Nonconvulsive status epilepticus in patients under intensive care: Should we view epilepsy as a sleep disorder?

Abstract

Nonconvulsive status epilepticus (NCSE) was initially described in patients with typical and atypical absence status epilepticus (ASE) characterized by states of confusion varying in severity and in focal epilepsies with or without alteration of consciousness. Continuous EEG monitoring of critically ill patients has further refined the classification of NCSE into two main categories: with coma and without coma. Hypnotic, soporific or somniferous epileptic seizures do not exist. On the contrary, patients usually awaken when seizures occur during sleep, and their eyes remain open during ASE. Excessive sleepiness and coma alone are not ictal signs but are observed in the postictal phase of convulsive seizures. On the other hand, excessive sleepiness evolving into coma is a cardinal sign of metabolic/toxic encephalopathies with triphasic waves evolving to burst suppression patterns and ultimately to cerebral inactivity and death. NCSE alone does not directly cause coma. Comas are related to the underlying etiology, patient age and comorbidities, as well as the administration of intravenous sedative drugs to control epileptic seizures. In cases of severe brain injury, NCSE can explain the failure to awaken after the withdrawal of anesthetics and is only an aggravating factor of the neurological condition. In typical ASE, which is characterized by sustained, rhythmic, bilateral, synchronous and unreactive discharges with evolving spatiotemporal patterns (the best example of NCSE), there is no vigilance impairment. This contrasts with metabolic/toxic encephalopathies, which exhibit monomorphic generalized periodic discharges in which patients may become comatose and die. The extended concept of NCSE in comatose patients may lead to an inflated assessment of NCSE, implying a potentially worse prognosis compared to convulsive status epilepticus.

Key points.

• Absence status epilepticus is characterized by confusion of varying severity, but patients do not experience excessive sleepiness.

• NCSE alone does not directly cause coma.

• The origin of the coma must be determined using a clinical approach that first integrates clinical and neuroradiological data, followed by EEG results.

• Generalized epileptiform activity in comatose patients primarily reflects severe encephalopathy rather than NCSE.

• NCSE in comatose patients is an aggravating factor of the neurological condition.

1. INTRODUCTION

Nonconvulsive status epilepticus (NCSE) was initially described in patients with epilepsy but is now described in other patients also, including those who are critically ill. ¹ , ² In 1945, Lennox introduced the term “petit mal status,” later replaced by “absence status epilepticus” (ASE), to specify the picture of mental confusion of varying severity associated with bilateral spike–wave discharges. ³ Based on semiology and ictal electroencephalograms (EEGs), NCSE has traditionally been subdivided into nonconvulsive and confusional status epilepticus (SE; typical and atypical ASE, focal SE with impaired consciousness), and nonconvulsive–nonconfusional SE (e.g., aphasic SE). ⁴ , ⁵ , ⁶ Continuous monitoring of critically ill patients with EEG has further refined the classification of NCSE into two main categories: with and without coma. ⁷

A key question that arises when analyzing EEG signals is whether altered consciousness is a consequence of EEG discharges. ⁸ , ⁹ , ¹⁰ This article investigates the clinical relevance of classifying NCSE based on the level of vigilance and explores whether broadening the NCSE diagnosis in comatose patients could cause overinterpretation of EEG data at the expense of clinical context. This may lead to an inflated assessment of NCSE, implying a potentially worse prognosis compared to convulsive SE. ¹¹

