Journal of Child Psychology and Psychiatry 48:11 (2007), pp 1102–1110 doi:10.1111/j.1469-7610.2007.01774.

Autism in adults: symptom patterns and early

childhood predictors. Use of the DISCO in
a community sample followed from childhood
Eva Billstedt,1 I. Carina Gillberg,1 and Christopher Gillberg1,2
1
Institute of Neuroscience and Physiology, Göteborg University, Sweden; 2University of Strathclyde, Glasgow, and
NCYPE, Lingfield, UK

Background: Few studies have looked at the very long-term outcome of individuals with autism who
were diagnosed in childhood. Methods: A longitudinal, prospective, community-based follow-up study
of adults who had received the diagnosis of autism (classic and atypical) in childhood (n ¼ 105) was
conducted. A structured interview (the Diagnostic Interview for Social and COmmunication disorders –
the DISCO) was used in order to evaluate symptoms and symptom patterns 13–22 years after original
diagnosis. Childhood measures, including IQ-level at time of childhood diagnosis and communicative
speech registered before age 5 years, were studied in relation to the presence of autism symptoms at
follow-up. Results: The classical and atypical autism groups were fairly homogeneously impaired in
terms of symptoms in the social interaction category whereas other common childhood autism symp-
toms, including maladaptive and stereotyped behaviours, were more variable in the study group at
follow-up. Odd responses to sensory stimuli were still extremely common. Speech before 5 years of age,
IQ, gender, diagnosed medical disorder and onset of epilepsy before 5 years were variables that corre-
lated to outcome on the DISCO algorithm for autistic spectrum disorders (Wing & Gould, 1979) con-
cerning style and quality of social interaction, communication style and pattern of self-chosen
activities. Conclusions: Social interaction problems were still present in the vast majority of adults
with autism/atypical autism, but behavioural impairments were much more variable in adulthood.
Almost all cases were reported to show persistent perceptual problems. Certain childhood measures
were found to prospectively predict adult social interaction style, communication type, and pattern of
self-chosen activities, which still met diagnostic criteria for autism/atypical autism in adult-
hood. Keywords: Autism, the DISCO, community sample, adult outcome. Abbreviation: DISCO:
Diagnostic Interview for Social and COmmunication disorders.

