This article [see below] just appeared this week on Medscape — and it is timely. I agree with the author that autism spectrum disorder (ASD) [a] has increased in frequency of diagnosis over the past five decades, [b] is an example of gene-environment interactions, [c] represents a good example of a multiplex complex human disease, and [d] the predominant component is genetic predisposition — although one or more environmental factors might tip the balance in “a genetically susceptible” individual, resulting in manifestation of the full-blown disease (phenotype, trait).
—DwN
We Are Seeing More Autism, and It’s Mostly Genetic
F. Perry Wilson, MD, MSCE
DISCLOSURES | 21 April 2025
C52326757-Wilson-IF-4-21-25-VIM
This transcript has been edited for clarity.
Welcome to Impact Factor, your weekly dose of commentary on a new medical study. I’m Dr F. Perry Wilson from the Yale School of Medicine.
The issue I’m discussing this week is bound to be sensitive, so let me try to focus just on the facts. Yes, there are many more diagnosed cases of autism now than there have been in the past.
photo of Autism Diagnosis Among US Children and Adults
The US Secretary of Health and Human Services (HHS), Robert F. Kennedy, Jr, cites that fact and suggests that it implies the existence of an environmental cause of autism— as of yet unidentified — that, through careful research, can be identified, perhaps as early as September. The focus on an environmental explanation for the increase runs counter to a recent Centers for Disease Control and Prevention (CDC) report suggesting that the increase is due to better and earlier identification. RFK Jr referred to that line of evidence as a “canard” — a useful excuse to protect entrenched interests that are actually the culprits. And although I think we know what entrenched interests he may be referring to, I’ll give him credit that his public statements keep an open mind about what environmental factors might be at play.
I thought, therefore, that it might be a good time to look at what we already know about environmental versus genetic triggers for autism to set expectations on what an HHS-sponsored investigation may be able to find. In other words, let’s look at the data. Autism: genes versus environment.
There’s a pretty interesting way to quantify how much of a disease may be caused by things external to our genetic makeup — versus inherent to that makeup: twin studies. And the intriguing fact is, these studies can tell us how much of a disease may be due to an environmental trigger — even when we don’t know what that trigger might be.
To show you how this works, I’ll give two extreme examples.
Let’s look at a condition that is basically purely genetic — Huntington disease — and examine how it should work among different types of twins.
Identical (monozygotic) twins share virtually 100% of their genetic code. Since Huntington disease is purely genetic, you’d expect perfect concordance among identical twins [Table 1], regardless of whether they were raised in the same environment or not. There are actually a few case reports of identical twins raised apart — a rare occurrence — where both developed Huntington disease, as expected.
Table 1
photo of Twin study of Huntington Disease
In contrast, fraternal (dizygotic) twins, share 50% of the same genetic information (just like nontwin siblings). As such, we’d expect some concordance between fraternal twins in terms of Huntington disease (since it’s a dominant gene, 50% of children of an affected parent are affected), but not perfect concordance. And, similarly, it shouldn’t matter whether these twins were raised in the same environment or not. I’m including the in utero environment here — basically, “environment” means anything not genetic [Table 1].
Almost no disease is purely environmental, because genetic variation might change the susceptibility to environmental toxicants, but as an example of the other side [Table 2], let’s take something like scarlet fever. This twin study examined concordance rates for scarlet fever between identical and fraternal twins and found them to be basically the same. In other words, susceptibility to scarlet fever is driven almost entirely by the environment in which the kids are raised, not by their genetic similarity.
Table 2
photo of Twin study of Huntington Disease
Then consider something like type-2 diabetes (T2D), which has both strong genetic and environmental components. Twin studies show a higher degree of concordance among identical twins than fraternal twins, but data from twins raised apart suggest a substantial environmental component [Table 3]. Studies vary, but most end up attributing about 60% of the risk to genetic and 40% to environmental exposures. And, to be clear, when I write “environment” here, I don’t mean toxicants in the air or something — I mean the environment we have created, which is loaded with calorie-dense, ultra-processed foods.
Table 3
photo of Twin study of Huntington Disease
So, what about autism? If it is caused by some environmental exposure — a vaccine or vaccine ingredient, a pollutant, energy beams from space, whatever — we should see similar concordance between identical and fraternal twins, provided they were raised together. What do we actually see? Concordance rates among identical twins range from 60%-90% and, for fraternal twins, from about 3% to 30%. That’s a strong genetic signal.
There are quite a few studies out there that dig in even deeper, but this one in JAMA is pretty solid since it comes from Sweden where, as I have mentioned multiple times before, healthcare is universal and so are electronic medical records. For what it’s worth, the estimates from this study are pretty similar to those from other studies.
