Why would the pediatric community not know about the study by Frisch and Simonsen, published in 2015? That study showed more than twice as much infantile ASD associated with circumcision, which is VERY concerning! (You can read all about this in our previous blog)
Sometimes scientific results can be straight forward, but the way in which humans respond to the science can be a bit convoluted. This is a fascinating example. Let’s dive in.
Making a Word Salad
Morris and colleagues wrote a review about the pros and cons of circumcision, published in 2017, just two years after the Frisch and Simonsen study. (https://pmc.ncbi.nlm.nih.gov/articles/PMC5478224/). That article is, I think, very informative. They dismissed any possibility that circumcision could cause autism, citing a study by Sneppen and Thorup, who “found ASD prevalence was 7.2% in uncircumcised Danish boys and suggested Frisch's study suffered from confounding”.
This statement by Morris and colleagues is basically a “word salad”, akin to those commonly used in politics, that makes no sense. It’s fascinating to see such a salad in a scientific article.
Breaking Down the “Salad”
In a nutshell, what Morris and colleagues are saying is that Danish boys without circumcision have an incredibly high prevalence of autism according to Sneppen and Thorup, so therefore, because they didn’t find high levels of ASD, Frisch and Simonsen must have been mistaken. Morris and colleagues also mention the term “confounding” as part of their word salad, and we will need to come back to that later.
On the surface, the argument by Morris and colleagues doesn’t make a lot of sense. Frisch and Simonsen looked at the RATIO of autism in circumcised versus uncircumcised boys. The fact that somebody else found much more autism in another study should not be an immediate cause for throwing out the study by Frisch and Simonsen. Finding much higher levels of ASD in a different study is definitely concerning, suggesting that Frisch and Simonsen might have missed lots of cases of ASD. But it’s not concerning enough to throw out their study. For the Frisch and Simonsen study to be in error, the authors would have somehow had to make a mistake in measuring ASD in uncircumcised boys, but not in circumcised boys. It is difficult to imagine how this could have happened, and certainly the study cited by Morris and colleagues does not provide any clues.
It There a Reasonable Prevalence?
But let’s follow the trail laid down by Morris and Colleagues: Let’s look at Sneppen and Thorup’s study that found 7.2% of all uncircumcised boys with autism (https://pubmed.ncbi.nlm.nih.gov/27244821/). Sneppen and Thorup were looking at 137 uncircumcised boys in Denmark, born between 1996 and 2013 (average about 2005) when they were 10 years old, on average. Frisch and Simonsen, by comparison, looked at almost 340,000 uncircumcised boys born between 1994 and 2003, and found a prevalence of about 1.5% for boys born around the same time frame as the boys in the Morris study. It’s important to note here that Frisch and Simonsen were looking at the entire national register for all children born in Denmark. They were literally looking at the entire population, not a small sample.
Is 7.2% a reasonable prevalence of ASD? That’s about one in every 14 boys. Definitely not. No way. By comparison, we were very shocked in 2023 when the prevalence of autism in the US reached about 4.3% of all boys (https://pubmed.ncbi.nlm.nih.gov/36952288/), or one in every 23. Here I’ll point out that the prevalence of autism in Denmark has been studied carefully through time: It’s steadily climbing, the same as in most Western countries, but was about 1.3% for all 10-year old children (boys and girls) born in 2003 (https://pmc.ncbi.nlm.nih.gov/articles/PMC6248103/), which corresponds to about 2.1 % of the boys, both circumcised and uncircumcised, with autism (assuming the male/female ratio of ASD prevalence is 4/1). This is close to the number reported by Frisch and Simonsen, which was 1.5%.
Is It Confounding?
So, how did Sneppen and Thorup find that 7.2% of typical, uncircumcised boys have autism in Denmark? The answer is… that is not exactly what they found. First, the prevalence of 7.2% was found for all boys, with and without circumcision. All boys had been sent to the surgery department because of problems with their penis, most (95%) because of a problem called “phimosis” that is usually treated effectively with anti-inflammatory drugs. More than a quarter of the boys had “severe voiding problems”, which often involves pain when urinating, and more than a quarter of the boys were suspected of having inflammation of their penile tissue. These were not average 10-year old boys in Denmark, who at that time had a prevalence of ASD around 1.5 or 2%. These were boys who had more pain and inflammation than the average boy. We expect them to have more ASD than average boys because inflammation and autism are connected, and because pain management with acetaminophen is connected with autism. It’s not surprising that these boys had so much autism.
