Clinical manifestations

Presentation and course

In a systematic review of nonconvulsive status epilepticus symptoms, the most common manifestation was impairment of consciousness, which occurred in 82% of patients (72). This altered consciousness could manifest as confusion (49%), coma (22%), lethargy (21%), and memory loss (18%).

Absence status epilepticus, sometimes considered a type of nonconvulsive status epilepticus, can be characterized by confusion, psychomotor retardation, and abnormal behavior (impulsivity, agitation, or aggression) (21; 72). These symptoms may start or end with a generalized convulsion (21). Absence status epilepticus is predominantly seen in ambulatory patients and the presentation and management is unique from other forms of nonconvulsive status epilepticus. Absence status is, thus, discussed in detail in a separate dedicated report.

Focal status epilepticus with impaired awareness is often reported as an “epileptic twilight state” with impaired responsiveness or confusion and bizarre behavior. Patients may also have automatisms (oral or manual) (21). Focal status epilepticus with intact awareness symptoms reflects the function associated with the ictal focus; common manifestations may include fear, mood changes, autonomic fluctuations, aversive eye movements, prolonged paralysis, or disturbed spatial perception simulating transient ischemic attacks (21; 72). Additional focal neurologic deficits have been reported, including aphasia, prosoposagnosia, catatonia, and catalepsy (03; 30; 59; 60).

Hospitalized patients with nonconvulsive status epilepticus may manifest other subtle clinical findings apart from altered mental status. Ocular movement abnormalities (sustained eye deviation, nystagmus, eye jerks), subtle myoclonic and intermittent focal or multifocal movements, and hippus increase the likelihood of nonconvulsive status epilepticus (74; 76; 38). None of these symptoms are specific for nonconvulsive status epilepticus, and an EEG is needed to confirm the diagnosis. Risk factors for nonconvulsive seizure activity in the critically ill include coma, history of epilepsy, convulsive seizure prior to the EEG, and remote brain injury (38; 14).

EEG patterns associated with nonconvulsive seizure and nonconvulsive status epilepticus can be subtle, and in the last 2 decades several attempts have been made to establish diagnostic criteria. Based on previous criteria, an expert panel at the 4th London Innsbruck Colloquium on Acute Seizures in Salzburg, Austria proposed the working criteria for nonconvulsive status epilepticus in 2013 (04); the criteria were later refined in 2015 (Table 1) (44). The modified criteria also used the standardized terminology established by the American Clinical Neurophysiology Society (33). However, in the setting of status epilepticus, these patterns may fall in the ictal-interictal continuum category, making a diagnosis of nonconvulsive status epilepticus even more difficult (12). Because of this uncertainty, a trial of benzodiazepines or rapid acting intravenous nonbenzodiazepine antiseizure medication during EEG recording can be helpful. The test is considered diagnostic for nonconvulsive status epilepticus only if there is resolution of the periodic EEG pattern and improvement in mental status; improvement in EEG without a clinical change is considered an indeterminant response (40; 37). Clinical improvement might take up to 24 hours in the critically ill patient with nonconvulsive status epilepticus (23).

Neonatal nonconvulsive status epilepticus is a common finding among critically ill neonates. Due to its unique clinical and electrophysiologic characteristics, as well as specialized management considerations, neonatal status epilepticus will be discussed in detail in a separate dedicated article.

Table 1. Modified Salzburg Consensus Criteria for Nonconvulsive Status Epilepticus

Prognosis and complications

Nonconvulsive status epilepticus is a broad term that includes different groups of patients, and incidence and outcome estimates vary significantly in the literature. Ambulatory nonconvulsive status epilepticus typically represents an exacerbation of a seizure subtype or is the initial manifestation of a subacute neurologic insult and is typically associated with a favorable outcome. The prognosis is best for patients with preexisting genetic generalized epilepsy or de novo absence status epilepticus (48).

