A 53-year-old man (160 cm, 60 kg) diagnosed with hepatic cholangiocarcinoma was scheduled for laparoscopic lobectomy under general anesthesia. His medical history was normal. During operation, the accumulation of water in the CO2 absorption device resulted in a sudden increase in PETCO2 and lasted approximately 12 min, with the maximum value of PETCO2 reaching 70 mmHg.

A 44-year-old woman (165 cm, 62 kg) diagnosed with hepatic hemangioma was scheduled for laparoscopic resection of the complex hepatic hemangioma under general anesthesia. Her medical background was normal. The PETCO2 suddenly increased to 64 mmHg and lasted approximately 2 min when the hepatic portal vein was unclamped.

A 52-year-old man (165 cm, 57 kg) who suffered from hepatocellular carcinoma was scheduled for laparoscopic resection of right complex liver cancer under general anesthesia. He had a history of hepatitis B virus infection, which led to cirrhosis (Child-Pugh A) and portal hypertension. Abnormal laboratory findings included an aspartate aminotransferase level of 109 IU/L, a glutamate aminotransferase level of 105 IU/L, and a platelet count of 82*109/L. Other laboratory examinations were normal. During the hepatectomy surgical process, after the portal vein had been clamped for 15 min, the PETCO2 started to increase from 45 to 56 mmHg within 1 min and fluctuated in this range for approximately 10 min.

A 52-year-old man (165 cm, 62 kg) diagnosed with hepatocellular carcinoma was scheduled for laparoscopic central hepatectomy under general anesthesia. He was previously diagnosed with hepatitis B virus infection and received routine antiviral treatment. Abnormal laboratory findings included an aspartate aminotransferase level of 117 IU/L, a glutamate aminotransferase level of 161 IU/L, and a platelet count of 87*109/L. Other physical and laboratory examinations were normal. There was a sudden increase in PETCO2 during the opening of the hepatic portal vein. The maximum value of PETCO2 was 58 mmHg, and the duration of increased PETCO2 was approximately 2 min.

The anesthesia protocols used were similar for all four patients. The perioperative monitoring included SEDLine frontal EEG (Masimo Inc., Irvine, CA, USA), invasive arterial blood pressure, electrocardiogram (ECG), pulse oxygen saturation (SpO2), PETCO2, body temperature, and train of four ratios (TOF) for neuromuscular blockage monitoring. Anesthesia was induced by intravenous injection of a 2 mg bolus of midazolam and propofol target control infusion (TCI) at a plasma target concentration of 3 µg/ml; 0.4 µg/kg sufentanil and 0.2 mg/kg cisatracurium were intravenously injected when the Modified Observer’s Assessment of Alertness/Sedation Scale (MOAA/S) score was 1. Tracheal intubation was performed when TOFcnt ≤ 2. Anesthesia was maintained with desflurane and remifentanil, and the dosage was adjusted to maintain the PSI in the range of 25–50. The perioperative target blood pressure was maintained not beyond ± 20% of the baseline. Cisatracurium (0.05 mg/kg) was added when the TOFcnt was ≥ 2. All patients received a lung-protective ventilation strategy, and the target PETCO2 was between 35 and 45 mmHg. When the PETCO2 increased suddenly, the respiratory parameters were immediately adjusted to increase the minute volume of ventilation.

After the PETCO2 abruptly increased, density spectral array (DSA) monitoring revealed that the bilateral spectral edge frequency (SEF) decreased in all four patients (Fig. 1A), and burst suppression (BS) even emerged in patient 1 (black vertical bars with blue bottom in Fig. 1A). BS analysis of patient 1 showed that the BS ratio was 16.25%, duration of suppression was 117s, and peak-to-peak voltage of the bursts was 27.72(2.42) µV. The maximum reduction of SEF were from 8.8 to 2.7 Hz (69.3%), 18.5 to 8.4 Hz (54.6%), 12.6 to 5.2 Hz (58.6%) and 11.5 to 8.5 Hz (26.2%), respectively in four patients (Fig. 1B). During this period, the change in the SEF was opposite to that in the PETCO2; that is, the SEF decreased as the PETCO2 increased and returned to its original state as the PETCO2 returned to normal. The duration of the SEF decrease was roughly consistent with the duration of the PETCO2 increase, with a 1-minute delay (Fig. 1B). Interestingly, the patient sedation index (PSI) increased in 3 out of 4 patients (patients 1, 3, and 4), which is very different from over sedation (Fig. 1B). The PSI of patient 2 remained unchanged as the PETCO2 increased. The phase space plot of CO2 vs. SEF is shown in Fig. 2. The overall clockwise trend when CO2 was beyond 40mmHg indicated that the elevated CO2 took precedence over the decrease of SEF in patients 2, 3, and 4. However, it was difficult to determine the direction of patient 1 probably because SEF remained at a low level and even EEG burst suppression occurred when CO2 was markedly elevated.

Changes in the DSA, PSI, SEF, and EEG power during the PETCO2 increase. (A) Changes in the DSA during the increase in PETCO2. (B) Changes in PSI and SEF during the increase in PETCO2. (C) Changes in EEG power and MAP during the increase in PETCO2. (D) Changes in power ratio of each EEG frequency range during the increase in PETCO2. (E) Changes in absolute power of EEG frequency range during the increase in PETCO2. T1: The time of initial increase in PETCO2; T2: The time of PETCO2 return to a normal level; E1: The time of lowest EEG total power. Abbreviations: DSA = density spectral array; PSI = patient sedation index; SEF = spectral edge frequency; MAP = mean arterial pressure

The phase space plot of CO2 vs. SEF. T1: The time of initial increase in PETCO2; T2: The time of PETCO2 return to a normal level; E1: The time of lowest EEG total power

After analyzing the EEG total power of the four patients, we found that the EEG total power decreased with the rise of PETCO2 level, and the total power recovered when the PETCO2 returned to a normal level (Fig. 1C). We also extracted the raw EEG of this period and compared them with those obtained within 10 min before the event, which showed that the EEG frequency slowed down while the amplitude decreased during this period (Fig. 3).

The comparison of depressed EEG (Fig. 3b) and EEG before 10 min of depression (Fig. 3a). EEG (Fig. 3b) frequency slowed down while the amplitude decreased. a: EEG 10 min before E1; b: EEG of E1. E1: The time of lowest EEG total power

The absolute power in each frequency range was reduced except in patient 4. Perhaps, the absolute power of patient 4 did not change due to the short duration of EEG depression (Fig. 1E). By further analyzing the power ratio and absolute power of each EEG frequency range during this period, the power ratio of δ band (0.1-<4 Hz) increased, while α band (8–13 Hz) and β band (14–30 Hz) decreased with the rise of PETCO2. The ratio of the θ band (4-<8 Hz) increased in patients 2, 3, and 4, but decreased in patient 1. The ratio of the γ band (> 30 Hz) did not change significantly in any of the four patients [6] (Fig. 1D).

The vital signs were stable before the PETCO2 changed, and there was no bolus injection of any medications before the occurrence of EEG depression. The mean arterial pressure (MAP) remained above 50 mmHg during the PETCO2 increase (Fig. 1C). The end-tidal desflurane concentration and remifentanil dosage are summarized in Table 1. Four patients were treated actively when PETCO2 elevation was found. Therefore, it returned to the normal level in a short time, so the duration of electroencephalographic depression lasted for a very short time. No postoperative delirium (assessed by the Confusion Assessment Method) occurred during the follow-up, and the length of hospital stay was similar to that of the same type of surgical patients.