Swedish researchers examined about 40,000 twin pairs, as well as 2.5 million sibling pairs and about a million half-sibling pairs. The data suggested that 83% of autism diagnoses could be attributed to genetics, based on a high degree of concordance between identical twins versus fraternal twins [Table 4]. About 4% could be attributed to what are called “shared environmental factors” — these are things that siblings are both exposed to — for fraternal twins that includes everything from intrauterine on up, until they leave the nest.
Table 4
photo of Twin study of Huntington Disease
In that study, 16% of the diagnoses could be explained by non-shared environmental factors. This is interesting; it reflects a concordance in relatives after accounting for shared genetics even if they weren’t raised in the same environment. This could reflect some interesting epigenetic causes of autism or exposures (like certain illnesses) that aren’t highly shared between siblings.
I think there is something for everyone in these data. It’s quite clear that the majority of cases of autism can be attributed to genetic factors, even if we don’t yet know entirely what those factors are. That does support the hypothesis that the observed increase in cases is due to more or better diagnosis — we just missed these kids in the past, or diagnosed them with something else.
But autism spectrum disorder is not Huntington disease. There clearly are environmental factors at play as well; not every pair of identical twins is concordant for autism, even if 90% are. And, even if there is concordance, the severity may differ between identical twins, indicating some gene-environment interactions.
So, is something in the environment the major factor leading to the observed increase in cases of autism spectrum disorder? Probably not, but it should not be ignored — just because it isn’t the sole explanation. We should try to figure out what these other factors are.
Which means that now, the real work begins, because “environmental factors” is a big tent that includes essentially everything that isn’t genetic: mom and dad’s medical history, toxicants and pollutants, infections, medications, and so on. Twin studies are not the way to figure out which of these is the culprit — more dedicated, cause-specific research is. I, for one, am happy to see some funding increases targeted in this area, but I will be quite surprised, and honestly skeptical, if we “get an answer by September.”
F. Perry Wilson, MD, MSCE, is an associate professor of medicine and public health and director of Yale’s Clinical and Translational Research Accelerator. His science communication work can be found in the Huffington Post, on NPR, and here on Medscape. He posts at @fperrywilsonand his book, How Medicine Works and When It Doesn’t, is available now.
COMMENTS:
The estimated occurrence of “autism” in the U.S. in 1960 was ~1 in 2,500, or 0.04% of the population. Yours truly probably saw 15,000 pediatric patients during the 1960s-70s and can recall two cases of “autism.” In 2024, the Centers for Disease Control and Prevention (CDC) now estimates that 1 in 36 children in the U.S. diagnosed with Autism Spectrum Disorder (ASD); this extrapolates to ~2.3% of all American children being autistic (with males identified at a rate four times higher than females). These data would suggest that the frequency of autism has increased more than 50-fold over the last six decades.
As stated by Linda Birnbaum and Paul Stidoph [below], ASD is a human complex disease reflecting the combination of genes and the environment; in other words, there is a multiplex genetic component that is likely influenced by multiple environmental factors. Or, alternatively, ASD is a “character trait created by our more technical environment.”
Unlike a monogenic (Mendelian) simple (dominant or recessive) trait that is largely the manifestation of a single gene — human complex diseases (i.e., phenotypes, traits) are usually multifactorial quantitative traits, influenced by [a] numerous small-effect variants of genes (alterations in DNA sequence); [b] epigenetic effects (DNA methylation, RNA interference, histone modifications, chromatin remodeling); [c] environmental factors (smoking, occupation, lifestyle); [d] endogenous influences (ancestral differences, diseases of e.g., heart, lung, kidney, brain); and [e] each person’s unique microbiome.
For example, studies on the multifactorial quantitative trait “height” have now identified >12,000 small-effect genetic variants linked to height — which explain 10% to 40% of the variability in human height, depending on ancestry. Hundreds of ASD studies have implicated several hundred gene variants and DNA sequence structures (e.g., numbers of repeats in some repetitive DNA regions) involved in the trait (phenotype) of ASD.
ASD might also be considered as “subtle alterations in the patient’s genomic architecture.” The “genomic architecture” refers to random organization of functional elements (e.g., genes and regulatory regions) within the genome; i.e., it’s not just a linear sequence, but also includes the 3-dimensional arrangement of DNA in the nucleus, affecting how genes are expressed and how cells function.
ASD might also be considered as “subtle alterations in the patient’s whole-genome phenomics (WGP).” WGP is a field that studies the association between an organism’s complete genetic makeup (genome) and its observable characteristics and behaviors (phenotype, phenome). WGP involves analyzing high-dimensional phenotypic data on a genome-wide scale — considering the complex interplay between genes and their expression in various traits; this field aims to understand the genetic basis of phenotypes, including complex traits and diseases, and might eventually be used to develop personalized clinical recommendations, as well as improve plant and animal breeding.
