Epilepsy & Behavior 47 (2015) 143–146

Contents lists available at ScienceDirect

Epilepsy & Behavior

journal homepage: www.elsevier.com/locate/yebeh

Current controversies in the relationships between autism and epilepsy

a r t i c l e i n f o a b s t r a c t

Keywords: The controversies that have arisen in endeavoring to establish the nature of the relationships between autism
Epilepsy and epilepsy might be summarized in a few simple questions, most of which do not yet have clear, complete an-
Autism swers. Does epilepsy cause autism? Does autism cause epilepsy? Are there underlying brain mechanisms that
Comorbidity predispose to both conditions? What is the role of genetics in this regard? What is the importance of prenatal,
Genetics
perinatal, and postnatal environmental factors? Do any of the proposed relationships between autism and epi-
lepsy provide insight into useful management or treatment? Is the prognosis of either autism or epilepsy differ-
ent when the other condition is also present? What is the role of additional comorbidities, such as intellectual
impairment or attention deficit hyperactivity disorder, in the relationship between the two conditions and in
influencing treatment choices? From the evidence currently available, it would appear that epilepsy can rarely
be the cause of autistic features but is not the cause of autism in most cases. There is currently no credible mech-
anism for suggesting that autism might cause epilepsy. There is strong evidence for an underlying predisposition
for both conditions, particularly arising from genetic investigations. However, many issues remain unresolved.
Considering the amount of research that has been published in this area, it is surprising that so few definitive an-
swers have been established. The papers in this issue’s special section provide additional insights into the rela-
tionships between autism and epilepsy; while they do not provide answers to all the questions, they represent
considerable progress in this area and, at the very least, give some strong indication of what research might, in
the future, provide such answers.
© 2015 Published by Elsevier Inc.

Introduction criteria were narrow, and awareness of autism was limited. The
prevalence of autism in that period of time was quoted as being a
The relationship between epilepsy and autism is the subject of few per 10,000 [3]. In recent years, the importance, in terms of prac-
ongoing debate [1]. Questions that have arisen in this debate are tical management, of diagnosing milder autistic features has become
not only tantalizing from the scientific point of view but are also increasingly recognized, with the result that the diagnostic criteria
challenging from the practical point of view because the answers are now broader and prevalence of “autism spectrum disorder”
might have a profound effect on the management of patients. If the (ASD) has recently been reported as being greater than 1% [4].
risk factors for an individual with autism developing epilepsy are Using the previous narrow diagnostic criteria of severe or “Kanner”
known, steps can be taken to ensure that the epilepsy is recognized autism led to a much higher associated prevalence of comorbidity
promptly and managed properly. In some cases, it might even be such as intellectual impairment and epilepsy. The broader diagnostic
possible to prevent the epilepsy from developing. For example, criteria of “autism spectrum disorder”, which now include high-
knowing that there is a high risk of an association between certain functioning autism/Asperger syndrome [5], imply that the rates of
metabolic disorders and autism could lead to early recognition and intellectual disability and epilepsy will be lower. Most papers do
treatment of such disorders, which might, in some cases, prevent not take account of this changing situation, quoting the results of
the development of epilepsy. older studies with narrower diagnostic criteria alongside the results
Articles in this special section address several of the most funda- of more recent studies with the broader diagnostic criteria, as if the
mental questions in the science and treatment of individuals with comorbidity rate would be the same. There is, consequently, consid-
autism and epilepsy. These questions have persisted since the initial erable uncertainty about the answers to such simple questions as:
description of autism, and the answers are continuously evolving, “what is the rate of epilepsy in people with autism spectrum disor-
along with the definition of autism itself. In the years following der?” Papers on the relationships between autism and epilepsy
Kanner's [2] first description of autism in 1943, the diagnostic need to be read not only with a critical eye on the science but also
with an understanding of the diagnostic criteria that were used at
that time and in that particular study.
Some of the specific questions addressed by contributors to this spe-
⁎Corresponding author: FBesag@aol.com.
cial section follow.

