PEDIATRICS Vol. 122 No. 4 October 2008, pp. e911-e916 (doi:10.1542/peds.2008-0257)
ARTICLE |
The Case Against Routine Electroencephalography in Specific Language Impairment
a Departments of Neurology/Neurosurgery
b Pediatrics, McGill University, Montreal Children's Hospital-McGill University Health Center, Montreal, Quebec, Canada
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
BACKGROUND. Specific language impairment is a primary developmental language disorder in which language is impaired disproportionately to other developmental domains. Electroencephalography is often conducted in the medical investigation of a child with specific language impairment; however, at present, there is uncertainty regarding necessary testing using electroencephalography.
METHODS. The cases of 111 children with the diagnosis of specific language impairment over a 10-year interval, who also underwent electroencephalography, were systematically reviewed in a retrospective manner. Children with a history of previous afebrile seizures, acquired language delay, or documented language regression, developmental delay, hearing loss, coexisting autistic features, and known central nervous system disorders were excluded.
RESULTS. The majority (76%) of the children were boys. Thirty-five (31.5%) children had abnormal electroencephalography results, including 7 (6.3%) children with epileptiform activity. This is higher than the prevalence rate of epileptiform activity in a historical cohort of 3726 (3.54%) children but not statistically significant. The epileptiform activity was deemed active in only 3 of 7 patients and was not related to the specific type of language delay observed.
CONCLUSIONS. Although abnormal electroencephalographic activity is seen frequently in children with specific language impairment, epileptiform activity is rare and without apparent impact on clinical care. Awake electroencephalography does not seem to be useful in the routine diagnostic evaluation of young children with specific language impairment, although further investigations of both wake and sleep electroencephalography in this homogenous population must be conducted before definitive recommendations can be made.
Key Words: specific language impairment • EEG • pediatric • language delay
Abbreviations: SLI—specific language impairment • ADHD—attention-deficit/hyperactivity disorder • EEG—electroencephalography • PA—paroxysmal activity
Specific language impairment (SLI) is known by many other descriptive terms, such as "developmental language impairment," "developmental dysphasia," and "congenital dysphasia." It is estimated to have a prevalence of 7.4%1 and is a diagnosis likely to persist over the childhood years.2
In its purest terms, SLI is a primary developmental language disorder in which language is impaired disproportionately to other developmental domains arising within the context of normal hearing and the absence of coexisting autistic features. The International Classification of Diseases, 10th Revision, categorizes SLI into expressive and receptive language subtypes. In general, SLI is defined as a language disorder with language skills 2 SDs below the mean, with at least a 1-SD difference between verbal and nonverbal skills.1
A proposed definition by Tomblin et al1,3 defines SLI as a combination of normal intelligence and language impairment 1.25 SD below the mean. A diagnosis of SLI should not be given to children with language regression, acquired aphasia, or autistic traits. The majority of children with SLI do demonstrate more subtle difficulties with motor function,4 and there is a higher-than-expected frequency of coexisting neurobehavioral disorders, such as attention-deficit/hyperactivity disorder (ADHD).5
Although SLI is common in the pediatric population, the underlying causes of this entity are largely unknown.6 Studies of language processing have demonstrated a genetic predisposition for difficulty with phonological short-term memory.7 Uncertainty exists regarding the medical testing to be undertaken in the investigation of a child with SLI. One investigation often conducted routinely in this clinical context is electroencephalography (EEG).
The primary objective of this study was to determine the proportion of children with SLI encountered in ambulatory practice who have abnormalities on EEG and, more specifically, the frequency of epileptiform activity. Secondarily, we sought to identify whether any specific clinical features were associated with the presence of an abnormal EEG.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
A systematic retrospective review was undertaken of children seen in a single pediatric neurology practice in a variety of settings: (1) community-based office practice; (2) ambulatory neurology hospital-based clinic; and (3) university hospital office practice. Each patient was evaluated by a single pediatric neurologist (Dr Shevell) between the years of 1991 and 2000 (inclusive). Subsequent to 2000, EEG was no longer routinely ordered for patients with possible SLI because of a perceived low yield in diagnosing underlying disorders.
