Evidence lacking for most autism treatments

There has been some media interest in a series of reviews published in the American Academy of Pediatrics journal 'Pediatrics' on the effects of various medication strategies and behavioural therapies being used with autism spectrum conditions. A link to the relevant abstracts can be found here, here and here. Some viewers may know that this is the same journal that published various guidelines on gastrointestinal issues related to autism a little while back that I touched upon in a previous blog entry.

I initially had mixed feelings about the conclusions drawn from these reviews and the subsequent press coverage. The underlying message seemed to be that for most 'treatment options' there was no convincing evidence that they "actually help kids get better" (their quote not mine) at least in the long-term; although accepting short-term gains for some. Mixed feelings because of how this represents the current state of management and intervention options for (children with) autism - i.e. how little we actually know despite literally millions of pounds, dollars, Euro etc of research money being spent on R&D.

I was however heartened by the call for further research into these areas and also the suggestion that 'best' and 'non-responders' to intervention should be a focus, going back to the notion of autism and n=1. Breaking down what was actually included as part of these reviews, the authors looked at various pharmacotherapies being used in autism as well as various behavioural therapies. Their conclusions were that whilst some interventions seemed to show often quite pronounced positive effects for individuals, very little is known about the factors governing response and the characteristics pertinent to a positive response in the longer term.

Medications used to tackle challenging and repetitive behaviours specific to autism did receive a general thumbs up. Important however was the highlighting of 'side-effects' of said medications; particularly dyskinesia and weight gain following use of drugs like risperidone and aripriprazole. Unfortunately also, the anti-depressant medications didn't fare well; following on from the recent meta-analysis of SSRI use for autism shown here detailing limited effectiveness.

The one area of review that did provide the most convincing findings was related to a lack of general efficacy for the gastrointestinal hormone secretin for autism and a strange call to research it no more - strange because for about 5-6 years it hasn't; or not at least under large-scale controlled-trial conditions. What more is there to say? Well not much really.

I'd like to think that if there are lessons to be learned from these reviews they are these: (1) autism is not autism but autisms - focusing on sub-groups, and more importantly looking at best- and non-responder characteristics to various interventions, is probably the best way forward; (2) 'short-term' gains are important - but longer study periods or follow-up studies are required to see how these gains translate into long-term outcome - how we measure 'outcome' is another matter; (3) co-morbidity is something that perhaps has been given too little attention in autism intervention research - is intervention affecting core symptoms or impacting on peripheral symptoms? Does it make any difference anyway?

All of this is perhaps more fodder for the upcoming NICE review of autism coming to a station near you soon (at least in the UK).