Advertisement

Assessment of treatment patterns and healthcare costs associated with probable Lennox–Gastaut syndrome

Open AccessPublished:June 10, 2017DOI:https://doi.org/10.1016/j.yebeh.2017.05.021

      Highlights

      • 8% of 10-year-old epilepsy patients had probable LGS.
      • Prevalence of LGS decreased as age cohort increased.
      • Healthcare costs in LGS patients were consistently higher than non-LGS patients.
      • Many LGS patients did not receive widest-spectrum AED and may be inadequately treated.
      • Clobazam and rufinamide use was lower in older than younger cohorts of LGS patients.

      Abstract

      Lennox–Gastaut syndrome (LGS) is a chronic and severe form of epilepsy characterized by intractable seizures, cognitive impairment, and abnormal electroencephalogram findings with slow spike-wave complexes. It typically presents before age 8, but symptoms continue into adulthood and require lifelong treatment associated with significant clinical burden. Data on LGS-associated healthcare utilization and costs are limited. In this study we use a claims-based LGS classifier based on random forest methodology to identify patients with probable LGS from the a Medicaid multi-state database and assess its prevalence across the age spectrum, healthcare utilization, treatment patterns, costs, and comorbid conditions.
      The classifier identified patients with probable LGS across all ages, with up to 8% of 10-year-old patients with epilepsy identified as having probable LGS. The prevalence of probable LGS was lower in older age cohorts, indicating that it may be under-recognized in older patients. Our analysis showed that probable LGS is associated with considerably higher total healthcare and medical costs than non-LGS patients. The costs were generally consistent between age cohorts, suggesting that the cost burden extends beyond childhood and has a lifelong impact. Analysis of treatment patterns suggest that while the majority of probable LGS patients in this study received widest-spectrum AEDs, a considerable proportion did not and therefore may have been inadequately treated. Further, usage of clobazam and rufinamide was decreased in older compared to younger patient cohorts, indicating that older patient cohorts are less likely to be receiving optimum treatment for LGS.
      These findings indicate the need for increased clinical attention to LGS beyond pediatric years, with a focus on optimization of treatment for LGS patients of all ages with widest-spectrum AEDs. Timely recognition and adequate treatment of LGS are likely to result in improved outcomes and less costly management of this condition.

      Abbreviations:

      ADHD (attention deficit hyperactivity disorder), AED (antiepileptic drug), EEG (electroencephalogram), ER (emergency room), ICD-9-CM (International Classification of Disease, 9th Revision, Clinical Modification), IDD (intellectual developmental disorders), LGS (Lennox-Gastaut syndrome), NDC (National Drug Code), PPPY (per patient per year), SSW (slow spike-wave), VNS (vagus nerve stimulation), yrs (years)