Hospitalized patients with nonconvulsive status epilepticus and coma have a much poorer prognosis. Symptomatic nonconvulsive seizure and nonconvulsive status epilepticus (after acute or remote brain injury) are associated with a longer hospital stay and high mortality rates (approximately 70%) (25). Some of this poor outcome may be due to the underlying cause of nonconvulsive status epilepticus and from the complications occurring in the clinical course. However, this association persists after controlling for age, etiology, neurologic exam, and organ dysfunction (17; 25). There are data to support the use of the ACD scoring measure to predict long-term mortality. This scoring system is named based on its inclusion of three key factors: age at onset, level of consciousness on admission, and duration of status epilepticus (64).

In comatose patients with subtle status epilepticus following convulsive status epilepticus, the resolution of the EEG discharges is associated with better outcome (75). Following the cessation of convulsive status epilepticus, normal EEG, burst suppression, after-status epilepticus ictal discharges (not including periodic lateralized epileptiform discharges), and periodic lateralized epileptiform discharges were associated with mortalities of 0%, 59%, 41%, and 40%, respectively (39). Additionally, in the Veteran Affairs Cooperation Study on the treatment of status epilepticus, 30-day outcome was significantly worse for patients with subtle status epilepticus than that with clinically overt generalized convulsive status epileptics (hospital discharged 50% vs. 8.8%; mortality 27% vs. 65%) (76).

Nonconvulsive seizures and status epilepticus may lead to secondary neuronal injury in critically ill patients. For instance, nonconvulsive status epilepticus and nonconvulsive seizure can cause hippocampal atrophy in patients with traumatic brain injury (81) and are associated with elevation in markers of brain injury measured using cerebral microdialysis (glutamate and lactate to pyruvate ratio) (82). Nonconvulsive status epilepticus in patients with intracranial hemorrhage has been associated with hematoma expansion (13) and in patients with subarachnoid hemorrhage and traumatic brain injury has been associated with transient brain hypoxia, metabolic crisis, increased intracranial pressure, and worse outcome (84; 18; 83). Despite these data, we should emphasize that nonconvulsive status epilepticus should not be considered a hopeless condition as some patients, after weeks to months of treatment, still recover with good outcomes (23; 43).

Clinical vignette

A 30-year-old woman started to complain of headache, altered mental status, and visual changes 3 days after cesarean section. Her physical examination was remarkable for decreased alertness, disorientation, and left hemianopsia. cEEG showed very frequent 5- to 10-minute-long electrographic seizures arising from the right posterior quadrant, maximal at O2, and focal slowing and attenuation in the same region. MRI revealed increased FLAIR and diffusion signal in the right temporal and occipital lobe. The patient was diagnosed with posterior reversible encephalopathy syndrome. She was treated with antihypertensive agents, magnesium, and levetiracetam with subsequent resolution of her symptoms.

Biological basis

Localization

Localization is highly variable, but a focal process is probably involved in most cases, even when the EEG appears to be generalized. Preliminary perfusion CT evidence suggests that hyperperfused cortical areas can be identified during nonconvulsive status epilepticus but are not seen in other postictal settings (85). Also, a study showed focal, regional, or diffuse areas of F-2-fluoro-2-deoxy-D-glucose (FDG) uptake on brain PET in patients with periodic EEG patterns (70).

Etiology and pathogenesis

The etiology of nonconvulsive status epilepticus varies by case. For many patients with preexisting epilepsy, sudden discontinuation of antiseizure medicines can lead to presentation with convulsive or nonconvulsive status epilepticus.

Among critically ill patients, etiologies often include infectious, severe metabolic derangements, traumatic, ischemic, or hemorrhagic cerebrovascular events, increased intracranial pressure, and inflammatory disorders, among others (79).

Pathophysiology

There is little information on the mechanisms underlying nonconvulsive status epilepticus, with most of the prior literature investigating convulsive status epilepticus. Although the physiologic differences between convulsive and nonconvulsive status epilepticus are incompletely understood, there is likely significant overlap between the two. As such, we discuss the current data as they relate to convulsive status epilepticus and the insight they offer into the pathophysiology of nonconvulsive status epilepticus.

