- Email: [email protected]
Stimulant therapy and seizure risk in children with ADHD
Original Articles
Stimulant Therapy and Seizure Risk in Children With ADHD Sarah A. Hemmer*, Joseph F. Pasternak, MD*†, Steven G. Zecker, PhD‡, and Barbara L. Trommer, MD*† Stimulants are an effective treatment frequently prescribed for attention-deficit– hyperactivity disorder (ADHD), but they commonly are believed to lower the threshold for seizures. Although several studies have revealed that stimulants do not exacerbate well-controlled epilepsy, there is a paucity of data about seizure risk in nonepileptic children treated with stimulants. Two hundred thirty-four children (179 males, 9.1 ⴞ 3.6 years of age; 55 females, 9.6 ⴞ 3.9 years of age) with uncomplicated ADHD received electroencephalograms (EEGs) performed in our institution. Thirty-six patients (15.4%) demonstrated epileptiform abnormalities, and 198 (84.6%) demonstrated normal or nonepileptiform EEGs. Rolandic spikes accounted for 40% of the abnormal EEGs and 60% of those with focal abnormalities. Stimulant therapy was elected by 205 of 234 patients (87.6%). Seizures occurred only in the treated group, in one of 175 patients with a normal EEG (incidence 0.6%, 95% confidence intervals 0%1.7%) and three of 30 treated patients with epileptiform EEGs (incidence 10%, 95% confidence interval 0%-20.7%). Seizures occurred in two of 12 children (16.7%) with rolandic spikes. These data suggest that a normal EEG can be used to assign children with ADHD to a category of minimal risk for seizure. In contrast, an epileptiform EEG in neurologically normal children with ADHD predicts considerable risk for the eventual occurrence of seizure. The risk, however, is not necessarily attributable to stimulant use. © 2001 by Elsevier Science Inc. All rights reserved. Hemmer SA, Pasternak JF, Zecker SG, Trommer BL. Stimulant therapy and seizure risk in children with ADHD. Pediatr Neurol 2001;24:99-102.
*Department of Pediatrics; Division of Neurology; Evanston Hospital; Evanston, Illinois; †Northwestern University Medical School; Chicago, Illinois; and ‡Northwestern University Program in Learning Disabilities; Evanston, Illinois.
© 2001 by Elsevier Science Inc. All rights reserved. PII S0887-8994(00)00240-X ● 0887-8994/01/$—see front matter
Introduction Attention-deficit– hyperactivity disorder (ADHD) is the most common neurobehavioral disorder of childhood with an estimated prevalence of 8-10% in school-aged children [1]. Stimulant medications are effective, frequently prescribed treatments for ADHD [2]. Their use is apt to increase as a result of the recent multimodal treatment study [1] demonstrating long-term efficacy and superiority over other treatment modalities. Lowering of the seizure threshold is commonly considered to be a risk of stimulant therapy. Whereas the safety of stimulants in patients with treated, well-controlled epilepsy and concurrent ADHD has been documented [3-5], there is little information to guide the clinician regarding seizure risk in uncomplicated ADHD. The present study examined the relationships among electroencephalographic (EEG) findings, stimulant use, and seizure occurrence in children with ADHD without known epilepsy. Methods and Materials The medical records were reviewed of all patients referred to a hospital-based child neurology practice for evaluation of suspected ADHD or attention-deficit disorder (ADD) between 1993 and 1998 (n ⫽ 765) under a local Institutional Review Board-approved protocol. EEGs routinely were recommended for patients when the diagnosis of ADHD was made and the use of stimulant medication was planned. To be included in the study, patients were required to meet diagnostic criteria for ADHD according to DSM-III (ADD with or without hyperactivity) [6], DSM-III-R [7], or DSM-IV (primarily inattentive, primarily hyperactive/impulsive, or combined subtypes) [8]. Diagnosis was made by a child neurologist on the basis of parental endorsement of symptom descriptors supplemented by abbreviated Conners Parent Symptom Questionnaires [9], ActERs Teacher Profile questionnaires [10], and additional history provided by patients, parents, and teachers. Children with neurologic deficits or neurologic diagnoses other than learning disability, headache, and tic disorders were excluded, as were children taking antiepilepsy medications for any reason. In addition to ADHD
Communications should be addressed to: Dr. Trommer; Evanston Hospital; Division of Neurology; 2650 Ridge Avenue; Evanston, IL 60201. Received June 6, 2000; accepted September 21, 2000.
