We successfully documented the clinical manifestations of AE in children with TSC. AE developed within 24 h after fever onset, followed by a prolonged coma. MRI revealed brain edema with reduced water diffusion, predominantly in the subcortical white matter. Outcomes were poor (severe neurological sequelae or death in most patients) despite various treatments. We also found that a history of FSE was associated with the development of AE.
It is remarkable that the clinical manifestations of AE were similar among the children. All experienced AE within 1 day of fever onset, seizures lasting for 30 min, a monophasic clinical course with coma, and widespread MRI abnormalities. We presume that excitotoxicity attributable to prolonged seizures is the principal cause of the irreversible brain lesions because all children had a prolonged seizure refractory to antiepileptic drugs. Hypercytokinemia may be also involved in AE pathogenesis, but the laboratory abnormalities of our patients were milder than those of patients with acute necrotizing encephalopathy,18,19 which is considered to be caused by a “cytokine storm” associated with multiorgan failure and disseminated intravascular coagulation. Marked elevations of enzymes such as AST and LD were common in children with acute necrotizing encephalopathy immediately after disease onset.20 Hypercytokinemia may play only a limited role in the development of AE in children with TSC. On the other hand, elevated blood glucose and serum ammonia levels were common in our patients, suggesting metabolic derangement; this may be a sequela of critical illness caused by AE.
MRI revealed widespread abnormalities in all children. Reduced water diffusion (indicating cytotoxic edema) was evident, predominantly in the subcortical white matter, and conventional MRI suggested edema in the cerebral cortex. A similar MRI pattern is seen in children with Dravet syndrome complicated by AE.15,21 Okumura et al. reported 15 such children and showed that brain edema and reduced water diffusion in the cortical and/or subcortical white matter were characteristic of the condition.15 Notably, a prolonged seizure is an initial symptom of AE in children with Dravet syndrome. Widespread MRI abnormalities with cytotoxic edema may be neuroimaging features of AE in children with TSC.
Initial laboratory abnormalities and CSF analysis abnormalities were mild (or absent) in our children. This implies that the brain disorders of children with AE exhibited sudden onset and rapid progression. All children presented with a seizure induced by fever; distinguishing AE from less severe seizures is clinically difficult on initial presentation. Laboratory data may not be helpful; no marked abnormalities are present. However, hyperglycemia was common in AE children on presentation. Hyperglycemia is correlated with adverse outcomes of status epilepticus and AE22-24 and may be a convenient predictor of AE.
The outcomes of children with TSC complicated by AE were poor, although intensive treatment was performed. Treatments included supportive efforts to stabilize the general condition, seizure control, and neuroprotection. Although all patients required intensive care and artificial ventilation, their general condition was appropriately maintained. No patient developed shock, serious multiorgan failure, or disseminated intravascular coagulation. Seizure control was achieved in all patients after the aggressive use of antiepileptic drugs. A recent consensus treatment for status epilepticus refers to prompt recognition and the need for very early treatment to reduce morbidity and mortality, drug requirements, and seizure duration.25,26 Studies employing buccal or intranasal midazolam found that delivery via non-intravenous routes was a practical, rapid, reasonably safe, and effective alternative to intravenous lorazepam or diazepam as a first-line treatment for early status epilepticus in out-of-hospital settings.27,28 No such rescue drugs (example: buccal midazolam) are yet available in Japan. Neuroprotective treatment will be a subject of a future study. Several pharmacological and non-pharmacological treatments including intravenous immunoglobulin, corticosteroids, neuroactive steroids, and hypothermia have been used to treat patients who presented with status epilepticus,29-32 but neither efficacy nor tolerability has been investigated.
Our index case died of AE. Shepherd et al. explored the causes of death of TSC patients and found that 9 of 40 TSC patients who died had status epilepticus.11 The age at death ranged from infancy to adulthood. Shehata et al. reported that 2 of 21 patients with TSC complicated by status epilepticus died.12 These reports did not give detailed clinical and genetic information, and it is uncertain whether the dead patients met our criteria for AE. Welin et al., who used national registry data to estimate the prevalence of epilepsy and mortality associated with TSC in Sweden.33 The causes of death were directly related to TSC in 15 of 30 patients who died, including 3 who died of epilepsy. No additional information was provided. Amin et al. reported that renal disease was a major cause of mortality in TSC patients and for sudden unexpected death from epilepsy.34 No information on status epilepticus was given. Although the frequencies of AE may be low, more attention should be paid to AE to improve the long-term outcomes of patients with TSC.
We found that a history of FSE was a risk factor for AE in children with TSC. Nearly half of children with AE had experienced FSE before AE onset. Little attention has been paid to the relationship between TSC and FS. No study has adequately investigated the rate or clinical manifestations of FS in children with TSC. Notably, a history of FS in our study was more frequent (16%) in children with TSC but without AE than in the general population (3–8% in Japan). This suggests several different scenarios. One possible explanation is that children with TSC may be intrinsically susceptible to FS. However, no data support this hypothesis. Experimental and/or epidemiological studies are required. Another possibility is that mTOR pathway plays a role in FSE. The association between mutations in mTOR pathway genes and epileptic network has reported, and studies in rodent models of status epilepticus demonstrate that mTOR signaling is activated by status epilepticus.35 However, this biological hypothesis is unclear because there have been no studies on the relation between mTOR pathway and fever. Another possibility is that genes other than TSC1/TSC2 may contribute to AE development. Mutations in SCN1A and PCDH19 are well known to cause several types of epilepsy that are associated with FS.36,37 Mutations in SCN1B, SCN2A, SCN9A, GABRG2, CACNA1H, and STX1B have been found in families exhibiting genetic epilepsy with FS.38 It is possible that some genetic variants may modify the phenotypes of TSC, increasing susceptibility to FS. It is also possible that initial FSE may precipitate FSE recurrence, increasing the risk of AE in children with TSC. Maytal et al. reported that development of FSE in an otherwise normal child did not increase the risk of subsequent FS during the first few years following the initial episode.39 By contrast, the FEBSTAT study revealed that the risk of subsequent FSE was significantly increased in those with an initial FSE compared to a simple FS and that any MRI abnormality increased the risk 3.4 fold.40 These results may support the hypothesis that FSE occurrence may increase the risk of later FSE /AE in children with TSC.
Our study has several limitations. The selection of control children with TSC may have affected the results. A distinct feature of TSC is that disorders of various organs appear at different ages. The clinical manifestations of TSC develop with age, and the extent of each symptom or complication changes constantly. The severity of clinical manifestations varies widely, even in a single patient, according to age. A neonate with TSC may have no epileptic seizures but may have seizures in the future. Similarly, a young infant with no history of AE may develop AE in the future. Therefore, we believe that the clinical variables should be compared at specific ages. We found that the age at the onset of AE in most cases was 4 years of age or younger. Thus, we excluded children with TSC under 4 years of age from the control group and compared clinical variables that were recognized by 4 years of age; the appropriateness of such exclusion may be controversial. The time at which clinical information was collected may affect our results. We could not perform genetic analysis of all children. It is possible that the risk of AE may be correlated with the type of TSC1/TSC2 mutation. Genetic analysis would yield useful information on AE development in children with TSC. Finally, this was a retrospective study with a small number of patients. The results of this study should be validated by prospective studies with more sophisticated designs.