Animal studies of untreated status epilepticus have identified five discrete electrographic stages of continuous seizure activity: discrete electrographic seizures, waxing and waning seizures, continuous ictal discharges, ictal discharges with flat periods, and periodic epileptiform discharges (77). Some human studies have found a similar sequence, but others have not, perhaps due to treatment effects.

Status epilepticus is thought to be due to the failure of endogenous CNS mechanisms to terminate a seizure. Animal models have provided the most insight into the roles of various neurotransmitters and their receptors in status epilepticus. GABA, glutamate, NMDA receptors, acidosis, Na-K-ATP’ase, potassium, adenosine, opioids, neuropeptide Y, substance P, and other factors have all been implicated in mediating seizure termination (50). For instance, adenosine is an endogenous anticonvulsant and may contribute to seizure cessation and the postictal state (05). Murine studies suggest that the adenosine A1 receptor has an important role in restricting seizure spread (05). Indeed, most animal knockouts involving adenosine receptor pathways display profound aggravation of status epilepticus (02; 41). In addition, prolonged seizures can also lead to the internalization of GABA-A receptors, thereby altering the balance of excitation and inhibition in neuronal circuits, promoting self-sustaining mechanisms and resistance to benzodiazepines as status epilepticus progresses (32). The accumulation of intracellular chloride or higher bicarbonate permeability may also play a role in the loss of GABA mediated inhibition (11). At the same time, prolonged seizures can lead to increased trafficking of AMPA and NMDA receptor subunits from intracellular stores to the cell surface, thereby altering the balance of excitation and inhibition in neuronal circuits (54). Periodic discharges are thought to involve oscillatory corticothalamic pathways that are unmasked when cortical inhibition is lost (08). They are likely to utilize some of the same subcortical-cortical projection pathways as sleep transients (K-complexes, spindles, and vertex waves) and arousal (including the upper brainstem), explaining why they are sometimes exacerbated by alerting stimuli (34). Metabolic abnormalities may play a role in mediating cortical hyperexcitability as well (56).

Seizures are terminated by various mechanisms acting at the membrane, local network, and remote level. Subcortical brain regions, particularly the substantia nigra pars reticulata (SNR), have been shown to have an important role in seizure termination. The SNR is positioned anatomically to receive seizure activity originating in the neocortical hemispheres and then can exhibit back inhibitory impulses via the thalamus, superior colliculus, and pedunculopontine nucleus in the rostral brainstem. Data from animal models show that prolonged seizures and status epilepticus may in fact impair the ability of the SNR to terminate seizures, further contributing to an ongoing epileptic state (42).

Epidemiology

The estimated incidence for status epilepticus is 10 to 41 per 100,000 per year; 5% to 49% of these cases are absence status epilepticus, focal status epilepticus with or without impaired awareness, and nonconvulsive status epilepticus in coma (49). According to the subtypes, from all forms of status epilepticus, 1% to 6% is absence status epilepticus in adults, 9% to 23% is focal status epilepticus with intact awareness, and 16% to 43% is focal status epilepticus with impaired awareness. These numbers are a compilation of multiple studies, and it should be noted that the incidence varies broadly due to the heterogeneous definitions and diagnostics criteria. The researchers also reported that the incidence of nonconvulsive status epilepticus ranges from 2 to 20 cases per 100,000, and 10% of the elderly with a protracted confusion will have late onset de novo absence status epilepticus (49).

In the hospital setting, nonconvulsive status epilepticus has been found to account for up to 20% of all cases of status epilepticus in general hospitals and up to 47% in the intensive care unit. Additionally, more than one third of patients in the emergency department and 8% in the intensive care unit with altered consciousness and no clinical signs of seizure were found to be in nonconvulsive status epilepticus detected by EEG (62; 72; 73).

