11

Crosscutting Remarks

This chapter summarizes the conclusions made for each of the four COVID-19 vaccines under review and then presents the conclusions, including those about shoulder injuries related to intramuscular administration of any vaccine, by causal category. It offers a summary of information regarding evidence in children. Finally, it identifies methodologic challenges the committee encountered during its review.

The committee makes 85 conclusions about the causal relationship or lack thereof between vaccines and possible harms. Although it lacked evidence to establish, accept, or reject a causal relationship for many possible harms, it identified sufficient evidence for 20 conclusions. It is not surprising that evidence is insufficient for the majority; National Academies of Sciences, Engineering, and Medicine (National Academies) committees conducting similar reviews had similar results. The literature on the relationship between the four COVID-19 vaccines and many of the adverse events in the Statement of Task is sparse, not directly applicable to the question of causality, or conflicting or unconvincing due to study design problems, such as sample size limitations or case ascertainment. Fortunately, important research that sheds light on both the benefit and the potential harms of COVID-19 vaccines is published regularly. As discussed in Chapter 1, the committee incorporated into its definitions of the causal conclusions the understanding that further research may change a conclusion, although the committee thinks it is unlikely for conclusions establishing causality.

Given that this review occurred shortly after vaccines were available, the information in this report is a snapshot in time. New COVID-19 vaccines will be developed, and research will continue on many fronts. Understanding causation is a dynamic process; conclusions are refined as information accumulates. For example, the evidence reviewed in this report does not address real-world use in which many individuals received a “mix and match” sequence (i.e., some with BNT162b2 for their primary series received mRNA-1273 as a booster). Many people vaccinated for COVID-19 received other vaccines (e.g., influenza) simultaneously, and the effect of combined vaccination is not yet well explored. Most of the evidence regarding COVID-19 vaccines was from the primary series. Because children were among the last vaccinees, less evidence exists about them, especially for the youngest age groups (see subsequent section). These areas pose great opportunities for future research.

COVID-19 VACCINE–SPECIFIC CONCLUSIONS

The committee was not charged to evaluate the benefits of COVID-19 vaccines. All conclusions must be assessed in the context of the established harms of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2)

infection and the well-documented benefits of COVID-19 vaccines in preventing those harms. Most of the evidence the committee reviewed included BNT162b2 (see Box 11-1), which is not surprising, as it was the first vaccine available in the United States and many other countries; mRNA-1273 quickly followed, and many studies addressed it as well (see Box 11-2). The U.S. Food and Drug Administration (FDA) revoked the authorization of Ad26.COV2.S, and the number of studies reflected that short availability (see Box 11-3). NVX-CoV2373 is the most recently available vaccine in the United States—FDA issued an emergency use authorization (EUA) in July 2022 (FDA, 2022)—and the committee identified no epidemiological studies relevant to its review (see Box 11-4).

The committee made separate conclusions for each vaccine, even if they were of the same platform. However, the conclusions for the two messenger ribonucleic acid (mRNA) vaccines were almost identical:

• Evidence establishing a causal relationship with both vaccines and myocarditis;
• Evidence favoring rejection of a causal relationship between both vaccines and thrombosis with thrombocytopenia syndrome (TTS), infertility, Guillain-Barré syndrome (GBS), Bell’s palsy (BP), and myocardial infarction (MI) (numerous studies support the conclusions about GBS, BP, and MI; the evidence for TTS and infertility was more limited but still suggested no effect); and
• Evidence favoring rejection of a causal relationship between BNT162b2 and ischemic stroke, but the evidence was inadequate to accept or reject a causal relationship for mRNA-1273, as the data were more limited.

Despite the limited use of Ad26.COV2.S in the United States and a limited number of published studies, the committee identified sufficient evidence to favor acceptance of a causal relationship with two specific adverse events, TTS and GBS. The evidence bases for these two conclusions were very different. The conclusion about TTS relied heavily on strong mechanistic evidence that vaccination induced anti-PF4 antibody to platelets in people with TTS. Although the mechanistic findings for ChAdOx1-S were strong, it was not used in the United States. The similar, although less striking, mechanistic findings with Ad26.COV2.S, combined with pharmacovigilance data, led the committee to Conclusion 5-3: the evidence favors acceptance of a causal relationship between it and TTS. The data supporting Conclusion 3-1 about GBS were based on strong epidemiological studies and pharmacovigilance data.

SHOULDER INJURY CONCLUSIONS

The committee concentrated on case reports as the primary source of analysis, evaluating individual cases to arrive at the conclusions (see Box 11-5). Here, the committee was not limited to COVID-19 vaccines. The committee has examined evidence regarding shoulder injuries post-vaccination, exploring three potential mechanisms: direct trauma from improper placement, injury following injection regardless of technique, and vaccine constituents inducing harm, aiming to determine primary causative factors.