Diagnosis and follow-up information were entered into a standardized computerized database. A systematic screen of this database initially identified children with the diagnosis of SLI that was confirmed on detailed review of medical charts to extract variables of interest. Patients were given a diagnosis of SLI if they had a substantial congenital delay in language outside of the normal distribution of language acquisition, without obvious delays in other milestones (other than mild fine motor delay) and normal intelligence. This was determined by the evaluating neurologist (Dr Shevell). This diagnosis did not require evaluation by a speech language pathologist. Children who had age-appropriate comprehension but delay in expression were classified as having "expressive language delay," and those who were able to express but did not have developmentally appropriate comprehension were categorized as having "receptive delay." A mix of both expressive and receptive language delay was defined as "mixed delay." Exclusion criteria included an acquired language delay (eg, postcerebrovascular accident), evident language regression, a clinical history of afebrile seizures, known central nervous system structural abnormalities, audiometric documentation of hearing loss, and coexisting autistic traits in the social domain.
Variables extracted from standardized review of the children's medical charts included gender, age at initial parental concern, age at first visit, type of language delay (ie, expressive, receptive, or mixed), handedness (right, left, or ambidextrous), family history, comorbid disorders (ie, learning difficulties, ADHD, or behavioral difficulties), physical examination findings (ie, macrocephaly, microcephaly, or fine motor/coordination difficulties), EEG results, evaluations (ie, speech therapy, audiology, and psychology), and other investigations (ie, imaging).
The routine EEG ordered consisted of a standard 30-minute EEG as per the American Electroencephalographic Society, which included both photic stimulation and hyperventilation as activation procedures. Sleep recordings or 24-hour telemetries were not requested at the initial evaluation. Until the year 2000, EEG was ordered as part of a screen by the evaluating neurologist for children diagnosed with SLI. EEG results were categorized as normal, abnormal without epileptiform activity (ie, background abnormalities), and abnormal with epileptiform activity. Those EEG results that were abnormal without epileptiform activity were further categorized as mild-to-moderate or moderate-severe disturbance of background or abnormalities brought out only by activation procedures. Each abnormal EEG was then categorized according to the region affected (ie, generalized, posterior quadrants, centrotemporal, centroparietal, multifocal, or temporal) and by laterality (right or left). Epileptiform activity was considered to be any abnormal activity consisting of repetitive spikes, repetitive sharp waves, or spike and slow wave discharges. EEG results were reviewed systematically by 1 of 3 pediatric electroencephalographers at our institution at that time.
Demographic characteristics of the population were described, and frequencies and cross-tabulations of comorbidities were conducted. Comparison with a historic control group8 was made by Pearson 
2 analysis. Variables were all categorical and binomial: gender, handedness (right versus left/ambidextrous), physical examination (normal versus soft signs), and type of speech delay (expressive versus receptive/mixed). It was hypothesized a priori that handedness, gender, the type of speech delay documented, and the evidence of fine motor delay may be associated with an abnormal EEG. Crude odds ratios and 95% confidence intervals of factors with a possible association with an abnormal EEG were determined by univariate binomial logistic regression. The 4 covariates were then entered into a backward regression model to determine independence. A P value of .05 was considered a priori to be significant. Statistical analysis was conducted by using SPSS 12.0 (SPSS Inc, Chicago, IL).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
A total of 130 patients were given a diagnosis of SLI at initial evaluation and met inclusion and exclusion criteria. Of these, 111 patients had EEG undertaken (note: 11 patients did not have EEG ordered, 5 had missing data, and 3 patients did not attend the requested appointment). There was no difference in the demographic features between the children who had EEG performed and those who did not (Table 1).
|
In the group of children who underwent EEG (n = 111), parents were first concerned about their child's language development at 2.14 ± 0.67 years of age (range: 1.0–5.0 years). EEG was performed at 4.40 ± 1.49 years of age (range: 2.0–10.0 years) invariably within 2 months of their initial evaluation by the neurologist. Right-handed children made up the majority of patients (n = 75 [67.6%]), 7 (6.3%) were ambidextrous, and data regarding handedness were missing in 20 (18%) patients. The types of speech delays were categorized as expressive, receptive, and mixed. Seventy (63%) children had an available evaluation by a speech therapist, because either reports were unavailable or missing in the remaining children. Most (n = 61 [55%]) children had a mixed speech delay, whereas 47 (42.3%) children had an expressive speech delay, which was determined by the evaluating neurologist (Dr Shevell). When available, the speech therapist evaluation uniformally confirmed the neurologist's classification assessment.
Of the 111 children with EEG performed, fine motor abnormalities were noted in 23 (20.7%) children. Mild behavior abnormalities were present in 11 (9.9%) children, and an ADHD was formally diagnosed in 15 (13.5%) children. Fourteen (12.6%) children also had comorbid learning difficulties. Soft neurologic signs, macrocephaly, or microcephaly were present in 21 (18.9%) children. It was surprising that none of the patients had a history of previous febrile seizures.
