Long Term COVID Risks in Kids

Have more questions after today’s newsletter? I’m planning to try out answering some in video form on Instagram. I’ve also got some videos over there about the J&J vaccine issues. Head over there and follow me to stay updated!

By now, you likely heard that kids are less affected than adults by COVID-19. They are somewhat less likely to get COVID-19 and much less likely to be seriously ill or to die from the disease. Compared to an adult in their early 70s, a child of 10 who has COVID-19 is 40 times less likely to be hospitalized and 1300 times less likely to die. For further context, I’m repeating below a table I’ve put in earlier newsletters, which compares the risks of COVID-19 in kids to other risks.

Even accepting these low immediate COVID risks, though, the fear of long term complications looms large for many parents. What if you think your child has a mild case, but then months later they are very ill? What if they have no symptoms but then later something bad happens? What if they do get sick and the symptoms last for months?

Today’s post is going to go through what we know, and what we do not, about long term COVID complications in kids focusing on both MIS-C and on “long COVID”. Three caveats before we start.

First, this post is narrow. It is very specifically about what we know about long term issues in kids. I’m not going to talk about broader issues of spread, or even about kids risk of getting infected in the first place. Making choices about what activities to do or not will require thinking about these considerations as well, and factoring in the community risks (see calculator, linked at bottom!!)

Second: I’m not going to talk specifically about variants, but I think the analysis here does apply. Check out this helpful article for some context on B.1.1.7 in particular and kids.

Third: Especially when it comes to our kids, we all have our own risk tolerance. Some people let their eight year old walk to school, some don’t. Some people let their kids play football, some don’t. And our approach to COVID-19 risks is similar. I can give you some limited numbers here to help make your choices, but there isn’t going to be just one right choice for everyone.


You might have first heard about MIS-C back in late spring or summer of last year, when it was sometimes called “Kawasaki-like illness” and there were a number of cases in New York City. MIS-C stands for “multisystem inflammatory syndrome in children” and you can see the CDC discussion of it here. The syndrome causes inflammation in various organs, fever, rash, and neurological symptoms. It can be very, very serious and often requires hospitalization.

MIS-C is related to, but presents slightly differently, from Kawasaki Disease, which is a rare complication of febrile illnesses like the flu. Kawasaki generally affects very young children and MIS-C has been seen in a broader range of ages. What they share is some features of the timing: both show up as complications often weeks after acute infection is over.

MIS-C appears to be a complication of COVID-19 infection. I say “appears to be” since there hasn’t been definitive linking of the two. But the circumstantial evidence is strong enough that people have generally accepted the relationship. And MIS-C is scary: it can show up even in kids with mild or asymptomatic infections and the complications are serious. There have been a small number of deaths.

This is a real risk. It’s not fake news. But it’s also extremely rare. The CDC counts about 3100 MIS-C cases in the US, with 36 deaths. A recent JAMA article attempted to summarize the characteristics of identified cases and calculate an overall incidence number (article here). They identified about 1700 cases of MIS-C in people under 21, with more serious illness in older teenagers than in small children.

The JAMA paper estimates and overall incidence of 2.1 in 100,000, meaning this illness affects 2.1 in 100,000 children over the last year. By comparison, the incidence of childhood cancer is several times higher (there are typically around 10,500 cases of childhood cancer diagnosed per year).

Bottom line: very rare, but serious, complication of COVID-19 in children.


MIS-C is an acute, serious illness which develops as a longer-term consequence of COVID. “Long COVID”, in contrast, refers to the phenomenon in which people have complications and symptoms from COVID which persist long after acute illness. Long COVID has been recognized in adults, and the NIH has started an initiative to study it. Not everyone is as sold on the idea, so this is an ongoing discussion.

There are also concerns about long COVID in kids. One piece of this is kid who are very sick from COVID-19 (rare, but happens) and take a long time to recover. There are certainly instances of this, as there are with other serious illness like the flu or RSV.

The more nebulous concern I hear is that even children who do not get very sick, or are totally asymptomatic, can have long term symptoms like fatigue, headache or brain fog weeks later. This concern has been bolstered by some recent data from the UK and Italy, which reported out something like 12 to 40% of children with documented COVID-19 cases still had symptoms weeks later.

It is worth digging into this data a little more for context.

Best documented is the data from Italy; pre-print available here. This study covers 129 children in Italy who were diagnosed with COVID-19 between March and November of 2020. The children experienced a range of illness, and the sample includes some asymptomatic kids. They were surveyed between 2 and 4 months after diagnosis and asked about various symptoms. The survey (which you can see here) was explicit that the study was about long term impacts of COVID-19 in children, and asked about symptoms in the last 7 days. The headline scary number is that 42% of the 68 kids interviewed 4 months after diagnosis reported symptoms.

