Long-Term Durability and Dynamics of Anti-S-RBD IgG Response in Healthcare Workers: A Comparative Analysis of Homologous and Heterologous SARS-CoV-2 Vaccination Schedules in a 1-Year Serial Cross-Sectional Study

Introduction

The COVID-19 pandemic, which emerged in December 2019, placed global health systems under unprecedented pressure and exposed healthcare workers (HCWs) at the forefront of this struggle to a high risk of infection [1,2]. The protection of this critical population has been one of the primary goals of vaccine development and implementation strategies. During the pandemic process, numerous vaccines across different platforms such as inactivated virus, mRNA, and viral vectors were rapidly developed and put into use [3].

The “heterologous prime-boost” strategy, in which different vaccine platforms are used sequentially, gained popularity with the COVID-19 pandemic. The fundamental logic of this strategy is based on the principle that different vaccine platforms stimulate different components of the immune response (eg, humoral and cellular immunity) more effectively. For example, while one platform may generate a strong T cell response, another may trigger high titers of neutralizing antibody production. The combination of these 2 platforms has the potential to create a more balanced, potent, and broad-spectrum immunity, characterized by the concurrent activation of both humoral and cellular responses that collectively enhance cross-protection against emerging SARS-CoV-2 variants. While neutralizing antibodies are essential for preventing initial infection, a robust and diverse T cell response is critical for recognizing conserved epitopes across different variants, thereby providing a more resilient defense against viral evolution [4,5].

During the COVID-19 pandemic, the waning of post-vaccination immunity over time and the emergence of new SARS-CoV-2 variants necessitated the need for booster doses [6,7]. At this point, heterologous vaccination strategies stood out as a promising approach. Numerous clinical studies have shown that an mRNA booster dose generates superior neutralizing antibody and T cell responses compared to homologous regimens [8–11]. This effect is particularly pronounced when the primary vaccination was with an inactivated or viral vector vaccine. This superiority, however, is not universal, with some studies reporting comparable outcomes between homologous and heterologous schedules or even superior neutralizing antibody responses with homologous mRNA boosting [11,12], highlighting the need for further research.

Despite promising findings on short-term (up to 3 months) and medium-term (3–6 months) immunogenicity, a significant knowledge gap exists regarding the long-term durability of these responses. Data on antibody persistence beyond 1 year, the quality of immunological memory, and the modulating role of natural infection (hybrid immunity, defined as immunity developed from both vaccination and natural infection) are particularly limited [13,14]. Understanding the long-term efficacy of specific heterologous combinations, such as those widely used in countries like Turkey that initiated vaccination with inactivated vaccines (CoronaVac) before transitioning to mRNA vaccines (BNT162b2), is critical for shaping future public health policies.

Therefore, this study was designed to address this gap with the primary hypothesis that a heterologous vaccination schedule provides a more durable long-term antibody response than homologous schedules. We aimed to quantitatively evaluate the effects of different homologous (inactivated and mRNA) and heterologous (mRNA-boosted inactivated) vaccination schedules on anti-S-RBD IgG levels over a 1-year period in a cohort of high-risk healthcare workers (HCWs). Additionally, we sought to examine the contribution of natural infection history to this response and to interpret the findings to offer evidence-based recommendations for optimizing future vaccination strategies.

Material and Methods

STUDY DESIGN AND ETHICS APPROVAL:

This single-center, prospective, serial cross-sectional study was conducted at Çukurova University Faculty of Medicine Hospital between December 2, 2021, and September 28, 2022. The study protocol was prepared in accordance with the ethical standards of the Declaration of Helsinki, and approval was obtained from the Republic of Turkey Ministry of Health Scientific Research Platform and Çukurova University Faculty of Medicine Non-Interventional Clinical Research Ethics Committee (Date: October 1, 2021, Meeting Number: 115, Decision Number: 18). Written informed consent was obtained from all participants.