Classic and severe cases of autism usually persist
into adulthood, and the majority of cases diagnosed
in childhood continue to meet criteria for the dis-
order in early adult life (Lotter, 1978; Nordin &
Gillberg, 1998; Howlin, Goode, Hutton, & Rutter,
2004; Billstedt, Gillberg, & Gillberg, 2005). Indi-
viduals affected continue to show impairments in
the typical autism triad domains of reciprocal social
interaction, reciprocal communication, and restric-
ted behaviour patterns in adulthood (Howlin, 2003).
In a telephone survey Ballaban-Gil, Rapin, Tuch-
man, and Shinnar (1996) found high rates of
behaviour problems among 45 adults who had been
given a diagnosis of autism in childhood. McGovern
and Sigman (2005) followed 48 individuals dia-
gnosed with autism in childhood to adolescence. All
but two still met criteria for autism. In contrast to
the continuity in diagnosis, parents described
improvements in social interaction, repetitive/
stereotyped behaviours, adaptive behaviours and
emotional responsiveness to others in adolescence
compared to middle childhood. This improvement
was more evident in high-functioning adolescents
with autism than in low-functioning individuals.
Conflict of interest statement: No conflicts declared.
Ó 2007 The Authors
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
Seltzer stated that symptoms of autism appear to be
different at different stages of life (Seltzer et al.,
2003).
Autism in childhood has been well described.
However, symptoms typical of the infant and pre-
school child with autism may have taken on different
qualities or may no longer be evident at all in the
young adult person with autism.
We have conducted a follow-up study into early
adulthood of 120 individuals diagnosed with autism
in childhood (recruited from three population stud-
ies). There was considerable stability over time as
regards the diagnosis of autism per se (Billstedt
et al., 2005). The present paper looks at distinctive
symptoms and symptom patterns within the broader
autism spectrum in this population-based cohort.
The main source of information was a semi-struc-
tured interview with an informant who knew the
person with autism well at age 16–40 years. The
interview used was the Diagnostic Interview for So-
cial and COmmunication disorders (DISCO) (Wing,
Leekam, Libby, Gould, & Larcombe, 2002), which
was chosen because it is particularly effective for
diagnosing disorders within the broader ‘autism
spectrum’ (Leekam, Libby, Wing, Gould, & Taylor,
2002) and because it yields more information about
 Autism in adults: symptom patterns and early childhood predictors
non-autism problems than most other instruments
currently in use.
The aim of the study is to present a descriptive
picture of the most common symptoms in adults
with autism as measured by the DISCO, to find out
to what extent the DISCO-algorithm and clinical
diagnosis correlate, and to find out whether or not
certain childhood measures, derived from the ori-
ginal studies, were associated with specific symp-
toms within the autism spectrum in adult age.
Method
The original studies
Our study group, born in 1962–84, was recruited from
three population-based studies from Sweden of indi-
viduals diagnosed with autism in early childhood (Gill-
berg, 1984; Steffenburg & Gillberg 1986; Gillberg,
Steffenburg, & Schaumann, 1991). All 120 individuals
(84 males, 36 females) were followed up in late adoles-
cence–adult life (Billstedt et al., 2005). The original
studies of young children comprised participants with
DSM-III-R (APA, 1987) autistic disorder/infantile aut-
ism (61 males, 17 females) and 42 with atypical aut-
ism/autistic-like condition (23 males, 19 females). The
group with atypical autism/autistic-like conditions had
met 6 or more of the 16 – but not full – DSM-III-R
criteria for autistic disorder.
The follow-up study
The individuals were followed prospectively for a period
of 13–22 years (mean age 17.8 years, 3.6 SD) and re-
evaluated at ages 17–40 years (mean age 25.5 years,
6.4 SD). Six families declined participation in the fol-
low-up study and six individuals had died at the time of
follow-up (Billstedt et al., 2005). Three of the remaining
108 individuals were included in the follow-up study,
but did not participate in the current DISCO-study.
Clinical diagnoses at follow-up
At follow-up, individuals who met criteria from at least
two of the three domains of the triad of social, com-
munication and imagination/behavioural impairments
received autism spectrum disorder (ASD) diagnoses
after clinical assessment. ASD cases were then subdiv-
ided into three groups: (1) those with autistic disorder/
childhood autism who met criteria for this diagnosis
according to the DSM-IV (APA, 1994)/ICD-10 (WHO,
1993), (2) those with atypical autism according to the
ICD-10 in cases showing 4 or more of the 12, but not
full DSM-IV/ICD-10 diagnostic criteria for autistic dis-
order/childhood autism, and (3) those with other ASD
not meeting criteria for autistic disorder or atypical
autism. In addition, the diagnostic criteria for Asperger
syndrome by the Gillbergs (Gillberg & Gillberg, 1989)
were also checked. Clinical diagnoses were made con-
jointly by the first (E.B.) and second (I.C.G.) authors of
this paper, who were blind to the original specific
childhood diagnosis and symptomatology, based on all
available information from psychiatric-psychological-
Ó 2007 The Authors
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
1103
medical assessment. The third author (C.G.) had
examined all cases in childhood and was aware of their
childhood diagnosis but did not take part in the
assessment in the follow-up study.
Participants
One hundred and five individuals (75 males, 30
females) with childhood diagnoses of autistic disorder/
infantile autism (56 males, 15 females) or autistic-like
conditions/atypical autism (19 males, 15 females) were
included in this follow-up study. At follow-up, 89 of the
105 participants were clinically diagnosed as having
autistic disorder. Fifteen individuals were diagnosed as
having atypical autism and one individual did not meet
criteria for an ASD.
Intellectual functioning
All participants had undergone testing before 10 years
of age. Only one-fourth of all individuals included were
able to take a complete IQ-test (Wechsler, 1981, 1992)
at follow-up. However, all participants were assessed
using the Vineland Adaptive Behavior Scales (VABS;
Sparrow, Balla, & Cicchetti, 1984) allowing a reason-
able estimate of whether or not an individual should be
diagnosed as belonging in the severe learning disability
(SLD) category (see Billstedt et al., 2005). In the autism
group, SLD was diagnosed in 50 cases (70%), mild
learning disability (MLD) in 17 cases (24%), near-aver-
age intelligence (NA) in 2 cases (3%) and average (A)
intelligence in 2 cases (3%). In the atypical group, SLD
was diagnosed in 26 cases (76%), MLD in 7 cases (21%)
