From Medscape Pediatrics
An Expert Interview With Jeremy Veenstra-VanderWeele, MD
Therapies for Children With Autism Spectrum Disorders
The Effective Health Care Program under the auspices of the Agency for Healthcare Research and Quality (AHRQ) was launched in 2005 with the goal of providing valid evidence about the comparative effectiveness of different healthcare interventions. In April 2011 AHRQ released Therapies for Children With Autism Spectrum Disorders: Comparative Effectiveness Review Number 26, which was developed by the Vanderbilt Evidence-Based Practice Center. Three systematic reviews developed in conjunction with this research and focusing on specific therapies for autism spectrum disorders (ASDs) were published the same day in Pediatrics.[2-4] Laurie Scudder, DNP, NP, of Medscape interviewed Jeremy Veenstra-VanderWeele, MD, Assistant Professor of Psychiatry, Pediatrics, and Pharmacology and Director, Treatment Resistant Autism Consultation Clinic at the Treatment and Research Institute for Autism Spectrum Disorders, Vanderbilt Kennedy Center for Research on Human Development, Nashville, Tennessee, and an author for all of these papers about this research.
Medscape: Dr. Veenstra-VanderWeele, can you briefly describe the impetus for AHRQ’s decision to undertake a study on ASDs at this time?
Dr. Veenstra-VanderWeele: There is a lot of information about autism treatments. That information is from quite a number of sources. Families obtain information from providers, the Internet, and word of mouth. A number of studies have been published, and although each individual study may look hopeful, it’s important to put all of the data together in order to understand how strong, or not, the evidence is for each treatment.
The goals were to provide families with a sense of what the evidence looks like, to provide clinicians with an evidence base to use in advising families, and to provide a blueprint for where more research is needed.
Pharmacotherapy for ASDs
Medscape: The research broadly categorized treatment as behavioral and medical. The discussion of medical treatments focused primarily on antipsychotics, selective serotonin reuptake inhibitors (SSRIs), and stimulants and other medications for hyperactivity. Although this is clearly a very large range of agents, the conclusions across the spectrum were that there were insufficient, and sometimes conflicting, data in regard to efficacy for the majority of these treatments. The exception was the antipsychotic agents risperidone and aripiprazole, which were found to have benefits in mitigating some challenging behaviors, including hyperactivity, noncompliance, and repetitive behavior — though at the cost of significant side effects. Can you speak a bit more about these agents and any role that they should play in management?
Dr. Veenstra-VanderWeele: Part of what happens in the medical literature is that the best data often correspond to those treatments that are the newest. The atypical antipsychotics, including risperidone and aripiprazole, were studied in large part by the pharmaceutical companies or with funding provided by the companies that owned the patents for those medicines. It’s easy to understand why there might be more evidence there, partly because they’ve been studied more.
The evidence is quite supportive of their benefit for particular symptoms, most notably what is defined as "irritability," which is not a core symptom of ASDs. Irritability actually corresponds to a subscale of the Aberrant Behavior Checklist, which includes items measuring changes in mood, aggression, agitation, and self-injury. When clinicians and families hear the term irritability, it’s important that they realize that this refers to irritability, agitation, and aggression. For those symptoms these medicines are quite effective, resulting in a dramatic improvement in many individuals with ASDs.
Unfortunately, the side effects are quite significant. The most common side effects include sedation and weight gain. With that weight gain, of course, comes the risk for diabetes, hypertension, and everything else that occurs with obesity. To give you a sense of the weight gain, some of the kids in these studies gained around 10% of their body weight in just 2 months. That’s pretty extreme, and if that would be carried forward, it would obviously have an impact on health.
Because of that and the movement disorders that can also occur, these medicines should be reserved for cases in which someone is at risk from aggression or self-injury or when individuals are very impaired from these behaviors.
Medscape: Although the data were inconclusive in regard to the utility of stimulants in management of symptoms of hyperactivity, are there circumstances in which these agents should be considered?
Dr. Veenstra-VanderWeele: One of the fine points in diagnosis of ASDs and co-occurring conditions is that our current Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) only allows a diagnosis of autism or a diagnosis of attention-deficit/hyperactivity disorder (ADHD), but not both. Therefore, if you have autism and significant hyperactivity and impulsivity, you get a diagnosis of autism, but often you need treatment for hyperactivity and impulsivity.
That complicates evaluating which medicines may be helpful and means that the medicines that have been studied quite extensively for ADHD may or may not be appropriate for somebody with autism because somebody with autism would not have been included in those studies. A lot of clinicians have been using these medicines in kids with autism and co-occurring hyperactivity, impulsivity, and inattention for a long time, but there are very little data. There’s one well-controlled study in which methylphenidate was used in a population of children with ASDs and co-occurring hyperactivity. That study was supportive, but it’s only 1 study, and so there really aren’t enough data to be confident.
