close

Вход

Забыли?

вход по аккаунту

?

Discontinuing antiepileptic drugs in children with epilepsy A prospective study.

код для вставкиСкачать
ORIGINAL ARTICLES
.
Discontinuing Antiepileptic Drugs in
CMdren with Epilepsy: A Prospective Study
Shlomo Shinnar, MD, PhD,*tly[-t+ Anne T. Berg, PhD,** Solomon L. Moshe, MD,*f$Y Harriet Kang, MD,*tB
Christine ODell, RN, MSN,$ll Marta Alemany,ll Eli S. Goldensohn, MD,"B and W. Allen Hauser, MDti
I n a prospective study, antiepileptic drugs were discontinued in 264 children with epilepsy after a mean seizure-free
interval of 2.9 years. They were then followed for a mean of 58 months to ascertain whether seizures recurred. Seizures
recurred in 95 (36%) of the children. Etiology was a significant predictor of outcome (relative risk [RRI = 1.81). O n
multivariable analysis, significant factors in the idiopathic group included age at onset above 12 years (RR = 5.4), a
family history of seizures (RR = 3,1), the presence of slowing on the electroencephalogram prior to medication
withdrawal (RR = 2.4), and a history of atypical febrile seizures (RR = 2.8). Specific epileptic syndromes such as
juvenile myoclonic epilepsy and benign rolandic epilepsy were also significant predictors of outcome. I n the remote
symptomatic group, significant predictors of outcome included age at onset older than 12 years (RR = 3.6),moderate
to severe mental retardation (IQ < 50) (RR = 2.8), a history of atypical febrile seizures (RR = 2.0), and a history of
absence seizures (RR = 0.4). The majority of children with epilepsy in remission while on antiepileptic drug therapy
will remain seizure free when medications are withdrawn. A few readily available parameters distinguish those with
a good prognosis from those in whom seizures are likely to recur. These data provide the framework for the clinical
decision making for withdrawal of medications in these children.
Shinnar S, Berg AT, MoshC SL, Kang H, O'Dell C, Alemany hi, Goldensohn ES,
Hauser WA. Discontinuing antiepileptic drugs in children with epilepsy:
a prospective study. Ann Neurol 1994;35:514-565
Previous studies have shown that approximately 70%
of children with epilepsy who are free of seizures for
a period longer than 2 or 4 years while on antiepileptic
drugs (AEDs) will remain so when medications are
withdrawn [I-20). Whether one can reliably differentiate subgroups of children with a relatively good prognosis from those with a poor prognosis remains controversial [l-6, 19-21). Factors suggested to distinguish
those with good prognosis in terms of successful discontinuation of medication versus a poor prognosis include the etiology of the seizures, the age of onset, the
electroencephalogram (EEG), and the seizure type(s).
This study was designed to test the predictive ability
of these variables in a large cohort including children
with both idiopathic and remote symptomatic epilepsy,
Materials and Methods
Subjects
In a prospective cohort study, AEDs were discontinued in
264 children with epilepsy who were in remission while on
medication and who were seen at hfontefiore Medical Center, Bronx Municipal Hospital Center, and our private practices between October 1983 and October 1990.
From the Departments of 'Neurology, tpediatrics, $Neuroscience,
and SNursinr. and Bthe MontefioreiEinsrcin Epilepsv
- _ . Manwement
Center, Montehore Medical Center, The Albert Einstein College of
Medicine, Bronx; the '*Deparrmenr of Pediatrics, Yale University
School of Medicine, New Haven, CT; and +?TheGertrude Sergievsky Center, Columbia College of Physicians and Surgeons, New
Candidates for this study were children with epilepsy who
were seizure free for 1 or more years and who were about,
as part of their clinical care, to have their antiepileptic medications discontinued. It is the general philosophy of our division to treat children until they have been seizure free for at
least 2 years prior to attempting medication withdrawal but
this interval may be reduced or prolonged for individual patients. To be eligible for entry into the study, children had to
have epilepsy, defined as two or more unprovoked seizures
occurring at least 24 hours apart {22). Children with neonatal, febrile, or other acute symptomatic seizures 1221 were
excluded if they did not also have a clear history of epilepsy
as defined above. Children with epilepsy who were seizure
free while o n AEDs but who were remaining on medications
at the request of their physician or the family were not eligible for entry into the study.
The decision to withdraw AEDs was made after discussion
of the relative risks and benefits with the family, including a
discussion of known risk factors based on results of prior
studies. The patient and the family were then asked to participate in the observational outcome study. Informed consent
and, when appropriate, assent were obtained and the patient
was enrolled in the study. Participation in the study carried
no risk and only four eligible families refused to participate.
Received Sep 9, 1993, ,and in revised form Dec 3 . Accepted for
publication Dec 6. 1993.
iddress
correspondence
ter, Montebore
10467.
York, NY.
534 Copyright 0 1994 by the American Neurological Association
Dr Shinnar,
Epilepsy
Center,
Cen-
E. ZIOth Street, Bronx, NY
Prior to withdrawal of medications, a detailed history including age at onset; cause, number, and type of seizures;
length of time on medication; and family history of seizures
was taken. A neurological examination was also performed
and included an assessment of intelligence.
Seizures and Epileptic Syndromes
Seizures were classified in accordance with the revised international classification [23]. Some patients had more than one
type of seizure. Atypical febrile seizures were defined as seizures with fever in children who had already had prior afebrile seizures, or complex febrile seizures [24,251 in children
who then had epilepsy [I]. Status epilepticus was defined as
a seizure lasting more than 30 minutes or recurrent seizures
lasting a total of more than 30 minutes without the patient
fully regaining consciousness [26, 271. Whenever possible,
an attempt was made to classify the epileptic syndrome using
the proposed international classification {28,291. The group
labeled as having temporal lobe epilepsy fit the criteria for
this syndrome, including the presence of focal temporal
spikes or slowing on at least one EEG, though not necessarily
the one prior to discontinuation of medications 1281.
Etiology
The etiology of the seizures was classified as remote symptomatic if the child was known to have a static encephalopathy from birth prior to the seizure andlor had sustained a
prior neurological insult such as a stroke or significant head
trauma (associated with a depressed skull fracture, loss of
consciousness for longer than 30 minutes, or intracranial
bleeding) 122,301. All other unprovoked seizures were considered idiopathic. A history of “birth trauma,” a stormy perinatal course, neonatal seizures or other provoked seizures,
or “mild” head trauma in a neurologically intact child did not
preclude the diagnosis of an idiopathic seizure. This epidemiological definition, while recognizing the difficulties in clearly
determining etiologies in children, especially in relationship
to previous neurological problems or perinatal problems, assigns a remote symptomatic etiology only in those patients
for whom the condition has been clearly associated with an
increased risk of subsequent seizures. It is consistent with
the recently published guidelines for epidemiological studies
o n epilepsy 1301.
