total of 54 AIS patients completed the MoCA assessment at 12 months. Among them, 26 patients showed stable or improved MoCA scores (stability group), while 28 patients had a decline in scores (progression group). The progression group had a higher proportion of females, older age, lower education levels, and more severe white matter changes. Factors with a p-value < 0.1 were included in a logistic regression model, which revealed that females were more likely to experience a decrease in MoCA scores compared to males. A comparison between the progression and stability groups is shown in Table 5.

Table 5 Comparison between the cognitive decline and non-cognitive decline group patients at 12 months.
Full size table

P-NfH predicting PSCI

ROC curves were plotted (Fig. 4), and the Youden index was calculated to determine the optimal predictive threshold for p-NfH. The ROC curve showed an Area Under the Curve (AUC) of 0.881, with a 95% confidence interval (CI) of 0.791–0.970. The optimal cutoff point was determined to be 166.03 pg/ml, with a sensitivity of 0.774 and specificity of 0.926 (p < 0.01).

Fig. 4
Fig. 4
Full size image

ROC Curve for p-NfH in Diagnosing PSCI.

Discussion

In this study, 58 patients who suffered from AIS were followed for cognitive function for one year. The analysis of the collected data yielded several key findings: (1) During the acute phase, serum p-NfH levels were significantly elevated in stroke patients compared to healthy individuals, and p-NfH was correlated with stroke volume and NIHSS scores. (2) Older Age, lower educational level, higher NIHSS scores, and more severe brain white matter lesions were associated with the occurrence of PSCI. (3) Despite having adjusted for relevant risk factors, p-NfH levels were still significantly higher in PSCI patients, indicating that p-NfH is an independent risk factor for PSCI. (4) p-NfH was also associated with cognitive decline progression from 6 to 12 months after stroke. Integrating p-NfH into a biomarker model that includes conventional risk factors (age, education level, hypertension, diabetes, and NIHSS scores) significantly improved predictive performance.

The core finding of this study is that p-NfH, as a biomarker, has potential in predicting PSCI. During AIS, reduced blood supply to the brain triggers various neuropathophysiological processes, resulting in irreversible neuronal damage. It has been observed through radiological studies that axonal spheroids are commonly present in AIS patients, indicating axonal injury23. This injury disrupts the structural integrity of cellular cytoskeletal components, leading to the release of p-NfH protein. The released neurofilaments can enter the bloodstream directly or through the lymphatic channels of the brain10,24. The previous studies had shown that the serum levels of neurofilaments (NfH) increase after axonal injury25, and this increment may reflect the destruction of neuronal structure after axonal injury, further leading to impaired cognitive function. This study revealed elevated levels of p-NfH in AIS patients, which exhibited significant correlations with the NIHSS scores and stroke volume. Furthermore, AIS patients with moderate to severe brain white matter lesions had higher serum p-NfH levels.

Iadecola C has established that age is a significant risk factor for PSCI, particularly in individuals over the age of 65, where the occurrence of PSCI rises dramatically26. Additionally, Pendlebury ST and Rost NS have identified being female and having a lower educational background are risk factors for cognitive decline. Furthermore, characteristics of brain white matter lesions and acute ischemic stroke (AIS), such as NIHSS scores and stroke volume, also contribute to the development of PSCI27,28. The outcomes of this study are consistent with prior research, underscoring that advanced age, lower education level, stroke volume and NIHSS scores are associated with PSCI. Moreover, the female gender is a risk factor for the progression of cognitive decline from 6 to 12 months after stroke.

