Cognitive Issues in the Treatment of Epilepsy Kimford Meador, MD Departments of Neurology & Pediatrics Emory University Atlanta, Georgia kimford.meador@emory.edu

International Bureau for Epilepsy: 2004 Cognitive Function Survey 44% Difficulty learning 45% Felt that they were slow thinkers 59% Felt sleepy or tired 63% AED effects prevented them from achieving activities or goals N = 425 Europeans with epilepsy www.ibe-epilepsy.org/whatsnew_det.asp

Factors Affecting Cognition and Behavior in Epilepsy Seizure- Related Variables Treatment- Related Variables All patients with epilepsy are at increased risk for cognitive and/or behavioral impairment, and the origin for this type of impairment appears to be multifactorial. These variables may be seizure- or non–seizure-related or related to treatment with certain AEDs. Clinician awareness of the factors that comprise these variables and their relationship with cognitive/behavioral impairment is vital within the context of managing epilepsy and choosing appropriate AED therapy. Non–Seizure- Related Variables

Lee KH et al, Neurology 2002;24:59(6):841-6 Patient with Complex Partial Seizure from Left Mesial Temporal Sclerosis This slide shows an example of CLOA group. Patient was a 17 year old male with complex partial seizure. MRI showed left mesial temporal sclerosis. Ipsilateral thalamic and midbrain hyperperfusion is demonstrated. Lee KH et al, Neurology 2002;24:59(6):841-6

Longitudinal Study of Hippocampal Atrophy 12 unilateral TLE patients Repeat MRI: mean 3.4 yrs (2.5-5.2yrs) Progressive hippocampal atrophy occurred only in patients with continuing seizures Mean 10% loss of hippocampal volume in patients with continued seizures TLE = temporal lobe epilepsy Fuerst D et al. Ann Neurol. 2003;53:413-416

Cross-Sectional Cognition Study in Temporal Lobe Epilepsy (TLE) FSIQ of WAIS-R in 209 patients with unilateral TLE. IQ lower if >30 year Seizure Duration than 15 - 30 years and <15 years. IQ for 15 - 30 years and <15 years seizure duration did not differ. Decline is in patients without seizure control. Jokeit H et al. J Neurol Neurosurg Psychiatry 1999;67:44-50

Cognitive Effects Wine AEDs Higher Dose/ABL Polytherapy Rapid Titration Habituation AED differences Individual differences AEDs = antiepileptic drugs

Cognitive Abilities Most Likely to be Affected by AEDs Processing Speed (e.g., reaction time) Complex or Sustained Attention Dual Processing Verbal learning Paragraphs more sensitive than word lists Verbal fluency Rate at which words beginning with a specific letter can be generated AEDs = antiepileptic drugs

Cognitive Effects of Older AEDs in Healthy Adults Carbamazepine (CBZ), phenytoin (PHT), or valproate (VPA) rarely differ. Phenobarbital significantly worse on about 1/3rd of tests than PHT or VPA. Patients statistically better on placebo than older AEDs for about 50% of tests. Meador KJ et al. Neurology. 1991;41(10): 1537-1540; Meador KJ et al. Epilepsia. 1993;34(1):153-157; Meador KJ et al. Neurology. 1995;45(8):1494-1499.

Healthy Volunteers: Newer AEDs vs Placebo % tests with placebo better than AED gabapentin 0 – 19% lamotrigine 1 – 17% levetiracetam 11% oxcarbazepine 46% topiramate 29 – 88% tiagabine 0% Kalviainen et al, Epi Res 1996;25:291-7. Dodrill et al, Neurology 1997;48:1025-31. Leach et al, JNNP 1997;62:372-6. Meador et al., Epilepsia 1999;40(9):1279-1285. Meador et al., Neurology 2001;56:1177-82. Salinsky et al., Epilepsy & Behavior 2004;5:894-902. Aldenkamp et al., Epilepsia 2000;41:1167-7. Meador et al., Neurology 2003;13;60:1483-8. Salinsky et al., Neurology 2005;64:792-8. Meador et al., Neurology 2005;64(12):2108-2115. Blum et al., Neurology 2006;67:400-406.

