Clinical predictors of treatment outcome in North Indian patients on antiepileptic drug therapy: A prospective observational study
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.237000
Source of Support: None, Conflict of Interest: None
Keywords: Age at onset, cryptogenic epilepsy, epilepsy, pretreatment seizure frequency, prognostic factors
Epilepsy is a chronic neurological disease affecting around 69 million people worldwide which constitute nearly 1% of the world population. In India, 6–7 million people are affected with a prevalence rate of 2.2–10.4/1000/year and an incidence rate of 25.2–49.3/100,000/year. Being an incurable disease, it can only be treated by reducing the frequency of seizures  using antiepileptic drugs (AEDs); however, the often observed resistance toward these AEDs is still a major challenge to understand. Nearly 40%–50% of the individuals fail to respond to the first line AED monotherapy , with 30% being refractory, a medically intractable condition wherein a patient does not respond to even a multidrug therapy. This pharmacoresistance may also be fatal to the patients, as in the cases of sudden unexpected death in epilepsy. This has shifted the focus of epilepsy research toward identifying factors predictive of seizure prognosis in patients with epilepsy (PWE) on antiepileptic treatment. Several studies have investigated the possible predictors of seizure prognosis in medically treated PWE such as gender, age of onset, seizure and epilepsy type, electroencephalogram (EEG) patterns, neuroimaging abnormalities, and the number of pretreatment seizures.,,, Apart from these, socioeconomic factors such as poverty, illiteracy, and ignorance are also being related to a poor seizure remission.
In this study, we aim to identify the clinical predictors determining response to conventional AEDs in the North Indian population. In India, the conventional AEDs, phenytoin (PHT), carbamazepine (CBZ), valproate (VA), and phenobarbital (PB) are available free of cost in government-funded hospitals. Consequently, these drugs are prescribed to a large patient population. We, therefore, studied the response toward these AEDs in our patient population. This enabled us to develop a preliminary understanding of the AED response using clinical information.
Patients with epilepsy
A total of 1141 PWE of North Indian ethnicity fulfilling the inclusion criteria were screened at the Outpatient Department of Neurology, Institute of Human Behaviour and Allied Sciences (IHBAS), Delhi, India, between 2005 and 2015 after obtaining an informed consent. The inclusion criteria were as follows: the patients were >5 years of age and all of them were prescribed either the four AEDs, that is, PHT, CBZ, VA, and PB, or their combinations in a multitherapy (MultiT) regimen. The exclusion criteria were as follows: patients with gross neurological deficits such as mental retardation and/or motor deficits; with imaging abnormalities including the presence of a tumor, tuberculoma, multiple neurocysticercosis, vascular malformations, and atrophic lesions; with severe hepatic or renal disorders and diabetes mellitus; and the cases of pregnant women with epilepsy. Of the total of 1141 PWE encountered during the study period, 1056 patients were enrolled in this study and the remaining 85 patients were excluded in accordance with the exclusion criteria. All the four drugs are widely prescribed conventional AEDs, available free of cost at IHBAS, and were prescribed for oral administration at a dose within the therapeutic range, that is, 150–1200 mg/day for PHT, 800–1600 mg/day for CBZ, 250–3000 mg/day for VA, and 30–200 mg/day for PB. The study was approved by the institutional biomedical research ethics committee. The patients were diagnosed and prescribed treatment by an experienced neurologist, based upon the latest guidelines of the International League Against Epilepsy (ILAE)., The study protocol is summarized in [Figure 1]. A specified questionnaire was conceived to collect information regarding the clinical evaluation, demographic details, drug-dose information, and other investigations such as the electroencephalographic (EEG) profiling and neuroimaging. Patients were subsequently followed up at 2nd, 4th, 8th, and 12th month within a course of 1 year and evaluated for the drugs being administered and their dosage, the serum drug levels, the adequacy of seizure control, the adverse drug reactions (ADRs), and their compliance to AEDs. Patients who did not comply with the treatment regimen or who switched to a second line AED/MultiT in between the course of the study were excluded from the study.
Serum antiepileptic drug quantification
Serum drug levels of the prescribed AEDs were assessed in an auto-analyzer from Logitech Pvt. Ltd. (Model Echo) using CEDIA ® II assay kits by Microgenics Corporation (Fremont, CA, USA), as described in our previous report. The serum drug levels of each patient at different follow-ups were averaged over a period when consecutive doses were recorded, before correlating with other clinical parameters.
