Friday, 22 April 2016

Developmental regression in autism affects screening results

In today's short post I'd like to bring the findings reported by Lotta Höglund Carlsson and colleagues [1] to your attention and a reminder that developmental regression accompanying autism onset is an important feature for quite a few people.

With the aim of looking at the "national, routine 18-month developmental surveillance at Child Healthcare Centres (CHC) on children later diagnosed with autism spectrum disorder (ASD)" in Stockholm County, Sweden, authors reported on the results of said surveillance for those (N=175) diagnosed with autism in terms of those who initially passed this screen and what might be done to increase it's diagnostic utility. Some interesting data emerged.

We are told that over a third of the autism group "did not pass the required number of items" included in the screen and this was particularly prominent in those later diagnosed with autism and learning (intellectual) disability. We are also told that adding in additional variables linked to "regulatory problems - crying, feeding and sleeping" (the sorts of things that Kanner himself talked about) might increase the diagnostic yield of the 18 month surveillance by about 10%. But also... "Of those with ASD and ID who had passed, more than one-third experienced developmental regression after 18 months of age."

There you have it. Regression is very much part and parcel of quite a few cases of autism and all the chatter about autism being [universally] present during the earliest points of infancy is not necessarily accurate and applicable to all cases. Of course this probably won't be a surprise to many as previous examples of later onset or 'acquired autism' litter the peer-reviewed literature (see here and see here for example). Appreciating that many factors might influence the age at which an autism diagnosis is given (see here) and how the Swedish findings might be important for future work coming out of the UK for example (see here), I'd like to think there is more to do in this area. Y'know everyone keeps talking about the more plural 'autisms' for example (see here), well I'd be minded to suggest that those who passed vs. those that didn't pass early developmental screens might be another comparison to throw into the autism research mix with a focus on phenotypes. And given the previous research conducted by some of the authors on the Höglund Carlsson paper (see here and see here), there may be quite a few other variables that could also be included in future work in this area...


[1] Höglund Carlsson L. et al. Autism spectrum disorders before diagnosis: results from routine developmental surveillance at 18 months. Acta Paediatrica. 2016. April 8.

---------- Höglund Carlsson, L., Westerlund, J., Barnevik Olsson, M., Eriksson, M., Hedvall, �., Gillberg, C., & Fernell, E. (2016). Autism spectrum disorders before diagnosis: results from routine developmental surveillance at 18 months Acta Paediatrica DOI: 10.1111/apa.13418