Healthy Volunteers: Newer AEDs vs Other AEDs LESS impact on cognition MORE impact on cognition % tests gabapentin carbamazepine 26% topiramate 50% lamotrigine 48% 80% levetiracetam 42% oxcarbazepine phenytoin 0% When comparing the newer AEDs to each other, performance while taking some medications are more affected than while taking other medications. This table demonstrates the impact of those differences. For instance, the first row demonstrates that patients taking gabapentin significantly outperformed patients taking carbamazepine on 26% of the tests taken by the groups. Meador et al., Epilepsia 1999;40(9):1279-1285. Meador et al., Neurology 2001;56:1177-82. Salinsky et al., Epilepsy & Behavior 2004;5:894-902. Meador et al., Neurology 2003;13;60:1483-8. Salinsky et al., Neurology 2005;64:792-8. Meador et al., Neurology 2005;64(12):2108-2115.

Incidence of Unprovoked Seizures in Developed Countries Cloyd et al. Epilepsy Res 2006;68 (Suppl 1): 39-48

Cognitive Effects of AEDs in the Elderly Phenytoin = Valproate Craig & Tallis, Epilepsia 1994;35:381-390 Elderly more sensitive to cognitive effects of AEDs: Carbamazepine, Phenobarb, Phenytoin, Primidone VA Cooperative Study

VA Coop Geriatric Epilepsy Study N = 593 >65 y/o New onset epilepsy Mean Dose (mg/d) ABL (mcg/ml) CBZ = 558 6.8 GBP = 1424 8.7 LTG = 152 3.5 LTG GBP CBZ Rowan et al, Neurology 2005;64:1868-73. CBZ=carbamazepine, GBP=gabapentin, LTG=lamotrigine

Cognitive Effects of AEDs in Children Loring & Meador, Neurology 2004;62:872-7 Pressler et al., Neurology 2006;66(10):1495-9. Donati et al, Neurology 2006;67:679-682.

Children AED Cognition Studies Vining et al, 1987 PB < VPA Farwell et al, 1990 PB < Placebo Forsythe et al, 1991 CBZ = PHT = VPA Chen et al, 1996, 2001 PB < CBZ = VPA Aldenkamp et al, 1998 Pressler et al, 2006 LTG = Placebo Donati et al, 2006 CBZ = OXC = VPA Kang et al, 2007 TPM < CBZ Levisohn et al, 2009 LEV = Placebo When comparing the newer AEDs to each other, performance while taking some medications are more affected than while taking other medications. This table demonstrates the impact of those differences. For instance, the first row demonstrates that patients taking gabapentin significantly outperformed patients taking carbamazepine on 26% of the tests taken by the groups. Vining et al, Pediatrics 1987;80:165-174; Farwell et al,,NEJM 1990;322:364-369; Forsythe et al, Dev Med Child Neurol 1991;33:524-534; Chen et al, Epilepsia 1996;37:81-86; Aldenkamp et al, Epilepsia 1998;39:1070-4; Pressler et al, Neurology 2006;66:1495-9; Donati et al, Neurology 2006; 67;679-682; Kang et al, Epilepsia 2007;48:1716-23 CBZ=carbamazepine, GBP=gabapentin, LEV=levetiracetam, LTG=lamotrigine, OXC=oxcarbazepine, PB=phenobarb, PHT=phenytoin, TPM=topiramate

MCG Stories: Delayed Recall % Compared to Non-Drug Average Healthy Volunteer Studies CBZ=carbamazepine, GBP=gabapentin, LTG=lamotrigine, PHT=phenytoin, TPM=topiramate. Meador et al, 1991, 1993, 2000, 2001, 2005

In Utero AEDs & Behavioral Neurodevelopment in Animals Phenobarb reduces brain weight & impairs behavior in mice. Phenytoin impairs coordination & learning in rats. Phenytoin can cause hyperactivity in monkeys. Neurobehavioral effects also found for valproate.

