CHAPTER 15 Autism Spectrum Disorders
Before the 1990s, autism was thought to be a rare disorder with a dismal prognosis whose victims rarely achieved independent living and enduring relationships. Instead, most affected adults lived with their parents or in state institutions.1 Most individuals in whom autism as diagnosed were nonverbal and assumed to have some degree of mental retardation even when standardized measurements of intelligence were not available. Treatment programs, if existent, were usually housed in facilities serving segregated populations. There were no published autism guidelines, and there was little interest in research outside a relatively small circle of dedicated investigators. The media, lay public, and members of Congress were generally not even aware of the term autism, much less concerned about it.
The 1990 decade was proclaimed “the decade of the brain,”2 partly because of advances in neuroimaging and expanding knowledge about how the central nervous system worked. However, the 1990s might also be called the “the decade of autism” because of the rapidly expanding body of knowledge in the field. The 1990s also marked a period of significantly heightened public awareness, attributable at least in part to the 1988 release of the Academy Award–wining movie “Rainman.” Although some advocates expressed concern about how autism was portrayed, the movie nevertheless brought autism to public awareness. Public attention has remained high because autism has been and continues to be engulfed in a sea of controversy. Concerns about a possible “epidemic” relating to vaccines and toxins and the media’s frenzy regarding miraculous cures (auditory integration therapy, facilitated communication, secretin injections, and mercury chelation) have made autism a household term. Furthermore the media, motivated by dedicated advocates, has played an important role in capturing the attention of Congress, which in turn has resulted in autism-specific activities mandated by national legislation: the Children’s Health Act of 2000,3 the New Freedom Initiative of 2001,4 and the Combating Autism Act of 2005.5
Because of mandated federal support, autism “centers of excellence” emerged and contributed to a rapidly expanding body of knowledge. Thus, interest in autism within professional circles paralleled that of the lay public; this attention is illustrated by the exponential growth of training activities and published articles during the 1990s. Whereas before 2000 professional organizations such as the American Academy of Pediatrics (AAP) offered no stand-alone course in autism at its national conferences, since then autism has been at the top of the AAP list of “hot topics” and now consistently appears as a topic on conference agendas. Similarly, approximately 3000 autism-related articles were published in scientific peer-reviewed journals between 1943 (when it was first described by Kanner6) and 1990, whereas more than 4000 appeared in the 1990s alone.7 Beginning at the end of the 20th century, the first policy statements and practice guidelines were published.8–13 In spite of this, the cause of autism is still not known, and there is no known cure.
It is impossible to discuss the voluminous literature surrounding the autistic spectrum disorders (ASDs) in a single chapter. The ambitious reader is referred to Handbook of Autism and Pervasive Developmental Disorders (Volumes 1 and 2),14 Neurobiology of Autism,15 Autism Spectrum Disorders in Children,16 and Autism Spectrum Disorders.17
TERMINOLOGY
Terminology, definitions, and diagnostic criteria have changed over the years. The concept of “autism” before the 1990s apparently represented only a small proportion of ASDs. Although frequently used by European investigators in the 1990s, the term autistic spectrum disorder did not become popular in the United States until about 2000.18 It has become an “umbrella term” that includes three of the five pervasive developmental disorders (PDDs) listed in the most recent revisions of The Diagnostic and Statistic Manual of Mental Disorders (DSM) of the American Psychiatric Association19,20 and the Diagnostic and Statistical Manual for Primary Care, Child and Adolescent Version21: autistic disorder, Asperger disorder (referred to as Asperger syndrome in this chapter), and pervasive developmental disorder, not otherwise specified (PDD-NOS). The remaining two PDDs, Rett syndrome and childhood disintegrative disorder, are not discussed in this chapter.
The ASDs are neurodevelopmental conditions characterized by one or a combination of the following: significant social skill deficits, both qualitative and quantitative language abnormalities, restricted interests, and repetitive motor mannerisms. Although ASDs appear to have a strong genetic basis,22,23 the precise cause is unknown; thus, there is no pathognomonic physical sign or laboratory test. Instead, the diagnosis is made by determining the presence of characteristic developmental and behavioral criteria described in the fourth edition of the DSM (DSM-IV)19 or in the text revision (DSM-IV-TR).20 Many clinicians use a standardized evaluation tool that operationalizes the DSM criteria. Nevertheless, considerable subjectivity still exists in making this diagnosis, largely because of the wide range of symptoms included within the scope of the spectrum.
Autism Disorder
Autism was first described in the third edition of the DSM (DSM-III) in 198024 as “Infantile Autism”; the current term, “Autistic Disorder” replaced “Infantile Autism” in the revised version of the DSM-III (DSM-III-R) in 1987.25 Although clinical patterns vary in regard to severity, age at onset, underlying cognitive deficits, and other features, diagnosis of “classic” autistic disorder is dependent on the presence of at least half (six) of the criteria (Table 15-1).20 Symptoms in at least one of these areas must have been present before the age of 3.
TABLE 15-1 Diagnostic Criteria for 299.00 Autism Disorder
Rights were not granted to include this table in electronic media. Please refer to the printed book.
Reprinted with permission from the Diagnostic and Statistical Manual of Mental Disorders, 4th ed, Text Revision. Washington, DC: American Psychiatric Association, 2000, p 75.
Asperger Syndrome
Asperger syndrome is characterized by the same impairment in social interaction and restricted interests as in autistic disorder; however, language skills are relatively normal (defined as use of single words by age 2 years and phrases by 3 years) (Table 15-2).20 Later language is characterized by pragmatic deficits (problems with the social use of language). In addition, cognitive and adaptive skills are normal. Depending on the child’s age, it is sometimes quite challenging to distinguish between children with Asperger syndrome and children with autistic disorder and normal intelligence. Because of this, controversy exists as to whether Asperger syndrome represents a high-functioning form of autism or a separate entity.26,27 Children with Asperger syndrome are usually not recognized until after 4 years of age, when social interactions with peers in preschool settings become a concern.
TABLE 15-2 Diagnostic Criteria for 299.80 Asperger Disorder*
Rights were not granted to include this table in electronic media. Please refer to the printed book.
