Febrile Seizures: Clinical Insights From Decades of the National Institutes of Neurological Diseases and Stroke (NINDS) Support

In honor of the continued support of epilepsy-related research by the National Institutes of Neurological Diseases and Stroke (NINDS) over several decades, this invited review discusses some selected topics related to febrile seizures (FS). FS, identified as a unique seizure type more than a century ago, are the most common seizures of childhood, affecting up to 5% of children in the age range of 6 months to 5 years.^1,2 Over time, many unique aspects of FS have been recognized and described. These include the existence of 2 broad groups of FS: simple and complex. Simple FS are generalized from onset, brief (<15 min), and occur once during a febrile illness. Approximately 20%–30% of FS are termed complex, defined as having a focal onset, more than 1 FS occurring within a febrile illness, or if the seizure is prolonged, lasting >15 min. ³ Febrile status epilepticus (FSE) comprises a subset of complex FS (∼5% of all FS) when seizure duration exceeds 30 min. It is presumed that most FS are benign and incur no lasting neurological or cognitive sequelae, but complex FS carry a higher risk for the later development of epilepsy and long-term cognitive impairments, especially following FSE. In particular, FSE portends a higher risk of hippocampal sclerosis and temporal lobe epilepsy (TLE; see below).

Despite the accumulated understanding of many features of FS, some aspects remain controversial, especially regarding prognosis and treatment. Most FS are outgrown and do not lead to later epilepsy. Therefore, it is generally felt that prophylactic treatment is unnecessary and does not affect prognosis. Yet not all FS are benign, and many questions remain about FS effects on subsequent neurological function. This brief review cannot cover all of the past and current controversies and issues related to FS, but 4 key studies, all supported by NINDS, were selected as representative of the knowledge that can be gained with sufficient scientific and financial support. These studies have transformed our understanding of FS and have contributed to patient care. Two of these articles harken back to the 1970s, related to the large NINDS-sponsored Collaborative Perinatal Project (CPP), and 2 more recent papers derive from the multicenter Consequences of Prolonged FS in Childhood (FEBSTAT) study of FSE.

Predictors of Epilepsy and Intellectual Outcome in Children Following FS

The CPP followed about 54,000 children born between 1959 and 1966; 1706 (3%) children experienced at least 1 FS, all of whom had detailed histories and neurologic examinations until age 7 years. Several important pieces of information were revealed by the first study, ⁴ which provided epidemiological data that has stood the test of time. ² Two percent of the FS subset developed unprovoked afebrile seizures by age 7 years, meeting the criteria for epilepsy. About 1% of children who were previously normal and had a noncomplex first FS developed epilepsy. Children with prior neurologic or developmental abnormalities and complex features at the initial FS had an 18 times higher risk for epilepsy than those without FS. By defining the risk factors and population characteristics of this cohort, critical baseline information about FS epidemiology was obtained.

The same authors analyzed the CPP data to determine intellectual outcome at age 7 years using a sibling comparison design. ⁵ Full-scale intelligence quotient (IQ) was assessed using Wechsler intelligence scales, and academic achievement was evaluated using the Wide Range Achievement Test. The data led to the conclusion that FS were not associated with a decrement in IQ or early academic performance compared to siblings without FS, suggesting that FS did not invoke major cognitive sequelae. These conclusions have largely held up over time. However, recent studies employing more sophisticated tests and measures ⁶ are now casting some doubt as to whether all FS, even simple ones, ⁷ are truly benign. Language-based tests can be inaccurate at the young ages at which children are often tested, and deficits might emerge later. Genetic, developmental, and environmental factors may affect outcomes, and it is well recognized that children with neurodevelopmental impairments preceding FS are at higher risk for cognitive deficits. Animal models of FS are being used to assess aspects of cognition that are difficult to assess in children. ⁸ While animal models have their own caveats, they allow investigation of mechanisms by which the maturing brain develops hyperexcitability and seizures in response to hyperthermia, inflammation, epigenetic changes, and a host of other mechanisms. ⁹

FSE: Viral Etiology and Hippocampal Sclerosis

The etiology of FS has remained unclear. A familial predisposition to FS is well accepted, but a longstanding question is whether the fever itself or fever-induced inflammatory changes, such as cytokine release, actually trigger the seizure. Furthermore, considering that a viral illness is the typical FS trigger, it is unclear whether a particular virus is more likely to initiate the fever leading to FS or if any viral illness is sufficient.

A notable subgroup of children with FS (about 5%) develops prolonged seizures (FSE); children with FSE have an increased risk of hippocampal injury and subsequent TLE. The FEBSTAT study of 199 children with FSE found that human herpesvirus-6B (and to a lesser extent human herpesvirus-7) together accounted for one-third of FSE, ¹⁰ but it could not be determined whether viral neuroinvasion or viremia was sufficient to precipitate the FSE.

Of the numerous studies resulting from FEBSTAT among the most important findings are those derived from neuroimaging. ¹¹ Brain magnetic resonance imaging (MRI) scans performed within several days after FSE demonstrated acute hippocampal T2-hyperintensities in about 10% of the study group. Follow-up MRI scans were obtained at 1, 5, and 10 years following FSE, representing a remarkable longitudinal span, and demonstrated a high incidence of hippocampal sclerosis in about half of that cohort, but not in children whose initial MRI scans did not show T2 hyperintensities. Furthermore, 6 of the 10 children with hippocampal sclerosis at age 10 had developed mesial TLE. The major conclusion of this study is that visual assessment of T2 signal in the post-ictal MRI scan can predict children at higher risk for hippocampal sclerosis and mesial TLE following FSE at long follow-up times. The lack of initial T2 changes correlated highly with the absence of hippocampal sclerosis or the emergence of TLE. In addition to neuroimaging, electroencephalograms (EEGs) were obtained in the early post-FSE period and at later times; temporal area focal slowing or voltage attenuation was found in about 30% of children, ¹² correlating with the acute MRI findings, thus complementing the ability to predict TLE.

Conclusions

NINDS support has contributed substantially to our present knowledge about the epidemiology, outcomes, and neurobiology of FS, including the important subgroup of FSE. The 4 articles briefly highlighted above were large clinical studies, thus labor- and cost-intensive, but yielded critical information about FS. Study designs have allowed utilization of wide-ranging techniques, including epidemiology, neuropsychological assessment, viral analysis, and neuroimaging, while other approaches have examined detailed genetic contributions and EEG characteristics to better understand their multiple manifestations and concomitants. In addition, other research, much of it also supported by NINDS, has evaluated the role of inflammation in FS occurrence and its consequences, the neurobiology of neuronal susceptibility to fever or cytokine/inflammatory triggers, and issues pertinent to treatment considerations.