Neurodevelopment in Children of Women with Epilepsy Maternal seizure type # of seizures during pregnancy IQ & education of parents AEDs & other drugs Other environmental factors

Factors Affecting Cognitive Neurodevelopment When maternal IQ is controlled, no other single environment factor has a large effect. Heritability: 30-50% of IQ variance Sattler JM, 1992

Cognitive Effects of In Utero AEDs PHENOBARBITAL 2 retrospective Danish cohorts without maternal IQ (n=114 PB total): PB vs. general population: -7 VIQ1 PHENYTOIN Prospective without maternal IQ (n=20 PHT): PHT vs. controls: -8 IQ2 Prospective cohort (n=34 PHT, 36 CBZ): PHT not different when analyses using maternal IQ; also no effect for CBZ 3 Swedish (?prospective) cohort without maternal IQ (n= 67 PHT): PHT vs. unexposed controls: -8 IQ4 Two Class III for PB worse One Class II and Two Class III for PHT worse See next slide note for CBZ 1. Reinisch et al. JAMA 1995;274:1518-1525. 2. Vanderloop et al. Neurotox Terat 1992;14:196-92. 3. Scolnik et al, JAMA 1994;271:767-70. 4. Wide et al. Acta Paediatr 2002;409-14.

Cognitive Effects of In Utero AEDs: VALPROATE 2 retrospective cohorts from UK, which controlled for maternal IQ): VPA vs. other monotherapy or no AED Special education: 30% vs. 3-6%1 VPA group 6-16 years old: -10-14 VIQ2 (n=41 VPA) VPA group <6 years old: greater delay on SGS II (Schedule of Growing Skills II)1 (n=21 VPA) Prospective Finnish cohort without maternal IQ): VPA vs. CBZ: -12 VIQ3 (n=13 VPA MonoTx) No difference for CBZ vs. unexposed3 Two Class II for VPA worse Two Class II for CBZ not worse 1. Adab N, et al. J Neurol Neurosurg Psychiatry. 2001;70:15-21. 2. Adab N, et al. Neurol Neurosurg Psychiatry. 2004;75:1575-1583. 3. Gaily E, et al. Neurology. 2004;62:28-32.

NEAD Study Neurodevelopmental Effects of Antiepileptic Drugs 25 sites: USA & UK http://www.neadstudy.com Funded by NIH/NINDS #2RO1 NS 38455

STUDY DESIGN Multicenter prospective, parallel-group observational study with statistical control. Pregnant mothers with epilepsy enrolled from late 1999 to early 2004. AED monotherapy: Carbamazepine (CBZ) Lamotrigine (LTG) Phenytoin (PHT) Valproate (VPA) Blinded cognitive assessments: 2, 3, 4.5, & 6 y/o Primary outcome: IQ at 6 y/o

Neurodevelopmental Effects of Antiepileptic Drugs 309 mother/child pairs from 25 centers in US & UK Meador et al. NEJM 2009;360:1597-605 Funded by NIH/NINDS #2RO1 NS 38455 and #1 R01050659 Multicenter prospective, parallel-group observational study with statistical control. Pregnant mothers with epilepsy enrolled from late 1999 to late 2004. Primary outcome IQ at 6 y/o. Fetal valproate exposure related with lower IQ. Carbamazepine Lamotrigine Phenytoin Valproate Mean IQ 98 101 99 92 Difference 6 9 7 (CIs) (0.6:12.0) (3.1:14.6) (0.2:14.0)

Child IQ vs. Maternal IQ r = .23 r = .54 p<.001 p<.04 r = .49 r = .09 NS Pearson correlations (p values) by AED Group from multiple imputation analyses for Child IQ vs. Maternal IQ

