Professional Documents
Culture Documents
Original article
a r t i c l e i n f o a b s t r a c t
Article history: Objectives: A weak correlation between symptom severity and antibody levels after primary immuni-
Received 11 January 2023 zation against COVID-19 has already been shown. This study aimed to describe the association between
Received in revised form reactogenicity and immunogenicity after booster vaccination.
17 May 2023
Methods: This secondary analysis of a prospective cohort study included 484 healthcare workers who
Accepted 22 May 2023
Available online 25 May 2023
received a booster vaccination with BNT162b2. Anti-receptor binding domain (RBD) antibodies were
assessed at baseline and 28 days after booster vaccination. Side effects were graded (none, mild, mod-
Handling Editor: L. Scudeller erate, or severe) and reported daily for 7 days after booster vaccination. Spearman correlation coefficient
(rho) was used to determine the correlations between the severity of each symptom and anti-RBD levels
Keywords: before vaccination and 28 days after. The Bonferroni method was used to adjust p values for multiple
Adverse events comparisons.
Antibody titers Results: Most of the 484 participants reported at least one local (451 [93.2%]) or systemic (437 [90.3%])
COVID-19 post-booster symptom. No correlations between the severity of local symptoms and antibody levels were
SARS-CoV-2
found. Except for nausea, systemic symptoms showed weak but statistically significant correlations with
Side effects
28-day anti-RBD levels (fatigue [rho ¼ 0.23, p < 0.01], fever [rho ¼ 22, p < 0.01], headache [rho ¼ 0.15, p
Symptoms
Vaccine 0.03], arthralgia [rho ¼ 0.2, p < 0.01], myalgia [rho ¼ 0.17, p < 0.01]). There was no association between
post-booster symptoms and pre-booster antibody levels.
Discussion: This study showed only a weak correlation between the severity of systemic post-booster
symptoms and anti-SARS-CoV-2 antibody levels at 28 days. Therefore, self-reported symptom severity
cannot be used to predict immunogenicity after booster vaccination. Anselm Jorda, Clin Microbiol
Infect 2023;29:1188
© 2023 The Authors. Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and
Infectious Diseases. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
https://doi.org/10.1016/j.cmi.2023.05.028
1198-743X/© 2023 The Authors. Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases. This is an open access article under
the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
A. Jorda et al. / Clinical Microbiology and Infection 29 (2023) 1188e1195 1189
included all post-booster symptoms, which showed a statistically thyroid disorders (51 [10.5%] of 484). As determined by the pres-
significant correlation with 28-day antibody levels in univariate ence of anti-nucleocapsid antibodies at baseline, few had previous
analyses. Diagnostic plots (Fig. S1) were visually assessed, and infections with SARS-CoV-2 (24 [5.0%] of 484). Of the 484 partici-
linear regression assumptions were considered to be adequately pants, primary immunization was performed with BNT162b2
met. No influential data points were detected. Residual distribution (Pfizer-BioNTech) in 148 (30.6%) participants and with ChAdOx1
was close to normal. Coefficients of the linear regression model (AstraZeneca) in 336 (69.4%) participants.
were exponentiated and interpreted as the percentage increase in
anti-RBD levels per-unit increase of the independent variable. To Symptoms and treatment after vaccination
assess the risk of multicollinearity in the model, we calculated the
univariate correlation coefficients between the independent vari- Most participants (473 [97.8%] of 484) experienced at least one
ables and the variance inflation factor (VIF). The VIF indicates how adverse reaction within 7 days of vaccination (Fig. 1). Any severe
much the variance of an estimated regression coefficient is post-booster symptom occurred in 170 (35.1%) of 484 participants.
increased because of collinearity. In this study, we considered VIF Any local symptom occurred in 451 (93.2%) of 484 participants. Any
values of >2.5 as meaningful collinearity and values > 10 as sig- systemic symptom occurred in 437 (90.3%) of 484 participants. The
nificant collinearity. most common local symptom was pain at the injection site (445
To assess the potential impact of selection bias on our findings, [91.9%]). The most common systemic symptoms were fatigue (388
we performed two sensitivity analyses, which assumed that pa- [80.2%]) and headache (321 [66.3%]). Analgetic or antipyretic
tients, who were excluded because they did not report symptoms, medication for post-booster symptoms was taken by 133 (27.5%)
experienced either no systemic symptoms (Model 1) or severe participants. Only 7 (1.4%) of 484 participants consulted a medical
systemic symptoms (Model 2). doctor.
