Abstract
Purpose
Treatment for epilepsy primarily involves antiseizure medications (ASMs), which can be characterized using the clinical data warehouse (CDW) database. In this study, we compared retention rates and time to successful treatment for various ASMs to reflect both efficacy and adverse effects in patients with newly diagnosed epilepsy.
Materials and Methods
We identified newly diagnosed epilepsy patients with ASM treatment for more than 12 months using CDW of a tertiary referral hospital. Clinical characteristics were compared between groups with successful and unsuccessful treatment. Cox regression analysis was performed to evaluate independent variables of age, sex, comorbidities, and attributes of ASM regimens.
Results
Of 2515 eligible participants, 46.2% were successfully treated with the first ASM regimen, and 74.7% with all ASM regimens with the median time-to-treatment success of 14 months. Participants with second-generation ASM as the first ASM were more likely to be successfully treated with the first regimen compared to those with first-generation ASM (51.6% vs. 42.3%, p<0.001) and more successfully treated [hazard ratio (HR)=1.26; 95% confidence interval (CI): 1.15–1.39]. Overall, valproic acid was the most common ASM across a wide range of ages under 65 years, while levetiracetam in patients aged over 65 years or lamotrigine in female adult patients. Clinical factors associated with less favorable treatment outcomes included renal disease (HR=0.78; 95% CI: 0.66–0.92), liver disease (HR=0.65; 95% CI: 0.52–0.81), depression (HR=0.70; 95% CI: 0.57–0.84), and mechanical ventilation (HR=0.58; 95% CI: 0.50–0.67).
Conclusion
Second-generation ASMs have the advantage of more successful treatment with fewer ASM regimen changes compared with first-generation drugs. Various comorbid conditions as well as age and sex should be considered when selecting ASMs.
Keywords: Epilepsy, seizure, treatment outcome
Graphical Abstract
INTRODUCTION
Epilepsy is a neurological disorder with a persistent predisposition to generate epileptic seizures.1 In a meta-analysis, the incidence of epilepsy was 61.44 per 100000 person-years and the lifetime prevalence of epilepsy was 7.60 per 1000 population.2 Although some patients with epilepsy may be treated by surgery (e.g., anterior temporal lobectomy), electrical stimulation (vagal nerve stimulation or deep brain stimulation), or ketogenic diet, the mainstay of epilepsy treatment is antiseizure medications (ASMs).
Since the introduction of bromide as an ASM, more than 40 ASMs have been licensed and prescribed for patients with epilepsy.3 However, first-generation ASMs had limited mechanisms of action. For example, phenytoin (PHT) and carbamazepine (CMZ), which are among the most commonly used first-generation ASMs, inhibit voltage-gated sodium channel. Although valproic acid (VPA), another common ASM, has multiple mechanisms, including inhibition of voltage-gated sodium channel and increase of γ-aminobutyric acid (GABA), patients with refractory epilepsy had no choice other than using multiple sodium channel blockers since there were limited number of ASMs in first generation era. Combining ASMs with common mechanism of action increases adverse effect, which leads to suboptimal seizure control.4 Also, many first-generation ASMs have enzyme-inducing or enzyme-inhibiting effects. This complicates the treatment of patients who receive more than two ASMs or have other comorbidities (e.g., chemotherapy for patients with cancer).
Fortunately, many new generation ASMs with diverse mechanisms have been launched since the 1990s. Lamotrigine (LMT), oxcarbazepine (OXC), and lacosamide (LCS) are sodium channel blockers; levetiracetam (LEV) binds to synaptic vesicle protein 2A; perampanel inhibits post-synaptic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; gabapentin (GBP) and pregabalin (PGB) is a ligand of the α2δ calcium channel subunit. In addition to various mechanisms of action, most new generation ASMs do not exhibit drug interactions, and treating patients with polytherapy or with comorbidities has been much simpler with new generation ASMs.
Although more than 13 ASMs have been available for the last 20 years, the overall outcome for seizure control has not markedly improved.5 Once individuals are newly diagnosed with epilepsy, about half achieve seizure-free status with the first ASM regimen and nearly 20% become seizure-free after receiving two or more ASM regimens. However, almost 30% of patients with newly diagnosed epilepsy suffer from recurrent seizures despite numerous polytherapy trials, and this proportion has only slightly decreased even after introduction of many new generation ASMs.
Patient’s comorbidities or medical conditions affect the selection of ASMs. Dosing of LEV, GBP, or PGB should be reduced in patients with renal impairment.6,7 VPA should be prescribed with reduced dosing or avoided for patients with hepatic impairment.7 LEV or perampanel should be used with caution for patients with psychiatric disorders.8,9,10 VPA should be avoided in females of childbearing age due to teratogenicity.11,12 Adverse effect profiles of new generation ASMs are not necessarily superior to those of first generation ASMs, but the development of new ASMs has provided alternative treatment options for patients who experienced adverse effects from first-generation ASMs.