Means (95% CIs) for Child IQ as Function of Dose and AED Group Median dosages: CBZ = 750 mg/day, LTG = 433 mg/day, PHT = 398 mg/day, and VPA = 1000 mg/day

Valproate Dose Effects NEAD Significant for both birth defects and IQ 24.2% > 900 mg/day vs. 9.1% < 900 mg/day North America Not significant 1033 mg/day (+434) with malformations vs. 983 mg/day (+431) without Australia Significant 34.5% malformations > 1400 mg/day vs. 5.5% at < 1400 mg/day Finland Significant 23.8% for doses >1500mg/day vs. 9.5% for doses <1500mg/day UK Not significant 9.1% >1000 mg/day, 6.1% 600-1000 mg/day, 4.1% <600 mg/day UK Liverpool Significant Reduce VIQ 15 points > 1500mg/d, 9.9 at 801-1500mg/d, 2.2 < 800mg/d Reduce VIQ 20 points > 1500mg/d, 16.6 at 800-1500mg/d, 4.2 < 800mg/d Sweden and GSK data Not analyzed for dose effect of VPA

Cognitive Effects of Levetiracetam Fetal Exposure Griffiths Mental Development Scale at age <24 mos Developmental Quotient in Children of: WWE on Levetiracetam (n=51): 100 WWE on Valproate (n=44): 88 Healthy women on drug (n=97): 99 Weaknesses: Young age at assessment Retrospective collection of seizures and alcohol & tobacco use during pregnancy Completer Rate: 58% LEV and 37% VPA Shallcross et al, Neurology 2011

Success with Antiepileptic Drugs Previously Untreated Epilepsy Patients (N=470) Not Sz Free Sz Free Kwan P, Brodie MJ. N Engl J Med. 2000;342(5):314-319

Anterior Temporal Lobectomy (ATL) 60-75% Seizure Free <5% Morbidity <1% Mortality Average duration epilepsy 20 years prior to surgery

Henry Gustav Molaison Patient HM Born: February 26, 1926 Surgery: September 1, 1953 (age 27) Died: December 2, 2008 (age 82) Severe anterograde declarative memory disorder Retrograde memory disorder back 11 years Intact: immediate memory, procedural memory, priming, & release from proactive interference Scoville WB, Milner B. Loss of recent memory after bilateral hippocampal lesions. J Neurol Neurosurg Psychiatr 1957;20:11-21.

Neuropsychological Effects of Anterior Temporal Lobectomy LEFT Naming Deficits Worsening of Verbal Episodic Memory RIGHT Non-Verbal Episodic Memory Deficits (less consistent & less clinically significance) Trenerry MR et al. Neurology 1993;43:1800-1805 Hermann BP et al. Behav Neurosci 1994;108:3-10 Helmstaedter C. Epilepsy & Behavior 2004;5:S45-S55.

Predictors of Greater Risk for Post-ATL Cognitive Decline ATL on language dominant side Older age of seizure onset Older age at surgery Higher pre-op cognitive performance No hippocampal atrophy/sclerosis Poor post-op seizure control Helmstaedter C. Epilepsy & Behavior 2004;5:S45-S55. ATL= anterior temporal lobectomy

Other Predictors of Post-ATL Cognitive Outcomes Wada test fMRI MRS PET Evoked Potentials from implanted electrodes ATL= anterior temporal lobectomy

Other Types of Epilepsy Surgery & Cognitive Risks Frontal Parietal Occipital Multiple Subpial Transections Callosotomy Hemispherectomy

Vagal Nerve Stimulator No cognitive side effects Apparent improvements in some patients probably related to reduced seizures & Antiepileptic Drugs. Dodrill & Morris, Epilepsy Behav 2001;2:46-53