Reprinted with permission from the Diagnostic and Statistical Manual of Mental Disorders, 4th ed, Text Revision. Washington, DC: American Psychiatric Association, 2000, p 75.
Pervasive Developmental Disorder, Not Otherwise Specified
Pervasive Developmental Disorder, Not Otherwise Specified is a subthreshold term that is used when a child demonstrates some but not all of the criteria necessary to make a diagnosis of one of the specific PDDs. Unfortunately, there was an error in the DSM-IV text19: It stated that PDD-NOS should be used when there was an “impairment in… social interaction” or in verbal and nonverbal skills. This allowed clinicians to apply the PDD-NOS label in the absence of social skill deficits, thus broadening the definition of PDD-NOS and causing loss of specificity. This error was corrected in the DSM-IV-TR.20 The PDD-NOS label is reserved for persons who demonstrate “severe and pervasive impairment in the development of reciprocal social interaction” and either communication deficits or restricted interests/repetitive behaviors. Confusion still exists regarding the actual number of criteria that are necessary to apply the PPD-NOS label; by convention, at least two but not more than five should be present. PDD-NOS also includes “atypical autism,” which refers to persons with at least one feature that is dissonant with traditional autism, such as later onset or absence of stereotypies.28,29
Throughout this chapter, ASD refers to all three disorders as a group. When a specific ASD is discussed, the appropriate term is used. Autism is used in reference to older literature published before the concept of a spectrum of autistic disorders emerged. A broad spectrum does appear to exist, although its external and internal boundaries are hazy.30,31 Family studies have shown that the entire spectrum may be expressed in the same pedigree. Sometimes the term broader autism phenotype is used for individuals with isolated social deficits, particularly in the context of extended-family relatives of probands with autism.32,33
HISTORY
In 1943, Leo Kanner, a psychiatrist at the Johns Hopkins University School of Medicine, first defined autism as it is known today.6 About the same time, Hans Asperger, an Austrian pediatrician, unaware of Kanner’s work, published an article in German34 describing four children who demonstrated symptoms similar to those described by Kanner with the exception of better verbal and cognitive skills. Asperger syndrome escaped recognition until it was popularized by Wing’s translation into English.35,36 In 1978, Rutter37 published the first set of “essential criteria” for autism. These were incorporated into the next edition of the DSM (DSM-III),24 and autism became recognized as a separate entity within the newly created category of Pervasive Developmental Disorders.
New criteria were developed for the DSM-III-R,25 published in 1987, and were criticized for being too inclusive and thus promoting overidentification of autistic disorder.38 The criteria still in use today were published in 1994 in the DSM-IV19; Asperger syndrome criteria appeared for the first time in this version. The DSM-IV Autistic Disorder criteria were the result of years of analyses to reduce the overinclusiveness of DSM-III-R. Furthermore, collaboration with European groups working on the manual for the revised International Classification of Diseases, 10th edition (ICD-10),39 promoted conformity between the two classification systems. Studies have revealed that the DSM-IV criteria have better specificity (0.87) than do DSM-III-R criteria.40 Criteria for autistic disorder and Asperger syndrome have not changed in the DSM-IV-TR.20
Although Kanner initially hypothesized that autism was an inborn, biological condition,41 misconceptions based on psychodynamic theory soon became prevalent. Probably the most important one was the mistaken concept that autism might be caused by cold and unnurturing parents (“the refrigerator theory”). Bettelheim42 promoted this concept in his book, The Empty Fortress: Infantile Autism and the Birth of Self. The refrigerator theory remained popular until the 1960s when Rimland hypothesized a neurological cause.43 This hypothesis was later supported by the demonstration of neuropathological abnormalities on magnetic resonance imaging (MRI)44 and documentation of a high rate of coexisting seizures.45
The science of ASD has advanced a great deal. Collaborative research centers and multidisciplinary diagnostic teams proliferated during the late 1990s and continue to do so with even greater momentum in the new millennium. Since the 1990s, there have been a number of rapid developments: the debut of the first screening tools, the development of evaluation tools that operationalize DSM-IV criteria, neuropathic studies that revealed an early prenatal onset, identification of multiple genetic susceptibility genes, recognition of the relative importance of social skill deficits in defining ASDs, and evidence that early and appropriate intervention is effective in improving outcomes.7,13
PREVALENCE
The apparent dramatic rise in prevalence of ASDs has become a focus for parent advocacy groups and the media and may well be one of the most controversial topics in the field of autism (Table 15-3). More than 30 studies documented an apparent increase in prevalence of these diagnoses.46–55c In 2000, the Centers for Disease Control and Prevention organized the Autism and Developmental Disabilities Monitoring (ADDM) Network, a multisite, records-based surveillance program, to study the prevalence of ASDs. The ADDM Network employs systematic screening of developmental evaluation records for autistic behaviors rather than depending on a medical or educational diagnostic label of an ASD. In 2007, the ADDM Network reported ASD rates ranging from 1 in 303 to 1 in 94 8-year-old children for two time periods (2000, 2002) in a total of 14 sites in the United States; the average rate was 1 in 150 or 6.6 per 1000 8-year-olds.56 Studies varied in methods, definition, and case ascertainment strategies, but overall there appeared to have been up to a 10-fold increase worldwide since the 1950s.
