- Email: [email protected]
Epilepsy After Resolution of Presumed Childhood Encephalitis
Pediatric Neurology 53 (2015) 65e72
Contents lists available at ScienceDirect
Pediatric Neurology journal homepage: www.elsevier.com/locate/pnu
Original Article
Epilepsy After Resolution of Presumed Childhood Encephalitis Neggy Rismanchi MD, PhD a, Jeffrey J. Gold MD, PhD a, Shifteh Sattar MD a, Carol A. Glaser DVM, MD b, Heather Sheriff BA b, James Proudfoot MS c, Andrew Mower MD d, John R. Crawford MD, MS a, Mark Nespeca MD a, Sonya G. Wang MD a, * a
Department of Neurosciences, University of California, San Diego, California California Department of Public Health, Richmond, California c Biostatistics Unit, Clinical and Translational Research Institute, University of California, San Diego, California d Children’s Neurology Center, Children’s Hospital of Orange County, Orange, California b
abstract OBJECTIVE: To evaluate factors associated with the development of epilepsy after resolution of presumed childhood encephalitis. METHODS: A total of 217 patients with suspected encephalitis who met criteria for the California Encephalitis Project were identified. Evaluable outcome information was available for 99 patients (40 girls, 59 boys, ages 2 months to 17 years) without preexisting neurological conditions, including prior seizures or abnormal brain magnetic resonance imaging scans. We identified factors correlated with the development of epilepsy after resolution of the acute illness. RESULTS: Development of epilepsy was correlated with the initial presenting sign of seizure (P < 0.001). With each additional antiepileptic drug used to control seizures, the odds ratio of developing epilepsy was increased twofold (P < 0.001). An abnormal electroencephalograph (P < 0.05) and longer hospital duration (median of 8 versus 21 days) also correlated with development of epilepsy (P < 0.01). The need for medically induced coma was associated with epilepsy (P < 0.001). Seizures in those patients were particularly refractory, often requiring longer than 24 hours to obtain seizure control. Individuals who required antiepileptic drugs at discharge (P < 0.001) or were readmitted after their acute illness (P < 0.001) were more likely to develop epilepsy. Of our patients who were able to wean antiepileptic drugs after being started during hospitalization, 42% were successfully tapered off within 6 months. CONCLUSIONS: Limited data are available on the risk of developing epilepsy after childhood encephalitis. This is the first study that not only identifies risk factors for the development of epilepsy, but also provides data regarding the success rate of discontinuing antiepileptic medication after resolution of encephalitis. Keywords: Epilepsy, childhood, encephalitis, seizures
Pediatr Neurol 2015; 53: 65-72 Ó 2015 Elsevier Inc. All rights reserved.
Introduction
Encephalitis is potentially devastating, particularly in children. The annual incidence worldwide ranges from 3.5 to 7.4 per 100,000 and increases to 16 per 100,000 in
Article History: Received February 4, 2015; Accepted in final form March 14, 2015 * Communications should be addressed to: Dr. Wang; Department of Neurosciences; UC San Diego; 3020 Children’s Way, MC 5009; San Diego, CA 92123. E-mail address: [email protected] 0887-8994/$ e see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.pediatrneurol.2015.03.016
children.1 Presenting signs and symptoms can be nonspecific and subtle or quite dramatic. Some children present only with low-grade fever, headache, or nausea, whereas others present with profound encephalopathy, vomiting, seizures, and/or focal neurological deficits.2-5 Those children who present with subtle signs can also unexpectedly worsen and later develop seizures and other severe neurological sequelae.5 Encephalitis can result in greater morbidity and mortality in children than in adults. In one study of 1570 patients in California, 62 had intractable seizures during the acute illness, with children comprising 69% of these
66
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
patients.3 Another study of 148 patients in India demonstrated that 61% of children had seizures during encephalitis in comparison to 37% of the adults.6 Seizures reportedly are more likely to occur with specific viral infections, with herpes virus (HSV) (75% of patients with HSV) more commonly associated with seizures than Japanese encephalitis (54%), followed by nonspecific encephalitis (43%).6 The postinfectious long-term complications of encephalitis include dystonia, spasticity, epilepsy, and neurocognitive problems.4 In fact, despite treatment with acyclovir in cases of herpes encephalitis, two-thirds of patients experienced significant neurological impairment.7 Unfortunately, literature on long-term outcome and prognosis following encephalitis is limited, but studies on HSV encephalitis quote long-term morbidity rates as 30% in adults and up to 67% in children, with seizures occurring in 44% of these patients in one prospective study in Toronto, Canada.8,9 In the general population, seizures are a common neurological problem, with up to 10% of the population having a seizure in their lifetime and up to 3% of the adult population developing epilepsy.10,11 Central nervous system (CNS) injuries such as stroke, viral infection, trauma, febrile seizure, and status epilepticus are considered common risk factors for developing epilepsy.12,13 A group in Texas reported the 20-year risk of developing an unprovoked seizure after encephalitis or meningitis to be 6.8%, noting that patients with a prior CNS infection had a sevenfold increase in unprovoked seizures compared with the general population.14 CNS injuries are reported to account for 30%49% of all unprovoked seizures and epilepsies. Long-term follow up studies conducted in Seattle of individuals with a history of brain injury reveal that after a single late seizure, the risk for recurrent seizure is typically greater than 80%.15 The risk of epilepsy was increased by 16 times with viral encephalitis when compared to the general population and this risk remained elevated up to 15 years postinfection.12 A correlation between encephalitis and later development of epilepsy has been shown; however, these studies were often limited by follow-up data, etiology of encephalitis, low patient number, or lack of assessment of other potential contributing factors that resulted in epilepsy. To identify factors potentially contributing to the development of epilepsy, we studied a large cohort of patients longitudinally at a single institution who were enrolled in the California Encephalitis Project (CEP). Methods A cohort of 217 patients who met diagnostic criteria for inclusion in the CEP at Rady Children’s Hospital San Diego between 2004 and 2011 were studied.16 To be included in the CEP, children had to be hospitalized with encephalopathy lasting 24 hours and meet at least one of the criteria: fever, seizure, cerebrospinal fluid pleocytosis, focal neurological signs, or with neuroimaging or electroencephalograph (EEG) evidence of encephalitis.3 Details of the CEP and the standardized testing associated with CEP have previously been described.2,3,16 This retrospective study was granted approval under the multiinstitutional protocol established by the State of California for the CEP and a waiver of consent was granted. Patients were excluded from our study if they had prior neurological abnormalities, tumor, or were found to have other medical conditions causing these similar encephalitic signs and symptoms. Patients who were lost to follow-up or died during hospitalization were also excluded, resulting in
99 patients remaining in the study. Fig 1 summarizes these excluded patients. In this study, factors from initial presentation, hospital course, and discharge that might predispose a patient to epilepsy were assessed. The electronic medical record was reviewed including emergency room encounters, hospital admission course, medication administrations, all prior and subsequent clinic visits, laboratory results, neuroimaging studies, and EEG results.16 Details of neuroimaging and EEG acquisition are described in a prior publication.16 In this study, the term “epilepsy” was defined as any patient with at least one unprovoked seizure after resolution of the patient’s acute CNS insult. This is now an internationally accepted alternative definition of epilepsy in comparison to the traditional definition of requiring two unprovoked seizures.17,18 For our assessment, “difficult to control seizures” was defined as patients who required medically induced coma and more than three antiepileptic drugs (AEDs) for seizure control. Continuous variables were reported as medians and interquartile ranges, and categorical data as counts and percentages. The Fisher exact test was used to conduct group comparisons between categorical variables (development of epilepsy, discharge on AEDs, etc.), and the Mann-Whitney U test to compare continuous variables (hospital duration). A logistic regression was used to assess the impact of the number of AEDs given with the odds of developing epilepsy. P-values less than 0.05 were considered a statistically significant result. All data analyses are conducted using the R statistical programming language.19
Results
Within our cohort of 217 patients, follow-up information was available for a total of 99 patients who had met the inclusion criteria (Fig 1). Our study cohort comprised 40 females and 59 males, with a mean age at presentation of 9 years old (range of 2 months to 17 years). Of the 99 patients in this study, 24 later developed epilepsy. Presenting features associated with developing epilepsy
According to our demographic analysis, race seemed to be correlated with later development of epilepsy, although this had borderline statistical significance (Table 1, P ¼ 0.049). This relationship seems to be largely driven by a much higher proportion of Caucasian subjects in the nonepileptic group (n ¼ 18, 24%) compared with the proportion of Caucasian subjects in the epileptic group (n ¼ 1, 4.2%). Because of this discrepancy between the number of individuals in each of the Caucasian groups, race has a statically significant correlation with development with epilepsy, but these data suggest that being Caucasian decreases likelihood of developing epilepsy rather than a particular race increasing likelihood of epilepsy. A seizure at presentation showed a positive correlation with the later development of epilepsy (P < 0.001). There was not a significant relationship between presenting symptoms of fever, headache, or altered mental status and the subsequent development of epilepsy in this study. Elements of hospital stay associated with developing epilepsy
One major factor found to be important in the development of epilepsy was the number of AEDs used to obtain seizure control. These AEDs included maintenance and abortive seizure medications given en route to the hospital, in the emergency room, or during the hospitalization as well as any continuous infusions needed to control seizures
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
67
FIGURE 1. Study characteristics. There were 217 patients who were enrolled in California Encephalitis Project (CEP). Ninety-six patients were excluded because of a preexisting/concurrent neurological problem, tumor, acute process found to be something other than encephalitis, or no evidence of encephalitis upon closer examination of patient’s data. Of the 121 patients that remained in the study, 19 patients were lost to follow-up and 3 died during the hospitalization, so no outcome measures were available to assess, resulting in a total of 99 patients that remained in our study. CNS, central nervous system; HA, headache; psych, psychiatric. (The color version of this figure is available in the online edition.)