Elect roencephalogram
An EEG was obtained for 250 (955%)of the children prior
to withdrawal of medications. Only EEGs obtained near the
time of withdrawal were used as study EEGs in the analysis.
For 228 (91%) of these children, one or more previous EEGs
were available for comparison. These prior EEGs were also
used to classify the epileptic syndromes. All EEG examinations were routine studies that usually included both an
awake and sleep record, induced hyperventilation, and photic
stimulation. No quantitative spike counts were performed. In
younger children and in some older, neurologically impaired
children, only a sedated sleep record could be obtained. With
the exception of a few records from outside institutions,
EEGs were interpreted by at least one of us (E.S.G., H.K.,
or S.L.M.), all of whom were blinded to the outcome in
terms of seizure recurrence.
EEGs were classified as normal or abnormal. Epileptiform
abnormalities including focal spikes, multifocal spikes, rolandic spikes, generalized spike and wave complexes, and photoparoxysmal responses as well as focal and generalized slowing
were each coded separately. Slowing was coded regardless of
whether epilepdforrn abnormalities such as focal spikes or
generalized spike and wave complexes were also present.
When previous records were available, the EEGs were also
compared for any major changes.
Family Hirto y
Family history was considered “positive” if there was a history
of one or more unprovoked seizures in a first-degree relative
(parent or sibling) C221. Family history of any seizures in
second-degree or more distant relatives was ascertained but
not considered in this analysis. Similarly, a family history of
febrile seizures or other provoked seizures even in a firstdegree relative was coded but not counted as a history of
epilepsy.
Intelligence
Estimates of intelligence were made clinically in all children.
When available, formal psychological testing results were
used. Due to the requirements of the special education system, formal assessments were available for almost all children
with mental retardation (MR) though not for all children
classified as normal. Intelligence estimates were as follows:
normal (1Q > 801, borderline ( I Q 70-80), mild MK (IQ
50-70), moderate MK (IQ 35-50), and severe MR (IQ <
3 5 ) . For purposes of analysis the children with an IQ less
than 50 were combined into the category “moderate-severe
MR.” A motor deficit was classified as present if there was
clearly an abnormality on motor examinacion (e.g., hemiparesis, quadriparesis, etc.) regardless of the severity of the deficit. Soft signs such as clumsiness or posturing during tandem
walking were coded but not classified as an abnormality for
this analysis.
Rate of Drug Withdrawal
The children usually had a fairly rapid withdrawal of AED
therapy. In a few cases, failure of the patients to follow instructions resulted in a longer period of tapering of dosage.
If a patient was taking more than one medication (n = 1I),
one drug was completely stopped before tapering of the next
WE. hegun. Of the 253 children taking one AED, 77 (-30%)
had a medication withdrawal period of less than 1 month;
139 (55CF), of 1 to 3 months; 26 (lo%), of 3 to 6 months;
of more than 6 months. Of the 11 children
and 11 (47;),
taking more than one AED, medication withdrawal took
place in less than 3 months in 5 ( 4 5 % ) , in 3 to 6 months in
3 (27%1), and in more than 6 months in 3 (27%).
FOllW-Up
After enrollment, patients were followed by telephone interviews at least every 3 months for ascertainment of any seizure
recurrence. The majority of children had several clinic follow-ups as well. As seizures in this population rarely occur
in the physician’s office, ascertainment of seizure recurrence
is done by history. Therefore, telephone determination
should be as reliable as that obtained by direct contact in
Shinnar et al: Discontinuing Antiepileptic Medications
535
clinic follow~-up.In those children with a recurrence, records
of any emergency medical care were reviewed and the children were seen for reevaluation. A recurrence was defined
as any unprovoked seizure that occurred after the initiation
of medication withdrawal. Entry into the study was defined
as the day medication withdrawal began.
0.9
"O
P
Analysis
As length of follow-up influences the probability of observing
a recurrence, the statistical methods of analysis took into
account the variable length of follow-up for each child 131371. Univariate analyses for dichotomous variables were performed using Kaplan-Meier survival analysis 1, 34, 36, 37)
and statistical significance was calculated using the method of
Mante {ji'). Results are displayed as Kaplan-Meier survival
curves. with the cumulative probability of seizure recurrence
plotted as a function of time from the first seizure. Ninetyfive percent confidence intervals (CIS)for the Kaplan-Meier
estimates of recurrence risks were calculated at 6, 12, 24,
and 4 8 months using an approximate Greenwood formula
for the standard error [34, 37). Continuous variables were
examined using t tests. Univariate analyses for all risk factors
and multivariable analyses were performed using the Cox
proportional hazards model [33, 341. All p values are twotailed.
Results
Cohort Characteristics
The study group included 264 children, 136 (525%)
boys and 128 ( 4 8 % ) girls. The predominantly innercity minority population was representative of the mixture of patients seen at our institutions. It consisted of
106 (405&)white children, 84 (32%) Hispanic children, 61 ( 2 3 $ 7 )black children, and 13 ( 5 % ) children
of other ethnic origins. The mean age at entry into the
study was 1 1.8 years. Twenty-six (10%) of the children
were under 5 years old; 7 5 (28c/r),between 5 and 10
years; 79 (30%!), between 10 and 15 years; and 84
(32%), older than 15 years at the time of study entry.
Other characteristics of the cohort are summarized in
Table 1.
Tuble I . Clinicul Charucteristzcs of Cohort
No. of Children
Idiopathic
Remote symptomatic
Mean age of onset (yr)
Mean duration of AED therapy (yr)
Mean age at seizure control (yr)
Mean age at medication withdrawal (yr)
Yearb beizure free prior to medication
withdrawal (mean)
1-2
2 64
165 (63%)
99 ( 3 7 % )
5 9 ISD 4.44)
5.2 (SD 4.0')
9.0 (SD 4 81)
11 8 (SD 4 97)
2- 7,
3-4
29
60
120
42
(16F)
24
42
(16%)
SU
=
standard devtarion; AED
536 Annals of Neurology
=
(SD 1.67)
(23%)
(457:)
antiepileptic drug
Vol 35
No 5
May 1994
.5
5
[r
c
0
._
0.6
h
4
I
0.5 0.4
~
0.3 -
Q
0
a
0.2 -
0
12
24
36
TIME (months)
48
60
72
Fig 1 . Probability of remaining seizure free follouing dircontinuation of antiepileptic drugs in children with seizures after a
seizure free period: overall recurrence risk. Kapbn-Meier curzr.
Overall Reczlrrence Risk
Of the 264 children in the study, 95 (36%')had experienced a recurrence as of February 1, 1993. The overall
Kaplan-Meier estimate of recurrence was 0.20 at 6
months (35% CI: 0.15, 0.25), 0.28 at 12 months (9595
CI: 0.23, 0.34), 0.32 at 24 months (75% CI: 0.26,
0.38), 0.35 at 48 months (9596 CI: 0.29, 0.40), and
0.37 at 72 months (9555 CI: 0.31, 0.43) (Fig 1). The
mean time to recurrence was 9.5 months, with a median of 4.3 months. The risk of recurrence was greatest
in the first few months after initiation of medication
withdrawal: 37 (39%) of the 95 recurrences occurred
in the first 3 months and 54 (57%))within 6 months
after entry into the study. Ninety percent of recurrences occurred within 25 months after study entry.