In a recent multinational prospective cohort study, a deep learning model for predicting epilepsy control with seven initial ASM monotherapy regimens was created.13 Among 16 clinical factors, the number of seizures prior to treatment, presence of psychiatric disorders, electroencephalographic findings, and brain imaging findings were the most important variables, and this model showed a predictive power of about 0.52 to 0.65. However, outcomes were followed-up for only the first 1 year of the initial ASM treatment, and long-term outcomes of ASM regimen including combination therapy were not evaluated. Moreover, the difference in ASM generation was not considered, and the concept of time-to-treatment success was also not included.
Therefore, ASM treatment outcome is determined by a combination of efficacy, drug interactions, comorbid conditions, etc. Although the efficacy of new ASMs has not dramatically increased, favorable adverse effect profiles and fewer drug interactions may yield better overall outcomes due to continued release of new ASMs. Using a clinical data warehouse (CDW) database, we investigated specific clinical characteristics associated with successful treatment of epilepsy, including sequential combination of ASM and time-to-treatment success.
In this study, CDW data were used to investigate a series of changes in ASM combinations in individual patients over long follow-up periods. In addition, detailed clinical factors related to successful treatment were identified, including various comorbidities.
MATERIALS AND METHODS
Data source
This study utilized the CDW of Ewha Womans University Mokdong Hospital, which provides an Electronic Medical Records healthcare database of all patients visiting the hospital from January 2001 to December 2019. This anonymized and standardized database includes clinical laboratory test results, patient demographics, pharmacy information, hospital admission data, discharge and transfer dates, and International Classification of Diseases, 10th Revision (ICD-10) codes for diagnosis. All databases have been transformed into the OMOP CDM, version 5. The full specification for OMOP CDM, version 5 can be accessed at http://github.com/OHDSI/CommonDataModel. Study participants were identified from the CDW database of Ewha Womans University Mokdong Hospital. Patients with ≥2 diagnoses of epilepsy (ICD-10: G40.x) or status epilepticus (ICD-10: G41.x) with ASM prescriptions were selected. Index date was defined as the first day of diagnosis of epilepsy or status epilepticus with ASM prescriptions. Only participants prescribed one ASM on the index date were included. This removed referred cases who had been treated with ASMs in other centers. Participants who were treated for more than 12 months were included as this study assessed 1-year retention rate. This study was approved by the Institutional Review Board of Ewha Womans University Mokdong Hospital (EUMC 2021-11-038).
Variable definitions
Age at the index date was used as an estimate for onset age. Follow-up duration was calculated as the time interval between the index date and the last day of ASM prescription. Comorbidities from the index date to the last day of follow-up were identified using ICD-10 codes.14 Charlson comorbidity index was calculated based on these comorbidities. Since the follow-up duration varied by patient, the duration of admission was calculated as days of hospital admission divided by follow-up duration in years. Participants treated with mechanical ventilation or inotrope were investigated to identify severe cases.
ASMs approved in South Korea before the year 2000 were classified as first-generation ASMs (VPA, CMZ, PHT, fosphenytoin, phenobarbital, vigabatrin, topiramate, zonisamide, ethosuximide, and clobazam), while those approved since 2000 were classified as new-generation ASMs (LEV, OXC, LMT, GBP, PGB, LCS, perampanel, and clonazepam). ASM regimen refers to a combination of all ASMs prescribed at some point during the follow-up period, irrespective of daily dose. All participants in this study were prescribed one ASM at the beginning of treatment (monotherapy). In cases of unsuccessful treatment (seizure relapse) or drug-related adverse event, new ASMs were introduced and/or previous ASMs were discontinued. Therefore, some participants were treated with ≥2 ASMs (polytherapy) later in the follow-up period. If the ASM regimen was maintained unchanged for ≥1 year until the last day of follow-up, the participant was considered to be successfully treated. The time-to-treatment success was defined as the duration from the index date to the last instance where the ASM regimen had been consistently maintained for more than 1 year. A participant was considered to be unsuccessfully treated if the duration of the last ASM regimen was less than 1 year after ASM regimen change.
Statistical analysis
Continuous variables are summarized as the mean±SD, and counts (N) and percentages (%) are used for categorical variables. The mean differences between successful treatment group and unsuccessful treatment group were assessed by Student’s t-test for normally distributed continuous variables and Mann Whitney U test for skewed continuous variables. For categorical variables, Pearson’s χ2 test was performed to compare the proportions between the two groups.
