Tuesday, 9 May 2017

"Medical Conditions in the First Years of Life Associated with Future Diagnosis of ASD"

I rank the paper by Stacey Alexeeff and colleagues [1] (open-access available here) as being in the 'pretty important' category when it comes to hierarchy in autism research. Not only because of their use of data derived from Kaiser Permanente (KP) (quite a large US healthcare provider that has some autism research history) including some "3911 ASD [autism spectrum disorder] cases and 38,609 controls" but also because some of the findings reiterate what is already 'known' about in relation to autism and how various medical conditions might be quite important to at least some 'types' of autism. These results are also being presented at IMFAR 2017 (see here) which will be opening its doors very soon...

The Alexeeff paper is open-access but a few choice details are worthwhile pulling out:

  • As per those participant numbers, the authors confirmed that "ASD cases were defined as children with either (a) an ASD diagnosis from an ASD specialist or (b) two or more ASD diagnoses from non-specialists, separated in time." This contrasted with controls who "were required to never have had an ASD diagnosis as of June 2012" (the period of study covered those born between 2000 and 2009).
  • The medical conditions screened for in participant records were quite wide-ranging: "Over 1000 ICD-9 codes were grouped into 79 medical conditions (e.g., constipation) within 19 domains (e.g., gastrointestinal)" and importantly, relied on physician/clinician input. Researchers ascertained how frequently said medical conditions occurred in each of the groups: "whether certain medical conditions occurring in early childhood were associated with a higher risk of future ASD diagnosis."
  • Results: "38 of the 79 medical conditions had statistically significant associations with ASD risk after multiple testing adjustment." These conditions ranged from various types/forms of developmental delay (language, learning, motor) to more somatic conditions: "nutrition, genetic, ear nose and throat, and sleep conditions." It's also worth noting that some types of diagnosis/conditions displayed only a weak association with subsequent risk of ASD including that related to asthma (previously discussed quite a bit on this blog).
  • Researchers also looked at what clusters of medical conditions might be associated with lower or higher likelihood of an ASD diagnosis being given. They concluded that: "Developmental delay and mental health condition clusters were associated with the highest relative risks of ASD." They further suggest that "combinations of medical comorbidities could aid in risk stratification for ASD prior to ASD diagnosis using a supervised clustering analysis based on machine-learning methods." Machine learning yet again, applied to autism research eh?

There's little more for me to discuss about this work other than to say that much more targeted research is needed to build on these latest findings from Alexeeff and colleagues. Coinciding with the ideas that (i) autism rarely exists in some sort of diagnostic vacuum (see here) and that (ii) autism genes are not necessarily just genes for autism (see here) (indeed a topic rising in research importance [2]) the lessons to be learned are becoming a little clearer. I might also focus a little more on those somatic diagnoses discussed in the Alexeeff paper as being quite important to the clinical picture when it comes to autism. Y'know how various medical comorbidities are indeed over-represented when it comes to autism and how appropriate screening and management/treatment [3] should be offered when and where they are identified (see here) save any further healthcare inequalities arising...

To close, I appreciate that there are quite a lot of opinions out there about a certain US autism organisation, but a recent document of theirs talking about the various conditions/comorbidities that can follow a diagnosis of autism (see here) deserves an airing. Not that they were however the first group to talk about this important part of the autism spectrum (see here) but better late than never...


[1] Alexeeff SE. et al. Medical Conditions in the First Years of Life Associated with Future Diagnosis of ASD in Children. J Autism Dev Disord. 2017 Apr 22.

[2] Diaz-Beltran L. et al. Cross-disorder comparative analysis of comorbid conditions reveals novel autism candidate genes. BMC Genomics 2017. 18: 315.

[3] Flor J. et al. Developmental functioning and medical Co-morbidity profile of children with complex and essential autism. Autism Res. 2017 May 5.


ResearchBlogging.org Alexeeff SE, Yau V, Qian Y, Davignon M, Lynch F, Crawford P, Davis R, & Croen LA (2017). Medical Conditions in the First Years of Life Associated with Future Diagnosis of ASD in Children. Journal of autism and developmental disorders PMID: 28434058