Fifteen children experienced a recurrence while medications were being withdrawn.
The mean follow-up period of the cohort was 58
months. Of the 264 patients, 260 (98%)were followed
for more than 2 years; 208 (79%)) for more than 4
years; and 168 (64$%)were followed for more than 5
years.
Risk Factors for Irelapse
Factors whose association with modification of recurrence risk was assessed are shown in Tables 2 and 3.
Etiology was a significant risk factor. In the group with
idiopathic epilepsy, the most important predictors of
recurrence were the age at onset, a history of atypical
febrile seizures, a family history of epilepsy, presence
of slowing on the EEG, and the 5pecific epileptic syndrome. In the remote symptomatic group, the important predictors were the age at onset, a history of atypical febrile seizures, a history of absence seizures, and
the severity of MR.
Table 2 Rzsk Factorc for Recurrente Following Withdrawal of An~zepileptzcDrug ThercEpy zn Children: UnriJarzateAnaIyJn
Proporrionate Hazards Model
Overall
Dichotomous Risk Facrors
Remote sympromaric etiology
Age at onser > 12 yra
Age at onset < 2 yr"
Age at medication withdrawal > 12 vr
Atypical febrile seizures
Absence seizurcsb
F;uniIy history of seizures
Slowing on EEG
Moderate-severe MK (1Q < 5 0 )
Motor deficit
Rare
Ratio
1.81
4.24
1.87
1.64
1.78
0.64
1.64
1.70
2.37
2.18
95%. CI
1.21,
2 34,
I 16,
1.07,
1.13,
0.40,
0.93,
1.08,
2.70
7.08
3.00
2.46
3.32
1.02
2.90
2.68
1.49, 3.83
1.03, 1.81
"Compared with age at onset 2-12.
hIncludes both simple and atypical absence seizures.
MR = mental retardation; EEG = electroencephalogram;Cl
Idiopathic
Rate
Ratio
p Value
0.004
-
< 0.0001
4.66
0 009
0.0 17
0.008
0.057
0.085
0 019
0.0002
0.027
0.88
2.11
1.94
0.93
2.13
2.31
-
=
Krnicw Sympromatic
Rate
Ratio
9374 C1
p Value
2.41, 8.39
0.44,2.93
1.18, 3.?9
0.83, 4.59
0.51, 1.67
1.06, 4.27
1.18, 4.55
-
-
-
< 0.0001
0 -9
0.012
0.12
08
0.03
0 012
3.08
1.55
1.16
-
1.64
0.40
1.20
1.05
1.91
1.12
95Y C1
1.35, ?.01
0.89, 3.13
0.65, 2.05
0.85, 3.16
0.17, 0.95
0.41, 3.37
0.56, 1.96
1.07, 3.38
0.80, 1.56
p Value
0.005
0.12
0.62
0.136
0.03
0. 72
0.8'
0.025
0.50
confidence interval
Table 3. Factors Not Associated witb a Daffererice in Recurrence Rijk Following Withdrawal of
Antiepileptic Drug Therapy in Children: UniiJariateAnaIysiJ
Proporrionate Hazards Model
Dichotomous Risk Factors
Epileptiforrn electroencephalogram
Duration of therapy (> 5 yr)
Years to seizure control (> 2 yr)
Status epilepticus
Multiple medications
Valproate therapy
Partial seizures
> 30 tonic clonic seizures
Female sex
Rate
Ratio
1.13
1.21
1.29
1.19
1.64
1.06
0.78
1.30
1.32
Remote Symptomatic
Idioparhic
Overall
Rate
Rate
95% CI
p Value Ratio
0.73,
0.80,
0.86,
0.55,
0.72,
0.67,
0.65,
0.63,
0.88,
0.53
0.37
0.21
0.66
0.23
0.81
0.92
0.48
0.18
1.76
1.82
1.04
2.56
3.16
1.67
1.47
2.67
1.77
95% CI
p Value Ratio
0.48.
0.62.
0.63,
0.91,
0.43,
0.76,
0.50,
0.26,
0.75
0.65
0.66
0.07
0.43
0.29
0.66
1.71
2.12
2.04
5.79
7.28
2.48
1.56
4.35
0.91, 2.91
0.90
1.15
1.14
2.29
1.77
1.3!
0.88
1.06
1.63
0.93
0.09
1.42
0.97
1.25
0.45
1.32
0.87
1.18
1.14
1.12
95rr CI
p Value
0.76.
0.55,
0.70,
0.11,
0.47,
0.41,
0.66,
0.48,
0.63,
0.27
0.92
0.45
0.25
0.60
0.72
0.58
2.65
1 72
2.21
1.84
3.67
1.87
2.09
2.68
1.98
0.7:
Q.71
CI = confidence interval
Risk factors described as being potential predictors
in the literature and not found to be associated with a
differential recurrence risk in either the overall group
or in the idiopathic or remote symptomatic subgroups
are shown in Table 3. Relative risks with CIS for the
overall group are shown to clarify the robustness of
the lack of association with recurrence risk. Individual
risk factors of particular importance are discussed
below.
mxxoC;Y OF SEIZURES. Individuals with remote symptomatic seizures had a higher recurrence risk than did
those with idiopathic seizures. The 2-year KaplanMeier risk of recurrence was 0.42 (95% CI: 0.32,
0.52) for those with remote symptomatic seizures, 47
(479f')of whom had recurrences, compared with 0.26
(959; CI: 0.19,0.33) for those with idiopathic seizures,
48 (29g1)of whom had recurrences ( p < 0.005)
(Fig 2).
.-
2m
0.6
-5
0.5
U
0.4
.-
E
"
41
I
0
Remote Symptomatic (n = 99)
I
I
12
24
36
TIME (months)
48
60
72
F i g 2. Probability of rmaining sezture free following discontinuation o[antiepileptic- drug.\ in children umith seizures after a
ieizurr-free period: e&t of etiology on outcome, Kaplan-Meier
cuwes mmparing idiokathii-cma uith remote .rynptomatic Lajej.
Shinnar et al: Discontlnulng Antiepileptic Medications
537
j
l0.9
o\
T
L
l0.9
’O
Age of onset 2-12 (n = 128)
*
I
Age of onsel c2
f onset 2-12
Y
Age of onset >12
7
o i-- ,
(n = 18)
a
n
0 027
0.1
0.1
0
12
,
T-
24
36
j
0~
I-
48
60
0
72
I
1
12
24
36
40
,
60
1
72
TIME (months)
TIME (months)
B
A
Fig 3, Probability of remaining seizurefree follmcing discontinuution of uniiepileptic drzigs in cbzldwn with seizures afier a
seizuye-freeperiod. liflect of age at onset on outcome: KaplanMeier cuwes. ( A ) ldiopatbic cases. (B) Remote syrnptotnatic
cues.