      Keywords

      1. Introduction

      Lennox–Gastaut syndrome (LGS) is a chronic and severe form of epilepsy characterized by multiple intractable seizure types such as tonic, absence, and atonic “drop attacks,” cognitive impairment or regression, and abnormal electroencephalogram (EEG) findings with slow spike-wave (SSW) complexes. The syndrome typically presents before age 8, with peak onset of seizures between 3 and 5 years of age [
      • Bourgeois B.F.
      • Douglass L.M.
      • Sankar R.
      Lennox-Gastaut syndrome: a consensus approach to differential diagnosis.
      ,
      • Hancock E.C.
      • Cross J.H.
      Treatment of Lennox-Gastaut syndrome.
      ,
      • Camfield P.R.
      Definition and natural history of Lennox-Gastaut syndrome.
      ,
      • Arzimanoglou A.
      • French J.
      • Blume W.T.
      • Cross J.H.
      • Ernst J.P.
      • Feucht M.
      • et al.
      Lennox-Gastaut syndrome: a consensus approach on diagnosis, assessment, management, and trial methodology.
      ]. While LGS sometimes develops in previously healthy children for unknown reasons, 75% of cases are associated with a congenital abnormality or insult to the brain from infection or injury [
      • Camfield P.R.
      Definition and natural history of Lennox-Gastaut syndrome.
      ]. LGS is estimated to account for 1–10% of all childhood epilepsies [
      • Bourgeois B.F.
      • Douglass L.M.
      • Sankar R.
      Lennox-Gastaut syndrome: a consensus approach to differential diagnosis.
      ,
      • Hancock E.C.
      • Cross J.H.
      Treatment of Lennox-Gastaut syndrome.
      ,
      • Camfield P.R.
      Definition and natural history of Lennox-Gastaut syndrome.
      ,
      • Arzimanoglou A.
      • French J.
      • Blume W.T.
      • Cross J.H.
      • Ernst J.P.
      • Feucht M.
      • et al.
      Lennox-Gastaut syndrome: a consensus approach on diagnosis, assessment, management, and trial methodology.
      ], but this prevalence may be underestimated given the challenges in its recognition. It is one of the most difficult epileptic disorders to identify and manage due to multiple etiologies, lack of specific biological markers, and varied presentation and progression of features, as the syndrome often emerges over several years and continues to evolve over the life course [
      • Ferlazzo E.
      • Nikanorova M.
      • Italiano D.
      • Bureau M.
      • Dravet C.
      • Calarese T.
      • et al.
      Lennox-Gastaut syndrome in adulthood: clinical and EEG features.
      ,
      • Ogawa K.
      • Kanemoto K.
      • Ishii Y.
      • Koyama M.
      • Shirasaka Y.
      • Kawasaki J.
      • et al.
      Long-term follow-up study of Lennox-Gastaut syndrome in patients with severe motor and intellectual disabilities: with special reference to the problem of dysphagia.
      ,
      • Oguni H.
      • Hayashi K.
      • Osawa M.
      Long-term prognosis of Lennox-Gastaut syndrome.
      ,
      • Ohtsuka Y.
      • Amano R.
      • Mizukawa M.
      • Ohtahara S.
      Long-term prognosis of the Lennox-Gastaut syndrome.
      ,
      • Pina-Garza J.E.
      • Chung S.
      • Montouris G.D.
      • Radtke R.A.
      • Resnick T.
      • Wechsler R.T.
      Challenges in identifying Lennox-Gastaut syndrome in adults: a case series illustrating its changing nature.
      ,
      • Yagi K.
      Evolution of Lennox-Gastaut syndrome: a long-term longitudinal study.
      ].
      Long-term prognosis of LGS is poor, with increased mortality rates, often due to seizure-related accidents [
      • Glauser T.A.
      Following catastrophic epilepsy patients from childhood to adulthood.
      ,
      • Camfield P.
      • Camfield C.
      Long-term prognosis for symptomatic (secondarily) generalized epilepsies: a population-based study.
      ,
      • Autry A.R.
      • Trevathan E.
      • Van Naarden Braun K.
      • Yeargin-Allsopp M.
      Increased risk of death among children with Lennox-Gastaut syndrome and infantile spasms.