and A in 1 case (3%).
The DISCO
The DISCO is a 2–4-hour standardised interview
intended for interviewing a person who has known the
individual with suspected autism spectrum disorder
from early childhood. The purpose of using it is to assist
clinicians in the diagnostic work-up and care of persons
with autism spectrum and other developmental dis-
orders affecting social interaction and communication
(Wing et al., 2002). Its algorithm enables one to
investigate whether the necessary criteria of different
diagnostic systems are met. The diagnostic systems
that are used are ICD-10 (WHO, 1993), DSM-III-R (APA,
1987) and DSM-IV (APA, 1994). Algorithms based on
Kanner’s ‘Early Infantile Autism’ (Kanner & Eisenberg,
1956), Gillberg and Gillberg’s definition of Asperger
Syndrome (1989), Wing & Gould’s ‘Autistic Spectrum
Disorder’ (ASD; 1979), and ‘Social Impairment’ (1979)
are defined and used.
From the DISCO one obtains an ‘ever’ diagnosis,
which is retrospective, as well as a ‘current’ diagnosis
(‘ever’ includes ‘current’). During the DISCO interview,
responses to more than 300 questions are coded for
computer entry. The questions are divided into three
groups pertaining to: (i) current level of function, (ii)
delay in acquiring relevant skills, and (iii) untypical
behaviours associated with the relevant skills. In this
paper the ‘current’ responses on untypical behaviour
are presented. A few items were not relevant to a larger
1104 Eva Billstedt, I. Carina Gillberg, and Christopher Gillberg
part of the group and were therefore not included in this
study. Included were 22 social interaction items, 17
communication items, 28 routines/stereotyped beha-
viour items, 20 sensory items, 26 emotional problem/
maladaptive behaviour items, and 19 items concerning
self-care, motor skills and activity. Problems were rated
as ‘marked’, ‘minor’ or ‘no problems’.
Other scales/measures used
The DSM-III-R Global Assessment of Functioning scale
(GAF; American Psychiatric Association, 1987), Lotter’s
criteria (1978) for the classification of outcome and a
psychiatric-medical assessment were also included.
The results from these measures have been reported
elsewhere (Billstedt et al., 2005).
Procedure
The DISCO interviewees were parents or other close
family members (grandparents, siblings or foster-par-
ents) in 91 cases. In the remaining 14 cases the inter-
viewees were professional caregivers (i.e., teachers, staff
in day-centre activities or group homes). The interviews
were conducted either in an outpatient clinic, in the
home of the individual with ASD or in the home of the
interviewee.
Study of adult symptoms and their links to certain
childhood measures
For the study of possible associations between some
specific ASD symptoms and certain childhood meas-
ures, the following approach was used. The algorithm
for Wing and Gould’s ASD is based on 7 items. All these
items are investigator-based ratings taking into account
all the information obtained from the whole DISCO
interview. One concerns the current quality of social
interaction (aloof, passive, active but odd, over-formal,
shy, electively mute, chatty/friendly (typical of Williams
syndrome) or appropriate). A second item concerns the
individual’s style of communication (communicating
only needs, one-sided and repetitive, or truly recipro-
cal). A third item concerns the current overall pattern of
activities (narrow and repetitive or varied and flexible).
The remaining 4 items are related to imagination and
pretence. Many parents/caregivers were able to provide
only partial information for these 4 items, and they were
therefore not included in the study of childhood meas-
ures. Thus, only the three first (social, communication
and pattern of activities) items of the DISCO ASD
algorithm were correlated to childhood measures in this
study. The childhood measures (derived from the ori-
ginal studies) were: (1) gender, (2) childhood IQ/DQ-
level (categorised in SLD, MLD, NA and A categories), (3)
presence of some communicative speech before 5 years
of age (as noted in childhood records), (4) epilepsy (on-
set before 5 years of age and onset after 5 years of age
up until one year before follow-up as evidenced by
recording in childhood files, see Danielsson, Gillberg,
Billstedt, Gillberg, & Olsson, 2005), (5) medical dis-
orders regardless of presence or not of epilepsy (as
detailed in medical records at the time of early child-
hood population studies), (6) childhood ASD diagnosis,
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
and (7) immigrant parents (at least one parent born in
foreign country).
Statistics
Spearman’s correlation was used to investigate the
relationship between DISCO items and symptoms. We
then used stepwise regression techniques to evaluate
the relative contribution of each background factor to
outcome in terms of specific DISCO ASD algorithm
symptoms. We used stepwise regression in the SPSS
statistical package, where 6 predictors were entered
consecutively to maximise each of the 3 autistic spec-
trum algorithm items. Probability level for entry or
exclusion was .05.
Ethics
The study was approved by the Medical Ethical
Committee of Göteborg University.
Results
DISCO-10 diagnoses in the study group at follow-up
Eighty-two of 89 (92%) in the adult clinical autistic
disorder group and 3 of 15 individuals (20%) in the
adult clinical atypical autism group met DISCO-
algorithm criteria for autistic disorder. Seven indi-
viduals in the adult clinical autistic disorder group
(8%) and 12 individuals (80%) in the adult clinical
atypical autism group met the DISCO-algorithm
criteria for atypical autism. In the autistic disorder
group, 15 (17%) met also Kanner’s criteria for early
infantile autism according to the DISCO, and an-
other 13 individuals (17%) met Gillberg and Gill-
berg’s criteria for Asperger syndrome (1989). In the
atypical autism group, 1 man (7%) met the criteria
for Asperger syndrome according to Gillberg and
Gillberg (1989) and none in this group met Kanner’s
criteria for early infantile autism. All individuals met
criteria for Wing and Gould’s definition for Social
Impairment and ASD.
Specific symptoms
Given that, at the symptom level, there were few
differences across clinical diagnostic groups, the
autistic disorder and atypical autism groups were
collapsed and referred to as the ASD group. Table 1
shows all the untypical social interaction, commu-
nication and behaviour symptoms that were pre-
sent in half or more of the study group.
Percentages are calculated on the basis of results
obtained in 98–105 individuals (missing data in a
few cases for certain items). A detailed presentation
of all DISCO symptoms is published in an online-
only Appendix.
Social interaction. Overall rates of abnormal social
behaviour were high. No interaction with age peers or
Ó 2007 The Authors
 Autism in adults: symptom patterns and early childhood predictors 1105

