More thoughts on ultrasound, questions about risk, and autism

Last week's essay on ultrasound generated a great deal of commentary.  Some of the questions raised by readers were beyond my expertise, so I turned to Dr. Manny Casanova from the University of Louisville. He and his colleague Emily williams have spent a good bit of time studying ultrasound and its effects on cells.  They were kind enough to write the following essay and will respond to your questions here on the blog:

One problem of which we've become poignantly aware is that ultrasound, especially since the early 1990s, has been deregulated and is nowadays used to excess. Ultimately we would like to see more research into its safety, as well as tighter regulations on its use so that its risks don't outweigh the benefits. We'd also like to clarify that we're not proposing that ultrasound is "the" cause of autism. What we're proposing instead is that ultrasound may be one of many risk factors for those who have a selective vulnerability.

Many people when they first hear about ultrasound as a possible risk factor in the development of autism think it sounds like pseudoscience. Who can blame them? We've been subjected to many different hypotheses about what may be causing autism.  It seems like everyone is ultimately seeking the "holy grail" of causation. So we're all skeptical when we hear something new, especially something which seems to contradict our understanding of how we view the world-- or in this case, how we perceive the safety of ultrasound. After all, ultrasound is just a picture, right?

That's what we in our laboratory used to think until we began studying what mediates the effects of ultrasound. In the following paragraphs we hope to offer a simple explanation on the rather complex effect of ultrasound on the living cell.

Ultrasound refers to sound that has a frequency above that which can be detected by the human ear. Sound itself is the force of pressure through a solid, liquid, or gas; it causes the movement of those particles. In the case of prenatal ultrasound, the ultrasound transducer emits sonic waves into the abdomen, the sound enters the body including that of the developing embryo/fetus, bounces off the tissue, reflecting back, and that echo is measured by the transducer to form a representative visual image.
Ultrasound is currently used in a variety of ways in medicine and research, and some of these include:

1) the production of lesions in neurosurgery, similar to the use of laser;
2) transcranial (across the skull) stimulation of brain activity, similar to transcranial magnetic stimulation (TMS) or the use of electrodes;
3) vasodilation, or the widening of blood vessels, which helps in both visualization of the vasculature as well as the delivery of important medications to tissue;
4) transdermal (across the skin) delivery of medications which would normally be unable to cross the skin barrier;
5) wound healing, such as on certain bone fractures and ulcers;
6) the purification of foods via its oxidative potential;
7) the purification of metals also due to its oxidative capacity;
8) transmembrane delivery of nonviral genes into target cells (mainly used in research).

These are just a few examples of how science and medicine apply ultrasound. As you can probably guess by now, given its capacity at different levels of intensity to promote cell growth, cell destruction, alter membrane fluidity (e.g., poke temporary holes in cell membranes), and alter a cell's activity such as causing a neuron to fire, ultrasound has an incredible range of effects. It turns out it's not just a picture after all.

The physical effects of ultrasound include both its pressure on the water within and surrounding a given cell, and through the creation, oscillation (spinning), and implosion of bubbles in that same liquid. The latter is referred to as "cavitation" or the creation of a gaseous cavity within the liquid. Cavitation and noncavitational effects together can poke transient holes in cells, activate certain molecular pathways within those cells, cause temperature increases when the bubble violently implodes, promote the creation of free radicals (oxidation) when that gas escapes into the surrounding medium which can subsequently damage or even kill a cell, can cause general disarray within the cell, and at certain intensities may even promote mutations of DNA.

Most of the deadly effects on cells are generally not seen at diagnostic intensities levels. However, there is still the potential that ultrasound is altering how these cells develop and behave; i.e., it doesn't kill them, it changes them. In the case of autism, we frequently find abnormalities in neuron number and growth patterns in the brain. Given that ultrasound has the capacity to promote cellular growth, as well as its overuse in obstetrics and the apparent rising numbers of autism diagnoses, this is a prime area for scientific study. Needless to say, this is a gross simplification of our hypothesis, so for anyone interested in more detailed accounts, please contact us for further materials and we'd be glad to supply them (see

