The COVID-19 pandemic has caused an unprecedented global crisis, and having lasted for more than 2 years, has resulted in over 524 million COVID-19 cases, claiming >6.27 million deaths by Mid-May 2022, causing huge levels of economic damage (1, 2). Although antiviral agents against SARS-CoV-2, the virus causing COVID-19, are becoming available, vaccines and public health interventions remain the most promising approach against this global peril. Notwithstanding anti-SARS-CoV-2 vaccines are successful in reducing the mortality and morbidity rates, and the level of neutralizing antibodies has correlated with protection against SARS-CoV-2 reinfection (3) as well as severe COVID-19 (4–6), breakthrough infections do continue to recur. Importantly, it still remains a conundrum, how long this vaccine-induced acquired immunity would last in an individual.

Several studies have investigated the dynamics and the duration of protection of neutralizing antibodies developed following the onset of natural SARS-CoV-2 infection or vaccination (7–11). The results thus far remain inconsistent, with some reporting rapid waning of antibodies months after virus exposure or following vaccine administration (7–13), whereas others report the prolonged presence of neutralizing antibodies (14–16). This discrepancy partly stems from numerous factors including differences in patient demography, vaccine type, and number of infections/vaccinations one has become exposed to, SARS-CoV-2 strains, intrinsic properties of immunity, underlying disease and many other miscellaneous factors. Given the complex interaction between intrinsic and extrinsic factors, the duration of protective immunity developed post-vaccination against COVID-19 still remains uncertain. Here, by using an appropriate statistical model, we investigated the factors such as age, gender, vaccine status, type of vaccine viz., a viral vector-based AZD1222 and an inactivated BBV152 vaccine as well as comorbidities and their association with antibody decay.

In this large population-based real-life investigation conducted in Chennai, Tamil Nadu, India, we studied 519 individuals examined for anti-SARS-CoV-2 S1 IgG antibody titers following either vaccination or recovery from documented COVID-19 infection. We investigated the factors associated with development of a breakthrough infection, as well as hospitalization, and correlated them with the dynamics of anti-SARS-CoV-2 S1 IgG titers, as well as factors that might be associated with the decay. We found that age and comorbid conditions were the two factors independently associated with development of a breakthrough infection and hospitalization in the study population. A combination of both age (>60 years) and underlying comorbid conditions were associated with increased risk for contracting a breakthrough infection and hospitalization by ~15 and ~10 times, respectively. Anti-SARS-CoV-2 S1 IgG decay was only observed among recipients of AZD1222, but not BBV152 and those who recovered from a natural SARS-CoV-2 infection. Due to the decay of anti-SARS-CoV-2 S1 IgG, we also reported that the risk of developing a breakthrough infection and hospitalization gradually increased by 0.85 times with each month.

It is pivotal to understand when and how a breakthrough infection with SARS-CoV-2 occurred in fully vaccinated individuals as it is paramount to determine how long the public health measures needs to be in place and whether or not a community required a booster dose (18). Immunity against viruses works primarily by inhibiting the infection phenomenon either by humoral (e.g., neutralizing antibodies) or by killing the infected cells via cell-mediated immune responses. While a vaccine works by generating immune memory in the form of memory B-cells and T-cells that permits a more rapid and intensified immune response against secondary infection; most vaccines are not completely designed to prevent exposure or transmission of an airborne pathogen such as SARS-CoV-2. Hence, acquisition of a breakthrough infection is determined by whether the vaccinated individual at the time of exposure has adequate levels of protective immunity to prevent the establishment of an infection (18). Many factors are known to influence immune surveillance including the age of the host, the dynamics of antibody responses (19), type/nature of vaccine used (4), interval between the vaccine doses (20, 21), underlying comorbid conditions and other health issues (viz., neoplasms and immunocompromised state) (22).

Several studies on the dynamics of anti-SARS-CoV-2 IgG levels post-vaccination and after recovering from a natural SARS-CoV-2 infection have revealed that the antibody levels induced by vaccines generally undergo rapid decay (over the months). One study reported that individuals who received a AZD1222 vaccine had a substantial decline in antibody levels after 6 months and that the decline was significantly associated with development of a breakthrough infection (23). Similarly, although individuals who received the BNT162b2 mRNA vaccine had high antibody titers compared to those who had survived a natural infection, the antibody titers experienced a rapid decay by up to 40% for every subsequent month; whereas the decrease of antibody levels was only <5% per month among convalescing individuals (24). This is consistent with our observation that individuals administered with AZD1222 experienced a more rapid attrition in IgG antibody levels.