All analyses were carried out using RStudio (R version 4.1.2 [1
November 2021]). Association between local symptoms and antibody levels
Results Fig. 2 depicts the associations between the severity of local post-
booster symptoms and antibody levels before and 28 days after
Study participants booster vaccination. No correlations between the severity of local
symptoms and antibody levels were found (Fig. 2).
In the primary study, 558 patients received the BNT162b2 vac-
cine. Nineteen patients were excluded because they did not receive Association between systemic symptoms and antibody levels
their primary immunization with AstraZeneca or BNT162b2.
Thirty-four patients had missing baseline (n ¼ 2) or follow-up Fig. 3 depicts the associations between the severity of systemic
(n ¼ 32) antibody levels. Twenty-one patients did not report post-booster symptoms and the antibody levels before and 28 days
symptoms and were removed from the analysis. A total of 484 after booster vaccination. The severity of fatigue, fever, headache,
participants were included in the final analysis. The characteristics arthralgia, and myalgia showed a very weak or weak but statisti-
of the study participants are provided in Table 1. Most participants cally significant correlation with 28-day antibody levels. The
were healthy and had no chronic diseases (340 [70.2%] of 484). The strongest correlation with 28-day antibody levels was observed for
most common conditions were hypertension (48 [9.9%] of 484) and fatigue (rho ¼ 0.23) and fever (rho ¼ 0.22).
No statistically significant correlation between systemic post-
booster symptoms and pre-booster antibody levels was found
Table 1 (Fig. 3).
Baseline characteristics of study population
Fig. 1. Relative frequency (%) of local and systemic post-booster symptoms, use of analgetic or antipyretic medication, and medical consultation within the first 7 d after booster
vaccination.
Sensitivity analyses findings indicate that absent or mild symptoms do not reflect an
inadequate immune response. Symptom severity can therefore not
Compared with the 484 patients in the main analysis, the 21 be used to predict the antibody response after booster vaccination
patients, who were excluded because of missing symptom report- with BNT162b2.
ing, showed significantly higher 28-day antibody levels (geometric Notably, we observed no correlation between local post-booster
mean [95% CI]; 27 712 [20 391e37 661] vs. 18 170 [16 893e19 544]; symptoms and antibody levels. Previous studies examining the
geometric mean ratio ¼ 1.5 [1.1e2.1]; p 0.011). correlation between reactogenicity and immunogenicity after pri-
To assess the potential impact of this imbalance on our main mary immunization with BNT162b2 also found no correlation be-
results, we performed two sensitivity analyses using extreme as- tween local symptoms and antibody levels [3,5,17,18]. It is unclear
sumptions about the symptoms of the excluded patients. Assuming why there is no correlation between local reactogenicity and
that all of the 21 excluded subjects experienced any severe systemic immunogenicity. In contrast to systemic reactogenicity, local reac-
post-vaccination symptoms (Model 1), the correlation coefficient togenicity may also be affected by injection technique or anatom-
between post-vaccination symptoms and 28-day antibody levels ical aspects, such as proximity to sensory nerve endings or damage
was 0.22 (p < 0.001) (Fig. S4). Assuming that all of the 21 excluded to blood vessels caused by the injection. These confounders may
subjects experienced no systemic post-vaccination symptom at all obscure the correlation between underlying immunological pro-
(Model 2), the correlation coefficient between post-vaccination cesses and subjective symptoms. The lack of correlation could also
symptoms and 28-day antibody levels was 0.14 (p 0.001) (Fig. S4). be explained by the fact that the local immunologic response differs
markedly from the systemic one. Intramuscular administration of
Discussion mRNA vaccines causes local inflammation that attracts neutrophils,
monocytes, and dendritic cells from the blood to the injection site
This is the first study to examine the association between by the production of chemokines and other inflammatory media-
reactogenicity and pre-booster and 28-day antibody levels in a tors, such as interleukin-1b, interleukin-6, and CXCL10 [19]. This
large cohort. Two previous smaller studies found no association innate immune response is likely the cause of local symptoms but
between symptoms and antibody levels after booster vaccination may not reflect the extent of the humoral response. Antibody
with BNT162b2 [6]. In contrast, we observed a weak but statistically production is a feature of adaptive immunity that requires the
significant correlation between the severity of systemic symptoms transport of mRNA-containing lipid nanoparticles or antigen-
and 28-day antibody levels. However, the associations were subtle expressing cells to lymph nodes, where dendritic cells and mono-
and do not suggest a clinically relevant relationship between sys- cytes contribute to antigen presentation and priming of T cells [19].
