Two significant risks are associated with febrile seizures: recurrent febrile seizures and later epilepsy (Freeman 1980).
Recurrent febrile seizures occur in about 30% to 40% of children, usually within a year of the first seizure (Freeman 1980; Berg and Shinnar 1996; Lee and Ong 2004; Pavlidou et al 2008). Young age of onset and family history of febrile seizures are the strongest predictors of another seizure (Berg et al 1990; Pavlidou et al 2008). Additional predictors of recurrence include duration of illness and temperature at the time of the seizure.
The earlier the age of onset, the greater is the risk of recurrence. Children with a first febrile seizure before 1 year of age have a 50% chance of recurrence, compared with 20% if the first seizure is after 3 years of age (Verity et al 1985b).
A family history of febrile seizures is consistently associated with recurrences (Bethune et al 1993; van Esch et al 1994; Pavlidou et al 2008). However, a family history of afebrile seizures has not consistently demonstrated this relationship (Nelson and Ellenberg 1978; Berg et al 1992; Offringa et al 1992). Compared to simple febrile seizures, complex febrile seizures are not more frequently associated with recurrences (Nelson and Ellenberg 1978; Verity et al 1985a; Berg and Shinnar 1996). However, children with prolonged complex febrile seizures are more likely to have another prolonged seizure, in the event of recurrence (Berg and Shinnar 1996).
A shorter duration of fever before the first seizure and a lower temperature at the time of the first seizure increase the chance of recurrence (Berg et al 1992).
Based on the work of Berg and colleagues (Berg et al 1997), risk factors can be combined to provide a useful prediction scheme. They followed 428 children who presented to an urban emergency room with a first febrile seizure. Over the next 2 years, 32% recurred. The recurrence risk for those with none of the 4 risk factors (age less than 18 months, family history of febrile seizures, low temperature at the time of the seizure, and short duration of illness) was 4%, with 1 factor it rose to 23%, with 2 factors 32%, with 3 factors 62%, and with all 4 factors 76%.
Only 2% to 5% of children with a first febrile convulsion subsequently develop epilepsy (Nelson and Ellenberg 1976; Verity et al 1985b; Annegers et al 1987; Pavlidou and Panteliadis 2013). Risk factors for later epilepsy include (1) an abnormal neurologic or developmental status prior to the first febrile seizure, (2) a family history of afebrile seizures, and (3) a complex febrile seizure. Sixty percent of children with a first febrile seizure have none of these risk factors and a subsequent risk of epilepsy of only 0.9%. About 2% of children with 1 risk factor (34% of children with febrile seizures) and 10% of those with 2 or more risk factors (6% of children with febrile seizures) will develop epilepsy (Nelson and Ellenberg 1978; Annegers et al 1987). Those whose onset of febrile seizures occurs after the age of 5 years do not have an increased risk of epilepsy (Webb et al 1999). In an animal model, female rats who were exposed to febrile seizures as immature animals appeared to have higher risk for seizures in adulthood than males (Dai et al 2014). This observation has not been reported in humans. Therefore, risk factors in the individual child are not useful clinical predictors of epilepsy.
When epilepsy does develop, the seizures can be of virtually any type, although the highest association is with generalized, rather than partial, seizures (Rocca et al 1987; Camfield et al 1994). As noted in the Differential Diagnosis section, the risk of a subsequent afebrile seizure appears to be heightened for children presenting with the “afebrile-febrile seizure” (7.8% vs. 1.6% with a first febrile seizure) (Zerr et al 2005), although those whose first seizure (febrile or afebrile) is associated with gastroenteritis appear to be at particularly low risk for recurrence (hazard ratio 0.28 compared with children with seizures associated with other acute illnesses) (Martin et al 2010; Verrotti et al 2014).
