Longitudinal study of the immunological response to the SARS-CoV-2 vaccine in Instituto Português de Oncologia of Coimbra professionals

Highlights

• •Immunoglobulin (Ig)G levels declined over time in seropositive and seronegative participants.
• •Seropositive health care workers showed consistently higher IgG titers after vaccination.
• •IgG levels after the second dose in seronegative were similar to the IgG levels after the first dose in seropositive participants.
• •Participants with first-dose reactions showed higher IgG levels than those without reactions.
• •Our real-world data reinforce the importance of booster doses in public health planning.

Abstract

Objectives

Vaccination is crucial for preventing the spread of the new wave of the COVID-19 pandemic. Messenger RNA (mRNA)–based vaccines were approved in Europe: Pfizer/BioNTech BNT162b2 and Moderna mRNA1237. This longitudinal study evaluated the humoral immune response to SARS-CoV-2 vaccination in a cohort of health care workers receiving two vaccine doses and a booster dose.

Methods

Blood samples were collected before and after vaccination to understand the dynamics of immunoglobulin (Ig)G levels over a long period. Vaccine adverse reactions (VARs) were monitored through a questionnaire.

Results

Seronegative and seropositive individuals presented a decline in IgG titters, with consistently higher titters in seropositive individuals. Vaccination may elicit a response similar in magnitude to natural infection because the IgG titters after the second dose in seronegative individuals were similar to the IgG titters after the first dose in seropositive individuals. When sex was analyzed individually, there was a greater peak in IgG titters in females. Volunteers who reported VARs to the first dose had higher IgG levels than did those who did not report VARs.

Conclusion

Importantly, our cohort was composed mostly of females, included few seropositive individuals, and elderly people were not included. This study, which was conducted over approximately 600 days, presents real-world data on the immune response to COVID-19 vaccination from a large cohort of health care workers.

Introduction

The emergence of the Wuhan COVID-19 outbreak at the end of 2019 led to a global pandemic, declared by the World Health Organization (WHO) on March 11, 2020. In 2023, the WHO declared the end of this pandemic on May 5. SARS-CoV-2 has caused substantial mortality, serious consequences for the economy, and health problems, including mental health problems [1,2]. These consequences will last for years to come [3]. These persistent impacts have highlighted the need to better understand immunity and optimize vaccination strategies as part of long-term pandemic preparation.

To prevent the spread of this pandemic, many countries implemented measures very early, such as promoting social distancing and reducing interpersonal contact. In the European Union, four vaccines have been authorized, based on two different principles: messenger RNA (mRNA) vaccines, where the genetic information about spike (S)-protein was encapsulated in a liposome, including Pfizer/BioNTech (BNT162b2) and Moderna (mRNA1237); and vector-based vaccines, such as Janssen (Ad26.COV2.S) and AstraZeneca (ChAdOx1), where the genetic information about SARS-CoV-2 protein was introduced into cells [4].

The mRNA vaccines were the first in distribution for the member states of the European Union, after being approved by the European Medicines Agency on December 2020. In Portugal, vaccination of priority groups, including health care professionals, began in late December 2020 [5]. Although the efficiency of these vaccines was demonstrated in clinical assays, it remains unclear whether the vaccine confers protection against asymptomatic infection, as well as the level or the time trend of the antibody’s response [6].

Several studies have been published on humoral immunity; however, research on vaccine efficacy and long-term protection remains scarce [7,8]. Understanding the protection of vaccines is crucial for monitoring the antibodies immunoglobulins (Ig)G that bind to the SARS-CoV-2 S protein receptor binding domain, which is considered the key target for SARS-CoV-2–specific antibodies [6,9]. The highest titters of IgG after the second dose of the Pfizer/BioNTech vaccine are observed between days 21 and 28, but these levels begin to decline after 4-6 months [7,10]. The behavior of humoral immunity varies from person to person, depending on different factors, such as age, sex, comorbidities, serostatus, vaccine types, and circulating variants [6,7,10,11]. People with associated comorbidities, older people, and men are more prone to produce lower antibody levels of anti-SARS-CoV-2 [10,12,13].

In this longitudinal cohort study, we aimed to characterize the vaccinated population from Instituto Português de Oncologia of Coimbra (IPOC) and determine the proportion of individuals who developed antibodies, describing the dynamics of immune response (IgG) before the vaccine, after the first dose, after the second dose, and subsequent doses. In addition, we examined the relationship between vaccine adverse reactions (VARs) and the immune response.

