State entropy and burst suppression ratio can show contradictory information: A retrospective study

Electroencephalogram monitoring during anesthesia and critical care: a guide for the clinician

Perioperative anesthetic, surgical and critical careinterventions can affect brain physiology and overall brain health. The clinical utility of electroencephalogram (EEG) monitoring in anesthesia and intensive care settings is multifaceted, offering critical insights into the level of consciousness and depth of anesthesia, facilitating the titration of anesthetic doses, and enabling the detection of ischemic events and epileptic activity. Additionally, EEG monitoring can aid in predicting perioperative neurocognitive disorders, assessing the impact of systemic insults on cerebral function, and informing neuroprognostication. This review provides a comprehensive overview of the fundamental principles of electroencephalography, including the foundations of processed and quantitative electroencephalography. It further explores the characteristic EEG signatures associated wtih anesthetic drugs, the interpretation of the EEG data during anesthesia, and the broader clinical benefits and applications of EEG monitoring in both anesthetic practice and intensive care environments.

Impact of Age on the Occurrence of Processed Electroencephalographic Burst Suppression

BACKGROUND

Patient age is assumed to be an important risk factor for the occurrence of burst suppression, yet this has still to be confirmed by large datasets.

METHODS

In this single-center retrospective analysis at a university hospital, the electronic patient records of 38,628 patients (≥18 years) receiving general anesthesia between January 2016 and December 2018 were analyzed. Risk factors for burst suppression were evaluated using univariate and multivariable analysis. We measured the incidence of burst suppression as indicated by the burst suppression ratio (BSR) of the Entropy Module, the maximum and mean BSR values, relative burst suppression duration, mean volatile anesthetic concentrations, and mean age-adjusted minimum alveolar concentrations (aaMAC) at burst suppression, and cases of potentially misclassified burst suppression episodes. Analyses were done separately for the total anesthesia period, as well as for the Induction and Maintenance phase. The association with age was evaluated using linear and polynomial fits and by calculating correlation coefficients.

RESULTS

Of the 54,266 patients analyzed, 38,628 were included, and 19,079 patients exhibited episodes with BSR >0. Patients with BSR >0 were significantly older, and age had the highest predictive power for BSR >0 (area under the receiving operating characteristic [AUROC] = 0.646 [0.638-0.654]) compared to other patient or procedural factors. The probability of BSR >0 increased linearly with patient age (ρ = 0.96-0.99) between 1.9% and 9.8% per year. While maximal and mean BSR showed a nonlinear relationship with age, relative burst suppression duration also increased linearly during maintenance (ρ = 0.83). Further, episodes potentially indicating burst suppression that were not detected by the Entropy BSR algorithm also became more frequent with age. Volatile anesthetic concentrations sufficient to induce BSR >0 were negatively correlated with age (sevoflurane: ρ = -0.71), but remained close to an aaMAC of 1.0.

CONCLUSIONS

The probability of burst suppression during general anesthesia increases linearly with age in adult patients, while lower anesthetic concentrations induce burst suppression with increasing patient age. Simultaneously, algorithm-based burst suppression detection appears to perform worse in older patients. These findings highlight the necessity to further enhance EEG application and surveillance strategies in anesthesia.

Continuity with caveats in anesthesia: state and response entropy of the EEG

The growing use of neuromonitoring in general anesthesia provides detailed insights into the effects of anesthetics on the brain. Our study focuses on the processed EEG indices State Entropy (SE), Response Entropy (RE), and Burst Suppression Ratio (BSR) of the GE EntropyTM Module, which serve as surrogate measures for estimating the level of anesthesia. While retrospectively analyzing SE and RE index values from patient records, we encountered a technical anomaly with a conspicuous distribution of index values. In this single-center, retrospective study, we analyzed processed intraoperative electroencephalographic (EEG) data from 15,608 patients who underwent general anesthesia. We employed various data visualization techniques, including histograms and heat maps, and fitted custom non-Gaussian curves. Individual patients' anesthetic periods were evaluated in detail. To compare distributions, we utilized the Kolmogorov-Smirnov test and Kullback-Leibler divergence. The analysis also included the influence of the BSR on the distribution of SE and RE values. We identified distinct pillar indices for both SE and RE, i.e., index values with a higher probability of occurrence than others. These pillar index values were not age-dependent and followed a non-equidistant distribution pattern. This phenomenon occurs independently of the BSR distribution. SE and RE index values do not adhere to a continuous distribution, instead displaying prominent pillar indices with a consistent pattern of occurrence across all age groups. The specific features of the underlying algorithm responsible for this pattern remain elusive.

