The recent advancement in computer technology has allowed continuous EEG recording at the patient’s bedside. This has become possible because of increased memory size, fast processing speed, and also improved video screen resolution. Continuous and long-term monitoring of EEG allows the evaluation of dynamic changes of brain function that may not be visible by clinical examination alone. With accumulation of continuous EEG data in the intensive care unit (ICU), it has become apparent that nonconvulsive seizures (NCS) or nonconvulsive status epilepticus (NCSE) are relatively common in acutely ill, comatose or stuporous patients. Because of the lack of motor manifestation, NCS can easily be overlooked without continuous EEG recording. In this context, we adopt the term CCEEG (continuous critical care EEG) to distinguish it from LTM EEG (long-term monitoring of EEG) used primarily for evaluation of epilepsy patients. Because of the intermittent and unpredictable occurrence of NCS, a single 20- to 30-minute EEG recording would likely miss capturing these seizures. The value of CCEEG has been further strengthened by concomitant video recording, networking capability, ability of storing large amounts of data for many hours (>24 hours), and also the automatic computer analyses of seizure/spike detection and various quantitative EEG analyses by power spectrum measurement.

The most common reason for CCEEG is to find out if the patient is having NCSE or intermittent episodes of NCS in the acutely ill comatose or obtunded patients. The following clinical conditions raise the question of NCSE or NSC.

Other justifiable reasons for performing CCEEG include episodes of seizure-like symptoms such as muscle twitching, body posturing, eye deviation, chewing, pupillary abnormalities, or autonomic symptoms, but without overt convulsion.¹,²,³,⁴ The CCEEG may also be indicated for evaluation of vasospasms or cerebral ischemia, especially after subarachnoid hemorrhage (SAH) or interventional neuroradiological procedures.⁵,⁶ CCEEG may also be used for the evaluation of progressively changing brain function, either worsening or improving, for the purpose of prognostication.⁷,⁸ CCEEG is also indicated for guidance in maintaining a therapeutic burst suppression pattern with barbiturates or other anesthetic/sedative drugs⁹,¹⁰ and for assessment of the treatment efficacy for NCS and NCSE. CCEEG is now a routine protocol for detection of NCS or NCSE in patients undergoing therapeutic hypothermia after anoxic brain insult secondary to cardiac arrest.¹¹,¹²

The decision to switch to CCEEG may also be based on the finding of certain abnormal EEG patterns on the routine or initial recording. They include generalized (GPD^*) (Video 13-1, see Figs. 10-33D and 10-35) or lateralized periodic discharges (PLEDs/LPD^*) (see Figs. 10-39 and 10-40) (Video 13-2A and B), SIRPID (Fig. 13-1A-C, see also Fig. 10-38A and B) (Video 13-3A and B), lateralized rhythmic delta activity (LRDA) (Fig. 13-2A and B) (Video 13-4), or brief (potentially) ictal rhythmic discharges (B(I)RD) (Fig. 13-3A and B) (Video 13-5), because these patterns are more likely to develop NCS or NCSE¹³,¹⁴,¹⁵ (see section of “Evaluation of CCEEG for NSC and NCSE” in this chapter). As shown in Video 13-2A and B, PLEDs/LPD between two regions may show synchronous (Video 13-2A) or asynchronous discharges (Video 13-2B, see also Fig. 10-41) within the same hemisphere. SIRPIDs may consist of variable waveforms, ranging from blunt sharp-wave triphasic form (Video 13-3A, see Fig. 13-1B) to frank spikewave discharges (Video 13-3B, see Fig. 10-38B). This often makes it difficult to decide if the patient should be treated as having seizure.

NCS is defined as an electrographic seizure without overt clinical signs of seizures, though some patients may have subtle face or limb twitching, eye deviation, eyelid twitching, chewing, muscle tone increase, nystagmus, pupillary changes, or other autonomic signs (blood pressure change, respiration change, tachycardia or bradycardia, sweating, etc.). These clinical changes are diagnosed as seizures only by observing repeatedly the same subtle clinical sign correlating with consistent EEG changes (Video 13-6A and B). NCSE was traditionally defined when NCSs occur continuously or near continuously lasting more than 30 minutes.¹⁶,¹⁷ But recent guidelines proposed in 2016 by the American Epilepsy Society define status epilepticus as when a seizure lasts longer than 5 minutes.¹⁸ This applies for both convulsive seizures and NCSs.¹⁸ Yet another consensus proposed by American Clinical Neurophysiology Society for use with neonates is “summated duration of a seizure comprised of greater than or equal to 50% of an arbitrarily defined one hour epoch.”¹⁹

The overall incidence of NCS or NCSE in patients who are in coma or unexplained altered mental status in ICU settings (inclusive of pediatric patients) varies from about 10% to 50%.²⁰,²¹,²²,²³ Of various etiologies, the postconvulsive status epilepticus patients have the highest incidence of NCS or NCSE; one study showed close to half the patients who had convulsive status epilepticus developed NCS (14% of these were NCSE) after the convulsions had stopped clinically.²⁴