Back in the 1960s, '70s, and '80s, the scientific community was very cautious about using prenatal ultrasound. As much as science knew in the day, they expressed due concern and performed a good number of safety studies. From these studies, they decided that ultrasound was ultimately safe to use in obstetrics. However, science is ever-changing and continually learning more about development. Back in the 1970s, the height of concern over ultrasound was whether it promoted spontaneous abortion or reduced postnatal survival rates, whether it promoted macroscopic growth abnormalities like differences in birth weight and overall size, and whether it caused genetic mutations. Nowadays, we know much more about the molecular biology of the cell, and more as to how development can be affected in microscopic ways which can have very big effects on behavior. Let's face it: when a postmortem examination is performed on an autistic person's brain, usually one of the most striking things about it from a macroscopic level is that there isn't anything unusual. So the differences in an autistic person's brain are indeed very subtle; they need to be teased out with various technologies, with a knowledge of the complexity of anatomical, cellular, and molecular biology, and a nuanced understanding of early development. Our science has continued to mature, but unfortunately the early safety studies on ultrasound were never updated to include this new understanding.

It's time we go back and reassess, with new knowledge, techniques, and technology, whether or not ultrasound is truly as safe as we assume it is. It's also time that the regulations on ultrasound be refined so that we can be doubly sure we're not putting our unborn infants at risk, be it for autism or some other condition.

Again, what we want to stress is that we're not advocating the disuse of ultrasound. It's an extremely vital and useful tool in medicine. But we are advocating that it be used more wisely. For those who are pregnant, we recommend that ultrasound should not be performed during the first trimester unless it is an at-risk pregnancy, and especially not within the first 8 weeks of gestation. The first 8 weeks is the period when the greatest intensity of growth occurs-- and therefore when the greatest damage can be done. Be cautious of early and unnecessary ultrasounds. In addition, don't use fetal heart rate monitors for private use because these are handheld ultrasounds.

                                                                                                            Manuel F. Casanova, M.D.
                                                                                                            Emily L. Williams


Pittmom48 said…
I have a huge amount of respect for Dr. Casanova. I teach with brain images from his research whenever I talk to groups about the brain basis of autism. He has generously shared time to answer questions via email and also sent some parts of unpublished chapters - and there are many researchers who wouldn't bother to take that time.

This blog post is very clear. Thank you, once again, Dr. Casanova for illuminating some of the harder to understand aspects of a theory that looks fairly (unfortunately?) promising.
Pittmom48 said…
(Lu Randall, Autism Connection of PA, AKA Pittmom!)
EJ Willingham said…
I have a couple of questions for Drs. Casanova and Williams: Most women I know who have ultrasounds during pregnancy do so at about 17-19 weeks of gestational age. Can you address this particular timing of ultrasounds in your commentary? The only time I've been referred for an ultrasound during the early period you describe was for a suspected missed abortion because of bleeding--which indeed it was. What are unnecessary reasons women might be having ultrasounds in the first 8 weeks?

Also, can you define "overuse" of ultrasound? I'm not familiar with the literature on that. We had level II ultrasounds at ~18 weeks' gestation with the pregnancies for each of our three sons (only the first of whom is autistic) primarily because of maternal risk factors. What are some examples of overuse? Regarding at-home fetal heart monitors, how do the power and tissue targeting of these handheld home monitors compare to power of clinical diagnostic ultrasounds? Thanks for your time.
Sarah said…
As a MD, mother of four unscanned babies, and now writer on pregnancy birth and parenting, I have been following this issue for 18 years. I share Dr Casanova's concern about ultrasound safety and possible relation to autism, especially in light of this study, which showed dose-related abnormalities in neuronal migration in mice exposed to diagnostic levels of ultrasound (whole article available here

Just yesterday I read the 2010 CONCLUSIONS OF World Federation of Ultrasound in Medicine and Biology SAFETY SYMPOSIUM- 

”Ultrasonic imaging ... cannot be unequivocally proven to be without risk from undetected bioeffects,... There are 
no scientific epidemiologic data from exposure to outputs and scanning procedures of modern ultrasound equipment on which to base the direct assessment of safety.


My 2005 summary of ultrasound safety is available on my website and is updated in my 2009 book Gentle Birth Gentle Mothering.
Anonymous said…
Having recently suffered a "blast" injury from a power surge through an amplifier, I can attest to the diffuse disruption caused to mature neurological structures from sound vibrating through a brain. It is no secret that certain illnesses can cause disabilities during specific periods of development, such as the Rubella epidemic of the '60s. The mass of Vets returning from recent wars suffer from this type of blast injury, caused by invisible means, leaving invisible scars which are not shown through ordinary means (i.e. CAT scans). Good discussion, John.
Emily Williams said…
Hi, this is Emily Williams, one of the authors of the above article. Thanks, everyone, for your comments so far and I'm glad we were able to explain a little more of the science of ultrasound in understandable terms.