Hemmer et al: Seizures, Stimulants, and ADHD 99
Table 1.
Distribution of EEG findings in 234 patients with ADHD
EEG
Stimulant Treated (n ⴝ 205)
Untreated (n ⴝ 29)
Total (n ⴝ 234)
175 (1)
23
198
6
36
Normal
Epileptiform
30
Focal
Rolandic
Nonfocal
Nonrolandic
20
12 (2)
10 (1)
8
2
1
1
4
22
14
13
9
EEGs were obtained in 234 patients for whom stimulant therapy was recommended. The number of subjects with each EEG finding is presented for those who elected stimulant treatment (left column), those who declined (middle column), and for the total group (right column). Seizures were only seen in the stimulant-treated group; the number of patients who had at least one seizure during the study period is indicated in parentheses in the corresponding EEG category.
diagnosis, patients were required to have received an EEG in our institution and follow-up either by office visit or by telephone in 1999. EEGs were performed after the initial visit during which diagnosis was made, before or up to 8 weeks after initiation of stimulant therapy, and were interpreted by a board-certified clinical neurophysiologist (J.F.P.). EEGs were classified as normal/nonepileptiform or epileptiform. The epileptiform group was subdivided according to whether the abnormalities were nonfocal or focal, and within the focal abnormal group those with centrotemporal or centroparietal “rolandic” spikes were tabulated separately. Methylphenidate (0.3-1 mg/kg per day) was the initial stimulant prescribed. In some patients, 0.1-0.5 mg/kg per day of dextroamphetamine sulfate or mixed amphetamine salts (Adderall, ShireRichwood) was substituted, with dose adjustments and medication choice guided by clinical response. Statistical analyses were performed using Systat (SPSS, Chicago, IL) and Microsoft Excel software (Redmond, WA).
Results Two hundred thirty-four patients met our inclusion criteria. They consisted of 179 males who were 9.1 ⫾ 3.6 years of age (mean ⫾ S.D., range ⫽ 3-20 years) and 55 females who were 9.6 ⫾ 3.9 years of age (range ⫽ 3-19 years). Normal or nonepileptiform EEGs were found in 198 patients (84.6%) and epileptiform EEGs in 36 (15.4%). The distribution of EEG abnormalities is summarized in Table 1. Stimulants were offered to all 234 patients, but the decision to use them was ultimately made by parents. Stimulant treatment was elected by 205 and declined by 29 patients. The incidence of epileptiform EEGs did not differ between the treated (30/205) and untreated (6/29) groups (14.6% vs 20.7%, respectively, P ⫽ 0.7, binomial test) suggesting that the presence of an abnormal EEG did not bias the decision to treat. Rolandic spikes occurred in 5.6% of the study population and accounted for 40% of the abnormal EEGs and 60% of the EEGs with focal abnormalities. Spikes were right-sided in five patients, left-sided in four, and bilateral
100
PEDIATRIC NEUROLOGY
Vol. 24 No. 2
or midline in four with no gender-related pattern in the distribution. Seizures occurred in four patients, all of whom were in the stimulant-treated group. Only one seizure patient had an initial normal EEG. She was a 6-year, 7-month-old female whose first seizure occurred 6 weeks after initiation of methylphenidate treatment. She experienced two episodes several hours apart characterized by unresponsiveness, chewing movements, and tonic eye deviation. An EEG immediately thereafter revealed mild right temporal slowing consistent with postictal change. She subsequently experienced seizures during treatment with lowdose (30 mg/day) imipramine monotherapy, during concurrent treatment with carbamazepine and dextroamphetamine, and during carbamazepine monotherapy. A repeat EEG revealed sharply contoured theta activity emanating from the right midtemporal region. Her seizures were managed with carbamazepine plus valproic acid, and stimulants were not reintroduced. The other three seizure patients demonstrated epileptiform EEG abnormalities. One was a 9-year-old female who had had a single seizure after abdominal surgery in the neonatal period and was seizure-free on no medication thereafter. Her EEG at the time of ADHD diagnosis revealed generalized epileptiform discharge. She was treated uneventfully with methylphenidate for 12 months but experienced a witnessed unprovoked 4-minute generalized tonic-clonic seizure 2 months after the methylphenidate was discontinued during a summer vacation. She was treated with valproic acid and, with the exception of one seizure during a transient period of noncompliance, subsequently remained seizure free on valproic acid and dextroamphetamine for the duration of follow-up (more than 5 years). The other two seizure patients were males, a 7-year, 9-month-old and a 6-year-old whose EEGs revealed left central temporal and right central and parietal
