Diagnosis and management of autism in childhoodBMJ 2011; 343 doi: http://dx.doi.org/10.1136/bmj.d6238 (Published 21 October 2011) Cite this as: BMJ 2011;343:d6238
- Stephanie Blenner, assistant professor of pediatrics1,
- Arathi Reddy, developmental behavioral pediatrician2,
- Marilyn Augustyn, associate professor of pediatrics, division director1
- 1Division of Developmental Behavioral Pediatrics, Boston University School of Medicine, Boston, MA 02118, USA
- 2Boston Medical Center, Boston, MA, USA
- Correspondence to: S Blenner
Autism spectrum disorders are neurodevelopmental conditions that share core impairments in social reciprocity, communication, and behaviour but have a range of presentations
The prevalence of autism has increased over the past 15 years, partly because the definition now includes milder forms of the disorder
Genetic factors and potential environmental causes are being studied; vaccines have been shown not to be associated with autism in multiple studies
Surveillance and screening by the general clinician allow affected children to be identified early and to gain access to crucial early interventions
Most management strategies aim to improve communication, cognitive abilities, and social and daily living skills, while decreasing maladaptive behaviour
Involvement of general clinicians and paediatricians in ongoing care helps families access services and prioritise and plan for future needs
Autism spectrum disorder is a commonly used umbrella term for a class of neurodevelopmental disorders characterised by a triad of deficits in social reciprocity, impaired communication, and repetitive restricted patterns of behaviour or interests. Symptoms are evident in early childhood, often before age 3 years, and result in functional impairment. Autism was first described in the 1940s; it was originally thought to be relatively rare because only the most severely affected people were identified. However, epidemiological studies have documented an increasing prevalence over the past 15 years, with one UK cohort study from 2006 reporting a prevalence rate of about one in 110 children compared with four to five cases per 10 000 before the 1990s.1 This reflects, in part, recognition of a broader phenotype of affected individuals who share impairments within the three core areas.2
It is crucial that general clinicians are familiar with the diagnosis and management of autism because the importance of early identification and intervention, and the benefit of supporting families in navigating the myriad of decisions once a diagnosis is made, are well recognised. We review the current nosology, identification, and evaluation of autism spectrum disorders, along with up to date research into their causes and medical management on the basis of recent studies, systematic reviews, and consensus statements.
Sources and selection criteria
We searched PubMed, PsychInfo, the Cochrane database of systematic reviews, and citation lists of relevant publications using the subject headings and key words “autism,” “autism spectrum disorder,” “(a)etiology,” “screening,” “diagnosis,” and “treatment.” We also searched guidelines from the National Research Council, the American Academy of Pediatrics, and the National Institute for Health and Clinical Excellence.
How are autism spectrum disorders classified?
Currently, autism spectrum disorders comprise three medical diagnoses, as outlined in the Diagnostic and Statistical Manual of Mental Disorders, fourth edition, text revision (DSM-IV TR) under pervasive developmental disorders, with corresponding diagnoses in ICD-10 (international classification of diseases, 10th revision). These are autistic disorder; pervasive developmental disorder, not otherwise specified (PDD-NOS); and Asperger’s disorder.
Autistic disorder is diagnosed when an individual exhibits six or more symptoms across the three core areas. PDD-NOS is not as clearly defined but is diagnosed when there are five or fewer symptoms or an atypical presentation. Children diagnosed with Asperger’s disorder do not have clinically significant delays in language development before age 3 and their greatest difficulties are in the areas of social interaction and restricted interests. Children with all three diagnoses have marked impairment in social reciprocity, which is considered the defining feature of all autism spectrum disorders. Importantly, there has been much debate in the field over recent years about terminology and construct validity, and it is anticipated that DSM-V—planned for release in 2013—will replace the three current diagnoses with the single diagnosis “autism spectrum disorder” accompanied by dimensional descriptors.3
What causes autism spectrum disorder?