SUMMARY BY CAUSAL CATEGORY

The committee made six conclusions that the evidence establishes a causal relationship with vaccination (see Box 11-6); the evidence fell into two broad categories. The conclusions regarding the mRNA vaccines, BNT162b2 and mRNA-1273, and myocarditis relied on large epidemiological studies that were consistent with well-supported mechanistic evidence. Studies in animal models and ex vivo human samples show a connection between myocarditis and the activation of immune pathways, such as Toll-like receptor 4/inflammasome/interleukin (IL)-1β, triggered by mRNA COVID-19 vaccines. In patients with vaccine-associated myocarditis, the spike protein has been detected in myocardial tissue and is accompanied by elevated blood levels. The conclusions regarding certain shoulder injuries after intramuscular vaccination (independent of type) relied heavily on numerous well-documented case reports and a good mechanistic understanding that injection directly into certain areas of the shoulder could lead to injury.

The committee also made two conclusions that the evidence favors acceptance of a causal relationship for Ad26.COV2.S and GBS and TTS (see Box 11-7). As described, the evidence bases for these two conclusions varied.

The committee made conclusions favoring rejection of causality for 12 vaccine–adverse event relationships (see Box 11-8). Although the committee concluded that the evidence establishes a causal relationship with Ad26. COV2.S for GBS and TTS, it concluded that the evidence favored rejection with each of the mRNA vaccines. This supports the understanding that vaccine platform distinctly influenced the adverse physiologic and immune response. The committee also favored rejection of a causal relationship for the mRNA vaccines and several other outcomes: female infertility, BP, and MI. The committee favored rejection of a causal relationship between BNT162b2 and ischemic stroke but found that the evidence was inadequate to accept or reject a causal relationship between mRNA-1273 and ischemic stroke. The evidence base varied widely for these conclusions. The committee made one conclusion related to shoulder injuries, favoring rejection of a causal relationship for chronic rotator cuff disease following vaccination with any vaccine.

For most of the potential harms studied, the evidence was inadequate. Reasons for this include a paucity of studies (e.g., capillary leak syndrome [CLS]), difficulty in diagnostic accuracy (e.g., tinnitus), or methodological flaws, such as difficulty controlling for confounders. For some outcomes, the evidence was inadequate even given a large body of literature because the studies had conflicting results (e.g., pulmonary embolism after mRNA vaccination).

EVIDENCE IN CHILDREN

As noted in Chapter 1, potential vaccine-associated harms may differ in children and adults. For this reason, the committee conducted an in-depth review of the literature on adverse events to vaccines against SARS-CoV-2 specifically in children (under 18 years of age). At the time of committee review, data were available only for BNT162b2 and mRNA-1273. EUAs were later than for adults, and decreased uptake of vaccines in children, particularly those younger than 11, has led to far less data. Among the potential harms evaluated by the committee, infertility is not relevant to the pediatric population and has not been studied in children.

Children Younger Than 12

Few data exist for any possible harm other than myocarditis in children younger than 12. For myocarditis, surveillance studies of risk after COVID-19 vaccination have been conducted in children aged 5–11 (Walter et al., 2022), in addition to the original randomized clinical trials (Creech et al., 2022), as noted in the chapter on myocarditis. Multiple studies provide point estimates of risk (e.g., 1.3 cases per million children after the first dose and 1.8 cases per million after the second dose in one systematic review evaluating risk after either BNT162b2 or mRNA-1273 [Simões et al., 2023] or 1–5 cases per million according to the surveillance database used in a study of BNT162b2 [Watanabe et al., 2023]). A Danish surveillance study after BNT162b2 estimated incidence at 4.8 cases per million and used historical background incidence data to calculate a vaccine-associated myocarditis risk ratio of 4.6 (95% confidence interval: 0.1–156.1) (Hause et al., 2022). The absolute increase in risk from BNT162b2 and mRNA-1273 in the 5–11 age group appears to be less than in the 12–17 years and young adult age groups, but because of the epidemiological evidence, the magnitude of risk in this age group is uncertain. Additional research could shed light on the risk in this age group.

Data are sparse on the risk of myocarditis after COVID-19 vaccine in children 6 months to 4 years. The largest surveillance study, using the Vaccine Adverse Event Reporting System data, documented no cases of myocarditis after 599,457 doses of BNT162b2 or 440,773 doses of mRNA-1273 (Hause et al., 2021). Data in children 6 months to 4 years of age are insufficient to evaluate risk of myocarditis after COVID-19 vaccination, since myocarditis is rare in this age group.