during the hospitalization. The number of patients who developed epilepsy per category of number of AEDs used was assessed. For no drug (n ¼ 38), there were no patients who developed epilepsy; for one AED (n ¼ 7), there was one patient; for two AEDs (n ¼ 13), there were two patients; for three AEDs (n ¼ 14), there were three patients; for four AEDs (n ¼ 6), there were four patients; for five AEDs (n ¼ 10), there were six patients; and for five or more AEDs (n ¼ 11), there were eight patients who developed epilepsy. An increase in the number of AEDs administered correlated with an increase in the percentage of patients who later developed epilepsy (Fig 2A). A logistic regression model was used for the development of epilepsy with the number of AEDs as a predictor, which identified the odds of later developing epilepsy to be increased by a factor of 2.2 with each addition of an AED (95% confidence interval 1.6-3.1; P < 0.001). Length of hospital stay was also found to be a significant factor in the development of epilepsy. The total length of hospital stay (in days) was compared between the individuals who did and did not develop epilepsy. Using the Mann-Whitney U-test to account for rare extreme values in
both groups, this study shows that the median hospital duration of those who developed epilepsy was 21 days in comparison to 8 days in the nonepilepsy group (P ¼ 0.006; Fig 2B). Other elements during hospitalization that were analyzed are summarized in Table 2A. These factors included whether antibiotics were administered, the presence of cerebrospinal fluid pleocytosis (defined as white blood cell count 10), abnormal head CT, abnormal brain magnetic resonance imaging, abnormal EEG, need for intensive care unit stay during admission, or need for medically induced coma for seizure control. An abnormal EEG during hospital stay (P ¼ 0.044) and the need to use medically induced coma (P < 0.001) were both found to have a significant relationship with the development of epilepsy. Of the eight children who died in our series, four had been placed in medically induced comas. All four of these patients did not obtain seizure control until more than 24 hours after seizure onset. To analyze acute hospital management and the sequelae of epilepsy, the early success of aggressive seizure control therapies was evaluated. Successful
68
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
TABLE 1. Demographics and Presenting Symptoms
Total (n ¼ 99)
Epilepsy No (n ¼ 75) Age at onset Sex Female Male Race Hispanic Caucasian Asian African American Unknown/other Fever No Yes Headache No Yes Altered mental status No Yes Seizure No Yes
9 (4, 12)
P
Yes (n ¼ 24) 6 (3, 10.25)
8 (3, 12)
28 (37.3%) 47 (62.7%)
12 (50%) 12 (50%)
40 (40.4%) 59 (59.6%)
36 18 6 3 12
12 1 5 3 3
48 19 11 6 15
0.306 0.341
0.049* (48%) (24%) (8%) (4%) (16%)
(50%) (4.2%) (20.8%) (12.5%) (12.5%)
(48.5%) (19.2%) (11.1%) (6.1%) (15.2%) 0.099
26 (34.7%) 49 (65.3%)
13 (54.2%) 11 (45.8%)
39 (39.4%) 60 (60.6%)
32 (42.7%) 43 (57.3%)
11 (47.8%) 12 (52.2%)
43 (43.9%) 55 (56.1%)
27 (36%) 48 (64%)
5 (20.8%) 19 (79.2%)
32 (32.3%) 67 (67.7%)
50 (66.7%) 25 (33.3%)
6 (25%) 18 (75%)
56 (56.6%) 43 (43.4%)
0.811
0.214
<0.001***
Significance: ***0.001, **0.01, *0.05.
aggressive seizure control was defined as seizure control within 24 hours or less. Statistical analysis did meet significance for mortality; however, all four patients who died during medically induced coma seized for more than 24 hours and this prompted further analysis (Fig 3 and Table 3). Within our medically induced coma group (n ¼ 18), the median hospital duration of the patients with early successful aggressive seizure control (n ¼ 5) was 13 days. In comparison, the patients whose seizures lasted longer than 24 hours and were eventually placed in medically induced coma (n ¼ 13) had a median hospital duration of 26 days (P ¼ 0.103). Two of the patients in the latter group died, likely shortening the mean hospital duration significantly. Last, we looked at difficult to control seizures, defined as requiring medically induced coma plus three or more AEDs. Although three of these eight patients died because of difficulty obtaining seizure control, this was found to be trending toward statistical significance (P ¼ 0.077). Factors after discharge associated with developing epilepsy
Patients who required discharge with an AED or who required readmission for concern of sequelae from encephalitis were also assessed (Table 2B). The patients evaluated for readmission included acute illness related to the initial admission within a 2-month period or concern of seizure within a year after discharge. This study shows that patients who needed to be discharged on an AED (P < 0.001) or having to be readmitted (P < 0.001) were associated with the later development of epilepsy. In total, 27 patients required readmission after discharge. Nineteen of the 27 patients were readmitted within 2 weeks. Twelve of the 27 were readmitted specifically for seizure or possible seizure.
Finally, of the 42 patients who started AEDs during hospitalization and the four others that were started shortly after hospitalization, a total of 24 patients were able to be tapered off of AEDs and remained seizure-free during the time course of our study. Assessment of total time on the AED after discharge was conducted by categorizing them as 0-3 months (n ¼ 4), 4-6 months (n ¼ 6), 7-12 months (n ¼ 2), 1-3 years (n ¼ 3), or 3þ years (n ¼ 9). The percent of total patients tapered off in each group is illustrated in Fig 4. This study shows that about 42% of the patients whose AEDs were successfully tapered were able to do so by 6 months, with the remainder continuing an AED for 7 months or longer. Discussion
Limited information is available regarding the outcomes of childhood encephalitis, particularly in evaluating the factors involved in the development of epilepsy. Although race was found to have a marginally significant relationship with the development of epilepsy in this study, we advise caution in interpreting this result because of the small sample size of Asian and African Americans and the presence of an unknown/other category. Additionally, the findings are largely driven by the disproportionate number of Caucasians in the no epilepsy group rather than a high percentage of any particular race in the epilepsy group. Seizure at the onset of the illness was directly associated with the development of epilepsy. Seizures have been previously well-described in association with encephalitis and, although this sign is not as common as fever or headache, it may be more specifically related to a cerebral insult that would result in abnormal neural reorganization, blood-brain barrier breakdown, glial activation, neuronal hyperexcitability, or other
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
FIGURE 2. Factors associated with epilepsy during hospitalization. The number of antiepileptic drugs (AEDs) used to control seizures correlated with development of epilepsy (A). Each of the patients included in the study (n ¼ 99) was assessed for number of AEDs given during the patient’s acute encephalitis (this includes abortive medications and continuous infusions). The percentage of patients who then developed epilepsy in each of these groups from 0 to 5þ AEDs were calculated. For no drug (n ¼ 38), there were no patients who developed epilepsy; for 1 AED (n ¼ 7), there was 1 patient; for 2 AEDs (n ¼ 13), there were 2 patients; for 3 AEDs (n ¼ 14), there were 3 patients; for 4 AEDs (n ¼ 6), there were 4 patients; for 5 AEDs (n ¼ 10), there were 6 patients; and for more than 5 AEDs (n ¼ 11), there were 8 patients who developed epilepsy. Using a logistic regression for the development of epilepsy with the number of AEDs resulted in an odds ratio (OR) of 2.2 with a 95% confidence interval (CI) of 1.6-3.1 (P < 0.001). Hospital duration correlated with development of epilepsy (B). The figure is a graphical representation of the median number of hospital days when comparing the outcome groups of epilepsy versus no epilepsy. The black lines within the bars depict the median hospital durations (8 days for “no epilepsy” and 21 days for “yes epilepsy”); the colored bars depict the interquartile range for each group (11 days for “no epilepsy” and 27.5 days for “yes epilepsy”). The black circles above the bars are outliers because they are more than 1.5 times the interquartile range above the third quartile. A longer hospital duration was found to be correlated with later development of epilepsy (P ¼ 0.006). (The color version of this figure is available in the online edition.)
