Stroke is one of the leading causes of death and disability in Thailand, ¹ and carotid stenosis is one of the leading causes of stroke. ² The surgical treatment to prevent stroke is carotid endarterectomy (CEA). It is associated with periprocedural risks, including stroke (embolic or hemodynamic), myocardial infarction, and death. Therefore, strict selection criteria are applied for patients undergoing CEA. Current selection criteria support CEA for symptomatic low-risk surgical patients with 50% to 99% stenosis and asymptomatic patients with stenosis of 70% to 99%. ³ However, the ability of the patient to tolerate the cross-clamp depends on the sufficiency of collateral flow through the circle of Willis. Inadequate collateral cerebral perfusion during the cross-clamp period increases the risk of perioperative stroke. ⁴ Despite routine intraluminal shunt during the temporary occlusion of the ipsilateral internal carotid artery being controversial, intraoperative electrophysiological monitoring, such as electroencephalogram (EEG), is a valuable tool to detect cerebral hypoperfusion and determine selective shunting. ^{5 , 6} When the neurosurgeon performed the non-shunt technique, adequate cerebral perfusion during carotid cross-clamping could be achieved using several methods to protect the brain. ⁷ Spetzler et al. ⁸ reported excellent non-shunt surgical outcomes using intra-operative barbiturate and microsurgical techniques. The clinical use of barbiturates is known for cerebral protection against the prevention of focal cerebral ischemia, ⁹ especially when barbiturate was administered before the ischemic insult with doses large enough to produce burst-suppression activity on the EEG. ¹⁰

Despite the lack of clarity, the rationale for inducing burst suppression is based on its theorized potential neuroprotective effects. Burst suppression reduces metabolic demand by decreasing intracellular adenosine triphosphate (ATP) concentration, leading to decreased neuronal activity and electrical signaling, ¹¹ reduction in cerebral blood flow (CBF), ¹² and preserving limited energy resources during critical situations. Although, experimental animal studies conducted by Warner DS ¹³ and Robert Schmid-Elsaesser, ¹⁴ indicate that EEG burst suppression may not be necessary for maximum cerebral protection. Additionally, Westermaier T ¹⁵ found no additional neuroprotection following mild hypothermic treatment of rats subjected to reversible focal ischemia by barbiturate-induced burst suppression. Despite this, animal studies may not be conclusive in humans due to inter-species differences in burst suppression effects ¹⁶ and differences in physiology and the human clinical context.

Human research is crucial for understanding burst suppression benefits in clinical settings. Doyle PW ¹⁷ suggests that if the flow-metabolism coupling is intact, complete EEG burst suppression (100% burst suppression) may provide more cerebral protection than 50% burst suppression. However, the study did not evaluate the cerebral protection effects. Thus, further human studies are still needed to fully understand the relationship between burst suppression and cerebral protection, also the definite predetermined amount of barbiturate-induced burst-suppression activity on EEG including the optimal dosage, timing, and administration mode varies among studies to reach the burst-suppression pattern.

The Neurological Institute of Thailand is one of the few medical centers with EEG for intraoperative surveillance. Thus, to fill the knowledge gap, the authors aimed to study the optimal dose of barbiturates as thiopental for inducing EEG burst-suppression patterns in anesthetized patients undergoing carotid endarterectomy with a non-shunt technique.

There were 69 carotid endarterectomies performed during the study period, with 12 cases excluded. Of the remaining 57 patients analyzed ( Figure 1), only 24 achieved burst suppression on intraoperative EEG despite receiving continuous thiopental infusion with additional titration. These 24 patients were classified as the burst suppression group (BS) for the analysis. The demographic data and related details of both the BS and non-BS group are presented in Table 1. The group had a significantly higher average age of 72.8±9.1 years than the non-BS group, with an average age of 66.7±7.2 years (p-value=0.007). However, there were no significant differences in gender, body weight, ASA physical status, comorbidities, or pre-operative investigation data between the two groups. Hypertension was a common condition in both groups. The percentage of patients who received thiopental or propofol as induction agents and the dosages were not significantly different between the two groups ( Table 2). Perioperative doses of fentanyl and end-tidal concentrations of sevoflurane or desflurane also showed no significant differences. The amount of thiopental required to achieve burst suppression on intraoperative EEG was significantly higher in the BS group compared to the non-BS group (26.3±10.1 mg/kg/hr vs. 18.7±8.8, p-value=0.004). Although the carotid clamp time was slightly shorter in the BS group, it did not reach statistical significance (73.2±23.7 min vs. 83.3±34.8, p-value=0.225).