Data in children under 12 on the association of COVID-19 vaccines with immune-mediated mechanisms (TTS, immune thrombocytopenic purpura. CLS), neurologic syndromes (GBS, chronic inflammatory demyelinating polyneuropathy, BP, transverse myelitis), postural orthostatic tachycardia syndrome (POTS), sensorineural hearing loss or tinnitus, sudden death, or thromboembolic events (MI, ischemic stroke, hemorrhagic stroke, deep vein thrombosis, pulmonary embolism, venous thromboembolism) are limited to small observational studies, case series, and case reports. In some cases, no data are available at all. The paucity of data on most adverse events in children, particularly children <12 years of age, highlights both the poor immunization rates in children, an equity issue that is important to address with improved access to COVID-19 vaccines for children, and the need for further study of acute and long-term adverse events after COVID-19 vaccination in children.

Children Older Than 12

More data are available on potential harms in children 12+ than in those younger. As outlined in Chapter 7, substantial data are available on the risk of myocarditis in children 12+ and show an increased risk of myocarditis for boys. Findings are summarized in that chapter; for example, in the systematic review Living Evidence Synthesis conducted by the Canada’s Strategy for Patient-Oriented Research, findings from 16 studies of children 12–17

years old provided an estimated range in boys of 13–390 cases per million, all estimates above population norms, but the range for girls was 1–50 cases per million (Su, 2021). The findings in children 12–17 are consistent with findings in adults of increased risk of myocarditis, particularly in boys.

For a number of other potential harms, data in large surveillance studies include children 12+ and adults, but pediatric-specific findings on potential harms were not analyzed. For example, some studies of neurologic outcomes included children 12–17, but none described findings specific to children. Similarly, a large study of multiple immune-mediated, thromboembolic, and neurologic outcomes included children 12–17 but did not include age-specific risk estimates (Hause et al., 2022). For these outcomes, multiple studies included individuals 16+, as outlined in the previous chapters, but made no separate evaluation of them. Data on chronic headache, POTS, or sudden death in association with the COVID-19 vaccine in children 12+ were limited to case reports.

The data reviewed highlight the paucity of information specifically in children on possible harms after COVID-19 vaccination. Because so little data are available for children, particularly those under 12, for most of the harms reviewed, and because of the insufficient time or immunization of younger children to detect infrequent harms, including harms that are infrequent or nonexistent in adults, ongoing and future pharmacovigilance and epidemiology studies will produce more definitive data on the risk and relative incidence of harms.

Shoulder Injuries in Children

Data on shoulder injury, as summarized in Chapter 10, are largely limited to case reports or small case series. Among specific shoulder injury diagnoses, pediatric case reports of the potential harm after vaccination were not found for subacromial/subdeltoid bursitis, axillary or radial nerve injury, bone injury, acute rotator cuff injury, or septic arthritis, suggesting that these are very rare in children. Pediatric cases or case series were reported for Parsonage-Turner syndrome, adhesive capsulitis, and complex regional pain syndrome after vaccination did not provide sufficient evidence for conclusions regarding risk.

CONCLUDING REMARKS

The COVID-19 pandemic resulted in voluminous research in many disciplines on many topics by many investigators conducted very quickly. Many factors complicated this research. Vaccines were approved or authorized for use at different times for different populations in different countries. Older people were among the first groups to receive the vaccine; they often have comorbidities that could have put them at risk for health problems simply concurrent with vaccination. The communities being vaccinated had widespread SARS-CoV-2 infection, so that few studies were able to exclude patients with an infection that occurred simultaneously with vaccination. Thus, some of the conditions might reflect harms from infection rather than vaccination. Epidemiological patterns of non-COVID-19 infections changed dramatically during the early days of the pandemic and the vaccination campaigns due to social distancing and other public health interventions. See the discussion on GBS in Chapter 3 as an example. This complicates the use of historical controls in some studies. Many publications report surveillance findings, which do not use comparison populations. Rather, comparisons are made to historical trends, not a true contemporaneous unvaccinated population. Other methodologic limitations in many of the studies include challenges in confirming vaccine receipt and in diagnostic validity.

Many studies reviewed by the committee in this report were not initiated to support causal inference reviews such as this. Thus, although a particular paper might have had limited utility to this committee, it likely has relevance and immense purpose to others. The committee appreciates the work of the researchers and participants involved in these studies often under the very difficult circumstances of an ongoing public health emergency and hopes that the information and conclusions in this report are useful to the vaccine research community at large.

