The medicines cabinet and autism

The review articles on autism interventions recently published in Pediatrics have certainly created some discussion on the web and blogosphere. I have read through quite a lot of the media content and discussions following the reports, and opinions have certainly been formed and communicated by many people for all the various reviews; particularly when it comes to the one about pharmacotherapy for autism.

This got me thinking about an important issue raised following the review, namely what medications are 'used' in autism and more specifically, the data on the cost-benefit ratio. I might add that by using the term 'cost-benefit' I am not so much talking about the financials, but more the positive effect vs. side-effects or positive effect vs. no effect.

I will hold my hand up at the this point and say that I am not a medical Doctor or Pharmacist. My opinions are my own and based on the available information/research on the topic. I am just a daft old-ish researcher, so please, don't go changing meds or anything like that on the basis of any information here. Any such issues, changes, alterations or other advice about medication need to be discussed with the appropriate supervising Physician. OK, you (or I) have been officially disclaimed.

The use of medication 'for' autism has a bit of a love-hate relationship. On the one hand, some medications used to 'treat' some of the core and peripheral symptoms associated with autism spectrum conditions have been very well received on the whole (on the whole) - particularly where accompanied by good medicines management. One good example is melatonin for sleeping problems - melatonin by the way in the UK, is still a prescription-only drug despite being quite widely available elsewhere in the world.

On the other hand, other preparations have not fared so well for autism. The previous use of fenfluramine is one particular example (which was eventually withdrawn following indications of heart problems associated with general use). The recent Cochrane review on the use of SSRI medication for autism is potentially another example; although I would stress that SSRI outcomes are more down to a lack of significant effect in autism rather than a general safety issue. In some cases, medication is implicated in the very saddest outcome.

Somewhere in between these examples are the wide range of other pharmacotherapies used for various aspects of autism and its co-morbidities. I would, at this point, provide you with a link to a published paper which details the extensive list of medications used for autism outside of the recent Pediatrics review. The only problem is that said research into medication changes so fast that the links I would provide are slightly dated.

One of the most comprehensive links I can find is this one from the Autism Research Institute (ARI) and their survey of interventions for autism which include medications. The ARI list is extensive. Some medications seem a more obvious choice than others but I suppose it depends what you are trying to 'treat'. Bearing in mind also that people with autism are at least as prone as the general population to health and ill-health and hence likely to use the same meds.

The thing I like about the ARI data, as 'unscientific' as it may seem, is the fact that it details parent / caregiver / first-person experiences of whether medication was beneficial, detrimental or showed no effect. I say 'unscientific' here to mean that data is not derived from a clinical trial, where other modifying variables are perhaps controlled for. I could equally argue that first- and second-person observations are important, particularly where we are talking about n=1. Indeed, most physicians using medication for autism or anything else tend to be reliant on what happens to the patient rather than what a clinical trial might describe.

Some people might also question the authenticity of the ARI data, if for example, derived via the Internet and hence not subject to appropriate controls. I would agree that this might be an issue in some cases, although I fail to believe that 26,000+ people would be so wrong or so malicious. One also has to recognise that such methods of data collection are also receiving some validation in autism research.

Anyway, the ARI data provides some interesting information.

• A suggestion that the anti-epileptic drugs (AED) tend to be quite good at controlling seizure activity - a definite relief for all given my slightly grim posting on epilepsy and autism recently. Medications such as valproic acid and carbamazepine came out top (based on several hundred reports). When it came however to looking at behavioural changes on such meds, there was a definite shift in terms of positive reports.
• Allowing for different numbers of people reporting, the 'top' meds in terms of ratio of improvers/got worse were: the anti-fungals fluconzole and nystatin, the anti-viral valaciclovir (valtrex) and IVIG.
• The not-so-good meds in terms of ratio included some of the AEDs effects on behaviour (outside of epilepsy) and pemoline (cylert).

I will reiterate my previous disclaimer at this point on not giving medical advice before going on with a short analysis. I might not be a Physician but even I realise that there is something unusual about the list of meds supposedly most helpful for autism listed in this survey: none of them are, what you might call 'the usual meds' for treating/managing a developmental/behavioural/psychiatric (delete as appropriate) condition.

Don't get me wrong, medications like the neuroleptics are included in the ARI list: risperidone, for example, posts quite an admiral 2.8:1 (better:worse) ratio (n=1038). Its just that they don't seem to do as well, according to the reportees on the ARI site, as some of the more unfamiliar meds. Looking at the scientific literature on the 'top' meds, there is very little in the way of controlled trials of the various medications listed. Yes, there are lots of web entries about the experience of using things like nystatin and other potentially-related meds in autism but nothing spectacular in terms of clinical trials at the current time.

A similar story for the anti-virals, despite some initial evidence of why they might work in some cases. IVIG has been the subject of quite a bit of research related to autism (as well as oral immunoglobulin with results here and here). Here the research is best described as 'equivocal'.

Interestingly also that co-morbidities of autism seem to have been a target of these top meds, particularly a role for immune and gastrointestinal functions related to autism. I am not going to make too much more of this data for now. Only to say that if any willing organisation has a few millions pounds/dollars to spare, and would like to investigate a few potential intervention 'targets', perhaps look no further.