Epilepsy: Prognosis and Treatment
William H Theodore MDChief, Clinical Epilepsy Section
National Institute of Neurological Disorders and StrokeNational Institutes of HealthBethesda, Maryland, USA
Prevalence and Incidence
• Third most common neurologic disorder• First seizure incidence: 20-70 / 100,000 • Epilepsy incidence: 30-50 / 100, 000• Prevalence: 5-10 / 1000
– Reported higher in some developing countries
• Cumulative adjusted lifetime risk: 1.3%–3.3%
Hauser WA, Hesdorffer DC. Epilepsy: Frequency, Causes, and Consequences. New York, NY: Demos; 1991:1.
Annegers 1993
Etiology of Symptomatic EpilepsyUSA
0
10
20
30
40
50
60
70
80
<15 15-24 25-44 45-64 >65
devel
infection
trauma
CVD
tumor
degen
Epidemiology by Seizure Types
Reproduced with permission from Hauser WA. Epilepsia. 1992;33(suppl 4):S10.
Complex Partial (36%)
Unclassified (3%)Myoclonic (3%)
Absence (6%)Partial Unknown
(7%)
Other Generalized (8%)
Simple Partial (14%)
Generalized TC (23%)
Prognosis After a Single Seizure
• Reported 30-70% recurrence over 3 years– sampling, etiology, seizure types– Increased if underlying lesion– Decreased if avoidable acute precipitant
• CBZ reduced recurrence in children (Camfield 1989)
– 1/3 stopped drug due to side effects• 18% Rx vs 38% no RX in 2 years
– PHT, CBZ, VPA, PB (First Seizure Trial Group 1989)
AED Peak Plasma concentration
Protein binding
clearance T ½ Drug interactions Therapeutic level (μmol/L)
lamotrigine 1-3h 55% hepatic 15-60 AEDs 10-60
gabapentin 2-3hdose
0 renal 6-7h# minimal 40-120
tiagabine 1-2h 96 CYP3A 5-8h AEDs #
vigabatrin 1-2h 0 5-7h# #
Topiramate 2-4h 15 mixed 18-23h Lithium, OCs, some AEDs 10-60
Oxcarbazepine(MHD metabolite)
1-2h 40 Non-CYP mediated
10-12 hr(MHD metabolite)
AEDsoral contraceptives
50-140 (MHD)
Felbamate 2-6h 22-25 hepatic 15-23hr AEDs 200-400
Phenobarbital 1-4 h 40-55 hepatic 80-130 extensive 50-130
Phenytoin 2-6 hr 90 Hepatic*** extensive 40-80
Carbamazepine Slow, variable 70-75 hepatic 18-55 hr*12 hr**
extensive 15-45
Levetiracetam 1-2 h 0 Renal 6-10 hr minimal #
Zonisamide 3-4 h 40-60 CYP3A 50-60 hr extensive 35-200
Valproic acid 1-2 h 90& Hepatic 10-15 hr AEDs 300-600
Ethosuximide 3-5 h 0 hepatic 30-60 hr AEDs 300-600
DrugSodium
channelsCalcium channels
GABA system
Glutamate receptors
Clinical Efficacy
LRE PGE SGE
Phenytoin ++ YN N
Carbamazepine ++ Y N N
Oxcarbazepine ++ Y N N
Lamotrigine ++ + Y Y Y
Zonisamide ++ + Y
Valproate + + + Y Y Y
Felbamate + + + + Y Y
Topiramate + + + + Y Y
Ethosuximide ++ + N Y N
Gabapentin ++ + Y
Levetiracetam + + Y
Phenobarbital + + + Y
Epilepsy Therapy in 525 Patients
Kwan and Brodie 2000
Veterans Administration Cooperative Study
Reproduced with permission from Mattson RH, et al. N Engl J Med. 1985;313:145-151.
Perc
en
t C
onti
nuin
g
phenobarbitalphenytoinprimidonecarbamazepine
100
80
60
40
20
00 3 6 9 12 15 18 21 24 27 30 33 36
Months
Marsan et al 2007
Time to 12 month remission % remaining on drug
SANAD Study
Mattson et al 1985
Reasons for AED FailureVA Cooperative Study
CBZN=101
PHTN=101
PBN=110
PRMN=109
AllN=421
Toxicity 12 19 18 36 85
Toxicity+ seizures
30 33 29 25 127
Seizures 3 4 1 3 11
Total 45 56 48 74 233
Callhagan et al 2008
Prognosis of ‘Drug-Refractory’ EpilepsyRe-evaluation of 246 patients
• Drug failure before index date:– maximum tolerated dose in
54%– idiosyncratic reaction in 6.5%– intolerable side effect in 19%– unknown reasons in 21%.