The next version of the DSM, the DSM-V, will almost certainly allow for the diagnosis of ADHD in the context of ASDs. That will, of course, allow for the examination of the question of whether these medicines have evidence to support their use in the context of these comorbid diagnoses.
Medscape: Data in regard to the efficacy of SSRIs was were also inconsistent. Again, are there subsets of children for whom you would consider initiating 1 of these agents?
Dr. Veenstra-VanderWeele: Here again, the complexity is around comorbidity vs core symptoms. One of the core symptom domains of ASDs is restricted interests and repetitive behavior. That repetitive behavior can range from simple repetitive behaviors, similar to motor or vocal tics, to quite complex repetitive behaviors, similar to obsessive-compulsive disorder. Because of the similarity to obsessive-compulsive disorder, people have been using these medicines for years in the belief that they would help with similar symptoms in children with autism.
We still don’t know the answer to that question. The largest study of an SSRI in ASDs lumped all simple and complex repetitive behaviors together as a group. There didn’t seem to be any benefit of the SSRIs for repetitive behaviors in the overall analysis. Of interest, they observed some benefit for the irritability subdomain of the Aberrant Behavior Checklist. However, that was a secondary measure, and although it was statistically significant by itself, it doesn’t hold up when you consider all of the different measures that were considered in that study.
At this point, we don’t have enough data to judge whether these medicines are helpful for anything in ASDs. You may have children who have autism and a comorbid diagnosis of generalized anxiety disorder or separation anxiety disorder; these are disorders of which there’s good evidence for SSRIs. Would you see the same response in a child who has autism and 1 of these comorbid diagnoses? For that, we need more data.
Medscape: Are there any agents on the horizon that have potential utility in management of ASD symptoms?
Dr. Veenstra-VanderWeele: A number of medicines are under study. It’s too soon to judge which of them may end up being helpful. I believe that there are 2 basic approaches to developing new data. The first approach includes medicines that are currently available for the treatment of other disorders. Examples of that would be ongoing studies of memantine, acamprosate, and some of the medicines that are included in the report but don’t currently have enough data to judge whether they’re effective.
The other approach is to try to develop an actual understanding of what’s happening in the brain in ASDs and then use that information to generate treatments that are not just adapted from another disorder but are specifically targeted to autism. One example of that is ongoing studies of oxytocin in children with ASDs; we don’t know whether that treatment may have some benefit or about the possible adverse effects.
Another example would be ongoing efforts to develop medicines to target the brain abnormalities that have been identified in fragile X syndrome. A number of studies are under way to target the mGluR5 pathway in fragile X syndrome. This pathway may be involved in other children with autism who may also benefit from this treatment approach. Again, the data aren’t there yet to evaluate whether those medicines are going to be helpful, but the data from mouse models are supportive of the idea.
Medscape: Is there any role for the primary care provider in providing pharmacotherapy, or should use of these options be limited to only children with the most severe symptoms who are managed in a specialty setting?
Dr. Veenstra-VanderWeele: Generally speaking, the atypical antipsychotics are medicines that should be prescribed by an autism specialist, be it a child psychiatrist, pediatric neurologist, or developmental pediatrician. Some pediatricians become quite expert in treating autism, but the message is that this should be done by somebody who has a chance to compare this child with previous children who they’ve seen with autism.
In the case of SSRIs, there are enough data showing adverse effects, primarily an activation syndrome that includes decreased sleep, increased energy, and agitation that primary care providers should steer away from prescribing those medicines to children with autism. It does not mean that they are not useful, but they should probably only be prescribed by people who see more children with ASDs.
With regard to treatment of hyperactivity, impulsivity, and inattention in children with ASDs, treatment may be fairly similar to the treatment of ADHD in a similar age range, although the data are incomplete. We’re not there yet, but it would not be unreasonable for primary care physicians to use their typical approach for treatment of ADHD in children with ASDs, provided that these children actually meet the criteria for ADHD.
At the same time, I think most primary care providers are going to want to use consultation to get more information and guidance.
Medscape: The research conclusively stated that there was no benefit to the use of secretin in the treatment of ASD symptoms. This was more strongly worded than the recommendations with regard to other agents. This is not a new recommendation and repeats a conclusion from a 2005 Cochrane Review. Do you consider the case to be closed with regard to secretin?
Dr. Veenstra-VanderWeele: Yes, there’s no reason to use secretin for the treatment of children with ASDs.
