And we think we might know exactly why.

Recent data indicate that as much as thirty percent of black boys born in 2019 and 2020 in Los Angeles County in California have autism ^[1]. That's correct… not three percent. Ten times three percent. Almost one in every three boys. And, if the plausible explanation I'll describe here is correct, that level of autism is preventable.

The most recent data on the prevalence of autism was just released from California and published by a team out of Rutgers ^[1]. The data are so new that they haven't even been peer reviewed yet. It's published in what we call a "preprint".  Preprints can be super helpful in getting important information out to the public quickly, especially since the scientific peer review process can take years.

The new data show profound changes in the prevalence of autism, starting in birth years 2015 or 2016. After 2015-2016, the prevalence of autism literally doubled within 5 years ^[1]. After forty years of increases, starting slowly in the 1980s, accelerating in the 1990s, and steadily increasing through the 2000s, suddenly, between 2015 and 2020, the prevalence literally doubled.

But only in some groups of children. Whatever is going on is affecting children with low socioeconomic status more than others ^[1]. And it's affecting black children the most ^[2]. Los Angeles appears to be the worst hit. Again, it appears that up to thirty percent, almost one in every three, of the black boys born in 2019 and 2020 is Los Angeles County have autism.

The figure, taken from the preprint my group published ^[2], shows an example of some of that data from the Rutgers study, with autism prevalence measured by county and race/ethnicity. In some wealthier counties, the pattern of increase starting after 2016 is noticeably absent ^[1]. The data shown are an average of the prevalence in boys and girls, and the methods used capture only more severe cases of autism. Since boys tend to have autism more frequently than girls, and since these data miss many mild cases of autism, the actual prevalence of autism in boys is much higher than shown in the graph ^[1]. The team at Rutgers estimates that the prevalence of autism in black boys born in Los Angelas in 2019 and 2020 is approaching thirty percent ^[1]. The data from children born in 2020 are not complete yet, so those are not shown in the graph.  

The burning question, of course, is why? What is causing this effect?

As a chemist working on the underlying cause of autism for over a decade, these new data from California quickly brought three facts to mind.

1. Based on overwhelming evidence, we have already concluded that exposure of susceptible babies and children to acetaminophen (Tylenol) induces many if not most cases of autism ^[3-5].

2. Acetaminophen and marijuana share something in common: they interact with the same biochemical pathways in the brain (the cannabinoid system) to affect pain.

3. California made recreational marijuana legal about the same time that the uptick in autism started, in late 2016.

So, a hypothesis, an educated guess, was immediately evident. Maybe marijuana interacts with acetaminophen to make the acetaminophen even more dangerous for brain development than it already is?

Some background work was required. Was this a reasonable hypothesis? After some searching and numerous discussions with colleagues in the fields of medicine, toxicology, pharmacology and neuroscience, the following evidence came to light.

4. A 2019 study from Arkansas Children's Hospital showed that non-lethal doses of acetaminophen combined with a marijuana extract were lethal in laboratory mice. Neither one by itself was lethal, but together they were lethal ^[6]. Marijuana is a potent modulator of human metabolism, so an adverse drug-drug interaction involving marijuana would not be surprising ^[7].

5. Several studies show that occasional marijuana use is NOT associated with autism, but one 2024 study showed that frequent daily maternal use of marijuana IS associated with a much greater risk of autism… a 4 to 5 fold increase in the risk of having a child with autism ^[8].

6. Very frequent use of marijuana is strongly associated with lower socioeconomic status ^{[9, 10]}, and blacks in California are much more likely than others to visit the emergency department as a result of marijuana overexposure ^[11].

7. Exposure to marijuana smoke profoundly increases oxidative stress ^[12], which makes acetaminophen much more toxic ^{[13, 14]}.

8. Historically, autism was not connected with communities who used marijuana extensively. That is one of the reasons we do not think that marijuana, by itself, even with very frequent use, can cause autism. If marijuana is causing autism, it is likely because it enhances the adverse effects of acetaminophen on brain development.