• 6-month terminal seizure remission: – 14% of AED-treated patients
(about 5% per year of study)– 52% of surgery patients
• persistent intractability: • Duration > 10 years, mental
retardation, status, > 6 AEDs No drug seemed superior
Some Emerging AEDsAED CPS
(> placebo)
PGE SGE toxicity
Brivaracetam ↓ 22% 78% ↓ photosensitity
GI
Carisbamate ↓18-20% CNS
Eslicarbazepine ↓ ~ 20% CNS, GI
Lacosamide ↓ 20-25% CNS, GI
Retigabine ↓ 20-25% CNS
Rufinamide ↓ 20% total
↓ 40% atonic
CNS, GI
Why do AEDs Fail?
• About 30% of patients do not respond at all• About 10% of patients with good initial AED
response cease to respond• Pharmacokinetic
– Drug interactions
– Enzyme induction
• Tolerance to non-BZP AEDs ?– Receptor, channel response changes
• Drug efflux transporters– PgP, MRPs,
AED Tolerance
Loscher & Schmidt 2006
• Long-term BZPs: ↓ allosteric GABA-BZP site interactions
• VGB tolerance in MES model: ↓ GAD due to GABA feedback inhibition
Altered NA+ Channel Responses?
Remy et al 2003
No MTS
MTS
Multiple Drug Transporters (p-glycoproteins)
• Pump lipophilic drugs and other xenobiotics out of cells– Role in cancer
chemotherapy resistance
• May be overexpressed in human epileptic tissue, especially TLE
• Unreplicated link between MDR gene polymorphisms and human AED resistance
Loscher 2007
Loscher 2007
PgP Affects Brain Phenytoin Levels
Possible Therapeutic Maneuvers
• Manage with drug holidays, dose adjustments?– Alternate AEDs?
• Lower starting doses?• Cross-tolerance ?
– Choose drugs with different mechanisms?
• PgP inhibition– verapamil– tariquidar
Natural History of Epilepsy
• Natural history of untreated epilepsy unknown– Bromides since 1857– PB available since
1912
Alfr
ed H
aupt
man
Cha
rles
Loco
ck
Natural History of Epilepsy
• Natural history of untreated epilepsy unknown.– Course may
‘fluctuate.’
• No difference in seizure-free rate if treatment begun after 1st or 2d seizure
• In ‘resource poor’ countries, spontaneous remission rate ~ 30%– prognosis not related
to pretreatment GTCS #
Hauser et al 1998
Berg et al 2003
Early onset LRE may not become clearly
intractable for many years • 7 centers: 333 patients evaluated for
resective surgery for localization-related epilepsy prospectively identified at initial evaluation
• Latency from epilepsy onset to 2 AED failure 9.1 years
• 26% reported at least 1 yr remission• 8.5% 5 year remission
Intractable Epilepsy:Comparison of Diagnostic Criteria
Berg et al Epilepsia 2006
Kwan et al Epilepsia 2009
ILAE Epilepsy Outcome CategoriesSeizure Control Side Effects Outcome
Seizure-free* No 1A
Yes 1B
undetermined 1C
Treatment failure No 2A
Yes 2B
undetermined 2C
Undetermined No 3A
Yes 3B
undetermined 3C*at least 12 months AND three times the longest interseizure interval in 12 months prior to new intervention
Kwan et al Epilepsia 2009
Drug Resistant EpilepsyILAE 2009
• Failure of informative trials of two tolerated and appropriately chosen and used AED schedules (whether as monotherapies or in combination) to achieve sustained seizure freedom.
Kwan et al Epilepsia 2009
Data Needed to Determine if a Therapeutic
Intervention is “Informative”
• Mode of application (e.g., formulation, dose, dosing interval)
• Compliance• Duration of exposure• Was there was effort to optimize dose?• Reason(s) for discontinuation
– Unsatisfactory seizure control – Adverse effects – Psychosocial reasons, for example, planning for
pregnancy – Administrative reasons, for example, lost to follow up – Financial issues, for example, cannot afford drug– Other reasons
Berg et al 2003
Early onset LRE may not become clearly
intractable for many years • 7 centers: 333 patients evaluated for
resective surgery for localization-related epilepsy prospectively identified at initial evaluation
• Latency from epilepsy onset to 2 AED failure 9.1 years
• 26% reported at least 1 yr remission• 8.5% 5 year remission
Predicting Intractable Epilepsy• Epilepsy ‘Pattern:’• Remittent
– KCNQ2 or KCNQ3 benign familial convulsions
– Some absence• Non-remittent ‘drug
responsive’– JME
• Non drug-responsive but treatable– Localization-related
• Poorly responsive– LGS
• Clinical Features at Onset:
• Early age of onset
• presentation in status epilepticus ?
