A previously healthy 48-year-old male patient (height: 182 cm, weight: 98 kg) was scheduled to undergo resection of a mediastinal goiter. Preoperative laboratory investigations were within normal limits. A chest radiograph revealed no lung infiltrates or cardiomegaly. An electrocardiogram (ECG) showed normal sinus rhythm at 60 bpm. Computed tomography of the neck and chest revealed a tumor measuring 191 × 84 × 68 mm, extending from the right thyroid lobe to the anterior region of the tracheal bifurcation. The tumor caused leftward deviation and compression of the trachea in the cervical region and posterior compression in the thoracic region. The patient presented with dyspnea in the supine position, although arterial oxygen saturation was maintained at 99% on room air. No central nervous system (CNS) abnormalities were noted preoperatively.

The patient received no premedication. In the operating room, standard biometric monitoring was performed. Additionally, invasive arterial blood pressure monitoring, processed EEG, and acceleromyography at the adductor pollicis muscle were implemented. Processed EEG monitoring was conducted using a BIS Quatro sensor (Medtronic, Minneapolis, MN) placed on the left forehead, with electrode impedance maintained at 5 kΩ or less throughout the procedure. EEG waveform data were recorded in a database at a sampling rate of 250 Hz and analyzed offline using MATLAB R2024a (MathWorks, Natick, MA). The power spectrum and spectrogram were generated using the Chronux mtspecgramc function with the following parameters: a window length of 3 s, a 0.5-s overlap, a time-bandwidth product of 3, and 5 tapers.

The anesthetic record and EEG spectrogram throughout the course of events are shown in Fig. 1A. Due to the presence of a potentially difficult airway, awake nasotracheal intubation was performed using a local anesthetic spray (total of 10 mL of 4% lidocaine) and a continuous infusion of remifentanil at 0.075 μg/kg/min. Following successful intubation, general anesthesia and neuromuscular blockade were induced with 10 mg of remimazolam and 60 mg of rocuronium, respectively. The initial frontal EEG waveform after induction of general anesthesia exhibited delta and alpha oscillations (Fig. 1B). General anesthesia was maintained with remimazolam (30–100 mg/h), remifentanil (0.075–0.25 μg/kg/min), and intermittent boluses of fentanyl (total dose: 700 μg). Neuromuscular blockade was maintained with intermittent boluses of rocuronium, achieving a train-of-four count of less than 1 throughout the surgery.

Anesthetic record with a spectrogram of the frontal electroencephalogram (EEG), and time-domain frontal EEG signatures. A Anesthetic record and spectrogram throughout the course of events, with time 0 marking the induction of general anesthesia. EEG waveform (B) after induction of general anesthesia with remimazolam, showing delta and alpha oscillations; C 30 min after topical airway anesthesia with a total of 10 mL of 4% lidocaine, displaying the emergence of continuous theta waves; D immediately after local anesthesia of the chest wall with 20 mL of 1% lidocaine containing 1:100,000 epinephrine, illustrating enhanced continuous theta waves; E after the transition of theta waves to a frequently repeating pattern, persisting for approximately 1 s every 5 s; F following the emergence of sharp 2–5 phase periodic discharges, exhibiting a similar duration and cycle to the theta waves; G after administering a total of 150 mL of 20% lipid emulsion, showing the return of delta and alpha oscillations

Thirty minutes after the administration of topical airway anesthesia, continuous theta waves appeared on the frontal EEG (Fig. 1C). These theta waves were further enhanced immediately after the administration of 20 mL of 1% lidocaine with 1:100,000 epinephrine for chest wall local anesthesia (Fig. 1D). Subsequently, the theta waves transitioned into a frequently repeating pattern, persisting for approximately 1 s every 5 s (Fig. 1E). Eventually, sharp 2–5 phase periodic discharges emerged, exhibiting a similar duration and cycle to the theta waves (Fig. 1F). Following the administration of 20% lipid emulsion in divided doses of 50 mL, totaling 150 mL, the EEG normalized to delta and alpha oscillations (Fig. 1G).

During the period of EEG alterations, the bispectral index ranged from 28 to 63. The predicted effect-site concentrations of remimazolam [4], remifentanil [5], and fentanyl [6] were 660–880, 2.5–8.5, and 0–2.3 ng/mL, respectively. Although the QRS duration and corrected QT interval on the ECG were slightly prolonged, with fluctuations observed in the corrected QT interval (Fig. 1A), both normalized following lipid emulsion therapy.

The surgery was completed uneventfully in 368 min. The patient was alert and calm 8 min after the discontinuation of remimazolam infusion. He followed commands 3 min after the administration of 0.2 mg flumazenil. Following uneventful extubation, the patient remained stable and exhibited no prodromal CNS symptoms of LAST, such as dizziness or tinnitus. His postoperative course was uneventful except for right ulnar nerve palsy, likely caused by intraoperative compression of the right upper limb. He was discharged on postoperative day 10.