Approximately 15% of children with epilepsy have 1 or more preceding febrile seizures, regardless of the cause of the epilepsy (Camfield et al 1994). A history of febrile seizures was reported in 10% of 200 patients with treatment-resistant epilepsy, compared with 2.5% of 200 age-matched children with well-controlled epilepsy (odds ratio 4.33, 95% confidence interval 1.59 – 11.79) (Tripathi et al 2011). These observations suggest that the tendency for febrile seizures plays an important role in a person’s seizure threshold. However, there is no evidence that 1 or multiple febrile seizures actually cause epilepsy.
There is also no evidence that a short febrile convulsion damages the brain. The National Collaborative Perinatal Project study included 431 sibling pairs discordant for febrile seizures (Ellenberg and Nelson 1978). Psychometric testing at 7 years of age included the Wechsler Intelligence Scale for Children as a measure of overall intelligence and the Wide Range Achievement Test as a measure of academic achievement. For those known to be normal before the first febrile seizure, there was no difference in intelligence or school achievement between sibling pairs, even in the 27 with febrile seizures lasting more than 30 minutes. Chang and colleagues conducted another study utilizing a prospective, population-based, case-control method to assess the learning, spatial, and sequential working memory of 87 school-aged children with a previous febrile seizure and 87 randomly selected age-matched control subjects (Chang et al 2001). The febrile seizure group performed significantly and consistently better than control subjects on mnemonic capacity and had more flexible mental processing abilities than their age-matched controls. A case-control study of children of low socioeconomic status (159 with first febrile seizure, 141 healthy controls) also demonstrated no significant differences in cognition, motor skills, or adaptive behavior within one year of a febrile seizure (Leaffer et al 2013).
Long-term follow-up into adolescence and young adulthood confirms these studies’ encouraging findings. A population-based Danish study of 507 18- to 20-year-old men with histories of febrile seizures, but not of epilepsy, who presented for obligatory evaluation by the military draft board demonstrated no difference in intelligence testing results compared to 17,769 men without history of febrile seizures (Nørgaard et al 2009). Finally, a Finnish study followed a random birth cohort of 900 children and found that academic and social accomplishments measured at 12 and 18 years of age were no different for those who experienced febrile seizures than for those who did not (Sillanpää et al 2011).
Starting with the sentinel work of Murray Falconer, an important connection has been drawn between prolonged febrile seizures, mesial temporal sclerosis, and intractable temporal lobe epilepsy. Studies have shown acute vasogenic edema in the hippocampus following febrile status epilepticus, which resolves over 3 to 5 days (Scott et al 2006), as well as markers of oxidative stress after febrile seizures (Günes et al 2009). The cause and effect relationship has been a source of intense controversy. It has been suggested that “2 hits” are required for this sequence of events–the first an initial injury or malformation of the temporal lobe and the second a vulnerability factor (possibly genetic) unique to the child that allows the febrile seizure to occur. Fortunately, the sequence of a prolonged febrile seizure, mesial temporal sclerosis, and intractable temporal lobe epilepsy is uncommon, occurring in not more than 1 of 75,000 children (Camfield et al 1994). A rat model suggested increased markers of epileptogenesis of febrile seizures in the setting of cortical dysplasia (Park et al 2010). Although in retrospect, patients with intractable epilepsy due to cortical malformations in addition to mesial temporal sclerosis are likely to have experienced febrile seizures in early childhood (Fauser et al 2006), no prospective studies document a normal MRI and then a prolonged febrile seizure followed by unilateral hippocampal swelling, mesial temporal sclerosis, and intractable temporal lobe epilepsy. Among 199 children with febrile status epilepticus, 22 (11.5%) had abnormal hippocampal T2 signal on acute brain MRI, and many had coexisting temporal lobe abnormalities (eg, hippocampal malrotation) (Shinnar et al 2012). After one year, 10 of the 14 children available for follow-up MRI had developed hippocampal sclerosis, whereas only 1 of the 116 subjects with normal hippocampal signal immediately after febrile status epilepticus went on to develop abnormal signal on follow-up MRI (Lewis et al 2014). These children are being followed to evaluate whether these MRI findings are predictors of later development of epilepsy (the FEBSTAT study).