alterations suspected to result in epileptogenesis.3,5e7,14,20 Human studies reveal that certain viral causes of encephalitis are more likely to initially present with seizure
69
and that seizures in the context of encephalitis tend to be more frequent and difficult to control, which may lend itself to the later development of epilepsy.6,21 In fact, animal models of epilepsy use this relation by inducing repeated seizures with direct electrical stimulation or introduction of an infectious agent that results in an immune response in the brain to study epilepsy.22,23 Investigators have demonstrated spontaneous seizures in murine models of encephalitis and that lasting hyperexcitability triggers further inflammatory processes in the brain that leads to epileptogenesis.13,24,25 Difficulty attaining seizure control predicted a higher risk of epilepsy. This was evaluated by assessing the total number of AEDs used in treatment as well as the need for medically induced coma for seizure control. This study shows that each increase in number of AEDs required to control seizures during the acute illness resulted in a twofold increased likelihood of developing epilepsy. Several animal models and human studies evaluated time to seizure control and the correlation with future epilepsy, particularly focusing on the efficacy of certain AEDs (including diazepam, levetiracetam, phenytoin, phenobarbital, valproic acid, and topiramate).7,20,25-29 Unfortunately, the results of these studies are variable and fail to suggest a clear path to approach acute seizure control after a cerebral insult with the goal of achieving the best long-term neurological outcome. Shortening the duration of status epilepticus modifies the epileptogenic process because it is known that 43% of patients who are in status epilepticus later develop epilepsy.30-32 Our attempts to further analyze the type and sequence of AEDs used in correlation with outcomes were unsuccessful because of varied treatment paths between patients. Other factors during the hospital course proved to be very revealing in providing correlations with the development of epilepsy. Increased length of hospital stay was correlated with the development of epilepsy; however, our analysis was unable to consider comorbidities that arose during hospitalization, which may have contributed to a longer hospital stay. Although our results were similar to those in a recently published, smaller-scale study, it is possible that the length of hospital stay acted as a surrogate for disease severity even when looking specifically at epilepsy as an outcome measure.33 An abnormal EEG during the hospital course (including slowing, spikes, and sharp waves) significantly correlated with the risk of subsequent epilepsy. However, 20% of patients with an abnormal EEG did not subsequently develop epilepsy. Children with normalization of EEG after discharge were unlikely to later develop epilepsy.34 Such children were also more likely to be tapered off AED medication because this is a common current practice pattern. In this study, though not statistically significant, patients who required medically induced coma for seizure control had a higher association with the development of epilepsy. Further examination of our medically induced coma cohort actually revealed that the patients who achieved seizure freedom within 24 hours had better outcomes in terms of death and epilepsy as well as hospital duration; however, our sample size was not sufficient to yield significance to these findings. Patients who did not achieve seizure control within 24 hours may be more ill
70
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
TABLE 2. Factors during and after Hospitalization Associated with Epilepsy
A.
Antibiotic use No Yes Cerebrospinal fluid pleocytosis No Yes Abnormal computed tomography scan No Yes Abnormal magnetic resonance imaging No Yes Abnormal electroencephalograph No Yes Intensive care unit stay No Yes Medically induced coma No Yes B. Discharge on antiepileptic drugs No Yes Readmitted No Yes
Total (n ¼ 99)
Epilepsy No (n ¼ 75)
Yes (n ¼ 24)
22 (30.1%) 51 (69.9%)
6 (27.3%) 16 (72.7%)
28 (29.5%) 67 (70.5%)
11 (14.7%) 64 (85.3%)
8 (33.3%) 16 (66.7%)
19 (19.2%) 80 (80.8%)
51 (76.1%) 16 (23.9%)
13 (59.1%) 9 (40.9%)
64 (71.9%) 25 (28.1%)
23 (33.3%) 46 (66.7%)
7 (30.4%) 16 (69.6%)
30 (32.6%) 62 (67.4%)
8 (17.8%) 37 (82.2%)
0 (0.0%) 24 (100%)
8 (11.6%) 61 (88.4%)
37 (49.3%) 38 (50.7%)
8 (33.3%) 16 (66.7%)
45 (45.5%) 54 (54.5%)
69 (92%) 6 (8%)
14 (58.3%) 10 (41.7%)
83 (83.8%) 16 (16.2%)
P
0.999
0.071
0.171
0.999
0.044*
0.239
<0.001***
Total (n ¼ 99)
Epilepsy No (n ¼ 75)
Yes (n ¼ 24)
51 (68.9%) 23 (31.1%)
1 (4.3%) 22 (95.7%)
52 (53.6%) 45 (46.4%)
60 (81.1%) 14 (18.9%)
10 (43.5%) 13 (56.5%)
70 (72.2%) 27 (27.8%)
P <0.001***
<0.001***
Significance codes: ***0.001, **0.01, *0.05.
independent of when the medically induced coma was initiated and would be expected to ultimately have poorer outcomes regardless of the interventions. It is also possible that, similar to drug-resistant epilepsy in the outpatient setting after adequate trials of two AEDs, there is a common pathway of epileptogenesis that must be considered. Furthermore, in the setting of acute encephalitis, the severity of neuronal injury is even more substantial. Although the use of medically induced coma is not without risk, there have been studies suggesting that the benefits of early seizure control by medically induced coma (less than 24 hours) outweigh the risks. Some of these studies also have limited sample sizes.35-37 Given the results of these studies as well as our own, early medically induced coma may be beneficial, but clinical trials are necessary to validate this hypothesis. Being discharged with antiepileptic therapy or readmitted after the acute illness were also correlated with the development of epilepsy. Although not all patients discharged on an AED developed epilepsy, a significant number of our patients did. The decision to prescribe AEDs, of AEDs following hospitalization, the selection, and the need to continue a certain number of AEDs were left to the discretion of the discharging physician, who was not always a neurologist. The details of discharge AEDs were variable and thus no useful conclusions could be made from this information. Our study found that the need for hospital readmission for seizure evaluation was also
correlated with the development of epilepsy. One possibility is that some of these patients were potentially discharged prematurely, but these individuals were documented to have improved or stable examinations at the time of the initial discharge. Another possibility is that readmission for seizure evaluation is the first indicator of sustained neuronal injury that predicts future epilepsy. Many of the patients in this study continued taking AEDs to prevent seizures during the acute phase of cerebral insult with the plan to discontinue the AED after recovery. There are limited data available to help physicians determine when to taper an AED. Our data describe different time points at which the AEDs were tapered with continued seizure freedom. These data are limited by reliance on retrospective chart analysis that is dependent on timely and proper documentation by the treating physicians. Nevertheless, our results indicate that, with a reassuring follow-up EEG and normal neurological examination, a subgroup of these patients can successfully taper AEDs. As with many retrospective studies, limitations to our data include reliance on appropriate chart documentation, which may result in an underestimation of our findings. For example, information regarding breakthrough seizures was not always well documented. Our sample size became problematic when trying to conduct specific analysis in subgroups of our cohort. Additionally, patient management was not standardized and was
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
FIGURE 3. Hospital duration in days, with the median number of days compared between the outcome groups of aggressive seizure control versus no aggressive seizure control. The black lines within the bars depict the median hospital durations (26 days for “no aggressive seizure control” and 13 days for “yes aggressive seizure control”); the colored bars depict the interquartile range (IQR) for each group (21-88 days for “no aggressive seizure control” and 9-26 days for “yes aggressive seizure control”). The black circles above the bars are outliers because they are more than 1.5 times the IQR above the third quartile. Aggressive seizure control did not meet significance for hospital duration (P ¼ 0.103). (The color version of this figure is available in the online edition.)
dependent on the discretion of the treating physician. For example, not all of the patients in this study had an EEG preformed during hospitalization. Our cohort was also limited to patients enrolled in the CEP project, which excluded patients with an early positive HSV polymerase chain reaction. Information regarding neuronally directed antibodies (N-methyl-D-aspartate, paraneoplastic etiologies, etc.) was not available at the time of our analysis. Moreover, as reported in prior studies, an etiology was not identified in a majority of cases.3 Because of the nature of our patient population, follow-up information after discharge was also limited. Finally, factors such as
TABLE 3. Association of Aggressive Seizure Control And Outcome
Epilepsy No Yes Death No Yes Hospital duration (days)
24 hours (n ¼ 5)
>24 hours (n ¼ 13)
Total (n ¼ 18)
4 (80%) 1 (20%)
6 (46%) 7 (54%)
10 (56%) 8 (44%)
5 (100 %) 0 (0 %) Median (IQR) 13 (9, 26)
0.278 9 (69 %) 14 (78 %) 4 (31 %) 4 (22 %) Median (IQR) Median (IQR) 0.103 26 (21, 94) 25 (16, 63)
P 0.314
Abbreviation: IQR ¼ Interquartile range
71
FIGURE 4. Time to taper antiepileptic drugs (AEDs). During the course of our study, there were a 24 of 46 patients started on AEDs who were able to successfully be tapered off. The total duration on the AED before cessation was assessed and grouped in durations from 0-3 months (n ¼ 4), 4-6 months (n ¼ 6), 7-12 months (n ¼ 2), 1-3 years (n ¼ 3), or 3þ years (n ¼ 9). Illustrated is the percentage of patients that were safely tapered off AEDs at the different durations. (The color version of this figure is available in the online edition.)