[Processed EEG for personalized dosing of anesthetics during general anesthesia]

Electroencephalogram (EEG)-guided anesthesia is indispensable in modern operating rooms and has become established as the standard form of monitoring. Many anesthesiologists rely on processed EEG indices in the hope of averting anesthesia-related complications, such as intraoperative awareness, postoperative delirium and other cognitive complications in their patients. This educational review aims to provide information on the five most prevalent monitors used to guide depth of sedation during general anesthesia. This article elucidates the principles underpinning the application of these monitors where known, which are generally based on power in various EEG frequency bands and on the burst suppression pattern. Convinced that EEG-guided anesthesia has the potential of benefitting many surgical patients, it is felt that many basic principles and shortcomings of processed EEG indices need to be better understood in the clinical practice. After discussing the different monitors and clinically relevant data from the literature, the article gives a short practical guidance on how to critically interpret processed EEG information and troubleshooting of confounded indices in the context of clinical situations.

Unwanted spontaneous responsiveness and burst suppression in patients undergoing entropy-guided total intravenous anesthesia with target-controlled infusion: An observational prospective trial

STUDY OBJECTIVE

To estimate the incidence of unwanted spontaneous responsiveness and burst suppression (BSupp) in patients undergoing state entropy (SE) and surgical pleth index (SPI)-guided total intravenous anesthesia (TIVA) with target-controlled infusion (TCI).

DESIGN

Observational, prospective, single-center study.

SETTINGS

Operating room.

PATIENTS

107 adult (<65 years) and elderly (≥65 years) women undergoing breast surgery.

INTERVENTIONS

Propofol-remifentanil TIVA-TCI-guided by SE for depth of anesthesia monitoring (target value 40-60) and SPI for antinociception monitoring (target value 20-50) without neuromuscular blockade.

MEASUREMENTS

Age; body mass index; American Society of Anesthesiologists physical status classification; concentration at the effect site of propofol (CeP) and remifentanil (CeR) at loss of responsiveness (LoR), median during anesthesia maintenance (MdM), and at return of responsiveness (RoR); propofol infusion duration; incidence of postoperative delirium (POD) with Confusing Assessment Method for the Intensive Care Unit.

MAIN RESULTS

During SE-SPI-guided TIVA-TCI, 13.1% of patients showed unwanted spontaneous responsiveness, whereas 45.8% showed BSupp. Unwanted spontaneous responsiveness was observed mainly in adults (p < 0.05), and higher CeP RoR (p < 0.05) was registered. BSupp was observed mainly in patients showing a lower CeP MdM (p < 0.01) and CeP RoR (p < 0.05). Unwanted spontaneous responsiveness and BSupp were not associated with significant differences in CeRs. An age-related hysteresis effect was observed, resulting in higher CeP LoR than CeP RoR (p < 0.001). 12.2% of patients showed POD. Only preoperative serum albumin was associated with increased likelihood of POD (p = 0.046).

CONCLUSIONS

The SE-SPI-guided TIVA-TCI did not prevent unwanted spontaneous responsiveness and BSupp. CeP RoR may be used as a proxy for anesthetic sensitivity.

Always Assess the Raw Electroencephalogram: Why Automated Burst Suppression Detection May Not Detect All Episodes

BACKGROUND

Electroencephalogram (EEG)-based monitors of anesthesia are used to assess patients' level of sedation and hypnosis as well as to detect burst suppression during surgery. One of these monitors, the Entropy module, uses an algorithm to calculate the burst suppression ratio (BSR) that reflects the percentage of suppressed EEG. Automated burst suppression detection monitors may not reliably detect this EEG pattern. Hence, we evaluated the detection accuracy of BSR and investigated the EEG features leading to errors in the identification of burst suppression.

METHODS

With our study, we were able to compare the performance of the BSR to the visual burst suppression detection in the raw EEG and obtain insights on the architecture of the unrecognized burst suppression phases.