The interval between the last vaccine dose and the blood sampling was calculated individually for each participant and categorized into weekly intervals for analysis. Antibody levels were measured during 3 distinct calendar periods. To accurately reflect antibody kinetics, the time elapsed from the most recent vaccine dose to blood collection was calculated individually for each participant and reported on a weekly basis. Period 1 (December 2021): Samples were collected at a median of 38.0 weeks (range: 3.3–47.1 weeks) following the last vaccine dose. Period 2 (April 2022): Samples were collected at a median of 32.1 weeks (range: 2.6–66.9 weeks) following the last vaccine dose. Period 3 (September 2022): Samples were collected at a median of 50.6 weeks (range: 7.9–89.9 weeks) after the last vaccine dose. This standardization enabled a precise comparison of antibody persistence provided by different vaccine regimens relative to the time elapsed after vaccination.

STUDY POPULATION AND DATA COLLECTION:

The study included 307 healthcare workers who volunteered to participate. Individuals receiving immunosuppressive treatment, pregnant women, and those with a known history of cancer, HIV, or significant autoimmune disease were excluded. Participants’ demographic data, professions, comorbidity status, and history of past SARS-CoV-2 infection were collected through face-to-face interviews. To ensure data accuracy, participants’ COVID-19 vaccination history (vaccine type, number of doses, and dates), SARS-CoV-2 PCR test results, and dates of infection were verified and recorded in real-time by accessing the participants’ official digital health records via the Ministry of Health’s e-Nabız and Central Physician Appointment System (MHRS) platforms during the interview. No participants were lost to follow-up during the study period. As shown in Table 1, participants were grouped according to the vaccine combinations they received, reflecting the real-world application of Turkey’s national vaccination program.

INACTIVATED SARS-COV-2 VACCINE BY SINOVAC (CORONAVAC™):

CoronaVac™ (Sinovac Life Sciences, Beijing, China) is a traditional, whole-virion inactivated vaccine platform. It is manufactured by propagating the SARS-CoV-2 CZ02 strain in Vero cell cultures, followed by chemical inactivation utilizing β-propiolactone to eliminate viral infectivity while maintaining the structural conformation of the viral antigens. The formulation is adjuvanted with aluminum hydroxide to potentiate the humoral and cellular immune responses. Because it utilizes the entire virion, CoronaVac™ exposes the immune system to a broad array of viral structural proteins, eliciting neutralizing antibodies not only against the Spike (S) glycoprotein and its receptor-binding domain (RBD) but also generating responses against the Nucleocapsid (N), Envelope (E), and Membrane (M) proteins. The vaccine is administered as a 0.5-mL intramuscular injection containing 3 μg of viral antigen and offers significant logistical advantages due to its stability at standard refrigeration temperatures (2–8°C). Participants received 2 doses (or 3, if applicable) of the vaccine intramuscularly into the deltoid muscle, typically with an interval of 28 days between the primary doses, as per the national vaccination protocol.

:

The BNT162b2 vaccine, marketed as Comirnaty^® (Pfizer-BioNTech), is a nucleoside-modified messenger RNA platform formulated in lipid nanoparticles (LNPs). The incorporated mRNA sequence encodes the full-length SARS-CoV-2 Spike (S) glycoprotein, engineered with 2 stabilizing proline mutations to maintain the highly immunogenic prefusion conformation. The LNP delivery system, composed of a specific lipid admixture including the ionizable cationic lipid ALC-0315 and the helper phospholipid DSPC, protects the transcript from enzymatic degradation and facilitates efficient intracellular delivery. Immunologically, BNT162b2 elicits robust and highly specific humoral and cellular responses exclusively against the Spike protein and its receptor-binding domain (RBD). One vial (0.45 mL) contains 6 doses of 0.3 mL after dilution. One dose (0.3 mL) contains 30 micrograms of COVID-19 mRNA vaccine (embedded in lipid nanoparticles). In June 2021, the national vaccination policy was updated to offer individuals the choice to receive a primary 2-dose series of either the inactivated vaccine (CoronaVac) or the mRNA vaccine (BNT162b2), administered with a 28-day interval. Consequently, our study population encompassed diverse vaccination trajectories. One subset of participants, who had previously completed a primary 2-dose series of CoronaVac, opted to receive the BNT162b2 mRNA vaccine as their subsequent booster dose (heterologous regimen). Conversely, another subset of participants was entirely naive to the inactivated vaccine platform and started their primary immunization schedule exclusively with the BNT162b2 vaccine (homologous mRNA regimen).