The incidence of nonconvulsive seizure and nonconvulsive status epilepticus in children and neonate patients who were admitted to the intensive care unit, has been reported as 10% to 40% and 14% to 30%, respectively. In children, convulsive seizure or status epilepticus prior to starting EEG monitoring, structural brain injury, and several electrographic patterns (interictal or periodic epileptiform discharges, absence of background reactivity, and abnormal EEG background) increases the risk for nonconvulsive seizure.

Nonconvulsive status epilepticus is common among critically ill neonates and will be discussed in detail in a separate report.

Prevention

Among ambulatory patients with a diagnosis of epilepsy, adherence to antiseizure medication regimens is an important cornerstone in the prevention of status epilepticus of all types. For the majority of inpatient and critically ill patients, nonconvulsive status epilepticus is not usually preventable, and thus, management is usually focused on the rapid recognition, diagnosis, and treatment of status epilepticus when it occurs.

Differential diagnosis

Most nonconvulsive status epilepticus lacks obvious clinical manifestations apart from an altered level of consciousness, and the differential diagnosis of nonconvulsive status epilepticus includes a broad range of causes of altered mental status, coma, and delirium, such as toxic-metabolic encephalopathy, psychiatric disorders, trauma, tumor, vascular disease, migraines, thalamic or brainstem strokes, and confusional states induced by many drugs (such as lithium, alcohol, or psychotropic medications). Using the broad definition, any patient with coma and no other signs could have nonconvulsive status epilepticus. A high level of suspicion is important for recognition of nonconvulsive status epilepticus, which is more common than previously recognized (27), and cEEG monitoring is important for all patients with coma related to brain injury or history of convulsive status epilepticus or unexplained mental status changes (15).

In recent years, nonconvulsive status epilepticus has been reported as the presenting symptom of infection with COVID-19, and thus viral testing in patients with status epilepticus should be considered when appropriate (53).

Confusing conditions

Subtle myoclonic movements may be seen in some cases of nonconvulsive status epilepticus and should be distinguished from nonepileptic myoclonus. Subcortical myoclonus (presumably brainstem-generated in most cases), also known as reticular myoclonus, is probably the most common alternative diagnosis, and both cortical epileptic and subcortical nonepileptic myoclonus can coexist, particularly after a hypoxic insult. Typically, epileptic myoclonus involves muscle groups with larger cortical representations, such as the face and hand, whereas nonepileptic, subcortical myoclonus tends to involve axial muscles, such as sternocleidomastoid or abdominal muscles, though this may be difficult to distinguish in practice (24). In patients with nonconvulsive status epilepticus following generalized convulsive status epilepticus, jerking or myoclonic movements are common and can be elicited by alerting stimulation (without clinical arousal) (80; 34). The myoclonus is not always directly correlated with EEG discharges, and these patients may have a mixed cortical and subcortical etiology of myoclonus, perhaps relating to both the underlying epileptogenic injury and secondary brain injury from prolonged status epilepticus or hypoxia. In addition, muscle twitching arising from the spinal cord, peripheral nerves, or muscle (including due to metabolic processes) can be confused with the twitching of status epilepticus; for example, we have seen clonus related to exaggerated deep-tendon reflexes mistaken for seizure activity. This type of tendon reflex clonus is usually at the ankle or wrist and can be both elicited and terminated by positional changes.

Cortical myoclonus can be caused by a wide spectrum of diseases, including progressive myoclonic epilepsy, juvenile myoclonic epilepsy, chronic postanoxic myoclonus (Lance-Adams syndrome), and others (58; 67). Metabolic encephalopathies including renal failure (uremia) and hepatic failure (asterixis, thought to be subcortical in origin) (66) as well as electrolyte, thyroid, and vitamin derangements can also be associated with myoclonus (06). Numerous forms of neurodegenerative diseases can demonstrate myoclonus, including Lewy body dementia, Creutzfeldt-Jakob, and Alzheimer disease (up to 50% of patients with Alzheimer disease will develop myoclonus) (09; 06). When trying to characterize myoclonic movements, correlating EEG and EMG can be useful but may require computerized timing measurements and averaging for definitive results. In cortical myoclonus, the EEG discharge will consistently precede the muscle jerk, and the order of contractions of muscles will proceed from rostral to caudal (including within the cranial nerve-innervated muscles). In typical brainstem (reticular) myoclonus, there may or may not be an EEG correlate. If one exists, the EEG-EMG correlation may be variable with either one preceding the other, and lower cranial nerve muscles tend to be the first to contract, particularly the sternocleidomastoid (if involved) (58; 67).

