Adverse outcomes following convulsive status epilepticus in children: Relationship with hippocampal injury

Authors

  • Rod C. Scott

    1. UCL Institute of Child Health, Great Ormond Street Hospital NHS Trust, London, United Kingdom; and National Centre for Young People with Epilepsy, Lingfield, United Kingdom
    Search for more papers by this author

Address correspondence to Rod C. Scott, Radiology and Physics Unit, UCL Institute of Child Health, 30 Guilford Street, London WC1N 1EH, U.K. E-mail: rscott@ich.ucl.ac.uk

Summary

Convulsive status epilepticus (CSE) is the most common medical neurologic emergency in childhood. It is associated with significant mortality and morbidity. The estimates for the frequencies of adverse outcomes have a very wide range, but even the lower estimates are clinically important. The major predictor of outcomes following CSE is etiology. The characteristics of the episode of CSE itself, for example, seizure length and type, are relatively poor predictors. Nevertheless, there is a relationship between CSE and hippocampal injury. This relationship is well established in animal models, and there is some prospective evidence from human studies that CSE (particularly prolonged febrile seizure) can cause acute hippocampal abnormalities. Further study is required to establish the clinical relevance of these acute findings and to determine whether they predict later mesial temporal sclerosis associated with temporal lobe epilepsy.

Convulsive status epilepticus (CSE) is the most common neurologic emergency in childhood and has been associated with significant morbidity and mortality. Although adverse outcomes have been attributed to brain injury, particularly to the hippocampus, the role of etiology has not received much attention. There has been a recent systematic review on the frequency and nature of adverse outcomes associated with CSE (Raspall-Chaure et al., 2006). Mortality, subsequent development of epilepsy, and cognitive/behavioral impairments were all evaluated. The most important finding is that these outcomes were best predicted by the etiology and that characteristics of the CSE event (e.g., seizure length and type of seizure) either predicted outcome poorly or not at all. Therefore, the aim of the current review is to critically evaluate the roles of etiology and seizure discharges in the pathogenesis of adverse outcomes and of hippocampal injury.

Etiology of CSE in Children

There have been three important epidemiologic studies evaluating the incidence and etiologies of CSE in childhood: the north London status epilepticus in childhood surveillance study (NLSTEPSS)(Chin et al., 2006), a study from Kilifi in Kenya (Sadarangani et al., 2008), and one carried out in Okahama, Japan (Nishiyama et al., 2007). The incidence of CSE in London is 17–23/100,000 children per year, 108/100,000 children per year in Kilifi, and 38/100,000 children per year in Okahama, reflecting important geographic differences in the nature of CSE (see Fig. 1). In north London, approximately half of children who experience CSE were previously neurologically normal. The most common etiology in these children is a prolonged febrile seizure or a central nervous system infection. The remaining children had CSE in the context of preexisting neurologic disorders including cerebral palsy, progressive neurological disorders, and epilepsy (in 12% of the cohort) (Chin et al., 2006). In Japan there is a higher incidence of prolonged febrile seizures (Nishiyama et al., 2007) and in Kilifi most of the CSE occurs in previously neurologically normal children who have cerebral malaria, meningitis, or encephalitis (Sadarangani et al., 2008). As outcome is dependent upon etiology, it is likely that there will be geographic variations in outcomes, although there are only few outcome studies that have been carried out in developing countries.

Figure 1.


Pie charts showing the distribution of etiologies of childhood convulsive status epilepticus in three geographically disparate population based samples. PFS, prolonged febrile seizure.

Outcomes

The overall mortality associated with CSE in childhood has been estimated to be between 0% and 16% (Raspall-Chaure et al., 2006). The mortality rate in children with preexisting epilepsy or who had a prolonged febrile seizure was 0–2% compared to rates of 12.5–16% in children with acute symptomatic etiologies (Raspall-Chaure et al., 2006). Mortality was also higher in younger children, but this was almost entirely due to the higher incidence of central nervous system infections in this group. There is also an increased risk of cognitive impairment following CSE in children, although the estimates of this risk have a very wide range. Again, etiology is the most important predictor, with >20% of children who have acute symptomatic CSE developing a new cognitive impairment compared to a rate of <10% in other types of CSE (Raspall-Chaure et al., 2006). Therefore, in terms of cognition and behavior, the adverse outcomes are a function of preexisting brain disease or a function of brain injury associated with the etiology rather than with seizure length.

Epilepsy Following CSE in Children

There have been approximately 20 studies that have investigated the relationships between CSE in children and the subsequent development of epilepsy (Raspall-Chaure et al., 2006). Many seizure types and epilepsy syndromes have been reported. Rarely, Lennox-Gastaut syndrome and infantile spasms are recognized and a small proportion of children with early onset prolonged febrile seizures will have Dravet syndrome (Dravet et al., 2005). Focal seizures are the most frequently observed seizure types following CSE, but the structural substrates underlying this have not been systematically evaluated. The structural substrates that underlie the focal epilepsy could be developmental or acquired, with the acquired lesions being a consequence of either the etiology or the CSE itself. The risk of having a second seizure following an initial short seizure is on the order of 35–40% (Berg & Shinnar, 1991; Shinnar et al., 1996), which is very similar to the risk of having a short seizure following and initial unprovoked episode of CSE (Eriksson & Koivikko, 1997; Maytal et al., 1989). However, the risk of developing epilepsy following a symptomatic CSE is >50% (Raspall-Chaure et al., 2006).

