Quantitative EEG as a diagnostic and prognostic tool in hemispheric stroke patients undergoing type A aortic dissection surgery

Abstract Objective The diagnostic and prognostic value of quantitative electroencephalogram (qEEG) parameters, specifically the symmetry of amplitude‐integrated electroencephalography (aEEG) and relative band power (RBP), in the postoperative stroke of the cerebral hemisphere following type A aortic dissection, remains an area of inquiry. Methods We analyzed and processed 56 patients with type A aortic dissection who underwent bedside qEEG monitoring and analyzed the qEEG indices, brain CT, and clinical data of these patients. qEEG (symmetry of aEEG and RBP, and affected/unaffected hemisphere) indices were analyzed at discharge and 60 days after discharge. Results A total of 56 patients were studied. The 60‐day mortality rate was 12.5%. The affected hemisphere's diagnosis and mortality after 1‐year follow‐up were evaluated, and RBP beta demonstrated the highest area under the curve values with 95% confidence intervals (CI) of .849 (95% CI: .771–.928) and .91 (95% CI: .834–.986), respectively. According to the results of the logistic regression analysis, we have identified the strongest predictors for cerebral hemisphere stroke and 1‐year mortality in stroke patients. Specifically, aEEGmin exhibited the highest predictive power with an odds ratio (OR) of .735 for cerebral hemisphere stroke, whereas DTABR was confirmed as one of the strongest predictors with an OR of 1.619 for 1‐year mortality in stroke patients, indicating a high level of reliability. Spearman correlation coefficients showed that aEEGmax and aEEGmin were positively correlated with Alberta Stroke Program Early CT Score (aEEGmax: rho = .50, p < .001; aEEGmin: rho = .44, p < .001). Conclusions QEEG has been proven to be a sensitive indicator for monitoring brain function and can be monitored continuously. It can help clinicians detect and treat these patients early and improve long‐term prognosis.


LIMITATIONS
This retrospective study presents some limitations that need to be considered. First, the sample size is limited, and there is no uniform timing for qEEG measurements among postoperative patients. Additionally, the lack of raw EEG data hinders the evaluation of stroke diagnosis and prognosis. To validate the diagnostic and prognostic effectiveness of qEEG in detecting brain injury after AAD, well-designed and large-scale clinical studies are required.

INTRODUCTION
Type A aortic dissection (AAD) is a grave condition that carries a mortality rate of 50% within 48 h, emphasizing the urgent need for emergency surgical intervention (Tang et al., 2021). Despite advancements in medical technology, the 30-day postoperative mortality rate for AAD has only decreased from 30% to 12.2% (Parikh et al., 2017).
Neurological complications remain a significant concern due to the concurrent low perfusion of brain tissue, extended periods of cardiopulmonary bypass (CPB), and deep hypothermic circulatory arrest (DHCA) (Centofanti et al., 2016).
Post-stroke complications are frequently associated with perioperative brain injury. Reduced cerebral blood perfusion caused by cerebral edema or stroke can lead to diminished oxygen and glucose supply, ultimately resulting in cerebral infarction (Berger & Hakim, 1986;Bossone et al., 2013). The diagnosis of ischemic stroke relies heavily on neuroimaging and clinical assessment (Jadhav et al., 2020). Techniques such as MRI or CT perfusion can identify ischemic cores and potentially salvageable hypo-perfused penumbras (Dashtbani Moghari et al., 2021). Unfortunately, these methods are not practical for monitoring the development of cerebral ischemia in the acute phase of patients at risk of transport after AAD surgery (Jordan, 2004). Electroencephalography (EEG) represents a noninvasive technology with high temporal resolution capable of a rapid assessment of transient brain function. It is highly sensitive to acute changes in cerebral blood flow and nerve metabolism, making it a valuable tool for bedside monitoring of brain function in emergency situations (Hellstrom-Westas & Rosen, 2005). The subacute and post-acute changes in EEG associated with ischemic stroke have been widely studied (Ajcevic et al., 2021).

qEEG acquisition and processing
This was a retrospective study that involved bedside qEEG monitoring of patients using a Nicolet Monitor brain function instrument (NicoletOne 5.9.4, Natus Neurology Incorporated). Each patient was monitored at least once for a minimum of 20 min. The study analyzed two parameters of qEEG, namely, the amplitude-integrated EEG (aEEG) and relative band power (RBP), which were used to assess and quantify changes in brain function after AAD surgery.
To evaluate the usefulness of qEEG in detecting cerebral infarction in AAD, we extracted all spectral parameters and exported them to Excel files. These files were used for further analysis to assess the differences in qEEG between the left and right brain in patients with cerebral infarction following AAD surgery.