We applied a Cox proportional hazard model to estimate the association between explanatory variables, including age, sex, comorbidities, and characteristics of ASM regimens, with treatment success while using ASMs. Unsuccessfully treated patients were censored when they were no longer prescribed ASMs according to the CDW database. A hazard ratio (HR) for each variable was calculated with an associated 95% confidence interval (CI) and a statistical significance threshold of p<0.05. Proportional hazard assumptions were evaluated using log (-log) plots, and no significant violations were observed for the outcome variable of treatment success (Supplementary Fig. 1, only online). Adjusted cumulative incidence curves for significant explanatory variables from the multivariate Cox model were plotted. All statistical analyses were performed using SAS statistical software version 9.4 (SAS Institute Inc., Cary, NC, USA).
RESULTS
Comparison between successful and unsuccessful treatment groups
A total of 2515 patients meeting the inclusion/exclusion criteria were identified. Among them, 1878 (74.7%) were successfully treated with ASM treatment for the last year of follow-up, and 637 (25.3%) were unsuccessfully treated (Fig. 1).
Fig. 1. Flow chart of study participants. ICD-10, International Classification of Diseases, 10th Revision; EUMC, Ewha Womans University Medical Center; ASM, antiseizure medication.
Comparison between successful and unsuccessful treatment groups is shown in Table 1. The successful treatment group was younger (37.8±23.2 years) than the unsuccessful treatment group (42.2±21.5 years). Sex and follow-up duration did not show significant difference between the two groups. The proportion of successfully treated patients remained unchanged across index year groups. Charlson comorbidity index was higher in the unsuccessful treatment group. Hypertension, diabetes, ischemic heart disease, renal disease, liver disease, any malignancy, and depression were more common in the unsuccessful treatment group.
Table 1. Characteristics of Patients with Successfully and Unsuccessfully Treated Epilepsy.
| Variables | Successful treatment (n=1878) | Unsuccessful treatment (n=637) | p value | ||
|---|---|---|---|---|---|
| Age at index date (yr) | 37.8±23.2 | 42.2±21.5 | <0.001* | ||
| Sex | 0.894† | ||||
| Male | 976 (52.0) | 333 (52.3) | |||
| Female | 902 (48.0) | 304 (47.7) | |||
| Follow-up duration (yr) | 6.0±4.8 | 6.1±4.8 | 0.978‡ | ||
| Index year | 0.963† | ||||
| 2001–2005 | 646 (34.4) | 216 (33.9) | |||
| 2006–2010 | 507 (27.0) | 172 (27.0) | |||
| 2011–2015 | 442(23.5) | 156 (24.5) | |||
| 2016–2019 | 283 (15.1) | 93 (14.6) | |||
| Comorbidities | |||||
| Hypertension | 322 (17.1) | 141 (22.1) | 0.005† | ||
| Diabetes mellitus | 239 (12.7) | 104 (16.3) | 0.022† | ||
| Dyslipidemia | 2 (0.1) | 1 (0.2) | >0.999§ | ||
| Ischemic heart disease | 81 (4.3) | 42 (6.6) | 0.021† | ||
| Renal disease | 186 (9.9) | 107 (16.8) | <0.001† | ||
| Liver disease | 83 (4.4) | 60 (9.4) | <0.001† | ||
| Malignancy | 91 (4.8) | 60 (9.4) | <0.001† | ||
| Dementia | 153 (8.1) | 66 (10.4) | 0.087† | ||
| Depression | 107 (5.7) | 69 (10.8) | <0.001† | ||
| Charlson comorbidity index | <0.001† | ||||
| 0 | 1128 (60.1) | 286 (44.9) | |||
| 1 | 363 (19.3) | 147 (23.1) | |||
| ≥2 | 387 (20.6) | 204 (32.0) | |||
| Generation of first ASM | 0.432† | ||||
| First generation | 1081 (57.6) | 378 (59.3) | |||
| Valproic acid | 548 (29.2) | 183 (28.7) | |||
| Carbamazepine | 314 (16.7) | 87 (13.7) | |||
| Phenytoin | 106 (5.6) | 47 (7.4) | |||
| Others¶ | 113 (6.0) | 61 (9.6) | |||
| Second generation | 797 (42.4) | 259 (40.7) | |||
| Levetiracetam | 355 (18.9) | 105 (16.5) | |||
| Oxcarbazepine | 180 (9.6) | 40 (6.3) | |||
| Lamotrigine | 83 (4.4) | 18 (2.8) | |||
| Others∥ | 179 (9.5) | 96 (15.1) | |||
| Number of ASM regimens | <0.001† | ||||
| 1 | 1162 (61.9) | 0 (0.0) | |||
| 2 | 181 (9.6) | 123 (19.3) | |||
| 3 | 187 (10.0) | 120 (18.8) | |||
| ≥4 | 348 (18.5) | 394 (61.9) | |||
| Number of ASMs in the last regimen | <0.001† | ||||
| 1 | 1679 (89.4) | 395 (62.0) | |||
| 2 | 148 (7.9) | 173 (27.2) | |||
| ≥3 | 51 (2.7) | 69 (10.8) | |||
| Duration of admission (days/year) | 0.9±3.3 | 4.2±19.4 | <0.001‡ | ||
| Mechanical ventilation | 250 (13.3) | 218 (34.2) | <0.001† | ||
| Inotropes | 45 (2.4) | 49 (7.7) | <0.001† | ||
ASM, antiseizure medication.