      ]. Seizures are generally intractable and usually continue into adulthood, requiring lifelong treatment with multiple antiepileptic drugs (AEDs), surgical interventions, and the use of protective headgear to prevent serious injury from falls. Almost all LGS patients have a progressive deterioration in cognitive function with associated severe behavioral disorders. Most adult patients with LGS are unable to live independently [
      • Camfield C.
      • Camfield P.
      Twenty years after childhood-onset symptomatic generalized epilepsy the social outcome is usually dependency or death: a population-based study.
      ,
      • Camfield P.R.
      • Gibson P.A.
      • Douglass L.M.
      Strategies for transitioning to adult care for youth with Lennox-Gastaut syndrome and related disorders.
      ,
      • Skornicki M.
      • Clements K.M.
      • O'Sullivan A.K.
      Budget impact analysis of antiepileptic drugs for Lennox-Gastaut syndrome.
      ]. Together, these challenges present a considerable negative impact on quality of life for LGS patients and their families, as well as significant economic burden [
      • Skornicki M.
      • Clements K.M.
      • O'Sullivan A.K.
      Budget impact analysis of antiepileptic drugs for Lennox-Gastaut syndrome.
      ,
      • Benedict A.
      • Verdian L.
      • Maclaine G.
      The cost effectiveness of rufinamide in the treatment of Lennox-Gastaut syndrome in the UK.
      ,
      • Verdian L.
      • Yi Y.
      Cost-utility analysis of rufinamide versus topiramate and lamotrigine for the treatment of children with Lennox-Gastaut Syndrome in the United Kingdom.
      ,
      • Clements K.M.
      • Skornicki M.
      • O'Sullivan A.K.
      Cost-effectiveness analysis of antiepileptic drugs in the treatment of Lennox-Gastaut syndrome.
      ].
      Healthcare utilization, costs, and treatment patterns associated with LGS have not been well studied, due at least in part to its relatively low prevalence and inconsistency of diagnostic information captured in patient charts and in medical insurance databases. Medical insurance databases offer comprehensive data on healthcare utilization and costs for large patient populations, but LGS patients can be difficult to identify in these databases, particularly prior to the implementation of the 10th revision of the International Classification of Disease, Clinical Modification (ICD-10-CM [
      • The International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM)
      ]). Prior to implementation of the ICD-10-CM in October 2015, there was no ICD diagnosis code specific to LGS [
      • National Center for Health Statistics
      The international classification of diseases, ninth revision, clinical modification (ICD-9-CM).
      ]). Previous published studies have assessed budget impact and cost-effectiveness of AEDs using trial- or decision-based models in hypothetical cohorts of patients with LGS [
      • Skornicki M.
      • Clements K.M.
      • O'Sullivan A.K.
      Budget impact analysis of antiepileptic drugs for Lennox-Gastaut syndrome.
      ,
      • Benedict A.
      • Verdian L.
      • Maclaine G.
      The cost effectiveness of rufinamide in the treatment of Lennox-Gastaut syndrome in the UK.
      ,
      • Verdian L.
      • Yi Y.
      Cost-utility analysis of rufinamide versus topiramate and lamotrigine for the treatment of children with Lennox-Gastaut Syndrome in the United Kingdom.
      ,
      • Clements K.M.
      • Skornicki M.
      • O'Sullivan A.K.
      Cost-effectiveness analysis of antiepileptic drugs in the treatment of Lennox-Gastaut syndrome.
      ], but none to date have provided real-world data.
      In this study a claims-based LGS classifier based on random forest methodology was used to identify patients with probable LGS and evaluate their healthcare utilization and costs. This robust machine-learning methodology allows assessment of the prevalence of probable LGS, as well as associated total, medical, and pharmacy costs and treatment patterns across the age spectrum, thus providing real-world data on the clinical and economic burden of LGS.