Telemetry or sleep-deprived EEG was not performed during the initial recording in the 111 patients evaluated by EEG (Table 1). Nevertheless, 31 (27.9%) of 111 children spontaneously attained stage II (n = 26) or stage III/IV (n = 5) sleep during the routine evaluation. Normal EEG results constituted 76 (68.5%) of 111 of the initial recordings. Of the 35 abnormal EEG results, 7 (6.3%) of 111 overall were epileptiform in nature, 18 (16.2%) of 111 overall had mild background disturbances, and 3 (2.7%) of 111 overall had moderate abnormalities, 1 (0.09%) with severe abnormalities on an awake EEG with some epileptiform activity with sleep deprivation (see Table 2). Three (2.7%) of 111 patients had excessive paroxysmal activity (PA) with hyperventilation, and 3 (2.7%) of 111 had excessive PA with photic stimulation.
|
The 7 children with epileptiform activity are described in Table 3. Only 2 of these patients reached stage II sleep during their EEG. One patient only had epileptiform activity when provoked by hyperventilation. There were no 2 epileptiform electroencephalograms that were affected in the same region. Two patients with active EEGs were treated with anticonvulsant agents at the specific request of the parents (valproic acid or carbamazepine) with an improved EEG on subsequent recordings but no apparent change in the child's clinical status. Neither child later developed clinical seizures.
|
Compared with a historic cohort,8 in which 3.54% of 3726 normal children aged 6 to 13 years had epileptiform activity on EEG, the proportion of children with abnormalities in our group was greater but not to a statistically significant extent (Fisher's exact test, P = .120). Female gender had a statistically significant association with an abnormal EEG (2.63 [95% confidence interval: 1.1–6.4]; P = .036) on crude analysis, but when adjusted for potential confounders, it was no longer statistically significant. Physical examination abnormalities, handedness, and type of speech delay were not associated with an abnormal EEG.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Our study demonstrated that 31.5% of the children with SLI had abnormal EEG results, with 6.3% of the children having actual epileptiform activity. The high prevalence of boys with speech and language impairment in our study (75.7%) is consistent with previous studies.6,9,10 The majority of our children had a mixed expressive and receptive SLI subtype. Only female gender seemed to have a small association with the presence of an abnormal EEG result on crude analysis. There seemed to be no association between the specific type of language delay documented or the handedness of the child and the presence of an abnormal EEG result.
The epileptiform activity seen did not consist of electrographic seizures, and, for those with sleep studies, did not have any present evidence of a possible electrical status epilepticus during slow wave sleep or Landau-Kleffner syndrome. Treatment in 2 children seemed to improve the EEG results but was not sufficient to affect overt language development. This is not surprising, because changes in EEG may reflect impairments in brain maturation that may manifest with various phenotypes.11 Other types of language disorders (reading disability and speech sound disorder) have been found in children with rolandic epilepsy, as well as in family members of children with rolandic epilepsy.12 The EEG findings in rolandic epilepsy, which are centrotemporal spikes, can anatomically correlate with difficulties in expressive speech disorders. In our case, the epileptic activity on EEG was not confined to a particular anatomic distribution.
Previous studies have shown no consistency between the severity of the language deficit documented and the extent of EEG abnormalities.13,14 We could not reliably categorize the severity of the SLI, because not all of the reports from speech language pathology were available, nor was a single standardized language assessment undertaken that would permit such objective stratification. Studies have also observed increased epileptiform activity in those patients with receptive dysphasia more so than expressive dysphasia.15 More than half of our group did have a mixed expressive and receptive dysphasia (53.8%), and a similar proportion of children (4 of 7) with epileptiform activity was diagnosed with a mixed dysphasia.
Our results can be contrasted to a previous study (historical control) of 3726 neurologically normal children ages 6 to 13 years without a history of seizures8 in whom 3.54% of children were found to have epileptiform activity on EEG. The majority of these children were boys (66.4%). These EEGs were recorded during wakefulness and with hyperventilation only. PA on standard wake EEG has also been observed in 2.7% of the children aged 1 to 15 years16 and 1.5% of 1000 normal children.17
Several EEG studies in various populations of children with SLI and other developmental language impairments have been conducted and invite comparison with ours. An initial description of focal and generalized epileptiform abnormalities was given in a case series of 7 patients14 who had a congenital developmental dysphasia, more receptive than expressive. A study by Nasr et al10 included 138 children with developmental dysphasia, of whom 13 patients had language regression, 11 had autistic features, 16 had global developmental delay, and 17 had a history of epilepsy. Both sleep and wake studies were performed in 90% of these patients. The majority (17 of 19 [89.5%]) of those children in this study with developmental dysphasia without language regression and epileptiform activity on EEG already had a diagnosis of epilepsy. In the group without regression, despite other comorbidities (ADHD, autism, or global developmental delay), they were unable to detect paroxysmal epileptiform patterns during sleep or wake EEG in patients without a previous history of seizures.