Methods

Study design and population

This longitudinal cohort study, spanning over 600 days, involved health care workers at the IPOC. Data were collected from December 2020 until September 2022.

Health care workers from the IPOC received their first dose of the mRNA vaccine COMIRNATY (Pfizer/BioNTech BNT162m2) or MODERNA (mRNA1273) at the end of December 2020, with the second dose administered 21 days later. This was a volunteer study, in which all professionals with at least one dose of the vaccine were included. The participants provided blood samples, which were collected in serum vacuum tubes containing a clot activator and gel separator (Vacuette tube 5 mL CAT Serum Separator Clot Activator). In this study, we considered different variables, such as age, sex, epidemiologic history, number of received vaccine doses, IgG results previous to vaccination, post–first dose (after 12 and 18 days of inoculation), post–second dose (after 7 and 14 days of inoculation), and subsequent serological results.

This study was approved by the IPOC’s ethics committee (TI 11/2021). Informed consent was obtained from all subjects involved in the study.

Questionnaire about VARs

Information on VARs was collected through questionnaires titled “Adverse Reactions to COMIRNATY vaccine - first dose” and “Adverse Reactions to COMIRNATY vaccine - second dose.” The questionnaires, produced on the Microsoft Teams platform, were sent via e-mail to all vaccinated workers, with participation being voluntary. Reactions were categorized and assigned numerical codes: 1 for “without reaction,” 2 for “local reactions,” and 3 for “systemic reaction.”

Anti-SARS-CoV-2–specific IgG and immunoassays

Blood samples from vaccinated workers were processed into serum by centrifugation of the drawing tubes (2000 g/10 minutes). The presence and quantification of binding IgG against SARS-CoV-2 in serum samples were established using an enzyme-linked immunosorbent assay (ELISA) directed to the S1 domain of the virus’s S protein, the EUROIMMUN Anti-SARS-CoV-2 QuantiVac ELISA (IgG) assay on the EUROIMMUN Analyzer I-2P. The manufacturer’s information states that the assay sensitivity was 90.3% from day 10 after the onset of symptoms and 93.2% from day 21, and the specificity was 99.8%. The upper detection limit of the assay was 120 IU/mL. For samples with results limits, dilution was performed according to the manufacturer’s specificities to ensure the results were within the linearity of the test.

Positive SARS-CoV-2 infections were detected via real-time reverse transcription–polymerase chain reaction.

Statistical analysis

Continuous variables are presented as the means, medians, and interquartile range (IQR), and categorical variables are presented as absolute frequencies and percentages. To compare the immune response between groups, the Mann–Whitney test was used. To access the evolution of the immune response over time, we employed mixed linear models. These tests were used because the data presented a longitudinal structure, in which the same measures were obtained repeatedly for the same individual over time. Two models were estimated: model 1 included only individuals who were not infected before or during the follow-up period of the study, and, model 2, where all volunteers were included. For the estimation, models with fixed effects were used to control for possible constant individual variables over time that could influence the results. Categorical variables, such as sex and VARs, were compared by using the chi-square test.

For statistical analysis, the SPSS software version 28.0 (IBM Corporation, NY, USA) was used. All the differences and associations with statistical significance were considered for P <0.05.

Results

Participant characteristics

A total of 381 health care workers were included in the study, 61 (16%) males and 320 (84%) females. The mean age ± SD was 45.3 ± 10.1 years (IQR: 38-53 years). Most of the cohort received three doses (77.4%) of the BNT162b2 or mRNA1273 vaccine. Among 94 infected individuals, 17 had COVID-19 before vaccination, whereas 77 were infected during the study follow-up (Supplementary Table S1). The professional roles of the participants included physicians, nurses, dentists, pharmacists, superior technicians, operational assistants, and technical assistants.

IgG responses following vaccination

To determine the kinetics of the IgG antibodies, measurements were taken before vaccination, after the first dose (after 12-18 days), after the second dose (measurements over 600 days), and subsequent measurements over the period of the study. The peak IgG response was reached after the second dose, with levels reaching 1136 RU/mL without previous SARS-CoV-2 infection and 1803 RU/mL in those with previous SARS-CoV-2 infection. IgG levels began to decrease after 50 days and were nearly undetectable after 290 days post–second dose. After the administration of the booster dose, there was a significant mean increase in IgG levels of 963.13 RU/mL (P <0.01) (Figure 1).

