PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women’s and Children’s Health (UK). Autism: Recognition, Referral and Diagnosis of Children and Young People on the Autism Spectrum. London: RCOG Press; 2011 Sep. (NICE Clinical Guidelines, No. 128.)

6Differential diagnosis

Introduction

Many neurodevelopmental, mental and behavioural disorders may present with symptoms that suggest the possibility of autism but which are not autism. These can be described as the differential diagnoses of autism. It is essential to consider the differential diagnoses at each stage of the autism pathway: when the possibility of autism first arises and consideration is being given to referral to an autism team (see Chapter 3 on Recognition); when the autism team is considering whether to proceed with an autism-specific diagnostic assessment (see Chapter 4 on Following referral); when undertaking an autism diagnostic assessment; and when considering the diagnosis on completion of the assessment (see Chapter 5 on Diagnostic assessment).

If there are concerns about a child’s or young person’s development or behaviour, and especially if the possibility of autism has been raised, parents, carers and the child or young person may be anxious to know without delay what the nature of the problem may be. It is important to establish an accurate diagnosis, whether that is autism or an alternative condition. An inaccurate diagnosis of autism may result in the use of an inappropriate treatment strategy and may cause anxiety and distress to the child or young person and their parents/carers. This chapter addresses the most important disorders to be considered in children and young people presenting with possible autism and how they may be differentiated from autism. A differential diagnosis may also be a coexisting condition (see Chapter 7 on coexisting conditions).

Clinical question

  1. What are the most important differential diagnoses of autism?
  2. What features observed during diagnosis reliably differentiate other conditions from autism?

6.1. Overview of the evidence: identifying differential diagnoses

Nineteen studies were included in this review. These studies were carried out in the Australia,66;67;137 Canada,138 Germany,139 Israel,140 Italy,141 Japan,142 Norway,143 Sweden,70;144 the Netherlands,145;146 the USA,73;74;108;147 and the UK.148;149 All were uncontrolled observational studies and were graded as very low quality.

Eight of the studies were in a preschool population,67;74;108;140;142;145;147;149 and one study was in primary school age children148: there were no studies in secondary school age children. Five used a mixed population of preschool and primary school age children,66;137;138;141;144 two used a mixed population of primary and secondary school age children70;146 and three included children or young people of all ages.73;139;143

Only one study reported the range of intelligence quotient (IQ).146 Four studies reported mean IQ scores but the proportion of children with intellectual disability was not reported. 73;137;139;147 Four studies reported the proportion of children with intellectual disability but no separate outcomes were provided for each IQ group.67;70;140;143 Intellectual ability was not reported in the remaining studies.

6.2. Evidence profile: identifying differential diagnoses

Table 6.1 reports the prevalence of each differential diagnosis in children with suspected autism.

Table 6.1. Prevalence of alternative diagnoses in children with suspected autism.

Table 6.1

Prevalence of alternative diagnoses in children with suspected autism.

Table 6.2 shows the prevalence of each differential diagnosis in children with suspected autism spectrum disorders (ASD).

Table 6.2. Prevalence of alternative diagnoses in children with suspected ASD.

Table 6.2

Prevalence of alternative diagnoses in children with suspected ASD.

The conditions are reported under five categories identified by the guideline development group (GDG). Limitations, inconsistencies and indirectness are not reported in the table because the quality is very low.

The evidence for autism is reported separately from ASD as it was expected that some differential diagnoses would have different prevalence rates for each category and so it would not be appropriate to pool these data. Subgroup analyses are reported in relevant evidence profiles and evidence statements.

6.3. Evidence statements: identifying differential diagnoses

All evidence was graded as very low quality.

Evidence for autism

All studies

Mental and behavioural disorders

Evidence on two diagnoses (ADHD and a behaviour problem) of children and young people with suspected autism was identified. One study reported the prevalence of ADHD and one reported the prevalence of behaviour problems. The prevalence for both was 8%.

Neurodevelopmental problems

Evidence on only one diagnosis (developmental disorder/delay)] was identified. The pooled prevalence was 6% (95% confidence interval [CI] 1, 15).

