There has been significant discussion everywhere in the media over the last few days about risks of COVID-19 to children and, in particular, about the possible link between current or past COVID-19 infection and a serious inflammatory illness. This has been described as Kawasaki-like or similar to toxic shock syndrome; the coverage has settled on the name Pediatric Multisymptom Inflammatory Syndrome, or PMIS, which is what we’ll adopt here.
In discussing the link between COVID-19, kids and PMIS, it is extremely important to separate out three questions.
- First: Can kids get COVID-19 and become very sick from it?
- Second: What is the typical course of COVID-19 in kids? How likely are they to get it and to get very sick?
- Third: What is PMIS and is there a link with COVID-19?
(There is a fourth question, of whether kids spread the virus, but we will leave that aside here).
These questions are fundamentally separate, and yet it is easy for them to become muddled. This goes in all directions. Observing that the disease is very mild in the vast majority of children can lead people to conclude that “Kids cannot get it.” On the flip side, seeing that some children get PMIS has led people to conclude, “This is extremely serious for a large share of children.” These are both mistakes, and can lead to excess complacency on one side and excess fear on the other. It is better to rely on facts.
Question 1: Can kids get COVID-19 and get very sick?
Yes. There are hospitalizations for some children with COVID-19 and, very sadly, a small number of deaths. Based on CDC estimates, slightly over 100 children in the US have been hospitalized over the course of the epidemic. It is possible for children to be very sick with this illness.
It is important to keep in context that this statement is also true about things like the seasonal flu. Based on CDC numbers, more kids this flu season have been seriously ill and died of the seasonal flu than of COVID-19. We say this not to scare people but rather to make clear the non-COVID-19 world is not one without risks.
Question 2: What is the typical course of COVID-19 in kids? How likely are they to get it and to get very sick?
For the vast majority of children, the course of the virus seems to be mild or even asymptomatic. We talk more here about details of early work in China. Over 90% of children with identified disease have no symptoms, or mild or moderate symptoms. In the CDC hospitalization data, children have serious illness at a much lower rate than adults. Data from Italy shows, similarly, that children are less affected.
We are learning more all the time, but at this point in the epidemic the fact that kids are a relatively less affected group seems to be the strong data consensus. This includes infants and children at least through the early teens at least. Generally, the severity of the infection and risk of hospitalization seems to increase with age over the entire life course.
(If you want to read more about why kids might be relatively unaffected, this paper has a good section).
Question 3: What is PMIS and is there a link with COVID-19?
What is PMIS?
Pediatric Multisystem Inflammatory Syndrome (PMIS) is the name we’ve given to a set of symptoms which some doctors think are linked to a small number of serious COVID-19 related illnesses in children.
To back up: some children with viral illnesses, like the flu or other seasonal respiratory illnesses, get very sick. We do not understand well why this happens. Kawasaki Disease is a different inflammatory disease that doesn’t have a known infectious cause. Inflammation in Kawasaki causes damage to the arteries and often to the heart muscle. It’s rare, but not that rare: there are about 5000 cases in the US in a typical year. Symptoms include a persistently high fever, bloodshot eyes, a rash and bumps on the tongue. Kawasaki disease can be very serious, but it is treatable and the vast majority of kids recover.
PMIS shares some features with Kawasaki disease, but seems to have a wider range of presentations. In particular, doctors have reported more abdominal symptoms (diarrhea, nausea). This picture has a nice summary of the symptoms and how to treat the illness (the latter more for doctors than parents). But persistent high fever and general inflammation seem to be the defining feature. Doctors also report seeing more of this in older children, whereas Kawasaki disease is typically limited to kids under 5.
It is important to know that there is no test for PMIS (or for Kawasaki disease). It’s a symptom-based definition: people with a particular set of symptoms are considered to have it.
Is this treatable?
Yes. The vast majority of kids with Kawasaki disease recover and do not have long term complications. Doctors have been treating patients with PMIS similarly to Kawasaki disease based on the shared symptoms. Our early data show that most kids with PMIS recover. Early treatment can improve outcomes.
Is it actually linked with COVID-19?
This is a more complicated question than you might think given the media coverage. In the absence of COVID-19 there are about 5000 cases of Kawasaki disease in the US in a typical year, so we’d expect about 300 a month across the entire country. New York is reporting perhaps 100 suspected cases, so that would clearly be more than expected, but not by huge orders of magnitude.
In most of the data we have seen, only slightly more than half of the children presenting with suspected PMIS have a positive test for COVID-19 or evidence of past infection (through antibody testing). This might make you think that they are not linked, and indeed it seems plausible that some share of cases are not linked. Some kids get Kawasaki disease in a normal period, even without a pandemic.
But given what we know about serious cases of COVID-19 in adults, and the fact that generally viral illnesses can cause this constellation of symptoms, it seems plausible that these are linked.
Bolstering this claim is some new data from Italy, published in the Lancet, showing that in Bergamo, Italy, there were more Kawasaki-like cases than usual showing up at the hospital in the last several months. This spike suggests a COVID-19 link. (Of note: All 10 cases in this study recovered). Though a note of caution, this hasn’t yet been described in the Chinese outbreak.
So, NIH to the rescue: The National Institute of Allergy and Infectious Diseases (NIAID), part of the National Institutes of Health, has already funded a study to get to the bottom of the PMIS.
Is it growing so fast? I hear more every day!
One feature of the recent coverage of PMIS is that it seems like every day we discover exponentially more cases. In New York it was 15, then 65, then 100. This makes it seem like explosive growth.
This is not likely to reflect actual growth for two reasons. First, some of the increase in cases reflects looking back at past cases which might not have been recognized as PMIS at the time. Second, and probably more important, when parents hear about this, some of them get scared and more of them are likely to show up at the hospital with their children, not all of whom will actually have the disease. The symptoms to watch out for — fever, rash — show up in kids for lots of reasons. When media outlets report “possible” cases, this number may be inflated by care seeking – many of these may not turn out to be PMIS.
Summary: How Should Parents React to This?
Putting this together, we have a few facts.
- Kids can get COVID-19.
- The vast majority of children with COVID-19 will have mild disease.
- Some children who get very sick seem to have an inflammatory reaction, labeled PMIS,, which is scary, but treatable in the vast majority of cases reported so far.
The American Heart Association has a thoughtful piece on how parents should process this. There is a line here between vigilance and fear. Parents should be aware of this set of complications, and know that if their child has a high fever for more than a few days, they should be seen by a doctor. This is true, by the way, pandemic or not.
But parents should also be aware that this set of complications is extremely rare.