.

Peter McCullough misinterprets study on COVID-19 vaccination in pregnancy; the study shows that vaccines are safe for pregnant women

CLAIM
“No discernible benefit of covid-19 vaccination in pregnancy”; “higher rate of neonatal and all-cause readmissions among babies born to vaccinated mothers”
VERDICT more about the rating framework
DETAILS
Inadequate support: Abundant scientific data shows that COVID-19 vaccines effectively reduce the risk of severe disease for pregnant women. The study by Jurgensen et al. cannot be used to assess vaccine effectiveness during pregnancy because it isn’t designed for it
Misrepresents source (Strawman): The claim uses crude data from Jurgensen et al. which don’t take into account all the possible confounding factors that may lead to incorrect interpretation of the data. The authors considered McCullough to have misrepresented their study.
KEY TAKE AWAY
Getting COVID-19 while pregnant is associated with a higher risk of developing severe disease and pregnancy complications. Abundant data show that vaccination during pregnancy is safe for the mother and the unborn child and effectively reduces the risks of complications from the disease.

FULL CLAIM: “No discernible benefit of covid-19 vaccination in pregnancy” ; “the crude data show a higher rate of neonatal and all-cause readmissions among babies born to vaccinated mothers”

REVIEW


COVID-19 can be dangerous to pregnant women. Indeed, scientific studies have shown that pregnant women who get COVID-19 are at higher risk of severe disease[1-3]. Furthermore, COVID-19 also increases the risk of negative pregnancy outcomes such as stillbirth, newborn mortality, and newborn admission to neonatal intensive care[2, 4-6]. Thus, the American College of Obstetricians and Gynecologists (ACOG) as well as the U.S. Centers for Disease Control and Prevention (CDC) recommend COVID-19 vaccination to pregnant women.

However, since the first rollout of COVID-19 vaccines, claims that vaccinating against COVID-19 was unnecessary or even dangerous for pregnant women and their unborn child circulated. Health Feedback explained on several occasions why these claims are inaccurate.

In another iteration of those claims, cardiologist Peter McCullough claimed that a study by Jorgensen et al., published in October 2023[7], showed “No discernible benefit of covid-19 vaccination in pregnancy”. McCullough also claimed that the study indicated that COVID-19 vaccination posed a risk during pregnancy because “the crude data show a higher rate of neonatal and all-cause readmissions among babies born to vaccinated mothers”.

McCullough has repeatedly shared misinformation on COVID-19 vaccines in the past and promoted unsupported “treatments”. Here again, a closer analysis of Jorgensen et al. shows that McCullough misrepresented the study’s findings. Health Feedback also reached out to the study’s authors, who refuted McCullough’s claims. We provide details below.

What did the study do and what did it find?

Jorgensen et al. aimed to “determine if maternal mRNA COVID-19 vaccination during pregnancy was associated with an increased risk of adverse newborn and early infant outcomes”. To do so, they collected data from health administration databases of Ontario, Canada, involving more than 140,000 pregnant women. They grouped women in two categories: unvaccinated and vaccinated during pregnancy. Those who were vaccinated before getting pregnant weren’t included in the study.

Then, they compared the risk of negative pregnancy outcomes such as severe neonatal morbidity and mortality, admission to neonatal intensive care units, and neonatal and six-month hospital readmissions. A hospital readmission is when the baby and their mother were able to leave the hospital after birth, but had to return owing to complications in the baby within 28 days or six months of age[7].

The authors found that the risk of neonatal morbidity, mortality, and admission to intensive care units was lower among babies born to vaccinated mothers. Furthermore, there were no differences in the risk of hospital readmission between vaccinated and unvaccinated mothers.

Based on their results, the authors concluded that “maternal mRNA COVID-19 vaccination during pregnancy was not associated with adverse newborn outcomes”. The authors also noted that their conclusions are consistent with other studies previously published by other scientists.

The study wasn’t designed to assess COVID-19 vaccines effectiveness

In his post, McCullough highlighted a part of Table 1 from Jorgensen et al., which summarizes the characteristics of the study population, such as the mothers’ age, the infants’ weight at birth, and the preexisting health conditions of the mothers.

In particular, Table 1 indicates that the same proportion of women got COVID-19 during pregnancy, whether they belonged to the vaccinated or unvaccinated category. At first glance, such statistics would appear to support McCullough’s claim that the study revealed no benefits from vaccination.