What about “confounding”? Morris and colleagues seemed to be worried that something was “confounding” the study of Frisch and Simonsen. The issue of confounding is difficult to understand in the context used by Morris and colleagues, but in general it means that something unrelated to circumcision must have been at the root of the increased prevalence of ASD observed by Frisch and Simonsen in circumcised boys. Other factors that may have caused the association between circumcision and ASD in the Frisch and Simonsen study are discussed in the blog on circumcision and ASD. No reasonable candidates for confounding exist, and certainly none were suggested by Morris and colleagues. Basically, in this case, “confounding” amounts to an accusation with no evidence at all. Not even a suspect.
Why a Study May Be Disregarded
The big question is my mind is, why did Morris and colleagues, reviewing the pros and cons of circumcision in 2017 (https://pmc.ncbi.nlm.nih.gov/articles/PMC5478224/), disregard the population-wide study by Frisch and Simonsen in 2015 (https://pmc.ncbi.nlm.nih.gov/articles/PMC4530408/) because of a small study that found a very high rate of ASD in patients with significant inflammatory problems?
That’s the same as ignoring an ovenfire burning in my kitchen because I have a candle burning in the bedroom. Yes, fire is involved in both cases, but one has literally nothing to do with the other. The argument is nonsensical.
I don’t know with any certainty why the Frisch and Simonsen study is largely ignored. In my experience in discussing the study with physicians, it’s often disregarded when somebody literally thinks that it can’t possibly be true. It’s literally inconceivable. Based on their writing, I think even Frisch and Simonsen were a bit shocked. Certainly they ran an extraordinary number of double-checks to see if the numbers were real. The numbers are real. And if you review the blog about circumcision and ASD, you’ll see that everything makes absolutely perfect sense.
Another example of a Word Salad that Obscures Science
In the process of evaluating evidence, before we include it in our list of accepted evidence, we carefully evaluate the original studies. In addition, we often see what other people say about the study in the medical literature. The Schultz study in 2008 (https://pubmed.ncbi.nlm.nih.gov/18445737/) is a great example of another study that, like the Frisch and Simonsen study, was largely ignored. The study by Schultz brilliantly showed that acetaminophen given with a vaccine, not the vaccine itself, was associated with ASD. Soon after that paper was published, a detailed criticism on the Schultz article that can only be described as a “hack-job” was published, allegedly pointing out numerous errors in Schultz’s work. The rebuttal had no scientific validity, amounting to a giant word salad with no substance. Once we discovered Schultz’s work and the criticism of that work, we published an analysis of the situation here: https://pubmed.ncbi.nlm.nih.gov/37321575/, showing that the word salad attacking the Shultz study had absolutely no validity, The tragedy is that, at least in part because of this invalid criticism, the Schultz study is still often disregarded in a similar manner to the circumcision study.
Word Salads Confuse Communities and Disregard Consideration of New Scientific Knowledge
Despite the fact that the hack job on Schultz’s study had no valid substance, it confused the issue enough so that people disregarded it, and most people forgot about it. Indeed, I saw a recent article published in a radical magazine describing a case of regressive ASD in a 5-year old boy following multiple, simultaneous vaccinations. The authors of the article mentioned that the child was in the hospital for 6 days with a high fever following the vaccinations. If they had been aware of Schultz’s work published 15 years ago, they would have focused on acetaminophen used in an attempt to treat the fever, not the vaccine itself. We are still dealing today with an issue that should have been laid to rest more than 15 years ago by Stephen Schultz.
We humans are built with a need to understand our world, which manifests in us as a “confirmation bias”. If something doesn’t add up, if we can’t see how it could possibly be correct based on everything we know, we explain it away and toss it into the garbage heap of our memories. We search for any reason to disregard or explain away something that doesn’t make sense. That’s the fundamental nature of a confirmation bias. We prefer, if possible, to have a valid reason to disregard an observation, but almost any reason for dismissal of things that don’t make sense to us can be accepted. This allows us to get on with life without getting bogged down with uncertainty and worry. We like a world that makes sense. That’s my guess as to why, based at least in part on a reason that makes no sense, the study by Frisch and Simonsen has been ignored.
In a nutshell, it seems likely that people are following Sir Arthur Conan Doyle’s saying “When you have eliminated the impossible, whatever remains, however improbable, must be the truth.” They’ve decided that circumcision can’t be causing autism, so therefore the Frisch and Simonsen study must be invalid, even if nobody can figure out why.
In reality, the Frisch and Simonsen study is perfectly valid, and we understand that the results of their study are not only possible, they’re predictable.
One of the cool things about science is that the observations that do NOT make sense are often the ones that lead to discovery.