Data are presented as mean±standard deviation or n (%).
*Student’s t-test; †Chi-square test; ‡Mann-Whitney U test; §Fisher’s exact test; ¶Fosphenytoin, phenobarbital, vigabatrin, topiramate, zonisamide, ethosuximide, and clobazam; ∥Gabapentin, pregabalin, lacosamide, perampanel, and clonazepam.
The prescription trend of anti-seizure medication (ASM) exhibited a gradual change over the course of the study period. The utilization of older generation ASMs, such as VPA, CMZ, topiramate, and PHT, exhibited a gradual decrease. Conversely, there was a progressive increase in the utilization of new generation ASMs, including LEV, OXC, GBP, LMT, clonazepam, and PGB (Supplementary Fig. 2, only online).
During the follow-up period, the unsuccessful treatment group experienced more ASM regimen changes compared to the successful treatment group. In the successful treatment group, 61.9% of patients were successfully treated with the first monotherapy regimen, and change of regimen was not required until the last day of follow-up. Meanwhile, all unsuccessfully treated patients, by definition, tried more than one ASM regimens. On the other hand, the proportion of patients who tried four or more ASM regimens throughout the follow-up period were higher in the unsuccessful treatment group than in the successful treatment group (61.9% vs. 18.5%). In terms of the number of ASMs on the last day of follow-up, the unsuccessfully treated patients were using a greater number of ASMs compared to successfully treated patients. In the successful treatment group, 89.4% was treated with monotherapy, 7.9% was treated with dual therapy, and only 2.7% was treated with three or more ASMs. Among the unsuccessful treatment group, however, 62.0% was treated with monotherapy, 27.2% was treated with dual therapy, and 10.8% was treated with three or more ASMs. The unsuccessful treatment group was hospitalized longer and more likely to be treated with mechanical ventilation or inotropes compared to the successful treatment group.
Successful treatment in patients who started new ASM as initial therapy
Of the total study cohort, 46.2% was successfully treated with the first regimen and the median time-to-treatment success was 14 months. The proportion and number of additional successfully treated patients decreased, but a significant number of patients were successfully treated as ASM regimen trials proceeded (Fig. 2). Of the total study cohort, 1769 (70.3%) were successfully treated within the first seven regimens, while 109 (4.3%) were successfully treated with eight or more regimen trials. First-generation ASMs were prescribed in 1459 (58.0%) patients as the first regimen (older ASM group), and 1056 (42.0%) patients were prescribed new-generation ASMs as first regimen (new ASM group).
Fig. 2. Numbers of patients and cumulative rates of successful treatment with successive antiseizure medication (ASM) regimens. Light gray bar indicates the number of all patients receiving each ASM regimen, while dark gray bar represents successfully treated patients receiving that regimen. Gray dotted line shows cumulative rates of successful treatment of all patients, solid line with black circles is patients with first generation of ASM as the first regimen, and solid line with black triangles is patients with new generation of ASM as the first regimen.
The cumulative rate of successful treatment until the last day of follow-up was not significantly different between the new ASM group and older ASM group (75.5% vs. 74.1%, p=0.432). However, the rate of successful treatment with a first ASM regimen was higher in the new ASM group than in the older ASM group (51.6% vs. 42.3%, p<0.001). Cumulative rates of successful treatment remained significantly higher in the new ASM group than in the older ASM group from second to fifth regimens (p<0.05). Cumulative rates of successful treatment eventually demonstrated no significant difference between the new and older ASM group for sixth or later ASM regimens (p>0.05).
Fig. 3 shows first ASM selection by age group and sex. VPA was the most common ASM used before 65 years of age. The use of LEV increased in adults; this was the most frequently used ASM in older adults. VPA was more frequently prescribed to males than to females 0–64 years of age. However, LMT was mostly prescribed for ages 7–64 years, with remarkable female dominance. The use of CMZ peaked in younger adults (age 18–64), then decreased in older adults (age ≥65 years). GBP was chosen almost exclusively for adults.
Fig. 3. First antiseizure medication (ASM) selection by age and sex groups. Light gray bar indicates the proportion of female patients receiving each ASM as the first regimen in each age group, while dark gray bar represents male patients. VPA, valproic acid; CMZ, carbamazepine; LEV, levetiracetam; LMT, lamotrigine; GBP, gabapentin.