Electroencephalographic phenomenology. Controversy surrounds many of the EEG findings seen in nonconvulsive status epilepticus with respect to whether or not the electrographic patterns represent ictal or interictal activity. Specifically, EEG findings such as periodic lateralized discharges, bilateral independent periodic discharges, and generalized periodic discharges are often associated with status epilepticus and nonconvulsive status epilepticus. However, although sometimes clearly ictal (especially if they fulfill the clinical criteria outlined in Table 1), these patterns are often thought not to represent definite seizures (12). Standardized terminology for these equivocal patterns was published as a guideline by the American Clinical Neurophysiology Society and shown to have acceptable levels of inter-rater reliability (33; 28). The real question is not what to call these EEG patterns (though standardized terminology is required for further research), but which ones are contributing to impaired mental status, cause neuronal injury, and will affect outcomes--as which ones need to be treated and how aggressively to treat them. Cerebral microdialysis, biomarkers, animal studies, intracranial recordings, and carefully designed clinical and imaging studies might help answer these questions (65).

Associated and underlying disorders

This seizure type can occur in any condition in which convulsive status epilepticus occurs.

Diagnostic workup

EEG or cEEG is paramount (in fact, required by most definitions) to establishing the diagnosis in a comatose or obtunded patient with suspected status epilepticus or subtle status epilepticus. cEEG monitoring for 24 hours detects up to 88% of seizures and another 5% and 7% are detected on monitoring days 2 and 3 (14).

In recent years, the use of rapid, bedside EEG including electrode caps or quick-apply headbands with limited EEG arrays have been successful in decreasing the length of time to diagnosis, particularly in centers with limited access to formal EEG (35; 47). Furthermore, for institutions that lack continuous video EEG, there is growing evidence that shorter duration studies still have clinical utility, with the majority of patients with nonconvulsive status epilepticus being identified within the first hour of EEG (19).

Although EEG remains the gold standard for the diagnosis of status epilepticus, many hospital systems are unable to obtain timely EEG monitoring. In the critical care setting, there is preliminary evidence that pupillometry with abnormal or asymmetric neurologic pupil indices may be a sign of concurrent nonconvulsive status epilepticus (31). In addition, numerous MRI changes can be seen during or as a result of focal nonconvulsive status epilepticus, including DWI abnormalities and cortical T2-weighted signal hyperintensity, which may provide evidence for ongoing or recent nonconvulsive status epilepticus (29). Although pupillometry and MRI can help support a diagnosis of nonconvulsive status epilepticus, they cannot definitively confirm the diagnosis and, thus, should be used only as an ancillary testing strategy if EEG monitoring is delayed.

Management

In contrast to convulsive status epilepticus, there is no evidence-based treatment consensus for nonconvulsive status epilepticus due to a lack of randomized controlled trials. Most of the treatment recommendations come from expert opinions and are based off evidence originally studying the treatment of convulsive status epilepticus.