Medical or neurological problems

Only one study was found that looked at medical or neurological problems and this reported on two diagnoses: Rett syndrome and motor problems. The prevalence was 10% and 3% respectively.

Studies of children referred on suspicion of autism only

Mental and behavioural disorders

Evidence was identified for two diagnoses (a behaviour problem and ADHD). One study reported the prevalence of a behaviour problem and one ADHD. The prevalence for each was 8%.

Neurodevelopmental problems

Evidence on diagnosis of developmental disorder/delay was identified in two studies. The pooled prevalence was 6% (95% CI 1, 15).

Medical or neurological problems

Only one study was found that reported on two diagnoses (Rett syndrome and motor problems). The prevalence was 10% and 3% respectively.

Studies of children and young people referred for developmental problems only

No study met the inclusion criteria for this review.

Studies of children and young people referred for behavioural problems only

No study met the inclusion criteria for this review.

Studies of children and young people referred for positive screening results only

No study met the inclusion criteria for this review.

Evidence for ASD

Complete analysis: all studies

Mental and behaviour disorders

The prevalence of six diagnoses (behaviour problems, ADHD, emotional difficulties, Tourette syndrome, selective mutism and attachment disorder) were identified from evidence. Only data of the most prevalent differential diagnoses (behaviour problem, ADHD and emotional difficulties) are reported here.

Two studies reported the prevalence of behaviour problems in children and young people suspected of having ASD, seven on ADHD and three on emotional difficulties. The pooled prevalence was 24% (95% CI 1, 80), 14% (95% CI 6, 24) and 6% (95% CI 2, 10) respectively.

Neurodevelopmental problems

The prevalence of three diagnoses (a language problem, developmental disorder/delay and disintegrative disorder) was identified from evidence. Only data on the most prevalent differential diagnosis (language problem and developmental disorder/delay) are reported here.

Twelve studies reported on the prevalence of a language problem in children and young people suspected of having ASD and 13 on developmental disorder/delay. The pooled prevalence was 21% (95% CI 5, 43) and 15% (95% CI 8, 23), respectively.

Medical or neurological problems

The prevalence of three diagnoses were identified from evidence. One study reported on the prevalence of Down’s syndrome and fetal alcohol syndrome and one on the prevalence of motor problems. The prevalence was 3%, 3% and 2%, respectively.

Other

One diagnosis was identified that did not fit the other categories, which was abuse/neglect. The study reported a prevalence of 26%.

Studies of children referred on suspicion of ASD only

Mental and behaviour disorders

The prevalence of six diagnoses (ADHD, behaviour problem emotional difficulties, Tourette syndrome, selective mutism and attachment disorder) was identified from evidence. Only data of the most prevalent diagnoses are reported here.

Three studies were identified that reported on ADHD, one on a behaviour problem, two on emotional difficulties and one on selective mutism. The pooled prevalence for ADHD and emotional difficulties was 6% (95% CI 2, 13) and 4% (95% CI 3, 6) respectively. The prevalence of the behaviour problem and selective mutism was 4% and 1%, respectively.

Neurodevelopmental problems

The prevalence of three diagnoses (a language problem, developmental disorder/delay and disintegrative disorder) was identified from evidence. Only data of the most prevalent differential diagnoses are reported here.

Six studies were identified that reported on a language problem, while four studies reported on developmental disorder/delay. The pooled prevalence was 9% (95% CI 3, 17) and 5% (95% CI 3, 6), respectively.

Studies of children referred on suspicion of developmental problems only

Mental and behaviour disorders

Evidence showed the prevalence of emotional difficulty was 16%.

Neurodevelopmental problems

The prevalence of four neurodevelopmental diagnoses was identified from evidence. Only data of the most prevalent differential diagnosis are reported here.

Four studies were identified that reported on a language problem in children and young people referred for developmental problems, and four on developmental disorder/delay. The pooled prevalence was 41% (95% CI 2, 89) and 28% (95% CI 21, 36), respectively.