REFERENCES

Creech, C. B., E. Anderson, V. Berthaud, I. Yildirim, A. M. Atz, I. Melendez Baez, D. Finkelstein, P. Pickrell, J. Kirstein, C. Yut, R. Blair, R. A. Clifford, M. Dunn, J. D. Campbell, D. C. Montefiori, J. E. Tomassini, X. Zhao, W. Deng, H. Zhou, D. Ramirez Schrempp, K. Hautzinger, B. Girard, K. Slobod, R. McPhee, R. Pajon, R. Das, J. M. Miller, and S. Schnyder Ghamloush. 2022. Evaluation of mRNA-1273 COVID-19 vaccine in children 6 to 11 years of age. New England Journal of Medicine 386(21):2011–2023. https://doi.org/10.1056/NEJMoa2203315.

FDA (U.S. Food and Drug Administration). 2022. Coronavirus (COVID-19) update: FDA authorizes emergency use of Novavax COVID-19 vaccine, adjuvanted. https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-authorizes-emergency-use-novavax-covid-19-vaccine-adjuvanted (accessed December 20, 2023).

Hause, A. M., J. Baggs, P. Marquez, T. R. Myers, J. Gee, J. R. Su, B. Zhang, D. Thompson, T. T. Shimabukuro, and D. K. Shay. 2021. COVID-19 vaccine safety in children aged 5–11 years—United States, November 3—December 19, 2021. MMWR: Morbidity and Mortality Weekly Report 70(51–52):1755–1760. https://doi.org/10.15585/mmwr.mm705152a1.

Hause, A. M., D. K. Shay, N. P. Klein, W. E. Abara, J. Baggs, M. M. Cortese, B. Fireman, J. Gee, J. M. Glanz, K. Goddard, K. E. Hanson, B. Hugueley, T. Kenigsberg, E. O. Kharbanda, B. Lewin, N. Lewis, P. Marquez, T. Myers, A. Naleway, J. C. Nelson, J. R. Su, D. Thompson, B. Olubajo, M. E. Oster, E. S. Weintraub, J. T. B. Williams, A. R. Yousaf, O. Zerbo, B. Zhang, and T. T. Shimabukuro. 2022. Safety of COVID-19 vaccination in United States children ages 5 to 11 years. Pediatrics 150(2). https://doi.org/10.1542/peds.2022-057313.

Simões, E. A. F., N. P. Klein, C. Sabharwal, A. Gurtman, N. Kitchin, B. Ukkonen, P. Korbal, J. Zou, X. Xie, U. N. Sarwar, X. Xu, S. Lockhart, L. Cunliffe, C. Lu, H. Ma, K. A. Swanson, K. Koury, P. Y. Shi, D. Cooper, Ӧ. Türeci, K. U. Jansen, U. Şahin, and W. C. Gruber. 2023. Immunogenicity and safety of a third COVID-19 BNT162b2 mRNA vaccine dose in 5- to 11-year-olds. Journal of the Pediatric Infectious Diseases Society 12(4):234–238. https://doi.org/10.1093/jpids/piad015.

Su, J. R. 2021. Adverse events among children ages 5–11 years after COVID-19 vaccination: Updates from V-Safe and the Vaccine Adverse Event Reporting System (VAERS). Centers for Disease Control and Prevention.

Walter, E. B., K. R. Talaat, C. Sabharwal, A. Gurtman, S. Lockhart, G. C. Paulsen, E. D. Barnett, F. M. Muñoz, Y. Maldonado, B. A. Pahud, J. B. Domachowske, E. A. F. Simões, U. N. Sarwar, N. Kitchin, L. Cunliffe, P. Rojo, E. Kuchar, M. Rämet, I. Munjal, J. L. Perez, R. W. Frenck, Jr., E. Lagkadinou, K. A. Swanson, H. Ma, X. Xu, K. Koury, S. Mather, T. J. Belanger, D. Cooper, Ö. Türeci, P. R. Dormitzer, U. Şahin, K. U. Jansen, and W. C. Gruber. 2022. Evaluation of the BNT162b2 COVID-19 vaccine in children 5 to 11 years of age. New England Journal of Medicine 386(1):35–46. https://doi.org/10.1056/NEJMoa2116298.

Watanabe, A., R. Kani, M. Iwagami, H. Takagi, J. Yasuhara, and T. Kuno. 2023. Assessment of efficacy and safety of mRNA COVID-19 vaccines in children aged 5 to 11 years: A systematic review and meta-analysis. JAMA Pediatrics 177(4):384–394. https://doi.org/10.1001/jamapediatrics.2022.6243.

This page intentionally left blank.