Vickrey BG. Epilepsia. 1994;35:597-607 Comparison of Quality of Life With Seizures, HTN, Diabetes, & Heart Disease Seizure-free Auras Seizures Hypertension/ Diabetes Heart Disease N = 166 61 58 55 T-SCORE 52 HRQOL of 166 adult patients who had previously undergone surgical treatment for intractable epilepsy was compared with that of outpatients with hypertension, diabetes, and heart disease. QOL scales were converted to T-scores (the Y-axis variable in this figure) by transforming raw scores linearly to produce a mean of 50 and SD of 10 in the combined sample of patients with chronic diseases (Medical Outcomes Study) and epilepsy patients. Results showed that 55 completely seizure-free patients scored higher than patients with hypertension in 6 of 9 HRQOL domains, higher than diabetic patients in 8 of 9 domains, and higher than those with heart disease in all 9 domains. Patients who continued to have seizures with altered consciousness postoperatively scored significantly worse than patients with hypertension, diabetes, or heart disease on overall QOL and emotional well-being, and worse than patients with hypertension or diabetes on social function. In contrast, patients with seizures scored significantly higher than patients with diabetes and heart disease on general health perceptions. From this study, it appears that epilepsy surgery patients who become seizure-free for at least 12 months report better HRQOL than patients with hypertension, diabetes, or heart disease on most HRQOL dimensions. 49 46 Overall Quality of Life Emotional Well-Being Social Function Role– Emotional Energy/ Fatigue Pain Role– Physical Physical Function Health Perception Vickrey BG. Epilepsia. 1994;35:597-607

Relationship of Subtle AED Toxicity Adverse Events Profile to Quality of Life QOLIE-89 Total Score QOLIE-89 Total Score The impact of adverse AED effects on quality of life was studied in 194 patients with refractory epilepsy considering evaluation for epilepsy surgery. This slide shows a scatterplot of the correlation of the Adverse Events Profile (AEP) Summary Score with QOLIE-89 scores. The AEP is a reliable and valid measurement of the burden of common AED side effects. The correlation of quality of life with AEP total scores was highly statistically significant (P<.0001). Adverse Events Profile Summary Score Average Monthly Seizure Rate N = 200 r = -0.76, P<0.0001 Gilliam, et al. Neurology 2004;62:23-27

Mood, Quality of Life, & Neuropsychological Function Subjective Mood Best All Objective Objective Test Tests Memory 17.2% 4.3% 7.9% Language 14.6% 4.9% 12.7% Attention 28.7% 3.6% 9.3% QOLIE-89 total 46.7% 5.2% 13.3% % Variance explained by each factor; N = 257 epilepsy patients Perrine et al, Arch Neurol 1995;52:997-1003

Summary: Cognition & Epilepsy Cognitive impairment in epilepsy is multifactorial. Least cognitive effects: GBP, LEV, TGB, LTG. Intermediate effects: CBZ, PHT, OXC, VPA. Most adverse effects: PB, TPM, Benzos. AED susceptibility can vary across patient groups as well as across individual patients. Subjective and objective measures of cognitive function can dissociate. Benzos=benzodiazepines, CBZ=carbamazepine, GBP=gabapentin, LEV=levetiracetam, LTG=lamotrigine, PB=phenobarbital, PHT=phenytoin, OXC=oxcarbazepine, TGB=tiagabine, TPM=topiramate, VPA=valproate.

Prevalence of Psychiatric Disorders in Epilepsy Depression 11%–60% Anxiety 19%–45% Psychosis 2%–8% Anthony, et al. Epidemiol Rev 1995;17: 240-2 Weissman, et al. J Clin Psychopharm 1986; Suppl 6:11-17 Kessler, et al. Arch Gen Psych 1994;51:8-19

Behavioral & Psychotropic Effects of Antiepileptic Drugs Most of the AEDs can produce untoward subjective side effects CBZ, LTG, & VPA have proven efficacy in bipolar disorder. GBP & TPM used in add-on. AEDs are used in variety psych. disorders (eg, VPA in agitation & GBP in social phobia) CBZ=carbamazepine, GBP=gabapentin, LTG=lamotrigine, TPM=topiramate, VPA=valproate.