Nonconvulsive seizures and nonconvulsive status epilepticus
Some seizures show on the electroencephalogram without the convulsions that make ordinary seizures recognisable. Nonconvulsive electrographic seizures and nonconvulsive status epilepticus are told apart mainly by how long they last. Both are most common among critically ill patients, generally involve an imbalance of excitatory and inhibitory neurotransmitters, present subtly and are found mainly by continuous recording. In 2014 they were treated with benzodiazepines and other antiepileptic drugs, while their own contribution to harm remained unsettled.
The two classes and their definitions
Recurrent seizures have been recorded as a medical condition for thousands of years, historically as the "falling sickness", and were considered readily recognisable because of their distinctive presentation, which often includes the violent involuntary muscle movements called convulsions. The study, diagnosis and treatment of nonconvulsive seizures, by comparison, was in its infancy. There are two major classes of nonconvulsive seizure, nonconvulsive electrographic seizures (NCES) and nonconvulsive status epilepticus (NCSE), distinguishable mainly by how long they last. NCES are spikes or rhythmic discharges of electrographic activity with clear evolution in frequency, location or morphology, lasting at least 10 seconds but no longer than 5 minutes. Their clinical features range from none at all to coma, which is why they are so often undetected and underdiagnosed, and patients sometimes show mental status changes and subtle myoclonic movements of a limb, the face or the eyes. NCSE is similarly electrographic activity without overt convulsions. Its definition had changed from a seizure lasting longer than 30 minutes to, by 2014, a nonconvulsive seizure lasting longer than 5 minutes, or two or more seizures without a return of consciousness between them, and this less stringent criterion brought more electrographic seizures within the diagnosis. The criteria defining the two were not universally accepted by practitioners.
Prevalence and risk
Prevalence varies with the patient population, and perhaps no population is more susceptible than critically ill patients, particularly those in a neurological intensive care unit. A large proportion of critically ill patients have seizures that are solely nonconvulsive. Young age substantially increases the risk for critically ill patients, and among children admitted to a paediatric intensive care unit because of altered mental status the prevalence is among the highest of any population. The incidence of each disorder by patient population or diagnosis was reported as follows.
| Patient population or diagnosis | NCES, % | NCSE, % |
|---|---|---|
| Patients in a neurological intensive care unit | 26.7–34 | 10–18 |
| Patients in a paediatric intensive care unit with altered mental status | 16.3–39 | 23 |
| Brain tumours in adults | 23 | 12 |
| Traumatic brain injury | 18 | 8 |
| Ischaemic stroke | 6 | 7 |
| Intracerebral haemorrhage | 18–28 | 9 |
| Subarachnoid haemorrhage | 8–18 | 8 |
| Central nervous system infection | 26 | 17 |
| Altered consciousness or subclinical seizures | 18–37 | 10 |
| Toxic-metabolic encephalopathy | 18 | 8 |
A generalised seizure is one with widespread discharges affecting both brain hemispheres. Brain tumours predispose to both disorders, and critically ill patients with traumatic brain injury, ischaemic stroke, intracerebral or subarachnoid haemorrhage, central nervous system infection, altered consciousness or toxic metabolic encephalopathy are at additional risk for NCES. Patients with a history of generalised convulsive status epilepticus and those who are comatose are also at increased risk.
Reported prevalence depends on how a study was done as well as on the disease, since the duration of the recordings and whether patients with prior convulsive activity were enrolled can both affect the figure. In recent years the disorders had gone from undetected or rarely diagnosed to being managed at least five times a year by 86% of surveyed neurologists, and the number of diagnoses had risen mainly because of better detection technology and the changed definition of NCSE. Increased continuous electroencephalographic monitoring of critically ill patients, particularly those with altered consciousness, had shown both disorders to be more common than previously thought.