spikes, respectively. The first male experienced a 2-minute generalized tonic-clonic seizure with focal onset (tonic rightward head deviation) 3 years after initiation of methylphenidate therapy. Carbamazepine was initiated but was replaced by phenytoin because of drug allergies; methylphenidate was reinitiated and there were no further seizures for the subsequent 14 months of follow-up. The second male had an episode 10 months after methylphenidate was initiated, during which he was heard to fall and was found unresponsive with upward eye deviation for approximately 2 minutes. He was treated with valproic acid, was subsequently able to tolerate reintroduction of stimulant therapy (combined amphetamine salts), and had no further seizures during 2 months of follow-up. Thus the seizure incidence was one of 175 patients (0.6%) with a normal/nonepileptiform EEGs vs three of the 30 patients (10%) with epileptiform EEGs (P ⬍ 0.003, Fisher Exact Test). The 95% confidence interval for the 0.6% incidence of seizure occurrence with stimulant use after a normal EEG was 0-1.7%. In contrast, the 95% confidence interval for the 10% incidence in the epileptiform group was 0-20.7%, suggesting that the true proportion of patients with abnormal EEGs predicted to have seizures with stimulant use might be considerably overestimated or underestimated by these data. Among patients with abnormal EEGs, the incidence of seizures in the stimulant treated group (3/30) did not differ from the incidence in the untreated group (0/6, P ⬎ 0.5, Fisher Exact Test), probably because the sample size in the latter group was small. Discussion The 15.4% incidence of epileptiform EEGs in this study population is considerably higher than the estimated incidence (2%) of EEG abnormalities in an unselected population of nonepileptic children [11] and is twice the estimated incidence (7%) in a previously reported series of 100 ADHD patients studied consecutively in a pediatric neurology practice [12]. As we eliminated patients with conditions likely to have associated EEG abnormalities (e.g., known epilepsy, history of meningitis), as well as patients treated with antiepilepsy medications for any reason, we believe the rate of detection was not biased within our sample. Despite the fact that seizures only occurred in the stimulant-treated group, the data do not necessarily suggest that stimulants provoke seizures in the absence of an underlying diathesis. Indeed, only one patient (the 6-year, 7-month-old female) demonstrated marked sensitivity to stimulants, and her clinical course and subsequent EEG were consistent with a diagnosis of localization-related epilepsy (complex partial seizures) despite the initially normal EEG. The other three patients who experienced seizures all exhibited epileptiform EEGs. One patient (9-year-old female), whose EEG was consistent with primary generalized epilepsy, had discontinued her meth-
ylphenidate 2 months before her first seizure, although she had been exposed to it continuously for the preceeding 12 months. In the two male patients the EEG abnormalities were centrotemporal or centroparietal spikes consistent with benign rolandic epilepsy. Although neither patient had a seizure before initiation of stimulant therapy, both had long latencies to seizure occurrence (3 years in one and 10 months in the other), suggesting that stimulants were not strongly provocative agents. On the other hand, these patients represented 16.7% (2/12) of our treated patients with rolandic spikes, an incidence considerably higher than the 5-6% seizure rate reported for nonepileptic patients with this EEG pattern followed for 3-5 years [13-14]. Thus chronic stimulant administration in the presence of rolandic spikes may contribute to the expression of seizures. Overall, seizures occurred in 