hospital duration, time to follow-up, and time to wean off an AED were arbitrarily decided by the managing physician. There is limited information to guide prognosis after resolution of the acute event. Although there is no standardized approach to the management of patients with suspected encephalitis, our data can be used in formulating an algorithm in which to approach these patients. There is currently no standardization for detection of seizures, use of medically induced coma, or AED administration in the treatment of encephalitis. As reported in our prior publication, using EEG is instrumental in identifying subclinical seizures and allows for intervention and potential change in prognosis for a subset of patients.16 Our current study can provide reliable prognostic information to families of children affected with presumed encephalitis. Another strength of our study is that this is a large long-term study that encompasses most types of encephalitis and, thus, is widely applicable. Several large studies available are focused on encephalitis resulting from a particular virus, such as HSV or Japanese encephalitis, whereas ours is all-encompassing. Finally, these data stress the need for a large-scale prospective multiinstitutional trial to better understand the optimal management of seizures in the acute setting. Such trials should specifically investigate: (1) the efficacy of specific AEDs to improve outcomes, (2) whether aggressive seizure management in a larger population yields better outcomes, (3) the degree of aggressiveness employed (seizures controlled within shorter durations than 24 hours), and (4) standardization of AED tapering after encephalitis. These lines of investigation may alter the acute management of seizures in encephalitis, which could greatly modify the later development of epilepsy and the success rate of tapering anticonvulsant medications after resolution of encephalitis.
72
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
This project was partially supported by the National Institutes of Health, Grant UL1TR000100.
References 1. Johnson RT. Acute encephalitis. Clin Infect Dis. 1996;23:219-224. quiz 225-226. 2. Glaser CA, Gilliam S, Schnurr D, et al. In search of encephalitis etiologies: diagnostic challenges in the California Encephalitis Project, 1998-2000. Clin Infect Dis. 2003;36:731-742. 3. Glaser CA, Honarmand S, Anderson LJ, et al. Beyond viruses: clinical profiles and etiologies associated with encephalitis. Clin Infect Dis. 2006;43:1565-1577. 4. Thompson C, Kneen R, Riordan A, Kelly D, Pollard AJ. Encephalitis in children. Arch Dis Child. 2012;97:150-161. 5. Fodor PA, Levin MJ, Weinberg A, Sandberg E, Sylman J, Tyler KL. Atypical herpes simplex virus encephalitis diagnosed by PCR amplification of viral DNA from CSF. Neurology. 1998;51:554-559. 6. Misra UK, Kalita J. Seizures in encephalitis: predictors and outcome. Seizure. 2009;18:583-587. 7. Pitkanen A. Therapeutic approaches to epileptogenesisehope on the horizon. Epilepsia. 2010;51(Suppl 3):2-17. 8. McGrath N, Anderson NE, Croxson MC, Powell KF. Herpes simplex encephalitis treated with acyclovir: diagnosis and long term outcome. J Neurol Neurosurg Psychiatry. 1997;63:321-326. 9. Elbers JM, Bitnun A, Richardson SE, et al. A 12-year prospective study of childhood herpes simplex encephalitis: is there a broader spectrum of disease? Pediatrics. 2007;119:e399-e407. 10. Hauser WA, Annegers JF, Rocca WA. Descriptive epidemiology of epilepsy: contributions of population-based studies from Rochester, Minnesota. Mayo Clin Proc. 1996;71:576-586. 11. Forsgren L, Bucht G, Eriksson S, Bergmark L. Incidence and clinical characterization of unprovoked seizures in adults: a prospective population-based study. Epilepsia. 1996;37:224-229. 12. Herman ST. Epilepsy after brain insult: targeting epileptogenesis. Neurology. 2002;59(9 Suppl 5):S21-S26. 13. Ravizza T, Balosso S, Vezzani A. Inflammation and prevention of epileptogenesis. Neurosci Lett. 2011;497:223-230. 14. Annegers JF, Hauser WA, Beghi E, Nicolosi A, Kurland LT. The risk of unprovoked seizures after encephalitis and meningitis. Neurology. 1988;38:1407-1410. 15. Haltiner AM, Temkin NR, Dikmen SS. Risk of seizure recurrence after the first late posttraumatic seizure. Arch Phys Med Rehabil. 1997;78:835-840. 16. Gold JJ, Crawford JR, Glaser C, Sheriff H, Wang S, Nespeca M. The role of continuous electroencephalography in childhood encephalitis. Pediatr Neurol. 2013;50:318-323. 17. Fisher RS, Acevedo C, Arzimanoglou A, et al. ILAE official report: a practical clinical definition of epilepsy. Epilepsia. 2014;55:475-482. 18. Hesdorffer DC, Benn EK, Cascino GD, Hauser WA. Is a first acute symptomatic seizure epilepsy? Mortality and risk for recurrent seizure. Epilepsia. 2009;50:1102-1108.
19. Team RC. R: A language and environment for statistical computing. In. Vienna, Austria: R Foundation for Statistical Computing; 2014.. 20. Loscher W, Brandt C. Prevention or modification of epileptogenesis after brain insults: experimental approaches and translational research. Pharmacol Rev. 2010;62:668-700. 21. Kalita J, Misra UK, Pandey S, Dhole TN. A comparison of clinical and radiological findings in adults and children with Japanese encephalitis. Arch Neurol. 2003;60:1760-1764. 22. Pitkanen A, Lukasiuk K. Mechanisms of epileptogenesis and potential treatment targets. Lancet Neurol. 2011;10:173-186. 23. Jung KH, Chu K, Lee ST, et al. Cyclooxygenase-2 inhibitor, celecoxib, inhibits the altered hippocampal neurogenesis with attenuation of spontaneous recurrent seizures following pilocarpine-induced status epilepticus. Neurobiol Dis. 2006;23:237-246. 24. Libbey JE, Kirkman NJ, Smith MC, et al. Seizures following picornavirus infection. Epilepsia. 2008;49:1066-1074. 25. Temkin NR. Preventing and treating posttraumatic seizures: the human experience. Epilepsia. 2009;50(Suppl 2):10-13. 26. Lowenstein DH. Epilepsy after head injury: an overview. Epilepsia. 2009;50(Suppl 2):4-9. 27. Pitkanen A, Immonen RJ, Grohn OH, Kharatishvili I. From traumatic brain injury to posttraumatic epilepsy: what animal models tell us about the process and treatment options. Epilepsia. 2009;50(Suppl 2):21-29. 28. Temkin NR, Dikmen SS, Anderson GD, et al. Valproate therapy for prevention of posttraumatic seizures: a randomized trial. J Neurosurg. 1999;91:593-600. 29. Dikmen SS, Machamer JE, Winn HR, Anderson GD, Temkin NR. Neuropsychological effects of valproate in traumatic brain injury: a randomized trial. Neurology. 2000;54:895-902. 30. Hesdorffer DC, Logroscino G, Cascino G, Annegers JF, Hauser WA. Risk of unprovoked seizure after acute symptomatic seizure: effect of status epilepticus. Ann Neurol. 1998;44:908-912. 31. Pitkanen A, Kharatishvili I, Narkilahti S, Lukasiuk K, Nissinen J. Administration of diazepam during status epilepticus reduces development and severity of epilepsy in rat. Epilepsy Res. 2005;63: 27-42. 32. Bortel A, Levesque M, Biagini G, Gotman J, Avoli M. Convulsive status epilepticus duration as determinant for epileptogenesis and interictal discharge generation in the rat limbic system. Neurobiol Dis. 2010;40:478-489. 33. Susaki J, Chegondi M, Raszynski A, Totapally BR. Outcome of children with acute encephalitis and refractory status epilepticus. J Child Neurol. 2014;29:1638-1644. 34. Gold JJ, Nespeca M, Wang S. Risk of recurrent seizures and duration of AED therapy after suspected childhood encephalitis. In: American Epilepsy Society. Washington DC; 2013. 35. Hayashi K, Osawa M, Aihara M, et al. Efficacy of intravenous midazolam for status epilepticus in childhood. Pediatr Neurol. 2007;36: 366-372. 36. Kim SJ, Lee DY, Kim JS. Neurologic outcomes of pediatric epileptic patients with pentobarbital coma. Pediatr Neurol. 2001;25:217-220. 37. Barberio M, Reiter PD, Kaufman J, Knupp K, Dobyns EL. Continuous infusion pentobarbital for refractory status epilepticus in children. J Child Neurol. 2011;27:721-726.