Figure 1.

Evolution of the median immune response (IgG) according to infection status before vaccination.

Ig, immunoglobulin.

Parameters affecting immune response

In Table 1, we present the parameters with a significant impact on the immune response measured by IgG levels after vaccination. Model 1, which included only participants without infection at any time during the study, revealed a high intercept (B = 633.01 [95% confidence interval {CI}: 474.63-791.40], P <0.001), suggesting a strong basal IgG response. Sex (B = –35.33, [95% CI: −125.95 to 55.33] P = 0.445) and age (B = −2.24, [95%: −5.61 to 1.13], P = 0.193) did not significantly affect IgG levels, suggesting that these parameters do not substantially affect the initial immune response. After the second dose, the IgG titters began to decline. The booster dose was strongly associated with increased IgG levels (B = 706.96 [95% CI: 8701.48-712.43], P <0.01), highlighting the effectiveness of this dose in increasing IgG levels. Regarding model 2, which included all volunteers, independently of the infection history, the intercept reflected also a basal response (B = 543.48, [95% CI: 393.49-696.47], P <0.01). Similarly to model 1, the variables sex and age did not present statistical significance (P = 0.077 and P = 0.973, respectively). However, when analyzing IgG levels individually, the female participants demonstrated a higher peak IgG titters than their male counterparts did (Figure 2). The decrease of IgG in this model was more accentuated (B = −0.79, [95% CI: −1.02 to −0.56], P <0.01), emphasizing the need for a booster dose (Supplementary Material Table S2).

Table 1.

Parameters that impact the evolution of the immune response (immunoglobulin G).

Parameter	Model 1			Model 2
	B	CI (95%)	P	B	CI (95%)	P
Intercept	633.01	(474.63-791.40)	0.000a	543.48	(393.49-693.47)	0,000a
Sex (male)	−35.33	(−125.98 to −55.33)	0.445	−77.04	(−162.32 to 8.23)	0.077
Age	−2.24	(−5.61 to 1.13)	0.193	0.05	(−3.05 to 3.15)	0.973
Days post–second dose	−0.33	(−0.60 to −0.11)	0.000a	−0.79	(−1.02 to −0.56)	0.000a
Booster dose	706.96	(701.48 to 712.43)	0.000a	506.98	(501.16-512.81)	0.019b
VAR first dose	79.82	(9.49-150.16)	0.026	87.13	(20.52-153.74)	0.010b
VAR second dose	−54.76	(−124.93 to 15.41)	0.126	−10.54	(−77.09 to 56.02)	0.756
Previous COVID-19				538.11	(84.46-991.76)	0.000a
COVID-19 during study				−514.96	(−853.16 to −176.76)	0.185

Note: B, regression coefficient; CI, confidence interval.

^aP <0.01

^bP <0.05.

Figure 2.

Median values of immune response evolution (IgG) according to sex (female and male).

Ig, immunoglobulin.

Previous infection led to a significant increase in the IgG response (B = 538.11 [95% CI: 84.46-991.76], P <0.01), demonstrating a potentially enhanced immunity. On the other hand, when the infection was acquired during the study, the impact on IgG levels was not significant (B= −514.96, [95% CI: −853.16 to −176.76], P = 0.185). This finding may reflect heterogeneity in the timing of infection relative to vaccination, severity of disease, and individual immune profiles, which were not fully captured in this analysis.

In Supplementary Material Table S3, the results show in more detail the IgG values produced by those who did or did not have an infection. The IgG titters after the first dose were statistically different between health care workers who had a previous SARS-CoV-2 infection and those who were uninfected, with values of 46.95 (25.3, 92.07) vs 1266.5 (131.08, 2717.88), P <0.01. Seropositive individuals presented a more robust response after the first vaccine dose than seronegative individuals (Table S3). However, IgG levels were not significantly different among seropositive individuals after two doses with infection-naïve people, who also received two doses, with titters of 1803.2 (1301.28, 2080) vs 1136 (729.6, 1802.24), P = 0.096 (Table S3). Over time, the kinetics of IgG for both groups was similar; however, as expected, the IgG levels in seropositive individuals were always higher than in seronegative individuals (Figure 1). Our results suggest that, in terms of IgG kinetics and peak titters, vaccination in infection-naïve individuals elicits a humoral response of similar magnitude to that observed after natural infection in previously infected individuals, since there were no statistical differences between IgG titters, with values of 1266.5 (131.08, 2717.889) vs 1136 (729.6, 1802.24) P = 0.606 (data not shown).