Age at onset was categorized into three groups:
those with onset before 2 years old, those with onset
between 2 and 12 years old, and those wich onset after
12 years old. In the overall group, seizures recurred in
735T) (22/30)of those with onset after 12 years old,
45% (27/60)of those with onset before 2 years old,
and 26% (461174) of those with onset between ages
2 and 12 years ( p < 0.0001). When stratified by etiology, the increased recurrence risk in the group with
onset after 12 years old was present in both idiopathic
and remote symptomatic patients (see Table 2, Fig 3).
However, the increased recurrence risk in the group
with onset before age 2 was limited to those with remote symptomatic seizures (see Table 2, Fig 3B). For
children with idiopathic seizures there was no significant difference in recurrence risks between the group
with onset before age 2 and the group with onset between ages 2 and 12 (see Table 2, Fig 3A).
In the idiopathic group, the Kaplan-Meier recurrence risk at 2 years was 0.72 (7596 CI: 0.52, 0.73)
for those with onset after age 12, compared with 0.26
(95%, CI: 0.07, 0.46) in the group with onset before
age 2 and 0.20 (75% CI: 0.13, 0.26) in the group with
onset between ages 2 and 12 ( p < 0.0001) (see Fig
3A). For the remote symptomatic group, the Kaplanhfeier recurrence risk at 2 years was 0.75 (9’39$CI:
0.51, 0.99) for those with onset after age 12 and 0.45
(7576 CI: 0.30, 0.60) for the group with onset before
age 2 but only 0.29 (95% CI: 0.16,0.43) for the group
with onset between ages 2 and 12 ( p < 0.001) (see
Fig 3B).
Age at last seizure and age at discontinuation of medications were also strongly associated with differentia1
AGE.
538
(n = 45)
(n = 19)
Annals of Neurology
Vol 35
No 5
May
1994
risks of seizure recurrence. However, they were less
highly correlated than was age at onset. O n multivariable analysis they did not add additional predictive
power to the information conveyed by age at onset and
did not enter the model.
~AiviILY HISTORY OF SEIZURES. A family history of unprovoked seizures in a first-degree relative was present
in 13% of children with idiopathic seizures and 9%
of those with remote symptomatic seizures and was
associated with an increased risk of recurrence in children with idiopathic epilepsy (see Table 2). The
Kaplan-Meier risk of recurrence at 2 years in children
with idiopathic seizures was 0.43 (9595 CI: 0.22, 0.64)
in those with a family history of epilepsy compared
with 0.24 (95%)Cl: 0.17, 0.31) in those wich no family
history of epilepsy ( p < 0.03) (Fig 4).
Negative Family History (n = 144)
$
0.7
N
.-
Positive Family History (n = 21)
C
n 0.2
0.1
4
0 1
0
I
I
I
1
12
24
36
TIME (months)
‘la
1
60
72
Fig 4. Probubzlity of remaining seizure freefbl(ou,ing discontinnation of unliepiieptic drugs in children with seizures after a
seizure-jireeperiod; dfect of furnib histoty of epilepsy in firstdegree rehztive on outcome in idiopathic cases, Kaplan-Meier
CUnJes.
he 4 . Ebi-troericephalographic(EEG) Abnomza1itie.c in Cohort of Children Being Withdrawn
m Antiepdeptzi Drug Therapy after a Seizure-Free Interual
Idiopathic
Abnormal EEG
Epileptiform abnormalities
Focal spikes (no slowing)
Benign rolandic
Other focal spikes
Slowing and focal spikes
Generalized spike and wave (no slowing)
Slowing and generalized spike and wave
Nonepileptiform abnormalities
Slowing (no spikes)
Focal slowing
Generalized slowing
Nonspecific abnormalities
Normal EEGs
No EEG available
Total
Normal EEG (n = 98)
a
No. with
Recurrences
N
EEG
60
20
47
23
11
12
3
13
3
0
3
20
1
7
4
3
13
13
4
7
7
2
5
0
21
1
9
0
78
7
165
48
,
4
0 1
0
12
I
I
24
36
TIME (months)
I
48
KO.with
5%
N
Recurrences
%
33
28
13
0
25
33
35
75
54
54
50
50
29
11
23
46
55
64
-
-
-
11
15
3
1
7
64
66
56
-
28
13
29
21
18
6
12
3
42
7
99
16
7
9
1
0
7
1
2
5
33
0
33
37
33
42
0
0
19
5
47
45
71
47
was worse than previous EEGs was associated with a
higher recurrence risk (rate ratio: 1.82; 95% CI: 0.57,
5.87) but the differences were not statistically significant.
In children with remote symptomatic epilepsy, EEG
0.2
0.1
Remote Symptomatic
60
72
F i g 5 . Probability of remaining seizure free following discontinuation of antiepileptic drugs in children with seizures &er a
seiturrjiree period: effect of slowing on the eltxtroencephaloRram
( E E G ) on outcome far idiopathic cajes, &plan-Meier cumws.
ELECTROENCEPHALOGRAM. The distribution of EEG
abnormalities is shown in Table 4. One hundred nine
(43%) of the EEGs were abnormal. In the group with
idiopathic epilepsy, slowing on the EEG, whether generalized or focal, was associated with a higher risk of
recurrence (Fig 5). The !&plan-Meier recurrence risk
at 2 years was 0.45 (955% CI: 0.23, 0.67) in patients
with idiopathic seizures with slowing on the EEG, compared with 0.25 (95% CI: 0.16, 0.33) for children with
idiopathic epilepsy and normal EEGs. Epileptiform
EEG abnormalities were not associated with a significantly higher risk of recurrence in the idiopathic group.
An EEG prior to discontinuation of medications that
SEIZURE TYPE.
The distribution of seizure types in this
population is shown in Table 5. Atypical febrile seizures were associated with an increased risk of recurrence (Fig 6; see Tables 2 , 5 ) in both the idiopathic
and remote symptomatic groups. A history of absence
seizures was associated with a lower recurrence risk in
the remote symptomatic group only. As these occurred
in neurologically abnormal children, they would be
considered atypical absence and not the classic absence
seizures seen in the primary generalized epilepsies C23,
28, 293. A history of complex partial seizures or of
tonic clonic seizures was not associated with a differential recurrence risk.
A total of 16 (65%)children had a history of status
epilepticus including 8 (4.8%) with idiopathic epilepsy
and 8 (8%') with remote symptomatic epilepsy. In the
idiopathic group, 5 (63';1) of 8 children with a history
of status had recurrences, compared with 43 (27%) of
157 children without such a history ( p = 0.03). In the
remote symptomatic group, 2 children (25%) with a
history of status had recurrences, compared with 45
(49%) of 91 without such a history Cp = 0.18). The
risk ratios from the proportional hazard model are
shown in Table 2.
