Febrile Status Epilepticus

Darrell V. Lewis

Shlomo Shinnar

Introduction

Febrile seizures (FS) are acute symptomatic seizures that occur in 2% to 5% of children, making them the most common form of childhood seizures. About 5% of FS can be classified as febrile status epilepticus (FSE). Due to the long-debated relationship between prolonged FS, mesial temporal sclerosis (MTS), and temporal lobe epilepsy (TLE), there is currently a great deal of interest in the causes, effects, and long-term prognosis of FSE.

Definitions

The International League Against Epilepsy defines a FS as a seizure in association with a febrile illness in the absence of a central nervous system (CNS) infection or acute electrolyte imbalance in children older than 1 month of age without prior afebrile seizures.²⁰^,²¹ FSE is defined as a FS of 30 minutes’ duration or greater, or a series of FS lasting more than 30 minutes, between which the infant does not return to normal level of alertness.²⁰^,⁷³ FS, and by extension FSE, are also divided into those without prior neonatal seizures and those with prior neonatal seizures. No upper age limit is specified; however, FS, and also FSE, are most common between 6 months and 3 years of age, with peak incidence at about 18 months of age.

Epidemiology

Although the majority of FS are simple, approximately 5% are of sufficient duration to be classified as FSE. In a study of 428 children with a first febrile seizure, 14% had duration of longer than 10 minutes, 9% of longer than 15 minutes, and 5% of longer than 30 minutes or FSE.⁸ Although only 5% of FS are FSE, these seizures nevertheless account for approximately 25% of all childhood status epilepticus (SE)²³^,⁷⁴ and more than two-thirds of cases of SE in the second year of life. FSE can also be classified as partial (focal) or generalized, depending on the presence or absence of focal features in the ictus. The age distribution of FSE is identical to that of simple FS, with 74% of episodes occurring younger than 2 years and 96% at ages younger than 5 years.⁷³

Anatomic Pathways and Pathophysiology

Anatomic Pathways

The site of origin and pathways of propagation of electrical seizure activity underlying FS and FSE are unknown. Generalized delta waves, spikes, and polyspikes have been seen in simple FS recorded fortuitously in human infants (Morimoto et al 1991; Maehara 1988). The electroencephalographic (EEG) sources and propagation patterns during FS will very likely depend on the type of seizure (e.g., focal vs. generalized) and on the genetic susceptibility factors in each case. There are a few anecdotal case reports of EEGs recorded during FSE, one showing ictal activity over a temporal lobe and several showing diffuse bilateral but asymmetric ictal activity during FSE.⁵⁷^,⁷⁷ Interestingly, in these few cases, MTS was ultimately documented by follow-up magnetic resonance imaging (MRI) in the temporal lobe ipsilateral to the predominant ictal EEG activity.

More information is available on the EEG correlates of hyperthermia-induced seizures in rats used to model FS in human infants. Hyperthermia to approximately 42°C core temperature reliably induced 20-minutes’ duration seizures in 10- to 11-day-old rats with electrodes placed on the cortex, and in the hippocampus and amygdala. The clinical seizure activity suggested limbic origin, and the EEGs documented onset in the amygdala or, less often, in the hippocampus.⁶ Others have also found evidence of limbic origin of hyperthermia-induced seizures in infant rats.⁶⁷ The paucity of information on anatomic pathways associated with FSE prevents any firm conclusions, however.

Pathophysiology

The pathophysiology of FS has been studied only to a limited extent in animals, and even less is written on models of FSE. To model FS, investigators have induced hyperthermia in rats using radiant heat, hot air, or hot water immersion. A consistent finding has been that hyperthermic seizures in infant rats result in lowered seizure thresholds when the affected rats are adults.²⁵^,³⁹ The most intensely studied model exposes 10-day-old rat pups to heated air, consistently evoking hyperthermic seizures of about 20 minute duration, apparently arising in the hippocampus or amygdala.⁶ Within about 24 hours and up to about 2 weeks following these seizures, silver staining revealed argyrophilic neurons in the central nucleus of the amygdala and in CA3 and CA1 sectors of the hippocampi in these rats.⁷⁶ Ultimately, the argyrophilia resolves, and no cell loss is noted when cell counts are done at later age.⁷ Nevertheless, hippocampal excitability is permanently altered, and the threshold for chemically induced seizures is permanently decreased.²⁵ This persistent decrease in seizure threshold after infantile hyperthermic seizures has been noted by prior investigators also.²⁵^,³⁹ Although the limbic seizure threshold is reduced, there is a paradoxical increase of perisomatic γ-aminobutyric acid (GABA)_A inhibition due to increased pre-synaptic GABA release.¹⁹ In addition, a permanent alteration occurs in the properties of the hyperpolarization-activated cation current (I_h) in hippocampal CA1 and CA3 pyramidal cells.¹⁸ I_h is a depolarizing current carried by sodium (Na) and potassium (K) that is normally activated by hyperpolarization of the cell and tends therefore to counteract hyperpolarization and to return the cell to a more depolarized level following hyperpolarization. After rat pups have experienced hyperthermic seizures, hippocampal I_h is activated at more depolarized levels, resulting in mild resting depolarization of pyramidal cells and an enhanced posthyperpolarization rebound with depolarizing overshoot and a resultant burst of spikes. Therefore, when a brief tetanic train of stimuli is applied to the pyramidal neurons, the enhanced GABA_A inhibition produces a hyperpolarization that is followed by a depolarization and burst of spikes due to the enhanced I_h. This tendency for bursting may underlie the reduced seizure threshold. The altered I_h properties result from a change in relative expression of isoforms of these channels that persists into adulthood.¹²