B-cells that encounter their cognate antigens during an infection upon activation migrate to the center of the B-cell follicle, where they form germinal centers (GCs) (25, 26). Within the GCs, B-cells compete for a limited amount of T-cell-derived signals, such as cytokines and CD40 ligand that promote further maturation and differentiation into memory B-cells or plasmablasts (27, 28). Some of these plasmablasts will mature in the secondary lymphoid tissue itself into antibody-secreting plasma cells with short life spans; the other plasmablasts may enter into the circulation together with memory B-cells, home to bone marrow and other mucosal tissues, where they mature into long-lived plasma cells or memory B-cell that reside in these tissues to secrete antibodies for prolonged periods (25, 29, 30). Although both infection and vaccination can induce memory B-cells and plasmablasts that participate in humoral immune response, due to subtle differences in the nature of antigen stimulation, the memory B-cells generated in each case may be different. One study compared the memory B-cells induced following inoculation of a BNT162b2 mRNA vaccine and recovery from a natural infection and found that the mRNA vaccine induced robust plasmablast responses as compared to a natural infection that more prone to memory B-cells, thereby generating more plasma cells as well as better antigen-binding maturation (31). Another study compared the immune responses generated by both mRNA-based vaccine and the inactivated whole-virion vaccine and reported that the mRNA-based vaccine induced stronger humoral immune responses than the inactivated whole-virion vaccine (32). Further, the inactivated whole-virion vaccine induced significantly higher levels of IFN-γ response in CD4⁺ and CD8⁺ T cells as compared to that by the mRNA-based vaccine (32). Because T cell-derived signals (e.g., cytokines, ligands) are pivotal in promoting B cell maturation in germinal centers (27, 28), the inactivated whole-virion vaccine likely induced long-lived plasma cells and memory B cells.

Of note, there appears to be a fundamental difference between the adenovirus-based vaccine (AZD1222) and the mRNA-based subunit vaccine. Since adenovirus-based vaccine presents as a whole virus, it likely induces B-cell maturation much similar to inactivated SARS-CoV-2, and hence be able to generate long-lasting antibody responses. However, because this is an adenovirus-based vaccine and given that adenovirus is common in the population, the presence of anti-adenovirus neutralizing antibodies and anti-adenovirus specific T-cell response can prevent the vector from transducing the target cells, thereby limiting the efficacy of the vaccine (33). In fact, this likely could be a universal concern with all vaccines because of the presence of T and B cross-reactive memory responses to seasonal coronaviruses. Hence, it is difficult for a subunit vaccine that uses spike protein alone without adjuvants to induce long-lived plasma cells (34). These warrants improved vaccine formulations with suitable adjuvants to enhance antigenic stimulation.

Our study has also highlighted that the elderly age group (>60 years) and those with underlying comorbid conditions are at risk for acquisition of a breakthrough infection and hospitalization. Further, we have also reported that individuals administered with the AZD1222 also appear to undergo an accelerated decline in the levels of anti-SARS-CoV-2 S1 IgG. This likely could stem from an ongoing aging phenomenon involving the immune system known as immunosenescence (immune aging), where the generation of new T-cells appears to undergo progressive decline owing to thymic atrophy. The attrition is compensated for by the homeostatic proliferation of mature T-cells in the periphery. Eventually, the continually replicating mature T-cells undergo exhaustion due to telomere shortening (35) resulting in the expansion of senescent T-cell phenotypes indicated by the loss of co-stimulatory receptors (e.g. CD28, CD69) (36, 37), and de novo expression of co-inhibitory molecules such as killer-like immunoglobulin receptors (KIR) and PD-1 (35, 37). Our findings also reflect that the elderly with underlying comorbid conditions represent a high-risk population that requires additional medical attention, and specific measures to boost anti-SARS-CoV-2 immune responses (such as administering a second booster dose of the vaccines) in these groups are urgently warranted.

The limitation of this study is that we have not taken into account, the genetic variants of SARS-CoV-2 (38–41), especially the Delta and Omicron variants that have led to increased viral loads and high transmissibility. Notwithstanding, our study has provided a detailed information via a relatively large cohort involving both vaccinated and convalescent individuals recovering from SARS-CoV-2 infection. Our results indicate that the declining slope of anti-SARS-CoV-2 S1 IgG levels in AZD1222-vaccinated individuals is much steeper than in convalescent individuals and those who had received the BBV152 vaccine in Chennai, India. We have also provided an estimation of the rate of IgG decline as well as corresponding development of breakthrough infection and hospitalization risks by taking age, underlying comorbid conditions and time-scales into account. Given the expanding quantum of therapeutics against COVID-19, incapacitating the pandemic successfully is equally reliant on active public cooperation (42), which is of paramount importance to development of herd immunity against SARS-CoV-2.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Human Ethics Committee of the Madras Medical College (EC No. 03092021). The patients/participants provided their written informed consent to participate in this study.

The study was designed by SS, YY, GS, KN, SR, and ES. SS, GS, and KN provided regulatory oversight. SS and KN provided project management. SS, GS, KN, MR, and SR collected study data and oversaw participant visits. Participant data analysis and interpretation were done by YY, HT, YZ, ES, SR, and ML. Patient data collected and analyzed by SS, KN, GS, KV, PJ, DR, SP, SR, YY, HT, YZ, ES, SR, and ML and was interpreted by SS, KN, SR, YY, ML, and ES. SS, YY, HT, KN, GS, and ES wrote the manuscript. Data were accessed and verified by SS, GS, and SR. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.