As the number of cases of acute hepatitis in children globally mounts, the list of potential culprits seems to have stagnated, with, perhaps, one exception: dogs.

A survey of such patients in the United Kingdom found that 70% of those responding own them, according to a May 6 technical briefing from the U.K. Health Security Agency.

But correlation doesn’t equal causation.

“Pet dog ownership is common in the U.K.,” the report cautioned. “The nature of trawling questionnaire investigations means that some responses may be high through the play of chance due to the large numbers of questions asked.”

The cause of the mysterious cases of liver inflammation—often accompanied by jaundice, sometimes requiring transplant—is currently unknown.

Health officials have been quick to implicate adenovirus, a pathogen that causes cold- or flu-like symptoms, as a potential trigger, as it’s been detected in more than half of children with acute hepatitis of unknown cause in the U.S. and U.K.

But adenoviruses are common. And there’s no history of association between adenovirus 41—the type most commonly discovered in acute hepatitis patients who test positive for adenovirus—and hepatitis, according to the World Health Organization. Other types of adenoviruses have been associated with liver failure, but only in immunocompromised kids.

Furthermore, adenovirus 41 is fairly rare among adenoviruses, ranking No. 14 for reporting frequency from 2003-2016 and comprising less than a third of 1% of cases, according to a May 19 presentation from the U.S. Centers for Disease Control presentation on the matter, citing data from the National Adenovirus Type Reporting System.

Is adenovirus a red herring, just as dogs likely are?

“Adenovirus is common, so the fact that it’s around a lot and the fact that we haven’t really seen the pattern before makes it important to differentiate” from other potential causes, said Dr. Georges Benjamin, head of the American Public Health Association. “We should not close our search for other viruses.”

The case for a COVID connection

Active COVID infections were found in a smaller percentage of U.S. and U.K. patients, but evidence of past infection wasn’t reported, or perhaps even sought, in the majority of cases. Such testing is now underway in the U.K., according to the May 6 U.K. technical briefing, though “the high population cumulative prevalence of SARS-CoV-2 will make the interpretation of this data challenging.”

Of the 173 U.S. pediatric acute hepatitis patients reported in the U.S. as of last week, 12% had an active COVID infection, and 74% of the 19 tested showed evidence of prior infection.

“One of the first things to do is not to look for active COVID, but evidence of a prior COVID infection,” said Arijit Chakravarty, a COVID researcher and CEO of Fractal Therapeutics.

“We know that around 75% of children in the U.S. have evidence of a prior COVID infection. Out of 30 patients with liver failure, did 27 or more test positive for a prior COVID infection?"

If so, Chakravarty said, “that would strongly suggest an association with COVID. There's less than a 4% chance of observing that result by random chance alone."

"Of course, COVID antibodies wane over time, so some children with a prior infection may actually test negative by serology. That would mean that you might miss an association with COVID even if one existed.”

More recent theories point to adenovirus and COVID-19 as potential cofactors, such as a co-infection of the two, or the inability of the body to properly respond to adenovirus after a prior COVID infection, according to the U.K. Health Security Agency.

Another possibility, it pointed out: a post-COVID syndrome similar to Long COVID or MIS-C—a rare, inflammatory-based illness that can occur in children after COVID infection or exposure.

“Is it possible it’s not COVID? Of course it’s possible,” Chakravarty said. “But we need to rule it out specifically first. You don’t need to differentiate it as a trigger or direct cause in order to take a practical course of action from it.”

Ruling out COVID first is only logical, given the U.S. public health approach of allowing unchecked spread of the virus, he said.

“How can you say we’re going to be nimble and deal with problems as they arise, but then say ‘let’s not rule out COVID first’ when it’s rampant?” he said. “It’s a public health policy question.”

Something else entirely?

Dr. Alexandra Brugler Yonts, an infectious disease specialist and director of the Pediatric Post-COVID Program at Children's National in Washington, D.C., agreed that a potential COVID link “needs to be explored.”

“But it should not be the only focus,” she told Fortune, “especially since these cases of hepatitis emerged after the world opened up again and all the other viruses came back to the party.”

Another possibility is a new hepatitis virus that hasn’t yet been identified, Benjamin said. Known hepatitis viruses, A through E, have not been detected in the recent outbreak of cases, according to the U.K. Health Security Agency.

“I would encourage [researchers] to make sure they’ve totally ruled out COVID, and totally ruled out an unknown hepatitis virus—every few years we find a new hepatitis virus that hasn’t yet been identified, first A, then B, then they suddenly discovered C, then E,” he said.

Yet another possibility, one that perhaps has been least explored: a toxic exposure from a source like chemicals, medicine, or supplements, he added.

Divorcing politics from science

Some worry that medical professionals have been quick to dismiss COVID as a cause or factor, in an effort to draw attention away from the pandemic—and perhaps from the lack of vaccines for children ages 0-4.

Benjamin cautions that politics of all types must be left out of the epidemiological investigation.

“You can’t do good science if you’re worried about the politics,” he said. “If it’s COVID, let it be COVID. SARS-CoV-2 can impact every single organ system we know so far. The liver would not be an exception by any means. We need to make sure we’re right about the [cause], otherwise we’re not able to fix this.”

Dr. Jay Varma, director of the Cornell Center for Pandemic Prevention and Response in New York and the Kroll Institute’s chief medical advisor, cautioned that investigators must keep an open mind and not jump to any conclusions—adenovirus, COVID, or otherwise.

“Similar to a murder investigation, you don’t hone in on a single suspect too early, otherwise you might risk missing the real culprit,” he said. Adenovirus “is the leading culprit, but it’s unclear if it’s acting alone or in combination with other factors.”

“It could be a combination of adenovirus plus another infection, such as prior COVID-19 infection, or a genetic factor, or an environmental/chemical exposure, or some combination. It’s also possible it could not be adenovirus 41 at all. … These investigations can take a surprising turn.”

Waiting on an answer is anxiety-inducing for parents, he said, “but the experience I’ve had is that sometimes it isn’t until very long into these investigations that you get an ‘a ha’ moment.”

This story was originally featured on Fortune.com