http://dx.doi.org/10.1016/j.yebeh.2015.05.032
1525-5050/© 2015 Published by Elsevier Inc.
144 F.M.C. Besag / Epilepsy & Behavior 47 (2015) 143–146

Does epilepsy cause autism? notably intellectual impairment, which is also associated with epilepsy.
However, this coexistence of autism and epilepsy is association, not
This might be considered the most fundamental question regarding causation.
the relationship between autism and epilepsy. It is acknowledged that,
in rare cases, the epilepsy itself might be responsible for autistic fea-
Are there underlying brain mechanisms that predispose to both
tures. For example, in the Landau–Kleffner syndrome of acquired epi-
conditions?
leptic aphasia [6], the marked deterioration in social communication
has been described as being autistic [7], but it appears to be the direct
The answer to this question, in many cases, is undoubtedly yes. Sev-
result of the epilepsy. If the epilepsy is treated promptly and effectively,
eral metabolic and mitochondrial disorders underlie both conditions
the “autistic features” might improve or even resolve in at least some
(Frye, this issue). Many genetic defects have also been identified as
cases. The Landau–Kleffner syndrome and other syndromes in which
causing both conditions as further described in several papers in this
the epilepsy itself seems to be directly responsible for the autistic
special section. The paper by Frye raises important questions with re-
features are rare. The debate does not stop at this point, however.
gard to the importance of mitochondrial dysfunction. He points to a re-
The aim of several studies has been to determine whether epileptiform
cent meta-analysis that established that 5% of children with ASD meet
abnormalities, not necessarily presenting as obvious seizures, nor
the criteria for mitochondrial disease but as many as 30% may manifest
presenting with syndromes such as the Landau–Kleffner syndrome,
mitochondrial dysfunction. His review has also drawn attention to the
might be responsible for the “autistic regression” that occurs in approx-
finding that treatments for mitochondrial disease reportedly improve
imately a third of children with (narrowly defined) autism under
ASD symptoms. These findings have implications for how individuals
36 months of age [8]. The results of these studies are conflicting and
with ASD might be both investigated and treated. This is another area
inconclusive. These topics are covered in more detail in several papers
of autism research that requires much more attention so that specific
in this special section.
guidelines for investigation and treatment can be provided.
An area of particular interest has been the association of epileptiform
discharges and autism. Several studies have shown that such discharges
are far more frequent in people with autism [9–11], even if they do not What is the role of genetics in determining conditions that underlie
have a diagnosis of epilepsy, than in the general population [12,13]. The both autism and epilepsy?
number of interictal epileptiform discharges recorded is greater if
prolonged EEGs, including overnight recordings, are performed. It is in- The remarkable advances in genetics over the recent years have con-
teresting to note the frequent occurrence of nocturnal epileptiform dis- firmed that many genetically determined disorders increase risk of both
charges and autism alongside the very high rate of sleep disturbance autism and epilepsy. Much attention has rightly been given to obvious
(estimated as 40% to 80%) in this condition, as emphasized in Accardo examples such as tuberous sclerosis, which results from defects in the
and Malow (this issue). hamartin gene on chromosome 9 or the tuberin gene on chromosome
If epileptiform discharges are indeed higher in individuals with 16, causing defects in cell growth regulation. The manifestations of tu-
autism, are such discharges associated with autistic regression? berous sclerosis affect many body systems, but it is the cerebral
Some studies have claimed to have found such an association [14, malformations (“tubers”) that are associated with both epilepsy and,