Several factors complicate the interpretation these data, making it very difficult to discern whether there has been a true rise in prevalence or simply an apparent one. Most investigators have demonstrated that the apparent rise in prevalence is attributable, at least in part, to changing broader criteria and increased public and professional awareness.51,54,57 This is supported by a greater increase in the numbers of milder cases (i.e., PDD-NOS and Asperger syndrome). Other factors contributing to the apparent rise include the emergence of screening tools in the 1990s, which resulted in improved ascertainment, and the development of better diagnostic tools that can more reliably identify children at younger ages. The media (e.g., the National Broadcasting Company’s Autism Speaks Campaign, April 2006) and advocacy groups have been successful in raising public awareness so that parents are now recognizing ASD symptoms in their children and voicing their concerns earlier to physicians. There has also been an increase in recognition of ASDs among children who have disorders unrelated to ASD, such as Down syndrome58 and the syndrome of coloboma, heart anomaly, choanal atresia, retardation, genital lesions, and ear abnormalities (CHARGE association).59 Autistic features can also be detected in some children with congenital sensory disorders, especially when severe vision and/or hearing deficits are not detected and intervention is not implemented early.60,61
Finally, public policies have also made a significant effect on reported prevalence rates obtained from public administrative data. The Education of All Handicapped Children Act (Public Law 94-142) of 197562 made schools accessible to children with some disabilities. Many children with more severe disabilities, such as those with both autism and severe mental retardation, continued to live in segregated state institutions. Later laws such as the Americans with Disabilities Act of 199063 promoted closure of institutions; which caused more children with disabilities to live at home and attend community schools. Autism first became a diagnostic category for which children could receive services with passage of the Individuals with Disabilities Education Act (IDEA) in 1991.64 Before 1991, children with autism were most likely to be served under categories established by older laws, such as “mental retardation,” “learning disabled,” “speech delayed,” “or emotionally disturbed.”65 Since the passage of IDEA,64 school eligibility diagnoses have usually conformed to medical diagnoses of an ASD. This phenomenon of “diagnostic substitution” may account for a substantial proportion of the apparent rise in prevalence; however, educational administrative data may not be completely reliable.66–71 IDEA amendments72,73 have made supplementary services available to children with ASD that are not available to children with other disabilities, such as “year-around school.” When criteria for ASD are marginal, professionals may be tempted to apply the label in order to secure these supplementary services, which often also provide additional support for the parents. Such strategies tend to inflate the “prevalence” when values are obtained solely from educational sources.
The reasons for the reported 10-fold rise in prevalence of ASD remain controversial.50,51,74–82 However, there is broad agreement that more boys than girls are consistently found to be affected with ASD; sex ratios range from 2 : 1 to 4 : 1.50,53,79,83–86 The male : female ratio is even higher for high-functioning autism and Asperger syndrome, ranging from 6 : 1 to 15 : 1.87 A 2006 study, in which most affected children (53.3%) had normal intelligence, demonstrated a male : female ratio of 9 : 1.55
ETIOLOGY
ASDs are now believed to be biologically based neurodevelopmental disorders that are highly heritable.88 Because of the wide phenotypic spectrum, most experts believe that many genes are involved.89 In a minority of cases, ASDs may be associated with a medical condition or a known syndrome characterized by dysmorphic features and some degree of comorbid mental retardation.47,48 Although ASDs are believed to be mainly genetic in origin, the lack of 100% concordance in monozygotic twins indicates that environmental factors may modulate the phenotypic expression.22,88 Thus it has become increasingly apparent that the cause is multifactorial, with a variety of genetic and, to a lesser extent, environmental factors playing a role.
Genetic Underpinnings
Studies of twins have revealed concordance rates of classic autism in 60% of monozygotic pairs and in 0% to 3% of dizygotic pairs.22,88,90 When the broader phenotype was taken into consideration, the rates were 60% to 92% and 0% to 10%, respectively. In addition, family studies have demonstrated a rapid decrease in prevalence among first, second, and third-degree relatives. Using these data, Bailey and colleagues22 calculated that the predisposition for autism was more than 90% heritable, with multiple interacting genetic influences and strong family clustering.91 In spite of these strong genetic underpinnings, the exact cause or causes are still unknown. The task has been daunting because of genetic complexity and phenotypic variation. First, ASDs appear to be complex heritable disorders involving multiple genes; estimates based on family studies range from 5 to 20 genes.89 Each gene or gene combination may result in somewhat different subtype but often with overlapping behavioral phenotypes. The number of genes contributing to the disorder and the relative prevalence of each will increase or decrease the probability of identifying the cause; that is, success is more likely if there are relatively few genes that are somewhat common than if there are many genes that are rarer.91 A second factor making gene identification more challenging is the variability in the ASD phenotype. The wide spectrum of symptoms sometimes promotes inclusion of participants in a study with various ASDs, sometimes even including individuals with disorders falsely categorized to be in the spectrum. This imprecise diagnosis contaminates the study group and makes identification of a unifying etiological agent elusive.92,93
Two major strategies have been used in the search for the ASD susceptibility genes: targeted cytogenetic studies and whole genome screens of families of children with ASD.91,94 The first strategy depends on developing a hypothesis regarding the pathogenesis of ASD, focusing on one or more potential candidate genes and testing them genetically for an association with ASD. Candidate genes in ASD include, among others, those that appear to play a role in brain development (e.g., cerebellar Purkinje cell proliferation)15,95 or in neurotransmitter function (e.g., serotonin transmitter).96 The second strategy entails an indirect method and does not require investigators to make assumptions regarding the mechanism of inheritance. Instead, families with multiple members demonstrating an ASD (multiplex families) are studied to identify recurring DNA markers (breakpoints, translocations, duplications, and deletions) present in affected, but not in unaffected, members. Unfortunately, progress has been limited because the phenotypic endpoints of ASD are not well defined. Changes in DSM criteria and inconsistency in ascertainment strategies, resulting in a hazy delineation between “affected” and “unaffected” family members, contaminate outcomes and challenge interpretation of results. This phenotypic heterogeneity has challenged molecular searches for the ASD gene or genes in spite of several genomewide screens (International Molecular Genetic Study of Autism Consortium [IMGSAC]) and multicenter collaborative efforts since the 1980s.93,97–99 The results are very enlightening; however, rather than providing conclusions based on replicated findings from multiple labs, these studies have often produced confusion and uncertainty as more susceptibility loci are described. Although at least one autism-linked abnormality has been found on almost every chromosome, few sites have been identified with any frequency. Table 15-4 provides a summary of some of the more consistent findings; however, the reader is advised to consult more extensive reviews.7,92,100–104 Large study samples with pooled data from multiple populations (maximizing homogeneous samples) are needed to confirm the validity of reported candidate genes and susceptibility sites in ASD.93
TABLE 15-4 Selected Autism Spectrum Disorder (ASD) Genetic Markers
|
The high male : female ratio and the discovery of genes for both fragile X syndrome and Rett syndrome on the X-chromosome made it a plausible target.116,362,553 Investigators have targeted a variety of possible roles for the X-chromosome in ASD:
Skewed X-chromosome inactivation is known to occur in X-linked mental retardation carriers554 and to be responsible for most of the phenotypic variability seen in Rett syndrome.362 X-chromosome inactivation patterns in female patients with ASD and controls (from the AGRE database) were studied to determine whether skewness might account for expression of possible autism genes on the X-chromosome.553 Indeed, statistically greater skewness was found in those with classic Autism disorder than in controls (33% vs. 11%). Furthermore, of 10 asymptomatic mothers of Autism daughters demonstrating skewness, 5 also had highly skewed X-chromosome inactivation; of the mothers of the four control daughters showing skewness, none showed skewed inactivation. These results warrant further study to determine the possibility of skewed X-chromosome inactivation and/or X-linked candidate genes in the etiology of ASD in both male and female patients.