Autism is a complex heterogeneous disorder that has an established genetic basis. Twin studies have identified high heritability, with median concordance rates across studies of 88% for monozygotic twins.4 Siblings of children with autism are at higher risk of being diagnosed with autism, or other developmental disorder.5 The reported risk of recurrence is at least 4-7% for an autism spectrum disorder, with one recent study making the risk as high as 18%, although this has not yet been replicated. Although single gene disorders, such as fragile X syndrome (FMR1 gene) and tuberous sclerosis (TSC1 and TSC2 genes), account for a small proportion of cases,6 the proportion of individuals who have an abnormality on genetic evaluation is increasing as more sophisticated testing becomes available. A larger proportion of people with autism are thought to have “idiopathic autism”—with no identifiable genetic abnormality. Although this partly reflects the limitations of available testing, it also suggests that autism is a final manifestation of multiple aetiological pathways. Autism probably develops when an array of possible genetic vulnerabilities, possibly in concert with epigenetic factors or gene-environment interaction, affect neural connectivity and neurodevelopment.7 8
Environmental factors, as well as prenatal and perinatal factors,9 that might play a role in the development of autism are being investigated. Childhood immunisations have been a highly publicised area of focus, and clinicians are often confronted with questions about vaccine safety. Theories about how vaccines could contribute to the development of autism include the thimerosal hypothesis; the role of measles, mumps, and rubella vaccine; and exposure to multiple vaccine antigenic components. These theories have been intensely studied in a range of countries and no association or aetiological link to childhood immunisations has been identified.10
Another controversial area has been the role of gastrointestinal atypicality and diet in the development of autism. Inadequate metabolism of gluten—a wheat protein—and casein—a milk protein—have been hypothesised to result in peptides that cross the blood-brain barrier, bind to endogenous opioid receptors, and negatively affect behaviour, cognition, and interaction. To date, however, no large well designed study has identified significant changes in core symptoms after the elimination of gluten and casein from the diet of children with autism.11
How do children with autism present?
About half of parents of children with an autism spectrum disorder report having concerns before 12 months of age, with many more reporting recognition of abnormalities between 12 and 24 months in retrospective studies.12 13 A recent prospective study of the emergence of early behavioural signs of autism in children found differences in social communication evident by the second year of life.14 Despite behavioural markers being identified within the first 2 years of life, the current average age of diagnosis remains around 3 years or more.15
The most common presenting parental concern in young children subsequently diagnosed with autism is delayed language development. Other common behavioural concerns include lack of or inconsistent use of eye contact, social smile, imitation, response to name, interest in others, emotional expression, directed vocalisations, joint attention skills (pointing to “show,” following a point, monitoring others’ gaze, and referencing objects or events), requesting behaviours, and gestures (such as waving, clapping, nodding, and shaking head). Pretend play is also deficient in toddlerhood in many children with autism.16
Regression is seen in about 25% of children according to several studies using structured parent interviews or parent report coupled with medical record review.17 18 This usually occurs between 15 and 24 months of age and can present with loss of previously attained language abilities or social and play skills. Regression can occur in apparently typically developing children or be superimposed on existing developmental delay.
Presenting concerns in older children
Although the average age of diagnosis for autism is 3.1 years, it is 3.9 years for PDD-NOS, and 7.2 years for Asperger’s disorder. This delay may occur because the child meets early language milestones but presents with other concerns such as difficulty with transitions, extremes of behaviour, and perseverative interests; these symptoms may be difficult to identify until the child is older. In one study of children in community mental health settings, school aged children with autism spectrum disorder were more likely to be referred for social interaction difficulties and strange behaviour but were less likely to be referred for behaviours like drug use, truancy, or running away.19
Associated medical problems
Seizures are seen in 20-35% of children with autism and are more common in those with dysmorphic findings or cognitive delay.20 Pica or mouthing of objects in the environment is also common. Studies have reported varying rates of gastrointestinal problems. A cohort study in 2009 found that constipation and feeding problems were higher in children with autism, but not reflux or diarrhoea, suggesting that the increased incidence has behavioural, rather than organic, causes.21 Sleep disorders are common and usually more severe in children with autism. Reported problems include difficulties with sleep onset, maintenance, and duration.
How should non-specialists assess children with suspected autism?
Developmental surveillance—which includes flexible questioning about a child’s development (box) accompanied by observation of the child in the surgery—conducted as part of comprehensive medical care at each visit will help the clinician identify concerns.22 Clinicians can also use autism specific screening to assess symptoms. Recommendations and practices around screening for autism vary by country; the American Academy of Pediatrics recommends using an autism specific screening tool at 18 and 24 months,23 whereas several other countries, including the United Kingdom, do not currently endorse routine population screening.