15], and others have not [9,11,16]. However, the results need to be particularly if they occur in temporal lobes in West syndrome, autism.
scrutinized carefully. For example, Baird et al. [17] carried out sleep Not only do these findings raise more questions with regard to the na-
EEGs on 64 children with autism, none of whom had a diagnosis ture of autism, they also open another avenue of treatment. Drugs
of epilepsy. No statistically significant difference in epileptiform such as sirolimus (rapamycin) and everolimus that limit the cell growth
discharges was found between 39 of these children who had might actually have a role to play in the prevention of both the epilepsy
regressed and the remainder who had not regressed. However, and the autism if administered early; however, much more research is
there was a trend towards greater numbers of epileptiform dis- required before clinical recommendations can be made in this regard.
charges in those who had regressed, suggesting that if larger num- Many of the metabolic conditions referred to in the previous section
bers had been examined, the result might have been statistically are the result of identified genetic disorders. In addition, several genetic
significant. There is another major methodological issue that could disorders have been identified that cause intellectual impairment and
affect the outcome of such studies. If frequent epileptiform dis- are associated with both autism and epilepsy. These are referred to in
charges cause regression, such as appears to occur, for example, in several papers in this issue, notably in papers 2, 6, and 8.
the cases of Landau–Kleffner syndrome and West syndrome, and if As pointed out by Blackmon (this issue), advances in chromosomal
the EEG investigations are carried out too late, the epileptiform dis- microarray analysis (CMA) have made the detection of small chromo-
charges may no longer be present and, consequently, will not be de- somal variations possible. This has changed the role of chromosomal
tected. This provides a strong argument for carrying out prospective testing fundamentally. In the past, clinicians would examine a patient,
studies on large populations, ensuring that comprehensive EEG in- perhaps with dysmorphic features, autism, epilepsy, and intellectual
vestigations, including at least a sleep EEG and preferably an over- impairment, and express the view that it was highly likely that a chro-
night EEG, are carried out on any child who regresses, at the time mosomal abnormality was responsible, only to find that the investiga-
of the regression, not at some later time when the damage may tion techniques at that time revealed no such abnormality. It could be
have been done and the epileptiform discharges may have resolved said that the reverse is true now. Chromosomal microarray analysis
(“burnt-out cases”). Such studies would not be easy to perform but often yields unsuspected abnormalities, the significance of many of
might provide the best way of establishing the relevance of epilepti- which is currently difficult to interpret. Several of the genetic/chromo-
form discharges to the regression that occurs under three years of somal abnormalities that have been associated with both ASD and epi-
age in about a third of children with autism. lepsy have been listed in two of the papers in this special issue and
are covered in previous reviews [18]. These include the copy number
Does autism cause epilepsy? variants (CNVs) listed by Blackmon: 1q21 deletions, 7q11.23 duplica-
tions, 15q11.1-q13.3 duplications, 16p11.2 deletions, 7q11.23 duplica-
There does not appear to be any credible mechanism for suggesting tions, 15q11.1-q13.3 duplications, 16p11.2 deletions, 18q12.1
that autism causes epilepsy. Some have argued that autism could be said duplications, and 22q11.2 deletions. A number of the genetic abnormal-
to cause epilepsy on the grounds that, in a large proportion of individ- ities associated with ASD and epilepsy are listed by Paciorkowski et al.
uals, autism is associated with other comorbid brain dysfunction, (this issue). Blackmon (this issue) has reviewed the genetic/
 F.M.C. Besag / Epilepsy & Behavior 47 (2015) 143–146 145