      2. Methods

      2.1 Data source

      This retrospective claims analysis used insurance claims data from a Medicaid multi-state database of six states over the following periods: Florida (1997 Q3–2012 Q2), Iowa (1998 Q1–2013 Q1), Kansas (2001 Q1–2013 Q1), Mississippi (2006 Q1–2013 Q4), Missouri (1997 Q1–2013 Q1), and New Jersey (1997 Q1–2013 Q1). The duration of observable data for individual patients depended on their enrollment in the Medicaid program during these time periods; average enrollment duration was about ten years (Table 1). Medicaid claims data capture demographic information (gender, age, and race), diagnosis, costs (paid amounts), healthcare utilization information (procedure codes, dates of outpatient and emergency department [ED] visits, and dates/lengths of stay for inpatient), and dispensed drug information (National Drug Code [NDC] codes, costs, quantity/days of supply, and date of dispensing).
      Table 1Demographic characteristics and claims data of probable LGS and non-LGS patients identified using LGS classifier.
      Probable LGS patients

      N = 14,712
      Non-LGS patients

      N = 353,281
      Age at first epilepsy diagnosis, yrs., mean ± SD16.3 ± 15.735.3 ± 24.2
      Length of observation period, yrs., mean ± SD11.1 ± 4.59.3 ± 5.0
      Male, n (%)7,718 (52.5)166,916 (47.2)
      Race/ethnicity, n (%)
       Caucasian8,367 (56.9)202,430 (57.3)
       African American2,346 (15.9)86,368 (24.4)
       Hispanic547 (3.7)16,233 (4.6)
       Other2,895 (19.7)37,253 (10.5)
       Unknown557 (3.8)10,996 (3.1)
      Common comorbid diagnoses, n (%)
       Delayed development9,881 (67.2)61,171 (17.3)
       IDD
      IDD was categorized in ICD-9 as mental retardation.
      8,633 (58.7)54,494 (15.4)
       Epileptic encephalopathy6,425 (43.7)59,192 (16.8)
       Bone fracture4,599 (31.3)90,307 (25.6)
       Mood disorder3,627 (24.7)108,546 (30.7)
       Anxiety disorder3,169 (21.5)98,636 (27.9)
       ADHD2,654 (18.0)36,556 (10.3)
       Autism2,511 (17.1)14,025 (4.0)
       West syndrome980 (6.7)1484 (0.4)
      Non-drug treatment, n (%)
       VNS implant3,004 (20.4)1828 (0.5)
       Helmut use979 (6.7)963 (0.3)
       Corpus callosotomy121 (0.8)0 (0.0)
      Drug treatment
       Number of AEDs received, mean ± SD5.8 ± 2.31.8 ± 1.3
       ≥3 different AEDs received, n (%)13,797 (93.8)76,503 (21.7)
       ≥1 prescription fill for an AED, n (%)14,333 (97.4)239,392 (67.8)
      Clobazam
      Clobazam and rufinamide are the only AEDs that are US FDA-approved only for adjunctive treatment of seizures associated with LGS. Additional wide-spectrum AEDs are approved for LGS as well as other conditions and/or seizure types.
      1,007 (6.8)353 (0.1)
      Rufinamide
      Clobazam and rufinamide are the only AEDs that are US FDA-approved only for adjunctive treatment of seizures associated with LGS. Additional wide-spectrum AEDs are approved for LGS as well as other conditions and/or seizure types.
      1,177 (8.0)44 (0.0)
      Healthcare resource utilization (PPPY), mean ± SD
       Outpatient visits11.8 ± 24.19.0 ± 20.7
       Inpatient visits0.6 ± 1.50.6 ± 1.7
       Emergency department visits1.4 ± 4.51.3 ± 4.1
       EEG procedures0.3 ± 0.70.1 ± 0.4
       Neurological procedures1.0 ± 1.90.3 ± 1.0
       Use of wheelchair or walker0.1 ± 0.40.0 ± 0.4
      ADHD, attention deficit hyperactivity disorder; AED, antiepileptic drug; EEG, electroencephalogram; IDD, intellectual developmental disorders; LGS, Lennox-Gastaut syndrome; PPPY, per patient per year; VNS, vagus nerve stimulation; yrs., years.
      a IDD was categorized in ICD-9 as mental retardation.
      b Clobazam and rufinamide are the only AEDs that are US FDA-approved only for adjunctive treatment of seizures associated with LGS. Additional wide-spectrum AEDs are approved for LGS as well as other conditions and/or seizure types.