SEROLOGICAL ANALYSES AND SAMPLING SCHEDULE:

Serum samples were collected at 3 strategic time points: December 2, 2021, April 20, 2022, and September 28, 2022. The study’s sampling timeline was strategically synchronized with the national COVID-19 vaccination program managed by the Republic of Turkey Ministry of Health. The dates were selected to correspond with the primary vaccination rollout for healthcare workers (inactivated vaccine in January 2021 and mRNA vaccine in June–August 2021) and the subsequent booster doses scheduled at 6-month intervals via the Central Physician Appointment System (MHRS). Accordingly, our sampling dates were synchronized with these national milestones to capture the anti-S-RBD IgG kinetics at clinically significant intervals, focusing on milestones such as 3, 6, and 12 months after vaccination. This approach, supported by individual week-based calculations of the time elapsed since the last dose, allowed for a robust evaluation of real-world antibody durability in correlation with the official vaccination timeline.

All serological analyses were conducted at the central laboratory of our tertiary university hospital. The central laboratory has been continuously accredited by the Joint Commission International (JCI) since 2006. Quantitative detection of IgG antibodies against SARS-CoV-2 was performed using the MAGLUMI 2000 series fully automated chemiluminescence immunoassay (CLIA) analyzer (Snibe Diagnostics, Shenzhen New Industries Biomedical Engineering Co. Ltd., Shenzhen, China). The MAGLUMI SARS-CoV-2 S-RBD IgG (CLIA) kit was utilized, which operates on an indirect chemiluminescence principle. During the assay, specific IgG antibodies in the serum samples were incubated with magnetic microbeads coated with purified recombinant SARS-CoV-2 S-RBD (Spike protein Receptor-Binding Domain) antigen to form immune complexes. Following a wash cycle to remove unbound substances, ABEI (N-(4-Aminobutyl)-N-ethylisoluminol)-labeled anti-human IgG antibodies were added. After a second wash, the chemiluminescent reaction was induced by the addition of trigger solutions (Starter 1+2), and the resulting light signal was measured as relative light units (RLU) by a photomultiplier.

The antibody concentration was automatically calculated based on a system-integrated “Master Curve” validated through periodic 2-point calibration. In accordance with the manufacturer’s instructions, results were interpreted using a cutoff value of 1.00 AU/mL; values <1.00 AU/mL were considered “negative”, while values ≥1.00 AU/mL were classified as “positive”. The measurement range of the assay was 0.180 to 100 AU/mL. The AU (arbitrary unit) used in this assay is fully standardized against the World Health Organization (WHO) international standard (NIBSC code 20/136). Consequently, 1 AU/mL is equivalent to 4.33 BAU/mL (binding antibody units). This ensures reliable data comparability at the international level. The choice of the S-RBD region as the target antigen ensures that the assay identifies antibodies critical for inhibiting viral entry into host cells via the ACE2 receptor. This CLIA method demonstrates a high degree of correlation with the viral neutralization test (VNT).

ADVERSE EFFECT DATA COLLECTION:

Data on local and systemic adverse effects were collected through face-to-face structured questionnaires. During these interviews, epidemiological data and clinical safety profiles – including the presence, duration, and severity of vaccine-related symptoms – were systematically recorded. To ensure the highest level of data integrity, participants’ COVID-19 infection history, including precise infection dates and SARS-CoV-2 PCR test results, were cross-referenced and verified via the Republic of Turkey Ministry of Health’s “e-Nabız” (Personal Health Record System) and “MHRS” (Central Physician Appointment System) with the participants’ consent.