Questions general clinicians might ask parents
Children age 4 years or under
How does your child communicate with you?
Does your child point using an index finger?
Does your child look you in the eyes when communicating?
What does your child do if he or she sees something interesting, such as a plane or a dog?
Does your child turn to look to you when you call his or her name?
What does your child like to play with? Describe what he or she does while playing
Has your child ever stopped doing something he or she could previously do (such as talking, using eye contact)?
Children over 4 years
Does your child have friends? Tell me about your child’s friendships
What does your child like to do with his or her time?
Can you give an example of a recent conversation you had with your child?
How does your child handle change? Describe how your child reacted to a recent change or new situation
Multiple screening tools have been devised to try to standardise screening for potential autism spectrum disorder prospectively (table 1⇓). The most common tools designed for use with children as young as 18 months are the checklist for autism in toddlers (CHAT), pervasive developmental disorders screening test (PDDST), screening tool for autism in two year olds (STAT), checklist for autism in toddlers-23 (CHAT-23), and the modified checklist for autism in toddlers (M-CHAT).24 25 All use questions that assess core symptoms such as joint attention, non-verbal communication, and play. The infant toddler checklist (ITC) is used for even younger children, and it focuses on social and communication skills specific to the younger child. It has been evaluated recently in a large sample and shown to have potential for identifying 12 month olds at risk of autism spectrum disorder, but this has yet to be replicated in other samples.26 When concerns present in school aged children, instruments such as the social communication questionnaire (SCQ)—designed for children aged 4 years or more—can be used.27 28
A positive score on a screening test can support referral to a specialist team for comprehensive evaluation, but a negative score does not rule out autism.
Unless there is an associated genetic or neurological condition, children with autism often have no distinct physical findings on examination. Some studies had identified acceleration of head size between 6 and 12 months, with macrocephaly in toddler years,29 although a recent population study did not find a higher rate of macrocephaly in children with autism at 24 or 36 months.30
When to refer to a specialist
If the clinician or family is persistently worried about communication or developmental problems or a screening test identifies that a child is at “high risk” of autism, referral to a specialist team for a comprehensive evaluation is indicated.
Before evaluation by a specialist, the primary care clinician should order a hearing assessment to rule out hearing loss as a contributory factor. If the child has pica with frequent mouthing of items, measurement of haemoglobin and lead will identify anaemia or a raised lead concentration. A detailed history will uncover chronic sleep difficulties; similarly, history and abdominal radiography if necessary can diagnose common gastrointestinal problems, such as constipation. Any other presenting medical symptoms are evaluated appropriately—for example, consider electroencephalography or magnetic resonance imaging in children with seizures.
How does the specialist team perform a diagnostic assessment?
Because there is currently no specific diagnostic “test” for autism spectrum disorders, the diagnosis is established through a comprehensive evaluation that includes lifetime and family history, review of medical and educational records, behavioural observation, physical examination, administration of standardised instruments such as the autism diagnostic observation schedule, cognitive and adaptive assessment, and review of established DSM or ICD diagnostic criteria. Given the complexity and range of presentation, diagnostic evaluation is best done by paediatric neurologists, developmental and behavioural paediatricians, child psychiatrists or psychologists, or ideally, a multidisciplinary team, with specific training and experience in evaluating children with autism. Involving other professionals—such as speech and language and occupational therapists, special educators, and social workers when available—allows for more detailed assessment of specific domains, such as communication skills, sensory and motor problems, and family stressors and coping abilities. Multidisciplinary input also helps tailor interventions to a specific child’s strengths and weaknesses.
Which additional investigations will specialists perform?