Shinnar et al: Discontinuing Antiepileptic Medications 539
Table 5 . Distribution of Seizure Types in Cohort of Children Keing Withdrawn from
Antiepileptic Drug Therapy after a Seizure-Free Interval
Idiopathic
Remote Symptomatic
No. with
No. with
Seizure Type"
N
Recurrences
9,
N
Recurrences
%
Total
165
48
29
13
29
99
62
47
29
47
47
Generalized tonic clonic
Simple absence
86
Atypical absence
43
22
Partialb
77
21
Atypical febrile'
12
6
9
4
10
2
34
29
32
27
30
44
20
8
5
2
20
63
33
33
47
6
38
28
M yoclonic
Akinetic
Other Seizure Parameters
Scatus epilepticus
2
7
6
26
16
30 tonic c h i c seizures
Multiple seizure types
Previously intractable
-
-
-
24
41
19
20
13
6
25
51
63
45
8
12
21
12
9
7
54
2
25
6
20
12
43
50
43
aPatients may have more than one seizure type.
hParrial includes both with or without secondary generalization.
<Mustalso have had at least two unprovoked afebrile seizures.
were no other significant differences attributable to
specific epileptic syndromes in the idiopathic group.
Of the 99 children
in the remote symptomatic group, 80% had a neurological deficit, including 4296 with MR, 17%) with motor deficits, and 21%) with both MR and a motor deficit. The severity of the MR was correlated with the
risk of seizure recurrence. In children with remote
symptomatic seizures, the Kaplan-Meier recurrence
risk at 2 years was 0.51 (95% CI: 0.35, 0.67) in the
children with moderate-severe MR (IQ < 50), compared with 0.38 (95% CI: 0.26, 0.50) in those with
normal intelligence or mild MR ( p < 0.03). The presence of' a motor deficit and its severity was not associated with an increased recurrence risk in the remote
symptomatic group (see Table 2). In the overall group,
the presence of a motor deficit had a statistically significant association with outcome as it is a marker for
remote symptomatic seizures, which are associated
with an increased recurrence risk (see Table 2).
INTELLIGENCE AND MOTOR DEFICIT.
0.1
j
"
l
" I
I
0
12
24
,
36
48
---
60
72
TIME (months)
Fig 6. Probability of remaining seizure free following discontinuation of antiepileptic drugs in children with seizures after a
seizure-free period: &ict of a histoqi of atypical febrile seizuves
on outcome in overull cohort, Kaplan-Meier curves.
EPILEIYI'IC SYNDROME. The identifiable epileptic syndromes in the children with idiopathic seizures are
shown in Table 6. In many cases, no definite epileptic
syndrome could be assigned as this was a population
in which many children had only a few seizures and a
normal EEG. Benign rolandic epilepsy and juvenile
myoclonic epilepsy were associated, respectively, with
particularly favorable and unfavorable prognoses. The
favorable prognosis in children with benign rolandic
epilepsy occurred even though the mean age at study
entry was 12.0 years after a mean seizure-free interval
of 2.8 years, and 10 (77%) of 13 still had epileptiform
EEGs at the time of medication withdrawal. There
540 Annals of Neurology Vol 35 No 5 May 1994
Mzlltivuriuble Analysis
A multivariable analysis was performed using a proportionate hazards model (Table 7). Separate models were
used for the idiopathic and remote symptomatic
groups. Age at onset was the single most important
variable in all groups, with those with adolescent onset
having a worse prognosis. A history of atypical febrile
seizures was also associated with a higher risk of recurrence in all groups. Note that all these children also
had at least two unprovoked seizures to be eligible to
enter this study. In the idiopathic group, slowing on
the EEG prior to discontinuation of therapy and a fam-
Table 6. Epileptic Syndromes in Cohort of Children Being Withdrawn from Antiepileptic Drug
Therapy after a Seizure-FreeInternal: Idiopathic Cases
No. with
Recurrences
N
Epileptic syndrome
Primary generalized epilepsy
Childhood absence
Juvenile absence
Juvenile myoclonic epilepsy
Other primary generalized
Partial epilepsy
Benign rolandic epilepsy
Benign occipital epilepsy
Temporal lobe epilepsy
Other partial epilepsy
Unclassifiable partial epilepsy
Unclassifiable idiopathic epilepsy
Total idiopathic cases
(3-
21
5
3
61
26
9
4
17
0
1
3
8
34
19
33
100
41
25
0
25
43
31
5
29
10
48
28
29
4
9
22
68
14
4
7
26
17
36
165
p Value
0.006
< 0.001
Table 7. Risk Factors for Recurrence Following Withdrawal of Antiepileptic Drug Therapy in Children:
Mdtivariable Andbsis Using Cox Proportional Hazards Model
Proportionate Hazards Model
Risk Factor
Overall group
Age at onset > 12 yr
Atypical febrile seizures
Family history of seizures
Moderate-severe MR (IQ < 50)
Slowing on electroencephalogram
Idiopathic group
Age at onset > 12 yr
Family history of seizures
Slowing on electroencephalogram
Atypical febrile seizures
Remote symptomatic group
Age at onset > 12 yr
Moderate-severe MR (IQ < 50)
Absence seizures
Atypical febrile seizures
MR
=
m e n d retardation; CI
=
Relative Risk
95% CI
p Value
5.1
2.5
2.4
2.1
1.6
3.0,
1.4,
1.3,
1.2,
1.0,
8.7
4.4
4.4
3.6
2.6
< 0.001
< 0.003
5.4
3.1
2.4
2.8
2.8,
1.5,
1.2,
1.1,
10.7
< 0.001
6.3
0.003
0.013
0.037
3.6
2.8
0.4
2.0
1.6,
1.5,
0.2,
1.0,
8.3
5.2
0.9
3.9
0.003
0.012
0.049
4.7
7.2
< 0.003
< 0.002
0.03
0.049
confidence interval.
ily history of seizures remained important but a history
of status epilepticus, which was of borderline significance in the univariate analysis, did not remain so in
the multivariable model. The epileptic syndromes such
as juvenile myoclonic epilepsy and benign rolandic epilepsy could not be entered into the model due to
100% and 0% recurrence risks, respectively, which
made calculation of the relative risk impossible.
In children with remote symptomatic seizures, additional prognostic factors included moderate-severe MR
and a history of absence seizures. Note that in the
overall group, once the effect of the severity of MR is
entered into the model, etiology, which was significant
in the univariate analysis, does not enter the model.
O n the other hand, presence of a motor deficit, while a
marker for remote symptomatic etiology and therefore
significant in the overall group, was not associated with
a differential risk of recurrence in the remote symptomatic group.