2% (4/205) of the stimulant-treated group. This rate is not exceptionally alarming given that an estimated 1% of unselected children will have at least one afebrile seizure by the 14 years of age and 0.4-0.8% will have epilepsy by 11 years of age [15]. However, within our population of ADHD patients, epileptiform EEGs identified a subgroup with seizure risk as high as 20%, whereas normal or nonepileptiform EEGs were strongly indicative of seizure risk of less than 1%. Although the sample size was relatively small, the incidence of epileptiform EEGs in our population was surprisingly high. Further, this is likely to be a conservative estimate of the true incidence among patients seeking relief from ADHD symptoms if those with concomitant neurologic problems associated with EEG abnormalities (e.g., developmental language disorders [16] or history of meningitis) are included. To physicians and parents weighing potential benefits of stimulants for children with ADHD, a seizure represents a markedly disturbing adverse event, particularly if attributed to a therapy usually considered elective. Our data suggest that a normal EEG can be used to assign children to a category of “minimal risk” for seizure. In contrast, an epileptiform EEG in neurologically normal children with ADHD predicts considerable risk for the eventual occurrence of seizure. The risk, however, is not necessarily attributable to stimulant use. This work was supported in part by a generous gift from the Crown Family. S.A.H. was a Hyndman Summer Fellow at the Evanston Northwestern Healthcare Research Institute. The authors thank Sunnah Kim RN, MS and Catherine J. Kosinski for their participation in this study.
References [1] MTA Cooperative Group. A 14-month randomized clinical trial of treatment strategies for attention-deficit/hyperactivity disorder. Arch Gen Psychiatry 1999;56:1073-86. [2] Du Paul GJ, Barkley RA. Medication therapy. In: Barkely RA, ed. Attention deficit hyperactivity disorder: A handbook for diagnosis and treatment. New York: Guilford Press, 1990:573-612.
Hemmer et al: Seizures, Stimulants, and ADHD 101
[3] McBride MC, Wang DD, Torres CF. Methylphenidate in therapeutic doses does not lower seizure threshold. [Abstract] Ann Neurol 1986;20:428. [4] Feldman H, Crumrine P, Handen BL, Alvin R, Teodori J. Methylphenidate in children with seizures and attention-deficit disorder. Am J Dis Child 1989;143:1081-6. [5] Gross-Tsur V, Manor O, van der Meere J, Joseph A, Shalev RS. Epilepsy and attention deficity hyperactivity disorder: is methylphenidate safe and effective? J Pediatr 1997;130:40-4. [6] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 3rd ed. Washington, DC: American Psychiatric Association, 1980. [7] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 3rd revised edition. Washington, DC: American Psychiatric Association, 1987. [8] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 4th ed. Washington, DC: American Psychiatric Association, 1994. [9] Goyette CH, Conners CK, Ulrich RF. Normative data on revised Conners parent and teacher rating scales. J Abnormal Child Psychology 1978;6:221-36.
102
PEDIATRIC NEUROLOGY
Vol. 24 No. 2
[10] Ullman RK, Sleator EK, Sprague RL. ACTeRS Teacher form, 2nd edition. Champaign: Meritech, Inc., 1991. [11] Petersen I, Eeg-Olofsson O, Sellden U. Paroxysmal activity in EEG of normal children. In: Kellaway P, Petersen I, eds. Clinical electroencephalography of children. Stockholm: Almquist and Wiksell, 1968:167-88. [12] Millichap JG. Tics, Tourette Syndrome, seizures, and headaches. In: Millichap JG, ed. Attention deficit hyperactivity and learning disorders. Chicago: PNB Publishers, 1998:95-118. [13] Verrotti A, Greco R, Altobelli E, Domizio S, Sabatino G, Morgese G, Chiarelli F. Centro-temporal spikes in non-epileptic children: A long-term follow up. J Paediatr Child Health 1999;35:60-2. [14] Okubo Y, Matsuura M, Asai T, et al. A follow-up study of healthy children with epileptiform EEG discharges. J Epilepsy 1993;6: 250-6. [15] Hauser WA, Hersdorffer DC. Incidence and prevalence. In: Epilepsy: frequency, causes, and consequences. New York: Demos, 1990:1-52. [16] Fejerman N, Caraballo R, Tenembaum SN. Atypical evolutions of benign localization-related epilepsies in children: Are they predictable? Epilepsia 2000;41:380-90.