Table 1 The untypical social interaction, communication and behaviour symptoms that were present in half or more of the study
group

DISCO symptoms (n ¼ 98–105) Current% Marked/minor%

Social interaction1
No interaction or inappropriate quality of interaction 92 75/17
One-sided social approaches or no approach at all 90 68/22
No interaction with age peers/interacts only if led by other 90 65/25
No or inappropriate conventions of peer interaction 88 83/5
Lack of or inappropriate emotional response to age peers 84 72/12
No or inappropriate sharing of interest and enjoyment 80 56/24
No adequate giving of comfort to others 79 63/16
Avoidance of age peers 78 44/34
Behaviours indicating no awareness of others’ feelings 77 54/23
No adequate seeking of comfort when in pain/distress 72 59/13
No appropriate response to physical affection 70 20/50
No or inappropriate reaction to others’ happiness 58 37/21
Tends to have a blank, unfocussed gaze 51 24/27
Poor or no eye contact 50 19/31
No adequate response to visitors 50 18/32
Verbal and non-verbal communication2
Little, if any reciprocity in verbal communication 83 47/36
Unable to smile natural on request 74 36/38
Limited or odd facial expression 67 43/24
Abnormal tone of voice 53 29/24
Routines, resistance to change, stereotyped movements and vocalisations/verbalisations3
Maintenance of sameness in routines 62 32/30
Responses to sensory, auditory and visual stimuli4
Unusual reaction to firm touch 60 33/27
Indifference to pain, heat, cold 58 33/25
Acuteness of hearing 53 39/14
Self-injury 51 27/24
Emotional problem and maladaptive behaviour5
Laughs for no reason 67 34/33
Needs constant supervision 59 37/22
Lack of cooperation 59 29/30
No psychological barriers 55 33/22
Temper tantrums 50 16/34
Self-care, motor skills, and activity6
Lack of awareness of suitability of clothing 75 64/11
Poor coordination in PE and games 67 31/36
Immature gait when walking 56 35/21
1
DISCO, part 3, section xii-C, xiii-C, xiv-C; 2DISCO, part 3, section ix, x and xi; 3DISCO, part 4, section i, and v; 4DISCO, part 4
section, ii, iii and iv; 5DISCO, part 5, section I, and part 6, section I; 6DISCO, part 3, section ii-C, iii-C, iv-C, and part 4, section xi.

only interaction if led by other was very common. The
quality of the interaction was, when it occurred,
inappropriate, and appropriate conventions were
often missing. Inappropriate (or lack of) emotional
response to age peers was also common. One-sided
social approaches or no approach at all was reported
in a large majority of the group. About 70% in the
group of individuals with ASD disliked, tolerated or
enjoyed, but did not reciprocate physical expressions
of affection. Poor eye contact was reported in half of
the group and half the group tended to have a blank,
unfocussed gaze.
Verbal and non-verbal communication. The most
common problem in communication in the ASD
group was the lack of reciprocity and odd or limited
non-verbal communication. Facial expression was
judged to be inappropriate or unvaried in more than
half of the study group. Lack of intonation and lack of
prosody in speech was often reported. When smiling
Ó 2007 The Authors
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
in request, a large group was judged to do so in an
unnatural way.
Routines, resistance to change and stereotyped
movements and verbalizations. The only symptom
in this category seen in over half the sample was the
concern about maintaining the sameness of routines.
Responses to sensory, auditory and visual stimuli,
and self-injury. A majority of the study group (93%)
was reported to have some kind of impairing sensory
abnormality. Various types of aberrant responses to
touch were the most common sensory abnormalities
reported. Sixty-six percent of the study group were
reported to have abnormal responses to auditory
stimuli, and 45% showed abnormal responses to
visual stimuli. Specific abnormal responses that
affected more than half of the study group were dis-
like of being held firmly (and almost half of all indi-
viduals reacted in unexpected ways to gentle touch),
1106 Eva Billstedt, I. Carina Gillberg, and Christopher Gillberg