chromosomal abnormalities associated with brain malformations that In a totally different area, Frye (this issue) has made a strong argu-
can lead both to ASD and epilepsy. ment for treating underlying metabolic disorders that lead to both
Two specific challenges arise from these studies. First, how inten- ASD and epilepsy, quoting studies that suggest that some of the autistic
sively should an individual be investigated for chromosomal/genetic ab- features may be improved as a result of the appropriate treatment. Of
normalities; second, when such abnormalities are detected, how should course, it could be argued that identified underlying metabolic disorders
they be interpreted in terms of prognosis, treatment, and genetic would be treated in any case; this is certainly true, where treatments are
counseling? available. However, the growing recognition that metabolic disorders
can lead to some cases of autism (with or without epilepsy) makes us
What is the importance of prenatal, perinatal, and postnatal envi- ask the question: how extensively should children with autism be in-
ronmental factors? vestigated for metabolic disorders? Again, this is a subject that remains
open to debate.
Some interesting research on prenatal environmental factors has
emerged over recent years but also remains subject to considerable de- Is the prognosis of either autism or epilepsy different when the oth-
bate. The topic of environmental factors and autism has recently been er condition is also present?
reviewed by Schmidt et al. [19]. They concluded that the evidence for
a protective effect of maternal folate during pregnancy was strong Strictly speaking, autism is a pervasive developmental disorder, im-
[20] and that the evidence for a deleterious effect of air pollution was plying that it is constant over time. Although it is pervasive, the way in
consistent. The evidence for other environmental factors, such as organ- which it affects quality of life is influenced, to a major extent, by addi-
ophosphate pesticides and phthalate exposure, was less convincing. tional factors, such as intellectual ability, level of communication, envi-
They also emphasized that gene–environment interaction should be ronment, reactions to the individual by other people (including, where
considered. appropriate, behavioral management), and in some cases, medication,
Intrauterine infections such as rubella have long been associated such as risperidone to decrease anxiety or stimulant medication to
with intellectual impairment, autism and, by association, epilepsy [21]. treat comorbid ADHD. One of these factors, often a key factor, is how
Medication taken during pregnancy can affect brain development. Par- well the epilepsy is managed.
ticular interest has been shown in the effect of valproate on the fetus. In contrast to what might be referred to as “true autism”, the autistic
There is now good evidence indicating that the fetal valproate syn- features that arise, as previously discussed, in association with frequent
drome is associated not only with cognitive impairment but also with epileptiform discharges in some epilepsy syndromes, such as the
autism [22,23]. The precise role of drugs taken during pregnancy lead- Landau-Kleffner syndrome, can decrease or even resolve with early ef-
ing to the combination of autism and epilepsy is another area of re- fective treatment. In these specific cases, the prognosis of the autistic
search that requires further attention. features is profoundly affected by the presence of the epilepsy and, in-
Brain damage arising during delivery and through factors such as in- deed, by how effectively the epilepsy is treated with medication or
adequately treated severe neonatal jaundice or untreated hypothyroid- surgery.
ism/metabolic disorders should be avoidable by providing good A major factor affecting the prognosis of the epilepsy and the out-
obstetric and neonatal care, including appropriate screening. For exam- come of the autism, in practical terms, is intellectual impairment. The