As for Emily's questions above. 1) At present there are no studies addressing risk by trimester, so to say what could be occurring at 17-19 weeks from ultrasound exposure would be hypothetical. At that time, the cortex is still developing and may therefore still be vulnerable to teratogenic agents like ultrasound. But, as we've said, studies are definitely needed to address this issue, and preferably studies which occur in the laboratory and ones which are not solely epidemiological since their sensitivity to subtle phenomena can be limited. 2) Regarding unnecessary ultrasounds within the first trimester, anecdotally we've noted a surprising number of women whose doctors (different doctors from different locations) have really pushed for early ultrasound in non-risk pregnancies. In fact, Dr. Casanova's two eldest daughters have each received early and unwarranted ultrasounds simply to date their pregnancies. (Something they told him after the fact.) Multiple unwarranted ultrasounds are also a considerable problem. 3) At present, guidelines for ultrasound recommend no screening in the first trimester and no more than 2 per pregnancy. At present, the average is over 3, close to 4, and that is generally no risk pregnancies. In addition, heart rate monitors are available as well as keepsake images/videos from private businesses who offer these services. Their are guidelines but these guidelines are not being followed unfortunately. As an extreme example, a friend of mine was considered an at-risk pregnancy due to diabetes: she received ultrasound every week for the third trimester. 4) As for Doppler fetal heart rate monitors, I'm uncertain as to the comparative energy output, however one neurodevelopmental animal study in particular showed more extreme effects on behavior using the pulse doppler ultrasound versus other forms of ultrasound.

Thanks for your questions. Please let us know if you have anymore.
Anonymous said…
Now this is really interesting. I'm wondering if we can measure the amount of cavitation generated by inserting nanodiamond particles into the effected cells? If I can image them, maybe I can very precisely understand what the effect of ultrasound is at the cellular level? Another potential application for Dr. Auciello's work at Argonne.
Anonymous said…
Guess the concern I have is in tarring ultrasound unfairly given that additional work needs to be done to clarify what the effects really are? For example, using early diagnosis, we are now able to carry out surgery on fetuses who have been diagnosed intra-euterally with spina bifida. This is a major win win, because of the debilitating nature of this illness and ultrasound is a key technique for verifying the presence of the spinal defect. The truth is, we need sensoring capability that can precisely quantify the changes the body experiences in diagnostic situations and those types of sensoring techniques lie right at the edge of quantum biophysics and quantum entanglement phenomena.
Emily Williams said…
We definitely agree that more work needs to be done. However, there is enough evidence to raise genuine concerns about its current state of use. Ultrasound has definitive benefits to healthcare and so this is not an instance in which to throw out the baby with the bathwater. But exercising greater caution in the application of this tool is warranted until at which time we can confirm there is no risk to the developing infant. This means minimizing the number and duration of exposures to that which is absolutely necessary. Such an approach should not prevent the diagnosis and surgical treatment of conditions like spina bifida. In the field of medicine, it is generally better to overexercise caution than to place patients wellbeing at risk. For instance, we use radiation for imaging but it is still dangerous to overuse it.
Emily Williams said…
We definitely agree that more work needs to be done. However, there is enough evidence to raise genuine concerns about its current state of use. Ultrasound has definitive benefits to healthcare and so this is not an instance in which to throw out the baby with the bathwater. But exercising greater caution in the application of this tool is warranted until at which time we can confirm there is no risk to the developing infant. This means minimizing the number and duration of exposures to that which is absolutely necessary. Such an approach should not prevent the diagnosis and surgical treatment of conditions like spina bifida. In the field of medicine, it is generally better to overexercise caution than to place patients wellbeing at risk. For instance, we use radiation for imaging but it is still dangerous to overuse it.
Unknown said…
I had the same question before but one thing is certain.. ultrasound is of great use to health care..

Ultrasound Technician Salary in Alaska
Anonymous said…
This sound plausible, and is certainly worthy of investigation.

I particularly note that it seems like a potential candidate for the cause for the California Austism clusters, identified from birth locations of those later diagnosed as Austistic.
Emily Williams said…
We have an article currently in press which should be released from the International Journal of Developmental Neuroscience within the coming months. It goes into a thorough but more general review on the potential teratogenicity of prenatal ultrasound and may help people to understand some of the biophysical mechanisms of the tool. It's entitled, "Reassessment of Teratogenic Risk from Antenatal Ultrasound".
meghan said…
Hypersensitivity to sound is an extremely common trait among autistic individuals. Could it be some kind of post traumatic stress trigger related to what they were subjected to in utero? Or do you think the intense sound waves of the ultrasound could have altered the way they perceive and process sound in general?
Emily Williams said…
Ultrasonic waveforms generally are not audible to the human ear. But they are a form of physical force (i.e., "sound" is interpreted by the ear from a domino of particles colliding with one another until they finally collide with the sensory receptors within the ear causing particular vibrations which our brain then interprets as sound). Placing such force on the developing embryo or fetus is more our concern and not so much promoting an unpleasant sound-sensory experience for the baby. Such force can damage cell membranes, cause local temperature increase, promote free radical production, and cause what is called "microstreaming" of the liquids within and surrounding the cell; all of these are dependent upon the intensity of the ultrasound beam (higher intensity = greater risk), frequency (lower frequency = greater risk), and duration of exposure (longer duration = greater risk).
Are there any studies that show the rates of autism in children whose mothers did not have any scans at all ie I am interested to know if autism still occurs in non scanned women and if so at what percentage of births. Intresting research and seems to make a lot of sense from your explanation. Will be following your results thanks for sharing them here.
this is a little offline of topic, but has there been studies done in countries where ultrasounds arent used at all? how high are autism rates there? how high are any issues there as far as children go?
Team MacKenzie said…
As a mother to a perfectly healthy 3 year old who had more than 30 ultrasounds during my pregnancy, I ask you to be very careful when talking about ultrasound risks. Pregnant woman have enough to worry about without having to be scared to have, sometimes very necessary ultrasounds, as well. My daughter would not be alive today if it were not for the ultrasounds I had during my very high risk pregnancy.
Anonymous said…
I agree, caution is absolutely necessary when discussing whether to change medical intervention in any drastic way. And this is not a proposal for the dissolution of ultrasound use in obstetrics, particularly in at-risk cases. This is solely a work of caution and to bring a better understanding of this tool which can sometimes be used to freely. As much as parents such as yourself would caution scientists to weigh their postulates seriously and wisely (a wise caution in itself), I also caution parents in taking single case examples as proof of no effect. One case, especially one's own, may be extraordinarily moving but it's not good science nor skepticism to base such on outlook on individual examples when science is built off of studying populations. This work is simply proposing that, given the risks inherent to ultrasound use, particularly early and frequent, and the neuroanatomical development particular to autism (e.g., proliferative abnormalities), there is overlap and so we have raised the question for further study. In the meantime, we have also proposed increased caution when using this technique JUST IN CASE it does have deleterious effects we were previously unaware of. Does anyone here truly feel that application of moderation is unadvisable in this instance, given that we still don't know enough about ultrasound's effects on development to say for certain it is innocuous? Does any parent here truly want to take that kind of risk? Ultrasound has its uses, especially for at-risk pregnancies as well as for checking on overall mother and fetal health closer to the due date. But we have gotten into too much of a habit, parents and doctors alike, in just wanting to have photographs of our babies. In my mind, and that of my colleagues, that is not a good enough reason to put a child at risk.

CoAuthor: Emily L. Williams
Anonymous said…
Ultrasound is really good especially for pregnant women to see their babies inside their tummies.. But I know there might be some consequences for it that we really don't know.
thanks, ultrasound technician salary
Unknown said…
Ultrasound is a vital and useful tool in medicine. Is ultrasound safe for the developing fetus? I hope that more large scale studies can be done to address this question in more depth. Thanks for sharing your opinions and knowledge!
harvoa said…
50 Human studies, in utero, modern studies, indicate that ultrasound is very hazardous to the fetus. The alternative, a life without ultrasound, needs much discussion. And much research must be conducted regarding the politics of medicine and childbirth. Details:
Rachel R. said…
A call for caution is not remotely close to "fear-mongering" among pregnant women. No sane person would recommend that we all get x-rayed regularly just because we can. But it would be completely unreasonable to conclude that someone who says, "we recommend against getting x-rayed regularly just because you can" is saying, "if you think you broke your leg, don't get an x-ray." These are clearly two very different scenarios.

When ultrasound technology was developed, it was intended for diagnostic use in situations where there was a suspected problem that ultrasound could reasonably detect. Today it has become simply a routine thing we take for granted. "Routine" is okay for something that carries no risk, but it's foolish for something that is either proven risky, or for which we are unsure of the risk. To make matters worse, it hasn't even been demonstrated to be accurate at detecting many of the things we assume it to be accurate for. (There's a useful summary here:

Popular Posts