Effectiveness and Safety of mRNA COVID-19 Vaccines in Organ Transplant Patients

The protocol for this ancillary systematic review was prospectively registered in the Open Science Framework [14]. A comprehensive literature search was conducted across three electronic databases: Medline (via PubMed), Cochrane Central Register of Controlled Trials (CENTRAL), and Google Scholar, to identify eligible studies that evaluated the efficacy and safety of COVID-19 vaccines in organ transplant patients. Details of the search strategy are summarized in the Electronic Supplementary Table S1. The search was last updated on 14 January 2025. We did not place any restrictions regarding the language or year of the publications. We excluded conference abstracts, reviews and editorials. References from the eligible articles were examined and references pertinent to this study were included.

Two authors independently were involved in screening the articles for their suitability to be included. The full texts of potentially relevant articles were retrieved and independently assessed for eligibility and the following details were extracted: study characteristics (first author, year, country, design, setting, follow-up period); population (sample size, specific organ transplantation), age, sex, comorbidities; intervention (vaccine name, number of doses, timing of doses); and outcome data (definition of seroconversion, number of events and total participants in each group, and safety assessment). The risk of bias of included studies was assessed using Cochrane tool for RCTs and Joanna Briggs Institute (JBI) tool for other study designs [15,16]. Statistical analysis was carried out using R (version 4.5.1). Publication biases were assessed using Egger's regression analyses. Single pooled proportion using random-effects modeling was estimated for seroconversion rate across studies. The following subgroup analyses were carried out: vaccine types, type of transplanted organs, number of vaccine doses and prior COVID-19 infection. Leave-one-out sensitivity analyses were conducted to assess the robustness of findings by excluding studies one by one and assessing their impact on the pooled estimates. Bootstrap analyses with 1000 iterations were carried out to improve the robustness of the findings.

This systematic review and meta-analysis included a total of 125 studies [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141] from 762 studies (refer to Figure 1). The key characteristics of included studies are summarized in the Electronic Supplementary Table S2. Most studies were observational in design (n = 115), with 7 randomized controlled trials and 3 non-randomized trials also identified. Kidney transplant recipients constituted the largest population subgroup, featured in 95 cohorts, followed by liver (n = 39), heart (n = 29), and lung (n = 25) transplant recipients, with several studies (n = 38) including multi-organ transplant populations. The age of participants across studies ranged from 13.5 to 84 years, and males consistently represented the predominant proportion of the study populations, accounting for over 60% of participants. The most frequently administered vaccines were mRNA-based (n = 123), specifically the BNT162b2 (Pfizer-BioNTech) (n = 111) and mRNA-1273 (Moderna) platforms (n = 62), evaluated across various dosing schedules from one to five doses, with two- (n = 87) and three-dose (n = 30) regimens being the most common. One study administered double dose mRNA-1273 vaccine. The timing of seroconversion assessment post-vaccination varied widely, ranging from 10 days to over 12 months, though most studies (n = 103). Risk of bias assessment revealed that majority of the included studies had low risk amongst RCTs and moderate quality amongst the observational designs.

Details of the seroconversion reported in each included study are summarized in the Electronic Supplementary Table S3. The overall pooled seroconversion proportion was estimated to be 0.49 (95% CI, 0.43 to 0.55), reflecting 12,136 events. To view the comprehensive forest plot representing the individual data points for all 125 included studies, please refer to Electronic Supplementary Figure S1. However, a significant degree of heterogeneity was observed across the studies, as evidenced by the high I² statistic of 94.2% (p < 0.001) due to which the overall pooled estimate should be interpreted with caution as a aggregate of diverse clinical contexts rather than a definitive predictor for individual patients.

The meta-analysis of seroconversion proportions stratified by vaccine type reveals distinct differences in pooled estimates and levels of heterogeneity across the four assessed vaccines. The BNT162b2 vaccine (refer to Electronic Supplementary Figure S2) demonstrated an overall pooled seroconversion proportion at 0.49 (95% CI: 0.41 to 0.57) across 7213 events, though it was associated with high heterogeneity (I² = 93.3%). The mRNA-1273 vaccine (refer to Figure 2) showed a high pooled proportion of 0.55 (95% CI, 0.46 to 0.63) based on 4093 events, also exhibiting significant heterogeneity (I² = 90.5%). In contrast, the Ad26.COV2.S vaccine (refer to Figure 3), analyzed with fewer events (n = 85), had a lower pooled seroconversion proportion of 0.41 (95% CI, 0.13 to 0.77), and similarly high heterogeneity (I² = 85.6%). The Coronavac vaccine (refer to Figure 4) displayed the lowest pooled seroconversion proportion at 0.29 (95% CI, 0.14 to 0.50) from 1303 events, while still being highly heterogeneous (I² = 94.1%).