STATISTICAL ANALYSIS:

Statistical analysis was performed using IBM SPSS Statistics Version 20.0. The normality of the data was assessed using the Shapiro-Wilk test. As the study utilized a serial cross-sectional design with non-normally distributed data, non-parametric tests for independent samples were employed. The Mann-Whitney U test and Kruskal-Wallis test were used for comparisons between 2 or more groups, respectively. Bonferroni correction was applied for post hoc analysis. To control for the potential confounding effect of time, data were categorized into 4 intervals based on the time elapsed since the last vaccine dose (≤12, 13–24, 25–48, and 49+ weeks). Spearman’s rho correlation coefficient was used to evaluate the relationship between the time elapsed since vaccination and antibody levels. Due to the serial cross-sectional design given the independent nature of the sampling periods, each data point was treated as an independent observation; thus, advanced longitudinal models (eg, mixed-effects models, GEE) were not statistically applicable. No missing data for the primary outcome were recorded; therefore, imputation was not necessary. Categorical variables were analyzed using the Pearson chi-square test. A P value of <0.05 was considered statistically significant. A formal sample size power calculation was not performed prior to the study, which is acknowledged as a limitation.

Results

DEMOGRAPHIC AND CLINICAL CHARACTERISTICS OF PARTICIPANTS:

The median age of the 307 healthcare workers included in the study was 42.0 years (IQR: 29.0–47.5), with mean age 39.0±11.6 years, and 76.5% (n=235) were women. A history of at least 1 chronic disease was present in 27% (n=83). No statistically significant relationship was found between sex, age groups, presence of chronic disease, and anti-S-RBD IgG levels (P=0.096, P=0.740, and P=0.598, respectively). The demographic characteristics are summarized in Table 2.

ANTIBODY RESPONSES ACCORDING TO VACCINE REGIMENS:

The median anti-S-RBD IgG level of HCWs with heterologous vaccination (n=201) was 109.1 (95% CI: 104.5–113.8), which was statistically significantly higher than those with homologous vaccination (n=106) at 103.9 (95% CI: 98.2–109.1) (P<0.001; rank-biserial correlation r=0.25, 95% CI [0.14, 0.35]).

When vaccine combinations were examined in detail, the lowest antibody response was observed in the group receiving 3 doses of inactivated vaccine (SSS) (median: 77.5; 95% CI: 70.1–84.9). The antibody level of this group was significantly lower than all other groups containing mRNA vaccines (BBB, SSB, SSBB, SSBBB) (P<0.001), with effect sizes for these comparisons ranging from moderate to large. This finding indicates that even a single mRNA dose added to an inactivated vaccine series significantly strengthens the humoral response.

Although heterologous vaccination provided higher antibody levels than homologous schedules overall (r=0.25), the primary source of this difference was the markedly lower performance of the SSS subgroup within the homologous cohort. The fact that both homologous mRNA (BBB) and heterologous schedules (SSB, SSBB, and SSBBB) produced high and comparable responses suggests that exposure to an mRNA dose, rather than the platform combination itself, is the main determinant of humoral immunity.

Our analyses indicate that there was no statistically significant difference in anti-S-RBD IgG levels between the homologous BBB group and the heterologous vaccination regimens (SSB, SSBB, SSBBB) (P>0.05). This confirms that the groups were comparable and had similarly robust humoral responses. The finding that effect size values in pairwise comparisons between the groups remain at ‘negligible’ or ‘very small’ levels indicates that utilizing the mRNA platform – whether in the primary series (BBB) or as a booster dose (heterologous) – establishes a similar immunogenic ceiling. Furthermore, the narrow confidence intervals between the high median values achieved by the BBB group and those of the heterologous groups (particularly SSBB and SSBBB) demonstrate that both strategies possess a high protective capacity among healthcare workers.

Intergroup analysis revealed that anti-S-RBD IgG levels in the homologous BBB group were comparable to those in the heterologous groups (SSB, SSBB, and SSBBB), with no statistically significant differences observed (P>0.05). Both homologous mRNA and mRNA-boosted heterologous regimens achieved high plateau levels of antibody titers, significantly outperforming the homologous SSS group. Antibody distribution according to vaccine groups is given in Table 3.