Duvelleroy-Hommet et al18 demonstrated that PA during sleep was higher (9 of 15 vs 2 of 39; P < .001) in patients with expressive developmental dysphasia without language regression or a previous history of epilepsy compared with a control group of children. Most of these patients only had PA in <0.1% of actual sleep time. The authors concluded that pure expressive developmental dysphasia was unlikely to be the consequence of the PA. None of the patients had electrical status epilepticus during slow wave sleep.
A prospective study including both wake and sleep-deprived EEG was conducted in a heterogenous group of children with developmental dysphasias with and without a history of seizures.13 Standard awake EEG revealed that slightly less than one third (10 of 32) had abnormal EEG results, of which only 2 did not have a history of seizures. Sleep-deprived EEG demonstrated abnormalities in almost half of the patients (14 of 32), of whom 6 already had abnormalities in the standard EEG, and overnight recordings demonstrated epileptic features in 30 of these 32 patients. This is consistent with Tuchman et al,19 who observed that 20% of dysphasic nonautistic children had an abnormal EEG result, with 8% having seizures and epileptiform activity. They found that those with a receptive subtype of dysphasia were at highest risk for an abnormal EEG result. They also found that the rate of epilepsy was higher in children with isolated dysphasia without regression.
Continuous 24-hour recordings were undertaken in a group of children with varying types of developmental dysphasias.15 Half had PA during sleep, the majority having a mixed receptive and expressive dysphasia. This was in comparison with only 10% of control subjects with PA. Recommendations by the authors of that particular study were to do sleep recordings in patients with developmental dysphasia with a comprehension component and to give antiepileptic treatment to patients with >8% PA for a minimum of 6 months. However, treatment in previous groups of similar patients9,14 did not show any apparent improvement.
In contrast, in patients with language regression, EEG results have been shown to be abnormal more frequently than in developmental dysphasia.10,20 Furthermore, in isolated language regression there is a higher association with seizures and epileptiform activity compared with children with both language regression and autistic regression at a highly significant level (P < .001).20 The authors conclude that, in isolated language regression, both the language regression and the seizures may be a part of the same underlying cortical pathophysiology.
Our population of children was purposely kept homogenous to reflect the typical patient with developmental language impairment without comorbidities other than mild behavioral problems and soft neurologic signs encountered in ambulatory pediatric practice. This differs from all of the previous studies except that of Maccario et al,14 which, however, did include some children with regression, autistic tendencies, seizures, and other delays that would predispose these patients to have abnormal EEG results.
Although our study did not involve sleep EEG at intake, 28.9% of patients with a normal EEG result and 22.9% of patients with an abnormal EEG result did have sleep components in their EEG. Most children spontaneously attained stage II sleep, with a minority reaching stage III and IV sleep, and only 2 of the 31 children had paroxysmal nonepileptiform activity on EEG during sleep.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Although abnormal EEG activity is frequently present in children with SLI, epileptiform activity is not present in a statistically significant number of these children compared with a historical control group. None of the children had electrographic seizures. It seems that routine EEG is not useful in the initial diagnostic evaluation of a young child with SLI, unlike in those children with language regression or those with autistic features. However, further prospective studies looking at homogenous populations of children with specific SLI subtypes objectively defined on this aspect of practice with both wake and sleep EEG recordings are necessary before definitive final recommendations can be made.
|
ACKNOWLEDGMENTS |
|---|
Dr Shevell was supported by the Montreal Children's Hospital Foundation during the writing of this article.
We thank Alba Rinaldi for secretarial assistance and Evelyn Constantin for assistance with statistical analysis.
|
FOOTNOTES |
|---|
Accepted Jun 4, 2008.
Address correspondence to Michael Shevell, MDCM, FRCP(C), Montreal Children's Hospital, Room A-514, 2300 Tupper St, Montreal, Quebec, Canada H3H 1P3. E-mail: michael.shevell{at}muhc.mcgill.ca
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject Previous studies of heterogeneous sample characteristics have been unclear about whether EEG adds useful information to the evaluation of children with an SLI.
|
| What This Study Adds This study demonstrates that routine EEG adds little additional information to the evaluation of children with an SLI.
|
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
- Tomblin J, Records N, Buckwalter P, Zhang X, Smith E, O'Brien M. Prevalence of specific language impairment in kindergarten children.