indifference to pain, heat and cold and acuteness of
hearing. Self-injurious behaviours were seen fre-
quently or occasionally in half of the group.
Maladaptive behaviour. A majority of the study
group was reported to laugh for no obvious reason.
The need for constant supervision and lack of co-
operation was reported being a major problem in
more than half of the ASD group. Unawareness of
psychological barriers and temper tantrums were
common.
Self-care and motor skills. A large majority of the
study group had only a limited number of activities of
their own choosing, or no interests at all. Lack of
awareness of needing proper clothes according to
weather or social occasion was a common problem.
Poor coordination in Physical Education (PE) and
games was common as well as a general immature
gait when walking.
Childhood predictors of adult quality of social
interaction
Speech before 5 years of age (prospective record
data) and higher childhood IQ showed a strong cor-
relation with the DISCO-ASD-algorithm item for
better quality of social interaction at follow-up
(Table 2). There were also weaker correlations be-
tween early onset epilepsy (onset before 5 years of
age) and female gender on the one hand and poorer
quality of social interaction on the other. However, in
the stepwise regression analysis (Table 3), only
speech before 5 years, IQ and gender remained as
significant predictors of social interaction. The
presence of other medical disorders, childhood
diagnosis (typical/atypical autism) or having immi-
grant parents were not associated with the quality of
social interaction in adulthood either in the correla-
tional or stepwise analyses.
Childhood predictors of adult reciprocal
communication
The strongest univariate childhood predictors for
outcome as regards the DISCO-ASD-algorithm item
for reciprocal communication were speech before
5 years of age and higher childhood IQ. Early-onset
epilepsy also correlated (negatively) on this item but
to a lesser degree. In the stepwise analysis other
medical disorders emerged as a negative predictor
together with speech before 5 years (positively).
Gender, childhood diagnosis and immigrant parents
were not associated with adult reciprocal communi-
cation.
Childhood predictors of limited adult pattern of
self-chosen activities
Speech before 5 years of age and higher childhood
IQ were the factors that correlated significantly
with the DISCO-ASD-algorithm item limited pattern
of self-chosen activities both according to Spear-
man’s correlations and stepwise analysis. The
remaining childhood predictors (gender, epilepsy,
other medical disorder, childhood diagnosis and
immigrant parent) were not associated with this
item.

Table 2 Spearman’s correlations of Wing and Gould’s autistic spectrum algorithm items and various predictors

Quality of social Reciprocal Limited pattern of self-chosen

Symptoms interaction (n ¼ 105) communication (n ¼ 105) activities (n ¼ 102)

IQ/DQ .57** .51** .36**

Female gender ).23* ).06 .03
Epilepsy onset <5 years ).20* ).23* ).09
Epilepsy onset >5 years ).08 ).03 ).03
Other medical disorders ).05 ).17 .10
Speech before 5 years .51** .58** .37*
Typical autism ).12 ).14 ).05
Immigrant parents .03 .00 .04

* ¼ <.05, ** ¼ <.01.

Table 3 Childhood factors predicting Wing and Gould’s autistic spectrum algorithm items

Social interaction B p-value R-square ¼ .45

IQ/DQ .82 .001
Gender ).74 .003
Speech before 5 years .83 .003
Reciprocal communication B p–value R-square ¼ .39
Speech before 5 years 1.06 .001
Other medical disorders ).42 .008
Limited pattern of self-chosen activities B p–value R-square ¼ .43
IQ/DQ .27 .03
Speech before 5 years .40 .04