The forest plots of pooled seroconversion proportions among solid organ transplant recipients revealed considerable differences in immune response across different organ groups. The pooled seroconversion proportions generally followed a decreasing trend, with the highest rates observed in liver transplant recipients (0.8; 95% CI: 0.72, 0.86; refer to Figure 5), followed closely by kidney transplant recipients (0.47; 95% CI: 0.39, 0.54; refer to Electronic Supplementary Figure S3) and heart transplants (0.46; 95% CI: 0.27, 0.66; refer to Figure 6). Conversely, the immune response was markedly blunted in lung transplant recipients (0.29; 95% CI: 0.08, 0.65; refer to Figure 7) and multiple organ transplant recipients (0.43; 95% CI: 0.34, 0.53; refer to Figure 8).

The analysis of seroconversion proportions stratified by prior SARS-CoV-2 infection status revealed a substantial difference in immune response following vaccination. In the cohort of patients with a prior history of COVID-19 infection (refer to Figure 9), the pooled seroconversion proportion was exceptionally high at 0.90 (95% CI, 0.82 to 0.94) across 96 events. This strong and consistent response is further supported by the minimal heterogeneity observed in this group (I² = 0%), suggesting a highly predictable serological response to vaccination among those with pre-existing natural immunity. Conversely, in the large cohort without a prior history of infection (refer to Electronic Supplementary Figure S4), the pooled seroconversion proportion was significantly lower at 0.49 (95% CI, 0.46 to 0.52) across 12,703 events. This group exhibited an extremely high degree of heterogeneity (I² = 94.7%). Strikingly, stratification by prior SARS-CoV-2 infection was the only analysis that successfully resolved study-level heterogeneity. In the prior-infection cohort, the I² was reduced to 0%, suggesting that natural infection acts as a powerful unifying immunologic stimulus that overrides the variability seen in infection-naïve populations. In contrast, the infection-naïve group retained extreme heterogeneity (I² = 94.7%), indicating that other factors (e.g., immunosuppressive intensity) continue to drive variance in this subgroup.

The analysis of seroconversion proportion as a function of the number of vaccine doses administered reveals a clear, positive dose–response relationship. The seroconversion proportion exhibited a progressive increase with each additional dose, commencing at 27% after a single dose (refer to Figure 10). This proportion nearly doubled to 45% after the second dose (refer to Electronic Supplementary Figure S5). A further significant gain was observed with the third dose, which raised the seroconversion proportion to 61% (refer to Figure 11). The highest proportion was achieved after the fourth dose, reaching 84% (refer to Figure 12).

The bootstrap analyses confirmed higher seroconversion rates observe with mRNA-1273 vaccine type, liver transplants, prior COVID-19 infections and higher number of vaccine doses (refer to Table 1).

Leave-one-out sensitivity analysis revealed that no significant impact was observed on the pooled estimates with the removal of data from each study included in most of the analyses except for CoronaVac vaccine, lung and heart transplant recipients, one dose and those with prior COVID-19 infection (refer to). Electronic Supplementary Table S4

Egger's regression analyses did not reveal presence of any publication bias (refer to Table 2).

When stratified by antibody specificity, anti-RBD antibodies demonstrated the highest seroconversion rate at 54.6% (95% CI: 48.1–61.0%), while anti-spike IgG antibodies showed a slightly lower rate of 48.7% (95% CI: 42.8–54.7%). Notably, the combination of both anti-spike IgG and anti-RBD antibodies yielded the lowest seroconversion proportion at 27.1% (95% CI: 15.4–43.2%). Quantitatively, our results indicate that the use of anti-RBD assays alone may overestimate vaccine-induced immunity by approximately 27% relative to dual-antibody testing. Conversely, the requirement for dual anti-spike and anti-RBD positivity introduces a conservative bias that likely underestimates biological priming. These findings suggest that the choice of assay and the stringency of cutoff thresholds can shift pooled seroconversion estimates by nearly 27% in either direction, contributing significantly to the observed inter-study heterogeneity.

