According to the National Audit Project number 5 (NAP5), the incidence of self-reported awareness during general anesthesia was ~1:20,000 procedures.1 There was a significant variation in the incidence of awareness depending on the anesthetic techniques and surgical specialties. As such, recommendations were made regarding the use of processed electroencephalogram (pEEG) monitors, especially in patients at high risk of awareness. Nevertheless, NAP5 recommendations alerted to the fact that the use of monitors such as bispectral index (BIS) did not reduce the incidence of this complication and that they should be interpreted with caution when reaching unexpectedly low anesthetic concentrations.
pEEG monitors with spectrogram provide a visual representation of the frontotemporal brain activity displaying patterns that represent the effect of different agents.2 The concept of multimodal general anesthesia is based on the use of multiple drugs targeting different nociceptive pathways. This strategy is used to reduce the dose of each agent, thus minimizing their side effects.3 Simultaneously, some of those drugs act centrally at different molecular targets and neural circuits altering the raw electroencephalogram (EEG), which can lead to the unreliability of pEEG indices as a guide for the depth of hypnosis.
We report a case of awareness during the maintenance phase of multimodal general anesthesia in a patient monitored with bilateral BIS. Written consent was obtained before the publication of the case report.
DESCRIPTION OF THE CASE
A 65-year-old male patient (90 kg, 1.77 m) with hypertension and colon cancer presented for laparoscopic colectomy. The patient was monitored according to the American Society of Anesthesiologists standards. Neuromuscular blockade (quantitative assessment), nociception (analgesia nociception index [ANI] monitor), and depth of hypnotic effect (bilateral BIS sensor) were also monitored. A thoracic epidural catheter (T9–T10) was placed before the induction of general anesthesia. The multimodal general anesthesia strategy included propofol and remifentanil administered by target-controlled infusion (TCI), lidocaine (2 mg·kg−1 bolus, 2 mg·kg−1·hour−1 infusion), ketamine (0.25 mg·kg−1 bolus, 0.25 mg·kg−1·hour−1 infusion), magnesium sulfate (2 g bolus), dexmedetomidine (progressively titrated from 0.2 to 0.7 μg·kg−1·hour−1), and epidural 0.2% ropivacaine (15 mg bolus, 10 mg·hour−1 infusion). Deep muscle relaxation was obtained with rocuronium boluses, with a total of 150 mg administered during the procedure. The remifentanil TCI (Minto model) was started before the epidural catheter placement with an effect site concentration (Ce) of 2.0 ng·mL−1 and stopped after induction of general anesthesia. The propofol TCI (Schnider model) was guided by BIS monitoring aiming for index values of 40–60. The Figure shows the spectrogram and the BIS trend of the dominant cerebral hemisphere.
The surgery lasted for 4 hours 20 minutes. Nearly 3 hours 30 minutes after induction, BIS values were between 30 and 40 and a downward trend was observed. Dexmedetomidine infusion was maintained at 0.7 μg·kg−1·hour−1 and propofol Ce was progressively reduced from 1.5 to 0.5 μg·mL−1. BIS returned to values between 40 and 50 and propofol Ce was maintained at 0.5 μg·mL−1. However, the spectrogram showed a progressive dissipation of its alpha oscillation pattern for approximately 30 minutes (Figure, 04:30–05:00). Suddenly, the heart rate rose from ~60 to ~90 beats per minute and the systolic blood pressure from ~100 to ~190 mm Hg accompanied by a rise in BIS values from 35 to 74 and spectral edge frequency (SEF) from 12 to 25 Hz. The BIS monitor also showed suppression ratio (SR) values rising from 0 to 10, and the raw EEG showed isoelectricity with sporadic low-amplitude and high-frequency waves. The spectrogram revealed a loss in power in every wavelength (Figure, 05:00–05:10). BIS electromyography was <30 dB. The train-of-four count was 0 and the patient remained immobile. ANI values were consistently >50 and remained between 53 and 65 during and after the event. The TCI pump was altered to TCI view mode and a propofol bolus (30 mg) was administered. Afterward, propofol Ce was adjusted to 2.0 μg·mL−1. The heart rate, blood pressure, and BIS parameters returned to the baseline values and alpha spindles reappeared. During anesthetic emergence, the propofol infusion was stopped and the same alpha spindle fading pattern was observed (Figure, 05:20–05:50). Nearly 30 minutes later, BIS values progressively increased and SR went from 0 to 8 before recovery of consciousness. Again, the same bluish pattern accompanied by an increase in SEF values was observed.
At the end of the procedure, the patient was transferred to the postanesthesia care unit. No spontaneous recall of the event was reported, although the patient mentioned a nightmare during the procedure when actively questioned. A modified Brice interview conducted on postoperative day 2 and 1 month after the surgery were compatible with an episode of awareness.
This case illustrates the difficulties in titration of hypnotics and antinociceptive agents in a multimodal general anesthesia strategy based exclusively on pEEG indices.
During general anesthesia, unconsciousness is maintained primarily by using hypnotic agents. The antinociceptive agents contribute to unconsciousness by attenuating nociceptive-induced arousal.3 Due to their action on different pathways, the hypnotic dose required to maintain unconsciousness is reduced.