However, this is a misinterpretation of what the study can and cannot reveal. Indeed, clinical or epidemiological studies are designed in order to answer specific scientific questions formulated in advance. They cannot always be used to analyze other questions than the one initially formulated. Doing otherwise requires caution and the understanding that it may introduce bias.

Sarah Jorgensen, a pharmacist at the University of Toronto and one of the authors of the study, told Health Feedback that their study “wasn’t designed to evaluate benefits to pregnant women”.

One possible source of bias in comparing the proportion of women who tested positive for SARS-CoV-2 is that the vaccinated and unvaccinated groups are different. The authors wrote that “compared with mothers in the unvaccinated group, those who were vaccinated during pregnancy were more likely to be older than 30 years, nulliparous, influenza vaccine recipients during either of the 2 previous influenza seasons [used as a proxy for health behavior], and residents of urban areas and areas with higher incomes”.

Thus, both groups are more likely to live in different areas and have different socioeconomic backgrounds. This may introduce differences in the way they interact with other people and in their health-seeking behavior, which might in turn affect their risk of exposure to SARS-CoV-2.

Another possible source of bias is that the group of vaccinated women who tested positive for SARS-CoV-2 doesn’t exclude women who tested positive before vaccination. The corollary is that the proportion of vaccinated women who were infected cannot reliably measure vaccine effectiveness, as the vaccine cannot have a retroactive effect.

Table 2 from the study suggests that this is a likely scenario, as it indicates that 20% of women received their first dose during the third trimester. This means that this group of women remained unvaccinated until the later stages of their pregnancy. If any of them tested positive for SARS-CoV-2 during the first or second semester, they would be counted both in the vaccinated group and in the SARS-CoV-2 positive group. As we pointed out above, these cases shouldn’t be used to infer vaccine effectiveness, because they weren’t vaccinated at the time of exposure.

This illustrates that McCullough used the study by Jorgensen et al. to answer a question that the study wasn’t designed to answer. By contrast, other published studies that are designed to assess vaccine effectiveness showed that COVID-19 vaccination during pregnancy reduced the risk of COVID-19[4,8-12].

The study showed that COVID-19 vaccination isn’t dangerous for pregnancy

The claim that COVID-19 vaccination was associated with an increased risk of newborn readmission to hospital is a misinterpretation of the study.

McCullough is referring to the crude hazard ratio (HR) of readmission. The HR measures the level of risk associated between an outcome like hospital readmission and vaccination in pregnancy. The 95% confidence interval (95% CI) measures the uncertainty around the estimation of that HR. When the 95% confidence interval of the hazard ratio is above one, it means that there is an association between hospital readmission and mother vaccination. But when the confidence interval crosses one, no association can be established.

Table 3 from Jorgensen et al. indicates that the crude 95% confidence interval of hazard ratio for readmission is higher than one: between 1.03 and 1.13 for neonatal readmission and between 1.00 for 1.08 for readmission up to six months of age. However, what actually matters is the adjusted HR, not the crude HR. This is because the vaccinated and unvaccinated groups are different, as we explained earlier.

The adjusted HR includes adjustments for possible confounding factors that could arise due to the differences between the unvaccinated and vaccinated groups. Confounding factors are variables that affect the outcome of an experiment, but aren’t the variables being studied in the experiment. Not accounting for confounders can lead to incorrect interpretations.

in this case, the authors had to adjust for “infant sex, calendar month and year of conception, maternal age at infant birth, parity, pre-existing maternal medical conditions […], maternal influenza vaccination during either the 2019 to 2020 or 2020 to 2021 influenza seasons (as a proxy for health behavior), adequacy of prenatal care […] maternal positive SARS-CoV-2 polymerase chain reaction (PCR) test result during pregnancy, neighborhood-level income quintile, neighborhood-level proportion of the population who self-identify as a visible minority quintile, rural residence”.

Once adjustments were made to account for all these possible confounders, the confidence intervals of the HR encompass one. So, there was no difference between the vaccinated and unvaccinated categories in terms of hospital readmission. “There was no increase in either outcome after adjusting for potential confounders ([…] adjusted hazard ratio 1.03 (95% CI, 0.98-1.09) and 6-month readmission: 8.4% [vaccine exposed] vs. 8.1% aHR 1.01 (95% CI, 0.96-1.05)), Jorgensen wrote in an email to Health Feedback.