An important role of the specialist evaluation is to look for causes and co-occurring conditions. Identification of genetic or metabolic diagnoses allows the family to be given more accurate information on prognosis and risk of recurrence. When a child is diagnosed with autism, the specialist will complete a detailed physical examination to look for dysmorphic and neurological findings and Wood’s lamp examination of the skin to assess for hypopigmented macules seen in tuberous sclerosis. High resolution karyotyping, DNA for fragile X syndrome, and chromosomal microarray testing have been recommended as standard laboratory tests, although the replacement of karyotyping with microarray analysis as first tier testing is being advocated by some consortium groups.31 Additional genetic testing is done in the context of specific findings—for example, MECP2 (methyl-CpG binding protein 2; Rett syndrome) testing in young girls and PTEN (phosphatase and tensin homologue) testing in children with macrocephaly greater than 2.5 standard deviations above mean values for age and sex. Metabolic testing may be indicated in children with serious neurological abnormalities, particularly those born in countries where newborn metabolic screening is not routine. Mitochondrial disorders are rare but will be considered if a child with symptoms of autism shows excessive fatigability, abnormities on neurological examination, marked delay in early gross motor milestones, or unusual patterns of regression (repeated regression or regression after age 3 years).32
The specialist will also look for medical and psychiatric comorbidities not identified during the general practitioner’s initial history and investigations. The specialist will make a detailed review of symptoms and previous laboratory results; order tests that have not been done; and use screening tools to identify clinically relevant behaviours, such as hyperactivity or aggression, or emotional concerns, such as anxiety or depression, and to assess family functioning and level of support. Cognitive tests to determine the degree of intellectual impairment—seen in about 50% of children—can be useful for making a long term prognosis.33 However, traditional cognitive testing is not always accurate in children with autism, and adaptive deficits often disproportionately affect functioning, even when cognition is within or above the average range.
What kinds of treatment are available for autism?
Although autism is a medical diagnosis, intervention is typically delivered in the educational setting. This requires the paediatric clinician to be familiar with community and educational resources and laws on educating children with disabilities. Clinicians need to be able to counsel families about appropriate treatments, in consultation with specialists involved in the child’s care.
Much recent research has focused on the development of intervention programmes for young children with autism. Multiple intervention approaches (table 2⇓) are currently used, such as applied behavior analysis (ABA), the Denver model, and TEACCH. The underlying principles of these approaches vary (for example, behaviour analytical, developmental, and structured teaching) but target common goals. In a recent systematic review, studies of ABA approaches, early intensive behavioural intervention variants, and the early start Denver model improved cognitive performance, language skills, and adaptive behaviour skills in some young children with autism, although the literature is limited by methodological concerns.34 Educational programmes may use one or more approach (see table 2).
Although approaches vary, there is growing consensus about the crucial components of an educational programme for children with autism.35 Intensive intervention should start as soon as the diagnosis is seriously considered. The Institute of Medicine’s National Research Council recommends year round, systematically planned intervention of at least 25 hours a week. Appropriate educational settings consist of classrooms with a high degree of structure and a low student to teacher ratio. The curriculum promotes academic skills as appropriate for the child’s developmental level. Therapists and special educators involved in the child’s care help parents incorporate the techniques being used at home and provide opportunities for interaction with typically developing peers. Disruptive behaviours are best targeted with functional behavioural assessment, which identifies potentially modifiable factors (things that happen before and after a challenging behaviour that when changed can decrease the problem behaviour). Appropriate treatment targets include the development of functional communication, as well as social and adaptive skills to maximise independence. Establishing time specific measureable goals helps to assess achievement and allows the intervention programme to be adjusted if a child is not progressing.
Many children, particularly those who have progressed with intensive intervention or who have milder symptoms, will benefit from consistent interaction with typically developing peers. Integrated classroom programmes, which include a mix of typical students and children with autism, may be appropriate for these children. For high functioning children with autism who have strong verbal skills and are achieving academically, a regular education classroom with supportive services may be an option. Support can include speech and language services to deal with pragmatic language and communication problems and social skills groups to teach appropriate social interaction.
Are drugs indicated in the management of children with autism?
Children with autism often have psychiatric comorbidities or exhibit difficult behaviours, such as aggression, self injurious behaviour, hyperactivity, impulsivity, sleep disorder, and anxiety.36 When a new difficult behaviour emerges, a thorough search must be undertaken for underlying treatable medical causes (such as otitis media, dental pain, constipation, gastero-oesophogeal reflux). If a medical condition is ruled out, implement a functional behavioural assessment and behavioural plan supporting positive alternative behaviours as first line intervention. This is best done in consultation with a behavioural specialist or psychologist. If the behaviour does not respond to this intervention and functioning is seriously impaired, consider drug treatment to deal with target symptoms such as aggression, anxiety, or emotional lability.37 Table 3⇓ summarises available drug options, target symptoms, and evidence of efficacy.