2. EEG PATTERNS WITHIN NCSE

Using the Glasgow Coma Scale (GCS), Lattanzi et al. identified three phenotypic clusters within NCSE: GCS 3–8 with spontaneous burst suppression (BS; cluster one); GCS 9–12 with lateralized periodic discharges (LPDs) or generalized sharp and/or triphasic periodic discharges (cluster two); and GCS 9–12 without LPDs, generalized periodic discharges (GPDs), or BS (cluster three). ¹² Except for infants with early infantile developmental and epileptic encephalopathy (DEE), patients with “spontaneous” BS patterns are always in a deep coma, aligning them with cluster one. Patients with GPDs consistently exhibit altered consciousness levels, whereas those with LPDs may experience normal or mildly impaired consciousness (Figure 1A; Data S1). In patients with LPDs, the level of consciousness depends mainly on the underlying causes and comorbidities rather than directly on the EEG pattern. During a 6‐year study period, Lattanzi et al. did not assess patients with bilateral independent periodic discharges (BIPDs) on their EEGs, which are typically observed in comatose patients with acute brain lesions. BIPDs can be associated with electrographic seizures in nearly half of these patients ¹³ and may indicate a pattern of NCSE (Data S1). However, BIPD prevalence is relatively low, estimated to be approximately .5%–1% among critically ill patients undergoing continuous EEG monitoring. ¹⁴ Although periodic and rhythmic discharges (RDs) may coexist on the same EEG (Data S1), ¹⁵ RDs may appear as the sole pattern of NCSE. In clinical practice, focal SE with subclinical seizures or seizures with minor motor signs is a frequently observed cause of NCSE. When RDs are present, the level of consciousness/vigilance is variable, ranging from full awareness to deep coma (Data S1).

FIGURE 1.

(A) A 59‐year‐old man treated with carbamazepine and levetiracetam. Brain surgery had been performed 2 years earlier for cerebral metastasis; the patient was hospitalized for language disturbance. Lateralized periodic discharges (LPDs) at 1 Hz with polyspikes and high‐voltage polyphasic complexes corresponding to a focal nonconvulsive status epilepticus (NCSE) are shown. (B) A 62‐year‐old man with surgery 1 month earlier for glioblastoma, who was hospitalized for aphasia and impairment of vigilance. Relapse was seen on computed tomography scan. LPDs in the left centroparietal regions with complexes of polyspike–waves at 1 Hz corresponding to a focal NCSE are shown. Comparing these two electroencephalogram panels, the first patient only had aphasia (Data S1), but the second one had aphasia and impairment of consciousness due not to the focal NCSE but to the relapse of glioblastoma. A pure electric phenomenon (focal NCSE) cannot be responsible for an isolated impairment of consciousness, but in the second case, maintained the patient in a coma.

ASE is the best example of NCSE. Atypical ASE may be observed in DEEs like Dravet syndrome (DS), Lennox–Gastaut syndrome (LGS), and ring chromosome 20 syndrome. In these patients, there is no severe impairment of consciousness and symptoms include a fluctuating confused state, sometimes referred to as “bad days.” ¹⁶ EEG findings in LGS include irregular, diffuse, or anterior slow waves or polyspike–waves, typically at lower frequencies than at the baseline. ¹⁷ , ¹⁸ In DS, EEGs reveal continuous anterior irregular high‐amplitude delta waves mixed with focal or diffuse spikes. ¹⁸ , ¹⁹ Typical ASE can occur in virtually all generalized genetic epilepsies with typical absences. EEG reveals bilateral synchronous, symmetric, rhythmic paroxysmal activity that is unreactive to sensory stimulation. ²⁰ Typical ASE is characterized by confusion of varying intensity, ranging from mild (only subjective) to profound obtundation, leading to catatonic stupor, but eyes are open (Figure 2A; Data S1). ²¹ , ²²

FIGURE 2.

(A) Typical absence status epilepticus (ASE) in a 24‐year‐old woman with juvenile myoclonic epilepsy precipitated by abrupt discontinuation of clonazepam. Clinically, she was in a catatonic stupor with subtle myoclonic jerks of the face and arms. She did not obey simple orders. The electroencephalogram (EEG) shows rhythmic, continuous bilateral activity at 5–6 Hz that was not reactive to stimulation. After administration of intravenous (IV) clonazepam 1 mg, there was normalization of the EEG and consciousness. (B) An 88‐year‐old woman with type 2 diabetes mellitus hospitalized for impairment of consciousness. Blood tests showed high creatinine (154 μmol·L⁻¹), high urea (14 mmol·L⁻¹), and hyperglycemia (11.5 mmol·L⁻¹). Cerebral atrophy with vascular leukoencephalopathy was seen on magnetic resonance imaging. The EEG shows generalized periodic discharges with triphasic morphology at 2 Hz with predominance in the mid regions. Comparing these two patients, the first one with typical ASE and with continuous rhythmic activity at 5–6 Hz corresponding to a sustained EEG pattern has her eyes open and breathes normally (Data S1), whereas the second patient with a metabolic encephalopathy with monomorphic generalized periodic discharges at 2 Hz on her EEG is drowsy.

