Brain Disorders & Therapy

Viral encephalitis most commonly affects children and adults. It may cause seizures or status epilepticus in the acute stage, with potentially devastating sequelae, including post-encephalitic epilepsy (PEE) and refractory epilepsy [1]. Survivors of viral encephalitis have been shown to have a ten-fold increase in the risk of developing PEE and other neurological sequelae [2]. It has been proposed that approximately 4-24% of patients who survived with encephalitic disease would later develop post-encephalitic epilepsy (PEE) [3-5]. Among those developing PEE, a substantial proportion will develop intractable epilepsy [1]. A New Zealand study assessed the long term sequelae of herpes simplex encephalitis after treatment with acyclovir and found that 24% of patients developed PEE (50%) of whom were refractory to anticonvulsant treatment) [6]. Therefore, as Ohtsuka et al. proposed, postencephalitis is one of the main causes of localization-related refractory cases [7], thus more attention should be given to the therapy of patients with PEE.

To date, seizures and status epilepticus in the acute stage have been well-documented with regard to their clinical profile, incidence, treatment, predictors and outcome [8-10]. However, surprisingly little research has been conducted on the efficacy and safety of anticonvulsant treatments for PEE related to presumed encephalitis in both adults and children. CBZ is a traditional antiepileptic drug. It has many years of experience for treatment of epilepsy. It is currently recommended as the first-line antiepileptic drug treatment for patients with partial onset seizures and secondary generalized tonic-clonic seizures which were the overwhelming majority of seizure patterns in PEE patients [11-12]. So, in this present study, we retrospectively reviewed the efficacy and safety of anticonvulsant treatments in 292 cases of PEE related to presumed encephalitis from January 1982 to August 2009, and compared CBZ to other AEDs including PB, VPA-Na, VPA-Mg, TPM.

Patient selection

From January 1982 to August 2009 (the cases could be traced from hospital computer system), 292 patients treated in the First Affiliated Hospital of Chongqing Medical University, the Second Affiliated Hospital of Chongqing Medical University and the Children’s Hospital of Chongqing Medical University were enrolled. The selection of patients was based upon the following criteria: (1) a diagnosis of postencephalitic epilepsy as proposed in the discussion; (2) a determination that a given patient’s seizures met the definitions and classifications of those proposed by the International League Against Epilepsy (ILAE) in 2001 [13]; (3) a finding that the frequency of seizures occurred more than twice in the 12 month-period prior to therapy; (4) the part of the patients who insisted on regular therapy.

The following criteria would exclude individuals from the study: (1) patients definitively diagnosed with primary epilepsy as defined by the ILAE in 2005 [14]; (2) patients suffering from epilepsy secondary to intracranial lesions, such as cerebrovascular disease, craniocerebral injury, intracranial tumors, encephalodysplasia and demyelinating encephalopathy; (3) patients who acquired epilepsy secondary to systemic diseases, such as hypoglycemia, shock and/or uremia; and (4) patients who displayed any obvious abnormality of liver and kidney function, kidney stones, metabolic issues or a history of psychiatric illness prior to therapy.

Study procedure

Patients visited us at the clinic at six-month intervals or based on their own volition. During these sessions, seizure frequency, AED dosage, efficacy and adverse events were recorded. Blood, urine, liver and kidney functions were examined. The dosage was adjusted according to the efficacy and adverse reactions. When patients had no response to AED or discontinued the drug due to side effects, we considered it a failure for the first AED. And the physician either replaced the current drug with another AED or added a new therapy. In the first case, the patients were assigned to a monotherapy group, according to the AED taken. In the second case, the patients were assigned to a combined therapy group. Therefore, the sample size used in the efficacy assessment was based on patient visits. AEDs were selected based on seizure patterns, adverse effects, efficacy and the doctors’ experience.

A standardized data form was developed. The variables included in the database were patient characteristics (gender, age of seizure onset, type of seizures, frequency of seizures before treatment, history of febrile seizures, family history of the patient and the results of any EEG and neuroimaging).

Evaluation of efficacy and safety

The primary efficacy outcome measure was retention rate, which was the proportion of patients who continued with AEDs at the end of the 12-month treatment period. The secondary clinical outcomes were 12-month remission rates and 50% responder rates. The baseline was the 12-month seizure frequency before AED treatment. If the course of epilepsy was less than 12 months, the frequency of baseline seizures was assessed using the following formula: frequency (times/year)=number of seizures/course of epilepsy (days) × 365. Patients who experienced a reduction in the 12-month seizure frequency of at least 50% were considered 50% responders [15]. Patients who were seizure-free for more than 12 months during the treatment period were said to be in seizure remission [16]. Safety was assessed based on the occurrence of treatment-emergent adverse events [17]. There were two levels of analyzed adverse effects, moderate and serious [18]. When the adverse effects led to AED discontinuation or hospitalization of patients, they were defined as serious adverse effects. The rest were regarded as moderate adverse effects [17].