Variables associated with successful treatment
Cox proportional hazard regression analysis was performed to evaluate successful treatment (Table 2). In multivariate analysis, older adults were more successfully treated than younger adults (HR=1.25; 95% CI: 1.09–1.43), and the new ASM group was more successfully treated than the older ASM group (HR=1.26; 95% CI: 1.15–1.39) (Fig. 4A). Patients on two ASMs (HR=0.39; 95% CI: 0.32–0.46) and three or more ASMs (HR=0.28; 95% CI: 0.21–0.36) were less successfully treated than patients on monotherapy at the last follow-up (Fig. 4B). Patients with renal disease (HR=0.78; 95% CI: 0.66–0.92) (Fig. 4C), liver disease (HR=0.65; 95% CI: 0.52–0.81) (Fig. 4D), depression (HR=0.70; 95% CI: 0.57–0.84) (Fig. 4E), and history of mechanical ventilator treatment (HR=0.58; 95% CI: 0.50–0.67) (Fig. 4F) were less successfully treated.
Table 2. Risk Factors Associated with Successful Treatment.
| Variables | Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|---|
| HR | 95% CI | p value | HR | 95% CI | p value | ||
| Age group (yr) | <0.001 | 0.008 | |||||
| 0–6 | 1.29 | 1.04–1.57 | 1.17 | 0.95–1.43 | |||
| 7–17 | 1.18 | 1.06–1.33 | 1.05 | 0.93–1.18 | |||
| 18–64 | REF | REF | |||||
| ≥65 | 1.24 | 1.09–1.40 | 1.25 | 1.09–1.43 | |||
| Sex | 0.438 | ||||||
| Male | 1.04 | 0.95–1.14 | |||||
| Female | REF | ||||||
| Generation of first ASM | <0.001 | <0.001 | |||||
| First generation | REF | REF | |||||
| Second generation | 1.31 | 1.19–1.43 | 1.26 | 1.15–1.39 | |||
| Number of ASMs in the last regimen | <0.001 | <0.001 | |||||
| 1 | REF | REF | |||||
| 2 | 0.38 | 0.32–0.45 | 0.39 | 0.32–0.46 | |||
| ≥3 | 0.27 | 0.20–0.35 | 0.28 | 0.21–0.36 | |||
| Comorbidities | |||||||
| Hypertension | 0.81 | 0.72–0.91 | 0.001 | 0.94 | 0.83–1.08 | 0.387 | |
| Diabetes mellitus | 0.79 | 0.69–0.90 | 0.001 | 0.94 | 0.82–1.08 | 0.397 | |
| Ischemic heart disease | 0.76 | 0.60–0.94 | 0.014 | 0.89 | 0.70–1.12 | 0.334 | |
| Renal disease | 0.72 | 0.62–0.84 | <0.001 | 0.78 | 0.66–0.92 | 0.004 | |
| Liver disease | 0.59 | 0.47–0.73 | <0.001 | 0.65 | 0.52–0.81 | <0.001 | |
| Malignancy | 0.72 | 0.62–0.84 | <0.001 | 0.98 | 0.78–1.21 | 0.830 | |
| Depression | 0.65 | 0.54–0.79 | <0.001 | 0.70 | 0.57–0.84 | <0.001 | |
| Duration of admission | <0.001 | ||||||
| 0 | REF | ||||||
| ≤1 week per year | 0.85 | 0.74–0.97 | |||||
| >1 week per year | 0.57 | 0.42–0.76 | |||||
| Mechanical ventilation | 0.52 | 0.45–0.59 | <0.001 | 0.58 | 0.50–0.67 | <0.001 | |
| Inotropes | 0.52 | 0.38–0.69 | <0.001 | ||||
HR, hazard ratio; CI,confidence interval; ASM, antiseizure medication; REF, reference.
Fig. 4. Cumulative probabilities of successful treatment by multivariate analysis using Cox proportional model. Generation of first ASM (A), number of ASMs in the last regimen (B), renal disease (C), liver disease (D), depression (E), use of mechanical ventilation (F) adjusted for age group, generation of first ASM, number of ASMs in the last regimen, use of mechanical ventilation, use of inotropes, and underlying hypertension, diabetes mellitus, ischemic heart disease, renal disease, liver disease, malignancy, and depression. Shaded areas represent 95% confidence intervals. ASM, antiseizure medication.
DISCUSSION
In this study, 74.7% (1878/2515) of newly diagnosed epilepsy patients were successfully treated with a stable ASM regimen including combination therapy with the median time-to-treatment success of 14 months. Of 2515 eligible participants, 46.2% (1162) participants were successfully treated with the first ASM regimen. Participants with a new-generation ASM as the first ASM were more likely to be successfully treated with the first regimen compared to those with first-generation ASM (51.6% vs. 42.3%, p<0.001) and more successfully treated in Cox proportional hazard regression analysis (HR=1.26; 95% CI: 1.15–1.39). LEV was the most common ASM as a first choice in older adults, while VPA was the most common ASM in younger age (0–64 years of age), especially in males. On the other hand, LMT was more frequently prescribed to females than to males aged 7–64 years. Participants with comorbidities, such as renal disease, liver disease, or depression, were less successfully treated.