An epigenetic phenomenon similar to the one occurring in Rett syndrome has been proposed because of overlapping clinical presentations (i.e., stereotypies and regression in social and communication skills) and the discovery that a few individuals with ASD also demonstrated Rett syndrome–like mutations.362,363,361a A mutation in a regulator gene on the X-chromosome may cause the inappropriate activation or inactivation of otherwise normal genes that affect brain development and, in turn result in ASD. Between 2% and 7% of children with Angelman syndrome and a rare child with ASD also have been found to have MECP2 mutations.124,361a
Imprinting on the X-chromosome has been offered as a possible explanation for the high male : female ratio. Investigations of girls with either Turner syndrome or partial deletions of the X-chromosome revealed an increase risk for social skill deficits similar to those seen in ASD.97,100,555,556 Paternally rather than maternally derived deletions were more strongly associated with poor social cognition. Thus, it appears the paternal X-chromosome is important for development of this skill, and because boys do not receive X-chromosomes from their fathers, they might be at higher risk for social deficits as a result of imprinting (parental origin effect).
Isolated findings of Xp22 deletions or duplications have been reported in a few individuals with ASD.557
Genome screens have found a linkage to the Xq13-21 region that contains the genes that code for neuroligin, a cell-adhesion molecule that is thought to be involved with synaptogenesis.558
Mutations of the angiotensin II receptor gene on the Xq22-23 region have been implicated in one of the X-linked mental retardation syndromes in which almost 20% also meet criteria for ASD.559
A site on 2q, which appears to contain a susceptibility gene for autism and language delay (2q37), was identified in several studies, including two IMGSAC screens92,96,97,559; however, a more recent screen entailing a different database failed to confirm this.93
In a genome screen of pooled data from two countries, the 3p24-26 region emerged as the most promising.93 This locus contains the oxytocin receptor gene (OXTR). The possible link between ASD and oxytocin regulation of social behavior has been noted since the mid-1990s,96,561,562 and oxytocin receptors have been found throughout the limbic system. Social deficits in oxytocin knockout mice563 reduced oxytocin plasma levels,564 and some evidence that synthetic oxytocin ameliorated repetitive behaviors in adults with ASD565 have all pointed toward a contributing role for oxytocin.
Genome screens of this chromosome have resulted in the most consistent findings.97,98,566 Researchers have postulated a susceptibility site, called the AUTSI locus (7q31-33), where mutations in one or more of the involved genes can potentially increase the risk of an ASD.198 The RELN (7q22-33) gene and its secretory glycoprotein, reelin, appear to play a role in migration and cell lamination in the brain, especially in the cerebellum, where some of the most consistent neuropathological abnormalities occur.198,567 Other genes also occur at this site and may play a role. The FOXP2 gene (7q31-35) appears to play a role in the embryonic development of neural pathways involved in the acquisition of expressive language. Several mutations have been found in patients with speech disorders.568 Finally, the WNT2 gene on chromosome 7 appears to play a role in social skills.569
A variety of cytogenetic abnormalities occur at the 15q11-13 locus (duplications, deletions, translocations). In regard to ASD, 1% to 4% of study cohorts may demonstrate a duplication, usually maternally derived, at this site.100,122,123,570,571 A “chromosome 15 phenotype” has begun to emerge that is characterized by hypotonia, joint laxity, global (especially motor) developmental delays, seizures, speech delay, social deficits, stereotypies, and a variable pattern of mild facial dysmorphisms.123 Other abnormalities (deletions) also occur at the site and produce either Angelman or Prader-Willie syndrome, depending on the parent of origin. A GABA receptor gene (coding for a neurotransmitter highly implicated in ASD) also occurs at this site.92,100
Because serotonin is pivotal during brain development213,572 and platelet serotonin is one of the most common laboratory abnormalities in children with ASD,96,573,574 the serotonin transporter gene (17q11-12) has become a popular target for study. A susceptibility site at 17p12-q21 was noted to be the second most promising one in a comprehensive genome screen involving two databases from different countries.93
Terminal deletions at 22q13 have been associated with hypotonia, developmental delay, autistic-like behavior, and subtle physical features (ear anomalies, short nose, smooth philtrum, and full lips).366 Another study reported that 14% of patients with a confirmed microdeletion of 22q11.2 also met criteria for an ASD.367 Two other well-known syndromes (velocardiofacial [Shprintzen] syndrome and DiGeorge syndrome) are also associated with microdeletions at this site.113,575
Older IMGSAC studies have consistently revealed possible sites on 2q, 7q, and 17q; a more recent combined analysis of two primary genome scans from AGRE (United States) and Finnish populations (314 autism-affected families) revealed the best loci to be 3p24-26 and 17p12-q21 with additional promising sites at 1p12-q25, 4q21-31; 5p15-q12; 6q14-21, 7q33-36; 8q22-24; and 19p13-q13.93
|
AGRE, Autism Genetic Resource Exchange; GABA, γ-amino butyric acid; IMGSAC, International Molecular Genetic Study of Autism Consortium.