After an year of enrollment, the patients were assessed for their response to the prescribed treatment regimen. The patients were divided into two groups based on the number of seizures experienced over the study duration, excluding the initial 2 months period required for achieving steady-state levels of the AEDs being administered. Patients were categorized as (1) the “no seizure” group, which included those patients who attained complete freedom from seizures during the past 1 year, and (2) the “recurrent seizures” group, which included those patients who experienced ≥ 1 seizure/s during the same period.
For analyzing different parameters, the patients were divided on the basis of AEDs prescribed, gender, age at onset, seizure and epilepsy type, and seizure control status. For comparison of nonparametric continuous data, Kruskal–Wallis test was used, and for comparison of categorical data, chi-square test was applied. The probable interactions between the parameters were assessed using logistic regression analysis, considering the response to AEDs (“no seizure” vs. “recurrent seizures”) as the dependent variable and gender (male vs. female), age at onset (≤5 vs. >5 years)
, seizure type (focal vs. generalized seizures), epilepsy type (idiopathic + symptomatic vs. cryptogenic), and pretreatment seizure frequency (<12/year vs. ≥12/year) as independent variables. A P value of < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism 7 (GraphPad Software Inc., San Diego, CA, USA) and Microsoft Office Excel 2013.
Demographic and clinical characteristics of patients
A total of 1056 PWE, with their age range being 5–67 years (median age = 20 years) with a male: female ratio of 665:391, fulfilling the selection criteria, were enrolled in this study. Patients were prescribed PHT (n = 247, 23.4%), CBZ (n = 369, 34.9%), VA (n = 271, 25.7%), PB (n = 50, 4.7%), and MultiT (n = 119, 11.3%). Epilepsy was found to be more prevalent in the age group of 10-30 years (80%) and began at 5–20 years of age in most of the cases (70%). The age at onset, pretreatment seizure frequency, and the number of patients with focal seizures were similar (P value: nonsignificant) among the five treatment groups; however, the sex ratio, age of the patients at enrollment, time gap between the disease onset and treatment initiation, the number of patients with generalized seizures, and the epilepsy type significantly differed among these groups [Table 1].
With respect to the disease phenotypes, 394 patients (37.3%) were experiencing focal/partial seizures, 628 (59.5%) were having generalized seizures, and the remaining 33 patients (3.2%) were experiencing either mixed or other types of seizures [Table 1]. Based on the complete clinical history, investigation reports, and detailed follow-ups, 786 patients (74.4%) were carefully chosen for further analysis. Among these 786 patients, symptomatic epilepsy (n = 387, 49.3%) was observed to be most prevalent, followed by idiopathic (n = 203, 25.8%) and cryptogenic epilepsy (n = 196, 24.9%).
Response to antiepileptic drugs
A total of 786 (74.4%) patients on PHT (n = 161), CBZ (n = 283), VA (n = 205), PB (n = 37), and MultiT (n = 100) were successfully followed up over the next 1 year [Table 2]. Only 46 patients (6%) experienced ADRs related to skin, gastrointestinal tract, central nervous system, hormonal imbalance, others, or their combinations. Based on response to prescribed AEDs, patients were successfully segregated into the “no seizure” and “recurrent seizures” groups. Seizure freedom was achieved by 59%–65% of the patients on monotherapy, with VA showing the highest efficacy (65% “no seizure”). Among patients on MultiT, treatment failure was higher, with 55% patients in the “recurrent seizures” group. Overall, seizure freedom was observed in 475 patients (60%). A number of factors were studied to justify this differential drug response among PWE.
Gender and age at onset
Seizure recurrence was similar among the two gender groups (male vs. female, 40% vs. 39%) in the followed-up population. However, patients with an early disease onset, that is, at ≤5 years of age, were found to have higher chances of seizure recurrence than those with onset at >5 years of age [55% vs. 38%, P = 0.0016, odds ratio (OR) = 2.02, 95% confidence interval (CI) = 1.31–3.13] [Table 3].
Seizure and epilepsy type
No difference was observed in the proportion of patients with focal seizure and those with generalized seizure experiencing seizure recurrence (40% vs. 39%) [Table 3]. As among the epilepsy types, a higher proportion of patients with cryptogenic epilepsy continued to have seizures during their treatment than patients with idiopathic or symptomatic epilepsy (48% vs. 33%, P = 0.0049, OR= 1.61, 95% CI= 1.16–2.24). No significant difference was found between patients with symptomatic epilepsy and those with idiopathic epilepsy in the “recurrent seizures” group (35% vs. 29%).