Statistical analysis

Statistical analysis was performed using SAS v8.2 software (IP: site 0052133001). Baseline characteristics were summarized using descriptive statistics. Retention rates, 50% responder rates and remission rates of AEDs were calculated using Kaplan-Meier survival analysis. Comparisons between the retention, 50% responder rates and remission rates of AEDs were analyzed using Mantel-Haenszel test and chi-square test. For analysis of correlation between remission rates and times of AEDs, Pearson correlation analysis was used. For analysis of risk factors, including gender, age and AED selection for adverse effects, a Cox proportional hazard model was established and log-rank tests were used. Overall odds ratios (OR) with 95% confidence intervals (CIS; adverse effects) were calculated.

Basic information

Results from 292 patients with PEE were analyzed. Table 1 listed the main characteristics of the patients. The follow-up period was 3.0 ± 2.7 years (mean ± standard deviation, ranging from 1 to 18 years). The age at the onset of seizures was 15.7 ± 15.1 years (ranging from 0.1 to 64 years). The patients were divided into three groups according to the age of onset: children (1-13 year-old, n=169, 57.9%), teenagers (14-17 year old, n=28, 9.6%) and adults (18-59 year old, n=93, 31.8%), elders (≥ 60year-old, n=2, 0.7%). It showed PEE manifestations occurred mainly in children .The overwhelming majority of seizure patterns in PEE patients were secondary generalized tonic-clonic seizures (SGTCS) (238 patients or 81.5%); there were very few people in the partial seizure (PS) (54 patients or 18.5%) groups. Therefore, we did not compare efficacy between AEDs for different seizure types in this study. Among these patients, 33 (11.3%) patients had history of febrile seizure, 170 (79.4%) patients had abnormal EEG and 82 (44.6%) patients had positive neuroimaging finds.

Table 1: The demographic, clinical and auxiliary examination characteristics of patients.

Neuroimaging + means: cerebral CT or MRI show abnormalities; - means no visible abnormalities in cerebral CT or MRI.

There were 283 (86.5%) patient visits in the monotherapy groups, including Carbamazepine (CBZ), Phenobarbital (PB), Sodium Valproate (VPA-Na), Magnesium valproate (VPA-Mg), Topiramate (TPM), Lamotrigine (LTG), Phenytoin sodium (PHT), Levetiracetam (LEV) and Oxcarbazepine (OCZ). Since there were too few patient visits resulting in therapy with LTG, PHT, LEV and OCZ, they were not included in the analysis of retention rates, 50% responder rates and remission rates. There were 44 (13.5%) patient visits that resulted in combined therapy due to poor efficacy of monotherapy. In the combined therapy group, the patterns of drug combinations varied; moreover, the patient visits for each combination were too few for statistical comparison. Therefore, no matter what kinds of AEDs were used in the combination therapy, all patients receiving two or more AEDs were assigned to a single combined therapy group. The combined therapy group was compared with other 5 monotherapy groups to see if combined therapy in PEE patients who were refractory to monotherapy could achieve the similar efficacy as other patients who received monotherapy. Thus, there were six groups in the further statistical analysis.

In addition, these patients could be divided into three categories: 58 (19.86%) patients had no seizures in the acute stage of illness; 195 (66.78%) patients had seizures and/or status epilepticus in the acute stage of encephalitis, and there was a latent period between the primary seizure and subsequent epilepsy; and 39 (13.36%) patients showed refractory, repetitive partial seizures persisting from the acute phase to the convalescence phase.

Retention rates

Retention rates for AEDs were shown in Table 2 and illustrated with Kaplan-Meier survival curves in Figure 1. All AED retention rates decreased as follow-up time passed. There were significant differences between different AEDs in the first year (P<0.05). AEDs were listed in descending order of retention rates: CBZ>TPM>PB>VPA-Na>Combined>VPA-Mg. CBZ had significant higher retention rate than VPA-Mg and Combined therapy, and there were no significant differences between rest AEDs. Two years later, there were no significant differences between AEDs (P>0.05).

Figure 1: Kaplan-Meier distribution curve for time between retention rate of different AEDs.