The only structural alteration so far noted in these animals is mildly increased mossy fiber density in the granule cell and molecular layer of the dentate gyrus.⁷ It is very interesting that many of these rats develop spontaneous limbic seizures in adulthood even though the prolonged hyperthermic seizures they suffered as infants did not produce hippocampal sclerosis.²⁶ Thus, this model raises the possibility that a single prolonged febrile seizure in a human infant could permanently alter hippocampal seizure threshold and even predispose to TLE in the absence of obvious structural alteration.

Many investigators have proposed that both prolonged FS and the epilepsy that occasionally develops after these seizures may be a result of preexisting brain pathology. A few animal studies have combined induced brain anomalies with hyperthermic seizures to test this concept. Rat pups exposed to methylaxozymethanol (MAM) during gestation display extensive cortical and hippocampal dysgenesis and a heightened vulnerability to hyperthermia-induced seizures.³⁷ Hyperthermia induced cortical electrographic seizures in both the MAM-treated and the control pups. However, the MAM-exposed pups had a higher incidence of behavioral (clinical) seizures, probably due to increased spread of seizure activity. Significant pyramidal cell loss, independent of seizure activity and presumably due to hyperthermia alone, was documented in the CA1 and CA3/4 hippocampal subfields of MAM rats 4 weeks after exposure to hyperthermia. It was concluded that extensive dysgenesis lowers the threshold for hyperthermia-induced behavioral seizures and irreversible neuronal injury. In a different model, more restricted neocortical freeze lesions causing a focal microgyrus have also been shown to predispose infant rats to more severe hyperthermia-induced seizures.⁶⁶ These lesions reduced the latency to hyperthermia-induced generalized convulsions, and the convulsions themselves were followed by prolonged electrographic seizure activity compared with hyperthermic seizures induced in controls. Finally, video-EEG recordings from the same rats as adults revealed a high incidence of spontaneous limbic seizures in the rats with both lesions and hyperthermic seizures but not in lesion-only or hyperthermic seizure–only rats.⁶⁶ As was the case in the pure hyperthermia model,⁶ the spontaneous limbic seizures occurred without any obvious evidence of hippocampal sclerosis induced by the hyperthermic seizures experienced in infancy. These animal studies support the hypothesis that preexisting brain abnormalities could predispose to prolonged FS and subsequent epilepsy in human infants.

Contribution of Magnetic Resonance Imaging to Understanding Pathophysiology of Febrile Status Epilepticus

FSE is an uncommon and unpredictable event and generally does not recur. Therefore, studies of pathophysiology of FSE are quite limited and consist of tests such as EEGs and MRIs that can be done during or shortly after the acute and emergent presentation of the affected infants.

Animal Models

A brief review of MRI changes observed with animal models of SE will help put the human data in perspective. MRI analyses of SE in rats usually employ limbic SE evoked by kainic acid (KA) or pilocarpine, and hence the anatomic distribution of abnormalities reflects the regional specificity of these agents, with changes being seen most commonly in the amygdala, piriform cortex, and hippocampus as well as in extralimbic areas of the neocortex and thalamus. Six to 24 hours following lithium-pilocarpine–induced SE in adult rats, the apparent diffusion coefficient for water (ADC) is usually reduced in the amygdala and piriform cortex, and thalamus and cortex if affected.³¹^,³⁴ Along with ADC changes, T2 signal intensity is increased in amygdala, piriform cortex, hippocampus, and often in thalamus and cingulate cortex. The areas of increased T2 signal intensity have, in general, correlated with areas of neuronal injury, neuronal loss, and tissue edema on histologic analysis.³¹^,⁶² Acute ADC changes in the hippocampus in these models have been inconsistent. Some have seen an increase in ADC initially using pilocarpine,³¹^,³⁴ and in others no significant change using KA.⁵²^,⁶⁰^,⁷⁵ When the hippocampal MRI changes have been followed beyond the acute phase for 2 or more weeks, increased ADC and increasing T2 signal have been seen along with atrophy and cell loss.⁵⁹^,⁶³^,⁷⁵ Acutely increased piriform and entorhinal cortex T2 intensity and T2 relaxation times were predictive of ultimate development of spontaneous seizures and increased hippocampal T2 signal in a study of pilocarpine SE in 21-day-old adolescent rats.⁶³ Thus, in adult and adolescent rats, MRI changes following status reflect tissue damage and may be surrogate markers for the development later of MTS and limbic epilepsy.