The spectral edge frequency 95% (SEF95%) of both the BS and non-BS groups tended to decrease after carotid clamping, as indicated in Table 3. However, the two groups had no significant difference regarding MED or SEF95% before and after the clamping. Similarly, neither group significantly differed between the left and right MED or SEF95%. After carotid clamping, the BS group had an average BSR of 36.0±20.4 (right) and 36.3±20.6 (left), but this difference was not statistically significant. In contrast, the non-BS group did not exhibit any burst suppression pattern.

The incidence of hypertension, hypotension, and arrhythmias did not show a statistically significant difference between the two groups. Following the operation, eight patients (33.3%) in the BS group and sixteen (48.5%) in the non-BS group were awake and extubated. Most patients in both groups were intubated and transferred to the neurosurgical intensive care unit. There was no significant difference in extubation time for patients who were initially unable to extubate between the two groups (BS group 872.4±593.3 min. vs. non-BS group 601.6±473.9 min). One patient experienced a mild neurological deficit. No deaths were reported one month after the operation ( Table 4).

This study aimed to determine the amount of thiopental required to induce burst suppression patterns on intraoperative EEG monitoring in patients undergoing carotid endarterectomy without a shunt and to evaluate patient outcomes. The main findings indicated that not all patients achieved burst suppression despite the intention to maximize cerebral protection through a continuous thiopental infusion and titrated intravenous administration. Patients who received significantly higher doses of thiopental (26.3±10.1 mg/kg/hr) were likelier to achieve burst suppression on EEG. However, no significant difference was observed in postoperative outcomes between the burst suppression (BS) and non-burst suppression (non-BS) groups. Currently, limited data is available on the efficacy and optimal dosage of thiopental for inducing pharmacological burst suppression to prevent perioperative stroke during selective shunting in CEA.

Barbiturates, such as thiopental, are commonly used to prevent cerebral ischemia during cerebrovascular surgery. Thiopental is a fast-acting, short-duration barbiturate anesthetic that may exert its neuroprotective effects through various mechanisms, including antioxidant activity, GABA-ergic activity, stimulation of protein synthesis, removal of free radicals, and modulation of excitatory synaptic neurotransmission via adenosine. ¹⁹ Animal studies have shown that barbiturates can decrease brain oxygen demand and the size of cerebral infarction. ^{20 – 22} In cerebrovascular procedures that require temporary clips, such as extracranial-intracranial bypasses, carotid endarterectomies, and aneurysm clipping, barbiturates have been demonstrated to reduce the cerebral metabolic rate of oxygen and increase blood flow to ischemic regions. ^{23 – 25}

The thiopental dose required for EEG burst suppression patterns during cerebrovascular surgery can vary depending on several factors, including variability in monitoring and assessing burst suppression levels among healthcare providers. Sreedhar and Gadhinglajkar ²⁶ have reviewed several dosing regimens of thiopental for cerebrovascular surgery, including a bolus dose (4 mg/kg), a low dose followed by IV infusion (1 to 3 mg/kg IV followed by 0.06 to 0.2 mg/kg/min), and a high dose followed by infusion (loading 25 to 50 mg/kg followed by 2 to 10 mg/kg/hr).

The initial bolus doses of thiopental used in our study did not result in EEG burst suppression for most patients, which differs from the findings of Ramesh VJ. ²⁷ According to Ramesh VJ, almost all patients who received a bolus dose of 3 to 5 mg/kg experienced EEG burst suppression with a BSR greater than 25%. The initial bolus doses of thiopental only resulted in temporary suppression durations, consistent with previous studies by Moffat et al. ²⁸ and Gelb et al., ²⁹ which provided limited cerebral protection during the intraoperative period.