9. We can perform a quick calculation to see if it's reasonable that marijuana could literally double the prevalence of autism. That crude calculation, which supposes that some numbers found in the medical literature are correct, is shown in the footnote at the end of this article. The calculation depends mainly on two numbers: how many people use marijuana very frequently, and how much that frequent use increases the risk of autism.  And, the answer is yes, if indeed the published literature is correct, then the very frequent use of marijuana in California could account for doubling the prevalence of autism in individuals with low socioeconomic status but not in other individuals. The numbers add up.

At this stage, scientists, including myself, still call this idea a "hypothesis". We haven't proven anything except that the idea is plausible, or appears to be reasonable based on everything we know at the present time. The idea is currently published in a preprint ^[2], and you can find more information there.

Some people think of a hypothesis as simply a guess, but that definition could be misleading here. To be clearer, we could call this possible connection between acetaminophen, marijuana, and autism a "plausible explanation of the observations based on what we know right now". That description probably communicates the idea much better than simply calling it a guess. Nothing is proven, and we haven't reached any conclusions, but we do have a leading suspect.  

This plausible explanation gives us some immediate questions that need to be addressed. Is there another possible explanation?  Among those children with low socioeconomic status and autism, what is the prevalence of very frequent marijuana use by their mother? By other family members? How does that frequency compare with the frequencies we see in the families of children without autism? If marijuana is interacting with acetaminophen to cause a problem with brain development, is that a result of second-hand smoke during pregnancy or early childhood, or direct exposure from maternal use during pregnancy? Or a combination? Is it the oxidative stress that's most important, or is there some sort of adverse drug-drug interaction at play? Or both?

With autism possibly affecting up to 30% of black boys in some places, we desperately need to find answers, and we need those answers fast. At the same time, public health officials should be warned of a potentially disastrous interaction between acetaminophen and marijuana. We don't know if such an interaction exists, but it just might explain some startling observations.

Footnote: A very crude calculation of the amount of autism that might result from very frequent use of marijuana in California.

Let us suppose that a survey study conducted in California in late 2022/early 2023 [9]is accurate, and that about 14% of average Californians are very frequent users (multiple times a day) of marijuana. Let us also suppose that very frequent users of marijuana are 4.5 times more likely than other individuals to have a child with autism ^[8] and that individuals with low socioeconomic status are 2.5 times more likely to be very frequent users ^[9].

So, if 14% of all people are frequent users of marijuana, then about 2.5 x 14% = 35% of people with low socioeconomic status are very frequent users of marijuana. (Here we assume that the individuals with low socioeconomic status are a small minority of the total.)

Then, if those 35% of the people have a 4.5-fold increased risk of autism, and if the prevalence of autism is 5% for the average person in the population, then that 35% would carry an additional risk of 5% x 4.5 = 22.5%.

Then, the average prevalence of autism for the population with low socioeconomic status would be (0.35 fraction of very frequent users of marijuana x 22.5%) + (0.65 fraction of non-very frequent users of marijuana x 5 %) = 7.875% + 3.25% = 11.125% of children with autism. That's a bit more than double the prevalence of autism for the average children in this simulated population, which we arbitrarily set at 5%.

References

1. Nevison C, Zahorodny W. Accelerating Autism Prevalence in California with Changepoint Circa Birth Years 2015-2016 and Divergence by County and Socioeconomic Indicators. Preprints: Preprints; 2026.

2. Parker W, Harker M, Jones Iii JP, Konsoula Z, Meza-Keuthen S. Hypothesis: Emerging Evidence Might Suggest That Increases in Cannabis Use Disorder Following Legalization of Recreational Cannabis Significantly Contribute to Socioeconomic-Dependent Increases in the Prevalence of Autism Spectrum Disorder. Preprints: Preprints; 2026.

3. Parker W, Anderson LG, Jones JP, Anderson R, Williamson L, Bono-Lunn D, et al. The Dangers of Acetaminophen for Neurodevelopment Outweigh Scant Evidence for Long-Term Benefits. Children. 2024;11(1):44. PubMed PMID: doi:10.3390/children11010044.

