Does being treated for infertility raise the risk that your child will have an autism spectrum disorder? A number of mothers I have interviewed have voiced this concern. Now two new studies–one that looks at in vitro fertilization and the other at ovulation-stimulating drugs like Clomid–suggest that it may be so. These are preliminary studies but provocative.
A study presented at the International Meeting for Autism Research in Philadelphia on May 19 found that autism was nearly twice as common among the children of women who were treated with Clomid-type drugs than women who did not suffer from infertility (4% vs 2%), and the link persisted even after researchers accounted for the women’s age.
A second paper presented at the conference by an Israeli team found an association between autism risk and in vitro fertilization, which also involves the use of drugs that stimulate ovulation. The study looked at 564 children with ASD, who had come to an autism center for in-depth evaluation. It found that 10.2% of the children were the product of IVF, much higher than the rate in the general population of Israel, which is 3.5%.
These studies are far from definitive. The drug treatment study, for example, while based on a large sample of nearly 4,000 nurses, drew its data from questionnaires rather than clinical records, so there was no way to confirm the history or timing of treatment for infertility or autism diagnosis. Nor did researchers have access to information on whether the affected children were born prematurely, whether they were twins or triplets, or whether they had low birth weights.
“Preterm delivery, low birth weight, twinning and maternal age are all associated with infertility treatment and they are all associated with the risk of autism. We need to understand how to tease these factors apart,” observes Lisa Croen, senior research scientist with Kaiser Permanente in Oakland, Calif., and director of its Autism Research Program, who was not involved in the study.
The IVF study was was even more preliminary and drawn from what researchers call “a sample of convenience.”
Still, it’s interesting to think about how IVF or ovulation promoting drugs might increase the risk of autism. Do the hormones used in fertility treatments cross the blood-brain barrier and affect fetal development? Might the the quality of the eggs the drugs stimulate to ripen be sub-par? Does the growth medium used in IVF procedures sometimes harm the fetus (there’s some evidence for this in animal studies.)? Could there be something about infertility itself that also somehow ups the odds for autism? No one has the answers at this point.
Parents of kids with autism commonly observe that their children suffer from gastric distress, toileting problems, along with, in many cases, fussy eating that borders on the obsessive compulsive. (For example, I’ve heard about youngsters who insist on eating nothing but chicken nuggets for breakfast, lunch and dinner.) Issues like these have led numerous parents to switch their kids to special diets like the gluten-free, casein-free regimen popularized by DAN (Defeat Autism Now). But research on G.I. issues, diet and autism has lagged.
A new study published in the current issue of Pediatrics addresses the question of G.I. issues and autism, but it seems unlikely to settle the matter. Using Mayo clinic data from Olstead County, Minnesota, the study looked back at the medical records of 121 young adults who had been diagnosed with autism and compared them with 242 healthy controls from the same population and of the same age. Researchers, led by pediatric gastroenterologist Samar Ibriham, found that there was no difference between the two groups in the overall rate of G.I. problems from birth to age 21, though those with autism were much more likely to be plagued by constipation and feeding/food selectivity issues.
A third of those with autism (33.9%) had constipation vs. 17.6% in the control group. And a quarter (24.5%) with autism had feeding issues or food selectivity issues v. 16.1% of the controls. Ibrahim thought these differences could be attributed to behavioral issues (for example, not taking in enough fiber and water due to food rituals) and medication side effects as opposed to intrinsic G.I. differences.
In a commentary on the study, also published in Pediatrics, Dr. Mark Gilger and Dr. Anne Redel of Baylor College of Medicine, praised the study but also pointed out a number of big limitations, beginning with the fact that the study is retrospective and missing a lot of data. For example, it id not examine the duration, severity or recurrence of G.I. problems. In other words, a child who suffered from gastric bloating for months at a time could be counted the same way as a kid with a transient episode. In summary, they say this:
Ibrahim et al are to be commended for a well-performed study, with which they attempt to put to rest the nagging suggestion that there exists a link between autism and gastrointestinal disease. Unfortunately, there is more work to be done.
A good summary of the study can be found at webmd.com here.
All sorts of environmental toxins–lead, mercury, pesticides, for example–are known to affect the developing brain. A specific link to autism, however, remains elusive. At the Mailman School of Public Health at Columbia, Frederica Perera has been conducting a longterm study that is following several hundred inner-city children from before birth through age 11, looking at the effects of various environmental exposures on their development, growth and behavior.
In the past, Perera has linked exposure to the pesticide Dursban (chlorpyrifos) to low birth weight, developmental delays and deficits in attention and memory. Her latest study looks at another environmental toxin–substances called polycyclic aromatic hydrocarbons (PAHs), which are found in car and truck exhaust, factory emissions, cigarette smoke and basically any fumes resulting from the incomplete combustion of fossil fuels.