Table 15-4 describes findings of genetic investigations that have, for the most part, targeted etiological possibilities for “idiopathic ASD,” which represents most cases of ASD. Although the literature provides multiple systems for characterizing ASDs, it is perhaps most helpful in a discussion of etiology to subtype ASDs as either idiopathic or secondary.103 For the purposes of this discussion, patients with idiopathic ASD are those who do not have a coexisting associated medical condition or syndrome known to cause an ASD. Most individuals with ASD (perhaps almost all of those with Asperger syndrome) have the idiopathic subtype. Children with idiopathic ASD demonstrate variable behavioral phenotypes, are less likely to have coexisting mental retardation, and do not have dysmorphic features heralding a recognizable syndrome. Nevertheless, twin and family studies have revealed that idiopathic ASD is very heritable, with a recurrence rate of 3% to 7%,22 although phenotypic expression may be modified by other variables.92
Patients with “secondary” ASD are children with a known identifiable syndrome or medical disorder believed to play an etiological role in ASD; this occurs in only 2% to 10% of cases.23,66,103–113 In a meta-analysis of 23 epidemiological studies, Chakrabarti and Fombonne47 reported that a recognizable condition was identified in an average of 6% of those with a confirmed ASD. The rate of coexisting mental retardation was 26%, the lowest reported prevalence to date. The presence of severe mental retardation, especially when associated with dysmorphic features, increases the likelihood of identifying a genetic etiology.10,104,110,114 Genetic syndromes associated with ASD and coexisting mental retardation include
Of these four entities, the fragile X syndrome is the most common known genetic cause for the autistic phenotype, present in 1% to 5% of children with ASD, whereas 30% to 50% of those with genetically confirmed fragile X syndrome demonstrate some characteristics of ASD.117,126 The presence of a known disorder does not automatically imply causation. A few children with genetic syndromes characterized by features quite different from ASD may also meet full DSM criteria. For example, investigators have reported that that 6% to 7% of children with Down syndrome (usually characterized by relatively good social skills and obvious physical stigmata)58,127,128 and almost 50% of children with the CHARGE association59,129 meet criteria for a diagnosis of either autistic disorder or PDD-NOS. Children with severe congenital sensory impairments (visual and/or auditory) are also at risk for the development of symptoms consistent with ASD, especially when appropriate early intervention is not provided.60,61 Advancing paternal age has been shown to be associated with an increased risk of ASD possibly due to de novo spontaneous mutations and/or alterations in genetic imprinting.129a
A variety of immunological abnormalities in T cells, immunoglobulins, and anti–brain autoantibodies have all been reported in retrospective case studies,130–134 but systematic studies have confirmed neither their existence nor their relevance.102,135 Prospective studies have revealed that, except for a few individuals with recurrent infections, healthy children with ASD generally have normal immune function.135 Epidemiological data revealed clustering of autoimmune disorders in ASD families; however there was no increase in autoimmune disorders of the central nervous system and the patients with ASD did not themselves exhibit autoimmune disorders.136 Food allergies have also been implicated to play an etiological role in a few case reports,137,138 but, again, this has not been confirmed with rigorous studies.12,102,139
Environmental Factors
Regardless of the mechanism, a review of studies published since the 1950s reveals convincing evidence that most cases of ASD result from genetic factors with possible interacting environmental factors.22,92,140,141 Environmental influences may represent a “second hit” or “trigger” phenomenon; that is, they may modulate/stimulate preexisting genetic factors to result in the manifestation of ASD in an individual child.
Environmental factors should have their greatest effect during the prenatal period, especially early in gestation, because the developmental brain abnormalities associated with ASD occur during the first and second trimesters.141–144 Factors already identified to play a role include maternal rubella145 or cytomegalovirus infections146,147 and treatment with valproate148,149 or thalidomide.144,150 An isolated report150 indicated that fetal exposure to thalidomide during days 20 to 24 of gestational age was associated with ASD symptoms; later exposure resulted in the more characteristic limb abnormalities but no ASD characteristics. Nelson and associates151 reported increased cord blood levels of brain-derived neurotrophic factor and other neurotrophins in newborns in whom ASD was later diagnosed, which may have implications regarding the mechanism of the characteristic early brain overgrowth. Some investigators feel that fetal toxin exposure might be implicated by studies that have demonstrated higher rates of ASD in offspring of mothers who resided in urban settings during pregnancy152,153; however, other factors such as better access to diagnostic services in urban areas may be operative.
PERINATAL
Perinatal events have also been investigated, but findings have not been consistent and need replication.154–157 Among other factors, the strongest associations have been with threatened abortion and advanced maternal age.158 An association has been suggested between full-term neonatal encephalopathy and later diagnoses of ASD.159 In one study, ASD were diagnosed in 5% of survivors, which represented an almost sixfold increase in comparison with matched controls. This association, if replicated and confirmed, may represent a genetically derived predisposition making the infants vulnerable to both encephalopathy and ASD or a causative mechanism.
POSTNATAL
Postnatal causes of ASD are less likely possibilities. Moderate environmental deprivation has not been found to play an etiological role in ASD.152,160 One study of Romanian orphanages revealed that severe environmental deprivation could lead to something resembling at least “quasi-autism”; however, most symptoms disappeared if the children were adopted into nurturing homes during early childhood.161 A causal association between the measles, mumps, and rubella (MMR) vaccine and the development of ASD was proposed in a controversial case series study of 12 children with autistic regression and colitis.162 The onset of autistic behavior reportedly occurred shortly after receipt of the MMR vaccination. The Institute of Medicine,163 the AAP,164 the British Medical Research Council,165 and the Cochrane Collaboration166 reviewed epidemiological studies and other published and unpublished evidence and concluded that there is no evidence of a causal association. In 2004, 10 of the 13 authors of the original article stated in a retraction that they did not believe the data supported the conclusions regarding a possible causal relationship.167 Most recently, a multisite Collaborative Programs of Excellence in Autism study of 351 children with ASD failed to reveal any evidence of a causal association between ASD symptoms and the MMR vaccine.168
Questions have been raised repeatedly80,81,153 about the possible effects of environmental mercury exposure and mercury-containing vaccines on the development of ASD and other developmental disabilities. From large data sets from the United States, Sweden, and Denmark, no consistent association has been found between vaccines containing thimerosal (the mercury-based preservative) and ASD or other neurodevelopmental outcomes.169–171 Statements that the neuropathological and clinical presentations of ASD and mercury poisoning are similar were also refuted by Nelson and Bauman.172 Although scientific evidence170 of no association between vaccines (specifically MMR and thimerosal- or mercury-containing vaccines) and ASD continues to accumulate, many parents, as well as some professionals, remain unconvinced of the validity of the evidence.81a In a best seller (Evidence of Harm, 2005), Kirby82 hypothesized that various governmental agencies and medical organizations have conspired to cover up the possible harmful effects of thimerosal. The disbelief in the scientific evidence regarding mercury and vaccines remains one of the most challenging public heath problems faced by pediatricians in the United States today.