Brettschneider, Wilke C and van der Ende EL discovered elevated levels of p-NfH in the CSF of patients diagnosed with Alzheimer’s disease, vascular dementia, and frontotemporal dementia13,14,29. Furthermore, they also observed elevated p-NfH levels in the serum of patients who with frontotemporal dementia14,29. Additionally, heightened levels of p-NfH were found in the serum of patients with primary progressive aphasia, making it a potential distinguishing factor between different subtypes of this condition30. In summary, p-NfH can serve as a reliable marker for neurodegenerative lesions. Longitudinal imaging studies have found that stroke survivors exhibit secondary neurodegenerative changes, characterized by degeneration in brain regions distant from the infarcted area31. A MRI study conducted by Tiedt et al. revealed extensive white matter changes in the hemisphere affected by the stroke six months post-event. These changes were found to be associated with the levels of NfL, indicating the occurrence of axonal degeneration after stroke32. The cascading reactions of axonal injury, activation of neurodegenerative pathways, vascular damage, and various other injuries collectively impact the development of PSCI33. Ding et al. found that hyperbaric oxygen therapy could improve cognitive function in ischemic stroke mice by reducing p-NfH levels34. While p-NfH has been associated with various neurological conditions14, its relationship with PSCI had not been previously reported. In our study, we conducted serum p-NfH tests in 58 AIS patients, and our results demonstrated elevated p-NfH levels in PSCI patients and those who showed cognitive decline one year after stroke, suggesting that p-NfH is an independent risk factor for PSCI. This is the first evidence that p-NfH can serve as a reliable blood biomarker for predicting PSCI, showing significant accuracy in distinguishing between PSCI and non-PSCI patients.

Neuropathology associated with white matter hyperintensities includes axonal injury, white matter rarefaction, ischemia, and myelin breakdown33. We found that p-NfH was associated with NIHSS scores, stroke volume, and brain white matter scores. Therefore, p-NfH partly influences the occurrence of PSCI due to brain white matter lesions and stroke volume. Even after adjusting for these risk factors, our study found that p-NfH levels remained difference between the PSCI and non-PSCI groups, reaffirming p-NfH as an independent risk factor. Thus, p-NfH can serve as a predictive indicator for PSCI.

AIS can trigger secondary neurodegenerative changes in brain regions remote from the infarcted area, such as white matter tracts connected to the infarcted area. Secondary damage may lead to cognitive impairment and disruptions in neurotransmitter synthesis, contributing to PSCI6. Synaptic plasticity is closely related to the recovery and improvement of cognitive function, as increased synaptic connection strength and enhanced synaptic transmission efficiency directly upregulate information processing and storage in the central nervous system, thereby improving cognitive function35. Impaired synaptic plasticity after stroke may be a contributing factor to PSCI. Previous studies have revealed the presence of NfH has been linked to the strength of postsynaptic connections and presynaptic activity36. In addition, MMP-9 levels played a crucial role in the occurrence and development of PSCI37, while MMP-9 may affect NfH degradation10, which could be one of the reasons why p-NfH influences the occurrence of PSCI.

Limitations

First, we did not comprehensively assess cognitive function at baseline before the stroke. However, patients with known cognitive impairment or neurodegeneration and impairment in daily activities before stroke were excluded at screening on the basis of medical history and the IQCODE questionnaire. In the future, we will include a comprehensive baseline cognitive assessment as a key component and further elucidate the relationship between prestroke cognitive function and PSCI. Second, the sample size of this study is limited. However, we took into account infarction size, NHISS score, age, education level and other factors related to cognition, conducted comprehensive cognitive assessment which demonstrated the relationship between NfH and PSCI. We will aim to recruit a larger, multi-center cohort with a more diverse participant pool to ensure broader applicability of the findings. Third, patients with complete aphasia, massive cerebral infarction, and other serious diseases were excluded in this study. The severity of stroke in the patients in this study was relatively small, and the proportion of patients with lacunar infarction was relatively high. Previous studies have shown a positive correlation between NfH and NIHSS score10, so the correlation between NfH and PSCI may be higher in patients with severe stroke. Further studies involving patients with different infarct volumes are needed. Fourth,We did not include other confounding factors such as genetic predisposition and lifestyle, which may have had an impact on the results. However, we adjusted for age, education level, NIHSS score, etc. In the future, we will combine genetic testing and detailed lifestyle data to more fully understand the predictive effect of p-NfH on PSCI. Finally, our study followed for 1 year, and we need longer follow-up in the future to discover the long-term predictability of p-NfH.

In conclusion, our findings suggest that p-NfH levels may serve as a reliable biomarker for predicting post-stroke cognitive impairment (PSCI). This has significant clinical implications, as early identification of patients at high risk for PSCI allows clinicians to implement timely interventions, such as cognitive rehabilitation programs, lifestyle modifications, or pharmacological treatments, to potentially mitigate cognitive decline.