Our study used a continuous thiopental infusion to maintain EEG burst suppression during the carotid cross-clamp procedure. We administered a high dose of thiopental, similar to previous studies by McConkey PP et al. ³⁰ and Frawley JE et al. ³¹ However, not all patients in our study achieved EEG burst suppression, unlike the abovementioned studies where incremental bolus doses of thiopental were titrated. Nonetheless, none of our patients experienced a significant period of ischemia, as defined by a decrease in SEF95% to 50% of the baseline. ³²

EEG-confirmed burst suppression is widely recognized as the preferred indicator of cerebral protection during cerebrovascular surgery with barbiturate therapy.

Our institute protocol primarily relied on electroencephalography (EEG) to assess burst suppression. While intraoperative monitoring of burst suppression is commonly performed using EEG or Bispectral Index (BIS), ^{33 , 34} it has been observed that BIS-derived Burst Suppression Ratio (BSR) may underestimate the duration of EEG suppression, thereby reducing sensitivity in detecting burst suppression. ³⁵ We directly visualized the EEG trace to ensure a more accurate and real-time evaluation of raw EEG changes and BSR values across both brain hemispheres. This approach enabled us to detect potential cerebral ischemia and determine cerebral protection levels.

However, we recognize the significance of monitoring cerebral hemodynamics and function, ³⁶ including sensory-evoked potential (SEP), motor-evoked potential (MEP), and near-infrared spectroscopy (NIRS). ³⁷ In 2014, the institute began implementing sensory-evoked potential monitoring (SEP) and motor-evoked potential (MEP) for spinal surgery. However, these methods are not routinely employed for carotid endarterectomy, partly due to the impact of burst suppression on EEG functional assessment. ^{38 , 39} Additionally, cerebral oximetry was first introduced in 2022.

The study found that older patients were more likely to achieve burst suppression on EEG with thiopental therapy, potentially due to age-related factors. ⁴⁰ Although not all patients achieved burst suppression, the BS and non-BS groups had favorable clinical outcomes without significant perioperative complications. The BS group did have a more extended postoperative intubation period. However, their Glasgow Outcome Scale was not significantly different from the non-BS group at discharge or one month after surgery. These outcomes could be attributed to barbiturate therapy, as research on rats suggests that EEG burst suppression may not be necessary for maximum cerebral protection.

The dosage of thiopental that does not lead to EEG burst suppression may impact cerebral protection during carotid endarterectomy. However, thiopental-induced cerebral protection is only one of several strategies for cerebral protection. Other factors, such as the degree of stenosis, pre-existing medical conditions, and perioperative risks associated with carotid endarterectomy, influence patient outcomes. ⁴¹

Limitations of this study include the fact that it only focused on the dosage of thiopental and its effects on intraoperative EEG monitoring as a direct tool for evaluating brain electrical activity during carotid endarterectomy. It is important to note that EEG only provides information on brain electrical activity and does not directly measure cerebral blood flow or cerebral oxygen saturation (SjVO ₂). Therefore, relying solely on EEG may not accurately reflect the actual level of cerebral protection. Multimodal intraoperative monitoring techniques, such as transcranial Doppler (TCD) ultrasound, somatosensory evoked potentials, and cerebral oxygen saturation monitoring, are recommended for a more comprehensive patient status assessment. Additionally, this study did not assess the thiopental serum levels for barbiturate coma treatment. The retrospective cohort design also limited the quality of data collection. Therefore, a prospective randomized study is recommended to investigate further the role of intraoperative monitoring and barbiturate therapy in optimizing cerebral protection during carotid endarterectomy.

In conclusion, thiopental at a dosage of 26.3+10.1 mg/kg/hr can induce burst suppression on intraoperative EEG during carotid endarterectomy, potentially providing cerebral protection. However, a cautious approach is recommended when using thiopental to induce EEG burst suppression during carotid clamping. While barbiturate-induced cerebral protection is still considered to have a role, the necessity of EEG burst suppression for cerebral protection requires further investigation. Considering the limitations of our retrospective study, the recommended strategy should be individualized for each patient, considering available resources based on institutional protocols. Further evidence from medical studies, neurophysiology studies, and mathematical modeling is needed to gain a comprehensive understanding of burst suppression and its potential therapeutic applications. Therefore, it is crucial to carefully evaluate the potential benefits and risks of using burst suppression in clinical practice and identify the optimal perioperative settings where it may be beneficial.