Studies of children referred on suspicion of behavioural problems only

Mental and behaviour disorders

The prevalence of only two diagnoses was identified from evidence. One study reported on the prevalence of a behaviour problem in children and young people referred for a behaviour problem, and one on ADHD. The prevalence was 53% and 35%, respectively.

Neurodevelopmental problems

Only prevalence data for developmental disorder/delay was identified from evidence and this reported specifically on the prevalence of emotional difficulties, which was 28%.

Studies of children referred for positive screening results only

There were four studies looking at children referred after a positive result in a screening test for ASD. They each used a different screening test: Early Screening of Autistic Traits (ESAT), Young Autism and other developmental disorders Checkup Tool (YACHT-18), Checklist for Autism in Toddlers (CHAT) and Autism Spectrum Screening Questionnaire (ASSQ).

Mental and behaviour disorders

The prevalence of two diagnoses (ADHD and Tourette syndrome) was identified from evidence.

Three studies reported on the prevalence of ADHD and one on Tourette syndrome. The pooled prevalence for ADHD was 17% (95% CI 11, 23) and the prevalence of Tourette syndrome was 4%.

Neurodevelopmental problems

The prevalence of two diagnoses (a language problem and developmental disorder/delay) was identified from evidence.

Two studies reported on the prevalence of a language problem and four on developmental disorder/delay. The pooled prevalence was 24% (95% CI 17, 33) and 12% (95% CI 6, 19), respectively.

Other

One study reported the prevalence of abuse/neglect, which it reported as 26%.

6.4. Evidence to recommendations: identifying differential diagnoses

See section 6.8.

6.5. Overview of the evidence: identifying features that differentiate ASD from other conditions

No studies were identified.

6.6. Evidence profiles: identifying features that differentiate ASD from other conditions

No evidence.

6.7. Evidence statements: identifying features that differentiate ASD from other conditions

No evidence.

6.8. Evidence to recommendations: identifying features that differentiate autism from other conditions

Relative value placed on the outcomes considered

The GDG identified two outcomes to measure whether a condition is ‘important’ in the differential diagnosis of autism:

  • the prevalence of that condition in children and young people with signs and symptoms considered suggestive of autism
  • the severity of the condition.

However, there is no standard index to reflect impact, so the systematic review focussed on conditions with the highest prevalence only.

Trade-off between clinical benefits and harms

The GDG considered that identifying other conditions in the differential diagnosis of autism was an essential element of the autism-specific diagnostic assessment.

The benefit is the accurate and early recognition of alternative conditions, leading to earlier appropriate management. For example, treatment of epileptic encephalopathy might alleviate language regression and avoid ineffective treatment regimens.

The potential harm includes distress to the child or young person and/or their family or carer on being informed of another diagnosis which might be of greater concern to them than a diagnosis of autism, for example a condition associated with significant morbidity or mortality. Nevertheless, the GDG consensus was that the advantages of accurate diagnosis outweighed any disadvantages.

Trade-off between net health benefits and resource use

No evidence was identified and a health economic analysis could not be undertaken for this question due to the lack of baseline data. The costs and benefits of identifying other diagnoses during the assessment were considered by the GDG. The view was that, although there would be an additional cost associated with establishing an alternative diagnosis to autism (resources to undertake clinical review and any testing), this was likely to be cost effective compared with missing important differential diagnoses in children and young people.

Quality of evidence

Few studies were identified on the prevalence of other conditions and the quality of the evidence was low. No studies were identified that reported the severity of alternative conditions identified in children with signs and symptoms.

The grouping of conditions into categories leads to some difficulties in comparing outcomes across the available studies. Sub-group analysis by ‘reason for referral’ reduced heterogeneity. But as the confidence intervals around the prevalence estimates were very wide, interpretation of the data was difficult.

The GDG was concerned about bias in these studies due to pre-selection of samples and missing sample recruitment information. Therefore the GDG believed they did not provide credible and clinically relevant evidence on important alternative conditions. It was difficult to interpret the findings for clinical practice.