Of the 125 included studies, safety profile has been reported only in 36. Details of safety concerns reported in the studies included are mentioned in the Electronic Supplementary Table S5. In general, the COVID-19 vaccines demonstrated a favorable safety profile. Local and systemic adverse events were predominantly mild to moderate in severity, with the mostly local reaction being injection-site pain (median report rate: 59%) and frequent systemic events including fatigue (median rate: 32%), headache (median rate: 22%), and myalgia (median rate: 19.5%). Serious adverse events and episodes of acute allograft rejection were infrequent, with only isolated cases reported, such as biopsy-proven acute rejection [36,42,56,73,126,133]. A substantial number of studies explicitly concluded that the vaccines were safe, with no vaccine-related deaths or significant safety signals, and several noted no observed episodes of acute organ rejection. However, the absence of safety signals reflects limited reporting rather than confirmed safety equivalence.

This systematic review and meta-analysis of COVID-19 vaccination in solid organ transplant recipients establishes that this vulnerable population exhibits a significantly blunted seroconversion rate compared to the public, with an overall pooled proportion of only 0.49 (95% CI, 0.43 to 0.55) across 125 studies, underscoring the challenge of achieving protective immunity in this highly immunocompromised group. Crucially, the analysis identified several powerful determinants of vaccine success: prior SARS-CoV-2 infection was the strongest predictor of a robust response, resulting in an exceptionally high seroconversion rate of 0.90, effectively normalizing the immune response in SOTRs with hybrid immunity. Furthermore, seroconversion rates demonstrated a clear, positive dose–response relationship, escalating progressively from 27% after one dose to 84% after four doses, highlighting the critical necessity of booster regimens. While the immune response varied significantly across vaccine types and transplanted organs, with Liver recipients and mRNA-1273 vaccines showing relatively higher proportions than Lung recipients and CoronaVac, the consistent finding of high heterogeneity across nearly all subgroups emphasizes that vaccination strategies must be highly individualized, considering both the patient's history and the specific organ and regimen.

Our findings not only align with the results of previous syntheses but also substantially advance the understanding of COVID-19 vaccine effectiveness in SOTRs by providing a single, comprehensive quantitative framework assessing organ type, dose number, and prior infection status. Earlier meta-analyses consistently reported pooled seroconversion rates across different dosing regimens ranging from 20% to 60% [6,7,8,9,12]. Our overall pooled seroconversion rate of 0.49 (95% CI: 0.43–0.55) closely mirrors these findings, validating the consistency of the attenuated immune response in SOTRs while offering finer granularity by dissecting outcomes by transplant type, vaccine platform, and dose sequence. Our study presents robust organ-specific gradients that reaffirm and provide more precise estimates than prior syntheses. Earlier pooled analyses identified liver transplant recipients as the most responsive subgroup and lung transplant recipients as the least responsive, reflecting fundamental differences in immunosuppressive intensity and graft immunobiology. For instance, Chen et al. reported second-dose seroconversion rates of approximately 64.5% in liver, 37.6% in kidney, 50.3% in heart, and 28.4% in lung transplant recipients [12]. Our pooled estimates confirm this established hierarchy but reveal a greater separation among groups. Liver transplant recipients demonstrated the highest seroconversion rate (0.80; 95% CI: 0.72–0.86), while kidney and heart recipients exhibited intermediate rates (0.47; 95% CI: 0.39–0.54 and 0.46; 95% CI: 0.27–0.66, respectively). Lung transplant recipients remained the least responsive, with a pooled rate of 0.29 (95% CI: 0.08–0.65). These patterns align with the biological plausibility that liver transplantation is often associated with a more tolerogenic immune environment, permitting reduced immunosuppression. In contrast, lung transplantation necessitates more intensive regimens due to a higher risk of rejection and persistent immune stimulation, factors that profoundly impair B-cell regulation and antibody production [14,15]. Our platform-specific analysis facilitates a robust assessment of relative seroconversion rates. Earlier meta-analyses [7,8,11] indicated that mRNA-1273 demonstrates superior seroconversion proportions compared with BNT162b2 after two or three doses. For instance, Manothummetha et al. reported seroconversion rates of 44% (0–79.1%) for BNT162b2 and 51.4% (29.9–76.2%) for mRNA-1273 [8,9]. Similarly, our analysis showed that mRNA-1273 achieved the highest pooled seroconversion rate (0.55; 95% CI: 0.46–0.63), followed by BNT162b2 (0.49; 95% CI: 0.41–0.57). The Ad26.COV2.S vaccine (0.41; 95% CI: 0.13–0.77) and CoronaVac (0.29; 95% CI: 0.14–0.50) consistently elicited lower responses. Several factors may contribute to this difference: mRNA-1273 delivers a higher mRNA dose and utilizes a longer dosing interval (four weeks versus three weeks), which is consistent with evidence that extended intervals improve antibody responses. Additionally, differences in lipid nanoparticle formulations (SM-102 in mRNA-1273 versus ALC-0315 in BNT162b2) may also play a role [11]. A major advance of this study is comprehensive stratification by prior SARS-CoV-2 infection status. Most previous analyses excluded these patients, limiting interpretation to infection-naïve cohorts. In contrast, our analysis demonstrates that prior infection significantly enhances vaccine-induced immunity, with pooled seroconversion increasing from 0.49 (95% CI: 0.46–0.52) in infection-naïve recipients to a robust 0.90 (95% CI: 0.82–0.94) in those with documented prior COVID-19. This suggests that the immune priming achieved through prior infection induces memory B and T cells more effectively than vaccination alone in immunosuppressed populations [16].