RESULTS

We showed that the BSR did not detect burst suppression in 13 of 90 (14%) patients. Furthermore, the time comparison between the visually identified burst suppression duration and elevated BSR values strongly depended on the BSR value being used as a cutoff. A possible factor for unrecognized burst suppression by the BSR may be a significantly higher suppression amplitude ( P = .002). Six of the 13 patients with undetected burst suppression by BSR showed intraoperative state entropy values >80, indicating a risk of awareness while being in burst suppression.

CONCLUSIONS

Our results complement previous results regarding the underestimation of burst suppression by other automated detection modules and highlight the importance of not relying solely on the processed index, but to assess the native EEG during anesthesia.

Targeted Interventions to Increase Blood Pressure and Decrease Anaesthetic Concentrations Reduce Intraoperative Burst Suppression: A Randomised, Interventional Clinical Trial

Background

It has been suggested that intraoperative electroencephalographic (EEG) burst suppression (BSupp) may be associated with post-operative neurocognitive disorders in the elderly, and EEG-guided anaesthesia may help to reduce BSupp. Despite of this suggestion, a standard treatment does not exist, as we have yet to fully understand the phenomenon and its underlying pathomechanism. This study was designed to address two underlying phenomena—cerebral hypoperfusion and individual anaesthetic overdose.

Objectives

We aimed to demonstrate that targeted anaesthetic interventions—treating intraoperative hypotension and/or reducing the anaesthetic concentration—reduce BSupp.

Methods

We randomly assigned patients to receive EEG-based interventions during anaesthesia or EEG-blinded standard anaesthesia. If BSupp was detected, defined as burst suppression ratio (BSR) > 0, the primary intervention aimed to adjust the mean arterial blood pressure to patient baseline (MAP intervention) followed by reduction of anaesthetic concentration (MAC intervention).

Results

EEG-based intervention significantly reduced total cumulative BSR, BSR duration, and maximum BSR. MAP intervention caused a significant MAP increase at the end of a BSR > 0 episode compared to the control group. Coincidentally, the maximum BSR decreased significantly; in 55% of all MAP interventions, the BSR decreased to 0% without any further action. In the remaining events, additional MAC intervention was required.

Conclusion

Our results show that targeted interventions (MAC/MAP) reduce total cumulative amount, duration, and maximum BSR > 0 in the elderly undergoing general anaesthesia. Haemodynamic intervention already interrupted or reduced BSupp, strengthening the current reflections that hypotension-induced cerebral hypoperfusion may be seen as potential pathomechanism of intraoperative BSupp.

Clinical Trial Registration

NCT03775356 [ClinicalTrials.gov], DRKS00015839 [German Clinical Trials Register (Deutsches Register klinischer Studien, DRKS)].

Burst Suppression During General Anesthesia and Postoperative Outcomes: Mini Review

In the last decade, burst suppression has been increasingly studied by many to examine whether it is a mechanism leading to postoperative cognitive impairment. Despite a lack of consensus across trials, the current state of research suggests that electroencephalogram (EEG) burst suppression, duration and EEG emergence trajectory may predict postoperative delirium (POD). A mini literature review regarding evidence about burst suppression impact and susceptibilities was conducted, resulting in conflicting studies. Primarily, studies have used different algorithm values to replace visual burst suppression examination, although many studies have since emerged showing that algorithms underestimate burst suppression duration. As these methods may not be interchangeable with visual analysis of raw data, it is a potential factor for the current heterogeneity between data. Even though additional research trials incorporating the use of raw EEG data are necessary, the data currently show that monitoring with commercial intraoperative EEG machines that use EEG indices to estimate burst suppression may help physicians identify burst suppression and guide anesthetic titration during surgery. These modifications in anesthetics could lead to preventing unfavorable outcomes. Furthermore, some studies suggest that brain age, baseline impairment, and certain medications are risk factors for burst suppression and postoperative delirium. These patient characteristics, in conjunction with intraoperative EEG monitoring, could be used for individualized patient care. Future studies on the feasibility of raw EEG monitoring, new technologies for anesthetic monitoring and titration, and patient-associated risk factors are crucial to our continued understanding of burst suppression and postoperative delirium.