STATISTICAL ANALYSES
In this study, variables were reported as either mean and standard deviation or median and range, depending on the distribution. The normal distribution of variables was evaluated using the Kolmogorov-Smirnov test. Categorical variables were compared using the Chi-square test.
The relationship between the Alberta Stroke Program Early CT Score F I G U R E 1 Flow chart of study.
(ASPECTS) and qEEG parameters was assessed individually using Spearman correlation analysis. A statistically significant result was defined as a p-value <.05. To identify the most relevant independent qEEG indices related to post-stroke outcomes, stepwise multiple regression analysis was conducted. The data analysis was performed using R software version 4.2.2.

RESULTS
Between November 2021 and April 2022, a total of 136 patients with AAD underwent aortic repair procedures. Among them, 80 patients were excluded from the study, including 77 who did not undergo qEEG, 2 who opted out of the operation due to diffuse brain edema, and 1 with poor qEEG image quality. Ultimately, 56 subjects were enrolled in this study, as shown in Figure 1. A total of 112 qEEG analyses were performed, with 56 conducted on the left cerebral hemisphere and 56 on the right. stay, and 60-day mortality data were comparable between the two groups (p < .05), except for cross-clamp, DHCA, and hospital stay. Table 2 shows that, besides theta waves, other qEEG parameters exhibited significant differences between the affected and unaffected groups. Violin plots of the qEEG indices for the affected and unaffected groups are presented in Figure 3. Table 3 presents the results in terms of the presence of stroke survivor and non-survivor groups after 1-year follow-up. In addition to aEEGmin and RBP theta, statistical differences F I G U R E 2 Asymmetry of amplitude-integrated electroencephalography (aEEG) in the stroke patients. The red arrow indicated that aEEG of F3-P3 started to decrease on the left cerebral hemisphere compared with aEEG on the right side, and the yellow arrow indicated large cerebral infarction on the left side.

F I G U R E 3
Violin plots of the quantitative electroencephalogram (qEEG) indices between unaffected and affected groups.
were observed in all other qEEG parameters. As Figure 4A,B shows, the highest area under the receiver operating characteristic of the RBP beta for stoke and death was 0.849 (95% confidence intervals [CI]: .771-.928) and .910 (95% CI: .834-.986), respectively.

DISCUSSION
The changes in brain activity during postoperative neurological impairment in AAD are believed to be linked to neurophysiological alterations that occur when brain tissue perfusion is inadequate (Rossini et al., 2004). This mechanism is known as neurovascular coupling (Stragapede et al., 2019). In particular, the decrease in CBF in the ischemic area results in changes in EEG activity, mainly an increase in delta frequency power and a reduction in alpha frequency power (Jordan, 2004). Extensive research has been conducted on EEG changes in the subacute and acute phases of ischemic stroke (Ajcevic et al., 2021;Finnigan et al., 2004Finnigan et al., , 2007Sheorajpanday et al., 2011;Wu et al., 2016).
The transition from ischemia to infarction in the acute stage of ischemic stroke is a rapid process, typically taking only a few minutes to several hours from the onset of hypoperfusion to cell death (Murri et al., 1998). EEG has been shown to be a sensitive indicator of brain injury and can detect early changes in neurological function (Ajcevic et al., 2021). Previous studies (Finnigan & van Putten, 2013;Finnigan et al., 2007) have demonstrated that qEEG technology is capable of quantifying the amplitude, frequency, and spatial distribution of electrical activity in the cerebral cortex.
Early detection of brain complications after surgery for AAD is critical to improve long-term prognosis (Bae et al., 2005). However, due to factors such as invasive mechanical ventilation and delayed awakening, patients may not be able to undergo an effective physical examination of the nervous system. Additionally, transporting patients for CT

F I G U R E 4
Comparison of receiver operating characteristic (ROC) curves to predict the stroke (A) and death (B) after 1-year follow-up between the quantitative electroencephalogram (qEEG) indices in this cohort.