Perera’s studies of PAHs have linked high prenatal exposures (greater than 2.26 nanograms per cubic meter) to low birth weight, developmental delays, motor problems and now, in her latest study, lower IQ at age 5. In the study published in the current issue of Pediatrics, Perara found that kids with the higher exposure had a mean IQ of 96.5 compared with a mean score of 101.6 for kids in the low-exposure group. “A difference in four points could be educationally meaningful in terms of school success,” Perera told me in an interview.
Read more in my article for Time.com.
Families who already have one child with autism have perhaps ten times* the usual risk that a second child will also have autism. That’s why researchers are increasingly interested in focusing on these high-risk families as a way of teasing apart the genetic and environmental factors that cause the disorder.
One of the bigger studies of these “enhanced risk” families officially kicked off today with an online press conference. The Early Autism Risk Longitudinal Investigation (EARLI) will follow 1,200 pregnant women who already have at least one child with an Autism Spectrum Disorder, throughout pregnancy and through the first three years of the child’s life. The infants will be assessed according to language, social and cognitive development at ages 6 months, 12 months, 18 months, 24 months and 36 months.
Researchers at four sites- Philadelphia, Baltimore and two sites in Northern California–will recruit the women, including those who plan to become pregnant but are not yet expecting. The teams will collect a wide range of materials and data throughout the perinatal and post-natal period, explained principal investigator Craig J. Newschaffer, an epidemiologist who chairs the Drexel University School of Public Health. That includes 1. biological materials: blood, urine and hair samples from the pregnant woman and later from her child, breast milk and meconium (the infant’s first fecal waste), 2. environmental samples at home including household dust, air samples that measure the presence of cleaning products, pesticides and the like, and 3. information about diet, lifestyle and medical history, including infections and vaccine history.
The plan is to look for patterns and combinations of factors–genetic, environmental, epigenetic, immunological– that correlate with autism. Many of the genes linked to autism are very common in the overall population. “That suggests that those genes alone are not the full story,” explained Irva Hertz-Picciotto, chief of division of environmental and occupational health at the University of California Davis School of Medicine, one of the study sites. “We will be combining the environmental data with the genetic data,” she said. “The combination will I think really be the avenue to making headway.”
Researchers stressed that a prospective study that collects data and biological evidence as it goes obtains much higher quality information than the more common and cheaper retrospective studies, which ask parents to remember and reconstruct information about pregnancy, birth and early childhood.
EARLI is funded by a $14 million grant from the NIH and a $2.5 million grant from Autism Speaks.
*The estimated risk in the U.S. of having a child with autism is 1 in 150, or .6% The risk for families that already have one child with autism has been estimated at anywhere from 3% to 12%. Current thinking is that it’s toward the higher end of this range.
Recovery from autism is possible, but it’s relatively rare, according to research presented yesterday by University of Connecticut psychologist Deborah Fein at the International Meeting for Autism Research in Chicago. In an admittedly small study, Fein found that somewhere between 10% and 20% of children with autism disorder or PDD-NOS eventually lose the diagnosis. “Kids with PDD-NOS are more likely to recover than kids with the autistic disorder,” Fein said in an interview.
Her estimates are based on an ongoing study of 20 kids who were diagnosed with autism before age 5 and later were no longer on the spectrum. Her criteria both for the initial diagnosis and the recovery seem rigorous. She’s still working toward a larger sample of children who have recovered. Meanwhile imaging experts are looking at the brains of the recovered kids to see if they look like typical brains or more like the brains of kids with ASD.
What predicts this fantastic outcome? A relatively high IQ, the ability to understand speech (receptive language), ability to imitate and, surprisingly, motor development–crawling, walking, etc.– that’s not too delayed. Kids who have a lot of extremely repetitive behavior, kids who are very frequently lost in self-stimulatory behavior (flicking their hands in front of their eyes, rocking and generally being disengaged from the world), tend to have a poorer outcome. Most of the “recovered” kids in Fein’s sample received early intervention with ABA therapy, but she emphasizes that ABA is the predominant method in Massachusetts and Connecticut, where much of her sample is drawn.
Predicting the outcome for any individual child is difficult, says Fein. “I try to tell parents to move into high gear, try to get the best services possible, and try to live with the uncertainty for a few years. By age 7 or 8, they will at least have an idea of the developmental path that their child is on. You can see the trajectory.”
Two new studies in Nature--the largest genetic studies to date– may help make sense of the confusing number of genes that have been linked to autism. Basically, they report that people with autism are more likely to have changes in genes that govern how the brain is wired. Some of the genes make neuronal cell adhesion proteins that literally fuse together brain circuits. Others are involved in maintaining those circuits: sweeping away old wires and clearing a path for new ones. Of course, nothing is very straight forward in autism. The suspect gene variations are also quite common in people who don’t have autism. The thinking, says lead author Hakon Harkonarson of Children’s Hospital of Philadelphia, is that these gene changes are part of a larger constellation of gene defects that cause autism.
Here’s a link to my story on the studies at Time.com.
