New Data on Pregnancy and Vaccines

In which I discover a statistical error in a NEJM paper, but agree with vaccine safety conclusions.

Quick note: CDC guidance on masks of course resulted in a lot of emails with the “What about kids” question. I will write more about this (of course) but for now I’ll reiterate that I continue to think a lot of this is going to be about calculating risks and evaluating your feelings about them. A bit more on Twitter here.

Everyone aged 16 and over now has access to a COVID-19 vaccine in the US. It’s increasingly the case that if you want a vaccine, you can get one. This doesn’t mean there isn’t work to be done increasing ease of access and decreasing hesitancy, but if you seek it out, in most places there is no longer the “refresh page fifty times to get last appointment” problem.

Despite availability, some people are still nervous or actively reluctant. The group I hear most from is pregnant people. There are many questions: vaccine during pregnancy or wait? When in pregnancy? Is it better to get vaccinated while breastfeeding and transfer antibodies? Would you transfer them better while pregnant? Is it safe?

Last week, the CDC director took stronger stand on the view that pregnant people should be vaccinated (not just that they could be). In a confusing turn, the CDC then backtracked on this, saying their stance hadn’t changed. I’m not entirely sure what is going on there, but my sense is that the initial statement was motivated by some new evidence. I’ll take a look at what we now know below, and then get into a few specific questions.

Evidence Piece 1: Risks of COVID in Pregnancy

One thing we have updated on is the risks of COVID-19 in pregnancy. Put simply: it seems that the consequences of COVID-19 are (on average) worse in pregnancy. There are a couple of ways to look at this.

First, a recent article in JAMA used data from 18 countries (Brazil, UK, Mexico, others) and compared pregnant people who got COVID-19 to those who did not. They found that adverse pregnancy outcomes — including both infant and maternal mortality — were higher among those with COVID-19 versus those without. In a sense, this simply demonstrates that COVID-19 is risky for pregnant people which shouldn’t be surprising. But it’s sobering given the very serious risks the paper highlights.

Second, the CDC has analyzed a very large dataset (400,000 women) and concluded that pregnant people are at higher risk from COVID-19 than non-pregnant people. They’re more likely to be hospitalized, put in an ICU, to need oxygen and to die. The article appropriately cautions that serious complications are uncommon, even among pregnant people, since this is a low risk age group. But it does seem that pregnancy increases those risks.

It’s not clear that pregnant people are any more likely to get COVID-19 than their non-pregnant counterparts, but these data together suggest they are at higher risk for serious illness. This higher risk is one reason for a vaccine push.

Evidence Piece 2: Vaccine Safety

The evidence of risks in pregnancy was largely known before last week, so I think the CDC statements were a result of new evidence on vaccine safety in pregnancy, notably from this NEJM article.

The article focuses on what we know — a few months in now — about the safety of the mRNA vaccines (Pfizer and Moderna) when given during pregnancy. Front line health care workers were an early vaccine group, and that group contains at least some pregnant people. Hence, we are starting to get sizable numbers of people to follow-up on.

The article mainly relies on data from “V-Safe”. Everyone who is vaccinated is invited to enroll in V-Safe. It’s an app which texts you with links to online surveys about how you’re feeling after vaccination. The V-safe surveys include question about pregnancy status for individuals who do not identify as male.

From this initial survey alone, V-Safe identified 35,691 pregnant people. With this group, it’s possible to start looking at baseline side effects among pregnant versus non-pregnant people. The first finding of the paper is that these are very similar, as illustrated in their very nice graph below.

“Pain” here refers to arm pain, which nearly everyone has and is slightly more common in pregnant than non-pregnant people. But most of the other symptoms are actually more common in non-pregnant people. Bottom line: after initial vaccination we don’t see any evidence of more risks in pregnant people.

To get more detailed data on pregnancy outcomes, the authors make use of another part of V-safe. For some of those who indicate pregnancy, there is a follow-up to ask whether they would like to enroll in the V-safe pregnancy registry. This will entail further questions on their pregnancy as it evolves. The registry contacted 5230 individuals, of whom 3958 were eligible and willing to participate in further follow-up (nearly all were health care workers, between 25 and 44 years of age, and 80% identified their race as non-Hispanic white). Vaccinations were pretty evenly split across time during pregnancy, with slightly more of the participants being vaccinated in the second trimester than the first or third.

Once people were enrolled in the registry they were all interviewed once; some had already given birth or miscarried, most were still pregnant. There are plans to follow-up with the still-pregnant people on an ongoing basis, but that isn’t in this paper. The authors therefore identify pregnancy outcomes from the set of people who have concluded their pregnancy.

At the time of the writing of the paper, there were 724 live births (from 712 women) from those vaccinated. The authors can look at the characteristics of these births — were they pre-term? Did they have other complications? It is worth keeping in mind: these are largely people who were vaccinated in their third trimester (that’s why they have already given birth).

The data is reassuring. 9.6% of the births were preterm, 3.2% were small for gestational age and 2.2% reported congenital abnormalities. No infant deaths were reported. These figures are in line with what we would expect in the general population. The bottom line is there isn’t anything here which points to vaccine complications.