ple, untreated phenylketonuria (PKU) can lead to intellectual impair- greater the degree of intellectual impairment, the more likely epilepsy
ment, ASD, and epilepsy, but neonatal screening for this inborn error is to occur and the more likely it is to be resistant to treatment. Similarly,
of metabolism and implementation of the appropriate low phenylala- the greater the degree of intellectual impairment, the greater the diffi-
nine diet has almost entirely eliminated these problems from babies culties in managing the autism are likely to be.
born with PKU in developed countries. In addition, because some of the manifestations of autism may be
mistaken for signs of complex partial (dyscognitive) seizures, difficul-
Do any of the proposed relationships between autism and epilepsy ties may arise in diagnosing the epilepsy, and these could result in de-
provide insight into useful management or treatment? lays in treatment or inappropriate treatment.
In individuals with more severe autism, difficulties may also arise in
The case for investigating the child with unexplained loss of skills obtaining a satisfactory history, particularly with regard to diagnosing
has been made strongly in several previous publications [1,24]. A simple partial seizures/auras. In some of these individuals who do not
small number of these children will have conditions such as the Lan- like change in routine, reliable administration of antiepileptic medica-
dau–Kleffner syndrome or continuous spike waves in slow-wave sleep tion may prove to be difficult.
(CSWS) syndrome. Some will have other conditions associated with Over recent years, there has been an increasing trend to consider at
gross epileptiform abnormality, such as the hypsarrhythmia of West least some of the abnormalities in autism as being the result of different
syndrome. In all these cases, it has been argued that early effective treat- connectivity. The SPECT findings reported in Sasaki (this issue) indicate
ment can, in at least some children, allow skills to return and might even that there is hypoperfusion particularly in the temporal and frontal
reverse some of the autistic features. This raises the important question: lobes. How this reduced cerebral blood flow and, by implication, re-
should any child who loses skills for no apparent reason undergo EEG duced connectivity might influence the manifestation, treatment, and
investigations, including an overnight EEG if a routine EEG does not re- prognosis of epilepsy in people with autism remains to be determined.
veal the cause, even if there is no other evidence of epilepsy? It should
be noted, in this context, that approximately one-third of those with What is the role of additional comorbidities, such as intellectual dis-
the Landau–Kleffner syndrome do not present with obvious seizures. ability or attention deficit hyperactivity disorder, in the relationship
Because loss of skills can have such profound consequences, the current between the two conditions and in influencing treatment choices?
author would argue that the answer to this question is, emphatically,
yes. Even if a small proportion of these children lose skills because of Reference to this issue has been made in the previous section. It has
CSWS or other major epileptiform abnormalities, for those children, been argued that, in the small number of cases in which epilepsy ap-
treatment might prevent further deterioration or even allow them to re- pears clearly to be the cause of the autistic features and the loss of intel-
gain some skills. Although, as indicated in an earlier section, there is de- lectual ability, prompt effective treatment of the epilepsy might reverse
bate about the role of epileptiform abnormalities in autistic regression, both the cognitive loss and the autistic features.
some of the apparently negative results may have been because of If the diagnosis of ASD, epilepsy, or intellectual impairment leads to
methodological issues. investigation and identification of a metabolic disorder, this could affect
146 F.M.C. Besag / Epilepsy & Behavior 47 (2015) 143–146