temic symptoms and 28-day antibody levels. Furthermore, these Subsequently, T follicular helper cells promote the differentiation of
1192 A. Jorda et al. / Clinical Microbiology and Infection 29 (2023) 1188e1195
Fig. 2. Relationship between local post-booster symptomsd(A) any local symptom; (B) local pain; (C) itching; (D) erythema, and anti-RBD antibody levels at baseline and 28 d. Data
are presented with box-whisker plots (median, IQR, and range). Empty circles indicate the geometric mean values. IQR, interquartile range; RBD, receptor binding domain.
B cells into antibody-producing plasma cells [20]. Co-stimulatory symptoms such as lymphadenopathy. Considering the important
mediator molecules of these cellular interactions, such as type I immunological role of local lymph nodes near the injection site,
interferons, TNF-alpha, and interleukin-6, also contribute to sys- qualitative data on lymph node swelling would have been desirable
temic symptoms such as fever and may correlate more strongly but not feasible in this study design. Third, although there is reli-
with the actual antibody response [19,21]. In addition, circulating able evidence of the correlation between antibody levels and pro-
SARS-CoV-2 antigens have been detected in human plasma after tection against COVID-19 [23], the clinical relevance of subtle
vaccination, which may also enhance systemic symptoms and the differences in antibody levels among the degrees of severity of
humoral immune response [22]. post-booster symptoms is unclear. Fourth, we had to exclude pa-
This study has several limitations. First, the explorative nature of tients who did not report their symptoms after vaccination. This
this analysis does not allow for proof of causality, which makes the exclusion may have introduced selection bias. However, two
interpretation of our results speculative. Second, we relied on self- sensitivity analyses in which assumptions were made about the
reported symptoms and did not record more challenging symptoms of excluded patients showed little effect on the
A. Jorda et al. / Clinical Microbiology and Infection 29 (2023) 1188e1195 1193
Fig. 3. Relationship between systemic post-booster symptomsd(A) any systemic symptom; (B) fatigue; (C) fever; (D) headache; (E) nausea; (F) arthralgia; (G) myalgia, and anti-
RBD antibody levels at baseline and 28 d. Data are presented with box-whisker plots (median, IQR, and range). Empty circles indicate the geometric mean values. IQR, interquartile
range; RBD, receptor binding domain.
1194 A. Jorda et al. / Clinical Microbiology and Infection 29 (2023) 1188e1195
Table 2
Multivariate linear regression analysis assessing the association between log-transformed anti-RBD antibody levels at day 28 and all included covariates
Variable Incremental per-unit change (%) Lower 95% CI Upper 95% CI p VIF
in anti-RBD levels at day 28
BMI, body mass index; COPD, chronic obstructive pulmonary disease; RBD, receptor binding domain; VIF, variance inflation factor.
correlation between post-booster symptoms and 28-day antibody Medical Scientific Fund of the Mayor of the City of Vienna. The
levels. Fifth, the independent variables, especially the systemic funding source did not influence the study design, conduct, or
symptoms, showed a moderate degree of collinearity, which may reporting.
otherwise have affected the validity of the regression model. Sixth,
data on reactogenicity after the first two vaccinations were not
Appendix A. Supplementary data
available. We reasoned that a retrospective assessment of post-
vaccination symptom severity months later would not have resul-
Supplementary data to this article can be found online at
ted in sufficient data quality.
https://doi.org/10.1016/j.cmi.2023.05.028.