TIME-DEPENDENT CHANGE IN ANTIBODY LEVELS (WANING IMMUNITY):

The time elapsed since the last vaccine dose significantly affected antibody levels (P<0.001). Antibody levels peaked in the first 12 weeks (median: 135.7) and showed a gradual decline. While there is no universally established protective threshold, these levels remained well above the 1.00 AU/mL reactivity cutoff even after 49 weeks (median: 97.8). The time-dependent change in antibody levels is shown in Table 4.

When time-dependent antibody changes were analyzed by vaccine group, the antibody waning dynamics of different vaccination regimens were found to diverge significantly. In the SSS group, median anti-S-RBD-IgG levels showed no statistically significant difference relative to the time elapsed since vaccination (P=0.987, ɛe2<0.01). These data demonstrate that antibody levels in the inactivated vaccine group followed a low but stable course, with negligible time-dependent changes. In contrast, a significant decline characterized by a large effect size was detected in the BBB group (P=0.002, ɛ2=0.38). The median antibody level in the ≤12-week interval (124.5 AU/mL; 95% CI: 118.2–130.8) for the BBB group was significantly higher than all other time intervals (adjusted P<0.01, r=0.42). Similarly, significant reductions were identified in heterologous vaccination groups. In the SSB (P=0.004, ɛ2=0.28), SSBB (P=0.002, ɛ2=0.35), and SSBBB (p=0.013, ɛ2=0.22) groups, antibody levels in the early phase (≤12 and 13–24 weeks) were significantly higher than those in the late phase (25–48 and 49+ weeks). Specifically, in the SSB and SSBB groups, antibody levels peaked following mRNA supplementation but exhibited a declining trend over time with a medium-to-large effect size. The absence of data for the 49th week in the SSBBB group is due to its late implementation in the national vaccination schedule. Type I error for multiple testing was controlled using Bonferroni correction in all post hoc pairwise comparisons. The distribution of anti-S-RBD IgG according to the time elapsed since vaccination for each vaccine status is presented in detail in Table 5.

A total of 307 HCWs were evaluated across 5 vaccination groups. The homologous inactivated group (SSS) exhibited significantly lower anti-S-RBD IgG levels compared to all other regimens throughout the 48-week period (P<0.001). In contrast, the homologous mRNA group (BBB) and heterologous regimens (SSB, SSBB, SSBBB) achieved comparable and high antibody titers, with no statistically significant difference between them in early-phase measurements. Significant time-dependent waning was observed in all mRNA-containing groups (P<0.05); for instance, in the BBB group, a large effect size was detected (epsilon2=0.38) as levels declined from ≤12 weeks to 48 weeks. However, the SSS group maintained low but stable titers throughout the study (P=0.987, epsilon2<0.01) (Table 5).

EFFECT OF NATURAL INFECTION (HYBRID IMMUNITY):

Among participants with post-vaccination infections, the median interval between the final vaccine dose and infection was 22 weeks (range: 3–54 weeks; SD: ±8.4 weeks). Analysis of the 142 participants with confirmed COVID-19 infection revealed that 33.8% (n=48) were infected before vaccination, while 66.2% (n=94) experienced breakthrough infections occurring after at least 1 vaccine dose.

Data are presented as median and mean±SD. P values were calculated using the Mann-Whitney U test to compare infection-naive and hybrid immunity subgroups within each vaccination regimen (Table 6). The SSS (inactivated vaccine) group was the only group in which natural infection led to a statistically significant increase in antibody levels (P=0.014), with infection approximately doubling the antibody titers. In contrast, a “ceiling effect” was observed in all mRNA-containing groups (BBB, SSB, SSBB, and SSBBB), where infection provided no significant additional contribution to antibody levels (P>0.05). We found that mRNA vaccines, whether administered as a primary series or a booster, elicit a high humoral response and establish a peak antibody level. Consequently, no statistically significant difference was found between the antibody levels of participants with a history of COVID-19 (n=142) and those without (n=165) in the overall cohort (P=0.359). This lack of significance across the general cohort is primarily explained by the “ceiling effect” seen in mRNA-based regimens. Conversely, natural infection still significantly bolstered antibody levels in inactivated regimens (SSS) (P=0.014), indicating that hybrid immunity is a substantial gain for this specific group (Table 6).