4. Jones JP, 3rd, Williamson L, Konsoula Z, Anderson R, Reissner KJ, Parker W. Evaluating the Role of Susceptibility Inducing Cofactors and of Acetaminophen in the Etiology of Autism Spectrum Disorder. Life (Basel, Switzerland). 2024;14(8). Epub 2024/08/31. doi: 10.3390/life14080918. PubMed PMID: 39202661; PubMed Central PMCID: PMCPMC11355895.

5. Parker W, Corrigan PT, Anderson R, Jones I, J. P., Konsoula Z, Williamson L, et al. Evidence that acetaminophen triggers autism in susceptible individuals has been ignored and mishandled for more than a decade. Journal of the Academy of Public Health. 2025.

6. Ewing LE, McGill MR, Yee EU, Quick CM, Skinner CM, Kennon-McGill S, et al. Paradoxical Patterns of Sinusoidal Obstruction Syndrome-Like Liver Injury in Aged Female CD-1 Mice Triggered by Cannabidiol-Rich Cannabis Extract and Acetaminophen Co-Administration. Molecules (Basel, Switzerland). 2019;24(12). Epub 2019/06/20. doi: 10.3390/molecules24122256. PubMed PMID: 31212965; PubMed Central PMCID: PMCPMC6630875.

7. Smith RT, Gruber SA. Contemplating cannabis? The complex relationship between cannabinoids and hepatic metabolism resulting in the potential for drug-drug interactions. Frontiers in psychiatry. 2022;13:1055481. Epub 2023/01/28. doi: 10.3389/fpsyt.2022.1055481. PubMed PMID: 36704740; PubMed Central PMCID: PMCPMC9871609.

8. Tadesse AW, Ayano G, Dachew BA, Betts K, Alati R. Exposure to maternal cannabis use disorder and risk of autism spectrum disorder in offspring: A data linkage cohort study. Psychiatry Res. 2024;337:115971. Epub 2024/05/25. doi: 10.1016/j.psychres.2024.115971. PubMed PMID: 38788554.

9. Hill L, Ageze D, Dell'Acqua R, Gold A, Lanin-Kettering I, Rybar J, et al. Cannabis Use in California Following Legalization of Recreational Use. Cannabis and cannabinoid research. 2025:can20240179. Epub 2025/06/09. doi: 10.1089/can.2024.0179. PubMed PMID: 40489356.

10. Rabiee R, Lundin A, Agardh E, Allebeck P, Danielsson AK. Cannabis use disorder in relation to socioeconomic factors and psychiatric comorbidity: A cluster analysis of three million individuals born in 1970-2000. Scandinavian journal of public health. 2023;51(1):82-9. Epub 2022/09/20. doi: 10.1177/14034948221122431. PubMed PMID: 36120841; PubMed Central PMCID: PMCPMC9903242.

11. Nicholas W, Greenwell L, Washburn F, Caesar E, Lee G, Loprieno D, et al. Health Equity Implications of Retail Cannabis Regulation in Los Angeles County - Health Impact Assessment. Department of Public Health, County of Los Angeles Center for Health Impact Evaluation; 2019. p. 1-76.

12. Sarafian TA, Magallanes JA, Shau H, Tashkin D, Roth MD. Oxidative stress produced by marijuana smoke. An adverse effect enhanced by cannabinoids. Am J Respir Cell Mol Biol. 1999;20(6):1286-93. Epub 1999/05/26. doi: 10.1165/ajrcmb.20.6.3424. PubMed PMID: 10340948.

13. Parker W, Hornik CD, Bilbo S, Holzknecht ZE, Gentry L, Rao R, et al. The role of oxidative stress, inflammation and acetaminophen exposure from birth to early childhood in the induction of autism. J Int Med Res. 2017;45(2):407-38.

14. Zhao L, Jones J, Anderson L, Konsoula Z, Nevison C, Reissner K, et al. Acetaminophen causes neurodevelopmental injury in susceptible babies and children: no valid rationale for controversy. Clinical and experimental pediatrics. 2023. Epub 2023/06/16. doi: 10.3345/cep.2022.01319. PubMed PMID: 37321575.