If the frequency of autism has increased more than 50-fold over the past six decades — WHAT (in our “total ENVIRONMENT”) might we consider that has also increased ~50-fold between 1965 and 2025? First and foremost, the 1960s-criteria for diagnosing “autism” were extremely limited; today, the basket of symptoms has expanded dramatically, and the label of this complex disease is “autism spectrum disorder (ASD).”
Second, the highest suspect in some people’s minds is the media (e.g., social media, television, movies); it seems that every child from age 3 onward carries his own cell phone and internet tablet — for (“24-7”) watching extremely fast-moving violent noisy adventures and cartoons — compared with a Doris Day or Cary Grant movie or Disney cartoon of 1960.
Third, environmental chemicals, as mentioned below by Linda, (“air pollution, certain pesticides, and various endocrine disrupting compounds”) are in the mix. For example, PFAS (Per- and Polyfluoroalkyl Substances) represent a group of thousands of synthetic chemicals that have been implicated by some in causing the ASD increases; however, these chemicals have been widely used (due to their resistance to grease, oil, water, and heat) in various products since the 1940s — which predates by several decades the dramatic increase in ASD risk frequency.
Lastly, microplastics and food dyes should be mentioned, in rounding out this discussion. In conclusion, the genetics of ASD are exceedingly complex, and associations with possible environmental factors are likewise exceedingly complex.
—DwN
Hiroto T, King BH. Autism spectrum disorder: A review. J Am Med Assn. 10 Jan 2023; 329: 157-168 —AFTER writing all the above, I checked and I found that this review is the latest, out of 493 PubMed reviews on this topic. ☹
Sent: Wednesday, April 30, 2025 6:46 AM
Hello Dan
Fair enough, but I do not accept that ASD is a disease. To my mind it is a character trait created by the environment. It has been made more manifest because our environment, as children and adults, has become more technical. There are significant advantages for children under the British social security scheme if they are diagnosed with ASD — in terms of benefits in cash and kind. We are stoking the fire. Perhaps it is the same in the USA…??
Paul
Sent: Sunday, April 27, 2025 7:03 PM
Dan, I couldn’t agree more that ASD is multifactorial, complex, and very much multiple health conditions, and is essentially ALL genetic AND environmental. The twin studies have clearly shown a high heritability role. And there are hundreds of genes that have been implicated. We also have strong associations for various environmental factors, including air pollution, certain pesticides, and various endocrine disrupting compounds.
Differential diagnosis plays a role in the increase in prevalence from 1/150 in 2000 to 1/31 today. But it is NOT the entire story, and as we all know that our genes don’t change that rapidly, but our environment does.
Linda
Sent: Sunday, April 27, 2025 4:22 PM
This article [see below] just appeared this week on Medscape — and it is timely. I agree with the author that autism spectrum disorder (ASD) [a] has increased in frequency of diagnosis over the past five decades, [b] is an example of gene-environment interactions, [c] represents a good example of a multiplex complex human disease, and [d] the predominant component is genetic predisposition — although one or more environmental factors might tip the balance in “a genetically susceptible” individual, resulting in manifestation of the full-blown disease (phenotype, trait).
—DwN
We Are Seeing More Autism, and It’s Mostly Genetic
F. Perry Wilson, MD, MSCE
DISCLOSURES | 21 April 2025
C52326757-Wilson-IF-4-21-25-VIM
This transcript has been edited for clarity.
Welcome to Impact Factor, your weekly dose of commentary on a new medical study. I’m Dr F. Perry Wilson from the Yale School of Medicine.
The issue I’m discussing this week is bound to be sensitive, so let me try to focus just on the facts. Yes, there are many more diagnosed cases of autism now than there have been in the past.
photo of Autism Diagnosis Among US Children and Adults
The US Secretary of Health and Human Services (HHS), Robert F. Kennedy, Jr, cites that fact and suggests that it implies the existence of an environmental cause of autism— as of yet unidentified — that, through careful research, can be identified, perhaps as early as September. The focus on an environmental explanation for the increase runs counter to a recent Centers for Disease Control and Prevention (CDC) report suggesting that the increase is due to better and earlier identification. RFK Jr referred to that line of evidence as a “canard” — a useful excuse to protect entrenched interests that are actually the culprits. And although I think we know what entrenched interests he may be referring to, I’ll give him credit that his public statements keep an open mind about what environmental factors might be at play.
I thought, therefore, that it might be a good time to look at what we already know about environmental versus genetic triggers for autism to set expectations on what an HHS-sponsored investigation may be able to find. In other words, let’s look at the data. Autism: genes versus environment.
There’s a pretty interesting way to quantify how much of a disease may be caused by things external to our genetic makeup — versus inherent to that makeup: twin studies. And the intriguing fact is, these studies can tell us how much of a disease may be due to an environmental trigger — even when we don’t know what that trigger might be.