      2.2 Patient population and study design

      Eligible patients had ≥2 medical claims ≥30 days apart for specified epilepsy (ICD-9-CM: 345.xx) or unspecified epilepsy (ICD-9-CM: 780.39). Patients with ≥1 medical claim for petit mal status (ICD-9-CM: 345.2x) were excluded.
      LGS patients were identified from the overall epilepsy population using a classifier developed using random forest methodology, a machine-learning technique, paired with clinical input from JEPG and GM. The methodology will be described in detail in another article and is briefly summarized here. Candidate variables for LGS identification were drug-related (e.g., number of distinct AEDs received, number of AED prescription-fills, ≥1 prescription-fill for clobazam or felbamate; medical treatment-related (e.g., ≥1 claim for corpus callosotomy, vagus nerve stimulation [VNS] implant, helmet use, group home, or inpatient stay); or diagnosis-related data (e.g., ≥2 claims for intellectual developmental disorders [IDD] or delayed development, or ≥1 claim for autism, aggression, bone fracture, West syndrome, or epileptic encephalopathy; for complete list of variables see Table 1 in accompanying paper).
      Using a machine-learning technique, automated classification and regression decision trees of randomly selected candidate LGS variables predicted the disease status as “LGS” or “non-LGS.” Thousands of decision trees were compiled to form a (random forest) aggregate, and the patient's final disease status was selected based on whether the aggregate probability of disease was greater than a predetermined probability threshold.

      2.3 Outcomes and analyses

      Probable LGS and non-LGS patients were observed from their first epilepsy-related claim to the end of study eligibility, data availability, or patient death. Descriptive statistics (mean, frequency, proportion) were used to assess demographic and clinical characteristics, including gender, age at first diagnosis, of likely LGS patients compared to those of non-LGS patients. In addition, common comorbid diagnoses, drug and non-drug treatments, and healthcare resource utilization, were compared between probable LGS and non-LGS patients.
      Additional outcomes included number and proportion of patients with probable LGS, mean healthcare costs per patient per year (including total cost, total medical cost, and total pharmacy cost) for probable LGS and non-LGS patients, proportion of probable LGS patients with ≥1 claim for narrow-, wide-, and widest-spectrum AEDs, and proportion of probable LGS with ≥1 claim for clobazam or rufinamide. For these outcomes, patient data were categorized into one or more 10-year cohorts, based on age at the time of claim, and plotted against time-series panels to depict changes across age cohorts. This was a retrospective analysis of Medicaid claims data and did not involve the use of human subjects; therefore ethics committee approval was not applicable.