It also important to considerer the virological context during the study period. Between 2020 and 2022, the predominant SARS-CoV-2 variants circulating in Portugal included Alpha, Beta, and Omicron. However, detailed variant-specific data were not incorporated into the present analysis because the primary objective of was to characterize longitudinal IgG dynamics after vaccination.

One of the aims of this study was to relate adverse reactions to immune responses. To analyze this parameter, questionnaires were elaborated focusing on adverse reactions to the first and second vaccine doses. Table 1 presents the results regarding the VARs for both doses. In model 1, those who experienced adverse reactions after the first dose presented significantly higher median IgG values than those who did not experience adverse reactions (P = 0.026). In model 2, for the first dose, those who displayed adverse reactions had statistically significantly higher mean IgG values than the group who did not have adverse reactions (P = 0.010). For both models, concerning the second dose, the IgG titters were lower in those who reported adverse reactions than those who did not, but this difference was not statistically significant (P = 0.126; P = 0.756, respectively). For this analysis, we did not differentiate between types of adverse reactions; we considered all individuals who experienced any reaction, whether local or systemic, alongside those who did not report any reaction.

Discussion

On March 11, 2020, the WHO declared a global pandemic due to the dissemination of the SARS-CoV-2 virus, which causes COVID-19 disease [14]. In response to the urgency in controlling the spreading of the SARS-CoV-2 virus and prevent severe disease, there was a quick development and distribution of mRNA-based vaccines against the SARS-CoV-2 S protein. These vaccines demonstrated significant responses in combating the pandemic, which has resulted in more than 7 million deaths to date [15]; however, questions remained regarding the durability of vaccine-induced humoral immunity and the factors influencing variability in immune response.

In this longitudinal study, we characterized the response to mRNA SARS-CoV-2 vaccination, focusing on the IgG kinetics over approximately 600 days and the relationship between VARs and antibody levels in a cohort of health care workers from IPOC. The development of anti-IgG antibodies was consistent among all health care workers. The peak of antibody levels were reached after the second dose for infection-naïve individuals and after the first dose for previously infected individuals. The titters began to decline and approached near-baseline levels around 290 days after the second dose. Seropositive individuals had always higher titters than seronegative individuals, which support the concept of hybrid immunity, in which previous infection primes the immune system and vaccination further amplifies the response [11].

Several studies have correlated antibody titter levels post-vaccination between seropositive and seronegative individuals [16,17]. The findings vary across studies, where some reported a robust response in seropositive individuals [11,16,18], whereas others highlighted stronger responses in naïve-infected individuals [9].

Our results showed a robust response in health care workers with previous SARS-CoV-2 infection, indicating the presence of a memory response. However, in seropositive individuals, the second dose did not result in a much greater response compared with COVID-19–naïve subjects, which was in agreement with other studies [17]. Moreover, the response to the second dose in infection-naïve individuals was similar to the response observed in previously infected individuals after their first dose, suggesting that mRNA vaccines in infection-naïve individuals can elicit an immune response similar to that induced by natural infection. IgG levels were not significantly different between seropositive and infection-naïve individuals after two doses (P = 0.096), aligning with findings from other studies [17].

In our study, we observed no differences in the rate of antibody decline between seropositive and seronegative individuals. The literature presents three different hypotheses regarding antibody decline: (i) no differences between the two groups [8], consistent with our findings; (ii) a more pronounced decline in seropositive individuals [10]; (iii) a greater decline in seronegative individuals [19].

The relationship between age, sex, and vaccine-induced IgG production remains unclear. Some studies have reported differences in IgG production between males and females [20,21] whereas others have found no such differences [19,22]. In our cohort, we did not identify significant differences between males and females over time. On the other hand, when analyzing IgG individually, after the first and second doses, females developed higher antibody titters, which are in agreement with other studies [23]. Some researchers attribute these sex differences to factors such as the X chromosome [24] sex hormones and environmental factors [25]. In our cohort, the lack of a strong sex effect may be partly explained by the limited representation of males, with only 16.1% male participants, which may have influenced our results.