Stimulant Therapy and Seizure Risk in Children With ADHD Sarah A. Hemmer*, Joseph F. Pasternak, MD*†, Steven G. Zecker, PhD‡, and Barbara L. Trommer, MD*† Stimulants are an effective treatment frequently prescribed for attention-deficit– hyperactivity disorder (ADHD), but they commonly are believed to lower the threshold for seizures. Although several studies have revealed that stimulants do not exacerbate well-controlled epilepsy, there is a paucity of data about seizure risk in nonepileptic children treated with stimulants. Two hundred thirty-four children (179 males, 9.1 ⴞ 3.6 years of age; 55 females, 9.6 ⴞ 3.9 years of age) with uncomplicated ADHD received electroencephalograms (EEGs) performed in our institution. Thirty-six patients (15.4%) demonstrated epileptiform abnormalities, and 198 (84.6%) demonstrated normal or nonepileptiform EEGs. Rolandic spikes accounted for 40% of the abnormal EEGs and 60% of those with focal abnormalities. Stimulant therapy was elected by 205 of 234 patients (87.6%). Seizures occurred only in the treated group, in one of 175 patients with a normal EEG (incidence 0.6%, 95% confidence intervals 0%1.7%) and three of 30 treated patients with epileptiform EEGs (incidence 10%, 95% confidence interval 0%-20.7%). Seizures occurred in two of 12 children (16.7%) with rolandic spikes. These data suggest that a normal EEG can be used to assign children with ADHD to a category of minimal risk for seizure. In contrast, an epileptiform EEG in neurologically normal children with ADHD predicts considerable risk for the eventual occurrence of seizure. The risk, however, is not necessarily attributable to stimulant use. © 2001 by Elsevier Science Inc. All rights reserved. Hemmer SA, Pasternak JF, Zecker SG, Trommer BL. Stimulant therapy and seizure risk in children with ADHD. Pediatr Neurol 2001;24:99-102.
*Department of Pediatrics; Division of Neurology; Evanston Hospital; Evanston, Illinois; †Northwestern University Medical School; Chicago, Illinois; and ‡Northwestern University Program in Learning Disabilities; Evanston, Illinois.
© 2001 by Elsevier Science Inc. All rights reserved. PII S0887-8994(00)00240-X ● 0887-8994/01/$—see front matter
Introduction Attention-deficit– hyperactivity disorder (ADHD) is the most common neurobehavioral disorder of childhood with an estimated prevalence of 8-10% in school-aged children [1]. Stimulant medications are effective, frequently prescribed treatments for ADHD [2]. Their use is apt to increase as a result of the recent multimodal treatment study [1] demonstrating long-term efficacy and superiority over other treatment modalities. Lowering of the seizure threshold is commonly considered to be a risk of stimulant therapy. Whereas the safety of stimulants in patients with treated, well-controlled epilepsy and concurrent ADHD has been documented [3-5], there is little information to guide the clinician regarding seizure risk in uncomplicated ADHD. The present study examined the relationships among electroencephalographic (EEG) findings, stimulant use, and seizure occurrence in children with ADHD without known epilepsy. Methods and Materials The medical records were reviewed of all patients referred to a hospital-based child neurology practice for evaluation of suspected ADHD or attention-deficit disorder (ADD) between 1993 and 1998 (n ⫽ 765) under a local Institutional Review Board-approved protocol. EEGs routinely were recommended for patients when the diagnosis of ADHD was made and the use of stimulant medication was planned. To be included in the study, patients were required to meet diagnostic criteria for ADHD according to DSM-III (ADD with or without hyperactivity) [6], DSM-III-R [7], or DSM-IV (primarily inattentive, primarily hyperactive/impulsive, or combined subtypes) [8]. Diagnosis was made by a child neurologist on the basis of parental endorsement of symptom descriptors supplemented by abbreviated Conners Parent Symptom Questionnaires [9], ActERs Teacher Profile questionnaires [10], and additional history provided by patients, parents, and teachers. Children with neurologic deficits or neurologic diagnoses other than learning disability, headache, and tic disorders were excluded, as were children taking antiepilepsy medications for any reason. In addition to ADHD
Communications should be addressed to: Dr. Trommer; Evanston Hospital; Division of Neurology; 2650 Ridge Avenue; Evanston, IL 60201. Received June 6, 2000; accepted September 21, 2000.
