Studies on Distribution of Symptoms Between Different Age Groups from Child to Adult Life
Autism spectrum disorder (ASD) is a lifelong developmental disability with a prevalence which is hard to measure because of its chronical nature. Earlier prevalence estimates were lower, centering at about 0.5 per 1000 for autism during the 1960s and 1970s as opposed more recent reports of 1–2 per 1000, which may be related to changes in diagnostic practices, referral patterns, availability of services, age at diagnosis, and public awareness (Newschaffer et al. 2007
). Symptoms of autism include social and communication impairment and restricted repetitive and stereotyped patterns of behaviour, interests and activities. In a normal population autistic traits are relatively stable from childhood into adulthood (Taylor et al. 2017
). The triad of symptoms in people with ASD seems to decrease with age (Magiati et al. 2014
), but little is known about the distribution of symptoms between different age groups, both in clinical and non-clinical. Autism in childhood has been well described, but the symptomatology in later age groups has been little subject of investigation, especially for the group of adults with high functioning autism (HFA), or Asperger’s syndrome. It appears that symptom trajectories have considerable individual variation, and should be viewed from a developmental perspective. More at group level, adolescents and young adults may improve more in social interaction than in the Restricted, Repetitive Behaviors and Interests domain (Seltzer et al. 2003
The ideal model to explore symptom changes over a longer period of time would be a longitudinal study with valid symptom measures. No such study has been performed, which is not surprising given the efforts that it would take. A way to shed some light on this, is to perform a study with cross-sectional symptom measures in different age cohorts. This of course will not provide individual trajectories, but may increase our understanding to some extent. The purpose of this article is to describe the distribution in symptom (clusters) between the different age groups of individuals with HFA using the subscales of the autism quotient (AQ). We hypothesize that the later adult age groups respectively have more problems in repetitive and stereotyped patterns of behaviour than the domains of communication and social skills.
Only few studies investigated the distribution of symptoms of individuals with ASD into adulthood. Reviews on the outcomes of children, adolescent and adults with ASD mainly describe results from studies that are mostly conducted on cognitive and social outcomes (Howlin et al. 2014
; Levy and Perry 2011
; Magiati et al. 2014
; Seltzer et al. 2004
). Most of these studies have large variation in IQ scores and small variation in age. They mostly cover a small period in life, and are retrospective using the ADI-R (Autism Diagnostic Interview, Revised). To our knowledge, no studies have been conducted that explore the distribution of symptoms between different adult age groups, which is unfortunate since the clinical importance of providing insight differences in symptomatology between different age groups is evident. The longest follow up study by Howlin et al. (2014
), is the only study where a large population of individuals with HFA were followed into late adulthood. This study was mainly focused on cognitive development and social outcomes, but it also describes the development of symptoms after a period of 40 years using the ADI-R. Overall, this study showed a general improvement in autism symptomatology with age. A subgroup which was also assessed 20 years earlier (mean age 26) showed that social outcomes after 40 years of follow up were poorer than the assessment at younger adulthood.
Billstedt et al. (2007
) followed a population with a childhood diagnosis of ASD for a period of 13–22 years using the Diagnostic Interview for Social and Communication Disorders (DISCO). A large proportion of the included individuals suffered from intellectual disability. In that study it was found that various types of symptoms in the social interaction category were still common whereas communication problems were much less pronounced on follow-up. Behavioural impairments were much more variable in adulthood. Only one single symptom from this category, maintenance of sameness in routines, was reported to be present in half or more of the study group at the follow-up study.
Apart from these longitudinal studies, several retrospective studies have been published, with mixed designs and patient samples. In a retrospective study by Seltzer et al. (2003
) showed that adolescents are more likely to improve in reciprocal social interaction domain than adults, whereas the adults were more likely to improve in the restricted, repetitive behaviours and interests. No differences in severity of symptoms between cohorts in the communication domain were found. However, no levels of IQ were reported in this study. The authors speculated about the possibility that the developmental course of the abnormal behaviours of autism is one of abatement of symptoms 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.
Other follow up studies have been performed in childhood, including a study by Gillespie-Lynch et al. (2012
). It was found that improvements on social domain occur with increased age, but that only minor changes occur with respect to non-verbal communication and repetitive/stereotyped symptoms. Additionally, a study by McGovern and Sigman (2005
) suggested improvement in all domains, with high functioning participants showing more extensive improvements. Studies on the transition of adolescence to adulthood found some improvements in restricted repetitive behaviours (Chowdhury et al. 2010
). Others (Taylor and Seltzer 2010
) found that autism symptoms and maladaptive behaviours were generally improving with age during secondary school, but this improvement slowed down significantly after high school exit.
Overall, it remains unclear whether there are differences of symptom distribution in patients with an ASD, especially in late adulthood. A possibility to investigate this is to use the autism spectrum quotient, which is an efficient instrument for assessing and quantifying autistic traits in individuals.
Autism Spectrum Quotient (AQ)
The AQ is a questionnaire, which can be used as self-reported or reported by a close relative (AQ-adolescent and AQ-child), was originally developed to identify ASD among adults with normal intelligence (Baron-Cohen et al. 2001
). It is translated and validated in Dutch, Japanese, Polish, Australian and Canadian populations (Broadbent et al. 2013
; Hoekstra et al. 2008
; Lepage et al. 2009
; Pisula et al. 2013
; Wakabayashi et al. 2006
). The AQ contains five theoretically defined subscales of autistic behaviour; social skills, attention switching, attention to detail, communication and imagination.
The AQ consist of 50 items, 10 items per subscale. Original cut-off score (Baron-Cohen et al. 2001
) is 32 points, however this cut-off differs per study group. It also quantifies autistic traits in adolescents and children with HFA or Asperger Syndrome (AS) (Auyeung et al. 2008
; Baron-Cohen et al. 2006
; Sonie et al. 2013
). In adolescents and children the AQ is completed by a parent report. For spouses of patients the AQ has not yet been validated yet, although the AQ appears to have high face validity for such a use. One study of Wakabayashi et al. (2006
) were the AQ was reduced to a 40 item questionnaire for parents (by 32AS/HFA and parent pairs), shows a mean difference of 2.1 points (SD = 0.5), if the self-reported AQ is compared to the parent reported AQ on these 40 items.
Studies of the use of the AQ in the clinical practice show different results. Woodbury-Smith et al. (2005
) found that the AQ is strongly predictive who receives a diagnosis of ASD in adults with AS with normal intelligence and high functioning autism. The cut-off score with the best specificity and sensitivity was 26 out of a total of 50 items. In a Dutch study Sizoo et al. (2015
) found that the AQ has no sufficient validity to reliably predict a diagnosis of autism spectrum disorder in outpatient settings.
While the AQ may be less sensitive in HFA for predicting an ASD diagnosis, it appears a valuable tool for assessing and quantifying symptoms of ASD at different ages. In the present study, we therefore used the AQ to measure ASD symptoms at different age stages in order to provide the field with a better understanding of symptom distributions during the lifespan.
Also, the AQ was used to evaluate the appreciation of symptoms of ASD patients compared to their spouses.