Saturday, 14 April 2012

Significantly over-represented in autism

Comorbidity in autism spectrum conditions is of great interest to this blog. Not only because of the possible impact that certain comorbidities might have on the presentation of some cases of autism but also because, for one reason or another, certain types of comorbidity just don't seem to get either the acknowledgement or necessary resources for potential remediation that they require. As I have said many, many times, having an autism spectrum condition is not seemingly protective of having any other condition.

It was with comorbidity in mind that I was particularly interested in this study by Kohane and colleagues* (full-text) highlighting several important conditions to be noted alongside a diagnosis of autism.

The paper is open-access - everyone's favourite kind of paper - but if you want a short summary...

  • Based on an analysis of electronic health records from 2001 - 2010, data from the US Shared Health Research Informatics Network (SHRINE) system developed by some of the same authors was analysed. With a total sample size of over 2.3 million patients, data for ICD-9 diagnosed participants with an autistic disorder, Asperger syndrome or other pervasive developmental disorder under 35 years of age (N=14,381) were retrospectively mined to assess the prevalence of various comorbidity. The final total of participants is likely to be an over-estimation given that patients were not individually indexed, so if they travelled to more than one hospital included in the dataset they would be counted twice.
  • The results: based on diagnoses made via ICD-9 classification, several conditions seemed to be pretty prevalent comorbid to a diagnosis of autism. In ascending order across the total autism participant groups these included: epilepsy (19.4%), CNS/head anomalies (12.4%), bowel disorders (not including inflammatory bowel disorders) (11.7%) and schizophrenia (2.4%).
  • Comparisons with non-autism controls showed that these and most of the other comorbidities identified were significantly more frequently reported in cases of autism. Indeed, even some quite rare associations with autism such as muscular dystrophy were nonetheless still more prevalent in autism than in not-autism. 
  • When looking at age, and in particular changes to prevalence of comorbidity across age, there were a few interesting details to emerge. So, bearing in mind the p-value used, epilepsy, sleep disorders and bowel problems were statistically just as prevalent in the 18-34 years group as they were in the 0-17 years group. That being said, the older group with autism (18-34 years) showed significantly more frequent presentation with type-1 diabetes, inflammatory bowel disease and schizophrenia than the younger group. 

I'm not going to say too much more about these findings than has already been discussed on this blog. Epilepsy and seizure-type disorders are important comorbidities. Schizophrenia - well, that has also been covered quite recently from a diagnostic and potential biochemical point of view. Inflammatory bowel disease and autism, see here and here.

Other bowel disorders. There are some interesting comparisons with other datasets, including that of quite a widely discussed study on gastrointestinal (GI) issues related to autism by Black and colleagues** (full-text). They suggested that based on similar UK data, bowel problems were no more likely to occur in autism than in not-autism. Allowing for methodological caveats, it seems participant numbers might have been a factor in their findings. Certainly other studies, using different, more direct 'look-and-see' methodologies have arrived at different conclusions. Interesting also that the paper by Horvath and colleagues picked up the suggestion of issues with carbohydrate metabolism in cases of autism over 10 years before the recent Kushak paper. Who'd have thought it?

What the current study from Kohane and colleagues does reinforce is the notion that autism can and does carry comorbidity; and whilst arguments about the how and why of those comorbidities still rage in some quarters, there is a need for parents and professionals alike to be vigilant for them. Indeed, type 1 diabetes carries some pretty critical outcome if not treated.

Tracking back to the opening paragraph of this post, one has also to question what effect such comorbidities if present, might have on the presentation of autism and whether tackling some of the comorbidity might just have some surprising knock-on effects to the presentation of core symptoms and overall quality of life.

* Kohane IS. et al. The co-morbidity burden of children and young adults with autism spectrum disorders. PLoS ONE. April 2012.
DOI: 10.1371/journal.pone.0033224

** Black C. et al. Relation of childhood gastrointestinal disorders to autism: nested case-control study using data from the UK General Practice Research Database. BMJ. 2002; 325: 419-421


  1. A significant problem conceptually, is that 'autism' is seen as the primary diagnosis and the co-morbidities as, well... co-morbid. Other somatic disorders are often simply overlooked. Significant numbers of parents report allergic/intolerant responses to proteins that don't appear to involve bowel function. Many autistic children have mobility problems suggesting the involvement of connective tissue or muscles.

    But because these are seen as minor problems by diagnosing doctors, they simply aren't recorded.
    The sooner we see 'autism' as an outcome of other factors, rather than the underlying problem, the better.

  2. Many thanks for the comment logicalincrementalism.

    You make some valid points with regards to the view and impact of 'comorbidity' on cases of autism.

    The issue of connective tissue issues with regards to mobility problems is a very interesting area that has been mentioned on another post: