Autism and Prenatal Ultrasound (more)

I blogged earlier about Caroline Rodgers’s idea that prenatal ultrasound may cause autism. She believes this idea isn’t getting the attention it deserves.

Recently she wrote to the head of Health and Human Services:

The latest autism prevalence figures released in December showed that while the overall autism rate increased more than 50% in the four years ending in 2006, there were significant differences across ethnic groups. White women had a much higher incidence of autism among their children than Black or Hispanic women. White mothers had 9.9 autistic children per 1,000, versus Black mothers who had 7.2 and Hispanic mothers who had 5.9.

There were also geographic differences. Among the 10 states with monitored sites, Alabama and Florida had the lowest autism rates, with averages of 4.2 and 4.6 per 1,000, respectively – far lower than the two states with the highest autism rates, Arizona and Missouri, which tied at 12.1 per 1,000. One interesting apparent statistical anomaly occurred among Alabama’s Hispanic population, which had a 68% decrease in autism while the overall national increase was 57%. In trying to understand why Alabama Hispanics had such a decrease in autism, I searched for evidence of public health policy changes. What I found was a surprise: according to a CDC multi-state surveillance report, Alabama and Florida were two of three states that had cutbacks in Medicaid funding for prenatal care during the time mothers in the study were pregnant. (The third state, West Virginia, was not among those monitored for autism in the latest study.)

Digging deeper, I turned up a CDC report on the timing of entry into prenatal care. The report showed that although most women started prenatal care in the first trimester, the percentages of both Black and Hispanic women who lacked early (first trimester) prenatal care were nearly twice that of White women . . . Over the span of the 10-year study, more women [in] all ethnic groups received early prenatal care, but the 2-to-1 ratio remained the same. . . .
Taken together, these three CDC reports tell a disturbing story: as more women . . .  received more early prenatal care, the autism rate among their children increased, with those women receiving the most early prenatal care having the highest percentage of autistic children. . . .

A rigorous UC Davis study, published in January, of California children born between 1996 and 2000 identified 10 autism clusters . . . Highly educated women were much more likely to have children diagnosed with autism than parents who did not finish high school. In six of the clusters, the rate was as high as 4 to 1. Returning to the CDC Entry into Prenatal Care report, it is striking to note that in 1997 only 8.5% of pregnant women with some college education had delayed prenatal care, versus 29.9% of women who were not high school graduates – further [linking] early prenatal care [and] autism.

A study published in November on prenatal ultrasound trends from 1995-2006 found that the odds of a woman receiving an ultrasound during a prenatal visit nearly doubled over [those] 10 years. . . . The geographical and ethnic differences . . . dovetail with many of the geographical and ethnic differences found in the latest autism prevalence report. For instance, Southern women were 40% less likely to receive an ultrasound during a prenatal visit than Northeastern women, which could help explain why Florida and Alabama had the lowest autism rates among the states monitored. Also, Hispanics, who had the lowest overall autism prevalence rates in both the 2004 and 2006 CDC reports, were 20% less likely to receive an ultrasound during a prenatal than White women.

Not all the statistics available in these reports support the idea that prenatal ultrasound is causing autism. For instance, Southern states such as Georgia and North Carolina did not have low autism rates, but [perhaps this is because] the ultrasound trends study did not take into account “keepsake” ultrasound . . .

She also notes that a study by Yale neuroscientist Pasko Rakic “found that prenatal ultrasound disturbed neuronal migration in mice.”

Here is the broad argument. 1. Autism is correlated with wealth. It is absurd that autism causes wealth; it is unlikely that both are caused by something else. Thus this correlation makes it plausible that autism is caused by something that rich people have more of than poor people. Obviously rich people have more prenatal ultrasound. 2. A localized decrease in autism happened at the same time autism almost everywhere was increasing. At the same place and time prenatal ultrasound screening surely declined. This correlation is very difficult to explain with other ideas about what causes autism. Dozens of things (e.g., genes, diagnostic criteria) previously proposed as explanations of autism remained roughly constant at the same time as the decrease. 3. The mice data make the linkage considerably more plausible, assuming (a) the ultrasound dosage was reasonable and (b) humans with autism have unusual neural wiring that resembles the changes seen in the mice.

The full letter is on her blog. An article by Rodgers about this

27 Replies to “Autism and Prenatal Ultrasound (more)”

  1. Not the definitive study, but here’s a quick idea: What about countries in which ultrasound is paid for by universal healthcare systems? I’m thinking specifically of my native Germany where, it seems to me, pretty much every pregnant woman gets a few ultrasounds – the number of ultrasounds mothers get seems to not be related to socioeconomic status. (I’m basing this solely on anecdotal evidence.) This suggests that the social gradient in autism should be less extreme or nonexistent in Germany and similar countries – if there is truth to the theory.

  2. I wonder what the Autism rates are in communities that decline medical care like Christian Scientists and Amish? Also, I have always been curious to see the accidental death adjusted life expectancy rates for those groups, I suspect that they have as much if not longer life spans as people with the full benefit of western medical care.

    How to explain the high rates of Autism in recent African immigrants? I suspect they do not get ultrasound at a high rate. Many attribute the difference to vitamin D status. I also think that there are in fact a range of physical syndromes that are lumped together as autism spectrum. I can’t imagine the cause will be a single environmental stressor, but rather a variety of stressors added to particular genetic traits.

  3. Thanks for blogging about this. I remember your original post, and this is still the only place I’ve heard anything about it. Any idea how to get the NYTimes or Washington Post or the like to pick it up? It seems incredibly important and I have the feeling linking this post to my FaceBook account won’t be enough to get people’s attention. Obama isn’t on my friends list.

    Perhaps a tangential thought, but in a competitive information environment, what are the ways to help things that are both important and reasonable enough to deserve wider audiences go viral? Obviously, you’re doing your part by blogging about it, but what else could launch this idea to another level of attention? Rodgers needs a PR agency to help her get her important message out.

  4. Seth, you say “[i]t is absurd that autism causes wealth” — but I thought there was at least one psychology study that suggested Fortune-nnn CEO’s were more likely than most to have at borderline autistic tendencies?

    And isn’t the wealth factor potentially part of the “worried well” syndrome? To have a child diagnosed autistic, isn’t it an advantage to be white, middle class, to have appropriate health insurance, with enough free time to worry about the possibility and to take the child round the specialists?

    I don’t mean to rain on anyone’s parade — I love your work on personal observations — I just wonder whether this correlation has rather more possible common contributing factors than most ideas that reach your blog.

  5. Alasdair, it is absurd that having an autistic child makes the parents or their neighbors wealthier. You are raising the “both X and Y caused by something else” possibility, which I do not say is absurd for the reason you say. But that possibility and your worried well idea only explain one of Rodgers’s three lines of argument, which I list at the end. This is why her case is so persuasive (compared to other theories of autism): It is supported by three disparate kinds of evidence.

    MT, I suppose the NY Times would write about this if something relevant was published in a good scientific journal.

  6. What if Stephen Levitt got interested and did a back of the envelope analysis for his blog? Seems like the data would be readily accessible. Maybe you could put Rodgers in touch with him 🙂

  7. jay, I don’t like the theory you mention because, in my experience, genetics has never been shown to cause any common problem, in spite of many claims to the contrary. To bet on that theory is to bet on a horse that has never won a race. Autism runs in families, the author says — yes, and so did pellagra.

  8. If you don’t think genetics plays a part in brain function or personal tendency, I invite you to do a little anecdotal research on hoarding in families — and not just parent to child, but with uncles and grandparents that have little contact with offspring and so on, when it skips a generation and so forth. I’ve seen it with my own eyes on several occasions. Certainly not the foundation for a theory, but, perhaps a hypothesis for an experiment…

    I’d still suggest reading Simon Baron-Cohen’s book, even if you think it’s wrong, because you never know.

  9. Interesting [beginning] look at data…however extrapolating that ultrasound is the cause of such a problem is pretty far fetched. there are a lot of other factors that most likely are playing into such a phenomenon.

    more education quite often means older maternal age, which is linked with a higher rate of birth defects, for women as well as men for many types of diseases. while none of these studies specifically point to this it may play a role.

    access to prenatal care can often be similar to access to consistent/frequent pediatric care. if children with subtle behavioral differences do not receive frequent care, their chance of being diagnosed with such a disorder is less likely. this is a fact. autism is a disease with a wide spectrum of presentations.

    these studies are also incapable of closely examining the myriad of variables such as the intricacies of diet, exposure to cell phones and other radiation, as well as any other thing to could be added to this potentially endless list.

    the truth is…autism is a vogue umbrella of a diagnosis. it encompasses what we will probably list as a hundred variations of behavioral disorders once the technology and true understanding of the pathophysiology is there. until then blogs like this poke vague assumptions using series of unrelated data.

  10. questioneverything, of course there are plausible alternatives. Who said otherwise? But you have failed to give a reason that prenatal ultrasound — which unlike more prenatal care, has been shown to cause neuronal abnormalities (in mice) — is NOT plausible. Autism has greatly increased over the last 20 years — maternal age has not. Nor has prenatal care. Prenatal ultrasound has greatly increased over the same period.

    jay, my claim is that genetics hasn’t been shown to be important in any common health problem (e.g., diabetes, obesity, heart disease). Hoarding isn’t a common health problem. Sure, there are gene by environment interactions for common health problems. But in the right environment, nobody would have the problem — at least that is what the data seem to say. A theory that ignores environmental causation is a hard sell, given that data. But I agree I should read Baron-Cohen’s book.

  11. Regarding genetics and its role (or lack thereof) in common health problems: Seth, this paper claims that BMI has a heritability of about 33%. Is that figure an over-estimate in your opinion, or is 33% below the threshold at which you would consider it to be important?

  12. The classic “Mere coincidence?” story. A couple of my favorites are that the obesity epidemic was caused by the metric system (kilojoules harder to keep track of than calories), and by electric blankets (nighttime energy use decreased).

    Make a timeline of changes in society and correlate them with whatever bad trend and see what aligns. It’s especially easy with autism since there is no real hard start date for the epidemic, to the point that many informed researchers think the whole epidemic may very well be fake.

    This is similar to the data mining for correlation research that 95 percent of “science” is these days. Dump 1,000 characteristics each for 1,000,000 people into a computer and discover that cancer correlates highly with people who wear blue paisley neckties.

  13. I can say from personal experience that having an autistic child has definately not improved my economic condition! It is very expensive to get all the treatments my son needs and the little bit of SSI he gets really isnt cutting it at all…

  14. In this particular study they are showing that women typically White and well educated recieve health care earlier and more often. So perhaps the link isn’t accurate that educated white women have autistic children more often, but rather well educated white women know what the signs are, and know that something may be wrong and can afford to get tests and seek further healt care for their child. Where women of other ethnicities or lower eduacation levels don’t know what they are looking for/do not have the financial means to get their child properly treated/diagnosed. This show’s nothing more than faulty statistics and lacks and scientfic evidence. What about a study showing the actual effects of sound waves on a fetus.

  15. Back in 1987 when I was pregnant with my last child I researched ultrasound which was not around for the first two. Not only was there a book out about the dangers but the head of radiology for a New England hospital said it was dangerous for the fetus and he would never have his wife use it. I was convinced. I would not use it for fear of harming my child.

    I am white and educated (if you overlook the fact I didn’t finish my dissertation at Stanford). I chose a birth center over a hospital. I was in at 4 am and home by 9 am. The less intervention the happier I was. I was an outlier and didn’t give a damn. I fail to understand why educated women bow down to the medical mafia instead of researching. I think the poor know better – danger lurks in the white coats.

    (My father was a board certified MD who taught at a med school, had a private practice and published many articles. Maybe this is why I know white coats can hide danger.)

    @Dorothy – all that has long ago been taken into consideration and found to be of no consequence

  16. As the expecting father of twins currently in the care of a high-risk OB practice that uses frequent ultrasounds, I read Rodgers’ Midwifery Today article with great interest. (And, before continuing, I want to disclose that my wife and I originally began our care with midwives and wanted desperately to deliver naturally, preferably in water, with little-to-no modern-medicine micromanaging.) In fact, I’ve spent the last 4.5 hours reading the full text of every linked reference (and then some) and found it to be little more than a fear-mongering propaganda piece full of logical fallacies, irresponsible generalizations and academic dishonesty (most notably in the distortion of quotations and scientific data from the original sources, as well as the selective ignoring of authorial conclusions that contradicted her own argumentative goals).

    I could produce a publishable response to Rodgers’ disinformation at least equivalent in length. Most notable, however is that:

    -The mice study has no scientific bearing on an argument re: human fetuses. Read the full text (I could quote a dozen or more passages discrediting Rodgers’ use of this data); then read the peer-reviewed commentary/response to the study’s findings, also published in the Proceedings of the National Academy of Sciences, found here:

    -Pay special attention to the way Rodgers equates the actual FDA language of the way ultrasound “heats the tissue slightly” to hyperthermia (a gross exaggeration) and then goes on to equate the negatives effects of hyperthermia (true) to fetal development.

    -Note that her direct quotation (from the Cochrane Database) about Doppler ultrasound was pulled from the headline of the summary/abstract without reading the full text of the findings (lazy would be a compliment here). Even the Main Results portion of the summary states, “there were unexpected findings suggesting possible harmful effects, but the explanation for this is not clear.”

    I could seriously go on and on (sources since pulled from publication for being outdated, 28 year-old WHO data, the lack of a single reputable source cited in the article that uses both the terms “ultrasound” and “autism”…).

    As a faculty member in a university composition program, I’m tempted to use Rodgers’ article in my courses as an analytic example of how writers use fallacy and manipulation to obscure truth in research-driven argumentative writing.

    I’m not claiming that ultrasound technology is 100 percent safe in all circumstances. (In fact, during my wife’s next ultrasound, I plan to hover over the technician and not only inquire as to whether she can accurately read the thermal monitor on the machine (doubtful, so the literature implies), but implore her to remove the transducer from my wife’s stomach at the exact moment that it’s no longer medically necessary.) No, I’m writing simply to say Rodgers’ hasty (another compliment) correlative conclusions about ultrasounds and autism are little more than the unfounded autism explanation du jour.

    Please: Read some of the cited sources, carefully, and compare them against Rodgers’ version of the facts. I think you’ll be as astounded as I am.

  17. Drew, I looked at the first two links you gave. I saw nothing reassuring in them. Then I looked at the PNAS link. It referenced an Australian experiment I didn’t know about. The experiment found no clear differences between children given five ultrasounds and children given only one. That is reassuring, yes. But it also referenced three studies I hadn’t known about that did find clear differences. In one, children exposed to ultrasound had lower birthweights. The other two studies were with animals and found damage due to ultrasound. (You can find details here.) So I ended up more concerned than I began.

    I don’t understand why the following quotation — “there were unexpected findings suggesting possible harmful effects, but the explanation for this is not clear.” — supports your case. Just as puzzling, you complain that Rodgers quoted a headline “without reading the full text”. You seem to be saying that the full text makes the headline misleading or wrong but I have no idea why.

    You don’t provide a good explanation of why autism is more common among rich parents than poor ones, or a good explanation of the anomaly in autism rates Rodgers noticed (in a certain place autism went down rather than up). I thought the anomaly in autism rates was her strongest point.

    In summary, I’m unable to understand the force of any of your criticisms. And you seem to have ignored the epidemiological evidence for what Rodgers says. I’m sorry we can’t communicate privately about this because I think you have something valuable to say — I just can’t figure out what it is.

  18. Hi Seth. Thanks for responding to my thoughts.

    I wonder if our links are shooting to different places. sends me to “Our Unborn Children at Risk?,” the aforementioned commentary critique of the mouse study. The first paragraph under the “How Well Matched…” subsection clearly explains the scientific invalidity of the mouse study for human understanding (bone density, brain size, concentrated wave length, duration of waves, etc.). The second paragraph of this subsection would nicely sum up my own conclusions about ultrasound on human fetuses — let’s be careful with this technology, use it prudently, but not assume it’s causing harm, because we have no reason to believe such.

    Do you not link to the information I”m referencing?

    I clicked to your link, which is also quite interesting, and I noticed the bit about lower birth weights, but you had to have noticed the follow-up sentence stating that none of the children remained underweight or at risk. Quite the contrary, that, “There were no significant differences indicating deleterious effects of multiple ultrasound studies at any age as measured by standard tests of childhood speech, language, behaviour, and neurological development.” Compounded with the other two animal studies you alluded to — which, if they were anything like the mouse study, have little-to-no bearing on humans because of the way they’re conducted — I don’t understand how you’re more concerned than you began.