To show you how this works, I’ll give two extreme examples.
Let’s look at a condition that is basically purely genetic — Huntington disease — and examine how it should work among different types of twins.
Identical (monozygotic) twins share virtually 100% of their genetic code. Since Huntington disease is purely genetic, you’d expect perfect concordance among identical twins [Table 1], regardless of whether they were raised in the same environment or not. There are actually a few case reports of identical twins raised apart — a rare occurrence — where both developed Huntington disease, as expected.
Table 1
photo of Twin study of Huntington Disease
In contrast, fraternal (dizygotic) twins, share 50% of the same genetic information (just like nontwin siblings). As such, we’d expect some concordance between fraternal twins in terms of Huntington disease (since it’s a dominant gene, 50% of children of an affected parent are affected), but not perfect concordance. And, similarly, it shouldn’t matter whether these twins were raised in the same environment or not. I’m including the in utero environment here — basically, “environment” means anything not genetic [Table 1].
Almost no disease is purely environmental, because genetic variation might change the susceptibility to environmental toxicants, but as an example of the other side [Table 2], let’s take something like scarlet fever. This twin study examined concordance rates for scarlet fever between identical and fraternal twins and found them to be basically the same. In other words, susceptibility to scarlet fever is driven almost entirely by the environment in which the kids are raised, not by their genetic similarity.
Table 2
photo of Twin study of Huntington Disease
Then consider something like type-2 diabetes (T2D), which has both strong genetic and environmental components. Twin studies show a higher degree of concordance among identical twins than fraternal twins, but data from twins raised apart suggest a substantial environmental component [Table 3]. Studies vary, but most end up attributing about 60% of the risk to genetic and 40% to environmental exposures. And, to be clear, when I write “environment” here, I don’t mean toxicants in the air or something — I mean the environment we have created, which is loaded with calorie-dense, ultra-processed foods.
Table 3
photo of Twin study of Huntington Disease
So, what about autism? If it is caused by some environmental exposure — a vaccine or vaccine ingredient, a pollutant, energy beams from space, whatever — we should see similar concordance between identical and fraternal twins, provided they were raised together. What do we actually see? Concordance rates among identical twins range from 60%-90% and, for fraternal twins, from about 3% to 30%. That’s a strong genetic signal.
There are quite a few studies out there that dig in even deeper, but this one in JAMA is pretty solid since it comes from Sweden where, as I have mentioned multiple times before, healthcare is universal and so are electronic medical records. For what it’s worth, the estimates from this study are pretty similar to those from other studies.
Swedish researchers examined about 40,000 twin pairs, as well as 2.5 million sibling pairs and about a million half-sibling pairs. The data suggested that 83% of autism diagnoses could be attributed to genetics, based on a high degree of concordance between identical twins versus fraternal twins [Table 4]. About 4% could be attributed to what are called “shared environmental factors” — these are things that siblings are both exposed to — for fraternal twins that includes everything from intrauterine on up, until they leave the nest.
Table 4
photo of Twin study of Huntington Disease
In that study, 16% of the diagnoses could be explained by non-shared environmental factors. This is interesting; it reflects a concordance in relatives after accounting for shared genetics even if they weren’t raised in the same environment. This could reflect some interesting epigenetic causes of autism or exposures (like certain illnesses) that aren’t highly shared between siblings.
I think there is something for everyone in these data. It’s quite clear that the majority of cases of autism can be attributed to genetic factors, even if we don’t yet know entirely what those factors are. That does support the hypothesis that the observed increase in cases is due to more or better diagnosis — we just missed these kids in the past, or diagnosed them with something else.
But autism spectrum disorder is not Huntington disease. There clearly are environmental factors at play as well; not every pair of identical twins is concordant for autism, even if 90% are. And, even if there is concordance, the severity may differ between identical twins, indicating some gene-environment interactions.
So, is something in the environment the major factor leading to the observed increase in cases of autism spectrum disorder? Probably not, but it should not be ignored — just because it isn’t the sole explanation. We should try to figure out what these other factors are.
Which means that now, the real work begins, because “environmental factors” is a big tent that includes essentially everything that isn’t genetic: mom and dad’s medical history, toxicants and pollutants, infections, medications, and so on. Twin studies are not the way to figure out which of these is the culprit — more dedicated, cause-specific research is. I, for one, am happy to see some funding increases targeted in this area, but I will be quite surprised, and honestly skeptical, if we “get an answer by September.”
F. Perry Wilson, MD, MSCE, is an associate professor of medicine and public health and director of Yale’s Clinical and Translational Research Accelerator. His science communication work can be found in the Huffington Post, on NPR, and here on Medscape. He posts at @fperrywilsonand his book, How Medicine Works and When It Doesn’t, is available now.