Table 2: Retention rate of different AED

50% responder rate

Follow-up times and 50% responder rates were illustrated in Table 3 and Figure 2. They showed that the 50% responder rates of all the AEDs studied gradually decreased with longer treatment times. No statistically significant differences between AEDs could be identified after years two of follow-up. However, there were statistically significant differences in the first year. In the first year, the 50% responder rates of AEDs, in descending order are: PB>VPA-Mg>VPA-Na>TPM>CBZ>Combined (P<0.05). PB and VPA-Mg had significant greater responder rate than Combined therapy, and there were no significant differences between rest AEDs.

Figure 2: 50% responder rate distribution curve for time between different AEDs.

Table 3: 50% responder rate of individual drug with following-up

Remission rate

The remission rates between AEDs for seizures were illustrated in Table 4 and Figure 3. The results showed that remission rates decreased yearly. A Pearson correlation analysis showed a negative correlation with treatment time (r=0.97, P<0.05). In the first years, there were statistically significant differences in remission rates between AEDs. AEDs were listed in descending order of remission rates: VPA-Mg>TPM>CBZ>PB>VPA-Na>PHT>Combined. TPM and VPA-Mg had the highest remission rates, and significant higher than combined therapy which had the lowest remission rates. In addition, VPA-Na had the lowest remission rates for monotherapy. In the later years of treatment, the six drugs studied showed similar remission rates.

Figure 3: Seizure remission rate with different AEDs for time

Table 4: Remission rate of different drugs in the follow-up periods

Adverse effect

The incidence of adverse effects attributed to AEDs during long-term follow-up are listed in Tab. 5. The main adverse effects (≥ 5% incidence of adverse effect) of the AEDs studied are listed below. The main adverse effects of PB included dizziness (2 patients, 5.6%), gastrointestinal symptoms (2 patients, 5.6%), somnolence (2 patients, 5.6%), fatigue and skin rash. The adverse effects of VPA-Na included gastrointestinal symptoms, fatigue, somnolence, abnormal blood test results, weight gain, dizziness, baldness, extra-pyramidal symptoms and memory impairment. Side effects related to CBZ included skin rash, somnolence, dizziness, abnormal liver function, fatigue, gastrointestinal symptoms, abnormal blood test results (white blood cell decreased), diplopia and memory impairment. The main adverse effects of TPM included gastrointestinal symptoms (3 patients, 6.4%), dizziness, headache, skin rash, somnolence, diplopia, nystagmus, abnormal blood test results and abnormal liver function. The main PHT-associated adverse effects included headache (4 patients, 30.8%), somnolence (3 patients, 23.1%), gingiva hyperplasia (2 patients, 15.4%), attention disturbance (2 patients, 15.4%), flustered (1 patient, 7.7%), fatigue (1 patient, 7.7%) and dyspnea (1 patient, 7.7%). In six patients who used LEV, there were no occurrences of adverse events associated with LEV. One patient in 3 (33.3%) got dermatitis exfoliativa associated with LTG. No adverse events were associated with the AED in 20 patients who took VPA-Mg.

Table 5: Incidence of adverse effects of AEDs

A chi-square test was used to analyze adverse effects between different AEDs. Total, moderate and serious adverse effects had statistically significant differences between AEDs. Patients in the PHT subgroup had the highest incidence of adverse effects, including total, moderate and serious adverse effects, of the six drugs studied. The VPA-Mg had the lowest number of adverse effects. The incidences of adverse effects between CBZ, TPM and VPA showed no significant differences. In decreasing order, the rates for serious adverse effects were: PHT>TPM>CBZ>PB>VPA-Na>Combined>VPA-Mg.

Risk with adverse effects

An established Cox proportional hazard model showed that AED selection was a risk factor for adverse events (Log-rank test statistic=3.98, df=4, p<0.05). In gradual regressive analysis, mated drug comparisons of risk with adverse effects are listed in Table 6.

Table 6: Pairwise drug comparisons of risk with adverse effects

Compared to PB, PHT showed more risk for adverse effects; VPA-Na, CBZ and TPM showed a similar risk for adverse effects. Compared to PHT, three AEDs, VPA-Na, CBZ and TPM showed less risk for adverse effects. Compared to VPA-Na, CBZ and TPM showed a similar risk for adverse effects. TPM was similar to CBZ in its risk for adverse effects.