Ó 2007 The Authors

Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
 Autism in adults: symptom patterns and early childhood predictors
Discussion
Symptoms and symptom patterns
All the participants in this follow-up study in late
adolescence and young adulthood, of a population
cohort of individuals diagnosed in early childhood
with autism or atypical autism, still showed impair-
ment related to their ‘autism features’. Autism/
atypical autism is diagnosed in individuals, whether
children or adults, who manifest a range of symp-
toms that can be subgrouped under the categories
currently believed to constitute separate portions of
the ‘triad of impairments of social interaction, com-
munication, and imagination/behaviour’ (Wing,
1988). Almost without fail, the adolescent and young
adult individuals included in the present study still
met clinical DSM-IV diagnostic criteria for autism (a
large majority) or atypical autism (a minority) at fol-
low-up, and all adults with a clinical diagnosis of
autism/atypical autism met DISCO-thresholds for
autism/atypical autism. This means that the vast
majority showed above-threshold symptoms in the
three triad domains. However, the detailed analyses
of symptom patterns performed in the present study
showed that there was a great degree of variability of
symptoms across subjects.
First of all, various types of symptoms in the social
interaction category were still extremely common.
Fifteen of 22 symptoms in this category were repor-
ted to still be present in half or more of the study
group, underscoring the pervasiveness of the social
reciprocity problems that have been characteristic of
the group from infancy through childhood and
adolescence into adult life.
Secondly, in the communication category, only 4
of the 17 symptoms included in the category affected
half or more of the study group. The most prevalent
current symptom from the communication domain
was the limitation of reciprocity in communication
(clearly difficult to separate from ‘reciprocity in social
interaction’). Several communication impairment
symptoms, including echolalia, typical of young
children, were reported to be much less prevalent
when these children had grown up.
Thirdly, the final category in the triad, the beha-
vioural/imagination domain, comprises a variety of
different impairing symptoms such as handicapping
routines, resistance to change, stereotyped move-
ments and stereotyped vocalisations/verbalisations.
Only one single symptom from this category, main-
tenance of sameness in routines, was reported to be
present in half or more of the study group at the
follow-up study.
Fourth, symptoms not belonging to the triad but
very often present in young individuals with autism
are unusual responses to sensory (e.g., touch, pain,
auditory, and visual) stimuli. Such problems were
still almost pervasively present in adult age. The
finding of an extremely high proportion in the autism
group still showing impairing problems as regards
Ó 2007 The Authors
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
1107
responses to sensory stimuli in adulthood has also
recently been reported by Wing and her group
(Leekam, Nieto, Libby, Wing, & Gould, 2006).
Fifth, various types of ‘non-core autism’ emotional
problems and maladaptive behaviours are often
present in young children with autism. The variation
in occurrence of such symptoms in adulthood was
very considerable, with only a few symptoms in each
category affecting half or more of the study group.
Sixth, one of the most striking findings of the study
was the continued severe limitation of self-chosen
activities, often manifested in only engaging in
repetitive activities. This, and prevalent problems
reported on several other items of the DISCO-cate-
gory ‘self-care, motor skills, and activity’, stresses
the complexity of the disabilities subsumed under
the label of autism and the need for individualised
support for adolescent and adult people with autism
and their families/carers.
Seltzer and colleagues (2003) described and com-
pared symptoms of autism spectrum manifested in an
adolescent (contemporary diagnostic definitions
used) and an adult cohort (more narrow diagnostic
criteria used). Data were collected using the Autism
Diagnostic Interview-Revised (ADI-R; Lord, Rutter, &
LeCouteur, 1994) to assess the pattern of autism
symptoms in adolescence and adulthood. As in our
study, the majority of the sample continued to mani-
fest symptoms of autism that met the diagnostic
threshold. The adult cohort appeared to be less
severely impaired than the adolescent cohort in the
manifestation of abnormal behaviour, including ver-
bal symptoms and many restricted, repetitive
behaviours. The authors speculated about the pos-
sibility that the developmental course of the abnormal
behaviours of autism is one of abatement of symp-
toms from adolescence into adulthood. On the other
hand, their adolescent cohort appeared to be less
impaired than the adult cohort in the manifestation of
prosocial behaviours, such as communication and
social interaction. Our findings partly support (e.g.,
abatement of abnormal symptoms, including verbal
symptoms and many restricted, repetitive behav-
iours), and partly contradict (e.g., generally large
improvements across domains of functioning) the
conclusions drawn by Seltzer’s group.
Childhood measures
The two variables that have consistently been iden-
tified as early predictors of outcome in autism are IQ
and language (Lotter, 1974; Gillberg & Steffenburg,
1987; Nordin & Gillberg, 1998). The present study
confirmed the importance of both these factors, and
particularly underscored the salience of speech be-
fore 5 years of age as a most important predictor of
outcome as regards all three of the ‘core’ childhood
autism features in adulthood.
Epilepsy, with an onset before 5 years of age,
correlated to outcome in the area of the quality of
1108 Eva Billstedt, I. Carina Gillberg, and Christopher Gillberg
social interaction (and reciprocal communication).
However, in the stepwise procedure, this factor ‘dis-
appeared’ and did not account for a significant pro-
portion of the total outcome variance. Instead, in the
stepwise analysis, ‘medical disorder’ (often associ-
ated with epilepsy) was found to account for a sig-
