Did CoVID cause autism (or vice versa)?
Did CoVID cause autism (or vice versa)?
What do we know from a correlational study from the pandemic?
Autism. Covid.
Hmmm...is there a relationship?
Savvy readers will already have realized that I’m posing a correlational question and that there are routine cautions about inferring causal relationships based on correlations. But, let’s just think this through a little bit. I mean, what’s the matter with spinning up a plausible theory and running out to promote it?
Starting place
The issue I’m describing might be illustrated in this way
Autism < == > CoVID
That is, some measure of CoVID increases and at the same time, some measure of autism increases. The two go “up” together—a “positive” correlation. Another possibility is that as one measure increases, the other decreases—a “negative correlation.” Either way, the two measured phenomena are correlated—they co-vary—and the value of the correlation indicates how closely or tightly they co-vary.
Now, despite oft-heard admonitions to avoid inferring causal relationships from correlations, we are humans and so we might start looking for patterns that explain stuff…account for the correlations. We might generate hypotheses!
Did having CoVID during pregnancy cause autism? Did the CoVID virus disrupt genetic processes so that genes expressed different proteins so that normal language and social development were sent just enough askew to lead to autism?
Did autism cause CoVID infections? Did having autism during fetal development lead to a special susceptibility to CoVID? Could it be that some infant behavior (let’s just say rubbing one’s eyes) make autistic kids more susceptible to getting CoVID?
Those are a couple of grand hypotheses about the Autism < == > CoVID relationship. But, wait...is there an Autism < == > CoVID relationship? Shouldn’t we start from some facts before we run off to ginning up theories?
Evidence
Researchers Morgan Firestein and colleagues wisely went after the fundamental facts about a Autism < == > CoVID relationship. They examined records for a couple of thousand births in New York City and published their results in JAMA Network Open (an outlet of the prestigious Journal of the American Medical Association) on 23 September 2024.
They assessed whether prenatal exposure to the CoVID-19 pandemic milieu (writ broadly, meaning environments here CoVID was active) or actual maternal infection with SARS-CoV-2 was correlated with higher scores on the Modified Checklist for Autism in Toddlers.
Nope. No correlation…so our beautiful explanations are…uhm…well…baseless.
This cohort study of approximately 2000 children in New York City revealed no significant association between prenatal exposure to the pandemic milieu and M-CHAT-R positivity. Prenatal exposure to maternal SARS-CoV-2 infection was associated with lower rates of M-CHAT-R positivity.
Please read that last sentence a second time. In fact, having CoVID or being in a CoVID-inflused environment may have actually protected against autism! So…new theory! Tell all your pregnant friends to go get CoVID so their babies won’t have autism! Hurry! Don’t delay…oh, wait? What? You mean that interpretation of the study results seems bogus? Oh….
One can read a press release from Columbia, the university of most of the researchers, to get an alternative summary of the study.
Reference
Firestein, M. R., Manessis, A., Warmingham, J. M., Xu, R., Hu, Y., Finkel, M. A., Kyle, M., Hussain, M., Ahmed, I., Lavallée, A., Solis, A., Chavs, V., Rodriquez, C., Goldman, S., Muhle, R. A., Lee, S., Austin, J., Silver, W. G., O’Reilly, K. C., Bain, J. M., Penn, A. A., ... & Dumitriu, D. (2024). Positive autism screening rates in toddlers born during the COVID-19 pandemic. JAMA Network Open, 7(9), e2435005-e2435005. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2823875