There are various complications in interpreting this. One is the frame of the question — people may be primed to report more symptoms by the topic of the survey. But more important, I think, is that it’s hard to know the base rate. For example: 12.4% of children report nasal congestion. This is counted as a long COVID symptom but, of course, some kids have a runny nose for other reasons. The same is likely true for diarrhea, headache, rash, insomnia and the other symptoms listed (there are a total of 20 ). These symptoms are mostly mild. Of note, roughly 80% of the cases report that the symptoms bothered the child “not at all” or “only a little.”

The authors report two things which might help differentiate long COVID-19 from base rate symptoms. The first is that they ask about fatigue relative to before diagnosis. There, they do not find a COVID-19 effect. Overall, 75.4% of children report the same level of fatigue, 13.2% report less than before and 10.9% report more than before. The second is they compare children who had symptomatic COVID to those who had asymptomatic COVID, on the theory that long term symptoms associated with COVID-19 would be worse in the symptomatic group. This is a good idea, although they have too little data for precision; none of these comparisons rise to statistical significance.

Overall, in my view it’s hard to draw strong conclusions here. The scary headline number is certainly misleading. What they can realistically claim is that 42% of children in this group report one of a constellation of very common childhood complaints on a given day. Could they be long term COVID complications? They could be. But a large share of them likely reflect other illnesses or general child malaise and without a comparison group it’s difficult to know what share.

I would levy a similar concern with data from the UK which reports 12% of children have symptoms 5 weeks after a positive COVID test. These data are, again, derived from surveys which ask about a set of common illness symptoms and the base rate is difficult to derive.

This base rate concern isn’t idle. To get a sense of magnitude, I pulled data from the 2017-2018 National Health and Nutrition Examination Survey, which has a module on current health. Among children under 18 surveyed, 20% report a head cold or chest cold during the last month. This isn’t directly comparable to the numbers in the Italian or UK study for a number of reasons (they didn’t ask the same questions, it’s not the same time frame, and so on) but it does give some sense that these kind of symptoms are common.

None of this is to say that long COVID isn’t possible and, in fact, it would be surprising if we didn’t see lingering symptoms in at least a small number of children, especially those who got really sick. But the suggestion that 10%, 15%, 45% of children have symptoms months later — this is alarmist. The true numbers are likely much, much smaller than this.

We do need better data. The study in Italy had 129 children! In the US alone we have at least 3.7 million children who have tested positive, probably many more who had undetected infections. A better study of this would include a wide range of children, some of whom had known COVID and others who did not, and would include antibody testing so we could detect asymptomatic infections. Questions would be asked on symptoms at several time points, and comparisons could be made between children who had detected COVID, undetected COVID and no COVID. Together, this would give us a much better sense of the size of these longer term risks.


Where does this leave us, other than in our typical frustrating situation of not knowing enough?

Clearly, it is possible for children to have long term or extended complications from COVID-19. MIS-C is serious, but the risks are small. The long COVID risk is unclear in magnitude and the symptoms are largely mild.

When it comes to kids it is easy to be drawn into what-ifs. Sometimes, falling asleep at night, I am gripped by a vision of one of my kids running into the street in front of a car. These moments are terrifying, but making the right decisions about risks we take with our kids (and which we will let them take on their own) requires thinking rationally.

In this particular case, the problem is compounded by uncertainty. It’s not impossible to deal with, but requires you to think through varying scenarios. You can calculate the risk of COVID-19 infection, based on your community rate and features of the interaction you’re considering. You can incorporate the risks of MIS-C — perhaps 1 in 1500 after COVID-19 infection.

And then you can layer on top of this some long COVID scenarios. What if there is, say, a 1% chance that if your child gets COVID-19 they’ll have a runny nose, headache or insomnia in 4 months? This is another risk to incorporate alongside the immediate illness risk. What if this chance is 5%, or 0.1%?

If your answer is the same for all these scenarios then, well, you’ve got your answer. If not, you’ve got to think a bit more about how to deal with this uncertainty, and perhaps give more thought to how you read the long COVID evidence.

Ultimately, as a parent, you have to make the choice that works for you. It’s never going to be an easy one.

Weigh in!

Keep the thoughts coming. I don’t always write back, but I read everything.

Write to Me!

Where to Find Me

COVID Risk Calculator!

COVID School Dashboard

My Website

Family Firm


Expecting Better