Prior classifications of status epilepticus have been suggested on the basis of etiology, clinical phenomenology, pathophysiology, and EEG characteristics [15]. More recently, the Commission of Classification and Terminology of the International League Against Epilepsy charged a Task Force with revising the classification of status epilepticus. This group in turn created a classification based on four axes: (i) semiology, (ii) etiology, (iii) EEG correlates, and (iv) age [16]. Various forms are differentiated along taxonomic lines into motor activity, and impairment of consciousness, producing two major groups: [1] status epilepticus with prominent motor symptoms, including all convulsive forms, and [2] those without prominent motor forms representing non-convulsive forms of status epilepticus (NCSE). Each group can be subdivided by the degree of consciousness impairment, while etiology divides into: (i) known or symptomatic (also subdivided into acute symptomatic, remote symptomatic and progressive symptomatic), and (ii) unknown or cryptogenic. Known causes of status appear as an appendix. The third axis comprises EEG correlates—more important in NCSE. Descriptors adopted (such as morphology, location, time-related features) were adopted from those developed by the American Clinical Neurophysiology Society [17]. Nonconvulsive status epilepticus (NCSE) had previously been divided into two groups, largely along electroencephalographic criteria:

Each of these groups has been subdivided further. GNSE includes:

The behavioral correlates of different types of NCSE arise from—or are at least associated with—different areas of maximal involvement of seizure activity identifiable by EEG. Hence the proclivity of seizure activity for frontal, mesial temporal, neocortical temporal, and temporo-parieto-occipital junctional regions has come to be correlated with particular symptomatic or behavioral features particular to those brain regions. Additionally, such presentations are seen to occur with certain underlying associated conditions such as mental retardation, uremia, or Alzheimer disease, and therefore achieve clinical expression within the framework of the age of the patient. There are, therefore, particular clinical features typically seen in infants, small children, adults, or the elderly, although there is considerable overlap. A classification of types of NCSE based on localization-related EEG criteria as well as age of expression and particular epileptic syndromic context is provided in Table 20.2. Many of these features are superseded by the latest classification of the Task Force, but are provided, as they are familiar to many clinicians of the past two decades. 

A major challenge in definition has been the correct EEG identification of seizures. Because the determination of what represents seizures, and thus SE, on EEG depends on somewhat subjective interpretation, the art of diagnosis depends on accurate EEG interpretation. A universal EEG definition of seizure activity has been hard to pin down, as evidenced by the problems that even sophisticated, computerized seizure and spike detectors have had in correctly identifying seizures and distinguishing them from artifact. Some typical themes, however, can be noted. Seizures captured in their entirety will typically show a progression from low-voltage high-frequency spikes to high-voltage lower-frequency spike slow-wave activity before stopping abruptly and being replaced by background suppression. This progression from fast to slow components can be used to identify an isolated seizure. Problems arise, however, when the patient is in SE and activity thought to be epileptic, precedes the beginning of the tracing and continues beyond the end of the tracing. In such cases, a rhythmicity, often with variability, typically faster than one-per-second is usually seen. Rhythmic activity may contain sharply contoured or spiky components, typical spike slow-wave or polyspike-slow-wave complexes, or even rhythmic theta or delta frequencies. The major differential diagnostic confounders are epileptiform morphologies usually seen at less than one-per-second, such as periodic lateralized epileptiform discharges (PLEDs [or LPDs]), bilateral independent periodic lateralized epileptiform discharges (BiPLEDs [or BIPDs]), or even bilateral synchronous epileptiform discharges (GPEDs [or GPDs]), all of which may be seen in cortical hyperexcitable states that may follow seizures in a patient with a structural abnormality, or in patients with encephalitis (For updated terms for the types of periodic discharges (LPD, GPD, BIPD), see Hirsch and colleagues [17]). These states representing cortical “irritability,” even with a clinical correlate of diminished level of consciousness, have historically not been classified as active seizures. The electrical activity has been judged to be insufficiently fast, and to lack a more flagrant clinical correlate (such as clonic activity) to be epileptic or “ictal.” Nonetheless, this borderline is a “gray zone” because epileptic activity may not particularly originate in the motor cortex so as to produce clonic activity; conversely it may represent the end of the electroclinical continuum that follows convulsive status epilepticus [18].

Toxic, metabolic, and infectious encephalopathies, benzodiazepine withdrawal states, and neuroleptic malignant and serotonin syndromes may all be associated with altered behavior and levels of consciousness, accompanied by an abnormal EEG, often with epileptiform features such as triphasic waves (GPDs with triphasic morphology). In this way, these states resemble, and can be confused with, NCSE, even to the point of suppression of “ictal” triphasic-wave activity after IV benzodiazepines.

Differentiation of types of NCSE along clinical lines—for example, differentiating cases of CPSE from GNSE—can be problematic because of the marked overlap among the clinical characteristics of the different types of NCSE [15, 16, 18–77]. Such blurring of the lines can be seen in the many publications that provide clinical correlates to clearly identified focal or generalized nonconvulsive status. For example, historically, total unresponsiveness was said not to occur in the absence status, but such patients have been noted. Impaired consciousness may be common to many types of NCSE, as may be fluctuation in the level of consciousness, bradyphrenia and bradykinesia, confusion, or even simple automatisms.

Nonetheless, some generalizations can be made regarding CPSE and ASE. Fear, aggressivity, irritability, and anxiety are seen more frequently with CPSE than with ASE [42]. Similarly, stereotyped, complex automatisms are also more frequent in CPSE [42]. Lip-smacking, other oroalimentary automatisms, lateralized limb automatisms, and dystonic posturing, eye deviation, and nystagmus are typical of CPSE [38, 42–44]. In both CPSE and ASE, patients may be agitated, violent, and aggressive, and may experience hallucinations. The following sections delineate the behavioral features, emphasizing distinguishing characteristics.

Taking a step back, it may not be as important to characterize NCSE into ASE or CPSE, as it is to recognize NCSE at all. At Johns Hopkins Bayview Medical Center, where some 300+ patients with NCSE have been identified over the past 30 years, the diagnosis was frequently delayed or even missed [12, 40]. Table 20.3 describes clinical examples with such scenarios. To state the obvious, the suspicion that NCSE is present must enter the mind to trigger a request for an EEG, enabling diagnosis. Although cases of NCSE may present initially on any floor of the hospital, there are particular presentations favoring the emergency room, intensive care units, and on neurology and psychiatry services. NCSE may resemble other disorders. Examples of some of these are given in Table 20.4 (see also Chap. 2, “Types of Status Epilepticus: Definitions and Classification”).