Similar, to convulsive status epilepticus, benzodiazepines are often recommended for initial treatment of most forms of nonconvulsive status epilepticus, including focal status epilepticus with impaired awareness and nonconvulsive status epilepticus in critically ill patients. Second-line agents may include nonsedating intravenous antiseizure medications such valproic acid, fosphenytoin, phenytoin, topiramate, levetiracetam, and lacosamide (49; 36). Although fosphenytoin and valproate have traditionally been the most commonly used agents, newer drugs such as levetiracetam and lacosamide have gained increasing favor in more recent years given their similar efficacy and more favorable side effect profiles (10; 61). Because head-to-head data are lacking, the choice of which antiseizure medication to use is frequently based on tolerability and safety as determined by a patient’s comorbidities and clinical status. Frequently, more than one antiseizure medication may be required, in which case combining various mechanisms of action is frequently recommended; however, evidence for improved efficacy is lacking. When intravenous medications are unavailable or are contraindicated, enteral options, usually administered crushed through a nasogastric tube, have been reported to be effective in some cases (07; 46). Because there are few or no clinical findings, treatment response is assessed by cEEG monitoring. Treatment should be advanced until improvement is seen in the EEG or the clinical condition of the patient. Correcting metabolic abnormalities or treating systemic infection may also improve nonconvulsive status epilepticus in some critically ill patients. In patients who fail to respond to the first nonsedating antiseizure medication, often a second nonsedating antiseizure medication is recommended.

In cases of focal status epilepticus with impaired awareness that fail to respond to nonsedating antiseizure medications at appropriate doses, use of continuous intravenous anesthetic agents (such as midazolam or propofol) should be considered with appropriate cardiopulmonary support (49).

In subtle status epilepticus following generalized convulsive status epilepticus, the EFNS guidelines suggest general anesthetic doses of sedatives or barbiturates achieving burst suppression on EEG for a minimum of 24 hours until therapeutic doses of antiseizure medications are initiated (48). There is significant controversy about how aggressively to treat nonconvulsive status epilepticus that is not associated with prior convulsive status epilepticus in critically ill patients (40; 48). Some authors advocate using additional intravenous anticonvulsants, such as levetiracetam, valproate, or phenobarbital, prior to proceeding to anesthetic agents (48) as use of intravenous anesthetic may be associated with increased mortality risk (71).

Continuous EEG is increasingly used in critically ill patients with altered mental status to screen for nonconvulsive status epilepticus and nonconvulsive seizure though there is limited evidence that this test leads to improved outcomes. A retrospective cross-sectional study conducted to analyze cost-effectiveness and outcome of cEEG monitoring showed that the use of cEEG in the intensive care unit was associated with decreased in-hospital mortality in mechanically ventilated patients, without adding cost to the hospital stay compared to routine EEG (57). Novel risk prediction scores based on clinical and EEG findings may help optimize patient selection and duration of monitoring to improve cost-effectiveness (68; 69).

For the most refractory of cases, novel approaches with immunotherapy, steroids, ketamine, ketogenic diet, neurostimulation, and even targeted surgery can be considered (63; 20).

Outcomes

Concluding comments. Nonconvulsive status epilepticus has many subtypes that can be seen in the ambulatory and inpatient setting. The diagnosis of nonconvulsive status epilepticus should be considered when a patient with history of seizures and altered mental status is encountered in the clinic or if a patient is not back to his baseline after convulsive seizures, especially if neuroimaging or laboratory analysis do not provide an equivocal etiology for altered mental status. Nonconvulsive status epilepticus should also be considered in any critically ill patient with coma or fluctuating mental status, especially if they have a history of acute or remote brain injury. cEEG should be performed for at least 24 to 48 hours, and a trial of benzodiazepine should be considered when EEG discharges don’t meet electrographic nonconvulsive seizure criteria. Lastly, there is no consensus in the management of nonconvulsive status epilepticus, and therapy should be tailored to patient clinical history, presentation, and EEG finings.

A number of prognostic tools have been developed to predict outcomes after the treatment of status epilepticus, including the STESS (status epilepticus severity score), mSTESS (modified STESS), EMSE (epidemiology-based mortality score in status epilepticus), END-IT (encephalitis, NCSE, diazepam resistance, imaging features, and tracheal intubation), and SACE (seizures, age, coma, EEG evolution) (51). Features of these scoring systems are further detailed in Table 2.

Table 2. Scoring Systems

Special considerations

This is one of several clinical summaries originally developed as official descriptions of epileptic seizures and syndromes under the auspices of the International League Against Epilepsy (ILAE) Task Force on Classification and Terminology. It is now maintained by MedLink by agreement with the ILAE and may no longer reflect the official position of the ILAE.