Another common cause for NCS is postanoxic cerebral insult following cardiac or respiratory arrest, ranging from 10% to 60% (see Fig. 10-36), although myoclonic convulsive seizures are not uncommon (Video 13-7).²⁵,²⁶,²⁷ NCS or NCSE may also occur during therapeutic hypothermia after anoxic cerebral insult secondary to cardiac arrest and could occur during the rewarming period.¹⁰,¹¹,²⁶ The NCS or NCSE associated with ischemic brain injury carries a grave prognosis.²⁷

NCS and NCSE are also common in toxic metabolic encephalopathy (uremia, hypertensive encephalopathy, drug intoxication or withdrawal, hepatic failure, hypo- and hyperglycemia)²⁸ and patients with postoperative brain surgery,²⁹,³⁰,³¹ traumatic brain injury,³²,³³,³⁴ intracerebral hemorrhage,³⁴,³⁵,³⁶ ischemic stroke,³⁵,³⁶,³⁷,³⁸ and SAH.³⁹,⁴⁰ The incidence of NCS in these conditions varies, ranging from 10% to 30%. The incidences of NCS and NCSE are generally higher in neonates, infants, and children in NICU (neonatal ICU) or PICU (pediatric ICU)⁴¹,⁴²,⁴³,⁴⁴ than in adults. One study showed 44% of patients in PICU or NICU who underwent CCEEG monitoring had NCS and 75% of these were NCSE.²¹ It is expected that the probability of capturing seizures (either clinical or nonconvulsive) would be higher with CCEEG than with a routine 30-minute recording in the ICU setting. EEGs of 30 to 60 minutes will likely miss more than half of the seizures captured in CCEEG²²,⁴³,⁴⁵ With CCEEG, about 80% to 95% of NCS was captured within 24 to 48 hours.²³,⁴⁵,⁴⁶,⁴⁷ It is therefore reasonable to stop monitoring if there are no seizures for 2 days after the last seizure.
However, more than 2 days of monitoring may be required for a patient who has frequent interictal epileptiform activity and periodic discharges or those withdrawing from pharmacological sedation. On the other hand, the chance of capturing a seizure in CCEEG (72 hours) without any epileptiform activity during the first 2 hours is less than 5%.⁴⁸ Although the study cohort was biased because CCEEG was ordered primarily for a question of NCS, one study disclosed 83% (49 out of 59 patients) had NSC and almost half (47% or 23 of 49) of these had NSCE.¹⁶ Similarly, NSC was found in 18% to 35%,²³,²⁴ and up to 75% of these patients had NCSE.⁴⁸ Also in postcardiac arrest patients undergoing hypothermia, similar results have been reported even with a 30-minute recording.⁵⁰

The electrodes used for CCEEG are the same as used in routine EEG (silver-silver chloride or gold). Because the EEG is recorded without the attendance of a technologist most of the time, it is important that the electrodes are attached securely enough for many hours of recording. In some cases, extra or redundant electrodes may be placed. This is especially true for the “system” or “common” reference; if the system reference is removed, the entire recording will be lost. In addition to EEG electrodes, EKG electrodes should always be placed because EKG contamination, especially abnormal EKG, can mimic epileptiform activity (Chapter 15, see Fig. 15-15). EKG can also sometimes provide important clues for cardiac-related spells such as vasovagal syncope (see Fig. 10-46A-D) or seizure causing cardiac abnormality (see Video 10-14). Other important electrodes are eye monitors to record the electrooculogram. This is necessary for the differentiation of frontal delta activity from eye movement artifacts (see Chapter 15, Eye Movement Artifacts, see also Figs. 8-16A and B, 15-3, 15-4, and 15-5). All electrodes are best secured with collodion. The use of paste is not recommended for long-term recording. Electrodes for recording EMG (electromyograph) may be applied over appropriate muscles, especially when the patient shows muscle twitches or involuntary movements.

Because ICU patients often undergo MRI or CT scan during the monitoring session, the electrodes must be removed before and replaced after the examination is done. This is a nuisance and is time consuming for technologists. MRI- or CT scancompatible electrodes have now become commercially available. Subdermal needle electrodes (disposable stainless steel) can be used for rapid application and may be preferentially used for patients with skin problems or for patients who have fresh surgical wounds on the scalp. However, needle electrodes attenuate low-frequency signals. Also, they are not suitable for long-term recording. Other recently introduced electrodes are subdermal wire electrodes, which are made of Teflon-coated wire with a silver-chloride tip. These are suitable for long-term recording⁵¹ and can be made compatible for CT or MRI scan. An electrode cap may be used for quick application when an EEG technologist is not available but is not suitable for longterm recording, so it should be replaced when the EEG technologist becomes available.