Pathophysiology
The frequency and pathophysiology of the disorders vary considerably with the underlying cause, and the pathophysiology generally involves an imbalance of excitatory and inhibitory neurotransmitters, including glutamate and γ-aminobutyric acid, that leads to neuronal excitotoxicity and damage and, ultimately, seizures. Epileptogenesis is the triggering of seizures, and an epileptogenic aetiology is the underlying diagnosis that results in seizures. The pathophysiology is complex because it involves several excitatory, inhibitory and metabolic pathways and can vary with the epileptogenic aetiology. Convulsive and nonconvulsive seizures, for example, contribute to further epileptogenesis by directly precipitating neuronal damage, while brain tumours do so through bleeding-mediated release of iron, oedema, necrosis, a rise in peritumoral pH, raised N-methyl-D-aspartate levels and imbalances between excitatory and inhibitory neurotransmitters. Some of these mechanisms are also prominent in other neurological and psychiatric disorders: the same neurotransmitter imbalances have been implicated in Alzheimer disease, schizophrenia, anxiety and depression, and the same putative mechanisms form the biological basis for treating those disorders.
The seizure threshold is the balance between excitatory and inhibitory forces in the brain, which affects how susceptible a person is to seizures. Most seizures are caused by abnormal electrical discharges in the brain, and raised extracellular excitatory amino acids can initiate seizure activity by triggering waves of neuronal depolarisation that exceed the seizure threshold and lead to prolonged neuronal damage. Glutamate, the most abundant excitatory amino acid in the central nervous system, has a key role in this process. It is normally stored in presynaptic vesicles, and its extracellular level is tightly regulated at several stages, and changes in that level are epileptogenic. Dysregulated ion exchange can depolarise the neuronal membrane so that excess glutamate is released into the synapse, where it triggers N-methyl-D-aspartate receptor–mediated release of excess calcium inside the postsynaptic cell, a process implicated in programmed cell death. Alkalisation of extracellular pH has also been associated with membrane depolarisation, so conditions that alter pH could be epileptogenic. Dysregulation of these excitatory pathways is believed to be a component of the underlying pathophysiology of NCES and NCSE. Dysregulation of the glutamate excitatory pathway in particular has also been called a major mechanism for both.
The brain has defences against seizures caused by excitotoxic damage. Extracellular glutamate prompts nearby interneurons to release inhibitory neurotransmitters such as γ-aminobutyric acid, which suppresses the excitability of the discharging neuron; synaptic glutamate activates nearby sodium–potassium adenosine triphosphatase transporters, which take glutamate back up from the synapse and so attenuate glutamate-mediated excitotoxicity; and glial cells buffer extracellular potassium, which also suppresses neuronal excitability. Failure of those inhibitory mechanisms, or injury to the inhibitory interneurons or glial cells, may also precipitate the disorders.
Clinical presentation
The clinical features are protean and subtle. Some patients show no cognitive effects, but the typical presentation involves some degree of altered mental status, such as confusion, catatonia or depressed mental acuity, and about 8% to 31% of patients with NCSE are comatose. Personality changes such as agitation and aggression have been noted. Involuntary eye movements occur: blinking, staring, nystagmus (spontaneous slow horizontal or vertical eye movements), sustained deviation or misalignment of the eyes, and hippus, an irregular ocular spasm with exaggerated dilation and contraction of the pupil. Some patients have subtle facial twitching, occasional muscular twitches of a limb, or automatisms, a set of brief unconscious behaviours.
Diagnosis and continuous electroencephalography
Both disorders had gone undetected and underreported because their symptoms are ambiguous, their diagnosis complex and the technology of local hospitals limited. The subtle presentation and the complexity of interpreting the electrographic signatures together complicate diagnosis. Clinicians must assess the medical history, concomitant medicines, family history of seizures, movements of the face, eyes and limbs, and level of consciousness. An electroencephalographer, often a specialist in clinical neurophysiology with several years of experience, reviews individual discharges and the overall pattern for the features distinctive of the disorders. The need to consult such specialists also lets the disorders go unnoticed and underreported, whereas convulsive seizures and status epilepticus can be relatively straightforward to diagnose because of their distinctive clinical manifestations.
Routine, noncontinuous electroencephalograms often miss the disorders, as do some newer seizure-detection tools, but continuous recording over a prolonged period has proved effective. Continuous electroencephalography is analogous to telemetry electrocardiography: arrhythmias, like nonconvulsive seizures, often occur sporadically without the patient knowing and can be missed by a routine electrocardiogram. Just as telemetry gave a more accurate picture of the true prevalence and clinical significance of arrhythmias, continuous electroencephalography was changing the perception of how frequent and how grave nonconvulsive seizures are. Given the rising incidence and perceived significance of the disorders, continuous recording was increasingly used in neurological intensive care units.