McCullough also claimed that the analysis by Jorgensen and colleagues didn’t take into consideration some negative pregnancy outcomes such as stillbirth or miscarriages. While Jorgensen acknowledged that this information wasn’t part of the data they studied, she pointed out that other published studies had already showed that COVID-19 vaccines didn’t increase the risk of miscarriage[13]. Health Feedback also addressed this topic on multiple occasions.

Jorgensen also pointed out that McCullough incorrectly assumed that some cases of readmissions were missing from the study, such as when a mother gave birth in a hospital but then went to a different hospital for readmission. However, these cases are taken into account. Jorgensen clarified that “all hospital admissions across Ontario were indeed included, i.e., if an infant was discharged after birth from a hospital in Toronto and readmitted to a hospital a week later in Ottawa, this was included in our analysis”.

In summary, McCullough erroneously interpreted the study from Jorgensen et al. published in October 2023. This study aimed at assessing whether COVID-19 vaccination during pregnancy was associated with a greater risk in pregnancy complications. The study’s results indicated that COVID-19 vaccination was safe. However, this study cannot be used to infer vaccine effectiveness, as it wasn’t designed for it. There is already a wealth of data from other published studies, establishing that COVID-19 vaccination is beneficial during pregnancy, as it reduces the risk of severe COVID-19 and doesn’t increase the risk of pregnancy complications.

SCIENTISTS’ FEEDBACK


Sarah Jorgensen, Postdoctoral fellow, University of Toronto:
SF: Did your research show no benefits of vaccination for pregnant women?
Our study evaluated safety outcomes of infants of mothers who received the COVID-19 vaccine during pregnancy and wasn’t designed to evaluate benefits to pregnant women.

SF: Did your research find an increased rate of neonatal and all-cause readmissions among babies from vaccinated mothers?
We evaluated all-cause neonatal (between discharge from the birth admission and 28 days of age) and all cause 6-month readmissions.
There was no increase in either outcome after adjusting for potential confounders (neonatal readmission: 5.5% [vaccine exposed] vs. 5.1% [vaccine unexposed], adjusted hazard ratio 1.03 (95% CI, 0.98-1.09) and 6-month readmission: 8.4% [vaccine exposed] vs. 8.1% aHR 1.01 (95% CI, 0.96-1.05) ).
We hypothesized that the slight numerical (but not statistically significant) differences were because a larger proportion of infants of unvaccinated mothers were excluded from the analyses of these outcomes because they either died during the birth admission (0.06% [vaccine exposed] vs. 0.16% [vaccine unexposed], adjusted risk ratio 0.36 (95% CI, 0.25-0.52)) or their birth admission was longer than 28 days for neonatal readmission (0.95% vs. 1.5%, aRR, 0.59 (95% CI, 0.53-0.65)) or longer than 6 months for 6-month readmission (0.01% vs. 0.04%, aRR 0.37 (95% CI, 0.17-0.82). Essentially the most vulnerable infants were excluded from this analysis and there was a higher proportion of these infants in the vaccine unexposed group. But again, there was no significant increase in either outcome among infants of vaccinated mothers. And there were no increases in other outcomes evaluated in this study (severe neonatal morbidity, neonatal death, neonatal intensive care unit admission).

SF: Did you exclude miscarriages and stillbirths from your analysis?
Yes. Unfortunately reliable data on miscarriages are not available in the provincial databases we used for this study. Stillbirths following maternal COVID-19 vaccination in Ontario had already been assessed by my colleagues in another study[13].

If I might just add, there seems to be some confusion in the article by Dr. McCullough about admission to other hospitals not being captured in our analyses. All hospital admissions across Ontario were indeed included, i.e., if an infant was discharged after birth from a hospital in Toronto and readmitted to a hospital a week later in Ottawa, this was included in our analysis.
Exclusions due to data linkage problems referred to infants whose birth records could not be linked a maternal record (to create the mother-infant dyad). This occurred for ~3% of infant records.

 

REFERENCES

   

Published on: 10 Nov 2023 | Editor:

Health Feedback is a non-partisan, non-profit organization dedicated to science education. Our reviews are crowdsourced directly from a community of scientists with relevant expertise. We strive to explain whether and why information is or is not consistent with the science and to help readers know which news to trust.
Please get in touch if you have any comment or think there is an important claim or article that would need to be reviewed.

ifcn-fact-checkers-code-of-principles-signatory