IMFAR 2012: Toward Better Psychopharmacological Management of Anxiety In ASD

This is one of four talks from the IMFAR 2012 symposium Anxiety in Autism Spectrum Disorders: From Biology to Treatment. The overall session theme: It’s clear that anxiety is a major problem for kids and adults on the autism spectrum, now we need to figure out how to treat it.

This summary is not intended nor should be taken as medical advice. Please consult a medical doctor for any decisions regarding psychopharmaceuticals. 

Any errors or omissions in the edited-down transcription below are on
me. Dr. Scahill also presented data and slides from several studies,
that information has either been summarized or omitted. INSAR members
can download the full audio file of Dr. Scahill’s talk.


Lawrence Scahill

School of Medicine, Yale University

There are so few data on this topic that he actually had to change the talk.

The story starts in 2002, with a study in the New England Journal of Medicine on Risperdal (Risperidone) treating aggression and self-injury in autism [PDF]. This was part of a series of steps that led to the official FDA approval of Risperidone to treat tantrums, aggression, and self-injury in autism. This was a big deal, it was the first time it had ever happened.

Notice he’s deliberately not saying that the drug is treating autism — it is treating the cluster of symptoms in children with autism that, when they occur, can cause big problems. The reduction of those symptoms improves the lives of these children. So this became a pathway for drug approval in autism. And that’s part of the reason that drug companies are now interested.

The FDA reckons that this is a safe and effective treatment for tantrums, aggression and self-injury — the aberrant behavior checklist, the irritability subscale, the focused symptoms. This gives us a handle on how to approach other outcomes. But when they look closely, they realize that they really do not have too many other good measures. They have lots of ideas for symptom clusters, but not data. And if you can’t measure it, you can’t study it.

He was part of a task force to survey, very carefully, all measures of anxiety in typically developing children and adults, and see if they might apply to children and adults with autism spectrum disorder. And after all was said and done, they thought there might be some problems, mostly because the scales they were using were reliable solely in typically developing children. Many of the scales, such as the MASC (Multidimensional Anxiety Scale for Children) relied too much on language, on asking people to talk to you.

So, what do we know about drugs and children with anxiety and ASD? Not much.

If we look at SSRIs, there are two large scale studies in children funded by the NIH (National Institutes of Health) that show SSRIs are effective in treating the big three anxiety disorders: separation anxiety, generalized anxiety disorder (GAD), and social phobia. But no evidence in children with ASD.

What about benzodiazepenes? They are effective in treating anxiety disorders in adults outside the world of ASD. They tested children in 1995, but stopped the study early because it was such a dud — the drug was Clonazepam — so he doesn’t think you should go there, he thinks that would be a waste of time.

What about the anti-convulsants? There’s an anti-convulsant called Pregabalin that is marketed in Europe for GAD and social phobia. There are six very well-done studies that show that drug is effective in anxiety (it’s been studied in the US but has not been approved for adults, and hasn’t been studied in children).

He described two NIH-funded studies about anxiety disorders in typically developing children. The two drugs were Fluvoxamine [Luvox], and Sertraline [Zoloft], both SSRIs. The main outcome measure was the Pediatric Anxiety Rating Scale (PARS). In both studies, the improvement was solid.

What about SSRIs in children with ASD? If we pull from three sources: the Fluvoxamine study, the Sertraline study, and a study we did with a different SSRI, Citalopram — not in anxiety, but in children with repetitive behaviors. With Citalopram, there was no difference between a placebo and a drug for repetitive behaviors. The only difference found was in adverse effects, in particular something called ‘activation’: increased motor activity, disinhibition, compulsive behaviors, insomnia.

In the Fluvoxamine study for typically developing children, 27% of the children had increased motor activity, 19% had insomnia — really no different than the placebo. In the Sertraline study, 5% had increased motor activity, 6.6% had disinhibition, 6.6% insomnia. In the Citalopram study: 38% increased motor activity, 20% disinhibition, 38% insomnia. Now, obviously some kids experienced one, two, or all three of these symptoms. And those rates are higher in kids with ASD than in typically developing kids with anxiety disorders. Is it something to do with the dose? No, a very conservative dosing strategy was used — ASD children just appear to be more vulnerable to these adverse effects.

So, what are we going to do to measure anxiety in this population? We need to do a much better job. We need a large-scale study.

(Childhood Anxiety Sensitivity Index) may not always be very useful [in
measuring anxiety in autism] because it’s a self-report tool, it would
be more helpful if it was revised. The CGI (Clinical Global Impression) may be more accurate.

Conclusions: If you use SSRIs [with autistic people], you are doing it without any empirical data. But if you do:

  • Start very low — start very slow. Because our data from the Citalopram study suggests that the dose and the rate of increase is what brings on the activation.
  • Monitor sleep — sleep is actually looks like the tip of the activation.
  • Activation doesn’t mean discontinuation. But the adverse affects won’t go away on their own, you have to lower the dose (that’s his experience).


Readers may want to also refer to Scahill’s 2009 paper Anxiety in Children and Adolescents with Autism Spectrum Disorders.