In a recent study, a deep learning model predicted seizure control rate of the first ASM monotherapy regimen over the first year of treatment.13 However, the area under receiver operating characteristic curve was modest, ranging from 0.52 to 0.65. In our study, ASM selection and sequential combination of ASM regimens in each patient were analyzed in detail in patients of various age groups. Out of 2515 patients, 46.2% were successfully treated with the first ASM regimen, and 74.7% were successfully treated with all ASM regimens including combination therapy with the median time-to-treatment success of 14 months (95% CI 13–17 months). These results were consistent with previous studies that reported a seizure-free rate of 60%–70% among newly diagnosed epilepsy patients and 45%–50% with the first ASM regimen.15,16
Despite the introduction of numerous new ASMs over the last decade, the seizure-free rate has not increased dramatically.16 Similarly, our study did not show significant improvement in the rate of successful treatment according to index year groups. On the last day of follow-up, the cumulative rate of successful treatment did not show significant difference according to the generation of the first ASM. However, the older ASM group and new ASM group showed significantly different rates of successful treatment with successive ASM regimens. The new ASM group was more likely to be successfully treated with the first ASM regimen compared to the older ASM group. The older ASM group, however, was more successfully treated with the second or later ASM regimen compared to the new ASM group. At the end of follow-up, the overall rate of successful treatment of the older ASM group was the same as that of the new ASM group.
Notably, the new ASM group was successfully treated with fewer ASM regimens and, thus, was successfully treated faster than the older ASM group. Although epileptologists aim to choose the best ASM for newly diagnosed epilepsy patients, more than half of new patients experience seizure recurrence. Although some of these recurrences happen in the process of ASM dose titration, a single episode of seizure recurrence may cause physical, psychological, social, and occupational disabilities for patients with newly diagnosed epilepsy. Therefore, it is beneficial to minimize seizure recurrence with new ASMs in the initial period of treatment.
Consistent with previous studies, unsuccessfully treated patients received more ASMs and had more comorbidities.17,18 Especially, renal disease, liver disease, and depression were associated with unsuccessfully treated epilepsy. Hypertension, diabetes, ischemic heart disease, and malignancy seemed to be associated with unsuccessfully treated epilepsy in univariate analysis, but these associations were not significant in multivariate analysis.
Seizures are affected by chronic kidney disease (CKD), as is ASM treatment.6,7 The incidence of uremic seizures with kidney failure is ~10%. These seizures are often non-convulsive and may mimic uremic encephalopathy. Furthermore, patients with CKD encounter challenges in the selection, loading, titration, and maintenance of ASMs due to potentially altered pharmacokinetics.
Liver cirrhosis is another condition often associated with seizures. In a previous study, epileptiform discharge was observed in 15% of patients with hepatic encephalopathy, and 10% of these had clinical seizures.19 The prevalence of status epilepticus in patients with cirrhosis was estimated to be 0.7%.20 As in CKD, impaired hepatic clearance complicates the selection of ASMs in patients with severe liver disease, such as liver cirrhosis.
An association between depression and seizures has been reported by many researchers.21,22 Adults with active epilepsy were three times more likely to report depression or anxiety than those without epilepsy.21 Furthermore, the severity of epilepsy was associated with severity of depression.23 A hyperactive hypothalamic-pituitary-adrenal axis and disturbance of neurotransmitters (glutamate and GABA) have been suggested as underlying mechanisms linking depression and epilepsy.10
This study had several limitations. First, it was based on claims data; therefore, we defined successfully treated epilepsy based on stable ASM prescription and the assumption that patients with intractable epilepsy would continually receive new ASM regimens. There is a possibility that some of the successfully treated patients in this study might be intractable cases. Still, stable ASM regimen reflects good tolerance and no abrupt change in patient status. Second, dose changes within the same regimen could not be identified due to the unavailability of ASM doses in the current CDW database. In a recent study that assessed the outcomes of ASM regimens using claims data, an unsuccessful regimen was defined as “any change other than a dose change (i.e., increase/decrease) in the subsequent 1–12 months after the change.”24 We adopted a similar approach in this study. However, it is important to note that “successful treatment” in this study may be overestimated compared to actual seizure-freedom. Third, this was a retrospective cohort study from a single healthcare center, so the results in this study may be biased due to ASM preferences of physicians. Validation using a multi-center common data model database or a prospective cohort study is warranted to evaluate the impact of ASM selection on epilepsy outcome. However, this study included 2515 patients over two decades. Also, completeness of our cohort and long follow-up duration (6.0±4.8 years) were strengths of this study. The adoption of survival analysis to evaluate treatment success in each subgroup was another strength of this study.