Conclusion
References
This study showed only a weak correlation between the severity
[1] Jubishi D, Okamoto K, Hamada K, Ishii T, Hashimoto H, Shinohara T, et al. The
of systemic post-booster symptoms and anti-SARS-CoV-2 antibody association between adverse reactions and immune response against SARS-
levels after 28 days. Although the association between reac- CoV-2 spike protein after vaccination with BNT162b2 among healthcare
togenicity and immunogenicity would be plausible based on our workers in a single healthcare system: a prospective observational cohort
study. Hum Vaccin Immunother 2022;18:2048559. https://doi.org/10.1080/
immunological understanding, the weak correlation precludes the 21645515.2022.2048559.
clinically meaningful use of symptom severity for predicting [2] Tani N, Chong Y, Kurata Y, Gondo K, Oishi R, Goto T, et al. Relation of fever
immunogenicity. Importantly, absent or mild systemic symptoms intensity and antipyretic use with specific antibody response after two doses
of the BNT162b2 mRNA vaccine. Vaccine 2022;40:2062e7. https://doi.org/
after booster vaccination do not indicate an inadequate immune 10.1016/j.vaccine.2022.02.025.
response. [3] Lim SY, Kim JY, Park S, Kwon JS, Park JY, Cha HH, et al. Correlation between
reactogenicity and immunogenicity after the ChAdOx1 nCoV-19 and
BNT162b2 mRNA vaccination. Immune Netw 2021;21:e41. https://doi.org/
Author contributions 10.4110/in.2021.21.e41.
[4] Braun E, Horowitz NA, Leiba R, Weissman A, Mekel M, Shachor-Meyouhas Y,
The primary study was conceptualized by H.R., D.S., D.A., and et al. Association between IgG antibody levels and adverse events after first
and second Bnt162b2 mRNA vaccine doses. Clin Microbiol Infect 2022;28:
M.Z.; the secondary analysis presented in this article was concep- 1644e8. https://doi.org/10.1016/j.cmi.2022.07.002.
tualized by A.J., F.B., and M.Z.; the manuscript was written by A.J., [5] Hermann EA, Lee B, Balte PP, Xanthakis V, Kirkpatrick BD, Cushman M, et al.
F.B., and M.Z.; analyses of the data were performed by A.J. and R.R. Association of symptoms after COVID-19 vaccination with anti-SARS-CoV-2
antibody response in the Framingham Heart Study. JAMA Netw Open
All authors contributed to data curation and the design of the paper, 2022;5:e2237908. https://doi.org/10.1001/jamanetworkopen.2022.37908.
revision as well as editing of the paper. [6] Bauernfeind S, Einhauser S, Tydykov L, Mader AL, Salzberger B,
Hitzenbichler F, et al. Association between adverse reactions and humoral
immune response no longer detectable after BNT162b2 booster vaccination.
Transparency declaration Vaccines (Basel) 2022;10:1608. https://doi.org/10.3390/vaccines10101608.
[7] Matsumoto N, Hagiya H, Nakayama M, Furukawa M, Mitsuhashi T, Takao S,
Conflict of interest et al. Examining the association between vaccine reactogenicity and antibody
titer dynamics after the third dose of BNT162b2 vaccine using a mixed-effects
model. J Infect Chemother 2023;29:39e42. https://doi.org/10.1016/
The authors declare that they have no conflicts of interest. j.jiac.2022.09.012.
The following authors received financial support within the last [8] Radner H, Sieghart D, Jorda A, Fedrizzi C, Hasenohrl T, Zdravkovic A, et al.
Reduced immunogenicity of BNT162b2 booster vaccination in combination
3 years: D.A. from AbbVie, Amgen, Janssen, Lilly, Merck Sharp, with a tetravalent influenza vaccination: results of a prospective cohort study
Novartis, Pfizer, Roche, Sandoz, Sobi Sweden; HR from Gilead, Sci- in 838 health workers. Clin Microbiol Infect 2023;29:635e41. https://doi.org/
ence GmbH, Janssen, Merck Sharp, and Pfizer. 10.1016/j.cmi.2022.12.008.
[9] Higgins V, Fabros A, Kulasingam V. Quantitative measurement of Anti-SARS-
CoV-2 antibodies: analytical and clinical evaluation. J Clin Microbiol
Funding 2021;59. https://doi.org/10.1128/JCM.03149-20. 031499-20.