nificant proportion of the variable in terms of poor
outcome of reciprocal communication in early adult
life. The generally poorer level of intellectual func-
tioning as a common denominator for epilepsy and
medical disorders might hold the key to explaining at
least part of the ‘predictive’ power of these variables.
If intellectual functioning was left out of the stepwise
regression model, the predictive power came down to
R2 ¼ .32 (from .45).
Female gender indicated greater abnormalities in
the area of social interaction. Associations between
gender and reciprocal communication and between
gender and limitation in self-chosen activities were
not found. This could be – at least partly – explained
by Wing’s (1981) theory – further developed by Gill-
berg (1992) – that girls who qualify for a full dia-
gnosis of autism may have more severe brain
dysfunction than boys at similar level of diagnostic
symptoms. The females in the present study more
often had epilepsy than did the males (Danielsson
et al., 2005). Those with more severe brain dysfunc-
tion in childhood would – obviously – seem to be at
higher risk of poor outcome in adult life than would
those with less pronounced brain dysfunction.
Methodological considerations and generalisability
of findings
How generalisable are these findings to other cohorts
of individuals with autism? Our study group was
diagnosed and delineated several decades ago when
the concept of autism was narrower than it is now.
Nevertheless, strict operationalised diagnostic cri-
teria were used when assigning diagnoses of autism/
atypical autism. However, it is likely, indeed highly
probable, that – with the methodology used at the
time (mostly the 1970s and the 1980s) – many
individuals with autistic symptoms and IQ-test re-
sults within the normal range would have been
excluded as falling outside the autism category in the
original study. Our results are therefore probably
only generalisable to the low- and middle-function-
ing groups with autism rather than the high-func-
tioning group. Nevertheless, in some areas of
functioning, particularly those related to impairment
of reciprocity, almost identical problems are seen
both in childhood and adulthood in low- and high-
functioning individuals with ASD – including those
with Asperger syndrome. This suggests that the
developmental trajectory in autism is partly inde-
pendent of IQ (Howlin et al., 2004; Cederlund &
Gillberg, 2004). Even in our own small subgroup of
relatively high-functioning individuals with autism
some symptoms were just as frequent as in the
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
low-functioning subgroups. Examples were odd or
limited facial expressions, none or only a limited
number of activities of own choice, and in-
appropriate behaviours indicating limitations in
awareness of the feelings of others. A major finding of
the study was that the study group was more
homogeneous concerning different symptoms in the
social interaction category, such as no interaction
or inappropriate quality of interaction and/or one-
sided social approaches, whereas other typical
autistic symptoms, such as maladaptive and ste-
reotyped behaviours, were much more variable. The
fact that sensory problems were almost pervasive in
adulthood raises the important issue as to whether
or not such problems (which were almost universally
present in childhood also) should be considered as
more ‘core’ than other ‘typical’ behaviour problems
for the delineation of the syndrome of autism.
Clinical implications
It is essential to compile a thorough clinical history
and, to cover childhood development and symptoms
in detail in order to be able to properly assess the
meaning of ‘outcome’ for an individual with ASD.
Several typical childhood autistic traits may have
disappeared or have become toned down in adult-
hood, leaving core autism features, such as limited
reciprocity in communication and social interaction
(and possibly aberrant responses to sensory stimuli),
as well as a very restricted range of self-chosen
activities as the main characteristics. The high rate
of persistent abnormal responses to sensory stimuli
needs to be taken into account when planning for an
autism-friendly environment.
It is possible that relatively higher IQ and relatively
early onset of some communicative speech in autism
may still be the best predictors of a better outcome,
but much more needs to be done in terms of quality-
of-life research in the field before clinical implica-
tions of these correlates can be fully understood. At
the present time it would appear that it is still
impossible to separate out which of the early child-
hood factors autism, intellectual level and verbal
ability have the best predictive power for adult
symptom outcome.
In summary, this study provides detailed infor-
mation about the symptomatic manifestations of
autism in adulthood. It demonstrates the very per-
sistent nature of the social-communication recipro-
city problems experienced by individuals with
autism, and their need for help and support in these
areas. The practical implications of this for child and
adult psychology and psychiatry, educational and
occupational services are considerable. It is import-
ant to recognise the many different behaviour pat-
terns in the field of autism in adults in order to
interpret and understand the ‘meaning’ of the
behaviour as correctly as possible and to provide
Ó 2007 The Authors
 Autism in adults: symptom patterns and early childhood predictors
correct support. In developing intervention ap-
proaches it is important to take account of the
characteristic behavioural patterns of the individual,
even in cases that can be reliably ‘classified’ as
belonging in this or that diagnostic category.
Supplementary Material
The following supplementary material is available for
this article:
Appendix. DISCO symptoms (Word document)
This material is available as part of the online ar-
ticle from:
http://www.blackwell-synergy.com/doi/full/
10.1111/j.1469-7610.2007.01774.x
(This link will take you to the article abstract).
Please note: Blackwell Publishing are not re-
sponsible for the content or functionality of any
supplementary materials supplied by the authors.
Any queries (other than missing material) should be
directed to the corresponding author for the article.