      3. Results

      3.1 Demographic and clinical characteristics

      Of nearly 400,000 patients meeting eligibility criteria for epilepsy, 14,712 patients were identified as having probable LGS based on the LGS classifier (Table 1). Patients with probable LGS were younger at age of first diagnosis than non-LGS patients (mean age 16.3 vs 35.3 years). Probable LGS and non-LGS patients had similar mean length of observations periods (11.1 vs 9.3 years). With the exception of mood disorder, patients with probable LGS had higher rates of comorbidities than non-LGS patients, with the most frequent being delayed development (67.2% vs 17.3%), intellectual developmental disorder (58.7% vs 15.4%), epileptic encephalopathy (43.7% vs 16.8%), and bone fracture (31.3% vs 25.6%). Probable LGS patients also had more treatment-related claims than non-LGS patients, including AED prescription fills and non-drug treatments such as VNS implant, helmet use, and corpus callosotomy. Prescriptions for clobazam and rufinamide were filled by 6.8% and 8.0%, respectively, of LGS patients. Mean healthcare resource utilization, including outpatient, inpatient, and ER visits, EEG and neurological procedures, and use of wheelchair or walker, were similar in probable LGS versus non-LGS patients, with outpatient visits and neurological procedures slightly higher in likely LGS patients.

      3.2 Prevalence and healthcare costs in probable LGS patients

      The proportion of probable LGS patients increased from early age cohorts to a maximum of 8.4% at age 10, and then steadily declined in the age-10 through adult cohorts. Only 2.0% of patients in the age-60 cohort were identified by the classifier as having probable LGS (Fig. 1).
      Fig. 1
      Fig. 1Number and proportion of epilepsy patients with probable LGS by age cohort.
      LGS, Lennox-Gastaut syndrome.
      Compared with non-LGS patients, probable LGS patients had substantially higher total healthcare costs (non-LGS, patients: $7170–$25,901 per patient per year [PPPY]; probable LGS, $28,461–$40,193 PPPY) (Fig. 2). Of all healthcare costs incurred by patients with probable LGS, medical costs were the main cost drivers ($25,303–$37,342 PPPY). Home-based care was the main cost driver in the 1–18 year-cohorts, totaling $12,396 to $18,360 PPPY, but decreased sharply in adult cohorts to $6978 PPPY at age 60. Long-term care (LTC) costs ranged between $1648 to $5933 PPPY in pediatric cohorts and increased to $16,215 PPPY in the age-60 cohort. Pharmacy costs contributed to a relatively small portion of all healthcare costs incurred by probable LGS patients ($1592–$5630 PPPY).
      Fig. 2
      Fig. 2Mean healthcare costs for probable LGS and non-LGS patients (PPPY by age cohort).
      LGS, Lennox-Gastaut syndrome; PPPY, per patient per year.

      3.3 LGS treatment patterns

      Across all age cohorts, most probable LGS patients (62.6%–82.3%) had ≥1 claim for an AED (Fig. 3). The majority of claims were for the widest spectrum AEDs, with the rate of claims remaining fairly consistent across age cohorts. Approximately 26.0–51.0% of probable LGS patients received narrow spectrum AEDs with the highest rates in the 30- to 40-year-old cohorts. Less than 10% of probable LGS patients had prescriptions for wide spectrum AEDs.
      Fig. 3
      Fig. 3Probable LGS patients with ≥1 claim for AEDs by age cohort.
      aWidest spectrum AEDs: clobazam, clonazepam, clorazepate, diazepam, felbamate, lamotrigine, levetiracetam, perampanel, rufinamide, topiramate, valproate, zonisamide.
      bWide spectrum AEDs: lacosamide, phenobarbital, pregabalin, primidone, vigabatrin.
      cNarrow spectrum AEDs: bromides, carbamazepine, eslicarbazepine, ethosuximide, fosphenytoin, gabapentin, oxcarbazepine, phenytoin, retigabine, tiagabine.
      AED, antiepileptic drug; LGS, Lennox-Gastaut syndrome; PPPY, per patient per year.
      Use of AEDs indicated only for seizures associated with LGS (clobazam and rufinamide) was not common in probable LGS patients (Fig. 4). The highest proportion of probable LGS patients using either clobazam or rufinamide was in patients under age 5 (maximum: 17.5% for clobazam, 7.4% for rufinamide), with rates declining steadily to about 5.5% of patients receiving either drug in the age-20 cohort and <1.0% of patients in the age-60 cohort. In contrast, the use of antipsychotics and selective serotonin reuptake inhibitors (SSRIs, not shown) increased sharply in the cohorts up to 18 years old, (antipsychotics: 0–19%, SSRIs: 0–11%), and reached maximum rates during adult cohorts (antipsychotics: 24.6%, SSRIs: 21.5%).
      Fig. 4
      Fig. 4Probable LGS patients with ≥1 claim for AEDs indicated solely for LGS by age cohort.
      aThe proportion of probable LGS patients with ≥1 claim for clobazam or rufinamide was computed as a fraction of the total number of probable LGS patients present in the sample after the approval date for each drug (rufinamide, November 2008; clobazam, October 2011).
      AED, antiepileptic drug; LGS, Lennox-Gastaut syndrome; PPPY, per patient per year.