Contents lists available at ScienceDirect
Pediatric Neurology journal homepage: www.elsevier.com/locate/pnu
Original Article
Epilepsy After Resolution of Presumed Childhood Encephalitis Neggy Rismanchi MD, PhD a, Jeffrey J. Gold MD, PhD a, Shifteh Sattar MD a, Carol A. Glaser DVM, MD b, Heather Sheriff BA b, James Proudfoot MS c, Andrew Mower MD d, John R. Crawford MD, MS a, Mark Nespeca MD a, Sonya G. Wang MD a, * a
Department of Neurosciences, University of California, San Diego, California California Department of Public Health, Richmond, California c Biostatistics Unit, Clinical and Translational Research Institute, University of California, San Diego, California d Children’s Neurology Center, Children’s Hospital of Orange County, Orange, California b
abstract OBJECTIVE: To evaluate factors associated with the development of epilepsy after resolution of presumed childhood encephalitis. METHODS: A total of 217 patients with suspected encephalitis who met criteria for the California Encephalitis Project were identified. Evaluable outcome information was available for 99 patients (40 girls, 59 boys, ages 2 months to 17 years) without preexisting neurological conditions, including prior seizures or abnormal brain magnetic resonance imaging scans. We identified factors correlated with the development of epilepsy after resolution of the acute illness. RESULTS: Development of epilepsy was correlated with the initial presenting sign of seizure (P < 0.001). With each additional antiepileptic drug used to control seizures, the odds ratio of developing epilepsy was increased twofold (P < 0.001). An abnormal electroencephalograph (P < 0.05) and longer hospital duration (median of 8 versus 21 days) also correlated with development of epilepsy (P < 0.01). The need for medically induced coma was associated with epilepsy (P < 0.001). Seizures in those patients were particularly refractory, often requiring longer than 24 hours to obtain seizure control. Individuals who required antiepileptic drugs at discharge (P < 0.001) or were readmitted after their acute illness (P < 0.001) were more likely to develop epilepsy. Of our patients who were able to wean antiepileptic drugs after being started during hospitalization, 42% were successfully tapered off within 6 months. CONCLUSIONS: Limited data are available on the risk of developing epilepsy after childhood encephalitis. This is the first study that not only identifies risk factors for the development of epilepsy, but also provides data regarding the success rate of discontinuing antiepileptic medication after resolution of encephalitis. Keywords: Epilepsy, childhood, encephalitis, seizures
Pediatr Neurol 2015; 53: 65-72 Ó 2015 Elsevier Inc. All rights reserved.
Introduction
Encephalitis is potentially devastating, particularly in children. The annual incidence worldwide ranges from 3.5 to 7.4 per 100,000 and increases to 16 per 100,000 in
Article History: Received February 4, 2015; Accepted in final form March 14, 2015 * Communications should be addressed to: Dr. Wang; Department of Neurosciences; UC San Diego; 3020 Children’s Way, MC 5009; San Diego, CA 92123. E-mail address: [email protected] 0887-8994/$ e see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.pediatrneurol.2015.03.016
children.1 Presenting signs and symptoms can be nonspecific and subtle or quite dramatic. Some children present only with low-grade fever, headache, or nausea, whereas others present with profound encephalopathy, vomiting, seizures, and/or focal neurological deficits.2-5 Those children who present with subtle signs can also unexpectedly worsen and later develop seizures and other severe neurological sequelae.5 Encephalitis can result in greater morbidity and mortality in children than in adults. In one study of 1570 patients in California, 62 had intractable seizures during the acute illness, with children comprising 69% of these
66
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
patients.3 Another study of 148 patients in India demonstrated that 61% of children had seizures during encephalitis in comparison to 37% of the adults.6 Seizures reportedly are more likely to occur with specific viral infections, with herpes virus (HSV) (75% of patients with HSV) more commonly associated with seizures than Japanese encephalitis (54%), followed by nonspecific encephalitis (43%).6 The postinfectious long-term complications of encephalitis include dystonia, spasticity, epilepsy, and neurocognitive problems.4 In fact, despite treatment with acyclovir in cases of herpes encephalitis, two-thirds of patients experienced significant neurological impairment.7 Unfortunately, literature on long-term outcome and prognosis following encephalitis is limited, but studies on HSV encephalitis quote long-term morbidity rates as 30% in adults and up to 67% in children, with seizures occurring in 44% of these patients in one prospective study in Toronto, Canada.8,9 In the general population, seizures are a common neurological problem, with up to 10% of the population having a seizure in their lifetime and up to 3% of the adult population developing epilepsy.10,11 Central nervous system (CNS) injuries such as stroke, viral infection, trauma, febrile seizure, and status epilepticus are considered common risk factors for developing epilepsy.12,13 A group in Texas reported the 20-year risk of developing an unprovoked seizure after encephalitis or meningitis to be 6.8%, noting that patients with a prior CNS infection had a sevenfold increase in unprovoked seizures compared with the general population.14 CNS injuries are reported to account for 30%49% of all unprovoked seizures and epilepsies. Long-term follow up studies conducted in Seattle of individuals with a history of brain injury reveal that after a single late seizure, the risk for recurrent seizure is typically greater than 80%.15 The risk of epilepsy was increased by 16 times with viral encephalitis when compared to the general population and this risk remained elevated up to 15 years postinfection.12 A correlation between encephalitis and later development of epilepsy has been shown; however, these studies were often limited by follow-up data, etiology of encephalitis, low patient number, or lack of assessment of other potential contributing factors that resulted in epilepsy. To identify factors potentially contributing to the development of epilepsy, we studied a large cohort of patients longitudinally at a single institution who were enrolled in the California Encephalitis Project (CEP). Methods A cohort of 217 patients who met diagnostic criteria for inclusion in the CEP at Rady Children’s Hospital San Diego between 2004 and 2011 were studied.16 To be included in the CEP, children had to be hospitalized with encephalopathy lasting 24 hours and meet at least one of the criteria: fever, seizure, cerebrospinal fluid pleocytosis, focal neurological signs, or with neuroimaging or electroencephalograph (EEG) evidence of encephalitis.3 Details of the CEP and the standardized testing associated with CEP have previously been described.2,3,16 This retrospective study was granted approval under the multiinstitutional protocol established by the State of California for the CEP and a waiver of consent was granted. Patients were excluded from our study if they had prior neurological abnormalities, tumor, or were found to have other medical conditions causing these similar encephalitic signs and symptoms. Patients who were lost to follow-up or died during hospitalization were also excluded, resulting in
99 patients remaining in the study. Fig 1 summarizes these excluded patients. In this study, factors from initial presentation, hospital course, and discharge that might predispose a patient to epilepsy were assessed. The electronic medical record was reviewed including emergency room encounters, hospital admission course, medication administrations, all prior and subsequent clinic visits, laboratory results, neuroimaging studies, and EEG results.16 Details of neuroimaging and EEG acquisition are described in a prior publication.16 In this study, the term “epilepsy” was defined as any patient with at least one unprovoked seizure after resolution of the patient’s acute CNS insult. This is now an internationally accepted alternative definition of epilepsy in comparison to the traditional definition of requiring two unprovoked seizures.17,18 For our assessment, “difficult to control seizures” was defined as patients who required medically induced coma and more than three antiepileptic drugs (AEDs) for seizure control. Continuous variables were reported as medians and interquartile ranges, and categorical data as counts and percentages. The Fisher exact test was used to conduct group comparisons between categorical variables (development of epilepsy, discharge on AEDs, etc.), and the Mann-Whitney U test to compare continuous variables (hospital duration). A logistic regression was used to assess the impact of the number of AEDs given with the odds of developing epilepsy. P-values less than 0.05 were considered a statistically significant result. All data analyses are conducted using the R statistical programming language.19
Results
Within our cohort of 217 patients, follow-up information was available for a total of 99 patients who had met the inclusion criteria (Fig 1). Our study cohort comprised 40 females and 59 males, with a mean age at presentation of 9 years old (range of 2 months to 17 years). Of the 99 patients in this study, 24 later developed epilepsy. Presenting features associated with developing epilepsy
According to our demographic analysis, race seemed to be correlated with later development of epilepsy, although this had borderline statistical significance (Table 1, P ¼ 0.049). This relationship seems to be largely driven by a much higher proportion of Caucasian subjects in the nonepileptic group (n ¼ 18, 24%) compared with the proportion of Caucasian subjects in the epileptic group (n ¼ 1, 4.2%). Because of this discrepancy between the number of individuals in each of the Caucasian groups, race has a statically significant correlation with development with epilepsy, but these data suggest that being Caucasian decreases likelihood of developing epilepsy rather than a particular race increasing likelihood of epilepsy. A seizure at presentation showed a positive correlation with the later development of epilepsy (P < 0.001). There was not a significant relationship between presenting symptoms of fever, headache, or altered mental status and the subsequent development of epilepsy in this study. Elements of hospital stay associated with developing epilepsy
One major factor found to be important in the development of epilepsy was the number of AEDs used to obtain seizure control. These AEDs included maintenance and abortive seizure medications given en route to the hospital, in the emergency room, or during the hospitalization as well as any continuous infusions needed to control seizures
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
67
FIGURE 1. Study characteristics. There were 217 patients who were enrolled in California Encephalitis Project (CEP). Ninety-six patients were excluded because of a preexisting/concurrent neurological problem, tumor, acute process found to be something other than encephalitis, or no evidence of encephalitis upon closer examination of patient’s data. Of the 121 patients that remained in the study, 19 patients were lost to follow-up and 3 died during the hospitalization, so no outcome measures were available to assess, resulting in a total of 99 patients that remained in our study. CNS, central nervous system; HA, headache; psych, psychiatric. (The color version of this figure is available in the online edition.)