F I G U R E 5
Matrix of Spearman's correlation coefficients between Alberta Stroke Program Early CT Score (ASPECTS) and nine quantitative electroencephalogram (qEEG) indices. A color-coded correlation scale is presented on the right of the plot. Based upon the scale, red ones stand for negative correlations and blue ellipses stand for positive correlations. Crossed out boxes illustrate insignificant correlations of a given variable with itself. Red star symbols represent statistical significance levels: "***" represents p < .001, "**" represents p < .01, "*" represents p < .05, no stars represent p > .05.
or MRI scans of the brain can increase risks and is not always feasible in a clinical setting (Smith-Bindman, 2010;Vilela & Rowley, 2017).
Therefore, alternative methods for early detection, such as qEEG, may be valuable for these patients.
Although continuous low voltage and electrocerebral silence on aEEG have been previously linked to poor prognosis, the use of leftright EEG activity asymmetry as a diagnostic and prognostic indicator for acute stroke has not been widely studied (Rundgren et al., 2006;Tao & Mathur, 2010;ter Horst et al., 2004). Our study shows that reduced EEG activity on the affected side results in bilateral EEG asymmetry, which can predict the extent of brain injury. The greater the asymmetry, the more severe the patient's condition, and the worse the long-term neurological prognosis. Additionally, RBP in qEEG uses color to indicate the proportion of δ, θ, α, and β in the trend map. In ischemic and hypoxic conditions, the proportion of δ waves increases in the affected brain, leading to noticeable asymmetry in both sides of F I G U R E 6 Kaplan-Meier survival curve for stroke and non-stroke patients. .000

TA B L E 2
Abbreviations: DAR, delta/alpha power ratio; DTABR, (delta + theta)/(alpha + beta) power ratio. the brain. This RBP-derived asymmetry may further aid in the diagnosis of acute stroke.
The Modified Rankin Scale and the Barthel Index are widely used tools to assess functional disability and dependence in patients with neurological conditions. However, due to missing data, these measures were not included in our retrospective study. In contrast, the ASPECTS scoring system is a widely recognized and reliable tool for assessing the severity of ischemic stroke (Menon et al., 2011;Pexman et al., 2001). This system requires only CT scan and has been extensively validated, providing crucial information about patient conditions and prognosis (Naylor et al., 2017). Our study revealed a linear correlation between aEEG and ASPECTS, suggesting that aEEG may serve as an important predictor of neurological prognosis. In particular, we found that aEEGmax and aEEGmin were linearly correlated with ASPECTS, underscoring the importance of aEEG as a crucial indicator of neurological outcomes. Overall, our findings highlight the value of accessible and effective diagnostic tools in the management of ischemic stroke patients.
Studies (Colombo et al., 2019;Lanzone et al., 2022) have shown that spectral exponent (SE), an index reflecting EEG slowing and quantifying power spectral density with power law decay, can be a valuable metric for assessing the neural physiological state of cortical circuits after focal ischemic injury. Comparing SE between the affected and unaffected hemispheres can aid in guiding neurorehabilitation efforts (Donoghue et al., 2020). Unfortunately, our qEEG equipment can only monitor SE for the entire brain and was therefore unable to be included in this study.
Previous research studies have suggested a negative correlation between alpha activity and stroke prognosis (Diedler et al., 2010;Schleiger et al., 2014). Additionally, studies of EEG activity in patients undergoing carotid endarterectomy have shown that alpha changes occur earlier than slow waves (Sharbrough et al., 1973). However, in the case of aortic type A dissection, DHCA is required, which results in the whole brain EEG being in a state of electrical resting inhibition, and bilateral brain RBP being mainly theta waves. During the awakening process in the intensive care unit, the changes in the overall pattern and proportion of theta waves in both sides of the brain were most prominent. This difference in findings from other studies may be attributed to these factors.
Our study revealed that patients with cerebral infarction had a lower 60-day survival rate compared to those without cerebral infarction, and this difference was statistically significant. Moreover, after 1-year of follow-up, the difference in survival rates remained significant, as shown in Figure 5. In contrast, the study has demonstrated that cerebral infarction increased the hospitalization rate of patients without affecting their long-term mortality (Bossone et al., 2013). However, our results may be attributed to the larger size of infarcts in our patient population, as well as lower medical compliance after discharge.

CONCLUSION
In summary, our study investigated the association between early EEG changes and neurological impairment in patients with ischemic stroke after aortic dissection. Our findings indicate that qEEG is a sensitive tool for monitoring brain function, and its continuous monitoring can aid clinicians in the early detection and treatment of brain complications, leading to improved long-term prognosis.

AUTHOR CONTRIBUTIONS
Ya-peng Wang, Wen-xue Liu, Yi Jiang, Shan Lu, Yang Chen, and Dongjin Wang contributed to study concept and design; Ya-peng Wangand Wen-xue Liu involved in acquisition of data; Yang Chen and Dongjin Wang contributed to drafting the manuscript; all authors read and approved the final manuscript. All authors have confirmed that manuscript complies with all instructions to us. All of authors confirmed that this manuscript has not been published elsewhere and is not under consideration by another journal.

CONFLICT OF INTEREST STATEMENT
All authors declare that they have no conflict of interests.

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