      4. Discussion

      The current longitudinal assessment of treatment patterns and healthcare costs of patients with probable LGS demonstrates its long-term impact, with probable LGS patients amassing considerably higher total healthcare and medical costs than non-LGS patients. Probable LGS-associated costs ranged from approximately $25,000 to $40,000 across age cohorts, suggesting that LGS is associated with a considerable cost burden that extends beyond childhood to the entire life course.
      Probable LGS existed across all age cohorts, with up to 8% of 10-year-old patients with epilepsy identified by the classifier as having LGS. The prevalence of probable LGS appeared to decrease with increasing age cohort, indicating that it may be under-recognized among older patients. Previous studies have described the changing nature of LGS features, such as decreased incidence of SSW discharges and changes in seizure type and/or frequency [
      • Ferlazzo E.
      • Nikanorova M.
      • Italiano D.
      • Bureau M.
      • Dravet C.
      • Calarese T.
      • et al.
      Lennox-Gastaut syndrome in adulthood: clinical and EEG features.
      ,
      • Ogawa K.
      • Kanemoto K.
      • Ishii Y.
      • Koyama M.
      • Shirasaka Y.
      • Kawasaki J.
      • et al.
      Long-term follow-up study of Lennox-Gastaut syndrome in patients with severe motor and intellectual disabilities: with special reference to the problem of dysphagia.
      ,
      • Oguni H.
      • Hayashi K.
      • Osawa M.
      Long-term prognosis of Lennox-Gastaut syndrome.
      ,
      • Ohtsuka Y.
      • Amano R.
      • Mizukawa M.
      • Ohtahara S.
      Long-term prognosis of the Lennox-Gastaut syndrome.
      ,
      • Pina-Garza J.E.
      • Chung S.
      • Montouris G.D.
      • Radtke R.A.
      • Resnick T.
      • Wechsler R.T.
      Challenges in identifying Lennox-Gastaut syndrome in adults: a case series illustrating its changing nature.
      ,
      • Beaumanoir A.
      The Lennox-Gastaut syndrome: a personal study.
      ,
      • Hughes J.R.
      • Patil V.K.
      Long-term electro-clinical changes in the Lennox-Gastaut syndrome before, during, and after the slow spike-wave pattern.
      ,
      • Roger J.
      • Remy C.
      • Bureau M.
      • Oller-Daurella L.
      • Beaumanoir A.
      • Favel P.
      • et al.
      Lennox-Gastaut syndrome in the adult.
      ]. These changes, which may reflect response to treatment, changes in brain plasticity with aging, or effects of prolonged epileptic activity, can complicate the diagnosis and treatment of LGS [
      • Pina-Garza J.E.
      • Chung S.
      • Montouris G.D.
      • Radtke R.A.
      • Resnick T.
      • Wechsler R.T.
      Challenges in identifying Lennox-Gastaut syndrome in adults: a case series illustrating its changing nature.
      ]. It is important for clinicians to continually reappraise clinical and EEG findings in patients of all ages with epileptic encephalopathy, bearing in mind the variable progression of LGS features.
      Our analysis of AED treatment patterns suggests that while the majority of probable LGS patients received widest-spectrum AEDs, a considerable proportion of patients did not and therefore may have been inadequately treated. Further, the decline of clobazam and rufinamide usage rates in older patient cohorts indicates that treatment optimization may decrease as patients with LGS age. The underutilization of widest-spectrum AEDs observed in this study, particularly LGS-specific AEDs such as clobazam and rufinamide, is likely due to lack of knowledge and/or experience with these newer AEDs, yet these widest-spectrum AEDs have demonstrated efficacy in treating atonic seizures, and therefore confer potential benefits in the form of reduced risk of injury from atonic seizures [
      • Montouris G.D.
      • Wheless J.W.
      • Glauser T.A.
      The efficacy and tolerability of pharmacologic treatment options for Lennox-Gastaut syndrome.
      ].
      Limitations to this study include possible inaccuracies in the Medicaid claims databases (e.g., diagnosis misclassification or underreporting). The database includes neither complete clinical data (therefore limiting assessment of disease severity), nor information on oral medications received during patient hospitalizations. Billing diagnoses in claims data reflect both actual clinical diagnoses and rule-out workup diagnoses; therefore, false-positive diagnoses are possible. In addition, while no systematic bias is expected, some differences in patient characteristics observed between age cohorts may be due to population turnover. Mean length of observation period was approximately 9–11 years; therefore individual patients observed in the study may not have been longitudinally present across all age cohorts. Finally, in the absence of an LGS-specific diagnostic code or other established method of predicting LGS status for comparison, the accuracy of the classifier in identifying true LGS patients could not be tested; however the proportion of probable LGS patients identified by the classifier was similar to that of previous studies [
      • Skornicki M.
      • Clements K.M.
      • O'Sullivan A.K.
      Budget impact analysis of antiepileptic drugs for Lennox-Gastaut syndrome.
      ,
      • Benedict A.
      • Verdian L.
      • Maclaine G.
      The cost effectiveness of rufinamide in the treatment of Lennox-Gastaut syndrome in the UK.
      ,
      • Verdian L.
      • Yi Y.
      Cost-utility analysis of rufinamide versus topiramate and lamotrigine for the treatment of children with Lennox-Gastaut Syndrome in the United Kingdom.
      ,
      • Clements K.M.
      • Skornicki M.
      • O'Sullivan A.K.
      Cost-effectiveness analysis of antiepileptic drugs in the treatment of Lennox-Gastaut syndrome.
      ], indicating that the algorithm is not likely to have missed LGS patients.
      The current findings indicate the need for increased clinical attention to LGS beyond pediatric years, with a focus on optimization of treatment for LGS patients of all ages with widest-spectrum AEDs. Furthermore, clinical attention should emphasize the use of appropriate AED treatment early on. Despite the availability of LGS-specific AEDs such as clobazam and rufinamide, LGS may not be appropriately treated with these widest-spectrum AEDs, leading to increased potential for long-term disease burden. Although LGS is typically considered to be a childhood encephalopathy and is most often diagnosed during childhood, its features continue into adulthood, requiring lifelong treatment and considerable healthcare resource utilization. Recognition and treatment of LGS in adults requires the same commitment to achieving seizure control and quality of life as in younger patients. Timely recognition and adequate treatment of LGS are likely to result in improved outcomes and less costly management of this condition.