In conclusion, our study suggests that second generation ASMs have an advantage of more successful treatment with less ASM regimen changes compared to first generation ASMs. Various comorbid conditions as well as age and sex should be considered when selecting ASMs. Carefully designed studies are needed to confirm these findings and to determine if the beneficial effect of first generation ASMs results from superior efficacy or favorable adverse effect profiles.
ACKNOWLEDGEMENTS
This study was supported by grants from the Basic Science Research Program, Convergent Technology R&D Program for Human Augmentation, and BK21 Plus Program through the National Research Foundation of Korea (NRF-2019M3C1B8090803, 2020R1A2C2013216, and RS-2023-00265524), the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number : HI23C1532), and Institute of Information & communications Technology Planning & Evaluation (IITP) grant (No. RS-2022-00155966) by the Korean government (MSIT) and the Artificial Intelligence Convergence Innovation Human Resources Development programs of Ewha Womans University.
Footnotes
The authors have no potential conflicts of interest to disclose.
- Conceptualization: Hyungmi An and Hyang Woon Lee.
- Data curation: Sungeun Hwang, Hyungmi An, and Hyang Woon Lee.
- Formal analysis: all authors.
- Funding acquisition: Hyang Woon Lee.
- Investigation: all authors.
- Methodology: Sungeun Hwang, Hyungmi An, and Dong Woo Shin.
- Project administration: Hyang Woon Lee.
- Resources: Hyungmi An and Hyang Woon Lee.
- Software: Hyungmi An.
- Supervision: Hyang Woon Lee.
- Validation: Hyungmi An and Dong Woo Shin.
- Visualization: Hyungmi An and Dong Woo Shin.
- Writing—original draft: Sungeun Hwang.
- Writing—review & editing: Hyungmi An, Dong Woo Shin, and Hyang Woon Lee.
- Approval of final manuscript: all authors.
SUPPLEMENTARY MATERIALS
Log negative-log survival curve plots to assess the proportional hazard assumption of age group, generation of first antiseizure medication (ASM), number of ASMs in the last regimen, use of mechanical ventilation, and underlying hypertension, diabetes mellitus, ischemic heart disease, renal disease, liver disease, any malignancy, and depression.
Antiseizure medication (ASM) prescription trend during the study period. Use of first generation ASMs (e.g., valproic acid, carbamazepine, topiramate, phenytoin) gradually decreased and use of new generation ASMs (e.g., levetiracetam, oxcarbazepine, gabapentin, lamotrigine, clonazepam, pregabalin) gradually increased.
References
- 1.Fisher RS, Acevedo C, Arzimanoglou A, Bogacz A, Cross JH, Elger CE, et al. ILAE official report: a practical clinical definition of epilepsy. Epilepsia. 2014;55:475–482. doi: 10.1111/epi.12550. [DOI] [PubMed] [Google Scholar]
- 2.Fiest KM, Sauro KM, Wiebe S, Patten SB, Kwon CS, Dykeman J, et al. Prevalence and incidence of epilepsy: a systematic review and meta-analysis of international studies. Neurology. 2017;88:296–303. doi: 10.1212/WNL.0000000000003509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Rho JM, White HS. Brief history of anti-seizure drug development. Epilepsia Open. 2018;3(Suppl 2):114–119. doi: 10.1002/epi4.12268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.French JA, Faught E. Rational polytherapy. Epilepsia. 2009;50(Suppl 8):63–68. doi: 10.1111/j.1528-1167.2009.02238.x. [DOI] [PubMed] [Google Scholar]
- 5.Brodie MJ. Antiepileptic drug therapy the story so far. Seizure. 2010;19:650–655. doi: 10.1016/j.seizure.2010.10.027. [DOI] [PubMed] [Google Scholar]
- 6.Títoff V, Moury HN, Títoff IB, Kelly KM. Seizures, antiepileptic drugs, and CKD. Am J Kidney Dis. 2019;73:90–101. doi: 10.1053/j.ajkd.2018.03.021. [DOI] [PubMed] [Google Scholar]
- 7.Lacerda G, Krummel T, Sabourdy C, Ryvlin P, Hirsch E. Optimizing therapy of seizures in patients with renal or hepatic dysfunction. Neurology. 2006;67(12 Suppl 4):S28–S33. doi: 10.1212/wnl.67.12_suppl_4.s28. [DOI] [PubMed] [Google Scholar]
- 8.Harden C. Safety profile of levetiracetam. Epilepsia. 2001;42(Suppl 4):36–39. [PubMed] [Google Scholar]
- 9.Ettinger AB, LoPresti A, Yang H, Williams B, Zhou S, Fain R, et al. Psychiatric and behavioral adverse events in randomized clinical studies of the noncompetitive AMPA receptor antagonist perampanel. Epilepsia. 2015;56:1252–1263. doi: 10.1111/epi.13054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kanner AM. Depression and epilepsy: a bidirectional relation? Epilepsia. 2011;52(Suppl 1):21–27. doi: 10.1111/j.1528-1167.2010.02907.x. [DOI] [PubMed] [Google Scholar]