during the hospitalization. The number of patients who developed epilepsy per category of number of AEDs used was assessed. For no drug (n ¼ 38), there were no patients who developed epilepsy; for one AED (n ¼ 7), there was one patient; for two AEDs (n ¼ 13), there were two patients; for three AEDs (n ¼ 14), there were three patients; for four AEDs (n ¼ 6), there were four patients; for five AEDs (n ¼ 10), there were six patients; and for five or more AEDs (n ¼ 11), there were eight patients who developed epilepsy. An increase in the number of AEDs administered correlated with an increase in the percentage of patients who later developed epilepsy (Fig 2A). A logistic regression model was used for the development of epilepsy with the number of AEDs as a predictor, which identified the odds of later developing epilepsy to be increased by a factor of 2.2 with each addition of an AED (95% confidence interval 1.6-3.1; P < 0.001). Length of hospital stay was also found to be a significant factor in the development of epilepsy. The total length of hospital stay (in days) was compared between the individuals who did and did not develop epilepsy. Using the Mann-Whitney U-test to account for rare extreme values in
both groups, this study shows that the median hospital duration of those who developed epilepsy was 21 days in comparison to 8 days in the nonepilepsy group (P ¼ 0.006; Fig 2B). Other elements during hospitalization that were analyzed are summarized in Table 2A. These factors included whether antibiotics were administered, the presence of cerebrospinal fluid pleocytosis (defined as white blood cell count 10), abnormal head CT, abnormal brain magnetic resonance imaging, abnormal EEG, need for intensive care unit stay during admission, or need for medically induced coma for seizure control. An abnormal EEG during hospital stay (P ¼ 0.044) and the need to use medically induced coma (P < 0.001) were both found to have a significant relationship with the development of epilepsy. Of the eight children who died in our series, four had been placed in medically induced comas. All four of these patients did not obtain seizure control until more than 24 hours after seizure onset. To analyze acute hospital management and the sequelae of epilepsy, the early success of aggressive seizure control therapies was evaluated. Successful
68
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
TABLE 1. Demographics and Presenting Symptoms
Total (n ¼ 99)
Epilepsy No (n ¼ 75) Age at onset Sex Female Male Race Hispanic Caucasian Asian African American Unknown/other Fever No Yes Headache No Yes Altered mental status No Yes Seizure No Yes
9 (4, 12)
P
Yes (n ¼ 24) 6 (3, 10.25)
8 (3, 12)
28 (37.3%) 47 (62.7%)
12 (50%) 12 (50%)
40 (40.4%) 59 (59.6%)
36 18 6 3 12
12 1 5 3 3
48 19 11 6 15
0.306 0.341
0.049* (48%) (24%) (8%) (4%) (16%)
(50%) (4.2%) (20.8%) (12.5%) (12.5%)
(48.5%) (19.2%) (11.1%) (6.1%) (15.2%) 0.099
26 (34.7%) 49 (65.3%)
13 (54.2%) 11 (45.8%)
39 (39.4%) 60 (60.6%)
32 (42.7%) 43 (57.3%)
11 (47.8%) 12 (52.2%)
43 (43.9%) 55 (56.1%)
27 (36%) 48 (64%)
5 (20.8%) 19 (79.2%)
32 (32.3%) 67 (67.7%)
50 (66.7%) 25 (33.3%)
6 (25%) 18 (75%)
56 (56.6%) 43 (43.4%)
0.811
0.214
<0.001***
Significance: ***0.001, **0.01, *0.05.
aggressive seizure control was defined as seizure control within 24 hours or less. Statistical analysis did meet significance for mortality; however, all four patients who died during medically induced coma seized for more than 24 hours and this prompted further analysis (Fig 3 and Table 3). Within our medically induced coma group (n ¼ 18), the median hospital duration of the patients with early successful aggressive seizure control (n ¼ 5) was 13 days. In comparison, the patients whose seizures lasted longer than 24 hours and were eventually placed in medically induced coma (n ¼ 13) had a median hospital duration of 26 days (P ¼ 0.103). Two of the patients in the latter group died, likely shortening the mean hospital duration significantly. Last, we looked at difficult to control seizures, defined as requiring medically induced coma plus three or more AEDs. Although three of these eight patients died because of difficulty obtaining seizure control, this was found to be trending toward statistical significance (P ¼ 0.077). Factors after discharge associated with developing epilepsy
Patients who required discharge with an AED or who required readmission for concern of sequelae from encephalitis were also assessed (Table 2B). The patients evaluated for readmission included acute illness related to the initial admission within a 2-month period or concern of seizure within a year after discharge. This study shows that patients who needed to be discharged on an AED (P < 0.001) or having to be readmitted (P < 0.001) were associated with the later development of epilepsy. In total, 27 patients required readmission after discharge. Nineteen of the 27 patients were readmitted within 2 weeks. Twelve of the 27 were readmitted specifically for seizure or possible seizure.
Finally, of the 42 patients who started AEDs during hospitalization and the four others that were started shortly after hospitalization, a total of 24 patients were able to be tapered off of AEDs and remained seizure-free during the time course of our study. Assessment of total time on the AED after discharge was conducted by categorizing them as 0-3 months (n ¼ 4), 4-6 months (n ¼ 6), 7-12 months (n ¼ 2), 1-3 years (n ¼ 3), or 3þ years (n ¼ 9). The percent of total patients tapered off in each group is illustrated in Fig 4. This study shows that about 42% of the patients whose AEDs were successfully tapered were able to do so by 6 months, with the remainder continuing an AED for 7 months or longer. Discussion
Limited information is available regarding the outcomes of childhood encephalitis, particularly in evaluating the factors involved in the development of epilepsy. Although race was found to have a marginally significant relationship with the development of epilepsy in this study, we advise caution in interpreting this result because of the small sample size of Asian and African Americans and the presence of an unknown/other category. Additionally, the findings are largely driven by the disproportionate number of Caucasians in the no epilepsy group rather than a high percentage of any particular race in the epilepsy group. Seizure at the onset of the illness was directly associated with the development of epilepsy. Seizures have been previously well-described in association with encephalitis and, although this sign is not as common as fever or headache, it may be more specifically related to a cerebral insult that would result in abnormal neural reorganization, blood-brain barrier breakdown, glial activation, neuronal hyperexcitability, or other
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
FIGURE 2. Factors associated with epilepsy during hospitalization. The number of antiepileptic drugs (AEDs) used to control seizures correlated with development of epilepsy (A). Each of the patients included in the study (n ¼ 99) was assessed for number of AEDs given during the patient’s acute encephalitis (this includes abortive medications and continuous infusions). The percentage of patients who then developed epilepsy in each of these groups from 0 to 5þ AEDs were calculated. For no drug (n ¼ 38), there were no patients who developed epilepsy; for 1 AED (n ¼ 7), there was 1 patient; for 2 AEDs (n ¼ 13), there were 2 patients; for 3 AEDs (n ¼ 14), there were 3 patients; for 4 AEDs (n ¼ 6), there were 4 patients; for 5 AEDs (n ¼ 10), there were 6 patients; and for more than 5 AEDs (n ¼ 11), there were 8 patients who developed epilepsy. Using a logistic regression for the development of epilepsy with the number of AEDs resulted in an odds ratio (OR) of 2.2 with a 95% confidence interval (CI) of 1.6-3.1 (P < 0.001). Hospital duration correlated with development of epilepsy (B). The figure is a graphical representation of the median number of hospital days when comparing the outcome groups of epilepsy versus no epilepsy. The black lines within the bars depict the median hospital durations (8 days for “no epilepsy” and 21 days for “yes epilepsy”); the colored bars depict the interquartile range for each group (11 days for “no epilepsy” and 27.5 days for “yes epilepsy”). The black circles above the bars are outliers because they are more than 1.5 times the interquartile range above the third quartile. A longer hospital duration was found to be correlated with later development of epilepsy (P ¼ 0.006). (The color version of this figure is available in the online edition.)