In both the idiopathic and remote symptomatic
groups, children with no risk factors had a very low
risk of recurrence at 2 years, whereas children with
multiple risk factors had a high risk of recurrence (Table 8). Among children with idiopathic seizures, 61%
fell into the best prognostic group, which had only a
12% recurrence risk. In the remote symptomatic
group, only 9% fell into the best prognostic category
and 46% had two or more risk factors.
Shinnar et al: Discontinuing Antiepileptic Medications
541
Table 8. Risk of Seizzlre Recuwence 2 Years aftw W'ithdraufal of Antzeprleptrc Drug Therupy rn
Children as u Function of Number of Risk Fuctors PreJent
No. with
No. of Risk Factors"
57
N
Recurrences
97
54
7
14
14
27
50
2-Year
Recurrence Riskb
Idiopathic seizures
0
1
2
2 2
0
6
86
-
-
1
11
Remote symptomatic seizures
0
1
2
3
>3
9
44
37
9
0
15
23
8
-
34
62
89
-
"Risk facrors are those from multivariable analysis (Table 7) and are differenr for idiopathic and remote symptomatic groups
'Calculated from Kaplan-Meier analysis
Discussion
Our results are consistent with those of previous studies showing that approximately 70% of children with
seizures who are seizure free for 2 or more years while
on AED therapy will remain seizure free when medications are withdrawn 11-20}.
Not surprisingly, children with remote symptomatic
epilepsy had a higher recurrence risk. However, even
in this group of neurologically abnormal children, more
than half remained seizure free after medications were
withdrawn. This suggests that even though children
with remote symptomatic epilepsy are far less likely to
achieve remission than those with idiopathic epilepsy
122,38-41}, those who are seizure free can often be
successfully withdrawn from medication. This is especially true for those without severe MR or other additional risk factors (see Table 8). The magnitude of the
effect of etiology on recurrence risk is consistent with
that found in a recent meta-analysis of the literature
c201.
Age at onset was the most important predictor of
recurrence in all groups. Those with onset after 12
years old had a high risk of recurrence. This was also
true in our previous study [l}, though the numbers
there were too small to permit meaningful analysis.
Most other studies also showed an elevated risk of
relapse in patients with adolescent-onset epilepsy compared with childhood onset [l, 2, 4-7, 14, 19, 201,
with an average relative risk of 1.78 [20]. The present
results also explain some of the discrepancies in the
literature regarding the importance of a very young age
at onset. This turns out to be a risk factor for recurrence, but only in those with remote symptomatic epilepsy. Thus, studies [3} that included large numbers of
neurologically abnormal children found it to be a risk
factor, whereas studies that included primarily children
with idiopathic epilepsy did not find it to be significant
[I, 191.
542 Annals of Neurology Vol .35 Nu 5 May 1994
As was the case in our previous study El), atypical
febrile seizures were a risk factor for recurrence. This
study included a sufficient number of neurologically
abnormal children to demonstrate that they are risk
factors in both children with idiopathic and those with
remote symptomatic epilepsy. Note that atypical febrile seizures do not correspond to complex febrile
seizures as defined by the National Institutes of Health
(NIH) consensus conference [24, 251. Rather, these
are children with afebrile seizures who also have seizures triggered by fever. The reason why the effect is
stronger in the multivariate analysis than the univariate
one is that children with atypical febrile seizures have
onset at a younger age. Therefore the protective effect
of younger age at onset is a negative confounder in the
crude association between atypical febrile seizures and
seizure recurrence.
EEG abnormalities were again found to be associated with a higher risk of recurrence but only in children with idiopathic seizures. In the current study, the
presence of slowing was associated with an increased
risk of recurrence while epiieptiform abnormalities
were not. Slowing included both generalized background slowing and focal slowing. Similar results for
the association of slowing in the EEG with an increased
risk of relapse were found in our previous study [l}.
Data regarding epileptiform abnormalities, while negative, must be interpreted with caution. Most children
with focal spikes had benign rolandic epilepsy, a syndrome with a particularly favorable prognosis [28, 291.
If one excludes children with benign focal epilepsies
of childhood, then only 11 children had focal spikes.
Our previous smaller study did find an increased risk
of recurrence in children with epileptiform EEG abnormalities [l}. In comparing the results of other studies, very few provide detailed information regarding
the nature of the precise EEG abnormalities. However,
the majority of the studies did show an increased risk
of recurrence in those with an abnormal EEG prior to
withdrawal of the drug therapy f l , 2,6, 8, 14, 15, 201,
though there have been several negative reports [4-6,
11, 12). A recent rrieta-analysis of the available literature concluded that the association of an abnormal
EEG with an increased relapse risk in both children and
adults is consistent but fairly modest, with an average
relative risk of 1.47 120f. However, the relative risks
in individual studies ranged from 0.82 to 3.73 E20).
The lack of detailed EEG data in published reports, as
well as the relative paucity of specific EEG abnormalities in a group of patients who are seizure free for
several years, precludes more detailed analysis of the
relative importance of specific EEG abnormalities.
The present study demonstrated that even in a large
and fairly diverse population, a few readily obtainable
variables can provide valuable prognostic information
about recurrence risks. These few risk factors can identify children with very low and very high recurrence
risks. The child with idiopathic epilepsy and fio other
risk factors has a 2-year recurrence risk of only 1295.
This is a substantial group, accounting for over 60%
of children with idiopathic epilepsy. While the children
with remote symptomatic epilepsy do have a higher
recurrence risk, they too can be divided into subgroups
with very different prognoses, though fewer will be in
the favorable prognosis subgroups.
Duration of the seizure-free period was not associated with the risk of recurrence in this group. It should
be noted that this was not a randomized study and that
children with perceived poor prognostic factors were
more likely to be treated for longer periods of time by
their physician. Since patients remaining on medications would not enter the study, it is difficult to know
precisely how many patients who were seizure free
at 2 years were not withdrawn from medications. A
qualitative review suggests that over 90% of children
who were seizure free while on AEDs were subsequently withdrawn from medications. However, the
etiology, the EEG, and the physician’s perception of
potential risk factors and the risk-benefit ratio of continued therapy may well have affected the duration of
therapy prior to discontinuation. This should not have
created a substantial bias in the other variables analyzed
such as etiology or the EEG, but may have obscured
any effect of the duration of the seizure-free interval
prior to medication withdrawal on outcome. Therefore, no clear conclusions regarding the effect of duration of the seizure-free interval o n recurrence risk can
be drawn from this study. The literature suggests that
the effect of a seizure-free period of more than 2 years
is associated with only a marginal change in the recurrence risks {19].
The actual decision as to how long to keep a patient
on medications must be individualized and take into
account the potential risks and benefits of continued
therapy. Factors to be considered include not only the
statistical probability of seizure recurrence but the consequences of a recurrence. The risk of injury from a
brief seizure is small. The main impact of a brief seizure
in children and adolescents is psychosocial [17,42,43].