The significance of MRI abnormalities is less clear in hyperthermic seizures of shorter duration studied in infant rats. There is only one MRI analysis of prolonged hyperthermic seizures in infant rats using P10 rat pups exposed to heated air and suffering approximately 20-minute-long seizures. These are relatively brief seizures, compared to the 2- to 4-hour-long SE typical of the kainate or pilocarpine models. Nevertheless, transiently increased T2 signal was noted in the dorsal hippocampus, amygdala and piriform cortex at 24 hours in six of eight rat pups and at 8 days in seven of eight with the hippocampal changes lagging behind the other areas. However, Fluoro-Jade staining failed to reveal evidence of dying neurons in these areas at 1, 2, 4, and 7 days after the hyperthermic seizures.²⁷ Recently, the same group has shown that their model of P10 prolonged hyperthermic seizures results in about a third of the rats developing spontaneous limbic seizures by 3 months or more of age.²⁶ Therefore, in these 10-day-old infant rats, transient acute postictal T2 hyperintensity did not correlate with cell death or persistent MRI abnormality.

Human Magnetic Resonance Imaging Studies

There are only a few MRI studies of human infants performed soon after an episode of FSE, and these have emphasized abnormalities in the hippocampus. Vanlandingham et al.⁷⁷ described 27 infants with FSE, the majority of whom were imaged within 4 days of the episode of FSE. Of 15 with lateralized seizures, six had definite hippocampal MRI abnormalities. Presumed chronic hippocampal injury with atrophy and increased T2 signal bilaterally were seen in two subjects with histories of perinatal asphyxia and evidence of periventricular leukomalacia. In the other four, presumed acute abnormalities of increased T2 signal and increased volume were seen on the side of presumed seizure origin. Of note in the lateralized group was the longer average seizure duration of 99 minutes in subjects with swollen, bright hippocampi versus 41 minutes in those without these findings.

There were 12 infants with generalized FSE, and only one had subtle bilateral increase of hippocampal signal and size. The authors concluded that both chronic and acute hippocampal abnormalities could be seen following FSE, and that focal and prolonged FSE was more likely to be associated with these abnormalities. Follow-up MRIs have demonstrated that hippocampi with definite increased size and T2 signal on acute imaging developed atrophy with persistently increased T2 signal months to years later, a picture consistent with MTS.⁴⁵

A more recent study of this type⁶⁸ carefully analyzed hippocampal T2 signal relaxation times and volumes in 21 infants with presumed generalized FSE. Visual assessment of the MRIs was normal in 16, three appeared to have hippocampal asymmetry (of which two were confirmed by volumetry), one had a left temporoparietal arachnoid cyst, and one had poor gray-white differentiation in the left midtemporal gyrus. Unlike the Vanlandingham study,⁷⁷ there were no hippocampi with grossly increased T2 signal intensity on visual inspection. However, the group mean calculated T2 relaxation time was increased compared with controls in those imaged within 2 days of the event but not in those imaged later at 3 to 5 days after FSE. In addition, the mean hippocampal volumes of the FSE group were increased compared with an age-matched control group. Follow-up imaging of 14 of these infants from 4 to 8 months later⁶⁹ showed that there was no longer any difference between mean T2 relaxation time or mean hippocampal volume when compared with the controls. However, the hippocampal volumes of the subjects had decreased from the initial studies, and the hippocampal asymmetry had increased and was outside the 95% of control limits in five of the 14. In one of these five, the asymmetry was visually apparent. The conclusion was that FSE could produce transient hippocampal edema and in some cases perhaps neuronal loss.

Grunewald et al.³⁸ looking at MRI findings in 10 infants imaged within 2 weeks of FSE found abnormal hippocampal volume ratios in nine when compared with volume ratios of controls. Because no abnormalities of hippocampal T2 signal or T2 relaxation time constants were found, the authors concluded that the abnormal right to left volume ratios could not be ascribed to edema. No follow-up scans were reported. However, based on the findings of Scott et al.⁶⁸ that increased T2 signal tended to reverse by several days after FSE, the MRIs done an average of 11 days after FSE in the Grunewald study might have been too late to detect increased T2 relaxation times.

Only gold members can continue reading. Log In or Register to continue