Pandian and colleagues compared a routine 30-minute recording with continuous recording of median 2.9 days and found that the routine recording would have missed more than half of all seizure cases. Their rate combined NCES with convulsive seizures, which are easier to diagnose, so it likely understates how many NCES diagnoses would be missed. Jirsch and colleagues found seizure activity in only 15% of critically ill patients at the start of recording but in nearly half by one hour. Because the disorders can occur without overt clinical signs, clinicians increasingly used video monitoring time-locked to the recording to catch subtle movements, and were training bedside caregivers to annotate the record with medication changes, details of the electrodes and recording, and suspected seizures. Continuous and video-continuous recording together had pivotal roles in the increased number of diagnoses and the corresponding increased prevalence.
Treatment
Treatment combined non-drug and drug measures. Comorbidities are concurrent disease states, which may include the underlying cause of the seizures. Intubation was indicated where comorbidities or the drugs themselves made respiratory failure likely. Intravenous access was key because many treatments are given intravenously, and continuous monitoring gave insight into diagnosis, response to treatment and the possible need to adjust therapy. Because the disorders are heterogeneous, laboratory tests were obtained to identify and treat the underlying cause. In patients with a history of epilepsy, antiepileptic drug levels should be measured and subtherapeutic levels corrected quickly, and higher maintenance doses may be considered, especially for adherent patients. Other critical steps include withholding medicines, correcting metabolic disturbances and treating infections that could be the trigger. The cause also affects the treatment options in some cases.
Events occasionally resolve without drugs, but most need them. Benzodiazepines were first-line for most events, lorazepam frequently and diazepam less commonly, because its effect is shorter and its elimination more complex. Second-line drugs, used when patients do not respond, included phenytoin, fosphenytoin, valproate and levetiracetam. Barbiturates such as phenobarbital, though typically considered as effective as benzodiazepines, were generally reserved for third line because their toxicity is less manageable. Adverse events were major considerations, and death from treatment often resulted from respiratory depression, hypotension, pneumonia related to respiratory depression and haemodynamic instability. Overtreatment with barbiturates and benzodiazepines can cause respiratory suppression and respiratory failure, and hypotension is likewise a toxicity of both. As in the treatment of some malignancies and infectious diseases, the side-effect profile is a key determinant of the choice of treatment. So, to minimise toxicity and exploit potential synergy, patients were often given combination therapy or switched immediately to another drug after a brief benzodiazepine course, and phenytoin and fosphenytoin sometimes joined benzodiazepines as part of first-line treatment. Because many events were inadequately treated, new drugs were under investigation, and existing treatments for epileptic and non-epileptic disorders were also being explored.
Clinical significance and open questions
Both disorders have been associated with mortality and with significant morbidity among critically ill patients, including brain damage, injuries, severe disability and a vegetative state. Because they occur in some of the most clinically vulnerable patients and are inherently hard to diagnose, their true clinical significance may be markedly understated. Robust research on the clinical outcomes of the disorders was needed. They are thought to cause considerable physiological and metabolic stress that can precipitate or worsen neurological injury after an acute insult. But concurrent medicines and underlying and concomitant systemic disorders, such as infection, traumatic brain injury and hypotension, can do synergistic damage, so only limited evidence of permanent dysfunction is attributable to the disorders alone. The disorders also often occur with underlying causes that are inherently neurotoxic or fatal. Case series and laboratory data nonetheless suggest that prolonged or continual ictal discharges, the electrical discharges that occur during the seizures, are associated with a significant increase in morbidity or mortality. In animal studies, preventing the seizures with antiepileptic drugs significantly reduced mortality. So even without other pathology, morbidity and mortality are serious potential concerns that should be investigated in adequately powered clinical studies.
Although much had been learned about the disorders, many important questions about their overall clinical significance and management remained unanswered. A better understanding of outcomes would aid evidence-based recommendations; the objective benefit of acute treatment, for instance, had not been quantified, and it was unclear which patients should be treated at all. There had been a recent surge in academic and pharmaceutical research into the mechanisms underlying the disorders, their detection and their treatment. Ongoing and planned research focused on diagnosis, including biomarkers, on treatment and on outcomes.