• abnormal neurological exam
• partial seizures at diagnosis
• mixed seizure types ~ developmental delay
• multiple seizures prior to treatment
• seizure clustering, ‘density’
• Structural lesion
Spooner et al 2006
New onset TLE in Children: MRI and Prognosis
Sillanpaa et al 1999
Prospective Study of Finnish Children
1964-1992
0
10
20
30
40
50
60
70
80
90
IdioPE IGE SympPE SecGE
%
remission no remission dead
Drug Therapy: Prognosis by Seizure Type (n=1102)
0
10
20
30
40
50
60
GTC Mixed CPS
VA118-12VA118-24VA264-12VA264-24
Mattson et al 1996
What is Intractable Epilepsy?(modified after DC Taylor)
• The Lesion or Disease:– mesial temporal sclerosis, malformation
• The Illness:– intermittent seizures
• The Predicament:– social– psychological– economic
• AEDs treat the illness, not the disease– Is that important?
Progression of Epilepsy• “The interparoxysmal
mental state of epileptics often presents grave deterioration.”
• “Each fit apparently leaves a change in the nerve centers, facilitating the occurrence of other fits.”– Gowers 1890
• “Mental deterioration follows relentlessly.’’
– Cecil’s Textbook of Medicine 1929 Edwin G ZabriskieAssociate Professor of Neurology, Columbia UniversityPhysician to the Neurological Institute
Silanpaa et al 1998; Jokeit et al 2000; Helmstaedter et al 2003
Neuropsychological and functional Prognosis in TLE
• Surgery accelerates decline if unsuccessful
• Stops or reverses it if successful
• In Finnish pediatric study, adverse socio-economic effects even in patients who entered adult life in remission off AEDs
Cramer et al 2003, Ettinger et al 2004, 2005, Kobau et al 2006
Depression and Epilepsy
• Depression in Population > 18 survey data– 36.5% epilepsy– 27.8% asthma– 11.8% control– Adults ever told of epilepsy: RR 2.5 – Adults with active epilepsy: RR 3.0
• Reduced quality of life• Increased medical resource use
Quality of Life• Seizure control usually considered most
important measure• Complete seizure-freedom usually has a much
greater effect on HRQOL measures than simply reduced frequency
• Depression has greater adverse impact than seizure frequency itself in some studies
• Drug side effects and unemployment– Issue of when to withdraw drugs after successful
surgery
Devinsky et al Neurology 2005; Baker Neurology 2006
Seizure Control, Depression, and Anxiety
• Several studies suggest seizure frequency predicts anxiety and depression symptoms
• Multicenter surgery study– ↓ depression ~ seizure
control– 6.1% new depression in
non-seizure free patients
0
5
10
15
20
25
% depressed % anxious
preop NSF SF
Silanpaa et al 1998; Sperling et al 1999
Death• Standardized mortality ratio is increased in
epilepsy, even if no underlying illness• Marked increase in sudden unexplained death
– SUDEP related to: – GTCS– > 2 AEDs
• Death after TLE – SMR for patients with recurrent seizures 4.69– seizure free patients: no difference vs age- and sex-
matched population of the United States• Persistent seizures ~ death in Finnish pediatric
study• Death is due to uncontrolled epilepsy
Approaches to Intractable Epilepsy
• Surgery– Focal resection– hemispherectomy– Callosotomy (palliative)
• Ketogenic Diet
• Experimental Drugs
• Brain Stimulation
Wiebe et al 2001
0
10
20
30
40
50
60
70
surgical medical
Seiz
ure-
free
One year2-10 years
Helmstaedter et al 2003
Controlled Temporal Lobectomy Trial
‘Intractable’ TLE:Comparison of Medical and Surgical Outcome
Non-randomized Clinical Series
The Ketogenic Diet
30% Fat
50% Carbs
20% Protein
85% Fat
5% Carbs10% Protein
Potential Mechanisms of Action
• Ketosis• Acetone• Aspartate, GABA• Polyunsaturated
fatty acids • Mitochondrial
uncoupling• Glucose modulation• Enhanced
glutamate transport• Opening KATP
channels• Acidosis• Caloric restriction• Decreased IL-1ß• Neurosteroids
Ketogenic Diet
• Traditionally started gradually in the hospital after a 24-48 hour fast– Families educated daily