Hippocampal Injury Associated with CSE

There is a long-standing hypothesis that CSE (and in particular prolonged febrile seizures) can cause hippocampal injury that matures into mesial temporal sclerosis (MTS) associated with temporal lobe epilepsy. Because MTS is the most common structural abnormality identified in people undergoing epilepsy surgery (Babb & Brown, 1993), it is important to understand the relationships between prolonged febrile seizures and MTS so that novel strategies that have the potential to reduce the incidence of MTS can be devised. There is a wealth of data from animal models that has definitively shown that brain injury can occur as a result of prolonged seizures. Many different experimental paradigms have been used to induce status epilepticus. These include systemic and focal injections of seizure-inducing drugs such as kainic acid (Ben Ari, 1985) and pilocarpine (Cavalheiro, 1995), as well as electrical stimulation models (Lothman et al., 1989). Injury has been observed in many brain areas including the hippocampus, entorhinal cortex, piriform cortex, thalamus, and cerebellum (Meldrum, 1997). Many of the animals with brain injury will subsequently develop epilepsy that resembles human temporal lobe epilepsy.

Therapeutic Strategies

If CSE in humans can cause hippocampal injury then it is important to devise therapeutic strategies that either prevent or minimize that injury. As such therapies will always need to be given after the onset of CSE, we wanted to determine whether there is a window of opportunity for effective administration of neuroprotective or antiepileptogenic agents. We have, therefore, monitored the time course of hippocampal injury from the day of the acute event through to 21 days later using multiparametric quantitative magnetic resonance imaging (MRI). CSE sets in train a process that leads to T2 (see Fig. 2), T1, and apparent diffusion coefficient changes that maximize 48 h after the acute event. The severity of these changes predicts hippocampal volume 21 days after the event. Therefore, there is potentially a 48 h window for therapeutic intervention and it is now important to understand the mechanisms underlying the MRI changes so that novel therapies can be produced.

Figure 2.


T2-Weighted images acquired at 9.4T in a control rat (left image) and a rat exposed to 90 min of pilocarpine-induced status epilepticus 48 h previously. There is increased signal in the hippocampus of the rat exposed to status epilepticus, particularly affecting the CA1 region (shown with the blue arrow). There is also increased signal observed in the piriform cortex (yellow arrow).

Discussion

It is important to translate the wealth of animal data to humans. There are many case reports of MRI changes associated with CSE in humans, but few systematic studies. It is nevertheless becoming increasingly clear that prolonged febrile seizures can cause MRI-identified changes in the hippocampus within a few days of the event (VanLandingham et al., 1998; Scott et al., 2002). Increases in T2 relaxation time and in hippocampal volume have been observed within 2–5 days of a prolonged febrile seizure. These findings are most consistent with hippocampal edema, although it is possible that the findings represent a preexisting hippocampal abnormality that predisposed the individual to having a prolonged febrile seizure (Scott et al., 2002). Follow-up imaging of this cohort of children 6–8 months after the acute event revealed that there had been a reduction of both T2 relaxation time and hippocampal volume (Scott et al., 2003). The combination of the acute and follow-up findings is best explained by the development and subsequent disappearance of hippocampal edema, but could also represent damage and volume loss to a hippocampus that was abnormal prior to the seizure. Either way the data support the view that prolonged febrile seizures can cause an acute hippocampal insult. Further follow-up is required to determine whether any of the children go on to develop MTS. As with the animal data there are now data that support the view that acute hippocampal findings can be used to predict final hippocampal volume in humans (Provenzale et al., 2008).

It is, therefore, clear that prolonged febrile seizures can cause an acute hippocampal insult but it remains uncertain how many children will go on to develop MTS associated with temporal lobe epilepsy. Ongoing studies following the population-based cohort of children with CSE ascertained in north London have the potential to accurately estimate the incidence of MTS following CSE in children.

In conclusion, CSE in children is associated with adverse outcomes that are primarily a function of etiology, although there is evidence that CSE itself can cause brain injury, particularly to the hippocampus. Although there are many studies evaluating the mechanisms of brain injury associated with CSE, there is a dearth of studies evaluating the impact of etiology. If the ultimate goal is to reduce adverse outcomes associated with CSE, then it is essential that studies not only investigate the impact of the epileptic discharge but also investigate how the impact of etiology can be minimized.

Disclosure

The author advises on the licensing of buccal midazolam for the emergency treatment of seizures. If the licence is granted, his employing institution will gain financially, but he will not receive any personal benefit. He has also received travel grants from GlaxoSmithKline, Janssen-Cilag, UCB Pharma, and SPL.

Ancillary