    You chose not to acknowledge the fact that Rodgers equated the slight thermal effect of ultrasound waves on humans with hyperthermia, then equated hyperthermia with autism. This would be like saying that every time I’m in the frozen-foods section of the grocery store my body temperature drops a degree (F) or two, and that hypothermia can cause the loss of extremities, so there’s as good a chance as any that the frozen-foods section is causing the loss of extremities.

    I will admit that I may have wrongly assumed that Rodgers did not actually read the Cochrane study because A) she quoted verbatim from the overly generalized headline/abstract (poor form at even the freshman undergraduate level) and the Cochrane study is the one that was removed from publication for being outdated. Considering Rodgers wrote her original article in 2006, it’s possible she clicked through and paid the fee to read the entire article. Possible. I’ll leave it at that.

    I cannot explain the correlations between socioeconomic status and autism rates because I’m neither social nor medical scientist. Neither is Rodgers.

    Besides evidentiary problems, I took great issue with the manner in which Rodgers’ original article was written. Before I get to the host of logical fallacies present in her argumentative assertions, consider the 6th paragraph of her article, which (referring to the mouse study) claims that, “The research…also implicated ultrasound in neurodevelopmental problems in children, such as dyslexia, epilepsy, mental retardation and schizophrenia…” Yet, the wording of the mouse paper actually states, “Furthermore, there are numerous human neuropsychiatric disorders that are thought to be the result of misplacement of cells as a consequence of abnormal neuronal migration.” Thought to be. In a study/paper about how ultrasound can cause abnormal neuronal migration in mice, the inference to connect the previous statement with ultrasound is there. I’ll admit that. But what Rodgers claimed and what the researchers claimed are two different things. It’s an irresponsible inference, in my opinion, to drop the buzzword “ultrasound” into the Midwifery Today article when it didn’t appear in the original context – just because most readers won’t notice.

    Finally, as for logical fallacies, I spotted all of the following:

    -Hasty Generalizations (not enough examples or untypical examples)

    -Post Hoc, Ergo Proper Hoc, which translates to “After this, therefore because of this.” The example here is, “Wealthy babies have better health care (and probably receive more ultrasounds) and have higher rates of autism, so wealthy babies have autism BECAUSE they have better health care (and probably receive more ultrasounds).

    -Non Sequitur (“It does not follow”). An example from Rodgers’ article might be, “Therefore, if repeated experiments show that elevated heat caused by ultrasound damages fetal brains in rats and other mammals, one can logically assume that it can harm human brains, too.” No, one can’t, for a number of reasons.

    -Ignoring the Question (Instead of dealing with the topic under discussion, the writer deliberately creates a diversion). See the sections on Hot Tubs/Saunas and the flashback to the vaccination debate, the former autism explanation du jour championed by Jenny McCarthy and taken seriously by no one except the socioeconomically privileged suburban housewives these fear tactics are meant to scare.

    -Question Begging (assuming the truth of a debatable point and basing the rest of the argument on that shaky assumption). I don’t think I need to even explain here.

    My problem with Rodgers’ article, you see, is about both scientific validity and presentation, though mostly with the latter. The bottom line for me is that Dr. Rakic (from the mouse study) clarified, “Our study in mice does not mean that use of ultrasound on human fetuses for appropriate diagnostic and medical purposes should be abandoned. Instead, our study warns against its non-medical use.” Yes. Okay. No more boutiquey, keepsake ultrasounds. Great. But for Rodgers to skew this data (along with the FDA’s and others’) into claiming that ultrasounds under the care of an Obstetrics professional (and for medical use) are causing autism is disingenuous at best, unethical propaganda for the Midwifery Way at worst.

  19. Drew, thanks for the clarification. I now understand your position much better. There’s no way to deal with your comments briefly so I hope at a later date to write about them in a blog post. If/when I do I will notify you by email.

  20. Go to you tube: type in “Behaviorally Fragile Autistics” on you tube. This is a most interesting caes of autism, self injurious behavior and seizures. Very complex, but not without hope. There are some very important points seen in videos that would help professionals and anyone working with this unique population.

Comments are closed.