Although the preceding discussion reveals the wide variety of coexisting conditions known to be associated with ASD, a thorough etiological investigation in the individual child with ASD rarely identifies a known cause, especially in the absence of mental retardation, dysmorphic features, a positive family history, and/or positive results of a focal neurological examination.10,12,47,104,107 Multicentered collaborative studies are needed and are currently being designed to systematically evaluate children with ASD for comorbid conditions.173 The inclusion of controls with mental retardation and/or other disabilities is important in determining whether the conditions are unique to ASD or equally prevalent among children with significant neurodevelopmental disabilities in general.
NEUROLOGICAL CORRELATES
A neurobiological basis for autism was first suspected in the 1970s when it was noted that approximately one third of persons with autism had epilepsy.174 The neurological basis was further supported by evidence that autism was associated with tuberous sclerosis, a neurocutaneous disorder.118 Since that time, a growing body of evidence has revealed that brain growth and organization are different between normal children and those with ASD and that the onset of these abnormalities begins early in prenatal brain development, in some cases as early as days 20 to 24 of gestation.15,95,144,150,175 Information regarding brain development, cytoarchitecture, and functioning has been accessed through a variety of methods, including evoked potentials, electroencephalography (EEG) and magnetoencephalography, structural and functional MRI (sMRI and fMRI), positron emission tomography, neurotransmitter levels, neuropsychological testing, and postmortem examination of brain tissue. Many findings have been nonspecific and inconsistent; neuroimaging studies sometimes reveal results that are disparate from those of microscopic tissue examinations. Study samples have been characterized by a wide spectrum of behavioral heterogeneity and IQ scores. The presence or absence of coexisting mental retardation has often confounded the interpretation of reports and has been responsible for conflicting results. The highly variable and complex ASD phenotype, coupled with changing DSM criteria, has prevented the recognition of a cohesive neurological mechanism to explain all core symptoms. Research has been further impeded by the dearth of available specimens, lack of an animal model, and logistical challenges in studying children who are nonverbal with behavioral challenges.
Although individuals have been evaluated with clinical neuroimaging since the 1960s, the first rigorous research-quality neuroimaging studies did not begin to appear in the literature until the 1980s.176 The first systematic postmortem study of tissue in an individual with known autism was reported in 1985.177 In 1988, the first MRI abnormality (hypoplasia of the cerebellum) was reported.44 Since then, studies describing neurophysiological and neuropathological abnormalities, as well as theories regarding their effect on behavior and learning, have mushroomed. For a more thorough review of the neurological aspects of ASD, the reader is referred to related chapters by Volkmar and colleagues14 (see also Anderson and Hoshino178) and reviews by Bauman and Kemper15 and Polleux and Lauder.179 In 2006, the neurological findings supported by the greatest degree of evidence included the following:
Although early investigators occasionally reported slightly enlarged ventricles, the changes were therapeutically insignificant.176 Studies of the cingulated gyrus, basal ganglia, thalamus, and brainstem are fewer in number and less conclusive. There are many additional findings and theories, but these are not as well studied, more controversial, or confounded by comorbid medical conditions and/or intellectual deficits.
Seizures and Electroencephalographic Data
Electrophysiological studies were among the earliest used to probe differences in the autistic brain and provided the first evidence that autism was a neurological rather than a behavioral disorder. Abnormal electroencephalographic results are more prevalent in autism than is frank epilepsy (50% vs. 30%, respectively)176,180 and reflect the disruption of balance between inhibition and excitation, which in turn can negatively affect attention and sensory processing.179,181
Brain Growth
Kanner6 himself noted that 5 of his 11 original patients had large heads, but it was not until the1990s that brain size/volume in ASD was systematically studied through both indirect methods (occipitofrontal circumference measurements) and direct methods (sMRI and postmortem examinations). Although these early studies revealed increased brain size182 and volume,183 the mechanisms and changes in growth velocity over the developmental period in children were not recognized until the late 1990s. Head size does not appear to be large at birth; in fact, it is sometimes reported as somewhat smaller than average.184 Acceleration of growth, as evidenced by serial occipitofrontal circumference measurements and confirmed with volumetric studies, begins to occur around 6 months and peaks between 2 and 4 years of age, thus occurring in concert with (or even before) the appearance of ASD symptoms.185–188 Approximately one third of children with ASD meet criteria for macrocephaly, and 90% have greater than average brain volumes.185 Growth appears to level off in late childhood in the majority; thus, brain volume is not significantly different from that in controls by age 12 years, although macrocephaly (as measured by occipitofrontal circumference) may persist.186,189,190 Furthermore, increased cortical folding resulting in abnormal gyral patterns, reflecting increased volume, has been noted in affected children but not adolescents or adults.191 These findings appear very consistent, especially when study subjects are matched for intelligence. The main contribution to increased size lies in nonuniform changes in the hippocampus, amygdala, and cerebral white matter and, less consistently, gray matter.185,192–194 Although several theories have been proposed for the abnormal growth pattern (e.g., increased neurogenesis, glial cell proliferation, abnormal myelin, and/or decreased apoptosis/pruning), it seems most likely that it is the outer radiate zone, related to intrahemispheric synaptogenesis and connectivity, that matures postnatally and produces the rapid growth in ASD. Interestingly, parents of children with ASD can have macrocephaly in the absence of ASD symptoms. Thus, macrocephaly might be caused by a susceptibility gene that works in concert with other genes to produce ASD.