      Disclosures and funding

      JEPG has served as a consultant and/or speaker for Allergan, Cyberonics, Eisai Co., Ltd., Lundbeck LLC, Sunovion Pharmaceuticals, Inc., Supernus Pharmaceuticals, UCB, and Upsher-Smith Laboratories, and has received research grants from Eisai, Co., Ltd., and UCB. GM has served as a consultant for Acorda Therapeutics, Inc., Eisai Co., Ltd., Lundbeck LLC, UCB, and Upsher-Smith Laboratories. PGD, WYC, ET, and MSD are employees of Analysis Group, Inc., and FV was an employee of Analysis Group, Inc. at the time the study was conducted. VS is an employee of Lundbeck LLC, and JI and TS were employees of Lundbeck LLC at the time the study was conducted. This study was funded by Lundbeck LLC.

      Acknowledgements

      Manuscript preparation, including writing, editing, and formatting the manuscript, incorporating author comments, and coordinating submission requirements, was provided by Jennifer Kaiser, PhD, of Prescott Medical Communications Group. This editorial support was funded by Lundbeck LLC .

      References

        • Bourgeois B.F.
        • Douglass L.M.
        • Sankar R.
        Lennox-Gastaut syndrome: a consensus approach to differential diagnosis.
        Epilepsia. 2014; 55: 4-9
        • Hancock E.C.
        • Cross J.H.
        Treatment of Lennox-Gastaut syndrome.
        Cochrane Database Syst Rev. 2013; 2: CD003277
        • Camfield P.R.
        Definition and natural history of Lennox-Gastaut syndrome.
        Epilepsia. 2011; 52: 3-9
        • Arzimanoglou A.
        • French J.
        • Blume W.T.
        • Cross J.H.
        • Ernst J.P.
        • Feucht M.
        • et al.
        Lennox-Gastaut syndrome: a consensus approach on diagnosis, assessment, management, and trial methodology.
        Lancet Neurol. 2009; 8: 82-93
        • Ferlazzo E.
        • Nikanorova M.
        • Italiano D.
        • Bureau M.
        • Dravet C.
        • Calarese T.
        • et al.
        Lennox-Gastaut syndrome in adulthood: clinical and EEG features.
        Epilepsy Res. 2010; 89: 271-277
        • Ogawa K.
        • Kanemoto K.
        • Ishii Y.
        • Koyama M.
        • Shirasaka Y.
        • Kawasaki J.
        • et al.
        Long-term follow-up study of Lennox-Gastaut syndrome in patients with severe motor and intellectual disabilities: with special reference to the problem of dysphagia.
        Seizure. 2001; 10: 197-202
        • Oguni H.
        • Hayashi K.
        • Osawa M.
        Long-term prognosis of Lennox-Gastaut syndrome.
        Epilepsia. 1996; 37: 44-47
        • Ohtsuka Y.
        • Amano R.
        • Mizukawa M.
        • Ohtahara S.
        Long-term prognosis of the Lennox-Gastaut syndrome.
        Jpn J Psychiatry Neurol. 1990; 44: 257-264
        • Pina-Garza J.E.
        • Chung S.
        • Montouris G.D.
        • Radtke R.A.
        • Resnick T.
        • Wechsler R.T.
        Challenges in identifying Lennox-Gastaut syndrome in adults: a case series illustrating its changing nature.
        Epilepsy Behav Case Rep. 2016; 5: 38-43
        • Yagi K.
        Evolution of Lennox-Gastaut syndrome: a long-term longitudinal study.
        Epilepsia. 1996; 37: 48-51
        • Glauser T.A.
        Following catastrophic epilepsy patients from childhood to adulthood.
        Epilepsia. 2004; 45: 23-26
        • Camfield P.
        • Camfield C.
        Long-term prognosis for symptomatic (secondarily) generalized epilepsies: a population-based study.
        Epilepsia. 2007; 48: 1128-1132
        • Autry A.R.
        • Trevathan E.
        • Van Naarden Braun K.
        • Yeargin-Allsopp M.
        Increased risk of death among children with Lennox-Gastaut syndrome and infantile spasms.
        J Child Neurol. 2010; 25: 441-447
        • Camfield C.
        • Camfield P.
        Twenty years after childhood-onset symptomatic generalized epilepsy the social outcome is usually dependency or death: a population-based study.
        Dev Med Child Neurol. 2008; 50: 859-863
        • Camfield P.R.
        • Gibson P.A.
        • Douglass L.M.
        Strategies for transitioning to adult care for youth with Lennox-Gastaut syndrome and related disorders.
        Epilepsia. 2011; 52: 21-27
        • Skornicki M.
        • Clements K.M.
        • O'Sullivan A.K.
        Budget impact analysis of antiepileptic drugs for Lennox-Gastaut syndrome.
        J Manag Care Spec Pharm. 2014; 20: 400-406
        • Benedict A.
        • Verdian L.
        • Maclaine G.
        The cost effectiveness of rufinamide in the treatment of Lennox-Gastaut syndrome in the UK.
        PharmacoEconomics. 2010; 28: 185-199
        • Verdian L.
        • Yi Y.
        Cost-utility analysis of rufinamide versus topiramate and lamotrigine for the treatment of children with Lennox-Gastaut Syndrome in the United Kingdom.
        Seizure. 2010; 19: 1-11
        • Clements K.M.
        • Skornicki M.
        • O'Sullivan A.K.
        Cost-effectiveness analysis of antiepileptic drugs in the treatment of Lennox-Gastaut syndrome.
        Epilepsy Behav. 2013; 29: 184-189
        • The International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM)
        (Available at:)
        • National Center for Health Statistics
        The international classification of diseases, ninth revision, clinical modification (ICD-9-CM).
        (Available at:)
        • Beaumanoir A.
        The Lennox-Gastaut syndrome: a personal study.
        Electroencephalogr Clin Neurophysiol Suppl. 1982; : 85-99
        • Hughes J.R.
        • Patil V.K.
        Long-term electro-clinical changes in the Lennox-Gastaut syndrome before, during, and after the slow spike-wave pattern.
        Clin Electroencephalogr. 2002; 33: 1-7
        • Roger J.
        • Remy C.
        • Bureau M.
        • Oller-Daurella L.
        • Beaumanoir A.
        • Favel P.
        • et al.
        Lennox-Gastaut syndrome in the adult.
        Rev Neurol (Paris). 1987; 143: 401-405
        • Montouris G.D.
        • Wheless J.W.
        • Glauser T.A.
        The efficacy and tolerability of pharmacologic treatment options for Lennox-Gastaut syndrome.
        Epilepsia. 2014; 55: 10-20