Herein, we report a case of high-amplitude slow delta waves, being detected during processed EEG monitoring, which was observed before the appearance of pupil dilatation in a patient with cerebral hemorrhagic infarction. The dose of sedative medication was not altered while we observed these EEG changes, and we consider it to have detected the onset of stroke.
Indications for continuous EEG monitoring include the detection of non-convulsive epileptic seizures, treatment evaluation of epileptic seizures, detection of cerebral ischemia, evaluation of sedation and therapeutic coma, and prognostic evaluation of brain damage [4]. EEG is useful for the diagnosis of acute ischemic stroke in the emergency ward [5]. Intraoperative arousal, postoperative delirium, and postoperative cognitive impairment may be reduced when the depth of general anesthesia is managed using processed EEG monitoring [6]. The SedLine® used in our case performed EEG processing using two electrodes on each side of the forehead. It showed a PSI value that indicated the depth of anesthesia, the PSI value from 25 to 50 indicates an appropriate depth of anesthesia. It can also monitor real-time raw EEGs.
EEG monitoring can detect an extensive ischemic stroke and estimate the size of the infarct lesion within hours of onset of a cerebral infarction [5]. EEG electrodes are generally placed according to a mapping system that relates the anatomy of the head surface to the underlying cortical regions of the brain. The standard EEG mapping is referred to as the 10–20 international system. This system measures the nasal root and occipital tuberosity, as well as the anterior point of the bilateral auricles, and obtains the vertex from each midpoint. The area between the nasal root and occipital tuberosity, and between the bilateral preauricular points is divided into 10% or 20%, and a total of 21 electrodes are placed systematically [7]. Standard critical care continuous EEG requires a minimum of 16 electrodes placed according to the 10–20 international system. If fewer than 16 electrodes are used, the interpretation of the critical care continuous EEG may be limited due to the inadequate spatial sampling and inability to distinguish artifacts from cerebral activity [4]. A few reports have shown that 2–7 electrodes are not sensitive enough to detect non-convulsive epileptic seizures [4, 8]. In our case, even with only four electrodes placed on the forehead, we could detect that the EEG changed to a delta wave, which is thought to appear at the onset of cerebral infarction. In cerebral ischemia, high-frequency activities, such as beta waves, are usually diminished; meanwhile, delta waves are increased on the affected side [9, 10]. Furthermore, the wave amplitudes can be increased or decreased [9]. In the present case, delta waves observed on POD 7, at 7 p.m., were initially presumed that there was a damage in the left frontal lobe or a large part of the left cerebral hemisphere because the amplitude of the waveform on the left forehead decreased. Processed EEG monitoring may allow for earlier detection of intracranial abnormalities, although it is not a sufficient substitute for subjective sedation scales such as the Richmond Agitation-Sedation Scale [11]. It was considered important for the processed EEG monitoring to possess the ability to display and record the raw EEG waveforms in addition to seeing the PSI value of sedation.
On POD 4, it was difficult to differentiate whether the patient's disturbance of consciousness was because of the confusion caused by his general condition or a small stroke. We considered the risk of moving the patient to the imaging suite to be relatively high owing to his general condition, hence, we did not perform imaging tests. At that time, EEG using the 10–20 international system should have been performed to detect abnormalities in a wide range of brain regions.
It would be difficult to detect occipital and temporal lesions, as Sedline® is a forehead monitor. Unfortunately, the fact that it only represents the lesions affecting the frontal lobe may be a limitation in detecting abnormal EEG signals. There are no reports on the sensitivity of EEG of the forehead in detecting stroke in our knowledge. However, EEG of the forehead may be useful during sedation of patients at risk of stroke, because at least when the EEG changes, immediate imaging tests can be performed to determine the indication for surgery, catheterization, or drug treatment.