Most studies are in concordance with the decreased levels of IgG titters in older individuals [10,13,21,22]. However, some studies indicate that older people may present higher antibody levels, potentially due to more severe COVID-19 symptoms and elevated cytokine levels related to comorbidities, which, in turn, may stimulate humoral immunity to SARS-CoV-2 [14]. In our study, we did not find a statistically significant relationship between age and anti-IgG levels. However, we have to take into account that our study population was relatively young, with no participants older than 67 years, which may not fully capture the effects of aging on immune response.

We also explored the association between VARs and IgG levels. As expected, individuals who developed adverse reactions, whether local or systemic, presented higher IgG titters than those who did not, although this difference was observed only after the first dose. In model 2, which included all the volunteers, for the first and second dose, the development of adverse reactions was associated with an increase of 87.13 RU/mL and a decrease of 10.54 RU/mL, respectively. Once again, the literature is inconsistent when it comes to the association of antibody production with VAR [26]. Some studies report a positive correlation between antibody titters and post-vaccination symptoms [9,27], whereas others find no such association [28]. Our data also showed that females presented severe symptoms more often than males, which is consistent with the findings of other studies [19,29]. The variability in how adverse reactions are reported across different studies could explain some possible bias related to this parameter. Furthermore, this study only included health care workers who may be more aware of symptoms. Importantly, the absence of adverse reactions should not be interpreted as an indicator of reduced vaccine effectiveness or efficiency.

The kinetics of antibody response is not yet fully understood. As previously discussed, studies show inconsistent findings regarding post-vaccination antibody levels, taking into account several factors such as age, sex, vaccine type, emerging variants, and previous infection. This suggests that antibody titters might be influenced mainly by temporal trends and specific antibody targets rather than by patient-related factors [10]. Generally, a decline in antibody levels is observed after full vaccination or natural infection, reflecting the typical kinetics of an immune response [18].

Our research, which relied on the voluntary participation of IPOC’s professionals, has potential biases that must be considered. As a single-center study, blood samples from volunteers were not collected simultaneously, which may affect the precision of the inferred kinetics. In addition, males were under-represented, individuals aged over 67 years old were not included, and the comorbidities were not taken into account. The number of health care workers with previous SARS-CoV-2 infection or breakthrough infections was small, and some volunteers discontinued participation during the study. The timing of previous natural infections varied among participants, potentially affecting post-vaccination antibody kinetics. Variant-specific data were not incorporated, and we cannot distinguish the impact of different circulating variants on the measured responses. Notwithstanding these limitations, our study provides valuable real-world data from a large sample size. Studies on dynamics of antibodies produced by vaccines are crucial because they offer essential insights into population-level protection and the necessity for booster doses to avoid the reappearance of new waves or future pandemics [29]. Overall, our findings support the notion that temporal factors and booster administration play a central role in shaping antibody titters, beyond baseline demographic characteristics alone. To the best of our knowledge, no other study in Portugal has followed up participants for such a long period while assessing antibody kinetics and adverse reactions. Longitudinal studies of vaccine-induced antibody dynamics are extremely valuable to support local health services and public health planning. Notably, even recent articles and reviews [30] on antibody kinetics report considerable heterogeneity across populations, highlighting the importance of conducting long-term region-specific studies.

Similar to influenza, SARS-CoV-2 has become a seasonal disease. The results presented in this study demonstrate that antibody titters increase with booster doses. Therefore, comparable to the influenza vaccine, an annual SARS-CoV-2 vaccine is advisable to ensure greater protection.

Conclusion

In summary, we concluded that neither age nor sex influenced antibody titters over time. Nonetheless, when we analyzed IgG levels separately, females developed higher anti-IgG titters compared than males. Individuals with or without previous SARS-CoV-2 infection showed a similar behavior on immune response during this study. In fact, after the second dose, there were no differences between these two groups. In both models, differences in immune response were observed only after the first dose.

Although this study has certain limitations, including a single-center design and an underrepresentation of certain demographics, it provides valuable insights from a large population group over an extended period, contributing to the understanding of immune responses to COVID-19 vaccination.

Declaration of competing interest

The authors have no competing interests to declare.