[10] Perkmann T, Perkmann-Nagele N, Breyer MK, Breyer-Kohansal R,
Burghuber OC, Hartl S, et al. Side-by-Side comparison of three fully automated
The study was funded by the Federal Austrian Ministry of Edu- SARS-CoV-2 antibody assays with a focus on specificity. Clin Chem 2020;66:
cation, Science and Research (21226.), the City of Vienna, and the 1405e13. https://doi.org/10.1093/clinchem/hvaa198.
A. Jorda et al. / Clinical Microbiology and Infection 29 (2023) 1188e1195 1195
[11] Food and Drug Administration. Guidance for industry: toxicity grading scale [18] Held J, Esse J, Tascilar K, Steininger P, Schober K, Irrgang P, et al. Reac-
for healthy adult and adolescent volunteers enrolled in preventive vaccine togenicity correlates only weakly with humoral immunogenicity after COVID-
clinical trials. Rockville, MD, USA: FDA; 2007. 19 vaccination with BNT162b2 mRNA (Comirnaty®). Vaccines (Basel) 2021;9:
[12] Overholser BR, Sowinski KM. Biostatistics primer: Part 2. Nutr Clin Pract 1063. https://doi.org/10.3390/vaccines9101063.
2008;23:76e84. https://doi.org/10.1177/011542650802300176. [19] Verbeke R, Hogan MJ, Lore K, Pardi N. Innate immune mechanisms of mRNA
[13] Zimmermann P, Curtis N. Factors that influence the immune response to vaccines. Immunity 2022;55:1993. https://doi.org/10.1016/
vaccination. Clin Microbiol Rev 2019;32. https://doi.org/10.1128/cmr.00084- j.immuni.2022.10.014. e2005.
18. 000844-18. [20] Teijaro JR, Farber DL. COVID-19 vaccines: modes of immune activation and
[14] Ooi EE, Dhar A, Petruschke R, Locht C, Buchy P, Low JGH. Use of analgesics/ future challenges. Nat Rev Immunol 2021;21:195e7. https://doi.org/10.1038/
antipyretics in the management of symptoms associated with COVID-19 s41577-021-00526-x.
vaccination. NPJ Vaccines 2022;7:31. https://doi.org/10.1038/s41541-022- [21] Trougakos IP, Terpos E, Alexopoulos H, Politou M, Paraskevis D, Scorilas A,
00453-5. et al. Adverse effects of COVID-19 mRNA vaccines: the spike hypothesis.
[15] Mahallawi WH, Mumena WA. Reactogenicity and immunogenicity of the Trends Mol Med 2022;28:542e54. https://doi.org/10.1016/j.molmed.
pfizer and AstraZeneca COVID-19 vaccines. Front Immunol 2021;12:794642. 2022.04.007.
https://doi.org/10.3389/fimmu.2021.794642. [22] Ogata AF, Cheng CA, Desjardins M, Senussi Y, Sherman AC, Powell M, et al.
[16] Li C, Bi H, Fu Z, Li A, Wan N, Hu J, et al. Retrospective study of the immuno- Circulating severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)
genicity and safety of the CoronaVac SARS-CoV-2 vaccine in people with vaccine antigen detected in the plasma of mRNA-1273 vaccine recipients. Clin
underlying medical conditions. Commun Med (Lond) 2022;2:151. https:// Infect Dis 2022;74:715e8. https://doi.org/10.1093/cid/ciab465.
doi.org/10.1038/s43856-022-00216-2. [23] Feng S, Phillips DJ, White T, Sayal H, Aley PK, Bibi S, et al. Correlates
[17] Zhang R, Leung KY, Liu D, Fan Y, Lu L, Chan PC, et al. Correlation of of protection against symptomatic and asymptomatic SARS-CoV-2 infec-
immunogenicity and reactogenicity of BNT162b2 and CoronaVac SARS-CoV-2 tion. Nat Med 2021;27:2032e40. https://doi.org/10.1038/s41591-021-
vaccines. mSphere 2022;7:e0091521. https://doi.org/10.1128/msphere. 01540-1.
00915-21.