It’s interesting that some of these proteins are known to be particularly active in wiring the frontal cortex during gestation. Geraldine Dawson, chief scientific officer of Autism Speaks, notes that children with autism appear to have poor connections between the frontal cortex and other parts of the brain. These poor connections could explain why it’s so difficult for children with ASD to coordinate a task like engaging another person’s attention. “Normal babies will point to an object and make some noise, look at the other person and look back to the object,” Dawson noted in an interview. “This involves the motor system, the language system, the gestural systemic all operating in a coordinated fashion.” Weak connections in the frontal cortex could explain why children with ASD have some much trouble with this kind of task.
Where are all the randomized controlled trials on treating autism—where are the well-designed studies that might clear up the confusion about what works? There are very few, and what few there are tend to be small, or of very short duration or badly flawed.
Why so? Because it’s almost impossible to randomize kids with autism for research purposes—i.e. randomly assign half to a control group and half to a group that gets a special, experimental intervention. Arguably, it’s unethical to do so, especially for a study that lasts more than a couple of weeks. Every parent with an ASD kid feels the clock ticking ominously. You’ve got to hurry and get the right intervention or opportunities for improvement may be lost forever. Or so it seems. No one wants their desperately vulnerable child assigned to a control group as opposed to the group getting the intensive intervention. There are some study designs that get around this: for instance, randomly assigning kids to two types of serious interventions, or assigning the control group to “standard care”-meaning what’s typically available in the community, or offering the control group the experimental intervention after the study is completed, but there’s no perfect solution.
Here’s a common problem: Even if you do randomize the subjects of a study, chances are that the control group will be “contaminated” by other interventions that worried parents will try while you’re conducting the study: drug interventions, behavioral, dietary and alternative treatments or whatever the latest fad brings. This problem was reported in at least two major studies: Sallows, Graupner (2005) and Tristam Smith (2000), both in the American Journal of Mental Retardation.
My interview with Ron Leaf in late November kind of blew my mind. It certainly changed my thinking about Applied Behavioral Analysis (ABA) and the history of autism treatment. Leaf (pictured) was part of O. Ivar Lovaas’s team at UCLA when the early work was done on ABA. Together with John McEachin and Mitchell Taubman—also Lovaas associates—he’s written an incendiary new book titled Sense and Nonsense in the Behavioral Treatment of Autism: It Has to Be Said.
I was expecting the usual diatribe against all methods other than ABA. Instead, Leaf rails against the way ABA has been misapplied. “As upset as I am about other approaches, I’m more upset about how ABA is being done. I feel the rigid approach with dogmatic rules is not good for children. ABA is its own worst enemy.”
More points: Back in the day, ABA therapists used “naturalistic language” instead of “talking like robots.” Now, says Leaf, “People doing bad ABA often don’t use the word ‘the.’ They say, ‘Point to boat.’ No wonder the kids don’t use the!” He says the stilted behavior so often seen in kids who have been through ABA programs stems from rigid, robotic therapy. “Kids are taught to say ‘Fine, thank you’ in response to ‘How are you?’ But what kid says ‘fine, thank you!’?”
To be sure, Leaf is no fan of unstructured therapy: “I don’t like the lackadaisical, non-systematic approach either,” he says, allowing that methods like Stanley Greenspan’s Floortime/DIR have more “curbside appeal” to parents.
Still, he insists that, done correctly, “ABA is systematic and fun.”
What are the very first signs of autism? Can it be detected prenatally by examining a pregnant woman’s blood for toxins, immunological changes or viral DNA? What’s the role of paternal age and health? Are there telltale patterns of behavior during the first 6 months of life?
These are just a few of the questions that stand to be answered by the enormously ambitious Autism Birth Cohort study (ABC) underway in Norway. The study, which involves 100,000 subjects, is funded for 5 years by the Norwegian government and the National Institute of Neurological Disorders and Stroke—part of NIH—which is contributing $13 million. It will release its first findings in early 2009 and probably produce additional findings for at least a decade to come.
Norway is an ideal place to conduct this study, says Ian Lipkin of Columbia’s Mailman School of Public Health. (Lipkin and Camilla Stoltenberg, of Norway’s Institute of Public Health are the study’s principle investigators.) Norway has a superb, national medical system, making it easy to enroll and track study subjects for the full five years. Tracking is key, since mothers are enrolled in the trial when they are 17 weeks pregnant and then their child is followed for five years—and longer, if the next phase of the study gets funded. A similar and much more costly study in the U.S.—the $10-billion National Children’s Study —will be hard-pressed to keep track of mobile Americans, especially since it aims to track its subject until age 21.
Huge amounts of data are being gathered in Norway. Lipkin gushes over the quality of the blood samples and storage techniques—technicalities that are exciting to researchers because optimal storage means the samples can be used again and again to search for new suspects. Parents complete questionnaires prenatally, when their child is 6 months old, 18 months old and 36 months old. Children who appear to have signs of autism are flown to Oslo for evaluation and videotaping that takes a day and a half.
The first findings will look at improving parent questionnaires to identify high-risk kids. Lipkin hints that there are also important findings in umbilical cord blood that may be useful for screening.