Pretreatment seizure frequency
To examine the correlation between pretreatment seizure frequency and the AED response, patients were categorized into five groups, namely, daily, weekly, monthly, yearly, and 5 yearly, based on the frequency of seizures experienced by the patients before initiating treatment [Table 1]. Of the 786 followed-up patients, 34 (4%) were clusterers, that is, those experiencing three or more seizures per day before treatment initiation. The likelihood of seizure recurrence during AED therapy was similar between the clusterers and non-clusterers; however, it was significantly higher in patients who experienced ≥12 seizures per year than in those who experienced lesser frequency of seizures (46% vs. 27%, P < 0.0001, OR = 2.21, 95% CI = 1.61–3.05) [Table 3].
Treatment delay was defined as the time interval between the disease onset and treatment initiation. Only 212 patients (27%) came for treatment within 1 year of their disease onset. No significant variation in seizure recurrence was found in patients who come for treatment after 1 year of seizure onset when compared with those who come within 1 year (42% vs. 36%, P = 0.1912, OR = 1.30, 95% CI = 0.88–1.92).
Serum drug levels
Serum drug levels were evaluated in 686 patients on different AED monotherapies and compared between the “no seizure” and “recurrent seizures” groups. On comparison, the two groups were found to have similar serum levels of the prescribed AEDs [Figure 2]. Furthermore, to account for dose variability, serum drug levels were corrected with respective doses; however, again, no difference was observed between the two groups.
A multivariate analysis was performed using logistic regression model to determine the possible interaction between age at onset, cryptogenic epilepsy, and pretreatment seizure frequency, considering response to AEDs as the dependent variable and gender, age at onset, seizure type, epilepsy type, and pretreatment seizure frequency as independent variables. In this model too, the three factors, early disease onset (OR = 2.28, 95% CI = 1.22–4.28, P = 0.0101), cryptogenic epilepsy (OR = 1.79, 95% CI = 1.07–3.00, P = 0.0277), and a higher pretreatment seizure frequency (OR = 2.63, 95% CI = 1.64–4.21, P = 0.0001) were found to be significantly associated with a poor treatment outcome in PWE. However, no link was revealed with gender (OR = 0.86, 95% CI = 0.55–1.34, P = 0.5110) and seizure type (OR = 1.09, 95% CI = 0.6868–1.72, P = 0.7258).
Apart from these parameters, other factors such as status epilepticus, involvement of precipitation factor/s, and presence of adverse drug reactions (ADR/s)were also examined for association with seizure recurrence during the AED therapy. None of the factors seemed to have influenced the AED response.
Our study observed an 1-year seizure remission rate of 59%–65% in patients on AED monotherapy, which dropped down to 45% in patients on MultiT. There is a debate currently going on regarding the efficacy and tolerability of AED monotherapy versus that of MultiT., Previous studies , on AED treatment outcome also reported a lower remission rate during MultiT when compared with monotherapy, which supports the need to prescribe AEDs in rational combinations, considering their mechanisms of action, pharmacokinetic interactions, and adverse effects along with their efficacy. However, a multicenter randomized study  found no significant differences between the two therapies. The type of therapy may differ based on the understanding and knowledge of one clinician to another and may lead to inconsistency in findings. Thus, we did not compare monotherapy with MultiT as one of the predictor variables. We used only intrinsic parameters such as gender and age at onset which could not be altered by external means.
Although gender, seizure type, treatment delay, and serum drug levels had no role in determining treatment outcome in our study, univariate and multivariate analyses revealed three factors to be predictive of a poor prognosis during AED therapy: an early disease onset, cryptogenic epilepsy, and a higher pretreatment seizure frequency. Our study observed poor seizure remission in patients with disease onset at ≤5 years of age, which is consistent with the results of previously published reports., Neuronal development occurs at an early age, hence any assault to the brain may result in permanent damage, leading to resistance to the prescribed drugs.
