Circumcision Does Not Raise a Boy’s Autism Risk, Study Finds

A new baby, a day or two old, and a decision that thousands of American parents make before they leave the maternity ward. For a while now, that decision has carried a nagging worry passed around online and amplified by comments from public officials last year: could circumcising a newborn boy somehow raise his odds of autism? A team of researchers reckoned they already had the numbers to check. So they checked.

The answer, drawn from nearly 3000 boys followed from birth into early childhood, is a plain one. No link.

The study, published in the journal JAMA Pediatrics by Monica McGrath and colleagues at the Johns Hopkins Bloomberg School of Public Health, leaned on a large national project, the Environmental influences on Child Health Outcomes cohort (ECHO for short). It is a prospective cohort, meaning children are enrolled and then tracked forward through time rather than dug up from old records after the fact. The analysis covered 2771 male children across 14 sites around the country. About 66 per cent had been circumcised before they went home from hospital, and roughly 7 per cent picked up an autism diagnosis, at a mean age of 3.8 years.

Here’s where it gets interesting, and where a quick glance at the raw figures could mislead you. Among the circumcised boys, 6 per cent were later diagnosed with autism. Among the uncircumcised boys, the figure was actually higher, at 9 per cent.

Why the Raw Numbers Mislead

Taken at face value, that gap makes circumcision look almost protective. But the two groups of boys were not alike to begin with: they differed by region, by family background, by all sorts of things that also shape whether a child is assessed and diagnosed. Once McGrath’s team adjusted for likely confounders, things like parental age and a family history of autism, the apparent difference melted away. What was left was no meaningful association at all (an adjusted odds ratio of 0.83, with a confidence interval straddling 1).

“Parents have been concerned about this potential circumcision-autism connection, but hopefully these findings may help allay their fears,” says McGrath, who directs the data analysis center for the ECHO project. The work grew directly out of last year’s public chatter. “We knew we had the data to help investigate this question,” she says.

The researchers didn’t stop at the yes-or-no of a formal diagnosis. They also pulled in two validated questionnaires, the Social Responsiveness Scale and a subscale of the Child Behavior Checklist, which pick up autism-related traits and behaviors that may never rise to a clinical label. Neither budged. After adjustment, circumcised and uncircumcised boys looked the same on both. The null held up when the team sliced the data by geographic region, by preterm birth, and by whether a baby had spent time in intensive care. One lone subgroup threw up a positive signal, in the West and only on one of the behavior measures, but the authors reckon that is most likely a statistical fluke thrown off by the sheer number of comparisons, traceable to a single site.

There was one thread they wanted to follow and could not. Some versions of the circumcision-autism idea point the finger not at the procedure itself but at the acetaminophen sometimes given for pain. Trouble is, hardly any of these babies got it: fewer than 5 per cent. Too rare to test as a pathway, so that particular question stays open, though the team judges an acetaminophen route unlikely.

Two older studies are what put this worry on the map, a 2013 ecological analysis and a 2015 Danish cohort that reported a 46 per cent higher autism risk among circumcised boys. Both, the authors point out, had problems: the Danish work came from a country where circumcision is rare and often done outside hospitals for religious reasons, tangling the exposure up with culture, migration and health-seeking behavior in ways that are hard to unpick.

What a Null Result Can and Cannot Say

It is worth being clear about what this study does and doesn’t settle. It’s observational, not a trial, so residual confounding can never be ruled out entirely. The families skewed well-educated and the circumcisions were all hospital based, so the findings say little about procedures done later in childhood or outside a clinical setting. A null result is an absence of evidence for a link, not a mathematical proof that no link of any size could exist. Against the backdrop of a genuine puzzle, US autism diagnoses have climbed from roughly 1 in 150 children in 2000 to 1 in 31 among eight-year-olds by 2022, and the CDC itself cautions that it is unclear how much of that reflects better awareness and diagnosis rather than a true increase. Whatever is driving that rise, circumcision does not appear to be part of the story.

“We observed no association between circumcision and autism in boys, but it remains important to evaluate the safety of common medical procedures,” says McGrath. Her team is still chasing the other factors that might explain the numbers. For the parents standing in the maternity ward, though, one item can probably come off the worry list.

Frequently Asked Questions