- 11.Honybun E, Thwaites R, Malpas CB, Rayner G, Anderson A, Graham J, et al. Prenatal valproate exposure and adverse neurodevelopmental outcomes: does sex matter? Epilepsia. 2021;62:709–719. doi: 10.1111/epi.16827. [DOI] [PubMed] [Google Scholar]
- 12.Vajda FJ, Hitchcock AA, Graham J, O’Brien TJ, Lander CM, Eadie MJ. The teratogenic risk of antiepileptic drug polytherapy. Epilepsia. 2010;51:805–810. doi: 10.1111/j.1528-1167.2009.02336.x. [DOI] [PubMed] [Google Scholar]
- 13.Hakeem H, Feng W, Chen Z, Choong J, Brodie MJ, Fong SL, et al. Development and validation of a deep learning model for predicting treatment response in patients with newly diagnosed epilepsy. JAMA Neurol. 2022;79:986–996. doi: 10.1001/jamaneurol.2022.2514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43:1130–1139. doi: 10.1097/01.mlr.0000182534.19832.83. [DOI] [PubMed] [Google Scholar]
- 15.Brodie MJ, Barry SJ, Bamagous GA, Norrie JD, Kwan P. Patterns of treatment response in newly diagnosed epilepsy. Neurology. 2012;78:1548–1554. doi: 10.1212/WNL.0b013e3182563b19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chen Z, Brodie MJ, Liew D, Kwan P. Treatment outcomes in patients with newly diagnosed epilepsy treated with established and new antiepileptic drugs: a 30-year longitudinal cohort study. JAMA Neurol. 2018;75:279–286. doi: 10.1001/jamaneurol.2017.3949. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kwan P, Brodie MJ. Refractory epilepsy: a progressive, intractable but preventable condition? Seizure. 2002;11:77–84. doi: 10.1053/seiz.2002.0593. [DOI] [PubMed] [Google Scholar]
- 18.Dash D, Aggarwal V, Joshi R, Padma MV, Tripathi M. Effect of reduction of antiepileptic drugs in patients with drug-refractory epilepsy. Seizure. 2015;27:25–29. doi: 10.1016/j.seizure.2015.02.025. [DOI] [PubMed] [Google Scholar]
- 19.Ficker DM, Westmoreland BF, Sharbrough FW. Epileptiform abnormalities in hepatic encephalopathy. J Clin Neurophysiol. 1997;14:230–234. doi: 10.1097/00004691-199705000-00008. [DOI] [PubMed] [Google Scholar]
- 20.Rudler M, Marois C, Weiss N, Thabut D, Navarro V. Status epilepticus in patients with cirrhosis: how to avoid misdiagnosis in patients with hepatic encephalopathy. Seizure. 2017;45:192–197. doi: 10.1016/j.seizure.2016.12.011. [DOI] [PubMed] [Google Scholar]
- 21.Kobau R, Gilliam F, Thurman DJ. Prevalence of self-reported epilepsy or seizure disorder and its associations with self-reported depression and anxiety: results from the 2004 healthstyles survey. Epilepsia. 2006;47:1915–1921. doi: 10.1111/j.1528-1167.2006.00612.x. [DOI] [PubMed] [Google Scholar]
- 22.Strine TW, Kobau R, Chapman DP, Thurman DJ, Price P, Balluz LS. Psychological distress, comorbidities, and health behaviors among U.S. adults with seizures: results from the 2002 national health interview survey. Epilepsia. 2005;46:1133–1139. doi: 10.1111/j.1528-1167.2005.01605.x. [DOI] [PubMed] [Google Scholar]
- 23.Josephson CB, Lowerison M, Vallerand I, Sajobi TT, Patten S, Jette N, et al. Association of depression and treated depression with epilepsy and seizure outcomes: a multicohort analysis. JAMA Neurol. 2017;74:533–539. doi: 10.1001/jamaneurol.2016.5042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Devinsky O, Dilley C, Ozery-Flato M, Aharonov R, Goldschmidt Y, Rosen-Zvi M, et al. Changing the approach to treatment choice in epilepsy using big data. Epilepsy Behav. 2016;56:32–37. doi: 10.1016/j.yebeh.2015.12.039. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Log negative-log survival curve plots to assess the proportional hazard assumption of age group, generation of first antiseizure medication (ASM), number of ASMs in the last regimen, use of mechanical ventilation, and underlying hypertension, diabetes mellitus, ischemic heart disease, renal disease, liver disease, any malignancy, and depression.
Antiseizure medication (ASM) prescription trend during the study period. Use of first generation ASMs (e.g., valproic acid, carbamazepine, topiramate, phenytoin) gradually decreased and use of new generation ASMs (e.g., levetiracetam, oxcarbazepine, gabapentin, lamotrigine, clonazepam, pregabalin) gradually increased.