The temporal distribution of post-vaccination infections remained notably low during the first half of 2021 (2.1%; n=3), coinciding with peak antibody responses. However, most infections occurred during the Omicron period, driven by the waning of immunity over time and increased variant escape. The distribution of primary infections according to the dominant variant periods in Turkey was: 30.3% (n=43) Wild-Type, 2.8% (n=4) Alpha, 23.9% (n=34) Delta, and 42.3% (n=60) Omicron. Despite the high infection rates recorded during the Omicron-dominant period, no participants had a severe clinical course requiring hospitalization. This observation confirms the robust clinical protection against severe disease afforded by intensive vaccination, independent of absolute antibody levels.

The high antibody levels detected during the major Omicron (BA.1/BA.2) wave in Turkey between January and April 2022 reflect both the immunogenicity of mRNA booster doses and the supplemental impact of natural infection (hybrid immunity) generated by this variant wave.

SAFETY PROFILE OF VACCINES:

When the adverse effects were examined, the rate adverse effects in those who received at least 1 dose of BNT162b2 vaccine (58.7%) was significantly higher than in those who received only CoronaVac vaccine (21.1%). The most frequently reported adverse effects for BNT162b2/CoronaVac (as%) were: pain in the vaccine arm 45.2/13.0; fatigue 15.1/4.9; fever/chills 11.1/1.6; pain at the injection site 10.7/3.0; tiredness 8.6/2.6; headache 8.6/4.0; joint pain 7.7/1.8; muscle pain 4.2/1.0; swelling at the injection site 3.5/0.6; redness at the injection site 3.5/0.6; nausea 2.5/1.4; back pain 1.6/2.0; cough 1.4/0.2; drowsiness 1.6/0.8; swelling in lymph nodes 1.2/0.0; dizziness 1.1/0.4; and vomiting 1.1/0.0. No serious adverse events were reported.

Discussion

Our study provides critical long-term, real-world data on SARS-CoV-2 vaccine schedules, revealing a key insight: receiving at least 1 mRNA vaccine dose, rather than the heterologous combination itself, is the primary driver of high-titer and durable humoral immunity. This main finding – that heterologous mRNA booster doses following a primary inactivated vaccine series generate a superior and more lasting anti-S-RBD IgG response compared to a homologous inactivated vaccine schedule – is consistent with many studies from Turkey and around the world, supporting the efficacy of this strategy [9,10,15]. However, our data allow for a more nuanced interpretation that moves beyond a simple heterologous versus homologous dichotomy.

One of the most significant findings of our study is the comparable long-term antibody levels observed between the homologous mRNA (BBB) group and the heterologous (SSB, SSBB, SSBBB) groups (P>0.05). This suggests that the robust immunogenicity of the mRNA vaccine platform itself, rather than a synergistic effect from combining different platforms, may be the dominant factor in achieving high and persistent antibody titers. While the heterologous prime-boost strategy is theoretically designed to elicit a broader immune response [4,5], our findings indicate that in the context of humoral immunity measured by anti-S-RBD levels, a full homologous mRNA schedule and an mRNA-boosted inactivated vaccine schedule can yield statistically indistinguishable outcomes over the long term. This aligns with recent studies suggesting that while heterologous boosting is highly effective, homologous mRNA boosting also provides a very strong and durable response, sometimes even superior in terms of neutralization capacity against certain variants [11,12].

A paradoxical but noteworthy finding was the antibody kinetics in the homologous inactivated vaccine (SSS) group. Despite having the lowest median antibody levels, this group was the only one to show no statistically significant decline over the 49-week follow-up period (P=0.987). This suggests that while inactivated vaccines may induce a less potent initial response, the resulting humoral immunity is remarkably stable. This could be hypothesized to result from the nature of whole-virion inactivated vaccines, which present a broader range of viral antigens to the immune system. This broader exposure might lead to the development of a more diverse B-cell memory repertoire that, while producing lower levels of anti-S-RBD–specific antibodies, demonstrates greater stability over time. This phenomenon warrants further investigation, as it could have implications for long-term protection against different viral components.