The risk of status epilepticus following gradual withdrawal of AED therapy is very low f 7 , 261. Even status
epilepticus in this setting has a very low morbidity and
mortality 126, 27, 441. An additional consideration
prior to medication withdrawal is the possibility that
previously well-controlled seizures will become refractory after a recurrence. This is a rare event.
Seizure control is promptly reestablished in almost all
patients who relapse following reinstitution of drug
therapy, including those in the present study 11, 2,
191.
On the other hand, the decision to maintain the patient who is seizure free on medication must also weigh
the risks. A small but real portion of patients who are
seizure free and remain on medication will still experience a relapse 16, 451. Other factors to be considered
are the toxicities of chronic AED therapy including not
only systemic toxicities, which are well known 1461,
but also the cognitive and behavioral side effects, which
are increasingly recognized as the major toxicity of
these medications [41, 42, 47-50]. In adolescent females, the effects of these medications on oral contraceptives and the potential for teratogenesis must also
be considered. One must also recognize that it is far
easier and safer to attempt withdrawal of medications
in adolescents before rather than after they start driving 121, 5 1-53}. We are thus particularly aggressive in
withdrawing adolescents with epilepsy from medications, even if they have risk factors for recurrence [43,
5 11.
In terms of a strategy for when to withdraw medications, the child with idiopathic seizures, a normal EEG,
and onset before age 12 who is seizure free for 2 or
more years while on AEDs is an excellent candidate
for medication withdrawal. Children with idiopathic or
remote symptomatic seizures and only one additional
risk factor (see Table 8) are also very appropriate candidates for medication withdrawal. In children with multiple risk factors, where the risk of recurrence is high,
one must individualize the decision based on the physician’s and family’s analysis of the risk-benefit ratio using
the considerations enumerated above (see also 119-2 1 ,
4 2 , 4 3 , 51-53]). A more detailed analysis of the risks
and benefits of medication withdrawal versus continued AED treatment is beyond the scope of this article,
whose emphasis is on defining the statistical risks of
seizure recurrence to be used as a basis for clinical
decision making. It is clear that the risk-benefit ratio
for the same statistical risk of recurrence is different
in children than adults and more in the direction of
withdrawing AEDs.
Shinnar et al: Discontinuing Antiepileptic Medications
543
Conclusions
The majority of children with seizures who are seizure
free for 2 or more years while on AED therapy will
remain seizure free when medications are withdrawn.
Etiology of seizures, age at onset, family history, seizure type(s), the presence of slowing in the EEG, and
the specific epileptic syndrome are significant prcdictors of outcome. Children who are neurologically
abnormal have a higher risk of recurrence but can still
often be withdrawn successfully from medications, especially if they do not have additional risk factors. The
final decision regarding medication withdrawal in a
child who is seizure free is an individualized one that
should be made jointly by the physician and the family.
The decision process must take into account the statistical risks of seizure recurrence, the potential consequences of such a recurrence, and the potential adverse
effects of continued AED therapy.
This work was supported in part by a Teacher Investigator Development Award 1 KO7 NS00930 and grant 1 R 0 1 NS26151 (to D r
Shinnar) and a FIRST award 1 R29 NS27728 and grant 1 RO1
NS31146 (Dr Berg) from the National Institute of Neurological
Disorders and Stroke.
We 'are indebted to all the housestaff and attending physicians who
allowed us to recruit their patients into the study. Particular thanks
are due to Drs Alfred Spiro, Isabelle Rapin, and Mitchell Steinschneider who as directors of the pediatric neurology clinic at Bronx
Municipal Hospital Center were responsible for the care of many of
these children. We also acknowledge the cooperation of the New
York Ciry Health and Hospitals Corporation and Bronx Municipal
Hospital Center and North Central Bronx Hospital. Thanks are also
due to our previous research nurses Ann Eckstein, RN, Nancy
Fremed, RN, Laurie Zeitlin-Gross, RN, and Maryann Perk, RN,
MSN, who assisted in the data collection and follow-up and to Ann
hlancini for editorial assistance.
Presented in part at the Child Neurology Society meeting, Atlanta,
GA, October 18-20, 1990, and American Epilepsy Society meetings, San Diego, CA, November 11-14, 1990; Philadelphia, PA,
December 8-11, 1991; and Seattle, WA, December 6-9, 1992.
References
1. Shinnar S, Vining EPG, Mellits ED, et al. Discontinuing antiepileptic medication in children with epilepsy after two years without seizures: a prospective study. N Engl J Med 1985;3 13:976-
980
2. Todt H. The late prognosis of epilepsy in childhood: resulrs of
a prospective followup study. Epilepsia 1984;25:137- 144
3. Emerson R, DSouza BJ. Vining EP, et al. Stopping medication
in children with epilepsy: predictors of outcome. N Engl J Med
1%1;j04:1125-1129
4. Holowach .I,Thurston DL, O'Leary J. Prognosis in childhood
epilepsy: followup study of 148 cases in which therapy had been
suspended after prolonged anticonvulsant control. N Engl J Med
1972;286:169-174
5 . Holowach-Thurston J H , Thurston DL, Hixon BB, et al. Prognosis in childhood epilepsy: additional followup of 148 children
15 to 2 1 years after withdrawal of anticonvulsant therapy. N
Engl J Med 1982;306:831-836
6. Medical Research Council Anriepileptic Drug Withdrawal Study
544 Annals of Neurology Vol 35 No 5
May 1994
Group. Randomised study of antiepileptic drug withdrawal in
patients with remission. Lancet 1991;337:1175-1180
7. Arts WFM, Visser LH, Loonen MCB, er al. Follow-up of 146
children with epilepsy after withdrawal of antiepileptic therapy.
Epilepsia 1988;29:244-250
8. Bouma PAD, Peters ACB, Arts RJHM, e t al. Discontinuation
of antiepileptic therapy: a prospective trial in children. J Neurol
Neurosurg Psychiatry 1987;50:1579-1583
9. Broeker H, Walter S. Zur frage der rezidive bei ausgeheilten
epileprien in erwachsenenalton. Psychiatr Neurol Med Psycho1
1981;33:536-542
10. Forster C H , Schmidberger G. Prognostische Wertigkeit von
EEG verlautsuntersuchungen bei Kmdern mit anfallsrezidiven
nach absetzen der antikonvulsiven therapie. In: Doose H, Gross
Selbeck G, eds. Epilepsie. Sruttgart: Thieme, 1978: 252-255
11. Juul-Jensen P. Frequency of recurrence after discontinuance of
anticonvulsanr therapy in patients with epileptic seizures. Epilepsia 1964;5:352-363
12. Juul Jensen P. Frequency of recurrence after discontinuance of
anticonvulsant therapy in parienrs with epileptic seizures: a new
followup study after 5 years. Epilepsia 1968;9:11-16
13. Keranen T. Discontinuation of antiepileptic drugs. In: Sillanpaa
M, Johannessen SI, Blennow G, Dam M, eds. Pediatric epilepsy.