The dose-stratified findings reinforce the reliability of additive immunologic gains. Previous syntheses reported increases from 9.5% after the first dose to 55.1% after the third [5,12]. Our results show similar stepwise increases: from 27% after one dose to 45% after two, 61% after three, and 84% after four doses. This progressive increase supports a model where repeated antigen exposure incrementally improves responsiveness, even under ongoing immunosuppression. These findings underscore the urgent need for repeated vaccination to achieve protective antibody levels, including the consideration of fifth or subsequent doses for non-responders [6,7,8,9,10,11,12,13]. We also identified that the choice of serological assay exerts a profound quantitative influence on pooled estimates. Seroconversion proportions fluctuated by nearly 27% depending on the target: anti-RBD assays were the most sensitive (54.6%), while requiring both anti-spike and anti-RBD positivity yielded the lowest proportion (27.1%). These discrepancies are likely driven by assay sensitivity and manufacturer-defined cutoffs optimized for the general population. This highlights the urgent need for standardized serological protocols to ensure that meta-analytic estimates accurately reflect the immunologic status of SOTRs. Considering the heterogeneity observed, we suggest not overinterpreting the pooled point estimates as these values represent aggregate measures amidst significant clinical and methodological variability.

Our findings stand in sharp contrast to the high-immunogenicity benchmark in immunocompetent adults, where pooled seroconversion rates reach 92.64% [142], with mRNA vaccines achieving near-universal (99.5%) neutralizing antibody generation [143]. Regarding safety, the theoretical concern that vaccination might trigger acute allograft rejection appears unsupported by current data. Reported cases of biopsy-proven rejection were rare and manageable with standard adjustments. This mirrors the safety profile of other essential immunizations, such as influenza and pneumococcal vaccines, where the risk-benefit ratio heavily favors vaccination. The risk of severe disease or death following natural infection in SOTRs is markedly higher than any potential risks associated with the vaccine.

The major strength of this meta-analysis lies in its comprehensive scope, encompassing 125 studies and providing the most robust quantitative evidence to date through novel stratifications by organ type, dose number (up to four), and, critically, prior SARS-CoV-2 infection status, which clarified the powerful effect of hybrid immunity. However, the study is subject to several limitations. Foremost among these is the pervasive, high level of heterogeneity across most subgroups, stemming from differences in immunosuppressive regimens, specific antibody assays and their defined thresholds, follow-up times, and patient comorbidities, which restricts the generalizability of the overall pooled estimates. Furthermore, most of the included articles were observational, limiting causal inference. Although most studies have evaluated mRNA vaccines, evidence on other vaccine platforms remains limited. Substantial heterogeneity was observed in the assessment of seroconversion rates, as different assays were used and cutoff values defining seroconversion or a positive response varied across studies, potentially introducing measurement bias. In addition, safety outcomes were insufficiently reported. The persistent heterogeneity across most subgroups suggests that standard clinical variables such as organ type or vaccine platform only modestly account for the variance. The reduction of I² to 0% in the prior-infection subgroup identifies hybrid immunity as a major primary driver of immunologic consistency. Other sources of quantitative bias include the lack of standardized international units; assays with lower analytical cutoffs tend to overestimate clinical protection, while those using higher 'protective' thresholds underestimate the actual rate of biological seroconversion. Consequently, most studies were unable to establish a clear association between vaccination and serious adverse events or mortality, limiting robust conclusions regarding vaccine safety.