Drugs—mainly risperidone and aripiprazole to deal with challenging behaviours—are currently viewed as adjuncts in the treatment of autism.38 Some studies have associated older age at diagnosis, low adaptive skills and social competence, and higher levels of challenging behaviours with increased use of psychopharmacology.39 It is important that families are aware that currently used drugs do not treat the core symptoms of autism and should be used only as part of a comprehensive treatment programme.
What about complementary and alternative treatments?
Parents of children with autism are often interested in complementary and alternative treatments.40 Many of these treatments claim to improve symptoms, but few have been fully evaluated and even fewer shown to be effective by rigorous scientific study. Melatonin is efficacious in the management of sleep disorders associated with autism, particularly in improving sleep onset.41 However, other treatments, such as chelation therapy, can be directly harmful, with case reports of deaths in children with autism receiving chelation. Clinicians should be open to discussing these treatments with parents, so they can counsel them appropriately and encourage them to seek additional information when necessary.
What is the prognosis for a child diagnosed with autism?
One of the challenging aspects of treating children with autism spectrum disorders is the great variability in long term outcomes. Several small longitudinal studies have found that children who have absence of language by age 5 and low cognitive functioning have poorer outcomes in adolescence and adulthood on measures of self sufficiency and adaptive functioning,42 and one noted that higher cumulative early intervention hours were associated with better outcome.43 Many people with autism require supportive living arrangements into adulthood and relatively few are independently employed without support as young adults.44
When counselling families, stress that recent outcome studies looking at adolescents or adults diagnosed with autism in childhood reflect identification and intervention practices in place 10-20 years ago. Children who are less severely affected or benefiting from current intensive interventions may have a better prognosis. The clinician can emphasise that—although many people with autism may experience challenges with independent living, employment, social relationships, and mental health over their life course—the expectation is for developmental progress over time and that all children with autism can be supported in maximising their potential.
Questions for future research
What is contributing to the prevalence of autism?
Do particular environmental factors play a role in autism and how do they interact with genetic factors?
Is there a biomarker for autism spectrum disorder?
Can we identify which educational approaches and techniques would be most beneficial on the basis of each child’s presentation?
What drugs are effective in treating the core symptoms of autism?
What are the long term adult outcomes in children who receive specialised intensive early intervention from the time of diagnosis?
Tips for non-specialists
Assess for developmental delay at every well child visit
The use of an autism specific screening tool will increase the detection of autism, particularly if you are concerned about developmental delay or if parents report concerns about delayed language development or impaired social interaction
A negative screen does not rule out autism
Refer children for whom there are persistent concerns or who score positive on a screening tool to a specialist or multidisciplinary team, if available, for a diagnostic assessment
Initial investigations before referral should include a hearing evaluation and a detailed history to identify comorbidities such as seizures or constipation
Additional educational resources
Resources for healthcare professionals
First Signs Video Glossary (www.firstsigns.org/asd_video_glossary/asdvg_about.htm)—free video clip examples of symptoms of autism
American Academy of Pediatrics Toolkit (www.aap.org/publiced/autismtoolkit.cfm)—Useful resource for clinicians caring for children with autism; available for purchase
OCALI Autism Internet Modules (www.autisminternetmodules.org/)—A series of online learning modules about autism related topics and interventions
Resources for families
First Signs (www.firstsigns.org/)—Free parental education about autism
First Signs Video Glossary (www.firstsigns.org/asd_video_glossary/asdvg_about.htm)—Free video clip examples of symptoms of autism
Autism Speaks First 100 Days Kit (www.autismspeaks.org/docs/family_services_docs/100_day_kit.pdf) and OCALI Autism Internet Modules (www.autisminternetmodules.org/)—Free online learning modules about autism related topics and interventions
Cite this as: BMJ 2011;343:d6238
Contributors: SB, AR, and MA all helped write this article and are all guarantors.
Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: SB, AR, and MA had support from Boston University School of Medicine and Boston Medical Center for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.
Provenance and peer review: Commissioned; externally peer reviewed.