Although there are numerous published sMRI studies, no consistent abnormalities have been reported with regard to the gross anatomy or growth of other brain structures such as the brainstem, basal ganglia, and cerebellum.44,182,183,192,193 Several studies have consistently shown impaired growth (caused by hypoplasia, not atrophy) in the body and posterior regions of the corpus callosum. Only one study has correlated these anatomical findings with deficits in interhemispheric cognitive tasks.195 Rather than a focal neurological abnormality, ASD seems to be characterized by abnormalities of neural distribution and connectivity with excessive intrahemispheric connectivity and deficient interhemispheric connections (corpus callosum).176
Cerebellar Purkinje Cells
Although comparisons of sMRI and volumetric studies of the cerebellum have been controversial, one of the most consistent findings over time has been the marked decrease in Purkinje cells noted in postmortem microscopic studies.15,95,196 The absence of empty baskets suggests that the process is one of hypoplasia rather than atrophy after a noxious event, but some authorities disagree.197 Reductions in reelin may contribute abnormal regulation of neuronal layering and microscopic abnormalities found in the cerebellum.198 The absence of glial hyperplasia indicates the pathological process occurs early in brain development, before the time when the brain is able to initiate a reaction to neuronal injury. Furthermore, the number of olivary neurons is preserved, which provides additional evidence that the process must occur before weeks 28 to 30 of gestation. After this time, tight neuronal unions form between the two areas, and cell loss in the cerebellum would prompt an obligatory retrograde cell loss in the ascending olivary neurons.15,95,143,175 Although it has long been known that the cerebellum played a role in motor learning, modulation, and coordination, there is growing evidence that it also plays a role in verbal processing, affective behavior, and shifting of attention.199 Bauman and Kemper175 suggested that early in embryological life, the climbing olivary neurons might form primitive unions with collateral cells in the lamina dessicans (which disappears at 28 to 30 weeks) of the cerebellar peduncles. These neural units are not as efficient as the primary pathways and cease to function after a brief period. An intriguing question is whether this process may contribute to “autistic regression.”
Decreased Cell Size, Increased Cell Number, and Increased Packing Density in Limbic Structures
Investigators have targeted the limbic system because it plays an important role in social behavior/cognition (amygdala) and associative social memory, especially relationships among the emotional aspects of an experience (hippocampus). Postmortem microscopic studies have consistently revealed abnormalities in cell number and size and packing density in both the amygdala and hippocampus. However, sMRI and volumetric data have been inconsistent, diverse, and often contradictory.193,200,201
Abnormal Minicolumns in the Cerebral Cortex
Abnormal cortical minicolumns (defined as the most basic unit of neural organization) have been added to the growing number of neuropathological abnormalities found in ASD. Although Bailey and colleagues202 described several abnormalities of pyramidal neuronal migration (ectopic neurons in white matter zones, misoriented apical dendrites, and disorganized cellular layers) in the superior temporal gyrus. Casanova and coworkers203 more recently introduced “minicolumn” terminology into ASD literature. Minicolumns in some areas of the autistic frontal cortex were found to be smaller (representing an underdeveloped system) and had abnormal patterning. Both findings are consistent with deviant processes that occur very early in the second trimester. These anatomical abnormalities may result in deficient neuronal “insulation” and serve as the structural basis for increased neuronal “cross-talk” and overstimulation. This, in turn, may cause the sensory gating and processing difficulties found in some individuals with ASD.203,204 Other autistic symptoms then might be explained by a disregulation of axonal outgrowth, dendritic arborization, and synaptic connectivity.179 Abnormalities described in cortical frontostriatal circuits may be associated with ritualistic and repetitive behaviors.204 Volumetric sMRI studies of cortical systems serving language functions have revealed the absence of the usual left hemispheric hypertrophy (representing left brain dominance and language specialization). Instead of a larger left hemisphere (specifically, the Wernicke receptive language processing area), the planum temporale volumes were equal in subjects with ASD.205 Furthermore, decreased gray matter in the left inferior prefrontal gyrus or Broca’s area (expressive language center) resulted in actual reversal of the typical hemispheric asymmetry (left larger than right) in language-impaired subjects with ASD.206
Hypoactivity in the Fusiform Gyrus during Face Recognition Tasks
Functional neuroimaging techniques, primarily positron emission tomography and fMRI, have confirmed clinical and neuroanatomical data depicting ASD as a disorder characterized by uneven rather than generalized deficits.207 The most consistent fMRI finding has been hypoactivity in the fusiform gyrus, particularly in the fusiform facial area, confirming the clinical impression that deficits in facial recognition are characteristic of ASD.208,209 fMRI has also demonstrated associated deficits in related areas of the “social brain,” such as the amygdala, which plays a critical role in emotional arousal and integration of emotional data.210 Persons with ASD appear to be less motivated to look at faces or to follow the point of conversational partners.208 Computerized eye tracking techniques have also revealed that they pay less attention to faces and more attention to inanimate details in the background.209,211,212 When they do look at the face, they target the mouth rather than the eyes. Because oral expressions provide less information about emotional states than the eyes, persons with ASD often fail to detect meaningful social information during interactions.208
Neurochemical Testing
Neurochemical abnormalities may also be present in children with ASD.178 Increased levels of 5-hydroxytryptamine in whole blood, chiefly platelets, has been a fairly consistent finding. Although 5-hydroxytryptamine is an important neurotransmitter for brain development and modulation of sleep, mood, body temperature, appetite, and hormone release, no consistent abnormalities have been found in central nervous system levels. Age-related differences in serotonin synthesis capacity have also been demonstrated between children with autism and nonautistic controls.213
In conclusion, it is widely accepted that ASD is a “neurodevelopmental disorder,” although the specific underlying abnormalities have not been identified. ASD may actually represent a disorder of neural distribution rather than frank structural abnormalities.176 A project to create the first ever atlas of the autistic brain at several ages is well under way.214 Newer functional brain studies have provided some intriguing links between the neuroanatomical substrate and characteristic clinical features. Well-designed studies with participants matched for IQ levels and with the most sophisticated technology (e.g., diffusion tensor imaging) are needed to unravel the mystery of anatomical differences and white matter “connectivity” and associated discrepancies in learning and neuropsychological functioning. Such studies may provide valuable information that can be used to design, implement, and evaluate new and effective intervention strategies.