The authors also look at miscarriage rates, hoping to ask the question: did people who were vaccinated miscarry at rates higher than expected? For some reference, miscarriage rates (especially in the first trimester of pregnancy) are high. I don’t say this to scare you, just to set expectations. The paper provides a reference range: 10% to 26% of known pregnancies end in miscarriage.

The authors would like to provide a comparable miscarriage rate among vaccinated people. The number they provide is 12.6%, which is in this range. However, on further reflection, the number they calculate actually isn’t comparable to their reference range and doesn’t have much meaning.

This is a bit in the weeds but bear with me. (This is why I get up in the morning, people.)

At the time of the first survey, 827 people have completed their pregnancy. The result is 712 women with live births and 104 pregnancy losses (largely before 13 weeks). The authors divide 104 by 827 to get 12.6%.

This would be a reasonable calculation and comparison if they started with a sample of 827 people who had newly conceived and then followed them to the end of their pregnancy. If they did that, and they observed 104 miscarriages, they could calculate a miscarriage rate of 12.6%.

However, that isn’t what they do. What in fact happens is people enter the study at some point in pregnancy and a few weeks later they call them and ask if they are still pregnant. They then take the 827 people who are not still pregnant and ask what share of that group miscarried versus had a live birth. But this number is largely driven by when in pregnancy the people entered the sample.

If most people are vaccinated right before the end of pregnancy, then you’ll see a lot of live births two weeks later, and very few miscarriages. If most people are vaccinated at the start of pregnancy, then we will see a lot of miscarriages and few births. This is just mechanical, a result of when people choose to be vaccinated.

The fact is, the 12.6% number in the paper has no interpretation. It is not scary, or reassuring, or anything. It’s just meaningless.

Once the study is fully completed, the authors will have followed the full sample over their whole pregnancy and be able to map out pregnancy loss rates as a function of timing of vaccine, and that will provide us a number that can be compared to a reference group. That will be helpful, but until then we are much more limited.

I do think the authors could have done this calculation slightly better, and it would also be reassuring. There are 1224 individuals who were vaccinated during the first trimester or in the periconception period, 96 of them had miscarried by the time they were interviewed. This is a miscarriage rate of 7.8%. This is lower than the true rate, since later interviews will reveal more losses, but it is a true “lower bound” rather than some random number. And it does seem in the range of what we would expect.

To be clear: there is nothing in these data to suggest miscarriage rates are increased by the vaccine. Their second dataset — from the Vaccine Adverse Event Reporting System or VAERS — is also reassuring. If anything, the number they report may be too high. My complaint is with the statistics, not the conclusion.

Overall: I think it’s fair to say we are starting to get data in which suggests that vaccines in pregnancy are safe. This data is not yet detailed enough to make very precise statements. But: if there were significant issues with vaccines in pregnancy I believe we would see them by now. Even putting aside the detailed data here, the fact is that a huge number of pregnant women have been vaccinated at this point, and we haven’t heard about significant issues. That itself is evidence for safety.

Other Questions!

Do I pass immunity to my baby when vaccinated?

To some extent, probably yes. The latest paper on this is here and shows, among other things, evidence of antibodies from vaccination in cord blood. This suggests antibodies passing in the bloodstream of infants. The antibody concentration in that study was higher for vaccinated people than those who were recovered from COVID. The upshot is that some immunity may pass from vaccinated mothers to their infants.

Of note: antibodies also pass through breastmilk, but this immunity is passive — it isn’t retained when breastfeeding stops. Antibodies passed through the placenta may persist. This is all in need of more research.

When in pregnancy should I be vaccinated? What about fever in first trimester?

There isn’t much reason to favor vaccination at any particular time. The study on antibodies above suggested that antibody passage was highest for those who were fully vaccinated (second dose) several weeks before birth. If you have a choice, this argues for earlier vaccination.

A number of people have asked me about concerns about fever in the first trimester and the risk of birth defects (related to, for example, concerns about hot tubs and birth defects). Case-control studies have shown links with fever and neural tub defects. However: such studies are subject to a lot of biases and studies with better methods don’t seem to show these links. An example is this one, following 8000 women in Denmark — which does not show links between first trimester fever and birth defects.

Overall, there isn’t any particular reason to favor waiting on vaccines. For some women — especially those who are low risk for COVID-19 exposure — there may be psychological value to waiting until the second trimester since it might “feel” safer. This would need to be weighed against the COVID risk.

Which Vaccine?

The CDC has supported all three vaccines (Pfizer, Moderna, J&J) for use in pregnancy. And all three are currently running studies to formally evaluate safety and efficacy in pregnant people. At this point, nearly all of our data (including in the NEJM article above) is from the mRNA vaccines (Pfizer and Moderna). This is largely because those are the vaccines used in health care workers and in early vaccine rounds that was the group who was likely to be pregnant.

But, seriously, should I get it?

You know I’m not going to tell you what to do! It’s not my jam.

What I will say is I would get vaccinated if I were pregnant and if you asked my advice as your friend, I would recommend you do so. The risks of COVID-19 in pregnancy are real, and the safety data is very reassuring. I’m further excited by the possibility of antibody protection for the baby.

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