Other considerations

The GDG recognised the importance of the differential diagnosis for any individual with a developmental or behavioural concern, including those in whom autism is suspected.

The evidence produced results that were not useful in clinical practice. For example, studies of ‘abuse/neglect’ included information about attachment disorder. The GDG chose to develop a more clinically relevant list of conditions based both on the evidence and the GDG members’ knowledge and experience. The final list does not reflect the reported prevalence of the condition in the included studies as these data were not sufficiently robust but it does reflect the wide expertise of the GDG. It takes account of the prevalence data and also the severity and impact on quality of life. The list should facilitate accurate and timely recognition of conditions with a similar presentation to autism.

The GDG also developed advice on how to differentiate between alternative diagnoses with similar features (see Appendix K). The table in Appendix K is designed to enhance the implementation of the recommendation to take account of alternative conditions as part of the differential diagnosis of autism and throughout the autism pathway. For each condition the key clinical features are specified. The table shows the way that each condition typically differs from autism along with the assessments and investigations that should be undertaken. It highlights the relevant components of each assessment that contribute to the process of differentiation. The table is not the result of a systematic review of the literature but the GDG took note of the studies available in the evidence in which differentiating features were reported.

The GDG acknowledged the difficulties in differential diagnosis, as the mental and behavioural disorders and developmental disorders can, and frequently do, coexist with autism. Attachment disorders present particular challenges. In looked after children, early developmental history, which is crucial in autism diagnosis, may be difficult to obtain: re-examination over time in a different environment may clarify a diagnosis that is often dependent on experienced clinical judgement. Expertise may be required for cases such as severe hearing and visual impairment in recognising what signs and symptoms can be attributed to the sensory impairment and what falls outside that attribution. In these situations, access to expertise and tertiary opinion from other professionals is warranted.

Conditions such as epilepsy are more common in children and young people with autism and require specific treatment. Epileptic encephalopathy is a particular clinical concern if there is a history of regression of developmental skills. This has led to concern among clinicians about how to decide what tests should be done. A careful history is required, as social and language stasis and/or regression with features of autism without motor impairment or other physical features in a child under three years is typical of the regression that occurs in approximately a third of cases of autism. Language regression in a child of over three years should be referred for a medical opinion. Late autistic regression after apparently normal development (childhood disintegrative disorder [CDD]) typically includes cognitive regression, regression of bowel and bladder control and behaviour symptoms of distress and overactivity.

A child with physical symptoms and signs including seizures requires further investigation beyond the scope of this guideline.

Language delay, cognitive delay, impaired motor coordination or behavioural concerns are all common presentations of autism but are also all common neurodevelopmental problems and disorders in their own right. There is often an overlap of symptoms and individual test scores by themselves (for example language or motor coordination test scores) may not differentiate these conditions. However, the process of a professional with expertise doing such tests and considering the diagnostic features of autism will help make an accurate diagnosis.

Intellectual disability is one of the conditions that coexists most commonly with autism and is a difficult differential diagnosis in a young child. The evidence shows that the validity of the autism-specific tools for eliciting the history from an informant is limited below a mental age of 18 months (see Chapter 5). Autism diagnosis is often delayed in those with intellectual disability but distinguishing the way that a child with autism learns and communicates has important implications for future management. The particular features of coexisting autism in a child with intellectual disability may suggest an aetiological diagnosis for the intellectual disability, for example fragile X (see Chapter 7 on Coexisting conditions).

Finally, the GDG considered that disorders associated with psychosis, including schizophrenia and bipolar disorder, might be potentially important in the differential diagnosis of autism in some individuals.

In order to indentify the important differential diagnoses in each individual child or young person who has an autism diagnostic assessment, specific assessments may be required, if not already undertaken. These assessments may also help to interpret the findings of the autism-specific interview and observations (see Chapter 5).

Recommendations

NumberRecommendation
46Consider the following differential diagnoses for autism and whether specific assessments are needed to help interpret the autism history and observations:
Copyright © 2011, National Collaborating Centre for Women’s and Children’s Health.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.