Encephalitis usually means viral encephalitis. The gold standard for diagnosis of encephalitis is virus isolation in cell culture, and is soon to be replaced by the detection of specific nucleic acids in cerebrospinal fluid (CSF) or brain [19]. However, in some patients with encephalitis, there are no positive findings in CSF culture. Therefore, in the literature presumed encephalitis was defined as having symptoms of an acute febrile illness prior to, or at the time of, the onset of epilepsy and no positive findings in CSF culture [20]. These cases were thus categorized as presumed encephalitis, and all cases in this study fell into this category.

Epileptic seizures may occur not only during the acute stage of encephalitis, but also later, following resolution of the acute phase of illness [21]. In our study, 292 patients with presumed encephalitis developed post-encephalitic epilepsy (PEE). Of these, 234 patients (80.13%) had seizures and/or status epilepticus in the acute stage of encephalitis. These observations were confirmed in studies of PEE reported from California and the Department of Neurology of Aarhus Kommunehospital, which showed that PEE chiefly occurred in patients who had epileptic seizures during the acute phase of the disease, but only rarely in patients who had not had such seizures [22-23].

In our study, 292 patients with PEE were mainly children, and there was no significant difference between the number of males and females; in addition, SGTCS was the predominant seizure type in PEE. We found that PEE manifestations mainly occurred in children, and only rarely in elders. These observations were in agreement with those reported by other investigators [22,24]. SGTCS was the predominant seizure type in PEE patients (81.5%), which was also in agreement with other investigator [8,25].

In this retrospective open study, retention rates were used to assess the overall efficacy and safety. In our study, retention rates followed the general rule that the retention rate of a drug treatment decreased gradually with time [25-28]. Our results show that there were significant differences between different AEDs in the first year (P<0.05). In decreasing order, the retention rates of AEDs were: CBZ>TPM>PB>VPA-Na>Combined>VPA-Mg.

In our study, efficacy was evaluated using 50% responder rates and remission rates. A 50% responder rate indicated a reduction in the 12-month seizure frequency of at least 50%. And 50% responder rates of the AEDs for the first year are listed in descending order: PB>VPA-Mg>VPA-Na>CBZ>TPM>combined therapy. They all showed decent 50% responder rates. The 50% responder rates of VPA-Na, CBZ and TPM nearly conformed to previous studies in patients with newly diagnosed epilepsy [29,30]. The remission rate, the least subjective efficacy measure [31], is the achievement of freedom from seizures and has the greatest impact on quality of life [32]. Both variables decreased with time no matter what kind of AED was used, which was in agreement with other reports [33,34]. The remission rates for AEDs in the first year, in descending order, were: VPA-Mg>TPM>CBZ>PB>VPA-Na>combined therapy.

In decreasing order, the rates for serious adverse effects were: PHT>TPM>CBZ>Combined>PB>VPA-Na>VPA-Mg. Among the total, PHT had significantly more adverse effects (both moderate and serious adverse) than other AEDs. The fact that PHT had the most frequent serious adverse effects was suggested by previous trials [35]. So, we do not recommend using PHT for PEE patients.

Compared to CBZ, TPM had similar retention rates (68.6% vs. 59.6%), 50% responder rate (82.9% vs. 83.0%), and remission rate (60.0% vs. 72.3%) in the first year and adverse effects (pairwise drug comparisons of risk: TPM to CBZ 0.99). As to retention rates, there were no significant differences between CBZ and TPM. Meanwhile, CBZ showed similar retention rates in our study compared with previous studies in patients with new onset epilepsy [36]. The retention rate of TPM (59.6%) showed no difference compared with other reports (40-60%) in patients with epilepsy after 1 year [37]. However, our researchers discovered lower retention rates after 2 years (from 2 to 4 years: 34.0, 19.2 and 12.8%) for TPM than reported by other investigators (45, 38 and 30%, respectively) [33]. It may be because we only looked at TPM used to treat PEE. TPM showed higher 50% responder rates and remission rates than CBZ, but there were no significant differences. And the remission rates for TPM and CBZ were agreed with other reports [31,38]. The pairwise drug comparisons of risk with adverse effects (0.99, p>0.05) showed that TPM had less risk with adverse effects than CBZ but without significant difference. The incidence of serious adverse events for TPM was 10%, including diplopia, serious skin rash, abnormal blood test results and damage to liver function, which was similar to previous clinical trials (11%-28%, Reife et al.) [39]. In patients receiving CBZ, 8.1% of cases presented serious adverse effects, including skin rash, bone marrow depression and damage to liver function. This was similar to other reports (3-9%) [35,40]. Therefore, we suggest that TPM is a good alternative for PEE patients.