Hemmer et al: Seizures, Stimulants, and ADHD 99
Table 1.
Distribution of EEG findings in 234 patients with ADHD
EEG
Stimulant Treated (n ⴝ 205)
Untreated (n ⴝ 29)
Total (n ⴝ 234)
175 (1)
23
198
6
36
Normal
Epileptiform
30
Focal
Rolandic
Nonfocal
Nonrolandic
20
12 (2)
10 (1)
8
2
1
1
4
22
14
13
9
EEGs were obtained in 234 patients for whom stimulant therapy was recommended. The number of subjects with each EEG finding is presented for those who elected stimulant treatment (left column), those who declined (middle column), and for the total group (right column). Seizures were only seen in the stimulant-treated group; the number of patients who had at least one seizure during the study period is indicated in parentheses in the corresponding EEG category.
diagnosis, patients were required to have received an EEG in our institution and follow-up either by office visit or by telephone in 1999. EEGs were performed after the initial visit during which diagnosis was made, before or up to 8 weeks after initiation of stimulant therapy, and were interpreted by a board-certified clinical neurophysiologist (J.F.P.). EEGs were classified as normal/nonepileptiform or epileptiform. The epileptiform group was subdivided according to whether the abnormalities were nonfocal or focal, and within the focal abnormal group those with centrotemporal or centroparietal “rolandic” spikes were tabulated separately. Methylphenidate (0.3-1 mg/kg per day) was the initial stimulant prescribed. In some patients, 0.1-0.5 mg/kg per day of dextroamphetamine sulfate or mixed amphetamine salts (Adderall, ShireRichwood) was substituted, with dose adjustments and medication choice guided by clinical response. Statistical analyses were performed using Systat (SPSS, Chicago, IL) and Microsoft Excel software (Redmond, WA).
Results Two hundred thirty-four patients met our inclusion criteria. They consisted of 179 males who were 9.1 ⫾ 3.6 years of age (mean ⫾ S.D., range ⫽ 3-20 years) and 55 females who were 9.6 ⫾ 3.9 years of age (range ⫽ 3-19 years). Normal or nonepileptiform EEGs were found in 198 patients (84.6%) and epileptiform EEGs in 36 (15.4%). The distribution of EEG abnormalities is summarized in Table 1. Stimulants were offered to all 234 patients, but the decision to use them was ultimately made by parents. Stimulant treatment was elected by 205 and declined by 29 patients. The incidence of epileptiform EEGs did not differ between the treated (30/205) and untreated (6/29) groups (14.6% vs 20.7%, respectively, P ⫽ 0.7, binomial test) suggesting that the presence of an abnormal EEG did not bias the decision to treat. Rolandic spikes occurred in 5.6% of the study population and accounted for 40% of the abnormal EEGs and 60% of the EEGs with focal abnormalities. Spikes were right-sided in five patients, left-sided in four, and bilateral
100
PEDIATRIC NEUROLOGY
Vol. 24 No. 2
or midline in four with no gender-related pattern in the distribution. Seizures occurred in four patients, all of whom were in the stimulant-treated group. Only one seizure patient had an initial normal EEG. She was a 6-year, 7-month-old female whose first seizure occurred 6 weeks after initiation of methylphenidate treatment. She experienced two episodes several hours apart characterized by unresponsiveness, chewing movements, and tonic eye deviation. An EEG immediately thereafter revealed mild right temporal slowing consistent with postictal change. She subsequently experienced seizures during treatment with lowdose (30 mg/day) imipramine monotherapy, during concurrent treatment with carbamazepine and dextroamphetamine, and during carbamazepine monotherapy. A repeat EEG revealed sharply contoured theta activity emanating from the right midtemporal region. Her seizures were managed with carbamazepine plus valproic acid, and stimulants were not reintroduced. The other three seizure patients demonstrated epileptiform EEG abnormalities. One was a 9-year-old female who had had a single seizure after abdominal surgery in the neonatal period and was seizure-free on no medication thereafter. Her EEG at the time of ADHD diagnosis revealed generalized epileptiform discharge. She was treated uneventfully with methylphenidate for 12 months but experienced a witnessed unprovoked 4-minute generalized tonic-clonic seizure 2 months after the methylphenidate was discontinued during a summer vacation. She was treated with valproic acid and, with the exception of one seizure during a transient period of noncompliance, subsequently remained seizure free on valproic acid and dextroamphetamine for the duration of follow-up (more than 5 years). The other two seizure patients were males, a 7-year, 9-month-old and a 6-year-old whose EEGs revealed left central temporal and right central and parietal