alterations suspected to result in epileptogenesis.3,5e7,14,20 Human studies reveal that certain viral causes of encephalitis are more likely to initially present with seizure
69
and that seizures in the context of encephalitis tend to be more frequent and difficult to control, which may lend itself to the later development of epilepsy.6,21 In fact, animal models of epilepsy use this relation by inducing repeated seizures with direct electrical stimulation or introduction of an infectious agent that results in an immune response in the brain to study epilepsy.22,23 Investigators have demonstrated spontaneous seizures in murine models of encephalitis and that lasting hyperexcitability triggers further inflammatory processes in the brain that leads to epileptogenesis.13,24,25 Difficulty attaining seizure control predicted a higher risk of epilepsy. This was evaluated by assessing the total number of AEDs used in treatment as well as the need for medically induced coma for seizure control. This study shows that each increase in number of AEDs required to control seizures during the acute illness resulted in a twofold increased likelihood of developing epilepsy. Several animal models and human studies evaluated time to seizure control and the correlation with future epilepsy, particularly focusing on the efficacy of certain AEDs (including diazepam, levetiracetam, phenytoin, phenobarbital, valproic acid, and topiramate).7,20,25-29 Unfortunately, the results of these studies are variable and fail to suggest a clear path to approach acute seizure control after a cerebral insult with the goal of achieving the best long-term neurological outcome. Shortening the duration of status epilepticus modifies the epileptogenic process because it is known that 43% of patients who are in status epilepticus later develop epilepsy.30-32 Our attempts to further analyze the type and sequence of AEDs used in correlation with outcomes were unsuccessful because of varied treatment paths between patients. Other factors during the hospital course proved to be very revealing in providing correlations with the development of epilepsy. Increased length of hospital stay was correlated with the development of epilepsy; however, our analysis was unable to consider comorbidities that arose during hospitalization, which may have contributed to a longer hospital stay. Although our results were similar to those in a recently published, smaller-scale study, it is possible that the length of hospital stay acted as a surrogate for disease severity even when looking specifically at epilepsy as an outcome measure.33 An abnormal EEG during the hospital course (including slowing, spikes, and sharp waves) significantly correlated with the risk of subsequent epilepsy. However, 20% of patients with an abnormal EEG did not subsequently develop epilepsy. Children with normalization of EEG after discharge were unlikely to later develop epilepsy.34 Such children were also more likely to be tapered off AED medication because this is a common current practice pattern. In this study, though not statistically significant, patients who required medically induced coma for seizure control had a higher association with the development of epilepsy. Further examination of our medically induced coma cohort actually revealed that the patients who achieved seizure freedom within 24 hours had better outcomes in terms of death and epilepsy as well as hospital duration; however, our sample size was not sufficient to yield significance to these findings. Patients who did not achieve seizure control within 24 hours may be more ill
70
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
TABLE 2. Factors during and after Hospitalization Associated with Epilepsy
A.
Antibiotic use No Yes Cerebrospinal fluid pleocytosis No Yes Abnormal computed tomography scan No Yes Abnormal magnetic resonance imaging No Yes Abnormal electroencephalograph No Yes Intensive care unit stay No Yes Medically induced coma No Yes B. Discharge on antiepileptic drugs No Yes Readmitted No Yes
Total (n ¼ 99)
Epilepsy No (n ¼ 75)
Yes (n ¼ 24)
22 (30.1%) 51 (69.9%)
6 (27.3%) 16 (72.7%)
28 (29.5%) 67 (70.5%)
11 (14.7%) 64 (85.3%)
8 (33.3%) 16 (66.7%)
19 (19.2%) 80 (80.8%)
51 (76.1%) 16 (23.9%)
13 (59.1%) 9 (40.9%)
64 (71.9%) 25 (28.1%)
23 (33.3%) 46 (66.7%)
7 (30.4%) 16 (69.6%)
30 (32.6%) 62 (67.4%)
8 (17.8%) 37 (82.2%)
0 (0.0%) 24 (100%)
8 (11.6%) 61 (88.4%)
37 (49.3%) 38 (50.7%)
8 (33.3%) 16 (66.7%)
45 (45.5%) 54 (54.5%)
69 (92%) 6 (8%)
14 (58.3%) 10 (41.7%)
83 (83.8%) 16 (16.2%)
P
0.999
0.071
0.171
0.999
0.044*
0.239
<0.001***
Total (n ¼ 99)
Epilepsy No (n ¼ 75)
Yes (n ¼ 24)
51 (68.9%) 23 (31.1%)
1 (4.3%) 22 (95.7%)
52 (53.6%) 45 (46.4%)
60 (81.1%) 14 (18.9%)
10 (43.5%) 13 (56.5%)
70 (72.2%) 27 (27.8%)
P <0.001***
<0.001***
Significance codes: ***0.001, **0.01, *0.05.
independent of when the medically induced coma was initiated and would be expected to ultimately have poorer outcomes regardless of the interventions. It is also possible that, similar to drug-resistant epilepsy in the outpatient setting after adequate trials of two AEDs, there is a common pathway of epileptogenesis that must be considered. Furthermore, in the setting of acute encephalitis, the severity of neuronal injury is even more substantial. Although the use of medically induced coma is not without risk, there have been studies suggesting that the benefits of early seizure control by medically induced coma (less than 24 hours) outweigh the risks. Some of these studies also have limited sample sizes.35-37 Given the results of these studies as well as our own, early medically induced coma may be beneficial, but clinical trials are necessary to validate this hypothesis. Being discharged with antiepileptic therapy or readmitted after the acute illness were also correlated with the development of epilepsy. Although not all patients discharged on an AED developed epilepsy, a significant number of our patients did. The decision to prescribe AEDs, of AEDs following hospitalization, the selection, and the need to continue a certain number of AEDs were left to the discretion of the discharging physician, who was not always a neurologist. The details of discharge AEDs were variable and thus no useful conclusions could be made from this information. Our study found that the need for hospital readmission for seizure evaluation was also
correlated with the development of epilepsy. One possibility is that some of these patients were potentially discharged prematurely, but these individuals were documented to have improved or stable examinations at the time of the initial discharge. Another possibility is that readmission for seizure evaluation is the first indicator of sustained neuronal injury that predicts future epilepsy. Many of the patients in this study continued taking AEDs to prevent seizures during the acute phase of cerebral insult with the plan to discontinue the AED after recovery. There are limited data available to help physicians determine when to taper an AED. Our data describe different time points at which the AEDs were tapered with continued seizure freedom. These data are limited by reliance on retrospective chart analysis that is dependent on timely and proper documentation by the treating physicians. Nevertheless, our results indicate that, with a reassuring follow-up EEG and normal neurological examination, a subgroup of these patients can successfully taper AEDs. As with many retrospective studies, limitations to our data include reliance on appropriate chart documentation, which may result in an underestimation of our findings. For example, information regarding breakthrough seizures was not always well documented. Our sample size became problematic when trying to conduct specific analysis in subgroups of our cohort. Additionally, patient management was not standardized and was
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
FIGURE 3. Hospital duration in days, with the median number of days compared between the outcome groups of aggressive seizure control versus no aggressive seizure control. The black lines within the bars depict the median hospital durations (26 days for “no aggressive seizure control” and 13 days for “yes aggressive seizure control”); the colored bars depict the interquartile range (IQR) for each group (21-88 days for “no aggressive seizure control” and 9-26 days for “yes aggressive seizure control”). The black circles above the bars are outliers because they are more than 1.5 times the IQR above the third quartile. Aggressive seizure control did not meet significance for hospital duration (P ¼ 0.103). (The color version of this figure is available in the online edition.)
dependent on the discretion of the treating physician. For example, not all of the patients in this study had an EEG preformed during hospitalization. Our cohort was also limited to patients enrolled in the CEP project, which excluded patients with an early positive HSV polymerase chain reaction. Information regarding neuronally directed antibodies (N-methyl-D-aspartate, paraneoplastic etiologies, etc.) was not available at the time of our analysis. Moreover, as reported in prior studies, an etiology was not identified in a majority of cases.3 Because of the nature of our patient population, follow-up information after discharge was also limited. Finally, factors such as
TABLE 3. Association of Aggressive Seizure Control And Outcome
Epilepsy No Yes Death No Yes Hospital duration (days)
24 hours (n ¼ 5)
>24 hours (n ¼ 13)
Total (n ¼ 18)
4 (80%) 1 (20%)
6 (46%) 7 (54%)
10 (56%) 8 (44%)
5 (100 %) 0 (0 %) Median (IQR) 13 (9, 26)
0.278 9 (69 %) 14 (78 %) 4 (31 %) 4 (22 %) Median (IQR) Median (IQR) 0.103 26 (21, 94) 25 (16, 63)
P 0.314
Abbreviation: IQR ¼ Interquartile range
71
FIGURE 4. Time to taper antiepileptic drugs (AEDs). During the course of our study, there were a 24 of 46 patients started on AEDs who were able to successfully be tapered off. The total duration on the AED before cessation was assessed and grouped in durations from 0-3 months (n ¼ 4), 4-6 months (n ¼ 6), 7-12 months (n ¼ 2), 1-3 years (n ¼ 3), or 3þ years (n ¼ 9). Illustrated is the percentage of patients that were safely tapered off AEDs at the different durations. (The color version of this figure is available in the online edition.)