3. WHEN COMA BECOMES A NEGATIVE ICTAL MANIFESTATION

Seizures are characterized by excessive neuronal discharges. In focal SE, this hyperactivity can be observed using different neuroimaging techniques. ²³ , ²⁴ , ²⁵ Few negative signs (e.g., aphasia, hemianopsia, motor deficits) can be observed during a seizure. ²⁶ Jirsch and Hirsch list coma as a possible negative symptom (semiological spectrum of nonconvulsive seizures and NCSE), ¹ , ² and several studies have reported NCSE with coma based on a false syllogism: (1) the EEG shows generalized epileptiform discharges; (2) the patient is in a coma; (3) therefore, this patient has NCSE with coma. ²⁷ , ²⁸ To assume that an isolated impairment of vigilance, such as coma, is a symptom of NCSE is analogous to suggesting the existence of “soporific,” “hypnotic,” or “somniferous” seizures. In contrast, patients often awaken from focal and generalized seizures, including nonconvulsive seizures with 3‐Hz spike–wave discharges in absence epilepsies. ¹⁸

NCSE alone does not directly cause coma, as coma is not an ictal manifestation. Rather, it may occur in the postictal phase of convulsive seizures or may be the terminal stage of untreated or insufficiently treated generalized tonic–clonic SE (subtle SE), ²⁹ which is a distinct and rare form of NCSE associated with coma. Comas are primarily related to the underlying etiology (e.g., hemorrhagic stroke and severe head trauma), rather than EEG patterns. Factors such as patient age and comorbidities also contribute to the altered level of vigilance observed in SE (Data S1). For instance, young adults may respond differently compared to older individuals with conditions like sliding syndrome, Alzheimer disease, Parkinson disease, stroke history, malnutrition, chronic renal impairment, or heart failure. The use of intravenous benzodiazepines to control epileptic seizures can rapidly lead to severe consciousness disorders in this population (Data S1). In older adults with brain atrophy and chronic anemia, a combination of focal SE and hypoxemia can rapidly lead to severe impairment of consciousness. Simple pneumonia in older adults can cause drowsiness, delirium, and increased risk of falls without clear evidence of fever, mimicking the presentation of NCSE. On EEGs, respiratory encephalopathies are characterized by slow background activity with possible triphasic waves (TWs). ³⁰ Other factors, such as sepsis, high doses of penicillin, and chronic renal impairment, can further contribute to the alteration of vigilance (multifactorial encephalopathy).

4. COMA IS A CARDINAL SIGN OF ACUTE ENCEPHALOPATHIES WITH TWs EVOLVING TO BS

As previously stated, “hypnotic” seizures do not exist. However, excessive sleepiness is a hallmark of toxic, metabolic, or multifactorial encephalopathies with TWs, characterized by a clinical progression from excessive sleepiness to coma (Figure 2B). The EEG patterns in such cases begin with slow background activity and consequent appearance of TWs, which become continuous. This is followed by a transition featuring the disappearance of the TWs and the appearance of BS patterns, progressing to cerebral inactivity and death. ²⁰ , ³¹ Misinterpreting BS patterns as NCSE poses a significant challenge. ¹² , ²⁸ , ³² Spontaneous BS in adults is not an EEG pattern of NCSE, but rather a marker of severe encephalopathy. BS is observed in severe postanoxic encephalopathies with poor outcomes ³³ and in metabolic, toxic, or multifactorial encephalopathies (Data S1). In patients without an evident history, BS patterns are often linked to drug overdose, such as baclofen intoxication or the effect of short‐acting barbiturates when used as hypnotics. ³⁴ , ³⁵ In contrast, subtle SE demonstrates periodic patterns on EEGs, ²⁹ including GPDs that can be likened to BS patterns due to the presence of low‐voltage background activity between the discharges. Subtle SE, although a form of NCSE, must be differentiated from other NCSE subtypes because of its unique clinical context, different prognosis, and specific treatment considerations. ²¹