Petersfield, Great Britain: Wrightson Biomedical, 1990:3?13340
14. Matricardi M, Brinciott M, Benederti P. Outcome after disconcinuation of antiepileptic drug therapy in children with epilepsy.
Epilepsia 1989;30:582-589
15. Nalin A, Galli V, Ferrari P, Ciccaron V. Sospensione della therapia antiepilettica in soggetti con epilessia insorta nell'eta infantile. Pediarr Med Chir 1982;4:601-606
16. Pestre M, Loiseau P, Larrieu E, e t al. Withdrawal of antiepileptic
drug therapy in adolescent epileptic patients. 1987;16:395-400
17. Tsuchiya S, Maruyama H, Maruyama K, et al. A follow up study
of 1007 epileptic children with anticonvulsant therapy for more
than 10 years. N o To Hattatsu 1985;17:23-28
18. Visser LH, Arts WFM, Loonen MCB, et al. Follow-up study of
166 children with epilepsy after withdrawal of anticonvulsant
therapy. Adv Epileprol 1987;16:401-404
19. Gross-Tsur V. Shinnar S. Discontinuing antiepileptic drug therapy in patients with epilepsy. In: Wyllie E, ed. The treatment
of epilepsy: principles and practice. Philadelphia: L e a d Febiger,
1993:858-866
20. Berg AT, Shinnar S. Relapse following discontinuation of antiepileptic drugs: a meta-analysis. Neurology 1994 (in press)
21. Pedley TA. Discontinuing antiepileptic drugs. N Engl J Med
1988;318:982-984
22. Hauser WA, Kurland LT. The epidemiology of epilepsy in
Rochester, Minnesota, 1935 through I96?. Epilepsia 1975;
16:1-66
2 1 Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised clinical and
electroencephalographic classification of epileptic seizures. Epilepsia 1981;22:489-501
24. Nelson KB, Ellenberg JH. Prognosis of children with febrile
seizures. Pecliatrics 1978;61:720-727
2 5 . National Institutes of Health. Febrile seizures: consensus development conference summary, vol 2, no 2. Bethesda, MD: National Instirutes of Health, 1980
26. Gross Tsur V, Shinnar S. Convulsive status epilepticus in children. Epilepsia 1993:34(suppl l):>12-520
27. Maytal J, Shinnar S, Mostie SL, Alvarez LA. The low morhidity
and mortality of status epilepticus in children. Pediatrics 1989;
8j:j23-331
28. Commission on classification and terminology of the International League Against Epilepsy. Proposal for revised classihca-
29.
30.
3 1.
32.
33.
34.
35.
36.
37.
38.
39.
40.
tion of epilepsies and epileptic syndromes. Epilepsia 1989;30:
389-399
Roger J, Drdvet C, Bureau M, et al, eds. Epileptic syndromes
in infancy, childhood and adolescence. London: John-Libbey Eurotext, 1985
Commission on epidemiology and prognosis, Ititerriational
League Against Epilepsy. Guidelines for epidemiologic studies
on epilepsy. Epilepsia 1993;34:592-536
Kaplan EL, Meier P. Nonparametric estimation from incomplete
observations. J Am Stat Assoc 1958;53:457-481
Mantel N. Evaluation of survival data and two new rank order
statistics arising in its consideration. Cancer Chemother Rep
1966;5O:163-1 7 0
Tarone RE, Ware J. On distribution-free tests for equality of
survival distributions. Biometrika 1977;64;156-60
Elandt-Johnson RC, Johnson NL. Survival models and data analysis. New York: Wiley, 1980:150-180
Cox DK. Regression models and life-tables. J R Stat Soc ( 8 )
1972;34:18?-220
Kalbfleisch JD, Prentice RL. The statistical analysis of failure
rime data. New York: Wiley, 1980:70-142
Dixon WJ, ed. BMDP statistical software 1981. Berkeley, CA:
University of California Press, 1981:555-5?5
Hauser WA, Hesdorffer DC. Epilepsy: frequency, causes and
consequences. New York: Epilepsy Foundation of AmericaDemos, 1990
Hutrenlocher PR, Hapke RJ. A follow up srudy of intractable
seizures in childhood. Ann Neurol 1990;28:699-705
Sillanpaa M. Social functioning and seizure status of young
adults with onset of epilepsy in childhood. An epidemiological
20 year follow up study. Acta ileurol Scand suppi 1983;96:
7-77
41. Goulden KJ, Shinnar S, Koller H. er al. Epilepsy in children
with mental retardation: a cohort study. Epilepsia 1991;32:690-
697
42. Freeman JM, Tibbles J, Camfield C, et al. Benign epilepsy of
childhood: a speculation and its ramifications. Pediatrics 1987;
73:864-868
43. Shiiiixu S. Tredtment decisions in childhood seizures. In: Dodson WE, Pclluck JM, eds. Pediatric epilepsy: diagnosis and therak>y.New, York: Demos, 1993:215-22 1
44. Dodson WE, DeLoretizo RJ, Pedley TA, et al. The treatment of
convulsive status epilepticus: recommcndations of the Epilepsy
Foundation of America’s working group o n status epilepticus.
JAMA 1993;270:851-859
45. Annegers JF, Hauser WA, Elveback LR. Remission of seizures
and relapse in patients with epilepsy. Epilepsia 1979;20:729737
46. Reynolds EH. Chronic antiepileptic toxicity: a review. Epilepsia
1975;16:?19-352
47. American Academy of Pediatrics, Committee on Drugs. Behavioral and cognitive effects of anticonvulsant therapy. Pediatrics
1985;76:644-647
48. Masur DM, Shinnar S. The neuropsychology of childhood seizure disorders. In: Segalowitz S, Rapin I, eds. Handbook of
neuropsychology, vol ?. Childhood neuropsychology. Amsterdam: Elsevier, 1992457-470
47. Trimble MR. Anticonvulsant drugs and cognitive function: a
review of the literature. Epilepsia 1987;28(suppl 3):S37-S45
50. Vining EPG, Mellits ED, Dorsen MM, et al. Psychologic and
behavioral effects of antiepileptic drugs in children: a doubleblind comparison between phenobarbital and valproic acid. Pediatrics 1987;80:165-174
5 1. Shinnar S. Seizures in adolescents. In: Friedman SB, Fisher M,
Schonberg SK, eds. Comprehensive adolescent health care. St.
Louis: Quality Medical, 1992:458-467
52. Chadwick D. Drug withcirawal and epilepsy when and how!
Drugs 1988;35:579-583
53. Scheuer ML, Pedley TA. The evolution and treatment of seizures. N Engl J hled 1990;3251468-1474
Shinnar et al: Discontinuing Antiepileptic Medications
545
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 082 Кб
Теги
drug, prospective, stud, children, epilepsy, discontinuity, antiepileptic
1/--страниц
Пожаловаться на содержимое документа