EEG recording can be carried out in the ICU without harm to the patient, has low costs, and can be performed repeatedly, or monitored continuously. Continuous EEG monitoring offers no significant advantages over intermittently repeated standard recordings [29].

AMSE is characterized by a stereotyped sequence of generalized discharges (generalized periodic discharges, GPDs) on a profoundly abnormal background (Figs. 13.1, 13.2, 13.3, 13.4c, 13.6a, 13.7, 13.8). Generalized bursts include spikes, spikes and waves, sharp waves, slow activity or combinations of these. Critical care EEG terminology has been standardized by Hirsch et al. [30], and post hoc surveys showed substantial inter-rater agreement in the use of these terms [31–33]. GPDs are time-locked to jerks, but in many cases, GPDs occur without visible motor signs. Jerks may be suppressed in the ICU by treatment with paralytic agents, sedatives, and anesthetic agents. Without superficial electromyogram recording, inconspicuous fine jerks may be missed. These coma cases have frequently been termed nonconvulsive status epilepticus (NCSE), but GPDs are also seen in advanced nonepileptic coma states, including those due to sedative-antiseizure drugs [1] or with hypothermia for surgical intervention [34]. Therefore, Bauer and Trinka [35] distinguished coma with GPDs after hypoxia from NCSE “proper” in the categorization of the epilepsies.

Such typical EEG patterns may be confined to selected segments of the EEG monitoring, and fluctuations and transitions are the rule in a given record or with repeated recordings [11, 36]. Generalized and focal seizure patterns may interrupt the periodicity (for a 2006 review, see Kaplan [37]) (see Fig. 13.3). Frequently, seizure patterns are clinically nonconvulsive or subtle and therefore termed electrographic or subclinical [38, 39]. Periodic lateralized epileptiform discharges (PLEDs) or lateralized periodic discharges (LPDs) after Hirsch and colleagues [30] can fluctuate with GPDs and suggest a localized lesion such as an infarction (see Fig. 13.6).

Suppression–burst activity represents a distinct type of GPDs. Periodic high voltage generalized bursts alternate with nearly isoelectric activity in a quasi-periodic fashion [40]. Suppression-burst activity is often an EEG correlate of AMSE; of coma without motor seizures; of periodic tonic eye opening (see above); or of oral, ocular, or appendicular subtle movements [24]. Modifiers include the length of intervals, the type of background activity, and burst morphology (see Figs. 13.1, 13.3, and 13.6a). Like other forms of GPDs, suppression-burst activity probably shares a common mechanism across diverse etiologies that cause oxygen and glucose deprivation [41].

Triphasic waves or continuous 2/s GPDs with triphasic morphology [30] were first described in hepatic coma but afterwards in a variety of encephalopathies, including post cardiac arrest syndrome (for summary, see Kaplan and Bauer, [1]). Typical and atypical forms have been described, but the diagnostic significance of any type remains to be established [42]. Triphasic-appearing waves have been recorded in myoclonic status epilepticus, in degenerative and spongiform encephalopathies, and in atypical absence status in the Lennox–Gastaut syndrome [1]. The clinical condition associated with triphasic waves after hypoxia is often considered NCSE; abnormal motor activity (see Figs. 13.2 and 13.7) has not been analyzed specifically.

Stimulus–induced rhythmic, periodic, or ictal discharges (SIRPIDs) were originally reported in critically ill patients [43]. Alvarez and colleagues analyzed the role of SIRPIDs in post cardiac arrest syndrome [44]; 14 of 105 patients had SIRPIDs, 5 of them with early myoclonus. The pattern lies somewhere along an ictal-interictal continuum [45].

The importance of EEG reactivity to exogenous stimulation has been stressed repeatedly [17, 46–48]. Hirsch and colleagues refined types of reactivity using their critical care EEG terminology [30], and a systematic description was proposed. Rossetti and colleagues assessed the reactivity (regardless of the appearance of epileptiform transients), and differentiated nonreactive records from reproducible changes in background EEG [17]. In a retrospective study of hypoxic-ischemic injuries, Howard and colleagues differentiated responsive from non- or poorly responsive EEG rhythms, and from low-voltage background activities [46]. Bauer and colleagues collected several types of EEG reactivity in the post cardiac arrest syndrome in a retrospective observational study [36]. Some studies showed that arousal can lead to a breakdown of electrical activity that may be difficult to distinguish from spontaneous EEG voltage attenuations. In summary, reactions to exogenous stimuli are manifold; their classification and correlation to clinical variables have not been studied systematically.

In AMSE, GPDs may alternate with rhythmic alpha–or theta frequencies (see Figs. 13.6b and 13.8). Furthermore, transitions from GPDs to continuous alpha/theta rhythms have been observed [49] (see Fig. 13.5). A long-standing debate involves the prognostic significance and differential diagnosis of different alpha frequencies in coma or in the locked-in syndrome [49]. Reactivity of alpha frequency activity predicted good outcome in an etiologically mixed sample [50]. Berkhoff and colleagues reinvestigated postanoxic alpha (theta) coma and distinguished complete and incomplete forms [51]. Complete forms correspond to the original description of alpha coma and signify a poor prognosis [52].

In the post CA syndrome, the bilateral absence of the N20 component of the somatosensory evoked potentials with median nerve stimulation, recorded on days 1–3 or later after CA, accurately predicts poor outcome, as do serum neuron-specific enolase levels of >33 μg/L at days 1–3 [9]. Neuroimaging methods in this setting have been summarized by Little and colleagues [53] and have their pros and cons [54]. Despite some reports claiming prognostic superiority over conventional tools [55, 56], the value and reliability of imaging methods remain inconclusive. Furthermore, imaging typically involves moving the patient, which can be difficult or may be harmful to a patient in the ICU.

All methods mentioned above have been used in the study of the post CA syndrome in general. No specific data can be found regarding clinically and electroencephalographically proven AMSE.

Treatment and prognostic statements in this section refer to postanoxic states in general. Coma with periodic epileptiform EEG abnormalities has been considered a type of NCSE, and these same EEG discharges are the hallmark of AMSE. The differential diagnosis between NCSE and AMSE depends on the observation of motor abnormalities––frequently abolished by neuromuscular blocking agents used in the ICU management of patients. Furthermore, in dosages used for treatment of these patients, sedatives and ASDs may produce periodic EEG changes themselves. Thus, diagnostic and prognostic statements regarding postanoxic coma depend on conceptual and observational uncertainties, and can be affected by the effects of sedating medication.

A long list of papers stresses the dismal prognosis of coma with AMSE and GPDs [16, 57–60]. Besides ICU treatment, ASDs and hypothermia have been tried. The Hypothermia after Cardiac Arrest Study Group confirmed the therapeutic effect of hypothermia in post cardiac arrest syndrome in general [61]. In AMSE, the beneficial role is less well established. AMSE can occur before and during hypothermia and with rewarming. Wijdicks and colleagues stated that patients with myoclonic status epilepticus within the first day after a primary circulatory arrest have a poor prognosis [9]. This is also true for hypothermia-treated patients with status epilepticus [62].

Several papers, however, have reported good outcome in early posthypoxic myoclonus in which the etiology for these coma states was respiratory arrest [63–66]. After the subject’s regaining consciousness, the diagnosis of Lance–Adams syndrome was established. A good neurologic outcome was reported in three cases of primary cardiac arrest syndrome and myoclonic status treated with hypothermia [67]. It was concluded that premature, pessimistic prognostic statements should be reconsidered.