5. WHEN GPDs BECOME A PATTERN OF EPILEPSY

The American Clinical Neurophysiology Society proposed changing TWs to GPDs with triphasic morphology due to their clinical implications. ³⁶ The change in terminology may have shifted the focus from encephalopathy to seizure activity or the ictal–interictal continuum. TWs are essentially EEG patterns of acute encephalopathies. However, similar findings associated with bilateral anterior slow spike–waves may be found in LGS ³⁷ and related DEEs. TWs can also occur in cases of severe brain injury (nonmetabolic TWs) ³⁸ accompanied by coma, occasionally reflecting NCSE. In such situations, it is crucial to differentiate typical from atypical TWs and to understand the circumstances of their appearance, such as asymmetrical and unreactive TWs in cases of hemorrhagic stroke or severe head trauma. It is also important to observe fluctuations in their frequency and morphology and their association with typical spikes (Data S1). GPDs with triphasic morphology are also observed in comatose patients following BS patterns induced by self‐poisoning with barbiturates, ³⁴ , ³⁵ or after weaning them off intravenous anesthetics, particularly pentobarbital and, ³⁹ , ⁴⁰ , ⁴¹ more rarely, propofol. ³⁹ , ⁴¹ This pattern of GPDs, which is related to anesthetic withdrawal, represents a transient and resolving encephalopathy. ⁴¹ Additionally, TWs may occasionally be seen in Alzheimer disease and Creutzfeldt–Jakob disease. ⁴² , ⁴³

Patients with GPDs with triphasic morphology do not have normal levels of consciousness, with states ranging from excessive sleepiness to coma, and in Lattanzi et al.'s study, it is logical to place these patients in the second cluster. ¹² In this series, younger people were not at risk of NCSE, with a median age at onset of 77 years (range = 67–84 years). Older people with brain atrophy, white matter disease, and chronic renal failure are at heightened risk of developing metabolic, toxic, or multifactorial encephalopathies with TWs compared with younger people. ³⁸

6. CONCLUSIONS

In contrast to narcolepsy, epilepsy is not a sleep disorder. An isolated nonconvulsive EEG phenomenon cannot be solely responsible for the alteration of consciousness. In typical ASE, characterized by sustained, rhythmic, bilateral, synchronous, and unreactive discharges with evolving spatiotemporal patterns (the best example of NCSE), there is no vigilance impairment. This contrasts with metabolic or toxic encephalopathies, which exhibit monomorphic GPDs with triphasic morphology. In a comatose patient, epilepsy may only be an epiphenomenon. ⁹ However, in patients with acute brain damage, epilepsy is an aggravating factor of the neurological condition, prolonging the coma after anesthetics withdrawal, and indirectly contributing to complications, such as pulmonary infection and death.

AUTHOR CONTRIBUTIONS

Philippe Gélisse: Conceptualization; writing—original draft; writing—review & editing. Arielle Crespel: Writing—review & editing. Both coauthors were substantially involved in the study and preparation of the manuscript. No undisclosed persons have had a primary role in the study or manuscript preparation.

CONFLICT OF INTEREST STATEMENT

P.G. has received support for teaching programs from UCB and Eisai and royalties for publishing from John Libbey Eurotext. A.C. has received support for teaching programs from UCB and Eisai and royalties for publishing from John Libbey Eurotext. She was an advisory board member for Eisai‐France. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Supporting information

Data S1.

Gélisse P, Crespel A. Nonconvulsive status epilepticus in patients under intensive care: Should we view epilepsy as a sleep disorder? Epilepsia. 2025;66:1041–1047. 10.1111/epi.18274

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in the supplementary material of this article.