As reported by several previous studies,,, we also observed patients with idiopathic epilepsy showing the highest efficacy (1-year seizure remission, n = 71%) to prescribed AEDs when compared with those having other types of epilepsy. Till date, a poor response in symptomatic epilepsy has been attributed to the presence of a neuroanatomic/neuropathologic abnormality in the brain. However, in our study, seizure recurrence did not significantly differ between patients having an idiopathic versus a symptomatic epilepsy. Instead, higher chances of nonresponse to treatment were observed in cryptogenic epilepsy. Cryptogenic epilepsy involves those epilepsies where the cause is unknown and involves a presumptive lesion which is difficult to detect with the current technology. Lack of knowledge regarding the etiology behind this class of epilepsy may be one of the reasons for treatment failure.
A higher pretreatment seizure frequency was observed to be another factor associated with seizure recurrence during AED treatment. Patients who had seizures on a daily, weekly, and monthly basis before treatment initiation were more prone to treatment failure when compared with those who were having them on an yearly or a 5 yearly basis. Sillanpää et al., also reported that having weekly seizures before treatment was significantly associated with the risk of not entering the 1-year terminal remission. Other studies ,, examined the number of pretreatment seizures, and the pretreatment seizure frequency was found to be positively correlated with treatment outcome. Based on their findings, Rogawski and Johnson  conceived 'the intrinsic severity hypothesis', which suggests that neurobiological factors (such as frequent seizures at disease onset) that confer increased disease severity lead to drug intractability.
So far, no correlation between serum drug levels and AED efficacy in PWE has been established. Our study too observed no difference in the serum drug levels between the “no seizure” and the “recurrent seizures” groups. To explain this disparity, researchers are looking for probable causal genetic variations that might be responsible for different drug resistance hypotheses. One hypothesis is the limited access of the drug to the epileptic focus within the brain due to seizure-induced P-glycoprotein expression and other adenosine triphosphate binding cassette (ABC) transporters.,, Another reason could be the role of drug-metabolizing enzymes in defining drug efficacy in PWE.,,, Previous reports from our group reported the association of an intronic single nucleotide polymorphism, IVS1 +606C>A (rs2606345) in drug-metabolizing enzyme, CYP1A1, with the efficacy of conventional AEDs in women with epilepsy., These reports showed a gender specificity in determining AED efficacy at the genetic level, which our clinical study could not reveal. Thus, studying the pharmacogenomics of AEDs may provide the impetus in identifying potential genetic markers for predicting AED efficacy. However, the inconsistencies in genetic findings emphasize the need to understand clinical heterogeneity and consider both clinical and genetic factors while identifying predictors of the response to AEDs.
Our study has some limitations. First, since our work is an observational study instead of a randomized, controlled one, the results may have been affected due to the selection bias. Second, the study duration is short, and thus, it cannot infer the results for a long-term prognosis. Finally, the study involves only conventional AEDs. The results, therefore, cannot be applied to a large patient population receiving newer AEDs. However, the strength of our study is in its prospective nature, collection of detailed information regarding the samples, and multiple time points for regular monitoring of doses and serum drug levels.
Our data suggests that 35%–41% of the patients do not respond to AED monotherapy and thus emphasizes the need to identify predictors that determine the treatment outcome in epilepsy. An early disease onset, cryptogenic epilepsy, and a higher pretreatment seizure frequency may serve as potential clinical predictors to identify AED treatment failure in epilepsy. This signifies that patients with these three variables may show resistance to the currently available AEDs. Thus, the development of novel personalized treatment approaches is required. Besides these clinical factors, genetic factors may also play a crucial role in determining the treatment outcome in epilepsies, which emphasizes the need to establish a rational approach considering both clinical and genetic factors for determining drug efficacy.
The authors express their thanks to patients and their family members for participating in the study. They gratefully acknowledge valuable scientific discussions and suggestions by Prof. Samir K. Brahmachari (Council of Scientific and Industrial Research [CSIR]- Institute of Genomics and Integrative Biology [IGIB]), Prof. M. Gourie-Devi (Sir Ganga Ram Hospital), and Dr. Abhay Sharma (CSIR-IGIB). They are thankful to Dr. Nimesh G. Desai, Director, IHBAS, for his support during project implementation. They also wish to thank Ms. Shama Parween for her assistance in the patient enrollment. CR acknowledges University Grants Commission (UGC) and DG acknowledges Indian Council of Medical Research (ICMR) for providing fellowships.
Financial support and sponsorship
This work was supported by Council of Scientific and Industrial Research (CSIR), BSC0123, and Indian Council of Medical Research (ICMR), GAP0091.
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]