• Ratio (fat: carbs and protein)– 4:1 more strict– 3:1 for infants, adolescents
• Calories 60-100% • Fluids 85-100%• Solid foods and/or formula• Requires dietician support• Strong family committment
Side Effects
• Constipation• Slowed weight gain• Acidosis when ill• Vitamin deficiency (if unsupplemented) • Renal stones • Impaired height and weight• Dyslipidemia • Gastrointestinal upset
Ketogenic Diet Randomized Controlled Study
Neal et al Lancet Neurology 2008
10/65 who stopped diet not included in analysis
Brain Stimulation for Epilepsy• Vagal Nerve
Stimulation• Transcranial Magnetic
stimulation• Intracranial stimulation
– Surface electrodes (‘responsive’)
– Deep Brain Stimulation
• Hippocampus• Thalamus• Cerebellum
Torpedo fuscomaculata
VNS
• Requires surgery, but extracranial
• Effects broadly comparable to new AED trials
• 30-40% ≥ 50% seizure frequency reduction
• In open label extension effect sustained ≥ 12 months
• Very rare patients seizure-free
• Only consider when no chance for resective surgery
-30
-25
-20
-15
-10
-5
0
% reduction versus
baseline
EO3 (p<.05) EO5 (p<.001)
high low
Refractory Generalized EpilepsyNei et al Epilepsia 2006
Transcranial Magnetic Stimulation
TMS in Epilepsy
• TLE: – Case reports and open
trials:• 30-70% seizure decreases
reported
– Blinded controlled trial • 16% reduction > placebo
(0.05<p<0.10)• Effect lasted 2-4 weeks
• Cortical Dysplasia– significantly decreased
the seizures in active compared with sham rTMS group
~4 cm
Thalamic Stimulation• Centromedian
– Uncontrolled studies reported improvement– Small controlled study: no effect
• Anterior– Recent controlled study showed seizure ↓
• 14.5% in the control group• 40.4% in the stimulated group
• Subthalamic– Improvement in uncontrolled studies
Long-term follow-up of patients with thalamic deep brain stimulation for epilepsy
• Long-term follow-up (mean, 5 years) – 6 patients with anterior (AN)– 2 centromedian thalamic deep brain stimulation
• Five patients (all AN) had 50% seizure reduction– benefit was delayed in two until years 5 to 6– after changes in antiepileptic drugs.
• Seizure reduction 1 to 3 months before active stimulation– Possibility of a beneficial microthalamotomy effect.
Andrade et al Neurology 2006
Hippocampal Stimulation
• Reduced CPS frequency reported in several uncontrolled studies
• One small controlled study: • Four patients with refractory MTLE
– Risk to memory contraindicated temporal lobe resection• Double-blind stimulation randomly turned ON 1 month and
OFF 1 month for 6 months• Median reduction in seizures of 15%
– Effects seemed to carry over into the OFF period
– Possible implantation effect. • No adverse effects. • One patient treated for 4 years has substantial long-term
improvement.Tellez-Zenteno et al NEUROLOGY 2006;66:1490–1494
Seizure Prediction
Energy level (red)decision threshold (blue)prediction output (green)seizure onset (black)Positive outputs (high level in green curve) observed ~ 2 h before seizures.
Esteller et al Clin Neurophysiol 2005
RNS™ Placement
Courtesy of Martha Morrell
Anterior Lead (A)
Posterior Lead (P)
ParahippocampalLongitudinal Strip (not connected)
Courtesy of Martha Morrell
Preliminary RNS Efficacy (n=65)Initial 84 days Most recent 84 days
Seizure-type
% with 50% ↓
Overall % ↓
% with 50% ↓
Overall % ↓
CPS 32 27 40 34
GTCS 63 59 55 66
Total Disabling
26 29 41 35
Barkley et al AES 2006
Risks of Brain Stimulation• TMS
– Rare seizures at high (>10hz) frequency• Epilepsy therapy trials are at ≤ 1 hz
– Mild headache, scalp discomfort• VNS
– Cough, Hoarseness when stimulator on– dyspnea, pain, paresthesia, and headaches– respond to alteration of stimulation settings– Very rare vocal cord paralysis, bradycardia during implant
• DBS– Bleeding– infarction– intracranial infection– All less likely with surface RNS