management and outcome, as has also been indicated in a previous [3] Wing L, Yeates SR, Brierley LM, Gould J. The prevalence of early childhood autism:
comparison of administrative and epidemiological studies. Psychol Med 1976;6:
section. 89–100.
Attention deficit hyperactivity disorder is particularly common in [4] Baron-Cohen S, Scott FJ, Allison C, Williams J, Bolton P, Matthews FE, et al. Prevalence
people with ASD and is also very common in people with epilepsy. of autism-spectrum conditions: UK school-based population study. Br J Psychiatry
2009;194:500–9.
However, there may be a reluctance to treat ADHD in the presence of [5] Association AP. Diagnostic and statistical manual of mental disorders. (DSM-5®):
epilepsy because of the mistaken idea that ADHD treatment exacerbates American Psychiatric Pub; 2013.
seizures, despite considerable evidence of the contrary. However, it is [6] Landau WM, Kleffner FR. Syndrome of acquired aphasia with convulsive disorder in
children. Neurology (Minneap) 1957;7:523–30.
important to ensure that the ADHD is not the result of the epilepsy itself [7] Mantovani JF. Autistic regression and Landau–Kleffner syndrome: progress or confu-
or inappropriate treatment of the epilepsy with antiepileptic drugs that sion? Dev Med Child Neurol 2000;42:349–53.
can cause features of ADHD, such as benzodiazepines, phenobarbital, or [8] Baird G, Charman T, Pickles A, Chandler S, Loucas T, Meldrum D, et al. Regression, de-
velopmental trajectory and associated problems in disorders in the autism spec-
vigabatrin.
trum: the SNAP study. J Autism Dev Disord 2008;38:1827–36.
[9] Chez MG, Chang M, Krasne V, Coughlan C, Kominsky M, Schwartz A. Frequency of
epileptiform EEG abnormalities in a sequential screening of autistic patients with
What research is recommended for the future? no known clinical epilepsy from 1996 to 2005. Epilepsy Behav 2006;8:267–71.
[10] McVicar KA, Ballaban-Gil K, Rapin I, Moshe SL, Shinnar S. Epileptiform EEG abnor-
Despite the enormous amount of research into autism, many unan- malities in children with language regression. Neurology 2005;65:129–31.
[11] Rossi PG, Parmeggiani A, Bach V, Santucci M, Visconti P. EEG features and epilepsy in
swered questions remain, including a number relating to the relation- patients with autism [see comments]. Brain Dev 1995;17:169–74.
ship between autism and epilepsy. However, the research so far has [12] Cavazzuti GB, Cappella L, Nalin A. Longitudinal study of epileptiform EEG patterns in
enabled us to make some recommendations for future focused studies. normal children. Epilepsia 1980;21:43–55.
[13] Capdevila OS, Dayyat E, Kheirandish-Gozal L, Gozal D. Prevalence of epileptiform ac-
During the initial phase of research, there was an important role for tivity in healthy children during sleep. Sleep Med 2008;9:303–9.
case reports and clinic-based studies. However, that phase is now in de- [14] Lewine JD, Andrews R, Chez M, Patil AA, Devinsky O, Smith M, et al.
cline. The current pressing need is for carefully conducted, prospective, Magnetoencephalographic patterns of epileptiform activity in children with regres-
sive autism spectrum disorders [see comments]. Pediatrics 1999;104:405–18.
population-based studies, particularly those that investigate the child as [15] Shinnar S, Rapin I, Arnold S, Tuchman RF, Shulman L, Ballaban-Gil K, et al. Language
soon as any regression is detected. These investigations should include regression in childhood. Pediatr Neurol 2001;24:183–9.
chromosomal studies, a search for metabolic/mitochondrial disorders, [16] Tuchman RF, Rapin I. Regression in pervasive developmental disorders: seizures and
epileptiform electroencephalogram correlates. Pediatrics 1997;99:560–6.
and EEGs, including at least a sleep EEG and preferably an overnight
[17] Baird G, Robinson RO, Boyd S, Charman T. Sleep electroencephalograms in young
EEG if no other cause for the regression has been found. The results children with autism with and without regression. Dev Med Child Neurol 2006;
should be related to possible environmental factors, including prenatal 48:604–8.
factors to which the mother has been exposed, such as pollutants, [18] Jeste SS, Geschwind DH. Disentangling the heterogeneity of autism spectrum disor-
der through genetic findings. Nat Rev Neurol 2014;10:74–81.
toxic chemicals, alcohol, and drugs, both illicit and therapeutic. Obstet- [19] Schmidt RJ, Lyall K, Hertz-Picciotto I. Environment and autism: current state of the
ric and neonatal factors should form an essential part of the data so that science. Cut Edge Psychiatry Pract 2014;4:21–38.
the importance of these can be determined. Because stratification into [20] Schmidt RJ, Tancredi DJ, Ozonoff S, Hansen RL, Hartiala J, Allayee H, et al. Maternal
periconceptional folic acid intake and risk of autism spectrum disorders and devel-
the various subcategories of the data is likely to result in small numbers, opmental delay in the CHARGE (CHildhood Autism Risks from Genetics and Envi-
there is a strong case for international research collaboration with ronment) case–control study. Am J Clin Nutr 2012;96:80–9.
agreed and standardized protocols. The hope is that these focused re- [21] Chess S. Follow-up report on autism in congenital rubella. J Autism Child Schizophr
1977;7:69–81.
search studies will provide answers that significantly improve the fu- [22] Williams G, King J, Cunningham M, Stephan M, Kerr B, Hersh JH. Fetal valproate syn-
ture outlook for people with autism and epilepsy. drome and autism: additional evidence of an association. Dev Med Child Neurol
2001;43:202–6.
[23] Christensen J, Gronborg TK, Sorensen MJ, Schendel D, Parner ET, Pedersen LH, et al.
References Prenatal valproate exposure and risk of autism spectrum disorders and childhood
autism. JAMA 2013;309:1696–703.
[1] Besag FM. The relationship between epilepsy and autism: a continuing debate. Acta [24] Sinclair DB, Snyder TJ. Corticosteroids for the treatment of Landau–Kleffner syn-
Paediatr 2009;98:618–20. drome and continuous spike-wave discharge during sleep. Pediatr Neurol 2005;
[2] Kanner L. Autistic disturbances of affective contact. Nerv Child 1943;2:217–50. 32:300–6.