hospital duration, time to follow-up, and time to wean off an AED were arbitrarily decided by the managing physician. There is limited information to guide prognosis after resolution of the acute event. Although there is no standardized approach to the management of patients with suspected encephalitis, our data can be used in formulating an algorithm in which to approach these patients. There is currently no standardization for detection of seizures, use of medically induced coma, or AED administration in the treatment of encephalitis. As reported in our prior publication, using EEG is instrumental in identifying subclinical seizures and allows for intervention and potential change in prognosis for a subset of patients.16 Our current study can provide reliable prognostic information to families of children affected with presumed encephalitis. Another strength of our study is that this is a large long-term study that encompasses most types of encephalitis and, thus, is widely applicable. Several large studies available are focused on encephalitis resulting from a particular virus, such as HSV or Japanese encephalitis, whereas ours is all-encompassing. Finally, these data stress the need for a large-scale prospective multiinstitutional trial to better understand the optimal management of seizures in the acute setting. Such trials should specifically investigate: (1) the efficacy of specific AEDs to improve outcomes, (2) whether aggressive seizure management in a larger population yields better outcomes, (3) the degree of aggressiveness employed (seizures controlled within shorter durations than 24 hours), and (4) standardization of AED tapering after encephalitis. These lines of investigation may alter the acute management of seizures in encephalitis, which could greatly modify the later development of epilepsy and the success rate of tapering anticonvulsant medications after resolution of encephalitis.
72
N. Rismanchi et al. / Pediatric Neurology 53 (2015) 65e72
This project was partially supported by the National Institutes of Health, Grant UL1TR000100.
References 1. Johnson RT. Acute encephalitis. Clin Infect Dis. 1996;23:219-224. quiz 225-226. 2. Glaser CA, Gilliam S, Schnurr D, et al. In search of encephalitis etiologies: diagnostic challenges in the California Encephalitis Project, 1998-2000. Clin Infect Dis. 2003;36:731-742. 3. Glaser CA, Honarmand S, Anderson LJ, et al. Beyond viruses: clinical profiles and etiologies associated with encephalitis. Clin Infect Dis. 2006;43:1565-1577. 4. Thompson C, Kneen R, Riordan A, Kelly D, Pollard AJ. Encephalitis in children. Arch Dis Child. 2012;97:150-161. 5. Fodor PA, Levin MJ, Weinberg A, Sandberg E, Sylman J, Tyler KL. Atypical herpes simplex virus encephalitis diagnosed by PCR amplification of viral DNA from CSF. Neurology. 1998;51:554-559. 6. Misra UK, Kalita J. Seizures in encephalitis: predictors and outcome. Seizure. 2009;18:583-587. 7. Pitkanen A. Therapeutic approaches to epileptogenesisehope on the horizon. Epilepsia. 2010;51(Suppl 3):2-17. 8. McGrath N, Anderson NE, Croxson MC, Powell KF. Herpes simplex encephalitis treated with acyclovir: diagnosis and long term outcome. J Neurol Neurosurg Psychiatry. 1997;63:321-326. 9. Elbers JM, Bitnun A, Richardson SE, et al. A 12-year prospective study of childhood herpes simplex encephalitis: is there a broader spectrum of disease? Pediatrics. 2007;119:e399-e407. 10. Hauser WA, Annegers JF, Rocca WA. Descriptive epidemiology of epilepsy: contributions of population-based studies from Rochester, Minnesota. Mayo Clin Proc. 1996;71:576-586. 11. Forsgren L, Bucht G, Eriksson S, Bergmark L. Incidence and clinical characterization of unprovoked seizures in adults: a prospective population-based study. Epilepsia. 1996;37:224-229. 12. Herman ST. Epilepsy after brain insult: targeting epileptogenesis. Neurology. 2002;59(9 Suppl 5):S21-S26. 13. Ravizza T, Balosso S, Vezzani A. Inflammation and prevention of epileptogenesis. Neurosci Lett. 2011;497:223-230. 14. Annegers JF, Hauser WA, Beghi E, Nicolosi A, Kurland LT. The risk of unprovoked seizures after encephalitis and meningitis. Neurology. 1988;38:1407-1410. 15. Haltiner AM, Temkin NR, Dikmen SS. Risk of seizure recurrence after the first late posttraumatic seizure. Arch Phys Med Rehabil. 1997;78:835-840. 16. Gold JJ, Crawford JR, Glaser C, Sheriff H, Wang S, Nespeca M. The role of continuous electroencephalography in childhood encephalitis. Pediatr Neurol. 2013;50:318-323. 17. Fisher RS, Acevedo C, Arzimanoglou A, et al. ILAE official report: a practical clinical definition of epilepsy. Epilepsia. 2014;55:475-482. 18. Hesdorffer DC, Benn EK, Cascino GD, Hauser WA. Is a first acute symptomatic seizure epilepsy? Mortality and risk for recurrent seizure. Epilepsia. 2009;50:1102-1108.
19. Team RC. R: A language and environment for statistical computing. In. Vienna, Austria: R Foundation for Statistical Computing; 2014.. 20. Loscher W, Brandt C. Prevention or modification of epileptogenesis after brain insults: experimental approaches and translational research. Pharmacol Rev. 2010;62:668-700. 21. Kalita J, Misra UK, Pandey S, Dhole TN. A comparison of clinical and radiological findings in adults and children with Japanese encephalitis. Arch Neurol. 2003;60:1760-1764. 22. Pitkanen A, Lukasiuk K. Mechanisms of epileptogenesis and potential treatment targets. Lancet Neurol. 2011;10:173-186. 23. Jung KH, Chu K, Lee ST, et al. Cyclooxygenase-2 inhibitor, celecoxib, inhibits the altered hippocampal neurogenesis with attenuation of spontaneous recurrent seizures following pilocarpine-induced status epilepticus. Neurobiol Dis. 2006;23:237-246. 24. Libbey JE, Kirkman NJ, Smith MC, et al. Seizures following picornavirus infection. Epilepsia. 2008;49:1066-1074. 25. Temkin NR. Preventing and treating posttraumatic seizures: the human experience. Epilepsia. 2009;50(Suppl 2):10-13. 26. Lowenstein DH. Epilepsy after head injury: an overview. Epilepsia. 2009;50(Suppl 2):4-9. 27. Pitkanen A, Immonen RJ, Grohn OH, Kharatishvili I. From traumatic brain injury to posttraumatic epilepsy: what animal models tell us about the process and treatment options. Epilepsia. 2009;50(Suppl 2):21-29. 28. Temkin NR, Dikmen SS, Anderson GD, et al. Valproate therapy for prevention of posttraumatic seizures: a randomized trial. J Neurosurg. 1999;91:593-600. 29. Dikmen SS, Machamer JE, Winn HR, Anderson GD, Temkin NR. Neuropsychological effects of valproate in traumatic brain injury: a randomized trial. Neurology. 2000;54:895-902. 30. Hesdorffer DC, Logroscino G, Cascino G, Annegers JF, Hauser WA. Risk of unprovoked seizure after acute symptomatic seizure: effect of status epilepticus. Ann Neurol. 1998;44:908-912. 31. Pitkanen A, Kharatishvili I, Narkilahti S, Lukasiuk K, Nissinen J. Administration of diazepam during status epilepticus reduces development and severity of epilepsy in rat. Epilepsy Res. 2005;63: 27-42. 32. Bortel A, Levesque M, Biagini G, Gotman J, Avoli M. Convulsive status epilepticus duration as determinant for epileptogenesis and interictal discharge generation in the rat limbic system. Neurobiol Dis. 2010;40:478-489. 33. Susaki J, Chegondi M, Raszynski A, Totapally BR. Outcome of children with acute encephalitis and refractory status epilepticus. J Child Neurol. 2014;29:1638-1644. 34. Gold JJ, Nespeca M, Wang S. Risk of recurrent seizures and duration of AED therapy after suspected childhood encephalitis. In: American Epilepsy Society. Washington DC; 2013. 35. Hayashi K, Osawa M, Aihara M, et al. Efficacy of intravenous midazolam for status epilepticus in childhood. Pediatr Neurol. 2007;36: 366-372. 36. Kim SJ, Lee DY, Kim JS. Neurologic outcomes of pediatric epileptic patients with pentobarbital coma. Pediatr Neurol. 2001;25:217-220. 37. Barberio M, Reiter PD, Kaufman J, Knupp K, Dobyns EL. Continuous infusion pentobarbital for refractory status epilepticus in children. J Child Neurol. 2011;27:721-726.