Behavioral and Developmental Strategies in ASDs
Medscape: One of the most extensive areas examined in the research is that of behavioral and developmental interventions in very young children. The conclusions were generally that the data around these therapies were insufficient and somewhat inconsistent. Studies were often small, nonblinded, lacked a comparison group, and did not control for confounding variables. However, the report emphasized that in selected populations of children, some therapies such as cognitive-behavioral therapy and social skills training have been associated with positive outcomes. However, many of these children, across all methodologies, did not appear to have dramatic improvements in social or cognitive functioning, and gains that did occur were often limited to the highest-functioning children and were not sustained. Although this is an enormous and heterogeneous body of research, can some global conclusions be drawn to assist primary care providers in helping to direct families to the most appropriate therapies? Is there any potential harm in these strategies?
Dr. Veenstra-VanderWeele: This is a difficult area. The data for the early intensive behavioral interventions are insufficient to provide guidance as to how big the effects are, which specific therapies are helpful, which specific children are most likely to benefit, and so on. However, the evidence for Early Intensive Behavioral Interventions (EIBI) is stronger than the evidence for any other intervention that may address either core symptoms or cognitive outcomes for children with ASDs.
These are the therapies for which most clinicians have the most hope. Unfortunately, we still don’t have enough data to have a lot of confidence. However, with more research, we’ll have a clearer sense of the effects. If you look at all of the early intensive behavioral interventions in aggregate, including the Lovass methodology and theEarly Start Denver Model, I would say that there is evidence that there is an effect, but there is not much specificity. If I were a parent of a child with an ASD, I would likely pursue 1 of these treatments. That’s because they look hopeful, albeit without complete data, and because the risk for harm appears lower than the risk with a medication — and that has to count for something.
Unfortunately, these treatments are very expensive and very time-consuming. If pursuing 1 of these treatments means that families cannot provide for their basic needs or function as a family, meaning that they just don’t have time to spend with each other or it leads to insurmountable family stress, then these treatments may lead to harm in the form of stress. Otherwise, the risk for harm is quite low.
Medscape: Parent training was included in this cluster of behavioral and educational therapies. Data here, too, were inconsistent. Again, are there any generalities in regard to parent training that can be drawn? Is there any potential harm to more thoroughly educate parents to manage these children?
Dr. Veenstra-VanderWeele: Again, we don’t have enough data to provide guidance on whether parent training, in general, is helpful, and which specific parent trainings may be most helpful, if any. Individual studies have supported this approach, but families are left weighing risk and potential benefit. Training that uses the EIBI model makes intuitive sense, but we don’t have data to say that it necessarily works.
I don’t think that there’s a lot of risk for harm, except within families who are stressed beyond their breaking point because of all of the interventions that they’re trying to put in place. As a clinician, I do see some families who are trying to do everything, and that means trying every single treatment that they hear about because they want to see things get better for their child. As a result, the family is near the breaking point, and I think that that is a real risk. That is something that primary care providers can help families to see. Choosing a couple of interventions and doing them to the best of your abilities makes sense, but trying to pursue everything is rarely the right approach.
Medscape: The researchers concluded with an appeal to practitioners, medical and behavioral, for the development of consistent, reproducible, objective outcome measures. Of equal concern is the need to better characterize children enrolled in trials in order to determine the best strategies for children along the entire spectrum. Will the DSM-V make that process of categorizing children more clear?
Dr. Veenstra-VanderWeele: Honestly, I don’t think that the DSM-V is going to simplify treatment research. I think that the DSM-V is an important move forward, but what really improves treatment research is the use of very careful diagnosis and categorization of individual study participants. One of the things that has happened, not just in the autism area, but throughout medical and treatment literature is that a single study will show very hopeful results, and people will conclude that the treatment works and doesn’t need to be studied again in that same way. However, we need replication; therefore, those individual studies that are quite hopeful do not lead us to clear evidence that a treatment works consistently, not just in a single study but across treatment teams, sites, and groups of patients. That is troubling and part of what has happened with the early intensive behavioral intervention literature. People have come to premature closure.
This is similar to what happened with stem cell transplants for breast cancer. Many women, including my mother, had stem cell transplants for breast cancer thinking that this worked. This is because well-meaning doctors and families looking for help thought that this was going to make a difference. This led to incorporation of this therapy into widespread use before it was fully studied.
Some of the studies being performed in autism right now are fine-tuning what we think already works. We are not ready for that yet. We need to go back and study more carefully whether some of these treatments work and which children benefit. Finding agencies want to fund new things, and that makes sense. Although we believe that current treatment is inadequate, we need to be careful to fund replication of the very best findings that are currently out there.