Acknowledgements
The authors wish to thank the individuals and their
families affected by autism for participation in this
study, and to acknowledge the support of the
Wilhelm and Martina Lundgren Foundation and the
Swedish Science Council (grant no. 2003–4581).
Correspondence to
Eva Billstedt, Göteborg University, Department of
Child and Adolescent Psychiatry, Kungsgatan 12, SE
411 19 Göteborg, Sweden; Tel: 0046 31 701 25 43,
Fax: 0046 31 701 25 45; Email: eva.billstedt@
vgregion.se
References
American Psychiatric Association. (1987). Diagnostic
and statistical manual of mental disorders (3rd edn,
rev.). Washington, DC: American Psychiatric Associ-
ation.
American Psychiatric Association. (1994). Diagnostic
and statistical manual of mental disorders (4th edn).
Washington, DC: American Psychiatric Association.
Ballaban-Gil, K., Rapin, I., Tuchman, R., & Shinnar, S.
(1996). Longitudinal examination of the behavioural,
language, and social changes in a population of
adolescents and young adults with autistic disorder.
Pediatric Neurology, 15, 217–223.
Billstedt, E., Gillberg, I.C., & Gillberg, C. (2005). Autism
after adolescence: Population-based 13–22 year
follow-up study of 120 individuals with autism
diagnosed in childhood. Journal of Autism and Devel-
opmental Disorders, 35, 351–360.
Cederlund, M., & Gillberg, C. (2004). One hundred
males with Asperger syndrome: A clinical study of
Ó 2007 The Authors
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
1109
background and associated factors. Developmental
Medicine and Child Neurology, 46, 652–660.
Danielsson, S., Gillberg, I.C., Billstedt, E., Gillberg, C.,
& Olsson, I. (2005). Epilepsy in young adults with
autism: A prospective population-based follow-up
study of 120 individuals diagnosed in childhood.
Epilepsia, 46, 918–923.
Gillberg, C. (1984). Infantile autism and other childhood
psychoses in a Swedish urban region. Epidemiological
aspects. Journal of Child Psychology and Psychiatry,
25, 35–43.
Gillberg, C. (1992). The Emanuel Miller Memorial
Lecture 1991. Autism and autistic-like conditions:
Subclasses among disorders of empathy. Journal of
Child Psychology and Psychiatry, 33, 813–842.
Gillberg, C., & Gillberg, I.C. (1989). Asperger syndrome
– some epidemiological considerations: A research
note. Journal of Child Psychology and Psychiatry, 30,
631–638.
Gillberg, C., & Steffenburg, S. (1987). Outcome and
prognostic factors in infantile autism and similar
conditions: A population-based study of 46 cases
followed through puberty. Journal of Autism and
Developmental Disorders, 17, 273–287.
Gillberg, C., Steffenburg, S., & Schaumann, H. (1991).
Is autism more common now than ten years ago?
British Journal of Psychiatry, 158, 403–409.
Howlin, P. (2003). Can early interventions alter the
course of autism? Novartis Foundation Symposia,
251, 250–259.
Howlin, P., Goode, S., Hutton, J., & Rutter, M. (2004).
Adult outcome for children with autism. Journal of
Child Psychology and Psychiatry, 45, 212–229.
Kanner, L., & Eisenberg, L. (1956). Child psychiatry:
Mental deficiency. American Journal of Psychiatry,
112, 531–534.
Leekam, S., Libby, S., Wing, L., Gould, J., & Taylor, C.
(2002). The Diagnostic Interview for Social and
Communication Disorders: Algorithms for ICD-10
childhood autism and Wing and Gould autistic
spectrum disorder. Journal of Child Psychology and
Psychiatry, 43, 327–342.
Leekam, S.R., Nieto, C., Libby, S.J., Wing, L., & Gould,
J. (2006). Describing the sensory abnormalities in
children and adults with autism. Journal of Autism
and Developmental Disorders, 3 (epub ahead of print).
Lord, C., Rutter, M., & LeCouteur, A. (1994). Autism
diagnostic interview-revised: A revised version of a
diagnostic interview for caregivers of individuals with
possible pervasive developmental disorders. Journal
of Autism and Developmental Disorders, 24, 659–
685.
Lotter, B. (1974). Factors related to outcome in autistic
children. Journal of Autism and Childhood Schizo-
phrenia, 4, 263–277.
Lotter, V. (1978). Follow-up studies. In M. Rutter, & E.
Schopler (Eds.), Autism. A reappraisal of concepts and
treatment (pp. 475–495). New York: Plenum Press.
McGovern, C.V., & Sigman, M. (2005). Continuity and
change from early childhood to adolescence in aut-
ism. Journal of Child Psychology and Psychiatry, 46,
401–408.
Nordin, V., & Gillberg, C. (1998). The long-term out-
come of autistic disorders. European Child and
Adolescent Psychiatry, 97, 99–108.
1110 Eva Billstedt, I. Carina Gillberg, and Christopher Gillberg
Seltzer, M.M., Krauss, M.W., Shattuck, P.T., Orsmond,
G., Swe, A., & Lord, C. (2003). The symptoms of
autism spectrum disorders in adolescence and
adulthood. Journal of Autism and Developmental
Disorders, 33, 565–581.
Sparrow, S., Balla, D., & Cicchetti, D. (1984). Vineland
Adaptive Behavior Scales. Circle Pines, MN: American
Guidance Service.
Steffenburg, S., & Gillberg, C. (1986). Autism and
autistic-like condition in Swedish rural and urban
areas: A population study. British Journal of Psychi-
atry, 149, 81–87.
Wechsler, D. (1981). Wechsler Adult Intelligence Scale-
Revised. San Antonio, TX: Psychological Corporation.
Wechsler, D. (1992). Wechsler Intelligence Scale
for Children–Third Edition. London: Psychological
Corporation.
Wing, L. (1981). Sex ratios in early childhood autism
and related conditions. Psychiatry Research, 5, 129–
137.
Journal compilation Ó 2007 Association for Child and Adolescent Mental Health.
Wing, L. (1988). The continuum of autistic character-
istics. In E. Schopler, & G. Mesibov (Eds.), Diagnosis
and assessment in autism (pp. 91–110). New York:
Plenum.
Wing, L., & Gould, J. (1979). Severe impairments of
social interaction and associated abnormalities in
children: Epidemiology and classification. Journal of
Autism and Developmental Disorders, 9, 11–29.
Wing, L., Leekam, S., Libby, S., Gould, J., & Larcombe,
M. (2002). The Diagnostic Interview for Social and
Communication Disorders: Background, inter-rater
reliability and clinical use. Journal of Child Psycho-
logy and Psychiatry, 43, 307–325.
World Health Organization. (1993). The ICD-10 classi-
fication of mental and behavioural disorders; dia-
gnostic criteria for research. Geneva: World Health
Organization.
Manuscript accepted 29 March 2007
Ó 2007 The Authors