Compared to CBZ, VPA-Na had similar retention rates (68.6% vs. 57.5%), 50% responder rate (82.9% vs. 83.9%), remission rate (60.0% vs. 52.9%) in the first year and adverse effects (pairwise drug comparisons of risk: CBZ to VPA-Na 1.70). In our study, CBZ and VPA-Na both showed good retention rate and remission rates that agreed with other reports in patients with epilepsy [31,38,41]. And VPA-Na showed lower retention rate and remission rate compared to CBZ, but there were no significant differences between the two drugs. VPA-Na showed less risk with adverse effects than CBZ. Especially in incidence of serious adverse events, VPA-Na showed lower incidences of serious adverse events (4.9%), including abnormal blood test results, baldness, extra-pyramidal symptoms and memory impairment. And compared to previous reports in patients with epilepsy (3-8%) [38,42], adverse effects of VPA-Na were similar. Combining the indices above, we propose that VPA-Na should be a better choice than CBZ for PEE patients.

Compared to CBZ, VPA-Mg had similar retention rates (68.6% vs. 42.1%), 50% responder rate (82.9% vs. 90%), and remission rate (60.0% vs. 75.0%) in the first year. VPA-Mg is one kind of salt of valproic acid which has anticonvulsant activity. It has not only been used for a long time in China, but also been studied in other contries, such as Italy [40-45], Mexico [44] and Canada [45]. Moreover it got high retention rate and low incidence of side effects [44]. In our study, it had a better 50% responder rate and remission rate than CBZ. However, it had a lower retention rate (42.1%) than CBZ (68.6%). Maybe the inconvenience and the marketing of VPA-Mg affected the retention rate. VPA-Mg should be prescribed three times a day. And CBZ is used wider than VPA-Mg in PEE patients, it gets more experience, and gets more approval by physician. So we got inconsistent conclusions. In our study, 18 patients that were prescribed VPA-Mg never complained of any adverse effects. A non-blinded study of 320 patients in Italy showed VPA-Mg had high 50% responder rate (95.2%), and only 4.3% of the adverse events were considered to be serious [46]. But our sample size for VPA-Mg was small, it was worthy of larger sample volumes of further tracking.

Compared to CBZ, PB had similar retention rates (68.6% vs. 58.3%), 50% responder rate (82.9% vs. 94.4%), remission rate (60.0% vs. 58.3%) in the first year, and adverse effects (pairwise drug comparisons of risk: CBZ to PB 1.13). PB showed lower retention rate than CBZ, which was lower than in another report in patients with newly diagnosed epilepsy [29]. Based on the 50% responder rates from the first and second years, PB was the best for treating PEE among all the AEDs we analyzed. And the rate (94.4%) was a little higher than other reports for PB (Placencia et al., 54%) [47]. However, the remission rate for PB, which matched previous reports was lower than CBZ, which led to low retention of PB. This may be because we studied PEE patients aged from 1 to 65 years old, while Placencia et al. only studied epilepsy patients aged from 6 to 65 years old [47]. In our study, after treatment with PB, 39 patients with barely-controlled refractory, repetitive partial seizures in the acute stage experienced a gradually decreased seizure frequency, but seizures were not completely eliminated. This state continuously evolved into PEE without a latent period. And PB showed less risk with adverse effects than CBZ. Combining the indices above, we suppose that PB is worse than CBZ. But PB has low cost, and in rural areas of some countries, including China [48], France [49], Kenya et al. [50], especially in China, there were a great number of studies for PB, and it showed good effect and has no severe side effect for epilepsy patients [48]. So for PEE patients, we suggest PB was also a good alternative.

In our study, combined therapy had the lowest efficacy. Compared to responder rate (82.9% vs. 70.5%), remission rate (60.0% vs. 29.6%) in the first year. A rational approach is commonly advocated: If two AEDs are used together, their combined effect will be greater if the known mechanisms of action are different [51]. However, there is no evidence in human beings to suggest that one particular combination of AEDs is more efficacious than another [39]. Our study showed that, in PEE patients who were refractory to monotherapy, the combined therapy didn’t achieved the similar efficacy as other patients who received monotherapy.

Considering the retention rates, remission rates, 50% responder rates and adverse effects together, we suggest that, compared to CBZ, TPM and VPA-Na should be the good alternatives for PEE patients. When considering the cost together, PB should be another choice in PEE. It is worthy to enlarge the sample size of VPA-Mg to further confirm its good efficacy and safety. And PHT was not recommended for PEE patients.

We thank Yida Hu from Department of Neurology, the First Affiliated Hospital of Chongqing Medical University, for his assistance in the initial data analysis.