Our analysis of hybrid immunity further refines this picture. While there was no overall significant effect of natural infection on antibody levels (P=0.359), a subgroup analysis revealed a crucial interaction. Natural infection nearly doubled the median antibody level in the SSS group (P=0.014), whereas it provided no significant additional boost in any of the mRNA-containing groups. This strongly suggests a “ceiling effect” in individuals who received at least 1 mRNA vaccine dose, where the vaccine-induced immunity is so potent that the additional stimulus from a natural infection does not significantly raise the antibody titer. Conversely, the lower baseline immunity from the inactivated vaccine was primed for a powerful anamnestic response upon encountering the live virus. This finding is particularly relevant, as studies like Ntziora et al have shown that hybrid immunity can offer significantly greater protection against infection than vaccination alone [16]. Our data suggest this benefit is most pronounced in those with a less potent initial vaccine-induced response.

These immunological phenomena can be interpreted through established cellular mechanisms, although it must be stressed that these were not directly measured in our study. The superior response elicited by mRNA vaccines is likely due to their ability to induce potent germinal center (GC) reactions and strong follicular helper T (Tfh) cell support, leading to extensive somatic hypermutation and the generation of high-affinity memory B cells and long-lived plasma cells (LLPCs) [17,18]. The durability of this response, as seen in our heterologous groups, may be a clinical reflection of this robust memory formation. For instance, Takano et al demonstrated that a heterologous booster could elicit not just durable but also broader antibody responses against different variants [19]. The long-term stability of immunity relies on LLPCs residing in the bone marrow, a compartment that natural infection is known to seed effectively [20]. Our findings on hybrid immunity in the SSS group may reflect the successful activation of a memory response that leads to establishment of such a durable cell population.

From a public health perspective, the clinical significance of our findings deserves consideration. While the absolute difference in median antibody levels between some groups was modest, the consistently higher antibody ‘set-point’ and more durable response in all mRNA-containing schedules could translate to a longer duration of population-level protection. This is a critical factor in mitigating the impact of future waves of infection. For countries like Turkey, which initiated their national vaccination campaigns with inactivated vaccines, our results provide strong evidence for the adoption of heterologous mRNA boosters as a standard and effective public health strategy to enhance humoral immunity in the population.

Regarding safety, our findings are consistent with the established literature. The higher reactogenicity observed in participants who received at least 1 dose of the BNT162b2 vaccine is a known characteristic of the mRNA platform and is often correlated with its strong immunogenicity [21]. The fact that all reported adverse effects were mild and transient, with no serious adverse events recorded, confirms the favorable safety profile of all vaccination schedules used in our high-risk HCW cohort.

Our study has several limitations, including its single-center design, the lack of cellular immunity and variant-specific neutralization assays, and the use of a serial cross-sectional analysis rather than true longitudinal modeling. A pre-study power calculation was also not performed. However, these limitations are balanced by significant strengths. This study was conducted in a homogeneous and high-risk cohort of HCWs, featured a long follow-up period exceeding 49 weeks, included multiple vaccine combinations that reflect real-world clinical practice in Turkey, and utilized a quantitative and standardized method for antibody measurement. These strengths enhance the external validity and relevance of our findings.

Conclusions

This study indicates that heterologous mRNA (BNT162b2) booster doses following primary vaccination with inactivated vaccine (CoronaVac) provide a significantly higher and more durable humoral immune response for over 1 year compared to a homologous inactivated vaccine schedule. Our findings align with the literature suggesting that the superiority of heterologous vaccination is not limited to higher antibody titers alone but may also reflect the formation of a more robust immunological memory. While our study has limitations, including the lack of cellular immunity data, the results have important implications. They support the adoption of heterologous booster doses as a viable and effective strategy in the long-term management of the pandemic, especially for high-risk groups. Future research should focus on using longitudinal modeling to analyze immune dynamics and should include comprehensive assessments of cellular immunity, including memory B and T cell responses, to fully elucidate the mechanisms behind the durability of heterologous vaccine responses. Understanding the quality and breadth of protection against future variants remains a key priority for pandemic preparedness.