CLINICAL SIGNS
Although emphasis has historically been placed on language deficits, they are not specific to ASD and are commonly also the presenting feature of children with mental retardation, hearing loss, and communication disorders. Stereotypies may be obvious and easily recognized, but they also occur in other conditions, primarily severe mental retardation and blindness. Furthermore, they often do not appear until after 3 years of age,215 and some forms (e.g., hand flapping) can be normal in certain situations (e.g., in an excited toddler). Thus, language deficits and stereotypies do not clearly distinguish ASD from other childhood disorders. During the 1990s, it became apparent that the social deficits, specifically those relating to “social communication,” were the most consistent and characteristic symptoms of ASD. The diagnosis of classic autistic disorder currently requires that at least one criterion be met in each of the language and restricted interests domains and that two criteria be met in the social skills domain.19,20 Several early recognizable social communication deficits (e.g., joint attention) appear to be fairly specific for ASD. The severity of these deficits varies significantly from patient to patient, thus creating diagnostic challenges.
Most parents first become concerned about their children’s development when they are between 15 and 18 months of age216,217; their first concerns usually focus on speech delays. Indeed, this has been the historical hallmark of ASD and will probably continue to be so because these deficits are easily recognized. However, with heightened public awareness about the early signs that occur before development of vocal speech, parents are beginning to voice concerns about more subtle receptive language skills (the child’s not responding to his or her name being called) and social skills (e.g., decreased eye contact, unusual attachments to objects, not caring whether parent is nearby). Studies have demonstrated that symptoms can appear before 1 year of age, although these may be subtle.218 Some infants appear to develop normally until approximately the second year of life, when they demonstrate regression in speech and social skills, withdraw, and become indifferent to their surroundings.219,220 Subtle abnormalities in social communication may be evident on careful examination of 1-year-old birthday video recordings.221–223 Expanded discussions can be found in chapters and reviews dedicated solely to early clinical characteristics.224–227
Social Skills Deficits
As noted previously, abnormalities in social communication skills are the most unique and consistent findings in infants with ASD. They appear earlier than identifiable speech deficits but are often more subtle. Children with ASD universally demonstrate deficits in social relatedness, defined as the inherent drive to connect with others and share complementary emotional states.228 Children with other types of disabilities (e.g., mental retardation, sensory disabilities) still attempt to connect with others: Those with hearing deficits compensate with eye gaze and gestures, and those with vision deficits compensate with voice and touch. Children with ASD do not appear to seek this connectedness; they are usually content with being alone, rarely make eye contact or bids for others’ attention with gestures or vocalizations, and react little to praise or bids for attention from others. In later years, they have difficulty in sharing the emotional state of others in cooperative games and group settings and have few, if any, friends.
DEFICITS IN JOINT ATTENTION AND SHARING OF INTERESTS
One of the most distinguishing characteristics of very young children with ASD is a deficit in “joint attention.”229–235 Joint attention is the desire coupled with the ability (facial expressions, gestures and/or speech) to draw another’s attention to objects, events, or other persons simply for the enjoyment of sharing experiences. Like other developmental skills, joint attention appears to develop in graduated stages, usually between 8 and 16 months. At approximately 8 months of age, a typically developing infant may participate in “gaze monitoring”: that is, when a parent looks away (e.g., to check the time), the infant follows the parent’s gaze and looks in the same direction. This social skill should be differentiated from simple auditory orienting, in which both the infant and the parent are stimulated by an environmental stimulus (e.g., clock alarm) at the same time. Children begin to “follow a point” at about 10 to 12 months of age. If a parent points in the direction of an interesting object or event and says, “Look,” the typically developing child looks in the direction that the parent is pointing. Upon seeing the object/event, the child looks back at the parent and smiles, frowns, or shows fear, whichever emotion is appropriate to the situation. An infant with ASD does not follow a point even when a parent tries repeatedly, calling the child’s name in a loud voice or with physical prompts such as touching the child’s shoulder before pointing.218
At 14 to 16 months of age, the typically developing child begins to point simply to “comment” about, or “share,” an interesting object/event (protodeclarative pointing). As the child points, he or she looks alternatively between the object/event of interest and the caregiver. The same triad exists (child, caregiver, object), but the goal is reversed. It is the shared social experience, not the object, that the child seeks. Children with ASD consistently fail to point to “comment” at age-appropriate times. If and when they do start to point, they are less likely to show positive affect and connectedness during the act. About the same time, typical children also begin to “show” items of interest to parents. This bid for attention is distinct from asking for help (e.g., bringing a jar of bubbles to the parent as a request to open it). Mastery of joint attention appears to be a reliable predictor of functional language development.13,236–238
The ability to disengage and shift focus of attention from one stimulus to a novel one is a very basic skill that can be measured in normally developing 6-month-olds. The inability to shift attention was proposed as a characteristic deficit in autism that possibly contributed to deficits in joint attention, inasmuch as it relies on shifting attention between an object/event and a partner.44,199 Multicenter studies of infant siblings of older children with ASD have revealed that the inability to shift one’s attention (from parent to object of interest and back to parent again) is measurable and perhaps the first reliable sign of ASD.218
POOR SOCIAL AND EMOTIONAL RECIPROCITY
One of the earliest developmental milestones is the ability to orient to social stimuli—in particular, turning to respond to one’s own name.239 At about 8 to 10 months of age, most children turn preferentially when their name is called. Like children with hearing impairments, those with ASD often fail to orient to their name. In fact, an early concern of parents of children later diagnosed with ASD is about their infant’s hearing. Parents are often puzzled because such children seem to attend to environmental sounds better than to human voices.240