In this case, numerous agents that may alter the patient’s EEG were used. Propofol acts at postsynaptic gamma-aminobutyric acid-ergic (GABAergic) receptors present in the cortex, thalamus, brainstem, and spinal cord. Its inhibitory effects on the thalamus and cortex are depicted on both raw EEG and spectrogram by delta waves (0.1–4 Hz) and alpha oscillations (9–12 Hz).2 At low doses, ketamine acts mainly on N-methyl-d-aspartate (NMDA) receptors on GABAergic inhibitory interneurons. This activity is marked by gamma oscillations (25–30 Hz).3 Magnesium enhances the effects of hypnotic agents by acting on arousal pathways with antinociceptive and hypnotic targets similarly to ketamine.3 Dexmedetomidine exerts an antinociceptive effect due to decreased noradrenergic excitatory inputs to the preoptic area of the hypothalamus, the thalamus, basal forebrain, and the cortex.3 Under dexmedetomidine, the EEG shows alpha spindles and slow-delta waves; higher doses may only reveal slow-delta oscillations.3 The effects of lidocaine on brain activity and raw EEG are not well established. Notwithstanding, alternative mechanisms through which lidocaine may enhance antinociception include blockade of sodium channels, NMDA receptors, and/or glycine receptors in the brainstem arousal circuits and in the amygdala, thus diminishing nociceptive transmission and inducing sedation.3
In our opinion, it seems clear that relying on BIS “standard” clinical end point values (ie, 40–60) to guide the hypnotic component of multimodal general anesthesia can be misleading, because the EEG patterns may differ from anesthesia with propofol or inhalational agents combined with opioids. In our case, before the episode of wakefulness, BIS values were between 30 and 40 which prompted a reduction in propofol Ce. Although an adequate increase of BIS values to 40–50 was observed, the alpha oscillation pattern started to fade for almost 30 minutes before the intraoperative event that we assumed to be the episode of awareness. In fact, according to the current literature, the loss of alpha oscillations is associated with the activation of arousal pathways.4 In this regard, the present case illustrates the importance of moving away from 1-dimensional indices and highlights the value of the spectrogram to assess the hypnotics’ effects and infer the level of unconsciousness. Retrospectively, there was an overreduction of propofol Ce which, even with coadministration of ketamine and dexmedetomidine infusions, was clearly insufficient to maintain the patients’ state of unconsciousness. Indeed, the NAP5 report recommends that the information conveyed by pEEG monitors should be interpreted with caution in the face of unexpectedly low anesthetic concentrations. Thus, the correct interpretation of all the available information, including the patient’s spectrogram, might have avoided the reported event.
Furthermore, this case highlights other important limitations of pEEG monitors, because conflicting information was observed during the wakefulness event. SR represents the percentage of time when the raw EEG reveals isoelectricity (defined by the BIS monitors as EEG amplitudes <5 μV) during the previous 63 seconds.5 Accordingly, it is expected that an increase in SR would be accompanied by a decrease in BIS. In our case, a simultaneous increase in BIS values and SR was observed. As previously described by Brown and Purdon,6 EEG baseline amplitudes are lower in elderly patients. We believe that these conflicting data were an anecdotal event that might have happened due to the patient’s high-frequency waves having very low EEG amplitudes and power (as shown in Figure A). The raw EEG revealed an atypical pattern of low-voltage, high-frequency waves alternated with periods of apparent isoelectricity, which differs from the typical burst suppression pattern characterized by high-voltage, high-frequency electrical activity alternated with periods of isoelectricity.7 Our hypothesis is that the monitor assumed higher frequency waves (ie, beta and gamma waves [lower amplitude, higher frequency]) as isoelectricity because both intraoperative and postoperative findings are compatible with an episode of awareness (implying undertitration of hypnotic and/or antinociceptive agents).
In addition, this case also reflects that awareness episodes are occasionally ignored by the patients, making these events underdiagnosed. When an episode of awareness is suspected, active search and use of validated questionnaires are essential to identify awareness events and refer patients for psychological therapies if necessary.
Name: Francisco Salgado-Seixas, MD.
Contribution: This author helped write the manuscript, research the references, and care for the patient during the perioperative period.
Name: Rui Pereira, MD.
Contribution: This author helped write the manuscript and research the references.
Name: Humberto Machado, PhD.
Contribution: This author helped edit the manuscript.
Name: Carla Cavaleiro, MD.
Contribution: This author helped edit the manuscript and care for the patient during the perioperative period.
This manuscript was handled by: BobbieJean Sweitzer, MD, FACP.
1. Pandit JJ, Andrade J, Bogod DG; Royal College of Anaesthetists; Association of Anaesthetists of Great Britain and Ireland. 5th
National Audit Project (NAP5) on accidental awareness during general anaesthesia: summary of main findings and risk factors. Br J Anaesth. 2014;113:549–559.
2. Purdon PL, Sampson A, Pavone KJ, Brown EN. Clinical electroencephalography for anesthesiologists. Anesthesiology. 2015;123:937–960.
3. Brown EN, Pavone KJ, Naranjo M. Multimodal general anesthesia: theory and practice. Anesth Analg. 2018;127:1246–1258.
4. Scheinin A, Kallionpää RE, Li D. Differentiating drug-related and state-related effects of dexmedetomidine and propofol on the electroencephalogram. Anesthesiology. 2018;129:22–36.
5. Aspect Medical Systems. BIS VISTA TM Monitoring System Operating Manual; 2005.
6. Brown EN, Purdon PL. The aging brain and anesthesia. Curr Opin Anaesthesiol. 2013;26:414–419.
7. Ching S, Purdon PL, Vijayan S, Kopell NJ, Brown EN. A neurophysiological-metabolic model for burst suppression. Proc Natl Acad Sci U S A. 2012;109:3095–3100.