spikes, respectively. The first male experienced a 2-minute generalized tonic-clonic seizure with focal onset (tonic rightward head deviation) 3 years after initiation of methylphenidate therapy. Carbamazepine was initiated but was replaced by phenytoin because of drug allergies; methylphenidate was reinitiated and there were no further seizures for the subsequent 14 months of follow-up. The second male had an episode 10 months after methylphenidate was initiated, during which he was heard to fall and was found unresponsive with upward eye deviation for approximately 2 minutes. He was treated with valproic acid, was subsequently able to tolerate reintroduction of stimulant therapy (combined amphetamine salts), and had no further seizures during 2 months of follow-up. Thus the seizure incidence was one of 175 patients (0.6%) with a normal/nonepileptiform EEGs vs three of the 30 patients (10%) with epileptiform EEGs (P ⬍ 0.003, Fisher Exact Test). The 95% confidence interval for the 0.6% incidence of seizure occurrence with stimulant use after a normal EEG was 0-1.7%. In contrast, the 95% confidence interval for the 10% incidence in the epileptiform group was 0-20.7%, suggesting that the true proportion of patients with abnormal EEGs predicted to have seizures with stimulant use might be considerably overestimated or underestimated by these data. Among patients with abnormal EEGs, the incidence of seizures in the stimulant treated group (3/30) did not differ from the incidence in the untreated group (0/6, P ⬎ 0.5, Fisher Exact Test), probably because the sample size in the latter group was small. Discussion The 15.4% incidence of epileptiform EEGs in this study population is considerably higher than the estimated incidence (2%) of EEG abnormalities in an unselected population of nonepileptic children [11] and is twice the estimated incidence (7%) in a previously reported series of 100 ADHD patients studied consecutively in a pediatric neurology practice [12]. As we eliminated patients with conditions likely to have associated EEG abnormalities (e.g., known epilepsy, history of meningitis), as well as patients treated with antiepilepsy medications for any reason, we believe the rate of detection was not biased within our sample. Despite the fact that seizures only occurred in the stimulant-treated group, the data do not necessarily suggest that stimulants provoke seizures in the absence of an underlying diathesis. Indeed, only one patient (the 6-year, 7-month-old female) demonstrated marked sensitivity to stimulants, and her clinical course and subsequent EEG were consistent with a diagnosis of localization-related epilepsy (complex partial seizures) despite the initially normal EEG. The other three patients who experienced seizures all exhibited epileptiform EEGs. One patient (9-year-old female), whose EEG was consistent with primary generalized epilepsy, had discontinued her meth-
ylphenidate 2 months before her first seizure, although she had been exposed to it continuously for the preceeding 12 months. In the two male patients the EEG abnormalities were centrotemporal or centroparietal spikes consistent with benign rolandic epilepsy. Although neither patient had a seizure before initiation of stimulant therapy, both had long latencies to seizure occurrence (3 years in one and 10 months in the other), suggesting that stimulants were not strongly provocative agents. On the other hand, these patients represented 16.7% (2/12) of our treated patients with rolandic spikes, an incidence considerably higher than the 5-6% seizure rate reported for nonepileptic patients with this EEG pattern followed for 3-5 years [13-14]. Thus chronic stimulant administration in the presence of rolandic spikes may contribute to the expression of seizures. Overall, seizures occurred in 2% (4/205) of the stimulant-treated group. This rate is not exceptionally alarming given that an estimated 1% of unselected children will have at least one afebrile seizure by the 14 years of age and 0.4-0.8% will have epilepsy by 11 years of age [15]. However, within our population of ADHD patients, epileptiform EEGs identified a subgroup with seizure risk as high as 20%, whereas normal or nonepileptiform EEGs were strongly indicative of seizure risk of less than 1%. Although the sample size was relatively small, the incidence of epileptiform EEGs in our population was surprisingly high. Further, this is likely to be a conservative estimate of the true incidence among patients seeking relief from ADHD symptoms if those with concomitant neurologic problems associated with EEG abnormalities (e.g., developmental language disorders [16] or history of meningitis) are included. To physicians and parents weighing potential benefits of stimulants for children with ADHD, a seizure represents a markedly disturbing adverse event, particularly if attributed to a therapy usually considered elective. Our data suggest that a normal EEG can be used to assign children to a category of “minimal risk” for seizure. In contrast, an epileptiform EEG in neurologically normal children with ADHD predicts considerable risk for the eventual occurrence of seizure. The risk, however, is not necessarily attributable to stimulant use. This work was supported in part by a generous gift from the Crown Family. S.A.H. was a Hyndman Summer Fellow at the Evanston Northwestern Healthcare Research Institute. The authors thank Sunnah Kim RN, MS and Catherine J. Kosinski for their participation in this study.