J Speech Lang Hear Res. 1997;40
(6):1245
–1260
[Abstract/Free Full Text] - Tomblin J, Zhang X, Buckwalter P, O'Brien M. The stability of primary language disorder: four years after kindergarten diagnosis.
J Speech Lang Hear Res. 2003;46
(6):1283
–1296
[Abstract/Free Full Text] - Tomblin J, Records N, Zhang X. A system for the diagnosis of specific language impairment in kindergarten children.
J Speech Lang Hear Res. 1996;39
(6):1284
–1294
[Abstract/Free Full Text] - Webster R, Erdos C, Evans K, et al. The clinical spectrum of developmental language impairment in school-aged children: language, cognitive, and motor findings. Pediatrics. 2006;118 (5). Available at: www.pediatrics.org/cgi/content/full/118/5/e1541
- Manor O, Shalev R, Joseph A, Gross-Tsur V. Arithmetic skills in kindergarten children with developmental language disorders. Eur J Pediat Neurol. 2001;5 (2):71 –77[CrossRef]
- Shevell MI, Majnemer A, Rosenbaum P, Abrahamowicz M. Etiologic yield in single domain developmental delay: a prospective study. J Pediatr. 2000;137 (5):633 –637[CrossRef][Web of Science][Medline]
- Bishop DV. The role of genes in the etiology of specific language impairment. J Commun Disord. 2002;35 (4):311 –328[CrossRef][Web of Science][Medline]
- Cavazzuti G, Cappella L, Nalin A. Longitudinal study of epileptiform EEG patterns in normal children. Epilepsia. 1980;21 (1):43 –55[Web of Science][Medline]
- Simms M, Schum R. Preschool children who have atypical patterns of development. Pediat Rev. 2000;21 (5):147 –158[CrossRef]
- Nasr J, Gabis L, Savatic M, Andriola M. The electroencephalogram in children with developmental dysphasia. Epilepsy Behav. 2001;2 (2):115 –118[CrossRef][Medline]
- Doose H, Neubauer BA, Petersen B. The concept of hereditary impairment of brain maturation. Epileptic Disord. 2000;2 (suppl 1):S45 –S49[Web of Science][Medline]
- Clarke T, Strug LJ, Murphy PL, et al. High risk of reading disability and speech sound disorder in rolandic epilepsy families: case-control study. Epilepsia. 2007;48 (12):2258 –2265[Medline]
- Echenne B, Cheminal R, Rivier F, Negre C, Touchon J, Billiard M. Epileptic electroencephalographic abnormalities and developmental dysphasias: a study of 32 patients. Brain Dev. 1992;14 (4):216 –225[Web of Science][Medline]
- Maccario M, Hefferen S, Keblusek S, Lipinski K. Development dysphasia and electroencephalographic abnormalities. Dev Med Child Neurol. 1982;24 (2):141 –155[Web of Science][Medline]
- Picard A, Cheliout Heraut F, Bouskraoui M, Lemoine M, Lacert P, Delattre J. Sleep EEG and developmental dysphasia. Dev Med Child Neurol. 1998;40 (9):595 –599[Web of Science][Medline]
- Eeg-Olofsson O, Petersen I, Sellden U. The development of the electroencephalogram in normal children from the age of 1 through 15 years. Neuropediatrie. 1971;2 (4):375 –404
- Peterson U, Koepp P, Solmsen M, Villiez T. Aphasie im Kindesalter mit EEG veranderungen. Neuropediatrie. 1978;9 (1):84 –96
- Duvelleroy-Hommet C, Billard C, Lucas B, et al. Sleep EEG and developmental dysphasia: lack of a consistent relationship with paroxysmal EEG activity during sleep. Neuropediatrics. 1994;26 (1):14 –18[Web of Science]
- Tuchman R, Rapin I, Shinnar S. Autistic and dysphasic children. II: epilepsy [published correction appears in Pediatrics. 1992;90(2 pt 1):264].
Pediatrics. 1991;88
(6):1219
–1225
[Abstract/Free Full Text] - McVicar K, Ballaban-Gil K, Rapin I, Moshé S, Shinnar S. Epileptiform EEG abnormalities in children with language regression.
Neurology. 2005;65
(1):129
–131
[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2008 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