References [1] MTA Cooperative Group. A 14-month randomized clinical trial of treatment strategies for attention-deficit/hyperactivity disorder. Arch Gen Psychiatry 1999;56:1073-86. [2] Du Paul GJ, Barkley RA. Medication therapy. In: Barkely RA, ed. Attention deficit hyperactivity disorder: A handbook for diagnosis and treatment. New York: Guilford Press, 1990:573-612.
Hemmer et al: Seizures, Stimulants, and ADHD 101
[3] McBride MC, Wang DD, Torres CF. Methylphenidate in therapeutic doses does not lower seizure threshold. [Abstract] Ann Neurol 1986;20:428. [4] Feldman H, Crumrine P, Handen BL, Alvin R, Teodori J. Methylphenidate in children with seizures and attention-deficit disorder. Am J Dis Child 1989;143:1081-6. [5] Gross-Tsur V, Manor O, van der Meere J, Joseph A, Shalev RS. Epilepsy and attention deficity hyperactivity disorder: is methylphenidate safe and effective? J Pediatr 1997;130:40-4. [6] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 3rd ed. Washington, DC: American Psychiatric Association, 1980. [7] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 3rd revised edition. Washington, DC: American Psychiatric Association, 1987. [8] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 4th ed. Washington, DC: American Psychiatric Association, 1994. [9] Goyette CH, Conners CK, Ulrich RF. Normative data on revised Conners parent and teacher rating scales. J Abnormal Child Psychology 1978;6:221-36.
102
PEDIATRIC NEUROLOGY
Vol. 24 No. 2
[10] Ullman RK, Sleator EK, Sprague RL. ACTeRS Teacher form, 2nd edition. Champaign: Meritech, Inc., 1991. [11] Petersen I, Eeg-Olofsson O, Sellden U. Paroxysmal activity in EEG of normal children. In: Kellaway P, Petersen I, eds. Clinical electroencephalography of children. Stockholm: Almquist and Wiksell, 1968:167-88. [12] Millichap JG. Tics, Tourette Syndrome, seizures, and headaches. In: Millichap JG, ed. Attention deficit hyperactivity and learning disorders. Chicago: PNB Publishers, 1998:95-118. [13] Verrotti A, Greco R, Altobelli E, Domizio S, Sabatino G, Morgese G, Chiarelli F. Centro-temporal spikes in non-epileptic children: A long-term follow up. J Paediatr Child Health 1999;35:60-2. [14] Okubo Y, Matsuura M, Asai T, et al. A follow-up study of healthy children with epileptiform EEG discharges. J Epilepsy 1993;6: 250-6. [15] Hauser WA, Hersdorffer DC. Incidence and prevalence. In: Epilepsy: frequency, causes, and consequences. New York: Demos, 1990:1-52. [16] Fejerman N, Caraballo R, Tenembaum SN. Atypical evolutions of benign localization-related epilepsies in children: Are they predictable? Epilepsia 2000;41:380-90.